vvEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Cincinnati, OH 45268
Office of Solid Waste and
Emergency Response
Washington, PA 20460
Superfund
EPA 540-2-90 010
September 1990
Second Forum on
Innovative Hazardous
Waste Treatment
Technologies: Domestic
and
Philadelphia, Pennsylvania
May 15-17, 1990
Technical Papers
-------
United States Office of Research and Office of Solid Waste an
Environmental Protection Development - Risk Reduction Emergency Response
Agency Engineering Laboratory September 1990
EPA/540/2-90/010 September 1990
&EPA Second Forum on Innovative
Hazardous Waste Treatment
Technologies: Domestic
and International
Philadelphia, Pennsylvania
May 15-17,1990
Technical Papers
U S. EfWhcmmenta^Prot^on Agency
Region 5, Library ^ ;] -•" ^ R .
77 West Jackson • ->. -.^ ^^ r.^,
Chicago, IL 60604-oo^U
-------
As a result of the high level of interest in innovative hazardous
waste control technologies, U.S. EPA's Office of Solid Waste and
Emergency Response (OSWER) and Risk Reduction Engineering
Laboratory (RREL) jointly conducted this conference. The
conference consisted of presentations of technical papers and
posters by international and domestic vendors of technologies for
the treatment of waste, sludge, and contaminated soils at
uncontrolled hazardous waste disposal sites.
The purpose of the 2 1/2-day conference was to help introduce
promising international technologies through technical paper and
poster displays, to showcase the results of the U.S. EPA Superfund
Innovative Technology Evaluation (SITE) program technologies, and
to present case studies of applied technologies from EPA's
Superfund contractors.
This compendium includes the papers that were made available by
authors and their institutions. The papers are published as
received and the quality may vary.
Although this document has been published by
the U.S. Environmental Protection Agency, it
does not necessarily reflect the views of the
Agency, and no official endorsement should be
inferred. Mention of trade names or
commercial products does not constitute
endorsement or recommendation for use.
-------
CONTENTS
Page
Remediation and Treatment of Superfund and RCRA Hazardous
Wastes by Freeze Crystallization
James A. Heist, Freeze Technologies Corporation 1-14
ARC/EPRI Process for Clean-up of Contaminated Soil
T. Ignasiak, Alberta Research Council 15-25
In-Situ Hot Air/Stream Extraction of Volatile Organic Compounds
Phillip N. LaMori, Toxic Treatments (USA) Inc 26-43
Extraction of PCB from Oil with EXTRAKSOL™
Diana Mourato, Environcorp Inc 44-50
DuPont/Oberlin Microfiltration System for Hazardous Wastewaters
Ernest Mayer, E.I. duPont de Nemours, Inc 51-68
Tyvek for Microfiltration Media
Hyun S. Lim, E.I. DuPont 69-74
Cleanup of Contaminated Soil by Ozone Treatment
Dr. E. WeBling, Chemisches Laboratorium 75-79
BioTrol® Soil Washing System
Steven B. Valine, BioTrol, Inc 80-90
Integrated Soil-Vapor/Groundwater Cleaning System at Selected
Sites in West Germany
Karlheinz Bohm, Ed. Zu'blin AG 91-115
Ultraviolet Radiation/Oxidation of Organic Contaminants in Ground,
Waste and Drinking Waters
Jerome T. Barich, Ultrox International 116-129
The Lurgi-Deconterra-Process Wet Mechanical Site Remediation
Eckart F. Hilmer 130-142
Developments and Operating Experience in Thermal Soilcleaning
H.J. van Hasselt, NBM Bodemsanering 143-162
Revolving Fluidized Bed Technology for the Treatment of Hazardous
Materials
Geoff W. Boraston, Superburn Systems Ltd 163-181
Thermal Gas Phase Reduction of Organic Hazardous Wastes in Aqueous
Matrices
Douglas J. Hallett, ELI Eco Technologies, Inc 182-191
-------
CONTENTS (Continued)
Page
X*TRAX™ - Transportable Thermal Separator for Organic Contaminated
Solids
Carl Swanstrom, Chemical Waste Management, Inc 192-204
Anaerobic Pyrolysis for Treatment of Organic Contaminants in Solids
Wastes and Sludges - The Aostra Taciuk Process System Thermal
Treatment
Robert M. Ritcey, UMATAC Industrial Processes 205-222
Algasorb®: A New Technology for Removal and Recovery of Metal Ions
from Groundwaters
Dennis W. Darnall, Bio-Recovery Systems, Inc 223-231
Biosorption - A Potential Mechanism for the Removal of Radio-
nuclides from Nuclear Effluent Streams
Peter Barratt, Biotreatment Limited 232-250
Biological Treatment of Wastewaters
Thomas J. Chresand, BioTrol, Inc 251-257
Biotechnical Soil Purification of Soil Polluted by Oil/Chemicals
Susanne Schi$tz Hansen, A-S Bioteknisk Jordrens 258-266
In Situ Physical and Biological Treatment of Volatile Organic
Contamination: A Case Study Through Closure
Richard Brown, Groundwater Technology Canada, Inc 267-306
Introduction to the GEODUR System
Svend Mortensen 307-315
The SITE Program: Heavy Metal Fixation in Soil
Philip N. Baldwin, Jr., Chemfix Environmental Services 316-323
Evaluation of Three Leading Technologies
Armand A. Balasco, Arthur D. Little 324-342
Soil Vapor Extraction and Treatment of VOCs at a Superfund Site in
Michigan
Joseph P. Danko, CH2M Hill ' 343-348
Sludge and Soil Treatability Studies at a Large, Complex Superfund
Site
Susan Roberts Shultz, Donohue & Associates, Inc 349-381
Conceptual Cost Evaluation of Volatile Organic Compound Treatment
by Advanced Oxidation
Glenn J. Mayer, CH2M Hill, Inc 382-409
-------
CONTENTS (Continued)
Recovery of Metals from Water Using Ion Exchange: A Case Study
Thomas A. Hickey, B&V Waste Science and Technology Corp 410-428
Critical Fluid Solvent Extraction
Cynthia Kaleri, U.S. Environmental Protection Agency 429-453
Bench-Scale Solvent Extraction
Joseph A. Sandrin, CH2M Hill 454-464
Field Demonstration of a Circulating Bed Combustor (CBC) Operated
by Ogden Environmental Services of San Diego, California
Nicholas Pangaro, Alliance Technologies Corporation 465-473
Low Temperature Thermal Treatment (LT3) of Soils Contaminated With
Aviation Fuel and Chlorinated Solvents
Roger K. Nielson, Weston Services, Inc 474-485
-------
Remediation
and
Treatment
of
Superfund and RCRA
Hazardous Wastes
by
Freeze
Crystallization
James A. Heist, President
Ken M. Hunt, Vice-President
Patrick J. Wrobel, Sr. Process Engineer NOVEMBER 1989
Robert W. Connelly, PhD, Sr. Process Engineer •w%*
FREEZE TECHNOLOGIES CORPORATION
Raleigh, North Carolina
-------
INTRODUCTION
Freeze crystallization is a separations process used to remove pure components
from solutions by crystallizing the materials to be removed. The technology has
proven advantages that make it economically attractive as well as uniquely able
to decontaminate many types of aqueous hazardous wastes. This paper will
illustrate how the process can be used in site remediation activities, by-product
recovery activities, and in the handling of mixed industrial wastes.
Freeze Technologies Corporation (FTC) has built a 10 gallons per minute (GPM)
DirCon™ plant to demonstrate the freeze crystallization technology. The plant is
contained in two modules that can be transported and requires less than 1 week
to set up. The company's current demonstration project is scheduled for the
Stringfeltow site in Riverside, California. Stringfellow is ranked high on the U.S.
EPA's National Priority List (NPL) of Superfund sites targeted for remediation.
The application is for a leachate from interception wells.
Freeze crystallization has several advantages over competing technologies for
remediation and waste recovery applications. First, it is an efficient volume
reduction process, producing a concentrate that has no additional chemicals
added to it. Volume reduction translates into reduced costs for wastes requiring
destruction by incineration or disposal in a landfill. When a large fraction of the
solvent (usually water) is removed from a waste, the remaining impurities often
begin to crystallize as well. These components are often sufficiently pure to have
by-product recovery value. Freeze crystallization has low processing costs
generally ranging from $.03/gal. to $.15/gal. for 40 and 5 GPM plants, respec-
tively.
-------
FREEZE PROCESS DESCRIPTION
The basic operation involved in freeze crystallization is the production of crystals
by removing heat from a solution. Crystals produced in this manner have very
high purities. Once small and uniform crystals have been produced, they must be
washed to remove adhering brine. The brine is recycled to the crystallizer, so that
as much solvent as desired can be recovered. The pure crystals are then melted
in a heat-pump cycle, which further improves the energy efficiency of the proc-
ess.
When one or more of the solutes exceeds its solubility, additional crystal forms
are produced, but they are formed separately from each other and from the
solvent crystals. Consequently is it easy to separate these different crystals by
gravity. This is because, in most waste applications, water is the solvent and with
ice being less dense and the solutes more dense, the separation can be accom-
plished by gravity.
A freeze crystallization process is composed of the following components, as
illustrated in the process flow diagram of Figure 1:
Figure 1:
Freeze Process
Flow Schematic
LfXfl to ™
^ VlTT—|—U^ REJECTION
HEAT f 1 CONDENSER
REJECTION I T
*• SLUDGE
OUTLET
ZH1
*CANTER M STRIPPER
O—Q^-^
CONCENTRATE
OUTLET
CRYSTALLIZER, where heat is removed to tower the temperature of the
material to the freezing temperature of the solution (usually crystallizing the
solvent first);
EUTECTIC SEPARATOR to segregate the crystals of solvent and solute
into different streams, so that each can be recoverd in pure forms;
-------
• CRYSTAL SEPARATOR/WASHERS to remove the crystals from the
mother liquor in which they are slurried, and to wash adhering brine to very
low levels so that the recovered crystals have high purity;
• HEAT-PUMP REFRIGERATION CYCLE to remove refrigerant vapor from
the crystallizer, and compress it so that it will condense and give up its
heat to melt the purified crystals;
• HEAT EXCHANGERS to recover heat from the cold effluent streams,
improving the efficiency of the process.
• DECANTERS and STRIPPERS are required in some processes to remove
volatile materials and/or refrigerant from the effluent streams before
discharge;
• UTILITIES, CONTROLS, ELECTRICAL SWITCH GEAR, PUMPS AND
PIPING are required to implement the freeze process in a continuous,
closed system.
The design, operating characteristics, capabilities and limitations are determined
largely by the type of crystallizer used, of which there are three basic types:
1) INDIRECT CONTACT crystallizers use a scraped surface or similar heat
exchanger that will crystallize by removing heat through the heat transfer
surface.
2) TRIPLE POINT crystallizers use the solvent as the refrigerant at its triple
point (where solid, liquid, and vapor phases are all in equilibrium). For
instance, with aqueous systems, the triple point occurs at less than 3 mm
Hg. absolute pressure at 30 degrees F, or below.
3) DIRECT CONTACT SECONDARY REFRIGERANT crystallizers use an
immiscible refrigerant that is injected directly into the process fluid, and
evaporates at several hundred to several thousand times the vapor
pressure of the solvent.
The direct contact process offers a number of advantages in treating hazardous
wastes. This process is more efficient, requiring less capital to build and less
energy to operate. Also, the direct contact process achieves a greater reduction
of waste volume and is capable of producing higher quality effluents. These are
the reasons FTC has chosen the direct contact process as the basis for its
technology.
-------
10 GPM DlrCon™ PLANT
The freeze crystallization process that has been accepted into the EPA's Super-
fund Innovative Technologies Evaluation (SITE) Program is a direct contact,
secondary refrigerant process. The mobile DirCon™ plant has been designed and
built for this program, and is illustrated in the photo shown in Figure 2. The
capacity of the plant is 10 GPM of ice produced from an aqueous waste stream
with a freezing point of 20° F.
The plant contains all of the components described earlier, including the crystal-
lizer, eutectic separator, crystal separator/washer, heat exchangers, heat pump
(open cycle screw compressor), decanters and strippers, and ancillary utility
related systems.
The plant is designed for transportability, using modular design concepts devel-
oped by Applied Engineering Co., Orangeburg, SC, the acknowledged leader in
this field. The plant is contained in two modules designed for transport on low-
boy trailers. Each measures approximately 50' long x 13' wide x 11.5' high. Upon
arrival at a site, a crane is used to stack one on top of the other. The three
electrical and thirty flanged-piping interconnections take about a day to complete.
The plant is self-contained except for electrical supply. Instrument air, cooling
water, electrical distribution, and electrical heating components are all contained
within the plant. The plant is controlled by a distributed digital processor that is
programmed to operate without attendance. On-line quality sensors are used to
evaluate process efficiencies, discern operating problems, and perform corrective
actions such as recycling effluent for re-processing if warranted.
Figure 2:
10 GPM DirCon™ PLANT
-------
DEMONSTRATION PROGRAM - STRINGFELLOW NPL
A demonstration of a commercial scale DirCon™ plant will be conducted at the
Stringfeltow NPL site at Riverside, California in December 1989. This demonstra-
tion is co-sponsored by the U.S. EPA and the state of California. The EPA
participation is through its Superfund Innovative Technology Evaluation (SITE)
program. The California participation is through its Waste Reduction Innovative
Technology Evaluation (WRITE) program sponsored by the Department of
Health Services, Alternative Technologies Office.
The goals of the SITE program are, among other things, to demonstrate applica-
bility of a technology on as general and as difficult a waste stream as can be
found. Procedures are designed to test for economic and technical performance.
Test results are analyzed to determine if the technology is applicable.
The FTC DirCon™ 10 GPM plant will be installed and ready to operate at the
Stringfeltow NPL site in Iate1989. The SITE tests and sampling program will
occur over a two to three week period and a public visitors day will be held while
the plant is at Stringfeltow. Further testing for reliability and longer term perform-
ance will continue for another several weeks before the plant is decontaminated
and removed from the site.
This demonstration program is designed to have minimal impact on the host site,
a condition that is paramount in EPA planning and with local communities.
Wastewater from the NPL site is piped from the leachate collection wells to the
on-site treatment plant. During the demonstration testing period, this flow will be
diverted to FTC's DirCon™ plant. The freeze plant will process this flow, produc-
ing separate streams of clean water and concentrate. These will be recombined
and pumped to the existing on-site treatment plant for routine processing. The
residence time in the DirCon™ plant, including storage before and after the actual
freeze processing, will be 72 hours.
Since the freeze process operates in closed vessels with recycle of refrigerants
and wastes at pressures less than atmospheric, there are no emmisions from the
process. An option exists at Stringfeltow for treating the concentrated effluent for
metals recovery, for organic by-product extraction, or for solidifying the concen-
trate for subsequent landfilling. The by-product recovery is an attractive option
because it would eliminate a large volume that is now landfilled.
-------
GENERIC APPLICATIONS FOR FREEZE CRYSTALLIZATION
Remediation does offer some opportunities for by-product recovery, but usually
these opportunities occur in on-going RCRA-waste generation facilities. There
are certain generic conditions that favor freeze crystallization. When used under
these generic conditions, the freeze crystallization process contributes signifi-
cantly towards accomplishing a more complete and more economically efficient
remediation through by-product recovery.
Freeze crystallization works by making pure crystals from water, or other compo-
nents, in a waste solution. The waste must first be in liquid form. In the case of
aqueous based wastes, the crystal produced is ice, and all impurities are ex-
cluded and remain in the concentrated liquid portion. The process is effective in
removing water from these wastes and reducing the volume of hazardous
contaminants.
Since crystallization excludes all impurities from the ice, all impurites are equally
removed. The freeze process is capable of treating wastes with heavy metals, all
types of dissolved organics, and radioactive materials. And all of the impurities
are reduced in concentration in the effluent by a factor of about 10,000.
Figure 3 shows conditions at which alternative treatment technologies are both
technically feasible and cost effective. The chart demonstrates that freezing
becomes more competitive as the waste becomes more concentrated and
complex. For instance, wastes with heavy metals require concentrations of 1,000
to 10,0000 mg/l to be economically recoverable with freezing. Aqueous streams
require organic concentrations of 3 to 7 wt-% before it is economicai to treat them
with freeze crystallization. When the waste contains both organics and heavy
metals at between .5 and 1.5 wt-% total contaminants, the freeze process
becomes more economical than multi-process treatment trains.
Figure 3:
Technology Comparison
Chart
I
N
C
R
E
A
S
E
D
C
O
N
C
E
N
T
R
A
T
I
O
N
Type of Contaminant
Volatile
Organics
Heavy
Organics
Salts
Heavy
Metals
Stripping
Ion Exchange
Carbon Adsorption
Biological & Chemical Oxidation
Electrodialysis
Reverse Osmosis
Evaporation
Freeze Crystallization
-------
CASE STUDY 1 - Stringfellow Leachates
Freeze Technologies performed treatability studies on Stringfellow leachates with
the feed composition shown in Table 1. These tests were performed on FTC's
batch process laboratory equipment. Test results confirm that freezing will
recover over 90% of the water, concentrating the impurities into a minimal
volume for disposal and/or recovery operations. It has been the experience of
FTC that a continuous process, which is used in commercial scale plants,
produces even better effluent qualities than those achieved in batch processing.
In the treatability tests, crystal clear melt was produced with a reduction in
conductivity from 39 mS/cm to less than .02, approaching a ratio of 2000:1. The
effluent (melt) qualities expected from a continuous processing plant at 90%
recovery of water from the leachate are also shown in Table 1.
Currently the leachate from Stringfellow is treated by a system consisting of
hydroxide precipitation of heavy metals and activated carbon adsorption of the
organics. The sludge is disposed of in a hazardous waste landfill, and the carbon
is returned for regeneration. Treatment costs for this system are $.363 per gallon.
A detailed listing of these costs is outlined in Table 2. Throughput at Stringfellow
has been slowly declining, with current production from the interception wells at
about 225,000 gallons per month. The annual operating costs to treat this volume
with the current system are $980,000, excluding equipment amortization.
By incorporating the freeze crystallization process into the existing treatment
system, costs for treatment can be substantially reduced. The costs to operate a
10 GPM DirCon™ plant are $.09 per gallon. (Costs as a function of DirCon™
plant size are summarized in Table 3.) The annual operating costs to treat the
Stringfellow leachate would be only $108,000 for the freeze process and
$300,000 to incinerate the concentrate. Total costs, including equipment amorti-
zation at $.05 per gallon, should not exceed $600,000 per year.
Constituent
PH
Conductivity, mS/cm
Total Organic Carbon
p-CBSA
Volatile Organics
Na
Total heavy metals
Al
Cd
Cr(+6)
Cu
Fe
Mg
S04
Cl
N03
Feed (Mg/l)
3.5
39
1400
2600
11
930
4100
2100
2.98
129
11
417
1180
20000
380
80
Melt (Mg/l)
6.5
0.02
<1.0
<1.0
<0.1
<1.0
<5.0
<1.0
<.01
<.01
<.01
<1.0
<1.0
<5.0
<1.0
<1.0
Table 1:
Stringfellow Leachate
Analyses
-------
Table 2:
Current Stringfellow
Treatment Costs*
'Costs shown do not include
equipment amortization
Cost Element
I/GAL
Chemicals
Activated Carbon
Supplies
Electricity
Labor
Effluent Post-treatment
Sludge Disposal
Total Cost
0.015
0.027
0.006
0.005
0.065
0.098
0.147
0.363
-------
CASE STUDY 2 - Mixed Industrial Wastes
Industrial facilities often have wastes that are collected from a variety of places
and combined for treatment at a common facility. In metal fabrication shops with
electroplating, metal cleaning, pickling, stamping, machining, and other opera-
tions, the combined waste might resemble the analysis shown in Table 4. A
typical treatment train might include an oil-water separator, neutralization,
hydroxide precipitation, and discharge to a POTW for organics removal. How-
ever, municipalities are looking much harder at toxic organic discharges, and
current treatment often fails to meet the effluent standards.
Freeze Technologies has performed laboratory testing on a waste such as this,
collected after an API separator. Over 90% of the water was converted to ice.
Alkalinity and suit ate salts were precipitated with heavy metals, hardness cations,
and probably to some degree with sodium ions. After about 75% water removal,
a second organic phase began to form, and a significant portion of the influent
organics partitioned into this phase that was composed of about 75% organics
and 25% water. The water phase contained less than 10% organics, but had
most of the dissolved salts.
The organic phase that forms this way will have a high heating value so it can be
incinerated directly, or perhaps recovered for its fuel value. The tests in the batch
treatability lab showed that this organic layer had a specific gravity much less
than ice, and it collected in the top of the ice drain column. This will allow its
recovery from a freeze crystallization process.
Cost Component
Amortization, 5 year SLD
Labor
Electricity
Supplies, Chem., etc.
Maintenance
Total Costs
Costs for Varying Plant Sizes ($/gal)
5GPM
0.080
0.040
0.008
0.010
0.007
0.145
10GPM
0.050
0.020
0.008
0.008
0.005
0.090
40GPM
0.015
0.005
0.006
0.005
0.004
0.035
Table 3:
Freeze Crystallization
Cost Summary
Constituent
Total Dissolved Solids
Dissolved Organics
PH
Alkalinity
Sulfates
Chlorides
Heavy Metals
Concentration (mg/l)
12500
8000
10
5000
4700
1500
3000
Table 4:
Typical Integrated
Manufacturing Effluent
10
-------
TREATABIUTY TESTING
Given the complex waste matrices encountered in most Superfund or RCRA
remedial actions, treatability studies should always be conducted. Wastewater
analysis is not sufficient to define the optimal treatment train for cleaning up liquid
waste streams especially when dealing with new technologies.
From our experience with 8 years of development of the freeze crystallization
process, and another 10 years of development of environmental process technol-
ogy, there are a number of guidelines we would propose for treatability testing in
EPA Superfund and other industrial waste programs.
1) Treatability studies should always be done by the technology developer.
2) Treatability studies should be incorporated into the RI/FS activities, and
replace the paper studies that are currently the basis for selecting be-
tween alternative treatment technologies. The RI/FS process is a way for
EPA to help the transfer of technology from the developers to EPA
contractors.
3) The EPA and its contractors should not attempt to define the treatment
technologies for individual sites. Instead, the EPA should conduct the pro-
curement activities in a way that elicits complete strategies from technol-
ogy developers. Most remediation sites have the need for a variety of
treatment technologies. The most efficient method for defining treatment
trains at individual sites is to first define a clean-up standard and let the
individual developers provide the train of technologies to meet those
standards.
The technology developer should have an institutionalized place in the remedia-
tion activities of Superfund and private clean-ups. The best process technology is
resident in this community, and this resource should be used for its expertise in a
way that ensures competition in procurement and full assessment of alternatives.
11
-------
THE FTC TREATABILITY TESTING LAB
Freeze Technologies performs treatability studies in the apparatus shown in the
photo of Figure 4. A simplified flow diagram for the unit is shown in Figure 5. The
unit operates in a batch mode. Feed is drawn under vacuum into the crystallizer,
and recirculation between there and the drain column is started with the slurry
pump. Refrigerant is pumped into the crystallizer, where it boils as it is heated by
contact with the waste. The refrigerant vapor passes through a refrigerated
condenser and is pumped back to the crystallizer as a liquid. When the drain
column is packed full of ice, the process is stopped. The concentrate is drained
from the ice and returns via gravity to the crystallizer for reuse. The ice is then
washed and melted. Since only 50% of the water in any batch can be converted
to ice, simulation of 90% recovery requires 4 stages of concentration. Twenty
gallons of sample is required to allow 12 of these staged tests.
These tests accomplish the following:
• interactions such as foaming or emulsification between refrigerant and
waste are observed,
• phase equilibria are confirmed, showing the operating temperatures that
will be required.
• the occurance of eutectic crystallization is observed.
• physical properties of the waste at the crystallizing conditions are deter-
mined.
The impact of this information is that more detailed preliminary process designs
are possible, which in turn allows more accurate cost projections. Problems that
would be seen in the field often show up at this stage of testing. Design adapta-
tions made in the batch lab are the first stage in modifying the full scale equip-
ment for use with new applications.
12
-------
Table 4:
Batch Treatability Test
Laboratory
Tables:
Batch Treatability Stand
Process Schematic
to vacuum
PRIMARY
COMPRESSOR CONDENSER
to
vacuum
REFRIGERANT
CONDENSER
AND RECEIVER
REFHK3ERANT
PUMP
drain
SLURRY
PUMP
13
-------
Freeze Technologies Corporation
2539-C Timbertake Road
Post Office Box 40968
Raleigh, NC 27629-0968
Office 919-850-0600
FAX 919-850-0602
14
-------
ARC/EPRI Process for Clean-up of
Contaminated Soil
T. Ignasiak, K. Szymocha, B. Ignasiak
Alberta Research Council
C. Kulik and H. Lebowitz
Electric Power Research Institute
1. Background
Soils contaminated with oily waste presents a serious environmental problem.
The management of this depends upon many factors such as contaminant
characteristics, topography, climate, ecology and soil characteristics. The
complex and diverse nature of oily wastes also causes severe technical and
economic limitations.
The Alberta Research Council (ARC) and the Electric Power Research Institute
(EPRI) have responded to this problem by developing a novel process for clean-up
of oily waste soils. The potential of this process has been demonstrated
through an extensive batch experimental program followed by verification in
6T/day pilot plant tests. A wide variety of oil contaminated soils has been
tested in the program with particular attention directed towards remediation of
soils from manufactured gas plant sites. Except for coal derived tars
(polycyclic aromatics hydrocarbons - PAH), such soils contain considerable
quantities of cokes, chars and coals as well as slags (molten mineral matter
with inclusions of organic solvent soluble and/or insoluble hydrocarbons).
The work carried out at ARC led to the development of a detailed conceptual
design for a transportable 200T/day demonstration plant for clean-up of oily
waste soils. The unit is engineered to handle a wide range of oily waste.
Process economics look very encouraging.
15
-------
2. Process Description
The process flow diagram for the ARC/EPRI process for clean-up of tar/oil
contaminated soil is presented in Fig. 1. A suspension of contaminated soil and
coal in water is subjected to tumbling in a specially designed drum at an
elevated temperature at optimized solids concentration. The mixture is
subsequently screened into two fractions: -1mm and +lmm. The -1mm fraction is
subjected to agloflotation which separates the coal floes (microagglomerates) in
the form of "froth 1". Depending on the characteristics of the treated soil,
the "middlings 1" are either reprocessed or combined with "froth 1". The
tailings from agloflotation are combined with a -1mm reject derived from
selective grinding of +lmm fraction. The combined material is subjected to
reprocessing in the presence of small quantities of coal and a suitable
collector and again is subjected to agloflotation, thus giving rise to "froth 2"
which, together with "froth 1", forms a combustible product. The tailings from
reprocessing yield clean soil. Depending on the characteristics of the treated
soil, the "middlings 2" are combined either with the combustible product or
clean soil.
3. Development Status
The process scheme described above has been developed based on extensive
batch and continuous 6T/day pilot scale tests carried out at the ARC. The flow
diagram of the Continuous Soil Clean-up Pilot Plant as designed, built and
commissioned at the ARC Devon Facility two years ago is presented in Fig. 2.
Based on results of those tests, a conceptual design of a 200T/day commercial
demonstration plant was developed in 1989 which has been used for detail process
cost evaluation.
The 200T/day demo plant is designed to be transportable. The unit is
engineered for handling the most difficult soils (for example, manufactured gas
plant sites).
ARC and EPRI intend to form a Consortium of Companies in the second half of
1990. The objective will be to demonstrate the ARC/EPRI technology for clean-up
16
-------
Figure 1:
Scheme of Processing Tar/Oil
Contaminated Soil
Contaminated soil/coal/water
Tumbling
Screening
+ 1.0 mm
Grinding
I
-1.0 mm reject
Clean coarse
reject
Collector
Reprocessing
-1.0mm
Collector
Agloflotation
Tailings
Middlings 1 Froth 1
Middlings 2 Froth 2
Clean soil
^Combustible product
17
-------
Figure 2:
(I
00
High shear
mixer 0
Flotation cells
oooooo
Low shear
mixers
Coarse soil
product
Fine soil
product
e
Dewatering screen
=h
Middlings Coal product
Flow Diagram of the 6 T/day Continuous Soil Clean-up Process
at ARC Devon Facility
-------
of contaminated soil during the next 2.5-3 years at a cost of about $7 MM (US).
This effort will involve detail engineering for the 200T/day unit, construction
of the unit, on-site commissioning and partial clean-up of two to three selected
sites with different soil characteristics.
4. Media and Pollutants Treated
The experimental work carried out at ARC, emphasized processing soils from
manufactured gas plant (MGP) sites. The experience generated at ARC indicates
that MGP soils are very difficult to clean. In total, seventeen samples of MGP
soils were tested. These samples originated from various contaminated sites in
the eastern and the northern parts of the United States.
The tar refuse (MGP) material is composed of tar, solids and water. Tar has
been defined as any organic matter extractable with organic solvents such as
benzene, toluene, xylene or dichloromethane. The solids represent a mixture of
mineral matter which contain non-extractable organic carbon (chars, coal, coke)
referred to as combustibles and slag. The solids are characterized by a wide
range of particle size distribution (from a few microns to a very coarse
material) and often contain various types and quantities of clays.
The composition of tar refuse sample (concentration of water, solids and
tar) was determined by Dean-Stark Soxhlet extraction with toluene which yielded
the water and tar concentrations. After drying and weighing, the solids were
ashed to give the contents of combustibles. The tar refuse samples that were
subjected to clean-up tests contained from 1-102 tar, 38-922 mineral matter,
2-222 combustibles and 4-542 water.
The tars extracted from tar refuse samples were subjected to detailed
analyses which indicate that the tars are highly aromatic and are characterized
by negligible contents of heteroatoms. About 60-80% of the tars were volatile
at temperatures up to 500"C and the number average molecular weight of the
volatile fraction was below 300. The volatile portion of the tar was subjected
19
-------
to detailed GC-MS analyses. The major species identified were various isomers
of mono-, di- and thri-methylated naphthalene, anthracene, phenanthrene and
pyrene and other polycyclic aromatic hydrocarbons.
In addition to MGP samples, a variety of oil spill samples (soils
contaminated with heavy oils, light oils, gasoline, diesel oil, as well as fuel
oils and intermediate API gravity oils) were tested for cleanability. All of
the oil spill samples tested so far could be cleaned much more readily compared
to the MGP samples. Among the oil spill samples, most of the work was done on
heavy oil contaminated soil samples. Compared to the tar, the heavy oils
contained negligible amounts of low boiling hydrocarbons and only traces of PAH.
The ARC/EPRI clean-up process shows promise for cleaning the soil
contaminated with a variety of organic compounds like creosote oils,
polychlorinated aromatic hydrocarbons, asphalt, etc.; however, the effectiveness
of this process for these particular compounds has not yet been determined.
Initial and Final Pollutant Concentrations
Essentially there is no upper concentration limit on coal and/or petroleum
derived pollutants present in soil in terms of the clean-up efficiency of the
ARC/EPRI process. A soil sample containing 50% (by weight) of a contaminant can
be cleaned as efficiently as a sample which contains only 0.5-5% of contaminant
concentration. Since, however, the clean-up requires the use of coal as an
adsorbent in quantities 2-4 times greater as compared to contaminant
concentration, it appears to be economically advantageous to treat samples
characterized by rather low or intermediate (up to 10-12% by weight) contaminant
concentrations. Low concentration of contaminants (0.5-5%) offers particular
opportunities due to limited application of pyrolytic and combustion techniques
for treatment of such samples.
Table 1 presents the results of clean-up of soils contaminated with
tarry/oily wastes. The concentration of "contaminant" in the "as received" soil
and in cleaned soil is based on the exhaustive extraction of the feed and clean
product with an organic solvent, as described above.
20
-------
Table 1
CLEAN-UP OF OILY WASTES -
TEST RESULTS
Sample
#
1
2
5
6
9
11
12
Type
Contaminant
MGP Tar #1
MGP Tar #2
MGP Tar #1-1
MGP Tar #2-2
Gasoline Spills
Heavy Oil Spills I
Heavy Oil Spills II
"Contam." cone., wU
Feedstock
8.6
1.2
5.6
10.6
3.1
43.0
0.2
Processed
Soil
0.07a
°-°°a h
0.05?'b
0.17b'c
0.00^
0.08C
0.00C
a) grinding and reprocessing
b) froth collector required
c) results not optimized
Samples contaminated with coal derived tars (MGP sites) were much more
difficult to clean. To bring the concentration of solvent extractable material
in these samples below 0.1% (1000 ppm), optimization of the treatment for the
individual samples is required. Optimization, grinding and reprocessing was
required for samples contaminated with petroleum derived products. One of them
(sample 11) could be readily cleaned to 800 ppm of total organic material. The
other two were cleaned below the detectibility level (100 ppm) of the
gravimetric analytical procedure used.
The concentration of PAH in cleaned soil samples (Table 1) originating from
MGP sites varied from about a few ppm to about 200 ppm. The concentration of
PAH's in clean soil samples from oil spills (Table 1) was below the detectivity
level.
None of the clean samples (MGP sites or oil spills) failed the EPA
Teachability test (TCLP). The TCLP tests as well as PAH's concentration tests
were carried out by Radian Corporation, Austin, Texas. Some verification of the
analyses were performed by the Laboratories of the University of Alberta,
Department of Chemistry, Edmonton, Canada.
21
-------
Process Limitations and Advantages
The ARC/EPRI clean-up process is based on utilizing the adsorption capacity
of coal with respect to the oily wastes. For oils and tars, this capacity is
high and the selectivity of the process is very good. This may not be the case
for other organics particularly those characterized by hydrophilic character.
Since the process requires that coal (or another solid carbonaceous with
adsorbent characteristics material) be available on a processing site,
processing soils contaminated with high concentration of oily contaminants makes
the processing economics less favorable. However, sometimes the contaminated
soil also contains carbonaceous materials (MGP sites and some petrochemical
plants) in addition to tars/oils. The presence of the carbonaceous material in
contaminated soil offers the opportunity of using this material as the adsorbent
for the ARC/EPRI process.
Intermediate scale combustion tests (samples size ~5T) carried out by
combustion engineers with the tar enriched coal has not revealed any problems;
most importantly no increase in PAH's presence in combustion gases was
identified. Certain contaminants (for example polychlorinated hydrocarbons)
adsorbed on coal would have to be combusted at a PCB disposal facility.
However, it would make more sense to adsorb the polychlorinated aromatics
contaminating IT of soil (at 1% concentration) and burn the 50 kg of
contaminated coal material (40 kg coal plus 10 kg contaminants) in a suitable
facility for PCB disposal rather than using IT of contaminated soil plus
additional fuel required for combustion.
An interesting approach to the utilization of the ARC/EPRI clean-up
technology is for the Alberta and Saskatchewan heavy oil/bitumen industry.
Economics of these two provinces rely to some extent on their enormously rich,
but low quality, oil/bitumen deposits. Currently the recovery of oil/bitumen is
being achieved by either mining and hot water separation or in-situ steam
flooding, using natural gas for steam/power generation. It is apparent that
these methods present a serious environmental problem reflected in the vast
accumulation of tailing ponds and oil spills. The adaptation of ARC/EPRI
technology for clean-up of contaminated ponds and spills using low cost, low
22
-------
quality coal and the utilization of this oil laden coal, instead of expensive
natural gas for steam generation, offers a very interesting option.
Furthermore, the combustible product (oil adsorbed on coal) generated in the
clean-up process can be thermally treated, releasing light oil, which can be
used as diluent for pipelining heavy oil. The thermally treated coal product
can then be combusted to generate the steam needed for oil/bitumen recovery.
Process Waste Streams
The ARC/EPRI process generates three streams: clean coarse reject, clean
soil and combustible product (contaminants adsorbed on coal). Subject to
state/provincial regulations and laws regarding the cleanliness of the product,
the clean coarse reject and clean soil can be used to generate various products,
for example, asphalt road, aggregate material or land filled either directly or
after some additional treatment (ozone treatment, bio-treatment). In most
cases, the oil/tar enriched coal can be combusted.
When the ARC/EPRI process is required to work at elevated temperatures (due
to high viscosity of the contaminant), the process leads to the production of
some quantity of low boiling hydrocarbons (BTX) which will have to be condensed
and the residual vapors will have to be quantitatively adsorbed on activated
carbon. The condensate could be readily combusted or sold to potential users.
The hydrocarbon-saturated activated carbon would have to be reactivated or
combusted.
The ARC/EPRI clean-up process is in most cases a net consumer of water;
about 15T water would be required for 200T/Day facility. Therefore, no problems
with discharging bulk quantities of process water would be encountered.
Economics of ARC/EPRI Clean-up Process
Based on the experience developed in the course of batch and 6T/D pilot
plant investigations carried out during the last three years on the ARC/EPRI
clean-up process, EPRI has undertaken conceptual engineering and economic
23
-------
studies for the demonstration clean-up plant. This work was completed in
November 1989. Follow-up work on detail engineering studies for the
demonstration clean-up plant have commenced in January 1990 and should be
completed in July/August 1990.
Since the soil clean-up demonstration plant would have to be moved from one
contaminated site to another, the plant with initial design capacity of
200T/day, would be built on trailer mounted modules. The plant would be capable
of treating the most intractable oil/tar contaminated materials (-4in) with
contents of organics ranging from 1-20%. The target for soil clean-up was set
at 0.1% by weight, or less of residual organic material extractable from the
processed soil with toluene as described above. The engineering design of the
plant is based on the principles of batch clean-up presented in Fig. 1 and
verified in a 6T/day pilot plant presented in Fig. 2.
The process of adsorption of organics on coal, combined with selective
grinding (limited to coke, char, slag, etc. but excluding rocks and mineral
matter), high shear agitation and agloflotation phenomena, are process
operations which are utilized in cleaning up the MGP soils.
In the cost analysis, it was assumed that the cost of coal will be equal to
revenue from agglomerates (coal + tar + other carbonaceous material recovered
from the feed) and the cost of oil (heavy oil) used in the process was
conservatively estimated at US $20/Bbl.
The capital cost for the 200T/day demonstration plant including site
installation cost was estimated at $2.7 MM (US), while the operating cost
(including electricity, collector and oil costs, labor, maintenance, supplies
and services, equipment rental and contingency) arrived at was US $35/T,
bringing the soil clean-up cost both capital and operating to about US $40/T of
soil processed. This cost includes only the cost of processing the soil and
does not include, for example, soil excavation and site restoration.
The case presented above is for the most intractable oil/tar contaminated
material requiring significantly more extensive processing compared to, for
instance, heavy oil spills.
24
-------
sanpisaj ^o [esodsip
pup uoi}djospp
poog A"j3^ mrt QOT> -puij6 aAi;o3[3S Jq/} 01 IHd3/3yV
===============================================================================
+021-06$ POog/jtPj lurr QOI> }ua6paa d3H3 UO ^/^ 02'OT d3^3 UO
+091$ P°°9 ^ 02> A"ouanbajj MOT jq/^ 02-91
08$ poog A"JSA uni Q9> uonB^°U q}°Jd jq/^ L2
OOI~08$ P°°9 i"* 89> sidujis A"JSA jq/^ 01
OIt~0^$ P0°9 u"* Ł9> J3}BM aAisua^xg jq/; 02 ZMH
jad
A"oug6v
A"q passasse pus adojn^ ui pgdo[9A9Q
S3I9010NH331 9NIHSVM 1IOS dO
2
SutqseM [ios ueadojn^ aq; o; pajeduioo
6uisiujojd A"jaA 55(001 A"6oiouqoa^ iyd3/3HV 3M^ ^5M^ sjeadde
iyd3/3JJV 3M} M^^ aiqpjpdujoo SL [ios jo auuo^ jad
" passassp s;ueid aq^ ^o
•oiqo 'i;puuPun 'A"ou96v uoip^^o^d [P^uguuuojtAU] sn '^uaiudoiaAdQ pup
^sanbaj aq; uo s^^asnqopsspw 'pjo^pag 'uon^jodjoo sai6oiouqoai
8861 ui suoneaiiddv pun;jadn$ jo; passassp pup adojng
ui padoiaAap sai6o[ouqoa^ 6uiqspM [tos >o A"jpujujns
6uissaoojd 9q; 6uiJ3MOi UL ;[n$9J 'pggpui 'ueo
3SOL3 ;e a^qpiiPAp ajp ^JPUJ [POO ps^puiujp^uoo 9L|; pup
3J3qM suoi^Pu°3 Japun LIOS ps^puiuup^uoo [io A"Apaq
;no paijjpo s9}eun}S9 ;soo pup u6is3p 6uiJ3aui6ua
-------
IN-SITU HOT AIR/STEAM EXTRACTION
OF VOLATILE ORGANIC COMPOUNDS
Phillip N. La Mori
Senior Vice President and Technical Director
TOXIC TREATMENTS (USA) Inc.
151 Union Street, Suite 150
San Francisco, California 94111
26
-------
SUMMARY
This paper presents results on using a new technology for the in-
situ removal of volatile organic compounds (VOCs) from soil. The
technology injects hot air/steam into soil through two 5-ft-
diameter counter rotating blades that are attached to drill stems.
The soil is cleaned as the drill stems are advanced into and
retracted from the ground. The evolved gases are trapped in a
surface covering called a shroud and contaminants are captured
aboveground via condensation and carbon absorption. The recovered
contaminants can be reprocessed or destroyed.
The results of the initial use of the commercial prototype show 85
to 99 percent removals of chlorinated VOCs from clay soils. The
technology was able to achieve its goal of less than 100 ppm over
80 percent of the time in three blind tests. The mass of material
recovered in the condensate was conservative (89 percent) of the
mass determined removed from the soil by chemical analysis and
physical measurements. The technology also removed significant
quantities (50 percent) of semi-volatile hydrocarbons (as
quantified by EPA 8270). This was unexpected. Subsequent analysis
has identified potential mechanisms for removal of semi-volatile
components. Further testing is planned to evaluate these
mechanisms.
The process has not been found to cause undesirable environmental
emissions as a result of its operation. Noise and air emissions
during operation are below the limits set by regional environmental
regulations in southern California. Soil hydrocarbon emissions
during treatment are not increased from background before
treatment. Soil adjacent to the treatment has not been found to
have increased VOCs after the process is operated. Environmental
emissions caused by operational problems can generally be cured
without shutdown.
27
-------
INTRODUCTION
This paper describes in detail the initial results on evaluating
and proving a new technology to remove organic compounds from soil
without excavation. The process removes volatile organic compounds
(VOC) from soil by hydrocarbons. The technology is described in
sufficient detail to understand its use to remove chlorinated VOCs
plus some nonchlorinated hydrocarbons from a former tank storage
facility in California. The unit is call "The Detoxifier11 and is
operated by Toxic Treatments (USA) Inc., (TTUSA).
The Detoxifier is a mobile treatment unit used in the in-situ
remediation of contaminated soils and waste deposits. The soil is
treated in place and is not excavated or removed to the surface.
The Detoxifier is capable of a wide range of site remediation
methods, including;
Steam-air stripping of volatile contaminants
Solidification/stabilization and construction of
containment structures by addition of chemicals or
physical agents (e.g., pozzolanic materials)
Neutralization or pH adjustment by addition of acids or
bases and oxidizing or reducing chemicals
Addition of nutrients, micro-organisms, and oxygen to
promote in-situ biodegradation.
DESCRIPTION OF TECHNOLOGY
The Detoxifier consists of a process tower, a control unit, and a
chemical process treatment train (Figure 1). These components are
configured to meet site-specific requirements. The process tower
is essentially a drilling and remediation agent dispensing system,
capable of penetrating the soil medium to depths of 30 feet, as
currently designed. Remediation agents (in dry, liquid, vapor, or
slurry form) are added to and mixed with the soil at various depths
by the drill head assembly. A box-shaped shroud, under vacuum,
covers environmental release. The drill assembly is composed of
two drill blades, each 5 feet in diameter, with injection
dispensers (Figure 2).
28
-------
ro
vo
p
R
0
C
E
S
T
R
A
N
Compressor
Boiler
Condensation
&
Separation
Carbon Beds
Distillation
&
Storage
V
A
P
0
R
OFFGAS
L
I
Q
U
I
D
CO
D
R
I
L
L
T
0
W
E
R
SHROUD
-------
OPERATING METHOD
The remediation of a large area is carried out by a block by block
treatment pattern. The area to be remediated is divided into rows
of blocks, with the process tower, control unit, and process
treatment train being moved from one block to the next after the
remediation of a block is completed. To assure complete coverage
of the area to be remediated, the drill assembly is positioned with
a 20 percent overlap of the previously treated block. The
effective surface area of a treatment block is approximately 30
foot2. The volume of each block is determined by the depth of
remediation (Figure 3).
INSTRUMENTATION AND CONTROL
On-line analytical instruments continuously monitor and record the
treatment conditions. The monitoring data are used to control the
treatment process and determine the achievement of remediation
objectives.
In applications involving steam/air stripping of volatile
contaminants, the off gas containing the contaminants is captured
in the shroud and sent in a closed loop (to prevent any
environmental release) to a trailer-mounted chemical process train
for removal of water and chemical contaminants. The clean air is
then recycled to the soil treatment zone.
The system control consists of process monitoring and control
instrumentation. Flame ionization detectors (FIDs) monitor the
concentration of total hydrocarbons at selected process locations,
including the process off gas from the shroud and the purified
return air. The level of contaminants in the off gas determines
the process control (Figure 4). A gas chromatograph (GC) provides
a periodic check of the identification and concentration of
specific compounds in the off gas stream and can be used to confirm
process control. The output of the FID, temperature sensors, depth
gauge, and other instrumentation is stored in a computerized data
logging system, displayed on a terminal, and recorded on a strip
chart recorder. This enables operator control of process
parameters to achieve the most effective treatment in the least
amount of time.
30
-------
TYP NET TREATMENT AREA
> n i • bn>* Mr r»t i
TREATMENT AREA
PER TREATMENT BLOCK
(
1
V
\
y i
7
1
\
\
„-• • . */
1
\
\
y ^
7\'
/ \
SHROUD | 1
LIMITS v^^/
"7
1
\
\
'---"' ' • X
/
1
\
\
V
TYP
•\
4
t
\
1
^* S*
/
1
\
\
\
v -
7
1
\
\
,~
/
tm
f
<
/
1 1
/ \
/ \
(ELLY BAR SPACING
X
TYPICAL /
NET TREAT- /
.MENT AREA | '
\ /; \
V"" '"'^"'"
\ CUTTER \
| BIT AREA |
\
! ^
ROW
^A-V^ .-
y^ira-T
iy | IX?) i
\
\
I 'fv>' \ '
\ / ^ KELLY BAR
x^ ^yf ^ ''
-/~\-
\ / v
1 I i
/ \ .
/ \ y
\
X
^
\
\
ROV CENTERLINE
/ 1 '
/ ' TYPICA
\" 1 OF OFf
\ ADJACl
CENTERLINE J *°VS
/
ROV CENTERLINC ,
1
A.
\TION
•SET
•Mt
TYPICAL
SEPARATION
1' OF PARALLEL
, CONSECUTIVE
ROWS
N\ DETDXIFIER TRE/
/' BLOCK PATTERN
-------
TOTAL LAB VOC vs. FID READING
AREA A POST TREATMENT DATA
co
ro
c
c
Q.
a
o
o
>
m
<
_i
i
<
i—
o
i
i
zuu
1 Q n
I ou
1 c n
I OU
1 A n
I 4U
1 i n -
1 ZU
1 nn
I UU
on
ou
c n
ou
A n
4U
o n
ZU
n
C
-f. . . . . . . .
...4- %^.
V.r»V
y« X%A. E?
C^&W "^^"^
^5 ^>^+
' '/*« ' '^*^
1<**S^ +
_/T , .J. i
>^ ^ -h
lj^ i _j_
>^ + T ^ ...
/ ^1 +
.Xh ... ^. ...... ^_ . + . . +
_i_ j. -+?" y& ^
-4- i*rb_ ~r~ -4-~ ^^ >t: -4-
,• T ^1- _|_ JL^ T- ^ -r
^^^ i ~~pn n
1 500 1000 1500 2000 2500 3000 35(
FID READING (ppm)
+ POST 8 WEEK DATA * PRE 8 WEEK DATA
-------
IN-SITU REMOVAL OF VOLATILE
ORGANIC COMPOUNDS
The Detoxifier presently is remediating a site in San Pedro,
California, contaminated mostly with chlorinated hydrocarbons. The
activities at the site were undertaken at the request of a client
and are under the jurisdiction of the California Department of
Health Services (CDHS), Toxic Substances Control Division.
The remedial design required that a Baseline Calibration and
Testing Program be conducted to develop operational procedures and
provide the data necessary to evaluate the effectiveness of the
Detoxifier steam/air stripping process to remove VOCs from the soil
at the San Pedro site.
The Baseline Calibration and Testing Program involved three phases
of effort. The first two phases determined the operating
parameters for process control over a wide range of site conditions
and was also used as a run-in for the apparatus. The third phase
was a blind test of the technology at three of the most
contaminated portions of the site. Following the Baseline Test,
additional testing was performed as part of the EPA SITE Program in
two distinct phases. These test efforts and their results are
described below.
BASELINE TESTING, 10 BLOCK
Purposes:
To evaluate the effectiveness of the Detoxifier in
removing hydrocarbons from the soil at the San Pedro site
To measure soil vapor emissions before and during
remediation
Treatment Protocol. The Baseline Test was conducted in three areas
of the site (A, B, and D) previously determined to be some of the
most highly contaminated. The treatment depth was 5 feet, even
through the contamination was known to extend below that. To avoid
potential contamination from soil below, the treatment protocol
called for treating to six feet to develop a one foot buffer zone
between the treatment depth and the remaining contaminated soil.
This was accomplished by treating the five to six foot zone first
and then treating the 0' to 5' soil according to the operating
parameters in Figure 4.
33
-------
Pre-Treatroent Sampling and Analysis. Soil samples were collected
from 24 borings located in the 10 treatment blocks and in 4
uncontaminated locations west of the treatment areas (for
background data). All drilling, sampling, and chemical analysis
were conducted according to legally defensible QA/QC procedures
approved by CDHS. (Sampling and Chemical analysis protocols are
available upon request.) A total of 91 soil samples were taken,
including:
68 samples collected from the 20 borings distributed
among 10 treatment blocks located in Treatment Areas A,
B, and D
12 samples collected from the 4 borings located to the
west of the current treatment areas (as a control)
11 duplicate samples were selected from the 20 borings
The 68 samples collected from treatment blocks in Areas A, B, and
D, and the 12 samples collected from the 4 borings located to the
west of the current treatment areas were analyzed for VOCs by EPA
Method 8240. Eight of the duplicate samples were also analyzed for
VOCs by EPA Method 8240; these duplicates were labeled in a manner
such that the laboratory could not distinguish them from the
previously analyzed samples. At the request of the CDHS, the
remaining three duplicate samples were analyzed for semi-volatile
hydrocarbons (SVHs) by EPA Method 8270 and for priority pollutant
metals by EPA Method 6020.
Following the above-described analyses, additional chemical
analyses were performed on soil samples that had been kept in
refrigerated storage. At the request of the CDHS, 65 soil samples
were analyzed for SVHs by EPA Method 8270. These 65 samples were
the remainder from the 68 samples collected from the 10 treatment
blocks located in Treatment Areas A, B, and D.
The composite average pre-treatment concentrations of VOCs and SVHs
varied among the treatment areas. Treatment blocks in Area A had
the greatest number of species or groups of VOCs and SVHs detected,
and treatment blocks in Area D had the least. The average
concentration of VOCs and SVHs was also dissimilar. Area D
exhibited the highest concentration of VOCs and Area A the lowest;
Area B had the highest concentration SVHs and Area D the lowest.
Analyses for physical properties were also performed on soil
samples collected from the 10 treated blocks sampled in Treatment
Areas A, B, and D. A total of 68 soil samples were analyzed to
evaluate the density and moisture content of the sample.
34
-------
Treatment. The remediation protocol was designed to achieve the
reduction of VOCs to less than 100 ppm. Some of the key data
recorded during treatment included:
Duration of treatment
Steam injected (total pounds)
Hot air injected (cubic feet per minute)
Air temperature
Concentration of volatile organics in the off gas from
the shroud to the process system
Depth of treatment
TTUSA did not have knowledge of the pre-treatment soil analyses of
the 10 blocks until completion of baseline testing. The decision
to terminate treatment of each test block was made when the
concentration of VOCs measured by the FID decreased to a value
previously determined as equating with less than 100 ppm VOCs in
the soil. This concentration level had been determined during
baseline calibration and extended baseline calibration. TTUSA did
not have the benefit of any other information to determine when to
cease treatment.
post-Treatment Sampling and Analysis. The same procedures were
used for post-treatment sampling and analysis. A total of 55 soil
samples tubes collected from 14 borings distributed among the 10
treatment blocks located in Treatment Area A, B, and D were
analyzed for VOCs and SVHs by EPA Methods 8240 and 8270,
respectively.
Results. Based on a comparison of the pre and post-treatment
chemical analyses, the major effects of treatment were:
A very substantial reduction in the concentration of VOCs
A significant and unexpected reduction in the
concentration of SVHs
Reduction of VOCs. A very substantial reduction in the
concentration of VOCs was calculated from chemical analysis in all
test blocks (see Table 1 for summary data). This conclusion is
reinforced by the fact that the mass of VOCs collected in the
chemical process train was roughly equivalent to the calculated
mass removed from he soil (see "Mass Balance" below). The major
mechanism that caused the reduction appears to be volatilization.
35
-------
The treatment techniques used were designed to reduce the
concentration of VOCs to less than 100 ppro. This level was
achieved in 8 to 10 treatment blocks remediated (see Table 1 for
summary data). As indicated in Table 1, all treatment blocks in
Areas A and B were remediated to substantially below this level,
but only 1 of 3 treatment blocks in Area D was similarly
remediated. The average pre and post-treatment concentrations
(ppm) for VOCs are:
Treatment Area Pre-Treatment Post-Treatment
A 1,114 12
B 1,353 30
D 3,954 139
The reduced effectiveness of the process in achieving a level of
100 ppm in Area D is believed to be caused by the high initial
concentration of tetrachloroethylene and the presence of more
clayey soil than encountered in Areas A and B.
Reduction of SVHs. An unexpected and major reduction in the
concentration of SVHs was indicated in all treatment blocks
remediated (see Table 1). The average pre and post-treatment
concentrations (ppm) for SVHs are:
Treatment Area Pre-Treatment Post-Treatment
A 3,775 627
B 12,116 1,766
D 1,014 85
Mechanisms that may account for this reduction include
volatilization, steam distillation, hydrolysis, and oxidation.
These reductions have been confirmed by laboratory testing. A
small quantity of SVHs (2.1 Ib) was collected in the process train;
however, the mass of SVHs collected does not account reasonably for
the mass reduction (1,018 Ib) indicated by post-treatment soil
sample analysis (see "Mass Comparison" below).
Soil Gas Emissions During and After Treatment. One concern
expressed by the CDHS was the potential for increased emissions of
VOCs into the ambient air during and after treatment by the
Detoxifier. This rould occur, for example, if the heat used to
volatilize hydrocarbons caused increased soil emissions adjacent to
the remediation or emissions occurred from just treated soil after
removal of the shroud. An evaluation of this was made during
36
-------
TABLE 1
SUMMARY OF
CHEMICAL ANALYSES
Treatment Concentration (ppm)
Block Pre-Treatment Post-Treatment
Volatile Organic Compounds:
A-8-g 1,149 18
A-9-g 824 7
A-10-g 1,368 11
B-50-n 1 , 1 23 23
B-51-n 1,500 13
B-51-m 1,872 55
B-52-m 917 29
D-92-b 2,305 53
D-93-b 3,720 163
D-94-b 5,383 203
Semi-Volatile Hydrocarbons:
A-8-g 1 ,794 637
A-9-g 2,510 653
A-10-g 7,020 592
B-50-n 22,829 1 ,670
B-51-n 14,924 2,304
B-51-m 10,040 2,495
B-52-m 669 594
D-92-b 707 55
D-93-b 1 ,465 90
D-94-b 869 111
o/
^
Reduction
98
99
99
98
99
97
97
98
96
97
64
74
92
93
85
75
11
92
94
87
37
-------
baseline testing. The test used was to be qualitative in nature as
a quantitative test would be expensive and time consuming/ per
agreement with the CDHS. A background value for soil vapor
emissions was established for each area prior to treatment, and
emissions were measured around the perimeter of the shroud during
baseline testing. A comparison of these data and the data obtained
during baseline testing indicated that the overall effect of the
hot air/steam stripping process was to increase slightly the rate
of soil gas emissions within and directly adjacent to the treated
area. Emissions from the just treated uncovered blocks were always
higher than background for several hours afterward.
A more recent test (under the approval of the CDHS) using improved
process vacuum and soil covering of remediated blocks showed no
increased emissions to the ambient air. In fact, the levels
remained below background during and after treatment.
We believe this means that soil gas emissions will not be caused by
the treatment process.
Mass Balance. The mass of VOC and SVH liquids collected by the hot
air/steam stripping process was measured directly. The mass of
VOCs and SVHs removed from the soil also was calculated using
measured values based on the chemical and physical analyses. The
mass of VOCs collected is essentially equivalent to the calculated
mass of VOCs removed from the soil. The following results present
the treatment blocks remediated, the mass of VOCs collected by the
process, the mass of VOCs removed from the soil, and the percent
VOCs recovered. (The six identified blocks were analyzed
separately because the CDHS was making also confirmatory
measurements.)
Calculated VOCs
VOCs Collected Removed from % VOCs
Treatment Blocks by the Process. Ib the Soil,. Ib Recovered
6(A8g, A9g, AlOg, 124.5 142.5 87.4
B50n, B51n, B51m)
10 (all Blocks) 264.8 296.6 89.3
This comparison indicates that the Detoxifier treatment removes
VOCs from the soil and captures them in the process. The closure
of the VOC mass balance is within what is possible for the
constraints of the experimental measurements. The reasons for the
difference in the mass of VOCs collected and the mass removed from
the soil can be attributed to the following:
38
-------
The size of the test block group compared with the size
of the process system (i.e., mass possibly lost through
process variables)
Uncertainties in sampling
Other unidentified causes, e.g., hang-up in the process
train
The mass of SVHs in the collected hydrocarbon liquid was not
equivalent to the mass of SVHs removed from the soil. After
remediation of six of SVHs had been collected, whereas it was
calculated that 1,018 Ib of SVHs had been removed from the
treatment blocks.
SVHs in SVHs Removed % SVHs
Treatment Blocks Process, Ib From the Soil, Ib Recovered
6(A8g, A9g, AlOg, 2.1 1,018 0.21
B50n, B51n, B51m)
The reason for this difference in mass has not been formally
demonstrated. However, examination of the compounds present, their
potential chemical fate in the presence of hot air/steam, and
chemical analysis suggest that hydrolysis catalyzed by the clay in
the soil plays an important role in their disappearance. It is
also believed that the hydrolysis reactions produce compounds that
are bound to the clay and cannot be identified by the analytical
techniques employed.
SITE DEMONSTRATION
The EPA SITE Program conducted a demonstration of the Detoxif ier at
the San Pedro site in September, 1989. Twelve blocks in Area A
were treated to depths of 5 feet. Pre and Post-Treatment samples
were taken from zero to 5 feet. Composite samples were analyzed
from the 36 post-treatment cores and discrete samples were analyzed
from four of the 36 cores. The report from this project is in
preparation. Preliminary results are given in Tables 5 and 6. The
results are in general similar to those in the above 10 Block Test.
Examination of the results showed several cores with very high
post-treatment VOC, e.g., in excess of 300 ppm. A careful review
of all the chemical analyses suggested that this was a result of
39
-------
Test Parameters
Treatment Depth
Buffer Zone
Cores per Block
Sample type
Deeper Samples
Pre/Post Sampling
8240/8270
10
Block
6'
Yes
1 & 2
D
4
Both
Both
12
Block
5'
No
3
C/D
No
Both
Both
6
Block
12'
Yes
3
C/D
2
Post
8240
Sampling Depth • 5'
-------
Post Treatment 12 Block
Total VOC's
Block ppm Block ppm
25e 14.4 31e 60.7
26e 12.3 32e 64.0
27e 28.5 33e 104.4
28e 34.1 34e 196.3
29e 81.5 35e 59.9
30e 145.4 36e 55.9
-------
the treatment protocol and not the technology. The 12 blocks for
the SITE Demonstration were treated to 5 feet without building a
buffer from 5 feet to 6 feet between the clean and contaminated
soil, thus, the process appeared to draw up materials from below
during treatment. This interesting result confirmed the need for
a buffer zone.
Shortly after the SITE Demonstration TTUSA initiated treatment into
the saturated zone to 12 feet of depth. The conclusion is that
the process operates with improved efficiency in the saturated
zone. The results for this work were favorable and are being
prepared as part of another report. Because of these results in
the saturated zone and a desire to obtain additional data with a
buffer zone, the SITE Demonstration sampled an additional six
blocks in February, 1990. Preliminary results are shown in Figure
7 for VOC. They are in good agreement with the previous work in
that the VOCs are effectively removed. Both sets of data will be
discussed in detail in the SITE Demonstration report scheduled for
summer 1990.
SUMMARY
This paper has presented the first detailed information on a new
technology for the in-situ removal of VOCs from soil. The data
indicate that VOCs are removed substantially from soil and that
they are captured above ground for disposal without adverse effects
to the ambient environment. No information is given regarding the
economics of this process, although it is believed to be
competitive with alternate remediation schemes. This technology
may be an attractive alternative to landfill disposal as landfills
become unavailable and passive vapor extraction schemes are slow or
uncertain in their efficiency. The technology has demonstrated
also that SVHs are removed from soil by the action of steam and air
in the soil environment. The mechanisms for this are not yet well
understood. The removal of SVH may add to the utility of this
technology in the future.
42
-------
Post Treatment 6 Block Data
Total VOC's
Block ppm Block ppm
26n 16 29n 80
27n 22 30n 119
28n 36 31n 45
-------
EXTRACTION OF PCB FROM SOIL WITH EXTRAKSOL™
Diana Mourato and Jean Paquin
ENVIRONCORP INC.
Montreal, Canada
H1K4E4
ABSTRACT
Polychlorinated biphenyls have been succesfully extracted from clay-bearing soil, sand and
Fuller's earth, by the Extraksol™ process. Extraksol™ process is a mobile decontamination
technology which treats soil and sludges by solvent extraction. Treatment with Extraksol™
involves material washing, drying and solvent regeneration. Contaminant removal is achieved
through desorption/dissolution mechanisms. The treated material is dried and acceptable to be
reinstalled in its original location.
The process provides a fast, efficient and versatile alternative for treatment of
PCB-contaminated soil and sludge. The contaminants extracted from the soil matrix are
transferred to the extraction fluids. These are thereafter concentrated in the residues of
distillation after solvent regeneration. Removal and concentration of the contaminants ensures
an important waste volume reduction.
Extraksol™ is a flexible process designed to extract a variety of organic contaminants from
unconsolidated solids. Polyaromatic hydrocarbons, oils and pentachlorophenols have also been
extracted from sludges, activated carbon, porous stones and gravel, with high removal
efficiencies.
44
-------
INTRODUCTION
Extraksol™ is a mobile decontamination process which extracts organic contaminants from
unconsolidated materials such as soil, sludge, gravel, etc. Polar and non-polar contaminants are
extracted from the soils by solvent washing through desorption / dissolution mechanisms.
Treatment with Extraksol involves material washing and drying, which generatesa
decontaminated, dry soil, ready to be returned to its original location.
Extraksol™ has been developed and commercialized by the Sanivan Group to complement its
hazardous waste treatment technologies. Initially developed to extract polychlorinated biphenyls
(PCBs) from contaminated soil, Extraksol™ has since demonstrated to be an efficient solution to
recycle sand, gravel, mixed soil and sludge contaminated with oils, greases, polyaromatic
hydrocarbons (PAHs), chlorinated organics such as pentachlorophenols (PCPs) and other
common organic pollutants.
Extraksol™ has been designed as a closed system (no gaseous or liquid discharges) which
considerably reduces the volume of contaminants. The solvents used are non-chlorinated,
non-toxic and non- persistent. Solvent regeneration is an integral part of Extraksol™ and is
carried in parallel to the extraction activities. Solvent regeneration allows to associate the
benefits of soil decontamination to the advantages of contaminant volume reduction.
In Eastern Canada, there is no alternative for PCB contaminated soil other than containment and
securisation. Incineration of PCBs is not yet an accepted solution whereas biodegradation
techniques are not available.
PROCESS DESCRIPTION
The process consists of the 3 phases: soil washing, solvent regeneration and soil drying.
Soil Washing
After introducing 8 totO drums of soil (or other contaminated solid material) within the
extraction vessel, the extractor is purged and the washing cycle is initiated.
During the soil washing phase, clean solvent is continuously pumped and withdrawn into and
from the extractor, creating a dynamic system. To further enhance soil-solvent contact, the
extractor is slowly rotated on its axis.
45
-------
The circulating fluid migrates through the soil and dissolves/desorbs the contaminants present
in the matrix according to the affinity of the solvent for the contaminant. The contaminant is
transferred from the soil matrix to the solvent and is carried out of the extractor with the fluid.
The used solvent is transferred to a storage tank.
The extraction phase is continued until the soil is thought to be decontaminated. Visual analysis
of the solvent sampled at the extractor's outlet provides an indication of the state of
decontamination. Circulation of 500 gallons of solvent is equivalent to one extraction cycle.
Although the washing time varies with the type and concentration of contaminant, it ranges from
1 hour to 2 hours for 4 tons of material.
After the soil washing cycles are terminated, the solvent is withdrawn from the extractor and
transferred to the contaminated solvent tank.
Solvent regeneration phase
Solvent regeneration is conducted in parallel to the soil washing phase. Regeneration is carried
by distillation and takes advantage of the differences in evaporation temperatures of the
extraction fluids and the contaminants. The contaminants are concentrated and recovered as
residues of distillation.
The solvent regeneration phase is an integral part of the Extraksol™ process which considerable
reduces the cost of decontamination. Since the solvent is regenerated to its original composition,
it can be reused without limitation and without the need of additives.
Soil drying phase
After treatment with Extraksol™, the soil is dry and can often be relocated in its original
location. Soil drying consists of removing the residual solvent from the soil. This is
accomplished by circulating hot inert gas through the soil within the extractor. Gas circulation
is induced by a vacuum pump in a closed system operation.
The hot gas evaporates the solvents and carries the vapors out of the extractor and into a
refrigerated condenser in which the gaseous solvent becomes liquid. The liquid solvent is
separated from the inert gas and conveyed into a storage tank.
46
-------
The solvent-free inert gas is heated by passage through a hot heat exchanger. The circuit is
completed when the hot gas is reinjected into the extractor. The drying phase extends from 1
hour to 1 1/2 hours depending on the type of matrix treated.
TYPICAL RESULTS - PCB EXTRACTION
Typical PCB extraction results in soils range from 95% to 99.9%, depending on the initial PCB
concentration and the residual PCB allowed to be left in the soil. The process has the capacity to
treat strongly contaminated material and declassify these materials as non-hazardous to levels
of PCb < 5 ppm. The residual PCB concentration left in the soil is a function of the number of
extraction cycles applied. In general, a 30% PCB removal is obtained in each extraction cycle.
Since Extraksol™ is a batch process, it offers the flexibilty to treat different types of materials
to different levels of decontamination.
To date, the Extraksol process has succesfully treated materials as various as Fuller's earths,
activated carbons, refinery sludges and wood treatment sludges. Organic contaminants such as
oils and greases, polyaromalic hydrocarbons (PAH), pentachlorophenols (PCP) and phenols
have been extracted from a variety of solid matrices.
CONCLUSIONS
Enough work has been done with the one ton per hour Extraksol™ unit to confirm that the
process can effectively extract oils, PCB, PAH, and PCP from soil and generate a "clean", dry
soil, which once mixed with top-soil for revegetation can be returned to its original location.
Decontamination of sand, mixed soils clayey soil, sludge, gravel, activated carbon and stones
have demonstrated that the process is very flexible and can be adapted to solve a range of
environmental problems.
The flexibility and mobility of the 1 ton/hour unit and its 3 day set-up period makes is a perfect
process to be used on small projects with a maximum of 300 tons of material to be treated. The
Sanivan Group is planning to build a larger Extraksol unit which will have the capacity to treat
6 to 8 tons per hour. This mobile unit will be designed on the same principles of operation as the
smaller unit. This larger unit will be operated by 2 operators as the smaller unit.
47
-------
TABLE 1, SUMMARY OF RESULTS OBTAINED WITH EXTRAKSOL'S PILOT SCALE UNIT
TYPE OF SOIL
Clay
Clay
Sand
Mixed Soil
Mixed Soil
Mixed Soil
Mixed Soil
TYPE OF FLUID
# 1
# 2
# 1
# 1
# 2
#3
# 4
REMOVAL OF OILS AND GREASES
Initial
Concent
(ppm)
1,970
8,040
470
13,890
14,400
34,300
21,540
Final
Concent
(ppm)
847
590
220
2,270
1,210
1,440
1,880
% Removal
(%)
56.9
92.7
53.2
87.7
92.0
96.0
91.0
REMOVAL OF PCB
Initial
Concent
(ppm)
7,925
2,055
600
3.6
5.3
5.2
4.8
Final
Concent
(ppm)
2,080
48.8
6.3
0.69
0.70
1.0
1.1
% Removal
( % )
73.8
97.6
98.9
89.0
87.0
81.0
77.0
TABLE 2, SUMMARY OF RESULTS OBTAINED WITH THE 1 TON/HOUR
EXTRAKSOL UNIT - PCB REMOVAL
TYPE OF
SOIL
clay-bearing
clay-bearing
clay-bearing
TYPE OF
FLUID
#2
# 1
# 2
Initial PCB
Concent.
(ppm)
150
163
54
Final PCB
Concent.
(ppm)
1 4
28
4.4
% Removal
(%)
91.0
82.0
92.0
48
-------
TABLE 3, SUMMARY OF RESULTS OBTAINED WITH THE 1 TON/HOUR
EXTRAKSOL UNIT - O&G REMOVAL
TYPE OF
SOIL
clay-bearing
clay-bearing
clay-bearing
refinery clayey soil
refinery oily sludge
refinery oily sludge
refinery oily sludge
refinery oily sludge
Fuller's earth
Fuller's earth
Fuller's earth
Fuller's earth
pulp & paper porous gravels
pulp & paper porous gravels
TYPE OF
FLUID
#2
# 1
#2
#2
#2
u r)
IT e.
a. o
TT Ł
" O
Tt Ł.
# 2
# 2
M O
it Ł.
M O
•tt f.
# 2
# 2
Initial O&G
Concent.
(ppm)
1,801
1,789
600
15,000
49,000
72,000
73,000
70,000
366,000
447,000
313,000
332,000
10,000
1 ,040
Final C&G
Concent.
(ppm)
1 82
166
80
800
4,200
2,000
4,800
340
4,200
5,500
3,700
4,000
3,690
207
% Removal
(%)
90
82
92
95
91
97
93
99
99
99
99
99
63
80
TABLE 4, SUMMARY OF RESULTS OBTAINED WITH THE 1 TON/HOUR
EXTRAKSOL UNIT - PAH REMOVAL
TYPE OF
SOIL
refinery clayey soil
refinery oily sludge
refinery oily sludge
refinery oily sludge
refinery oily sludge
TYPE OF
FLUID
# 2
2
2
2
2
Initial PAH
Concent.
(ppm)
332
81
240
150
1,739
Final PAH
Concent.
(ppm)
55
1 6
1 0
1 9
"30
% Removal
(%)
83
8 1
96
87
92
49
-------
TABLE 5, SUMMARY OF RESULTS OBTAINED WITH THE 1 TON/HOUR
EXTRAKSOL UNIT - POP REMOVAL
TYPE OF
WASTE
porous gravel
porous gravel
porous stones
activated carbon
TYPE OF
SOLVENT
# 2
# 2
#2
#2
Initial PCP
Concent.
(ppm)
8.2
81.4
38.5
744
Final PCP
Concent.
(ppm)
<0.82
<0.21
19.5
83
% Removal
(%)
>90
>99.7
50
89
50
-------
DU PONT/OBERLIN MICROFILTRATION SYSTEM FOR HAZARDOUS
WASTEWATERS
Paper to be presented at U.S. EPA Second Forum on Innovative Hazardous
Waste Treatment Technologies
May 15-17, 1990
Philadelphia, PA USA
Dr. Ernest Mayer
E. I. du Pont de Nemours, Inc.
Engineering Department Louviers 1359
P. O. Box 6090
Newark, DE 19714
(302) 366-3652
Abstract
Du Pont has developed and recently commercialized a new filter media based
on Tyvek® flash spinning technology. This new media has an asymmetric pore
structure, a greater number of submicron pores, and a smaller average pore size (1-3).
As a consequence, it has superior filtration properties, longer life, and in many
instances can compete with microporous membranes, PTFE laminates, and various
melt-blown media. When coupled with an automatic pressure filter (APF), it provides
an automatic wastewater filtration process that is a dry-cake alternative to conventional
crossflow microfilters and ultrafilters. This Tyvek®/APF process has proved extremely
useful in filtering heavy metal and other hazardous wastewaters to meet strict EPA
NPDES discharge limits. Specific examples and actual case histories will be
highlighted to illustrate its benefits. In addition, this Tyvek®/APF technology has
recently been selected for EPA's SITE-3 program which is also discussed.
51
-------
TYVEK® T-980 MEDIA
Du Pont has recently commercialized a new filter media based on Tyvek® flash
spinning technology. It has an asymmetric pore structure, a greater number of
submicron pores, and a smaller average pore size.(1-3). As a consequence, it has
superior filtration properties and longer life, and in many instances it can compete with
microporous membranes, PTFE laminates, and various melt-blown media (3). Its key
property is its tight pore structure at a very low cost compared to competitive products.
Table I outlines the media cost per gallon of waste filtered and shows that Tyvek® T-
980 grade, which is the lowest basis weight manufactured (0.9 oz/yd2) and the only
grade evaluated here, is very cost effective. In most applications the T980 grade was
sufficient so the added cost for a higher basis weight grade was not warranted. For
example, Tyvek® T-980 produced slightly poorer effluent quality than the 'standard'
0.45p. microporous membrane at a fraction of the cost; equivalent effluent quality to the
PTFE laminate at a fraction of the cost; and much better effluent quality than typical 1-
and 5- micron melt-blowns at equal or significantly lower cost (depending on
application). Tests with actual wastes showed the greatest cost benefit (Table I). A
similar cost benefit was obtained in operation with a low-level radioactive plating waste
at the Savannah River Plant (4,5). This installation realized almost a $200 M annual
savings when Tyvek® T-980 media was used with a more efficient filter aid. An added
benefit of the Tyvek® is its superior strength compared to microporous membranes and
the PTFE laminate media. This strength permits its use in robust automatic pressure
filters (6). The high strength (7) coupled with the tight, ~1- micron nominal pore
structure (3) led to Tyvek®'s use in the EPA Superfund SITE program (8).
OBERLIN AUTOMATIC PRESSURE FILTER (APF^
Tyvek® T-980 media requires a suitable filter housing. The Oberlin Filter
Company's automatic pressure filter (APF) was chosen for its simplicity and fully
automatic operation (9). The Oberlin APF has other advantages, namely:
• Completely automatic, unattended operation save for Tyvek® roll
replacement and chemical treatment makeup.
• Enclosed operation tor safety and handling of hazardous wastes.
• Fairly high operating pressure (up to 60 psig).
Mayerl 52
-------
• Completely in-line treatment-chemical addition, ie, filter aids and polymer
flocculants.
• Automated shutdown flushing capability.
• Cake washing capability to remove hazardous filtrate, if required.
• Automatic, positive dry cake discharge.
• Direct submicron filtration without the need for further downstream
processing.
• Reliable, low-maintenance performance.
• Completely automatic safety interlocks and enclosures and/or purging, if
required.
• Explosion-proof design, if required.
• Completely integrated pumping system(s), if required.
• Dirty-media takeup and doctoring, brushing, or washing, if required
(automatically accomplished during takeup).
• Standard PLC control.
TYVEK®/OBERLIN APF COMBINATION
Thus, the Tyvek®/Oberlin APF combination has some unique advantages,
especially its completely automatic submicron, low-cost filtration and its dry cake
discharge (8). This dry cake discharge feature is precisely why the resultant cakes
pass the modified EPA "Paint Filter Test" for land disposal (10) and in some instances
pass the new EPA TCLP (Toxic Characteristic Leaching Procedure) test for hazardous
components (11). Dry cake/submicron filtration in one operation is why the
Tyvek®/Oberlin APF combination was selected at Savannah River over conventional
crossflow microfilters and ultrafilters. In plating-waste treatment the simple, one-step
Tyvek®/Oberlin APF combination replaced the conventional three-step
clarifier/overflow sand filter/underflow recessed filter press process. However, the
purpose of this paper is to highlight case histories where the Tyvek®/Oberlin APF
filtration process has been successfully applied. The technology is most suitable for
hazardous wastewater where the solids loading is not too high. Examples include
53
Mayerl
-------
contaminated groundwater, plating wastewaters, low-level radioactive wastes, plant
equipment/floor washings, cyanidic wastes, plant wastewaters that contain heavy
metals, and metal grinding wastes. (7)
SAVANNAH RIVER SIMULATED WASTE
Table II summarizes a series of Oberlin bench-scale tests with simulated plating
wastewater (to duplicate existing plating line effluents without uranium). The Tyvek®
T-980 media matches all competitive media in rate and filtrate turbidity, but produces a
drier cake because of better air-blowing due to less media blinding. Tyvek® provides
excellent cake release and the lowest (15 mg) "A Wt Gain" after the cake is removed.
Furthermore, the Tyvek® T-980 is much stronger than the competitive media so it
could be rerolled in the Oberlin APF without a carrying belt. All the others required
careful handling. This feature simplified the operation and reduced maintenance.
ACTUAL SAVANNAH RIVER OPERATION
Table III details actual Savannah River plant wastewater treatment specifically
aimed at aluminum and uranium removal (4,5). The data were obtained from two
Tyvek® T-980/Oberlin APF units which had been operating for about three years.
Aluminum forming and metal finishing operations generate a high content of solids,
aluminum, and turbidity. These solids can be reduced below the National Pollution
Discharge Elimination System (NPDES) limits with the T-980/Oberlin combination at
very high 1200 gfd (gallons/sq ft/day) rates. This rate is about seven times higher than
the ultrafilter (UF)/reverse osmosis (RO) system (200 gfd) originally considered. Pilot
testing also showed that the UF/RO system repeatedly fouled with this waste, requiring
aggressive cleaning agents which significantly added to the waste volume.
ELECTRONICS MANUFACTURING PLANT WASTEWATER
This plant's effluent exceeded the local sewer authority's lead discharge limit.
As a consequence, the plant was mandated to cease discharge and dispose of their
wastewater in an off-site hazardous waste landfill at a $0.55/gallon cost. The Tyvek®
T-980/Oberlin combination was installed instead of a crossflow microfilter because its
dry cake feature significantly reduces waste volume. The T-980/Oberlin units repaid
their cost in three months of operation based on disposal cost savings alone.
Mayerl 54
-------
Table IV details tests similar to those with the Savannah River plating waste.
Tyvek® T-980 again compares favorably to the competitive media in rate and filtrate
turbidity (except for the 0.45 - micron microporous membrane). However, Tyvek®
T-980 produces the driest cake (75% solids), gives excellent cake release, and has
the lowest media "A Wt Gain" (0 mg). Again, the Tyvek® T-980 media was the only
one that could withstand the rigors of Oberlin APF rerolling without a support belt.
Table V details actual plant operation and compares the Tyvek® T-980 effluent
with the plant's discharge limits. As shown, the T-980/Oberlin units reduced the
effluent Total Suspended Solids (TSS) and lead levels to well below the plant's
required discharge limits at very high 1500 gfd flux ratios. In addition, the Oberlin APF
routinely achieved >60% solids dry cakes that could be disposed of in a RCRA-
approved landfill (at significant cost savings compared to the concentrate from the
crossflow microfilter).
ELECTRONICS PLANT CARTRIDGE REPLACEMENT
This plant's effluent had to be polished by absolute 0.45-micron cartridge filters
to meet heavy-metal discharge limits. Cartridge costs exceeded $1200 daily plus
significant labor charges. A Tyvek® T-980/Oberlin APF unit was installed at significant
cost savings compared to these cartridges or a crossflow microfilter which was also
considered. Payback was on the order of three months; the Oberlin APF produced dry
cakes suitable for landfilling off-site, and significant labor savings resulted.
Table VI demonstrates that the Tyvek®/Oberlin system easily met the TSS and lead
discharge limits at very high 1500 gfd flux. Cakes were also quite dry at -50% solids.
MUNITIONS PLANT WASTEWATER
This plant was faced with severe RCRA restrictions on heavy metals discharge.
They hired an environmental engineering firm to solve their waste problem, namely to
remove all metals (primarily lead, antimony, and barium) to well below discharge limits
and to produce dry cakes that could be hauled off-site to a RCRA-approved landfill.
This firm opted for the conventional crossflow microfilter/press approach which was
expensive, required manual labor, and was prone to fouling. I proposed the less
costly, single-step, Tyvek® T-980/Oberlin approach which was installed about a year
ago. Table VII shows that effluent TSS, lead, antimony, and barium levels are well
Mayerl 55
-------
below the discharge limits. In addition, the system demonstrated very high capacity
(15,000 gallons per day, gpd), very high flux (3700 gfd), and fairly dry, 40%, solids
cakes. The unit operates at 200° F which would have been troublesome for the
crossflow filter and explains the somewhat higher effluent lead level.
CLARIFIER UNDERFLOW HEAVY METALS REMOVAL
This plant was faced with a land ban of their main clarifier underflow sludge
(-2% solids) because it did not pass the EPA "Paint Filter Test" (10). This clarifier
treated the entire plant effluent and removed primarily lead, zinc, and copper. To
satisfy the RCRA land ban restrictions, the plant hired an expensive ($400/day) mobile
dewaterer who used a manual, recessed-chamber filter press that produced sloppy
cakes. These cakes had to be shovelled into dumpsters for off-site disposal. The
Tyvek®/Oberlin combination was installed about six months ago. Table VIII shows
excellent effluent quality at very high, 10,000 gpd, capacity and 900 gfd flux. The
cakes were sufficiently dry (50% solids vs. 40% requirement) to pass the "Paint Filter
Test" and were acceptable to the hauler/landfill operator at a significant cost saving to
the plant. The Tyvek®/ Oberlin system operates automatically at significant labor
savings compared to the manual press.
GROUNDWATER BARIUM REMOVAL
Contaminated leachate from a landfill required silt and barium removal to pre-
vent fouling in downstream organics removal equipment. A Tyvek®/Oberlin system
was chosen for this application because of its automatic operation and dry cake
feature. Table IX shows that TSS and barium levels are well below required limits at
an extraordinarily high 2700 gfd flux. This was substantially higher than the flux
obtained from an ultrafilter also considered for this application.
BATTERY MANUFACTURING HEAVY METALS REMOVAL
This plant consistently exceeded their permitted discharge limits from their con-
ventional clarifier/underflow press system and considered an overflow polishing sand
filter. Simultaneously, they heard of our Tyvek® T-980/Oberlin APF technology and
decided to evaluate it as a polishing filter. Testing showed that it could replace the
entire clarifier/press installation. Table X shows excellent metals removal, TSS
reduction, and excellent effluent turbidity (0.5 NTU). Flux is quite high at 2400 gfd
Mayer 1 56
-------
and capacity from the single Tyvek®/Oberlin system exceeds 60,000 gpd. These
benefits are in addition to dry cakes that could be easily disposed of in a RCRA-
approved landfill.
SUMMARY
These cases and actual operating applications demonstrate the utility of the
Tyvek® T-980/Oberlin APF technology to remove heavy metals at very high flux rates
and to simultaneously produce dry cakes that pass the EPA "Paint Filter Test." This
technology is quite competitive when compared to microfilter cartridges, crossflow
microfilters, and ultrafilters. In some instances, the Tyvek®/Oberlin system can replace
the conventional three-stage metal treatment process of clarifier, underflow filter press,
and overflow polishing sand filter. Thus, the waste engineer/consultant now has a
simpler, one-step process for treating hazardous metal-bearing wastewaters.
Mayerl 57
-------
REFERENCES
1. H. S. Lim and E. Mayer, Paper presented at 1987 Membrane Technology
Conference, Boston, MA, October 22, 1987.
2. H. S. Lim and E. Mayer, Paper presented at the International Technical
Conference on Filtration and Separation, Ocean City, MD, March 22, 1988.
3. H. S. Lim and E. Mayer, Fluid/Particle Separation Journal. 2(1), 17-21, (March,
1989).
4. H. L Martin, Paper presented at 10th Annual AESF/EPA Conference on
Environmental Control for the Metal Finishing Industry, Orlando, FL,
January 23-25, 1989.
5. H. L. Martin, P. K. Gurney, and L. P. Fernandez, Paper presented at 8th Annual
AESF/EPA Conference on Pollution Control for the Metal Finishing Industry,
San Diego, CA, February 11, 1987.
6. E. Mayer, Filtration News. 24-27 (May/June, 1988): "New Trends in SLS
Dewatering Equipment."
7. Du Pont Tyvek® Bulletin E-24534, 1988: "Tyvek® Engineered Specifically for
Filtration."
8. E. Mayer, Request for Proposal, SITE-3 Solicitation, Demonstration of
Alternative and/or Innovative Technologies, "Groundwater Remediation Via
Low-Cost Microfiltration for Removal of Heavy Metals and Suspended Solids,"
February, 1988.
9. Oberlin Filter Co.'s Bulletin, February, 1988: "Oberlin Pressure Filter."
10. N. J. Sell; Pollution Eng.. 44-49 (August, 1988).
11. Federal Register, Vol. 51, No. 114 (June 13, 1986).
Mayerl
58
-------
TABLE I
MEDIA COST PER GALLON WASTE FILTERED
ActualCosts/Gallon Filtrat fcents/aair
on
10
Media,
Microporous Membrane
Tyvek® T-980
PTFE Laminate
Melt-blown PP
Melt-blown PP
Nom.
Rating
(um)
0.45
1
0.8
1
~5
Overall
Filtrate
Quality
Excellent
V. Good
V. Good
Good
Poor
60 ppm
ACFTD**
6.7
0.5
8.2
0.7
0.3
600 ppm
ACFTD
29
1.3
41
2.4
1.5
Low-Level
Radioactive
Waste
69
1.7
63
6.5
4.0
Lead-
Bearing
Waste
14
0.3
12
1.3
0.8
*Based on actual media costs, flux rates, and measured life cycles.
**AC Fine Test Dust (ACFTD) challenge tests.
-------
TABLE II
en
o
SAVANNAH RIVER SIMULATED WASTE *
TESTING IN QBERLIN PRESSURE FILTER
Type
Microporous Membrane
Tyvek® T-980
PTFE Laminate
Melt-blown PP
Melt-blown PP
Media**
Nom.
Rating
W
0.45
1
0.8
1
~5
AWt.
Gain
(Mg.)
50
15
60
560
445
Cake
Release
Good
Excellent
Fair
Fair
Good
Cake
o/
/o
Solids
30
34
28
32
29
Filtrate
Turbidity
rNTLM
0.46
0.29
0.51
0.49
5.0
Avg.
Rate
(gfd)
1050
1070
1120
1040
1150
* Simulated waste with 600 ppm metal hydroxide solids (except uranium) plus 3:1 ratio of Celite 577 filter aid.
** A Wt. Gain is media weight gain after cake removed; and cake release judged visually after bending media
to simulate discharge around roll.
-------
TABLE III
ACTUAL SAVANNAH RIVER PLANT OPERATION
WITH TYVEK® T-980/OBERLIN PRESSURE FILTER
Propertv
Turbidity (NTU)
TSS (ppm)
Aluminum (ppm)
Lead (ppm)
Zinc (ppm)
Copper (ppm)
Uranium (ppm)
Capacity (gpd)
Flux (gfd)
Raw
Waste
110
687
127
1.6
0.5
2.0
2.3
35,000
—
T-980/
Oberlin
APF
0.32
1.4
0.95
0.2
<0.1
<0.1
0.01
60,000
1,200
NPDES
Limits *
—
31
3.2
0.43
0.32
0.21
0.5
—
._
*Actual State discharge permit values.
61
-------
TABLE IV
CTv
ro
ELECTRONICS WASTEWATER TESTING *
IN OBERLIN PRESSURE FILTER
Type
Microporous Membrane
Tyvek® T-980
PTFE Laminate
Melt-blown PP
Melt-blown PP
Media**
Nom.
Rating
W
0.45
1
0.8
1
-5
AWt.
Gain
(Ma.)
20
0
20
50
135
Cake
Release
V. Good
Excellent
Excellent
Fair
Good
Cake
%
Solids
63
75
72
64
63
Filtrate
Turbidity
(NTU)
0.47
2.0
1.4
2.6
2.6
Avg.
Rate
(gW)
1730
2160
2050
1900
1900
*Actual electronics manufacturing wastewater that contains about 6500 ppm heavy metals (Pb, Ba, etc) plus
1950 ppm Superaid filter aid (0.3 ratio) and 50 ppm high-charge cationic polymer flocculant (Praestol K-122L).
**A Wt. Gain is media weight gain after cake removed; and cake release judged visually after bending media
to simulate discharge around roll.
-------
TABLE V
ACTUAL ELECTRONICS MANUFACTURING PLANT WASTEWATER*
WITH TYVEK® T-980/OBERLIN PRESSURE FILTER
Property
Turbidity (NTU)
TSS (ppm)
Lead (ppm)
Barium (ppm)
Capacity (gpd)
Flux (gfd)
Cake % Solids
Raw
Waste
>1000
6500
500
-100
2500
—
...
T-980/
Oberlin
x APF
3.4
6.0
<0.25
<1
7500
1500
>60
Discharge
Limits**
N.R.
26
0.7
N.R.
—
—
—
* Includes floor washings also
** N.R. = not regulated.
63
-------
TABLE VI
ELECTRONICS PLANT CARTRIDGE FILTER REPLACEMENT*
WITH TYVEK® T-980/OBERLIN PRESSURE FILTER
Property
Turbidity (NTU)
TSS (ppm)
Lead (ppm)
Capacity (gpd)
Flux (gfd)
Cake % Solids
Raw
Waste
2400
5500
2300
2000
—
...
T-980/
Oberlin
APF
0.43
1.0
0.18
11,000
1500
50
Discharge
Limits**
N.R.
20
0.7
—
—
...
* Includes floor washings also
** N.R. = not regulated.
64
-------
TABLE VII
MUNITIONS MANUFACTURING PLANT WASTEWATER*
WITH TYVEK® T-980/OBERLIN PRESSURE FILTER
Property
Turbidity (NTU)
TSS (ppm)
Lead (ppm)
Antimony (ppm)
Barium (ppm)
Capacity (gpd)
Flux (gfd)
Cake % Solids
Raw
Waste
>1000
1100
540
4.4
50
3000
—
...
T-980/
Oberlin
APF
0.8
0.5
1.8
0.4
0.6
15,000
3700
40
Discharge
Limits**
—
30
5.0
1.0
50
—
—
•__
* Includes floor washings also
** N.R. = not regulated.
65
-------
TABLE VIII
CLARIFIER UNDERFLOW HEAVY METALS REMOVAL*
WITH TYVEK® T-980/OBERLIN PRESSURE FILTER
Property
Turbidity (NTU)
TSS (ppm)
Lead (ppm)
Zinc (ppm)
Copper (ppm)
Capacity (gpd)
Flux (gfd)
Cake % Solids
Raw
Waste
>1000
17,000
40
410
1050
6800
—
...
T-980/
Oberlin
APF
1.0
2.5
<0.01
0.2
1.2
10,000
900
50
Discharge
Limits**
N.R.
20
5.0
5.0
5.0
—
—
40
* Existing plant clarifier underflow sludge was previously hauled off-site to hazardous
landfill.
** N.R. = not regulated.
66
-------
TABLE IX
CONTAMINATED GROUNDWATER BARIUM REMOVAL*
WITH TYVEK® T-980/OBERLIN PRESSURE FILTER
Property
Turbidity (NTU)
TSS (ppm)
Barium (ppm)
Flux (gfd)
Raw
Waste
>1000
1200
500
—
T-980/
Oberlin
APF
9.0
6.6
6
2700
Discharge
Limits**
N.R.
20
50
—
* From hazardous landfill leachate.
** N.R. = not regulated.
67
-------
TABLE X
DIRECT FILTRATION FOR HEAVY METALS REMOVAL
FROM A BATTERY MANUFACTURING PLANT*
WITH TYVEK® T-98Q/QBERLIN PRESSURE FILTER
Prooertv
Turbidity (NTU)
TSS (ppm)
Nickel (ppm)
Cadmium (ppm)
Zinc (ppm)
Cobalt (ppm)
Capacity (gpd)
Flux (gfd)
Raw
Waste
175
1600
30-50
15-40
0.2-1.0
0.4-1.5
40,000
—
T-980/
Oberlin
APF
0.5
<1
<0.10
<0.05
<0.10
<0.05
60,000
2400
Discharge
Limits**
N.R.
20
2.27
0.4
1.68
0.22
—
...
* Replaced clarifier and underflow press.
** N.R. = not regulated.
68
-------
FLUID/PARTICLE
SEPARATION JOURNAL
TYVEK FOR MICROFILTRATION MEDIA
Hyun S. Lim, E.I.DuPont, Fibers Department
P.O. Box 27001, Richmond VA 23261
Ernest Mayer, E.I. DuPont, Engineering Department
P.O. Box 6090, Newark, DE 19714
THE AMERICAN FILTRATION SOCIETY
VOLUME 2, NUMBER 1, MARCH 1989
69
-------
TYVEK FOR MICROFILTRATION MEDIA
Hyun S. Lim, E.I. DuPont, Fibers Department
P.O. Box 27001, Richmond, VA 23261
Ernest Mayer, E.I. DuPont, Engineering Department
P.O. Box 6090, Newark, DE 19714
Tyvek™ spunbonded olefin is flash spun from high density
polyethelene and is used in the medical packaging and
protective apparel markets for its microbial barrier and air
permeability properties. However, use of Tyvek as a filtration
media has been limited because of poor permeability and low
dirt holding capacity. Recently, a new T-980 series Tyvek nas
been developed specifically for filtration applications by
creating greater number of submicron pores and a smaller
average pore size.
Permeability and filtration efficiency and capacity of Tyvek
measured by AC Fine Test Dust challenge tests are discussed
and its filtration performance parameters are compared with
melt blown media and microporous membranes. The results of
actual treatment of waste waters containing metal oxide,
uranium paniculate, and lead are presented to illustrate the
benefit of Tyvek for meeting the strict Environmental
Protection Agency discharge limits.
Fluid/Particle Sep. J, Vol.2, 17-21 (1989).
FILTER MEDIA CHOICES AND
MANUFACTURING
A proper selection of media is the most important factor
that determines the filler performance. The criteria for selecting
the optimum media are listed below:
1. How much fluid can flow through it (permeability)
2. How small a particle can it retain (nitration efficiency)
3. How long can it last (filtration capacity or life)
4. How much does it cost (nitration cost)
An extremely high standard of separation of submicron
particles requires use of microporous membranes. Most
membranes can only be used with adequate substrate support
and must be protected against rough handling. The major
microporous membranes are manufactured by a wet-cast
process or biaxial stretching which are relatively slow,
inefficient and generally expensive (costs on the order of $7-
40/sq yd). A new technology utilizing ultraviolet light for
crosslmking of monomer and oligmers at a high operating
speed has been developed to lower the media cost However, in
many nitration applications requiring fine separation of
submicron particles, the use of microporous membranes is
limited due to the high costs.
An alternative lower cost choice is available with
nonwoven media. Three basic technologies exist in the
nonwoven industry to produce filter media: melt spinning.
melt blowing, and flash spinning.
Melt spinning involves extruding filaments through
spinneret orifices followed by quenching the filaments with
cross-flow air and drawing through an aspirator jet by cocurrem
air streams. The descending filaments are electrically charged to
separate them and lay them down in a completely random
configuration onto a moving bell, under which is a suction
box. The fibers are relatively coarse as fineness is limited by
the capillary size in the spinneret.
Fibers of very fine diameter with better uniformity at lower
basis weight than the melt spun fibers can be produced by melt
blowing processes. Melt blowing technology was pioneered by
the Naval Research Laboratory to develop micro-denier fiber to
collect radioactive panicles in the upper atmosphere. This
technology was refined and licensed for commercial use by
Exxon Company. The process consists of extruding molten
polymer through spinneret orifices and fibrillating the extrudate
by high temperature and high velocity air streams. Fine
diameter, discontinuous fibers produced on the order of a few
inches in length are laid onto a moving belt equipped with a
suction box. The melt blown fibers have lower strength than
melt spun fiber and are often used in combination with other
strong fiber webs. In spite of small diameter fiber, the surface
area per unit weight is much lower than flash spun
plexifilamentary fibrils.
The high surface area of flash spun fiber results from
internal voids in the fibrils created by rupturing solvent vapor
globules within the polymer solution during the flash spinning
process. The flash spinning process involves decompression of
pure solvent droplets and highly saturated polymer/solvent
mixtures through a spin orifice. The fibrils become highly
oriented which accounts for their high strength. The fibrils
formed are of a micro-denier and are interconnected in a
continuous network described as a form of plexifilamentary
fiber structure. The fibril webs are collected on a moving belt
and are convened to a finished form using heat and pressure to
promote self-bonding. The immediate consequence of this fine
plexifilamentary fibril morphology includes high filtration
efficiency and good sheet tensile and tear strength.
Tyvek spunbonded olefin is flash spun from high-density
polyethylene and is used in medical packaging and apparel
markets for its microbial barrier and air permeability
properties. However, use of Tyvek as a filtration media has
been limited because of poor permeability and low din holding
capacity. Recently a new Tyvek product T-980 has been
developed specifically for filtration applications. A new
bonding process allows us to engineer structures to achieve
higher permeability by creating a greater number of submicron
pores and a smaller average pore size. An asymmetric structure
has been achieved which increases the din holding capacity and
filler life.
The selection of filter media based on cost/performance is
generally an empirical process that requires thorough testing
and evaluation. Rigorous prediction is difficult since many
variables interact such as fluid properties; contaminant type,
size and distribution; flow rate; and the objectives of the
filtration task. Thus, single-valued characterizations such as
nominal rating, absolute rating, and contaminant capacity are
not that useful in media selection (Ostreicher. 1986, refers to
cartridges). However, certain standards have been more or less
accepted by industry such that any new media must be
70
-------
evaluated by these so called "standard" tests. These can be
categorized into three areas: permeability, filtration efficiency,
and capacity (or life).
The purpose of this paper is to categorize a new filter media
developed by DuPont (Tyvek™) which we believe is unique
and exhibits interesting properties when evaluated by the
"standard" industry accepted tests. This new filter media of
Tyvek will also be compared to a typical microporous
membrane, a 5-micron "nominal" melt-blown PP
(Polypropylene) media, a 1-micron "nominal" melt-blown PP
media, and a PTFE (porytetrafluorethylene) membrane laminate
used extensively in industrial filters.
PERMEABILITY
The major deficiency with respect to filtration of the T-10
Tyvek that is used for envelopes, housewrap and disk sleeves is
insufficient air and water permeabilities which result in a tight
structure from micro-denier fibers spun by the flash spinning
technology. A new T-980 series Tyvek is developed by a new
bonding technology which increases the permeability by
creating greater numbers of submicron size pores and a smaller
average pore size. A scanning electron microscopy (SEM) of
T-980 Tyvek shows more open depth, finer fibers and an
asymmetric structure. Consequently, the new T-980 series
Tyvek exhibits much higher air and water flow rate than the
typical T-10 products (1042B); and, in fact, are comparable to
microporous membranes and PTFE membrane laminates and
begin to approach a typical melt-blown PP media, which is
known to have excellent permeability (Fig.l).
10»c-
r 10'-
10*
Figure 1: Air Flow Rale Comparison
FILTRATION EFFICIENCY
Table I lists typical media properties. The new Tyvek
bubble point (4 psa-which is a general indication of the largest
pore size) is substantially higher than the T-10 Tyvek grades
and begins to match the "0.8 micron" PTFE membrane
laminate. But, it is still much lower than the 0.45-micron
nominal rated membranes. Table II details some supplementary
latex bead challenge tests at 0.624 and 1.1-micron sizes and
one notes that:
• 0.624-micron and OJ-micron retention efficiencies range
between 93.0 and 99.6% and 92% and 99.4%, respectively
depending on basis weight, which are quite good
considering their 1-micron "nominal" 98% rating (for T-
980 at 1.1 micron and even higher efficiency for the higher
basis weight styles).
Retention efficiencies for the 1.1-micron beads are even
better (>99.0% for all grades), which indicates that the 1-
micron nominal rating is conservative and that single-
valued ratings are misleading.
MEDIA PROPERTIES
Media
Meh-Blown PP
Melt-Blown
Tyvek 1073B
Tyvek 1042B
New Tyvek T-980
New Tyvek T-980
(Hyekophilic)
PTFE Laminate
PTFE Laminate
Microporotu
Membrne
MilliporeHA
Norn.
Micron
Size(fi)
-5
1
1
2
1
1.
0.8
0.5
0.45
0.45
Prizier
Air
Plow
•CTM/ft1)
18
5
0.25
0.35
1.0
1.0
1.0
0.5
0.25
Bubble
PlinHjO
(PSI)
0.4
0.7
2.0
1.7
4
4
5
12
23
33
Water
Permeability
(GPM/h1
@ I PSI)
30
14
0.10
0.12
1.0
2.0
1.5
0.8
0.8
1.0
Style No.
Table I.
LATEX AND OOP CHALLENGE TESTS
New Tyvek
T-980 T-984 T-988 T-989
0.9
Baiu Weight
(otfytP)
% Retention 93.0
% Retention 99.0
% Retention 92.0
1.26
99.7
99. S
94.0
1.6
996
1.9
99.6 0.624
>999 1 1»
99.4 994 0.3"
I em/i«c nelyilvrcne Ittex tpbcra
•10 Ofeua OOP Mraol
Table n
Thus, these outside challenge tests confirm that new
Tyvek™ medias are quite efficient at low micron ratings, are
much tighter than melt-blown PP medias. compare favorably
to PTFE membrane laminates at similar permeabilities, and
begin to approach typical microporous membranes in
performance. Hence, the next logical step in our program was
to conduct full scale ACFTD challenge tests (as well as
comparable 47mm disc tests) to compare the various medias on
an actual commercial filler. The commercial filter chosen was
the Oberlin Pressure Filter (OPF) which is unique, is fully
automatic, can use both disposable media and recleanabte
ckxhs, and can be completely contained to prevent exposure to
toxic materials (Fig. 2). DuPont has used many of these
Oberlin units in a variety of applications, even in radioactive
service. This Oberlin unit is also ideally suited to use Tyvek
disposable media; and as a consequence, can achieve tight,
"nominal" 1-micron filtration automatically at very low cost
(more on this later).
71
-------
Wasttwiitr
FMd
Figure 2: Oberlin Pressure Filter
OBERLIN (AND 47MM DISC) ACFTD TESTING
All ACFTD testing was conducted at constant flux (0.75
but primarily at 1.0 gpm/sq ft) by either peristaltic pump
control to a Pall filterability kit or by air pressure control to a
Wilden M-2 double-diaphragm pump which feeds the OPF.
Filtered (0.2 micron Gelman). deionized water was used as
makeup and two ACFTD levels (60 and 600 ppm) were used
to represent typical wastewater treatment applications. Effluent
turbidities were measured with a Hach Model 2100A
turbidimeter on samples collected throughout the filtration life
cycle. Composite filtrate samples were collected in particle-free
bottles for subsequent panicle counting by both a Coulter
Counter and a Panicle Measuring System (PMS) laser counter.
Some samples were analyzed for total suspended solids (TSS)
by nitration through blanked Nuclepore 0.4 and 0.2-micron
membranes.
Fig. 3 shows that the Tyvek (0.9 oz/sq yd. Style T-980)
has effluent turbidities that are between the PTFE membrane
material and a typical 0.45-micron microporous membrane; and
much better than melt-blown PP medias.
Fig. 4 plots the Coulter Beta ratio against the ACFTD
micron rating (only for the 47mm disc tests since the OPF
tests were slightly corraminated by stray panicles even though
unit was thoroughly cleaned prior to use). Nevertheless. Fig. 4
100 e-
3
I
. 0.41 urn MenporM*
TVM (ffltaittl)
indicates that the new Tyvek again has higher Beta ratios (and
hence, higher efficiencies) than both the melt-blown PP and
the PTFE membrane media and approaches the 0.45-micron
membrane. Thus. ACFTD full-scale tests essentially
corroborate the outside tests in that Tyvek is much more
efficient than both melt-blown PP medias. compares favorably
with the PTFE membrane laminate, and begins to approach a
typical 0.45-micron microporous membrane.
100 c-
10
Ł 0.45 urn Mkraperaut
10
too
CaUMr B«U RiUo
10000
Figure 3: ACFTD Challenge Tests (60 PPM)
Figure 4: ACFTD Challenge Tests (60 PPM)
FILTRTATION CAPACITY (OR LIFE)
The third property of filter media that most end-users are
interested in is din capacity (or life). We conducted such an
evaluation by completing the ACFTD challenge tests (at both
60 and 600 ppm levels) to 30 psi differential AP (a nominal
value that is sufficient for most applications). These tests were
conducted on both the Pall 47mm disc filterability kit and the
full-scale Oberlin 2.4 sq ft HA filter. Table III summarizes the
ACFTD measured life of each of the medias evaluated here and
basically shows that:
The new Tyvek has about 24% the life of the 0.45-
micron microporous membrane at 60 ppm ACFTD; and
about 37% at 600 ppm ACFTD. This shorter life is
consistent with Tyvek's lower permeability and porosity
(ie, less depth penetration).
• HydrophiUc Tyvek has about 70% longer life than the
untreated Tyvek because of better wetting.
• Hydrophilic Tyvek compares favorably in life to the
PTFE membrane laminate, particularly since its cost is
only a fraction of the PTFE membrane laminate.
• The two melt-Wowns do not have substantially longer
life dun the microporous membrane; and their filtrate
quality it at best marginal.
Thus, Tyvek (both untreated and hydrophilic grades) has a
much shorter life than a 0.45-micron microporous membrane,
but its low cost helps » offset this penalty. Furthermore, cake
filtration is dominant in many cases such that media life
72
-------
becomes less important (ie, compare 60 to 600
and see applications testing below).
ppm results.
MEDIA FILTRATION CAPACITY BASED ON ACFTD CHALLENGES
60 PPM Tine To
30 PSI (Mini)
Norn.
Riling
Media
Microporoui
membrane
Ty»ek T-910
(Hydrophilic)
PTFE Membrane
Laminate
Melc- Blown PP
Melt-Blown PP
GO
0.45
1
o.s
1
_ r at
47MM
92
23
72
It
140
HA
67
15.5
52.3
-
Avg.
Floor
1.0
0.24
0.7S
0.95
1.50
600 PPM Time To
30 PSI (Mini)
47MM HA
12.0 17.5
6.5
10.0 12.0
13.5 20.0
1.45 21.3
Av|.
Factor
1.0
0.37
0.69
1.14
1 22
Table m
FILTRATION COSTS
Any filtration process chosen must meet one's objectives
as well as be the most economical. Generally speaking, most
processes are selected based on initial capital investment and
operating costs are usually ignored. Table IV highlights some
typical operating costs based on the work here. Basically, new
Tyvek is far more economical than any of the other
competitive medias for four typical applications (based on
cents per gallon filtered). In fact, its cost is a fraction of the
microporous membrane cost at only a slight deterioration in
filtrate quality. And, it is also cost-competitive to low-cost
melt-blown PP medias, but produces far better filtrate quality.
Tyvek also exhibits superior strength and cake release, which
are significant advantages in Oberiin filter operation. Thus,
Tyvek provides economical, low-micron filtration comparable
in many instances to microporous membranes and PTFE
membrane laminates.
MEDIA COST PEK GALLON WASTE FILTERED
Media
Microporoui
Tyvek T-9W
Tyvek T-9tO
(Hydropbilic)
PTFEManbnM
Mek-BtowBPP
Mrt-BtowePP
Notn.
Rattaf
00
0.45
i
I
o.t
1.0
-5
General
FUnt
Quality
V Good
V Good
V. Good
Good
Peer
10 PPM
ACPTD
6.7
0.5
0.3
(.2
0.7
0.3
COO PPM
AdTD
29
1.3
O.t
41
2.4
1.5
•»* i.ceowgaa>-
Sua
SPP Vatey
69 14
1.7 0.3
1.0 0.2
63 12
65 13
4.0 01
COM. (ha raMa, aad meeeured life cyck».
Table IV
APPLICATIONS TESTING
Table V summarizes a series of Oberiin bench-scale tests
with simulated Savannah River M-Area wastewater (to
duplicate plating line effluents) and the various media evaluated
above. Basically, the new Tyvek (hydrophilic) matches all
competitive medias in rate and filtrate turbidity, but produces a
drier cake because of better air-blowing (ie, lower final air-blow
pressure since higher air flow due to less media blinding). As a
result, cake release is excellent and residual media weight gain
is reduced (IS mg vs > SO mg for others).
SAVANNAH RIVER M-AREA STMIULATED WASTE
CHALLENGE TESTS IN OBERUN PRESSURE FILTER
Media**
Fowl
TVpe
Mkropoftxat
jjtjLji-J-_rjj|j
miillll BUT
Tyvek T-9W
Tyvek T-9W
(Hydrophilic)
PTFE Membrane
| a»lgng«M-
Melt-Blown PP
Melt-Blown PP
Norn.
Rating
00
0.45
I
1
O.S
1.0
-5
AWL
CUD
(Bf)
SO
10
15
60
560
445
Cake
CUe %
Rfheir Solid*
Good
Excellent
Excellent
Fa»
Fair
Good
30
34
34
28
32
29
FflBite Avg. Cake*"
Tinbtdity Rue An- Blow
(NTU) (jptn/ft2) Press
-------
CASE HISTORIES SAVANNAH RIVER PLANT
TyvckT-MO*
OMtoAPF.
Fnanol K-144L
MH Pobmo Ptritow 30 NPDES
•Turbidity
•TSSfopn)
•Ahmuwn (ppm)
•Laid (ppm)
•Urmium (ppm)
•Zinc (ppm)
•Copper (ppm)
no
«7
127
2.'3
0.3
02
0.32
1.4
0.95
0.2
0.01
<0, 1
<0. 1
31
3.2
0.43
0.5
0.32
0.21
Fhul200GPD
TabteVII
ACKNOWLEDGEMENTS
The authors wish to express their thanks to the following
people who contributed immensely in the data collection on
filtration tests and preparation of this paper: J.G. Wood,
DuPont Engineering Test Center and Dr. T. Oberlin, President
of Oberlin Filter Company.
LITERATURE CITED
Martin, Hi.. Paper presented at 10th Annual AESF/EPA
Conference on Environment Control for the Metal
Finishing Industry, Orlando, FL, Jan., 1989.
Ostreicher, Fluid Filtration: Liquid. Vol. II. ASTM STP 975,
Johnston, P.R. and H.G. Schroeder, eds, Amer. Soc. for
Test Mat., Philadelphia (1986).
74
-------
Chemisches Laboratorium Dr. E. WeBling
Oststra(3e2 • 4417 Altenberge • Telefon 02505/89-0
Cleanup Of Contaminated Soil By Ozone Treatment
1. Process Description
Volatile aliphatic and aromatic hydrocarbons are easily removed with
high efficiency from contaminated sites by soil gas sucking. This
method, however, fails when the soil is polluted with less volatile
substances like mineral oil or polycyclic aromatic hydrocarbons. In
these cases ozone treatment is a suitable technique to reduce the
pollutant concentration drastically.
Ozone, the most powerful technically applied oxidizing agent, is able
to destroy hydrocarbons when a gas stream enriched with ozone is
passed through the contaminated soil. The pollution may be reduced by
up to 98 % of the original concentration but the efficiency depends on
several parameters, e.g. the nature of the pollutants, the condition of
the soil, especially its permeability to gas, and the presence of
accompanying substances (humic material). Ozone treatment may be
carried out in-situ as well as on-site and off-site.
2. Experimental Results And Development Status
Experiments have been carried out so far in laboratory and pilot-plant-
scale with treated soil masses of 2 kilograms to 3 tons. In laboratory
scale the contaminated soil was placed in a cylindric glass reactor.
Barrels and skips were used in larger scale treatments. An ozone
containing gas stream was prepared from compressed air or oxygen in
an ozone generator by electrical discharge.
75
-------
Figures 1 and 2 show typical plots of pollutant concentration versus
treatment time received in laboratory experiments.
origin: tank farm
pollutant: PAH
•oil type: sand
treated quantity: 2 kilograms
specific ozone demand: 2.9 mg/mg PAH
fig 1: treatment of PAH contaminated soil
*
During 20 days the pollution of a soil derived from a tank farm con-
taminated with polycyclic aromatic hydrocarbons was reduced from about
2 300 mg/kg to 50 mg/kg (fig. 1), which is a reduction by nearly 98 %.
Similar results were obtained with a mineral oil containing soil from a
refining plant (figure 2).
76
-------
origin: refining plont
pollutant: mineral oil
eoil type: eand
treated quantity: 2 kilograms
•pectfic ozone demand: 3.5 mg/mg oil
time (h)
fig 2: treatment of mineral oil contaminated soil
By ozone treatment the concentration of hydrocarbons decreased from
12 000 mg/kg to 150 mg/kg within 2,5 days. Experiments in pilot-plant-
scale gave similar results.
The first in-situ-treatment project at a former gasoline station was
now started. Figure 3 shows the corresponding flow diagram. Pollutants
are mineral oil and polycyclic aromatic hydrocarbons in this case.
77
-------
gasproof barrier
ft
central control equipment 1
t
k >
\ *
1 emissic
1 active carbon filter I
control I-
pment I
*
1
1 emission contro
fl
r
1
r^
HL
measurement of ozone r"
concentration and flow |—
r injection weU 1
. vacuum well I
fig 3: in-situ-treatment, flow diagram
After sucking off all volatile hydrocarbons ozone treatment starts. By
using multilayer injection and vacuum wells it is possible to direct
the ozone containing gas stream. Volatile pollutants are adsorbed on
active carbon filters, which also destroy excess ozone. A central
equipment controls the ozone concentration, gas flow and ozone emission.
78
-------
3. Oxidation Products And Influence On The Soil Microflora
It is known that organic compounds are oxidized by ozone to alcohols,
aldehydes, ketones and carboxylic acids depending on their molecular
structur and reaction time. All these substances, however, are highly
biodegradable.
Oxidation products of selected polycyclic aromatic hydrocarbons have
been studied in more detail in our laboratory. Primary degradation
products are those, which have been expected by mechanistic con-
siderations. These substances are transferred into the main products of
ozone treatment, oxalic acid, carbondioxide and water (figure 4).
PAH > H02C - C02H > C02 + H20
fig H: oxidation products of polycyclic hydrocarbons
There is no indication for the formation of any toxic degradation
products.
During ozone treatment soil microorganisms are destroyed to a large
extent. A restauration of the soil microflora can, however, easily be
achieved for example by addition of the effluent of a sewage plant.
Treatment Costs
A considerable amount of the total treatment costs is needed for ozone
preparation by electrical discharge. In the Federal Republic of Germany
these costs are today estimated at 50 - 300,- DM (30 - 175 US-Dollar)
per ton of treated soil depending on the pollutant concentration. This
amount does not include installation costs and costs for running the
equipment as well as analytical control.
79
-------
BlOTROC
... for environmental solutions, naturally!
BioTrol® Soil Washing System
by
Steven B. Valine
Dennis D. Chilcote
Thomas J. Chresand
BioTrol, Inc.
Presented at
U. S. Environmental Protection Agency
Second Forum on Innovative Hazardous Waste Treatment
Technologies: Domestic and International
Philadelphia, Pennsylvania
May 15-17, 1990
ABSTRACT
Soil washing is an innovative, volume reduction process for
treating contaminated soil. In the process, soil is slurried with
water and subjected to several stages of intensive scrubbing with
interstage classification. Separation of the washed, coarser soil
components from the process water and contaminated fine particles
(silt and clay fraction) results in a significant reduction in the
volume of material requiring additional treatment or disposal.
A pilot-scale soil washing system with a treatment capacity
of 500 to 1000 pounds per hour was operated at a Superfund site in
Minnesota contaminated with wood preserving wastes, including
pentachlorophenol, polyaromatic hydrocarbons, petroleum hydro-
carbons, copper, chromium, and arsenic. The results of this on-
site pilot work are discussed, including preliminary results
obtained during an EPA Superfund Innovative Technology Evaluation
(SITE) demonstration. The results of bench-scale treatability
studies covering a variety of contaminants are presented, as well
as estimated costs for full-scale treatment.
BioTrol, Inc -11 Peavey Road • Chaska, Minnesota 55318 • 612/448-2515 • FAX 612/448-6050
80
-------
INTRODUCTION
Basic Principles
Soil washing is an innovative volume reduction technology for
the remediation of contaminated soils. The effectiveness of soil
washing relies upon two principal observations. First, contam-
inants tend to be concentrated in the fine size fraction (silt and
clay) of soil. Second, contamination associated with the coarse
size fraction (sand and gravel) is primarily surficial.
In the process, soil is mixed with water and subjected to
intense attrition scrubbing. This scrubbing action disintegrates
soil aggregates, freeing contaminated fine particles from coarser
sand and gravel. Contaminants may also be solubilized as dictated
by applicable solubility characteristics or partition coefficients.
Concurrently, the coarser soil particles are scoured by the
abrasive action of the coarse particles themselves, transferring
surficial contamination to the fine soil fraction. The process
water containing the contaminated fine soil material is then
separated from the washed coarse soil. These three mechanisms:
disintegration of soil aggregates, dissolution of contaminants, and
scouring of coarse particle surfaces, each operate to varying
degrees, depending upon the characteristics of the soil and contam-
inant (s) .
To improve the efficiency of soil washing, the process may
include the use of surfactants, detergents, chelating agents, pH
adjustment, or heat. In many cases however, water alone is
sufficient to achieve the desired level of contaminant removal
while minimizing cost.
Process Description
A simplified flowsheet of the BioTrol® Soil Washing System is
shown in Figure l. A multiplicity of unit operations are involved,
most of which are well known in the mineral processing industry.
Following excavation, debris is removed from the soil by
screening at 1/2 to 2 inches. Now free of large debris, the soil
is slurried with water in a mixing trommel or pug mill, depending
upon the nature of the soil and the amount of energy required to
achieve dispersion of the soil in water. Next, the soil slurry is
wet screened at approximately 1 to 2 mm to remove oversize
material, such as small rocks and pebbles. This oversize product
can be further washed and/or segregated, if necessary.
The undersize product from wet screening is fed to a froth
flotation circuit where hydrophobic contaminants (e.g., hydro-
carbons) and/or contaminated fine solids are transferred to a froth
81
-------
FIGURE 1
SIMPLIFIED FLOWSHEET
BIOTROL SOIL WASHING SYSTEM
EXCAVATE
FEED SOIL
SCREEN
SLURRY
SCREEN
FROTH
FLOTATION
MULTI-STAGE
COUNTERCURRENT
ATTRITION/
CLASSIFICATION
CIRCUIT
DEWATERING
WASHED SOIL
MAKE-UP
WATER
DEBRIS
RECYCLE
WATER TREATMENT
SYSTEM
WASHED
OVERSIZE
PROCESS
WATER
FROTH
FINE SILTS,
CLAYS,AND
ORGANICS
THICKENING
CONTAMINATED \
FINE |
SOLIDS /
DEWATERING
RESIDUALS
MANAGEMENT
82
-------
phase and removed. Flotation underflow then enters an intensive,
multi-stage, countercurrent attrition scrubbing circuit with
interstage classification. Each attrition scrubbing cell is
equipped with two or more propeller-type impellers of opposing
pitch to achieve the desired scrubbing intensity. Hydrocyclones or
spiral classifiers are used to achieve classification following
each scrubbing stage. The soil exiting the scrubbing circuit is
dewatered in a spiral classifier or vacuum belt filter and
discharged as the washed product.
Flotation froth and the fine soil particles suspended in the
process water from the scrubbing circuit are fed to a thickening
operation. During thickening, a polymeric flocculant is normally
used to aid in the settling and separation of the fine solids from
the process water. Following thickening, the fine solids can be
further dewatered using a belt press or centrifuge. Treatment
alternatives for the contaminated fine solids could include
incineration, biological degradation, stabilization, on-site
containment, or off-site disposal.
The clarified process water exiting the dewatering operation
is treated, if required, and recycled back to the soil washing
system to minimize water consumption. Make-up water is necessary,
since the products exiting the soil washing system usually contain
a higher level of moisture than the feed soil. The degree of
process water recycle, and therefore the make-up water requirement,
is dependent upon the characteristics of the soil and contamin-
ant (s) .
LABORATORY TREATABILITY TESTING
A laboratory treatability protocol has been developed to
assess the performance of soil washing on contaminated soil
samples. Testing is normally conducted on two samples from a site,
one that represents "average" site conditions, and one that
typifies "worst" conditions.
Samples are screened to remove debris, blended, and then
divided into representative aliquots for subsequent tests.
Chemical, particle size, and optionally, mineralogical analyses are
conducted to characterize the samples prior to soil washing tests.
Leaching studies determine the degree to which contaminants
partition between the soil and leaching solution. Data are
generated from a series of low intensity agitation tests covering
a range of solids concentrations. Studies can be designed to
compare the effect of physical or chemical "additives" on
contaminant solubility with baseline data from tests with water
alone.
83
-------
Froth flotation and attrition scrubbing tests are accomplished
with a bench-scale flotation machine. This type of apparatus has
been in use for several decades in the mineral processing industry
and can be equipped with either a flotation-type or an attrition
scrubbing impeller. Design parameters used in scale-up from this
"bench" apparatus to full-scale are well established.
Results of treatability testing with various client soil
samples are shown in Table 1. For organics, levels of contaminant
TABLE 1. RESULTS OF BENCH-SCALE TREATABILITY TESTING
mg/kg
Site Description
Wood Preserving
(California)
Wood Preserving
(Florida)
Chemical Plant
(Michigan)
Wire Drawing
(New Jersey)
Parameter
Total PAH (1)
Arsenic
Chromium
Copper
Zinc
Pentachlorophenol
Pentachloropheno 1
Dichlorobenz idine
Benzidine
Azobenzene
TPH (2)
VOC (3)
Copper
Nickel
Silver
Feed
Soil
4,800
89
63
23
345
380
610
770
1,000
2,400
4,700
2
330
110
25
Washed
Soil
230
27
23
13
108
4.0
25
13
6
7
350
0.01
100
60
4
Percent
Reduction
95
70
63
43
69
99
96
98
99
>99
93
>99
70
45
84
Town Gas
(Quebec)
Pesticide
Formulation
(Colorado)
Chemical Plant
(California)
Total PAH (1) 230 11
Chlordane 55 4.7
Aldrin 47 7.5
4,4-DDT 25 5.0
Dieldrin 46 7.0
PCB (4) 290 <0.1
(Aroclor 1260)
95
91
84
80
85
>99
Notes: (1) Polynuclear aromatic hydrocarbons
(2) Total petroleum hydrocarbons
(3) Volatile organic compounds
(4) Polychlorinated biphenyls
84
-------
reduction between the feed and washed soil generally ranged from 80
to greater than 99 percent. Reductions in the levels of metals
ranged from 45 to 84 percent. The type of metallic compound, i.e.,
oxide, insoluble salt, etc., will greatly impact the degree to
which metals can be removed from a soil. Removal efficiency could
be significantly improved by employing common hydrometallurgical
techniques. The data shown are insufficient to draw any
conclusions on the effectiveness of soil washing on metals removal.
PILOT-SCALE TESTING
A mobile, pilot-scale soil washing system with a nominal
treatment capacity of 500 to 1000 pounds per hour was constructed
for use in process development and for conducting on-site
demonstrations. The system consists of pilot-scale mineral
processing equipment assembled on a 42 foot long drop-down semi-
trailer. Included in the pilot plant system are a multi-component
automatic feed system (including a conveyer belt scale and
totalizer), a mixing trommel, vibrating screens, froth flotation
cells, attrition machines, spiral classifiers, hydrocyclones, and
dewatering equipment.
Clarified process water from the soil washing system is
treated prior to recycle. For biodegradable compounds, a pilot-
scale, fixed-film, biological treatment system is used. The bio-
reactor, contained in a 20 foot long, enclosed trailer, is a multi-
cell, submerged, aerated, packed bed reactor and uses an amended
bacterial consortium. Each cell is filled with a packing material,
usually structured PVC trickling filter media, which serves as the
substrate for microbial attachment. Porous diffuser pipes mounted
beneath the packing support grid provide aeration. The packing
volume is 72 cubic feet, or 540 gallons. Flow capacity can be up
to 10 gallons per minute.
From 1987 to 1989, extensive process development and field
pilot testing of the soil washing system was conducted at the
MacGillis and Gibbs Company site in New Brighton, Minnesota. This
Superfund site is contaminated with wood preserving wastes,
including pentachlorophenol, creosote (polynuclear aromatic
hydrocarbons) , petroleum hydrocarbons, and to a lesser degree, with
arsenic, chromium, and copper. Only water (no additives) was used
in the test program.
To illustrate the type of results achieved during pilot
testing, a typical mass balance for pentachlorophenol is presented
in Table 2. Chemical analyses and distribution are shown for both
the dry solid and aqueous phases. Note that 47 percent of the
contaminant was solubilized in the aqueous phase of the fine
particle slurry. The washed soil, containing 34 mg/kg penta-
chlorophenol, accounted for 75 percent of the soil fed to the
85
-------
system on a dry weight basis. On a mass basis, 94 percent of the
contaminant was removed from the washed soil.
TABLE 2. PENTACHLOROPHENOL MASS BALANCE
PILOT-SCALE TESTING AT MINNESOTA WOOD PRESERVING SITE
Dry Cone., mg/kg Distribution, percent
Solids
Weight, Dry Dry
Process Stream Percent Solids Water Solids Water Total
Feed Soil 100 680 100 100
Washed Soil 75 34 20 5.0 1.0 6.0
Oversize 12 1000 82 24 3.0 27
(Mostly Wood)
Fines Slurry 13 820 72 20 47 67
PRELIMINARY EPA SITE PROGRAM RESULTS
The U. S. Environmental Protection Agency conducted an
evaluation of BioTrol's soil washing technology in the Superfund
Innovative Technology Evaluation (SITE) program during September
and October of 1989. The MacGillis and Gibbs wood preserving site
was chosen as the location for the test.
The capability of the system was evaluated by testing with two
types of soil from the site. The first sample was excavated from
the vicinity of a former pole treatment tank and designated as the
"low" contaminant level soil. A second sample, designated as the
"high" contaminant level soil, was excavated from a "hot spot" in
the site disposal area. Tests with the "low" soil lasted two days
while tests with the "high" soil lasted about 6 days. The soil
washing system operated 24 hours per day for both tests. As
before, only water was used (no additives).
All product streams exiting the system were collected in 55
gallon drums and weighed to determine mass flows. Feed rate was
determined by recording conveyer belt totalizer readings. Grab
samples of all input and output streams were taken at roughly 2
hour intervals; individual grab samples were then composited to
represent about 6 hours of operation.
86
-------
Preliminary results of the demonstration are given in Table 3.
As shown, pentachlorophenol levels were reduced 90 to 91 percent in
the washed soil compared to the feed soil. Reductions in the level
of total polynuclear aromatic hydrocarbons ranged from 92 to 96
percent while reductions in total petroleum hydrocarbon levels
ranged from 94 to 95 percent.
TABLE 3. COMPARISON OF FEED AND WASHED SOIL
PRELIMINARY RESULTS OF EPA SITE PROGRAM
PILOT-SCALE TESTING AT MINNESOTA WOOD TREATING SITE
mg/kg
5011
Contaminant
Level
Low
(Test 1 of 1)
Parameter
Pentachlorophenol
Total PAH (1)
TPH (2)
Arsenic
Chromium
Copper
Feed
Soil
130
240
3,300
14
17
15
Washed
Soil
12
8.6
210
5.0
9.0
6.2
Percent
Reduction
91
96
94
64
47
59
High
(Test 1 of 2)
Pentachlorophenol
Total PAH (1)
TPH (2)
Arsenic
Chromium
Copper
540
290
8,800
28
49
39
56
23
470
7.2
8.5
5.2
90
92
95
74
83
87
Notes: (I) Polynuclear aromatic hydrocarbons
(2) Total petroleum hydrocarbons
Preliminary Toxicity Characteristic Leaching Procedure (TCLP)
results for pentachlorophenol are presented in Table 4. With both
soil types, the leachate solutions contained less than 1 mg/L
pentachlorophenol. Since the system is a water-based, intensive
scrubbing process, leachability (TCLP) of the washed product will,
in most cases, be very low. This fact should allow washed soil to
be placed back on-site.
Test results demonstrating the effectiveness of the biological
water treatment system on the process water from the soil washing
system are illustrated in Figure 2. For days 1 through 3, the
system treated water generated from the "low" soil test. Water
from the "high" test was fed to the unit on days 4 through 10.
87
-------
TABLE 4. RESULTS OF TCLP TESTS
COMPARISON OP PENTACHLOROPHENOL IN WASHED SOIL AND LEACHATE
PRELIMINARY RESULTS OF EPA SITE PROGRAM
PILOT-SCALE TESTING AT MINNESOTA WOOD PRESERVING SITE
Soil
Contaminant
Level
Pentachlorophenol
Washed Soil,
mg/kg
Washed Soil TCLP
Leachate, ntg/L
Low (Test 1 of 1)
High (Test 1 of 2)
10
19
59
70
0.23
0.32
0.74
0.92
50
40
30
20
10
FIGURE 2. BIOLOGICAL TREATMENT OF
PROCESS WATER FROM SOIL WASHING
(PRELIMINARY DATA - EPA SITE PROGRAM)
Pentachlorophenol Concentration (ppm)
Influent
012345678
Day
10 11
-------
COST
Estimated treatment costs for a mobile, commercial-scale, 20
ton per hour soil washing system are shown graphically in Figure 3
as a function of tons processed. Costs include capital recovery
(charged as an equipment leasing rate) and water treatment; not
included are costs for excavation, debris removal, and treatment or
disposal of residuals generated during treatment.
Total cost per ton is the sum of the mobilization and treat-
ment cost components. The estimated treatment cost of roughly $60
per ton varies only slightly with tons processed. Mobilization
cost per ton has the most impact at relatively small soil volumes
(less than 3,000 to 4,000 tons).
For illustration, from Figure 3, the estimated soil washing
cost for 20,000 tons of soil would be approximately $71 per ton or
$1.42 million. Costs for treatment or disposal of the residuals
generated during soil washing must also be calculated to estimate
total remediation costs. For example, soil washing of 20,000 tons
of soil with the characteristics shown in Table 2 would generate
2,400 tons of oversize and 2,600 tons of fines, both requiring
additional treatment. Using incineration at $200 per ton for the
oversize and slurry-phase biodegradation for the fines at $100 per
ton, treatment of residuals would cost an additional $740,000 or
$37 per ton based on 20,000 tons. Total soil treatment costs,
including treatment of residuals, is therefore estimated at $108
per ton or $2.16 million.
Soil washing unit costs will be significantly lower for
relatively large soil volumes (greater than 100,000 to 250,00 tons)
or for a fixed facility with an expected operating life of greater
than 5 years. Larger throughput capacities and use of longer
periods for capital recovery would lower total treatment costs to
the range of $25 to $40 per ton.
89
-------
120
FIGURE 3. ESTIMATED TREATMENT COST FOR
20 TON PER HOUR SOIL WASHING SYSTEM
(TYPICAL)
COST PER TON, $
100 -
80
60
40
20
0
. Total Cost (Treatment + Mobilization)
Treatment Cost
„ Mobilization Cost
10 20 30 40 50
TONS PROCESSED (Thousands)
(Cost excludes excavation, debris
removal, or disposal of residuals.)
60
CONCLUSION
Laboratory treatability studies and on-site pilot-scale
testing have demonstrated the effectiveness of the BioTrol® Soil
Washing System. Removal of organic contaminants typically ranged
from 90 to greater than 99 percent. Removal of metals was somewhat
lower, but could be improved through the use of hydrometallurgical
techniques.
Since the system is a water-based, intensive scrubbing
process, leachability (TCLP) of the washed product will, in most
cases, be very low. This fact should allow washed soil to be
placed back on-site without endangering human health or the
environment. As a cost effective, volume reduction step, soil
washing will play a significant role in the remediation of
contaminated soil.
90
-------
INTEGRATED SOIL-VAPOR/GROUNDWATER CLEANING SYSTEM
AT
SELECTED SITES IN WEST GERMANY
Authors
Karlheinz Bbhm, Ph.D.
and
Gerrit Rost
Department of Environmental Engineering
Ed. Ztiblin AG
Albstadtweg 3
7000 Stuttgart 80
Federal Republic of Germany
and
Martin R. Griek
ZDI Construction Services
607 Herndon Pkwy.
Suite 100
Herndon, VA 22070
Presented at
The Second Forum on
Innovative Hazardous Waste Technologies:
Domestic and International
Philadelphia, Pennsylvania
May 15 - 17, 1990
91
-------
INTEGRATED SOIL-VAPOR/GROUNDWATER CLEANING SYSTEM
I. INTRODUCTION
Ed. Ziiblin AG is an international civil engineering firm headquartered in
Stuttgart, West Germany. It is represented in the U.S. by its operating
subsidiary, ZDI Construction Services, Herndon, Virginia.
Ed. Zu'blin has provided the European and International markets with
construction related services for over 92 years. For the past decade their
Department of Environmental Engineering has been involved in the treatment
and containment of hazardous material. These areas include:
Asbestos - removal and disposal
Hazardous waste site containment - HOPE lined slurry walls
Thermal incineration of contaminated soil with rotary kilns
Integrated soil-vapor/groundwater cleaning systems
Vacuum extraction of soil-vapor is not a unique process. It is, in fact, the
standard method in Europe today for cleaning up sites contaminated with
VOC's. [Chlorinated Hydrocarbons (Solvents TCE-PCE-DCE) and Aromatics
(Petroleum Products - BTEX)]
Combining a soil-vapor extraction and treatment system with a groundwater
treatment system for the remediation of a VOC's contaminated site is a
sensible decision. The vadose zone can be remediated by vacuum extraction
with relative ease and economy. This process removes the source of the
groundwater contamination. Thus the groundwater only needs to be cleaned of
residual contamination. Removal of the source pollution shortens the
duration of the treatment process since percolation of volatile substances
through the soil horizons into the groundwater can take decades. The costs
and effectiveness of remediation can be vastly improved by using the combined
phase treatment process.
This paper presents two case studies utilizing Zliblin's integrated treatment
process incorporating this procedure. Zublin has installed and operated
dozens of these systems during the past decade. With typical German
efficiency, the company designed a highly flexible, efficient and functional
system to recover solvent contaminants and minimize discharge into the
biosphere.
II. SCHERWIESEN SITE
This study deals with a landfill, for industrial sludges, animal processing,
and household waste which had been operating since 1967. Following its
closure in 1980, the landfill was excavated to a depth of three (3) feet and
capped with a one meter thick compacted clay seal.
Engineering studies performed in early 1984, showed that hydroxide sludges
and chlorinated hydrocarbons were present throughout the site and had started
to contaminate the groundwater. The contamination plume had extended over
650 meters (2,100 feet from the site). At this point it had contaminated a
natural spring which was the primary drinking water source of a nearby
village. Remediation of the site was ordered by the appropriate authorities.
92
-------
In 1987 Zublin's Department of Environmental Engineering was awarded the
project to extract and treat both the soil-vapor and the contaminated water
in the perched water table (source contaminants) as well as the contaminated
spring water. Starting with commencement of the operation, a study was
conducted to record the effects of the treatment process. The data collected
during this study is presented herein. The plant is still in operation
today. Figure 1 is a schematic diagram of the treatment system. A site plan
of the project is shown in Figure 9.
III. TREATMENT PROCESS
A. Soil-Vapor
Initially, ten (10) soil-vapor extraction wells were constructed with varying
depths of 13 to 17 feet at the points of highest contaminant concentrations
at the site. A set of vacuum pumps extracted 250 cubic meters per hour from
a five (5) well manifold with a negative pressure of approximately 200 mbar.
The manifold was shifted to different wellheads when the vapor concentration
decreased. The soil at the site is a heavy clay. Its zone of influence for
each soil-vapor well was only 10 to 16 feet. With a permeability of the soil
of 10"6 cm/s, the need to drill additional wells was apparent. As a result,
in 1988, a year later, twenty (20) additional soil-vapor extraction wells
were installed.
The extracted soil-vapor passes through a water separator which removes the
water particles picked-up by the vacuum, along with the soil-vapor, from the
soil. The water from the separator, which may contain trace contaminants,
mixes with and is cleaned in the water treatment phase of the system. The
vapor, which has been heated by the energy produced by the vacuum pump, is
cooled prior to entering the activated carbon filter for maximum adsorption
by the carbon.
Two carbon filters are used: As one filters the influent air the other is
either on stand-by or being regenerated. Prior to the on-line filter
reaching break-through, the order is reversed. In this fashion the treatment
continues uninterrupted without the need to exchange and dispose of spent
carbon filter material.
During the regeneration process, steam desorbes the solvent from the
activated carbon. The contaminant laden steam is then condensed.
Subsequently, the solvent and water are separated resulting in recovery of
the solvent. This solvent is 99% pure and can then be recycled or properly
disposed. The contact water is fed back into the liquid phase of the system
for cleaning. During the first year over 3,000 kg of solvent was recovered
at this site.
93
-------
Contaminant concentrations of the soil-vapor were reduced from 8,000 ppm to
between 30 - 50 ppm. Figure 3 shows the influent concentrations of the soil-
vapor. Air volume was measured with an integrated flywheel anemometer. The
total volume extracted and treated over this period of time exceeded 2.5
million cubic meters. This is graphically demonstrated in Figure 4. The air
discharge concentrations were well below the German standards of 100 ug/1.
Figures 5 and 6 show the air discharge concentrations and quantity of air
treated. Figures 7 and 8 illustrate quantities of solvent recovered by the
plant.
As can be seen in looking at these curves there are instances of several
upward shifts. The primary cause for these shifts are two. The first is the
result of climatic conditions. As the clay dries, cracking occurs: These
cracks provide paths for the extraction of ambient air and lower the
concentration measured in the influent air stream. Further evidence of this
occurrence is the "decrease" in the negative pressure measured at the vacuum
pump.
The second reason for a shift is caused by interruption of the extraction
process at one or more of the wells. The phenomena of increased contaminant
concentration in the extracted soil-vapor after a shut-off well resumes
extraction is common. During this shut-off period the volatiles have a
chance to re-establish equilibrium in the "vacuumed" section of the soil.
That is, the volatiles inside the soil have a chance to diffuse into the
soil-air space. Consequently, concentrations following start-up are higher
than those recorded prior to shut down.
Other minor factors affecting the peaks are due to: shut downs of the plant
which occurred in December 1987, and June 1989 depicted on Figure 6; and,
precipitation - since a partial saturation of the vadose zone impedes removal
of the VOC's. The important point of the graphs, however, is to note the
trend of decreasing concentration in the contaminants.
B. Water
In order to expedite remediation of the site, a groundwater extraction and
treatment phase was implemented in conjunction with the soil-vapor treatment.
Five (5) of the original ten (10) wells were used to extract the perched
groundwater. A shallow ditch was constructed encompassing the site to
capture and treat the surface run-off.
The perched water is part of the aquifer that feeds the spring. It is
greatly influenced by precipitation as shown in Figure 2. The perched water
with a flow rate of 200 liters per hour it is pumped into a storage tank.
From there it is fed through a sand filter into two activated carbon filters
operating in series. The initial concentration of the water measured at the
plant was 23,500 ug/1 with point measurements at the monitoring well (P-12)
as high as 166,000 ug/1 of chlorinated hydrocarbons. Following treatment,
effluent concentrations of the discharged water were well within the German
drinking water standards of total CMC's less than 20 ug/1•
It should be noted that after constant and continuous remediation of 2-1/2
years, the contaminant concentration of the perched water was reduced to 1.9
mg/1.
94
-------
IV. ECONOMIC ANALYSIS
A cost analysis of the plant follows: Costs are 1n 1987 dollars at a
conversion ratio of 1.6 DM to $1.00 US.
Site Operations DM US
1. Soil-Vapor System
a. Installation 17.025 DM $10,640
b. Operation and Maintenance (12 mos.) 5.235 3,275
2. Solvent Recovery System
a. Installation 52.200 32,625
b. Operation and Maintenance (12 mos.) 18.070 11,295
3. GAC System
a. Installation 13.500 8,440
b. Operation and Maintenance (12 mos.) 2.160 1,350
Total 108.190 DM $67,625
Installation Costs
1. Site
a. Soil-Vapor System 17.025 DM $10,640
b. Solvent Recovery System 52.200 32,625
c. GAC System 13.500 8,440
Total 82.725 DM $51,705
Operation and Maintenance
1. Site
a. Soil-Vapor 5.230 DM $ 3,270
b. Solvent Recovery 18.070 11,295
c. GAC 2.160 1.350
Total 25.460 DM $15,915
These Costs Include:
1. Site preparation for the plant
2. Mechanical and electrical installation charges
3. Labor, material and equipment
4. Operating supplies, utility costs, and maintenance
These Costs Exclude:
1. Well installation
2. Permitting costs
3. Replacement charges (depreciation)
4. Carbon replacement - liquid phase
5. Engineering charges for the sampling/testing study
95
-------
conlominaled
water
condensolor
ii Ii tin n
conk
air
II
no
0
r|
SI
ted
r
DCK
^L_
***•
**~*
^
1
i
!
i
:
•n —
---
\
-r-
\
•
•
1
1
IL
V
^
-~-~
; |
I
IT
--
—I/
j||
*
nJn
1- •' '•; activated
• . ! I • carbon filler
fZ^J 'or a'r
:--jj y
! •
* ".r.r<"VJL..""fi*ff"**r""j*"".T.i".'."i"
Sl(
r water -seporalor
— 1
sand filler
sleam generator
pump
sloragelank
Irealed waler
waler softener
tundensale solvents
disposal
water
i": oir
v/.v.v.: sleam
SOIL-AIR AND WATER TREATMENT - SYSTEM DIAGRAM
ZUBLIN
Figure 1. Schematic of the soil - air and water treatment system
-------
Ł.
X
1.-:
l.ii
I.-I
i.::
0,6 -
0,4 H
0.2 -
0
O±s=»i>osts.."JL Si
Wntcr quantity
'4
fi
Wf
G/87 9/87 12/87 3/88 6/88 9/88
Date
12/88 3/89
6/89
9/89 12/89
Figure 2. Quantity oT perched wal.or extracted and treated by the plant.
-------
86
CHC concentration (ppm)
(thousand)
n
3
O
0
~s
f
9
n
o c
O *-- N
o o
01 --
00
~J
so
CO
CO
CO
CO
CO
CD
CO
3
r*
!3
CO
CO
CO
35
co •
CSJ
30
13
' | '
Li fJ: --I Ct1
J_._l L_
J
O
3
o
0
3
3
C8
O
3
w
0
H-
(5
H-
ffl
P
H
W
H-
- (D
-------
66
Soil vapor (cbm)
(million}
13
O
i
a
CO
o
o
**
'")
o
c
3
3
9
r-
w
o
0
1
fO
X
o
f
**•
o
-------
S:L tea
EffluonL air concontrntions
100
o
o
o
a:
o
p.
o.
6/87 9/87 12/87 3/88 0/88 9/88 12/88 3/89 G/89 9/89 12/89
Figure 5. Efriuont air conccnl.ral.iotis of Lhc soiJ vapor exl.rnc:t i on system.
-------
. .1. S i. 1;. e»
Soil vapor quantity treated
JC
X
K-. .0
o u
400 -
3Oi.i -
.•".i1 -
... _
100 -
r..o -
U.iO -
00 -
}
1
'|
i
\
\
\
\ }
'-/ '|
i
1
1
r.
"\ n A '-I ',-
V ' V ',
\r-\j\fa;
^V
W,A
i V'
. .' ^
1 \ /;
t 1
1 1
V
,
'(•'
If
ff
rt
V !
\
i
, l;
s
i.
Mil
' i
v J\
V-/ f.f '
'1 'l Irf
6/87 9/87 12/87 3/88 G/88 9/88 12/88
Da I o
0/89 9/89 12/89
Figure (5. Soil vapor quantity l.roal.rd over t i mo.
-------
O.i.Ł5|posei.J. Si. to
Recovered chlorinated hydrocarbons
O
ro
-------
Total recovered chlorinated hydrocarbons
o
CO
O
a:
u
1 -
0
6/87 9/87 12/87 :«/88 G/88 0/88 12/88 3/89 G/89 9/89 12/89
I)a l.o
Figure 8. Cummul alive solvent rocovorc»d by soil vnpor trr-ntmont.
-------
Ubersichtsplan Scherwiesen
Aktiv-x,
kohle-
filter
Biotop
ZOBL!
FIGURE 0
104
-------
V. NUREMBERG SITE
The second study involves a company located in Nuremberg, West Germany. It
was a used oil recycling plant for a number of years which declared
bankruptcy in the early 1970's. Several years later in 1979, the buildings
were razed and the number of hazardous chemicals removed from this site along
with several feet of surface soil. Eight (8) groundwater monitoring wells
were installed to monitor potential contamination of the underground aquifer.
During the drilling operations, it was discovered that the groundwater was
highly contaminated to a depth of 11 meters (33 feet) however, no action was
taken at this time to remediate the site.
Four (4) years later, following a period of substantial precipitation, a
large neighboring sand pit filled with oil due to a rise in the groundwater
level. Once visual contamination became readily apparent more investigations
were ordered which showed that the groundwater was contaminated with
chlorinated hydrocarbons and residual petroleum products. The groundwater
concentration was calculated as high as 18,000 M9/1•
In addition, located under a former barrel storage area, soil-vapor
concentration as high as 8,000 ug/1 consisting mostly of PCE, TCE and CIS 1,
2-DCE were noted. The area underneath the site was contaminated in the range
of 100 - 700 ug/1.
In September 1988, almost 10 years after the buildings were razed, a
treatment contract to remediate the site was awarded to ZUblin's
Environmental Engineering Department. Figure 10 is a plan view of the
Durbanol site. Figure 11 is a larger scale map showing the relative
location of the extraction wells and the plant.
VI. TREATMENT PROCESS
The treatment process at this site consisted of soil-vapor and groundwater
extraction and treatment of the VOC's and oil. The process was complicated
by a substantial amount of iron and manganese present in the groundwater.
Figure 12 is a schematic diagram of the various treatment process at this
particular site. Figure 13 is a plan view of the plant itself.
Soil-vapor extracted from well B9 using a vacuum blower with a capacity of
300 cubic meters per hour at a pressure of 50 mbar. The soil is very sandy
with a permeability of 10~3 cm/s. Range of influence for each well exceeds
50 feet. Wells Bl, 2, 5, 6 and 8 are used to monitor the groundwater and
offer additional vacuum extraction points, if required.
The soil-vapor is passed through a water separator to precipitate any water
picked-up by the suction. In this instance, the soil-vapor does not need to
be cooled because the blower, using less energy, does not heat this air up
significantly.
Wells SI and S2 are the two collection wells constructed in 1988 northeast of
the original plant site in the sand pit. These wells were each drilled 15
meters (45 feet) deep. Groundwater, at a flow rate of 1.5 liters per second
for each pump (measured by in-line gauges), is continually extracted from
both wells and fed into a two-stage air stripping plant.
105
-------
As the water trickles down through the towers, the counter flowing air strips
out the VOC's. The process air from the stripper is mixed with the soil-
vapor. The combined flow is then fed through the activated carbon filter
system which contains dual filters operating alternately in a fully automatic
environment. The system includes an on-site steam regeneration process for
the carbon. It is designed so that prior to reaching break-through it
switches over to the clean filter. At that point, the steam which is
generated by an oil fired burner, desorbes the VOC's from the carbon in the
off-line filter. Once the carbon is desorbed, which takes approximately an
hour, the filter remains off-line while the carbon cools down. At a
calculated predetermined time the automatic cycle switches'the clean filter
into the on-line mode.
Water for the steam, both to feed the boiler and condense the steam, consists
of contaminated water which has been cleaned in the liquid phase of the
operation. On this particular site, due to hardness problems, it is softened
before heating to eliminate scale build-up.
The steam, after it cleans the carbon, is condensed and flows into a cyclone
which separates the solvents from the condensate. The condensate is pumped
back into the system, mixed with the groundwater prior to the stripping tower
where its trace VOC's are removed.
After the stripping columns, the water flows into a collecting basin and
then through a sand filter which removes the flocculating iron and manganese.
The sand filter is piped so it can be back-flushed to remove the
flocculation when a drop in pressure is noticed.
In order to remove the trace VOC's and oil remaining in the water, a two stage
liquid phase carbon filter of approximately 5 cubic meters is utilized. Both
this discharge water and the cooling water used in the solvent recovery process
are subsequently discharged into a local sewage system.
The oil that is found floating on the surface of the groundwater in the
extraction wells is skimmed off by a membrane pump. The oil-water mixture
is fed into a separation tank which removes the oil from the water. The oil
is collected in containers and recycled. The water is pumped into the
treatment system prior to the stripping columns.
VII. ANALYTICAL MONITORING
Routine sampling and analysis of the system was performed during the study.
Four (4) sample ports measured the following information:
BL1 - Soil-vapor from well B9 after the vacuum blower but
before the activated charcoal filter.
BL2 - Consisted of a mixture of soil-vapor from B9 and
stripping air from the towers before the activated
charcoal filter.
BL3 - The discharge after the activated charcoal filter.
BL4 - Stripping air only before the activated charcoal filter.
106
-------
From start-up through mid-December 1988, only three sampling ports were
available; BL1, 2 and 3. Sampling port BL4 was installed in mid-December
1988. Once it was installed no further sampling was taken from port BL2.
VIII. OPERATION
Installation of the plant commenced on November 8, 1988. Ten days later the
first test run was conducted. On November 24 (3 weeks later) the groundwater
and soil-vapor treatment system commenced full operation and has been running
continuously since. During the period of November 24, 1988 through December
21, 1989, approximately 13 months, the following results have been achieved.
Soil Vapor
Hours of operation
Volume of soil-vapor
Total vapor extracted (approx.)
Recovered solvent
June 1989
4,854 hours
250 cu meters
1,213,500 cu meters
2,300 kg
December 1989
9,065 hours
250 cu meters
2,266,250 cu meters
2,700 kg
With the site concentration averaging 500 ug/1. The total quantity of
solvent received during the 13 months of operation was approximately 2,700
kilograms; about 2.5 kilograms per day.
Treated Groundwater
SI
S2
TOTAL:
Stripped VOC's
June 1989
18,315 cu meters
17,206 cu meters
35,521 cu meters
145 kg
December 1989
40,470 cu meters
34.210 cu meters
74,680 cu meters
175 kg
The results of the treatment program are pictured on several graphs. The
liquid extraction wells show a substantial contaminant reduction on Figure
14: SI initially increased in concentration from 5,050 ug/1 to over 6,980
ug/1 in December 1988 and then dropped to approximately 700 yg/1 in December
1989. S2 concentration was reduced from 11,500 ug/1 to under 700 ug/1 in the
same time frame.
It should be noted that in February and August 1989, clogging of the packing
material, due to the flocculating iron and manganese, resulted in an
inefficient stripping process which is readily apparent on Figure 14. The
packing material was exchanged - cleaned and the stripping efficiency
resumed. At all times the discharge concentration was below 1 ug/1 after
treatment.
On Figure 15, the obvious and expected decrease in soil-vapor concentration
from 6,900 ug/1 to values of approximately 500 ug/1 after one year is readily
apparent.
107
-------
IX. ECONOMIC ANALYSIS
The cost analysis for this project is divided up into six distinct phases as
follows:
DM US
A. Installation of Plant 35.110 DM $21,945
B. Soil-Vapor System
1. Capital Expense 33.845 21,155
2. Operation and Maintenance 17.200 -10,750
C. Stripping Tower
1. Capital Expense 45,920 28,700
2. Operation and Maintenance 11.830 7,395
D. Oil Stripping / Separator
1. Capital Expense 66.790 41,745
2. Operation and Maintenance 8.485 5,305
E. Iron / Manganese Removal
1. Capital Expense 40.565 25,355
2. Operation and Maintenance 8.855 5,535
F. Groundwater - ACS
1. Capital Expense 22.085 13,805
2. Operation and Maintenance 5.375 3,360
Total Capital Expense 244.315 DM $152,705
Total Operation and Maintenance 51.825 DM $ 32,345
These Costs Include:
1. Site preparation for the plant
2. Mechanical and electrical installation charges
3. Labor, material and equipment
4. Operating supplies, utility costs, and maintenance
These Costs Exclude:
1. Well installation
2. Permitting costs
3. Replacement charges (depreciation)
4. Carbon replacement - liquid phase
5. Engineering charges for the sampling/testing study
Costs are in 1989 dollars with a conversion rate of 1.6 DM to $1.00 US
108
-------
o
c:
m
t
'5
\
\
\
\
\
\
\
\
\
to \
*• \
^ \
00 \
o \
s \
^
.'
/
{'•.
\ \
\ '•
j
i
-------
-------
FORCED AIR
INJECTION
DISCHARGED
TREATED AIR
FEEDING WATER
ACTIVATED CHARCOAL
FILTERS
OIL SEPARATOR
STRIPPING
COLUMNS
PRESSURE
KETTLE
DISCHARGE-TREATED WATER
DURBANOL-PROCESS DIAGRAM
-------
10.3 METERS
AIR TREATMENT
COLLECTION BASIN I
COLLECTION BASIN II
t-> "• WATER STORAGE
ro .•
ACTIVIATED
CARBON
FILTERS
STEAM GENERATOR
WATER-SOFTENING
SCALE I > 29
DATE 10-28-1988
PLAN VIEW
-------
10000
5OOO
1000
500
100
m
co
10
9
N
1986
M
M
J J
1989
N
J
1990
LEGEND
DURBANOL
VOC CONCENTRATIONS IN WATER
o
•
VOC CONCENTRATION SI
VOC CONCENTRATION 52
VOC CONCENTRATION AFTER AIR STRIPPING
VOC CONCENTRATION AFTER FIRST A.C. FILTER
VOC CONCENTRATION AFTER DISCHARGE WATER
DATE: 1/31/1990
-------
ug/l
5000
4OOO
3000
2000
O
c
3>
tn
^
Ol
1000
1989
LEGEND
BL I SOIL VAPOR
BL 2 OFF GAS AFTER TREATMENT
DURBANOL
ANALYTICAL RESULTS
TOTAL VOC CONCENTRATION IN AIR
1990
-------
X. CONCLUSIONS
The following conclusions can be drawn from these two case studies:
A. The combined soil-vapor/groundwater extraction and treatment system
proved very effective in remediating a site contaminated with
chlorinated hydrocarbons. One major objective of this process was
to minimize impact on the surrounding environment during
remediation. This goal was admirably achieved.
B. The in-situ combination (water/air treatment) permits a wider range
of remediation. It also provides for recycling of the contaminated
condensate produced during on-site regeneration so that a complete
clean-up is achieved.
C. The on-site regeneration of activated carbon eliminates the cost
and problem of exchanging or disposing of the spent filter material
thus eliminating the "transfer of pollution" so common in today's
treatment strategy.
D. Concentrations of both the discharge water and air were well within
the regulatory requirements.
E. Residual soil and water concentrations at the site have continued
to decrease. It is certain that complete remediation will be
obtained with this process in the very near future.
F. Precipitation decreases the efficiency of a soil-vapor extraction
system.
G. An integrated extraction system is effective in both sandy and clay
soil. However, clay soil will be more expensive to treat.
SUWARY
The successful remediation of sites, containing both soil-vapor and
groundwater contaminant problems, can be best accomplished by the integrated
extraction and treatment system described in this paper. Dozens of projects,
throughout West Germany and Western Europe, have proven it to be the
economical method to remediate sites contaminated with volatile organics.
115
-------
ULTROX
INTERNATIONAL
S435 South Anne Street
Senta Ana. California 927O4
TEL: (714] 545-5557
FAX: (714) 557-5396
ULTRAVIOLET RADIATION/OXIDATION
OF ORGANIC CONTAMINANTS IN
GROUND, WASTE AND DRINKING WATERS
By: Jerome T. Barich
Presented at the EPA'a SECOND FORUM ON
INNOVATIVE HAZARDOUS WASTE TREATMENT TECHNOLOGIES:
DOMESTIC AND INTERNATIONAL
May 15-17, 1990
An ON-SITE Toxic Control, Inc. company
116
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
DESCRIPTION OF THE UV/OXIDATION PROCESS 2
ULTROX® UV/OXIDATION EQUIPMENT 2
APPLICATION OF UV/OXIDATION 3
CASE STUDY: EPA SITE PROGRAM - LORENTZ BARREL AND
DRUM SITE, SAN JOSE, CALIFORNIA 4
TYPICAL O&M COST GUIDELINES 5
PROCESS INFLUENT DESIGN CONSIDERATIONS 5
SUMMARY 6
Table I. Compounds Treated with UV/Oxidation 7
Table II. Applications of ULTROX® Technology 7
Table III. Typical O&M Costs by Compound Type 10
Table IV. Process Influent Design Consideration 10
Figure 1. Isometric View of System 11
Figure 2. System Flow Diagram 12
117
-------
INTRODUCTION
Regulatory standards for environmental clean-up are becoming
increasingly stringent. The U.S. EPA requires greater reliance on
permanent remedies. Local air quality districts are refusing to
allow water-borne contaminants to be transferred to the air. The
more conventional technologies (such as air stripping or activated
carbon) which transfer contaminants from one physical state to
another are becoming less well suited for solving today's problems.
In order to keep step with these tougher standards, Ultrox
International has developed its ULTROX® UV/oxidation technology
for the treatment of organic contaminants in water.
The ULTROX® UV/oxidation process is currently being applied to a
wide range of groundwater, wastewater, drinking water and process
water problems. Chlorinated solvents, BTEX compound, pesticides,
PCBs, phenols, and many other organic compounds, as well as BOD and
TOC, can be economically reduced to mandated levels.
Conventional chemical oxidation has been used in the treatment of
various waters polluted by organic chemicals for a number of years.
Hydrogen peroxide with a catalyst such as ferrous sulfate (Fenton's
Reagent) has been used for oxidizing phenol and other benzene
derivatives. Processes utilizing iron catalyzed peroxides and
chlorine compounds are attractive in that they utilize relatively
low-cost treatment equipment. The disadvantages of these processes
are that they can attack only a limited number of refractory
organics, and they produce iron sludges or chlorinated organics.
Commercial application of the patented ULTROX® process utilizing
ultraviolet light catalyzed ozone plus hydrogen peroxide
(UV/oxidation) as a water treatment technique is rapidly expanding.
It offers a means of destroying on site many of the toxic water
soluble organic chemicals found today in groundwater, wastewater,
leachate and drinking water supplies.
This paper describes the experience of Ultrox International in
applying the patented ULTROX® process to the full-scale treatment
of organic chemicals in wastewaters, drinking waters, leachates and
groundwaters. Ultrox International was issued a process patent in
1988 covering the application of UV light, ozone and hydrogen
peroxide to the treatment of a broad range of organic compounds in
water.
118
-------
DESCRIPTION OF THE UV/OXIDATIQN PROCESS
Ultraviolet light, when combined with O3 and/or H202 produces a
highly oxidative environment significantly more destructive than
that created with O, or H,02 by themselves or in combination.
UV light significantly enhances ozone or H2O2 reactivity by:
I. Transformation of 03 or H2O2 to highly reactive
(OH) radicals
II. Excitation of the target organic solute to a higher
energy level
III. Initial attack of the target organic by UV light
The importance of the conversion of the ozone or H202 to (OH) "can be
more easily understood after studying the relative oxidation power
of oxidizing species. Hydroxyl radicals have significantly higher
oxidation power than either hydrogen peroxide or ozone.
Oxidation Relative
Potential Oxidation
Species Volts Power*
Fluorine 3.06 2.25
Hydroxyl Radical (OH)" 2.80 2.05
Atomic Oxygen 2.42 1.78
Ozone 2.07 1.52
Chlorine Dioxide 1.96 1.44
Hydrogen Peroxide 1.77 1.30
Perhydroxyl Radicals 1.70 1.25
Hypochlorous Acid 1.49 1.10
Chlorine 1.36 1.00
* based on chlorine as reference (= 1.00)
ULTROX® UV/OXIDATION EQUIPMENT
The equipment components in ULTROX® systems: (1) have very few
moving parts, (2) operate at low pressure, (3) require a minimum
of maintenance, (4) operate full-time or intermittently in either
a continuous or batch treatment mode, (5) utilize energy efficient,
low temperature, long life UV lamps, and (6) can employ the use of
a micro-processor to control and automate the treatment process.
A typical process schematic and a reactor side view are illustrated
in Figures l and 2. The process incorporates highly reliable, time
119
-------
tested components to ensure high equipment utilization rates
(minimal down time) and continuous operation at performance levels
satisfying or exceeding performance specifications.
APPLICATION OF UV/QXIDATION
UV/oxidation equipment developed by Ultrox in recent years has been
used to treat a wide variety of waste streams. Table I lists toxic
compounds found in groundwaters, wastewaters and drinking waters
that have been successfully treated with the ULTROX® process. Case
histories from commercial systems and pilot and design studies for
applications in private industry and municipalities are presented
in Table II.
Modular equipment designs are used in ULTROX's commercial
installations. Reactor size varies from 325 gallons to 5,200
gallons. Ozone generators range from 21 Ib. to 250 Ib. per day.
In most applications, hydrogen peroxide is used with ozone.
Commercial ULTROX® systems now treat flow rates from 1.0 gpm to 950
gpm.
Generally, laboratory treatability studies and/or on-site pilot
plant studies are performed to determine the feasibility of
treating the water with UV/05, UV/H2O2, or UV/0,/H202. The ULTROX®
data base on the treatment of a wide range of compounds has now
expanded to where a minimum amount of laboratory or pilot plant
testing is required to design a commercial system for some priority
pollutants.
Full-scale systems, in most cases, are automated using micro-
processor control. Generally, the system is monitored (once per
shift or once per day). We have built commercial systems that are
monitored only once or twice a week. The systems are designed to
operate in a batch or continuous mode depending upon treatment
requirements.
In a number of cases, UV-oxidation is used as a component in a
larger treatment train. For example, at wood treating sites prior
to the UV-oxidation treatment, the wastewater or groundwater
requires upstream processing to break oil/water emulsions and
remove suspended matter. Process influent design considerations
are listed in Table IV.
CASE STUDY: EPA SITE PROGRAM - LORENTZ BARREL AND DRUM SITE,
SAN JOSE, CA
The EPA has established a formal program to accelerate the
development, demonstration, and use of new or innovative
technologies to be used in site cleanups. This program, called the
120
-------
Superfund Innovative Technology Evaluation (SITE) program, has four
goals:
• TO identify and, where possible, remove impediments to the
development and commercial use of alternative technologies.
• To conduct a demonstration program of the more promising
innovative technologies for the purpose of establishing
reliable performance and cost information for site
characterization and cleanup decision-making.
• To develop procedures and policies that encourage selection
of available alternative treatment remedies at Superfund
sites.
• To structure a development program that nurtures emerging
technologies.
Ultrox International applied and was selected for inclusion in the
SITE Program and in 1989 demonstrated its UV/oxidation technology
at the Lorentz Barrel and Drum Superfund site in San Jose,
California.
The Lorentz site was used for drum recycling for nearly 40 years.
Over this period of time, the site received drums from over 800
private companies, military bases, research laboratories, and
county agencies in California and Nevada. Drums arrived at the
site containing residual aqueous wastes, organic solvents, acids,
metal oxides, and oils. The groundwater beneath the site was
contaminated with a number of chlorinated solvents, chlordane,
toxaphene and PCBs.
An ULTROX® P-150 pilot plant was moved in on February 21, 1989.
Thirteen (13) tests were conducted between February 24 and March
9, 1989, on extracted groundwater from the site. The SITE Program
evaluation criteria were:
• Compliance of the treated water with regulatory
requirements
• Efficiency of the off gas (DECOMPOZON™)
Treatment Unit in Removing Ozone and VOCs
The organic compounds analyzed during the evaluation were:
• Aqueous Phase:
TCE 1,1-DCA 1,1,1-TCA
• Vapor Phase:
Vinyl Chloride 1,1-DCE 1,1-DCA
1,2-DCE 1,1,1-TCA 1,1,2,2-TCA
TCE Acetone Benzene
121
-------
Upon completion of the pilot plant demonstration and evaluation of
the analytical data from samples of the influent and effluent
(treated) water and the off gas into and out of the DECOMPOZON™
unit, it was concluded in the EPA's Technology Evaluation Report
that:
• The treated ground water met the applicable NPDES
standards for discharge into Coyote Creek
• There were no volatile organics detected in the exhaust
from the DECOMPOZON™ unit
• The DECOMPOZON™ unit reduced ozone in the ULTROX® reactor
off gas to levels less than 0.1 p.p.m. (OSHA) standards.
The U.S. EPA's Technology Evaluation Report fully explains the
details of this study.
The remainder of this paper is focussed on full-scale applications
of the ULTROX® process as shown in Table II, Cases A through L.
TYPICAL O&M COST GUIDELINES
Operating and maintenance costs can vary widely depending upon:
Electrical Power . Type and Concentration of Contaminants
H202 Cost . Concentration of Competing Species
Oxygen Cost . Treatment Goals
O&M costs can vary from as low as $.10/1000 gallons in municipal
drinking water applications to as high as $.10/gallon for leachate-
type or highly concentrated organic-laden waste waters. See Table
III.
PROCESS INFLUENT DESIGN CONSIDERATIONS
UV/oxidation is used often as a stand-alone treatment process.
Many times, however, other characteristics of a waste water or
ground water dictate that UV/oxidation be combined into a treatment
train with other technologies to obtain the desired result.
At one wood treating site, oil removal and filtration precedes the
ULTROX® process. Metals precipitation and free hydrocarbon removal
equipment is upstream of an ULTROX® system in a marine terminal
evaluation. pH adjustment of the influent to the ULTROX® process
can, at times, result in faster reaction times and lower oxidant
requirements. Table IV lists several parameters that could impact
UV/oxidation efficiency and necessitate the use of additional
processes.
122
-------
SUMMARY
The ULTROX® process is used commercially in a variety of
applications. We expect this range of applications to increase
greatly based on the results of many recently completed laboratory
and pilot plant studies. The process destroys or detoxifies
contaminants ON SITE without the generation of sludges or the use
of adsorbents that require additional treatment. UV/oxidation
process economics are competitive on a life cycle cost basis in a
direct comparison with these older processes that have the residual
problems cited above. Yet, these economic comparisons usually do
not assign any credit for the fact that residual liability is not
associated with UV/oxidation. Any economic analysis that would
assign credit for the elimination of residual liability (certainly
this must be considered in the case of real estate transfer) would
further improve the competitiveness of the technology.
123
-------
TABLE I: COMPOUNDS TREATED WITH UV/OXIDATION
ETHERS
BTEX
PHENOL
TCE
PCE
DCE
POLYNITROPHENOLS
KETONES
VINYL CHLORIDE
PESTICIDES
CITRIC ACID
TCA
DCA
MeCl2
CRESOLS
PCBS
PCP
TNT
AROMATIC AMINES
COMPLEXED CYANIDES
POLYNUCLEAR AROMATICS
DIOXINS
HYDRAZINE
RDX
1,4 DIOXANE
EDTA
HYDRAZINE
TABLE II: APPLICATIONS OF ULTROX® TECHNOLOGY
CASE A:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
DISPOSITION:
GROUNDWATER
METAL PLATER
450 p.p.b. COMPLEXED CYANIDES
<20 p.p.b.
< 2 p.p.b.
P.O.T.W.
CASE B:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION:
WASTEWATER
CHEMICAL MANUFACTURER
BOD5 70 p.p.m.
<30 p.p.m.
BOD5 <20 p.p.m.
650 g.p.m.
SURFACE WATER
CASE C:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION:
WASTEWATER
PETROCHEMICAL MANUFACTURER
PHENOLS >135 p.p.m.
PHENOLS <5.0 p.p.m.
PHENOLS <1.0 p.p.m.
70 GPM
BIOTREATMENT
124
-------
TABLE II: APPLICATIONS OF ULTROX® TECHNOLOGY
CASE D:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION:
WASTEWATER
PHENOLIC RESIN PLANT
BIOTREATMENT EFFLUENT PHENOLS >3.0 p.p.m.
14 p.p.b.
8 p.p.b.
25 GPM
P.O.T.W.
CASE E:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION:
GROUNDWATER
AUTOMOTIVE FOUNDRY
CHLORINATED VOCs >6000 p.p.b.
<20 p.p.b.
< 5 p.p.b.
200 GPM
SURFACE WATER
CASE F:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION:
GROUNDWATER
SEMICONDUCTOR MANUFACTURER
12.5 p.p.m. BTEX
<20 p.p.b.
< 5 p.p.b.
5.0 GPM
P.O.T.W.
CASE G:
SOURCE:
CONTAMINANT;
GOAL:
RESULT:
FLOW:
DISPOSITION:
GROUNDWATER
PHOTOCOPIER MANUFACTURER
TOLUENE + TCE >6000 p.p.b.
99% DESTRUCTION
99+%
150 GPM
STORM SEWER
CASE H:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION
WASTEWATER
PHARMACEUTICAL FIRM
COD >700 p.p.m.
COD <250 p.p.m.
COD <200 p.p.m.
7.0 GPM
P.O.T.W.
125
-------
TABLE II: APPLICATIONS OF ULTROX* TECHNOLOGY
CASE I:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
DISPOSITION:
PROCESS WATER
STEAM CONDENSATE
TOC >1000 p.p.b.
TOC < 50 p.p.b.
TOC < 30 p.p.b.
25 GPM
BOILER FEED WATER
CASE J:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
WASTEWATER
SPECIALTY CHEMICAL MANUFACTURER
150+ p.p.a. 1,4 DIOXANE IN BIOTREATER
EFFLUENT
90% - 95% DESTRUCTION
98+%
10 GPM
CASE K:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
MUNICIPAL DRINKING WATER
MANUFACTURING FIRMS
PCE 25 p.p.b.
< 5 p.p.b.
0.5 p.p.b.
950 GPM
CASE L:
SOURCE:
CONTAMINANT:
GOAL:
RESULT:
FLOW:
GROUNDWATER
WOOD TREATING FACILITY
6000 p.p.b. PHENOL
<200 p.p.b.
<100 p.p.b.
40 GPM
126
-------
TABLE III: TYPICAL O&M COSTS BY COMPOUND TYPE
S/1000 GALLONS*
TCE BTEX BASE NEUTRALS KETONES
PCE _r = -
$.45 $.80 $1.10 $1.50
* Costs based on influent contaminant concentration * 5.0 mg/1
TABLE IV: PROCESS INFLUENT DESIGN CONSIDERATIONS
OPACITY IRON
SUSPENDED SOLIDS MANGANESE
FREE HYDROCARBON ALKALINITY
HARDNESS pH
127
-------
ro
CO
ISOMETRIC VIEW OF
ULTROX SYSTEM
TREATED OFF GAS
CATALYTIC OZONE
DECOMPOSER
REACTOR OFF GAS
OZONE GENERATOR
DRYER
TREATED EFFLUENT
ULTROX
UV/OXIDATION REACTOR
HYDROGEN PEROXIDE
FROM FEED TANK
FIGURE 1
-------
ftotemtter
Ozone
from OZOM
generator
1
from SMtam
Ground-Water
Uonttoring Weta
f Effluent
1 Sample Tap
Hydrogen Peroxide
Feed Tank
Contaminated Water
Feed Tank (Bladder)
FIGURE 2
ULTROX* SYSTEM
FLOW DIAGRAM
CREATED: 11/4/88 ) REVBED: 02/02/89 | PIOT.CRF
-------
The Liaxrgi — Deoonteir ra— Piroc<
Wet Mechanical Site RemecL i si t i on
Eckart F. Hilner
Philadelphia
May 1990
130
-------
1. Introduction
2. Soil washing strategy
3. Process description
4. Comparison to other soil decontamination systems
5. Cleaning results
6. Costs
131
-------
Introduction
Contaminated soils are destined to become a major problem for
all densely populated countries. For instance this problem is
one of the most urgent political and technological challenges
facing Germany today.
The protection of the environment from pollution and the reme-
dial treatment of contaminated soil is an important objective
for the years to come.
In future natural land will hardly be available for new hou-
sing areas or industrial complexes. Old buildings give way to
new buildings. For new industries only old industrial sites
will be available.
These old sites often are highly contaminated being partly
dangerous for the inhabitants.
Before building up new industries in many caces the site has
to be decontaminated including soil and groundwater.
Because landfill capacities in Europe are more and more
limited, land disposal prohibitions will soon require that
these wastes be treated using best demonstrated available
technology.
Soil washing strategy
An important strategy in operating a soil washing system de-
pends on the fact/ that contaminants are primarily adsorbed
by clay and silt-minerals, hydrous oxides, and organic mat-
ter and are therefore concentrated in the fines fraction of
the soil. The cation exchange capacity of clay minerals
132
-------
enables large amounts of certain elements, inorganic and or-
ganic matter, to be adsorbed. On the other hand contaminants
may stick on the coarser material, which may be scrubbed by
either water alone or mixtures of water with some detergents.
In general separating organic and inorganic contaminants from
the soil is a purely pysical procedure.
The main problem is the input of sufficient energy into the
soil to liberate the contaminants and to transfer these into
a fine material suspension.
Lurgi's Deconterra-Process was developed using extensive ex-
perience in the field of wet mechanical processing of mine-
rals, especially of ores and salts, as well as of sludge from
loaden waters.
The separation into a clean soil fraction and a concentrate
of contaminants takes place in an attritioner drum. Conta-
minated soils undergo a severe scrubbing process, which in-
termingles coarse and fine material, using the coarser pebbles
like grinding bodies with a net energy input of 4 to 16 kW
hr/ton of throughput, during which the rotation speed of the
drum is between 50 and 90 % of the critical speed of
" 42'4 (min'1)
The suspension of the drum discharge will be screened in
coarser and finer fractions. These fractions are further
treated.
133
-------
Process description
The Deconterra-process for soil washing is illustrated in the
following flow sheet (picture 1).
The contaminated soil is reclaimed/ classified at 300 mm and
screened into a 0 - 100 mm fraction and a 100 - 300 mm frac-
tion.
The coarse fraction (100 - 300 mm) is crushed to under 100 mm
in a crusher and, together with the fine fraction, is trans-
ferred to a wet attritioner drum.
In the attritioner drum, the bulk of the adhering contaminants
is rubbed off with a controlled input of energy and first of
all suspended in the liquid and then bound adsorptively to
the fine particles of the soil.
The energy required for the attrition can be matched to the
type of soil and the character of the contamination, up to 16
kW hr/ton being transferred to the material.
The material discharged from the drum is screened in several
stages.
The fraction over 20 mm is returned via the braker cycle once
again to the attritioner drum.
The 1 - 20 mm fraction is either discharged from the process
as purified end product or subjected to a gravimetric sizing,
the heavy material being discharged clean, whereas the light
material, such as contaminated coarser coal, tar and wooden
particles, being obtained as a contaminants concentrate.
134
-------
Of the fraction under 1 nun, the finest fraction is cleared
from mud and transferred to a hydrocyclone, the overflow of
which is concentrated.
The cyclone underflow, together with the de-slimed coarse
fraction of the classifier, is introduced into the 2nd attri-
tion step and subjected to a further purification.
The material discharged from the 2nd attrition drum is pumped
to a froth-flotation system, the special reagents of which
are matched to particular requirements of the contaminants
involved.
The suspension discharged from the flotation system contains
the purified material, which is subsequently dewatered to a
residual moisture content of about 15 - 20 % and can be dis-
posed or piled as cleaned product.
The contaminants are concentrated in the foam discharged du-
ring the floatation. This foam reaches the thickener for pre-
dewatering. The thickener underflow is brought to a moisture
content of about 30 % in a dewatering step and represents the
bulk of the so-called contaminants concentrate, which con-
tains contaminants removed from the bottom in a concentrated
form.
The contaminants concentrate is composed of the following
materials, which are separated in the process:
1. Light material from the gravimetric sizing (such as
wood, tar fragments, coal, coke, etc.)
2. Overflow from the hydrocyclone, esentially the fines
portion of the soil with the adsorbed contaminants.
3. Froth from the flotation.
135
-------
Fractions 2 and 3 are obtained together as filter cake.
It has turned out that the process can also effectively clean
the fine grain fraction. The process therefore is also able
to decontaminate soils with a high proportion of fine grain
(less than 63 microns in diameter). Depending on the composi-
tion of the soil and the nature of the contamination, the
yield of decontaminated soil is between 65 % and 85 % of the
untreated soil.
The scrubbing water, resulting from the Deconterra-process,
is slightly contaminated, so that the bulk of the process
water can be pumped back into the plant. To avoid accumula-
tion of the contaminants in the water, a portion of it is
diverted to a water processing stage, in which the very fine
solid materials, as well as the dissolved contaminants are
reduced to the required minimum amount.
Comparison to other soil decontamination systems
Besides soil washing systems thermal methods suit very well
for organic contaminants to be destroyed by incineration. On
the other hand inorganic contaminants can be removed only
partly by incineration. Heavy metals can be solidified or
leached by extraction methods. The costs of treatment inclu-
ding exhaust gas cleaning and residue disposal will be very
significant.
Another alternative for decontamination of soil is the bio-
logical treatment. But this methods is limited to some simple
build up organic compounds and to a very small amount of in-
organic compounds.
136
-------
In general all various soil washing systems operate on the
principle that contaminants are concentrated in the fine com-
ponents of the soil. Materials coarser than for instance
100 y m are discharged essentially free of contamination. The
systems differ by the method of putting energy into the con-
taminated soil and by using different washing agents.
Two companies from the Netherlands use hydroclones and scrub-
bing with detergents and oxidants.
A dutch company and a german company use a high pressure wa-
ter beam to remove the contaminants from other soil partic-
les. Problems exist in high contents of fines in the soil.
One german company is using a vibrating screw for energy
transfer while another german company uses a centrifugal se-
parator.
The Lurgi-Deconterra-Process is using only water as extrac-
ting agent without addition of detergents, solvents, acids,
bases or similar materials. Included into the process is the
further treatment of the + 1 mm soil fraction.
After classification and the second attritioner step the
fraction 20 um to 1 nun is purified by pneumatic flotation.
During this step an optimum of different surface characte-
ristics of the soil components and the actual contaminants
can be achieved by dosing agents into the three flotation
cells.
This is a major difference and an essential advantage of the
Lurgi-Deconterra-Process compared to other soil washing pro-
cesses, which are mainly based on classification as a sepa-
rating process and not so much on a sorting process.
137
-------
Cleaning results
Lurgi has investigated and purified various samples from many
different contaminated soils, for instance from sites of old
coking plants, old ammunition plants, various chemicr1 indu-
strial sites and contaminated waste locations.
The following tables give some results out of various test-
work.
material from different sites of old coking plants in
the Ruhr Area and the Saar Area, Germany
Component contaminated soil decontaminated
(mg/kg) soil (mg/kg)
Dutch reference
(mg/kg)
A B
Naphthalene
Phenanthrene
Anthracene
Fluor anthene
Pyrene
Benzo-a-pyrene
I Polyaromatic
hydrocarbons
(PAH)
Ł Hydrocarbons
Cyanide (total)
18
237
40
220
170
49
1357
4000
94
- 0,2
1,5
- 0,2
0,3
- 0,2
- 0,2
7,1
30
17
0,1
0,1
0,1
0,1
0,1
0,05
1,0
100
5
5
10
10
10
10
10
20
1000
50
138
-------
Component contaminated soil decontaminated
(mg/kg) soil (mg/kg)
Dutch reference
(mg/kg)
A B
Arsenic
Lead
Copper
Nickel
Zinc
Vanadium
Chromium
Mercury
435
80 - 250
60 - 200
30
330
50 - 70
1400
700
22
< 50
< 50
< 20
<100
< 40
130
22
20
50
50
50
200
100
0,5
30
150
100
100
500
250
2
During the second world war several large ammunition works in
germany produced explosives/ mainly Trinitrotoluene TNT. To-
day these old sites are highly contaminated with Nitrotoluene,
Soil from some of these works was treated in a Deconterra-
pilot-plant with good results.
component
contaminated
soil (mg/kg)
decontaminated
soil (mg/kg)
Nitrotoluene
106,000
0.34
139
-------
Several possibilities are available how to treat the conta-
minants fines concentrate. The non satisfying method is ha-
zardous wast landfilling.
The thermal treatment of the concentrates in incineration
plants may be a good solution specially for organic contami-
nants. For inorganic contaminants solidification is a solu-
tion. In Germany thermal methods bear a lack of acceptance by
the population and therefore are very difficult to prevail.
At present Lurgi develops a process to destroy organic conta-
minants from the Deconterra washing concentrate by pressure
oxidation in an autoclave.
Costs
The economics for a Deconterra-system operating in the United
States of America are very difficult to estimate.
The costs are not only caused by the process itself but also
to a large extend by the local conditions, the contaminants,
the analytical expense, the safety conditions and environmen-
tal restrictions.
140
-------
Based on a Deconterra plant, which is a modular construction
and semimobile with a capacity of 25 t/hr wet material or
80.000 tons/year the treatment fee per ton is in the range of
$80-90. Excluded from this amount are the costs for re-
sidue disposal.
141
-------
tss
CONTAMINATED SOL (FEED)
ION
EXCHANGE ACTIVATED
CARBON FILTER
LIGHT MATERIAL
(CONTAMMATEO)
HEAVY MATERIAL
(CLEAN)
CONTAMINATION
DECONTERRA Process
-------
NBM BODEMSANERING
Developments and operating experience in thermal soilcleaning
NBM Bodemsanering B.V.
P.O. Box 16032, 2500 BA The Hague
Netherlands
Phone 31 70 3814331 / Fax 31 70 3834013
Ir. H.J. van Hasselt/Ir. A. Costerus
1. Abstract
2. Introduction
3. System description
3.1 Process
3.2 Equipement
4. Operational Costs and Production factors
5. Legislation
6. Development
143
-------
NBM BODEMSANERING
Developments and operating experience in thermal soilcleaninq
1. Abstract
This paper discusses the development and operating experience of
an indirect heated soil cleaning plant.
In The Netherlands, the soil pollution in the village of Lekker-
kerk, was one of the news highlights in 1979.
As a direct result of this, the Dutch government decided to start
a soil sanitation project for The Netherlands.
NBM Bodemsanering B.V., subsidiary of NBM Amstelland N.V., the
second building and construction group in The Netherlands, has
developed an indirect heated thermal treament system and became
one of the leading firms in The Netherlands in the cleaning of
soil and the execution of projects for soil pollution.
The equipment, operational costs, production factors and legisla-
tion are presented in this paper, and have, in 3 years of operati-
on, produces over 200.000 tons of cleaned soil.
The gas conditions in the rotary tube furnace, the absence of
oxygen and the presence of hydrogen, create optimal conditions for
the cleaning of soil containing halogenated hydrocarbons.
At the Technical University of Delft, laboratory work was done to
determine the process conditions required to bring the pollution
level below the accepted ("A" value, < 0,1 mg organochlorine
compound/kg) level and to test the process conditions for incine-
ration. Special attention was paid to the formation of CDD and
CDF.
Based on the results of this laboratory work, NBM decided to
carry out practical experiments.
First, in 1989, the incineration conditions were tested in the
full scale plant in Schiedam. While running the plant with an
input of soil contaminated with fuel oil a known amount of a
mixture of trichloroethene and tetrachloromethane was injected
into the incinerator, and stack emissions were measured. The
destruction efficiency was higher than 99,999 % .
144
-------
NBM BODEMSANERING
Further experiments were carried out in April 1990. About 700
tons of soil containing a.o. aldrin, dieldrin, and lindane were
processed during a three day test run. In these tests, the remai-
ning rest concentrations of contaminants were below the detection
limit of 10 micrograms/kg dry solids.
2. Introduction
In The Netherlands, the soil pollution in the village of Lekker-
kerk, was one of the news highlights in 1979. New houses had been
built on an old dumping site, and the soil was polluted with
benzene, among other substances. The public was alerted by a
stream of news about other sites that were contaminated.
As a direct result of this, the Dutch Government decided to start
a soil sanitation project for The Netherlands.
A first general investigation was made, and the Dutch Government
published the reference values for soil contamination, known as
the ABC values.
The government made Hfl. 2,000,000,000 available for the investi-
gation of suspected sites, to make plans for the sanitation, to
develop cleaning technologies, and to execute the first priority
jobs.
The road construction company NBM, a subsidiary of the NBM-
Amstelland group, was one of the firms that started the develop-
ment of a thermal cleaning technology.
The system had to meet the following requirements:
- It must be possible to clean any type of soil (sand, clay,
peat).
- It must be possible to remove all organic compounds and pollu-
tants that can be volatilized at temperatures up to 600 - 650 C
or can be removed by pyrolysis reactions at temperatures of 300
- 650 C.
Thermal process are not expected to remove heavy metals from
soil. Their very high boiling ooints require the development of
other cleaning process.
NBM decided to choose the indirect heated thermal process.
145
-------
NBM BODEMSANERING
See figure 1: schematic presentation of direct and indirect
heated thermal process.
Soil input
_
r--^^
r i
Rotary tub* lurnac*
II 1 1 II I! || ||
_
chlmrwv
nclncntoi
\
.^-^
r^
S
n
•oil input
burntr
Schematic indirect heating
, Rotary drum
~~*
chimney _
Inclntrttor
-H
\
Schematic direct heating
The principal arguments for the choice between direct and
indirect heated process are given in the following table.
Direct heated
Advantage
Unlimited
heat tansfer
capacity
relative short
heating and
cooling time
of equipment
Drawback
High volume of
gas stream in
rotary kiln
High volume of
gas steam in
incinerator
Dust difficult
to remove from
gas stream
Indirect heated
Advantage
Low volume of
gas stream
Low volume of
gas stream in
incinerator
Filtration of
dust at high
temperature
possible
Drawback
Limited heat
transfer
Long heating
and cooling
time of the
equipment
146
-------
NBM BODEMSANERING
3. System description
3.1 Process
When a soil contaminated with organic compounds is heated the
evaporation of volatile compounds will occur. The initial evapora-
tion rate will depend on the vapor pressure of the compound at
the surface of the matrix (soil), which is a direct function of
the temperature, the physical chacteristics of the compound, and
the adsorbtive interactions (physical and chemical) between the
compounds and the soil.
The aim of the equipment is to achieve a soil temperature high
enough to obtain a vapour pressure of the compound above. 1 atmosp-
here. In that case the transport in the vapour phase will be by
convection transport even in the pores of the soil particles.
The temperature of the tube furnace and the soil is given in
figure 2.
Figure 2
Temperature distribution of soil in tube
furnace for pilot - and commercial plant
1000
Temperature
800
600
400i
200
0 15 30 45
Residence time (min)
1 3
Pilot pltnt Comm. plant norm*)
Comm. plant max. Wall ttmptratur*
60
147
-------
NBM BODEMSANERING
For compounds which are not thermostable, thermolytic fission
reactions occur before the vapour pressure of 1 atmosphere will be
achieved. It is known that thermolylic fission reactions occur
for the following compounds:
- Iron cyanide complex ( Prussian blue)
- hexachloro cyclohexane
- longer aliphatic hydrocarbons
- natural humic materials
The thermolytic fission reaction of the natural humic materials
plays an important role in the process in the rotary tube furnace.
The following reaction is suspected to take place :
Formula
+H2OT
Mol.
quant. 1 —> 31 +16
+COT +H2T +CH4T +C2H6T
+5
+2 +2 +4 +1
Weight
kg. 1 --> 0,30 +0,22 +0,17 +0,12 +0 +0,03 +0,09 +0,08
Gasprod.out of
1 kg gas volume
in liters --> - +311 +100 +100 +40 +40 +80 +40
The rotary tube furnace is a highly gastight furnace, therefore no
explosion occurs with the inflammable gases because of the absence
of oxygen. The oxygen which however could enter the system reacts
instantaneously with the hydrogen present.
The absence of oxygen in the polluted gas stream is also an
important factor in treating chlorinated hydrocarbons. In the gas
phase dioxins and dibenzofurans could easily be formed in the
temperature range of 400 - 700 C , but the absence of oxygen makes
these formations impossible.
The polluted gas stream which has a small volume (300 - 500
nm3/hr) is treated in an incinerator at a temperature of 1100-
1200 C. All organic compounds will be oxidized to CO- and H-0,
with small quantities of NO
148
-------
NBM BODEMSANERING
At a temperature of 1100 - 1200 C, with a residual time of at
least 1 second and good turbulence (a small relative residence
time distribution) are required to destroy chlorinated hydro-
carbons too, without forming undesired products of incomplete
combustion.
Laboratory tests were conducted to establish the undesired forma-
tion of dioxins and dibenzofurans in the oxidation of hexachloro
cyclohexane and of drins as a function of residence time, tempera-
ture and oxygen content. The formation of 1 ng of undesired
product of incomplete combustion from 1 gram of chlorinated
hydrocarbon was accepted as maximum value.
Formation of PCDD + PCDF by combustion of dieldrin
time /temp.
0,5 s
0,7 s
1,2 s
1,5 s
1,7 s
750° C
• • • •
0+17
• • • •
21+2700
0+0
950° C
• • • •
93+21000
7+357
0+0
1150° C
50+1100
0+20
0+0
Formation of DCDD + PCDF by combustion of HCH
time/temp.
0,5 s
0,9 a
0,9 s
1,2 s
1,5 s
2,1 s
2,3 s
680° C
• • • *
• • * •
• • • •
• • • •
* • • •
• * • *
4500+274000
750° C
• • • •
• * « •
• • * •
* • • •
0+0
0+95
900° C
* * • •
0+95
0+19
10+465
1100° C
0+0
0+0
The values are the quantities expressed in ng of PCDD + PCDF
formulated out of 1 gram Dieldrin or 1 gram HCH.
149
-------
NBM BODEMSANERING
In the commercial plant, the average temperature in the incinera-
tor is 1100 C, and the average residence time 2 seconds. The stack
emissions of the incinerator meet the limits of the licenses.
Before the oxidation of the polluted gas stream out of the rotary
tube furnace the gas stream has to be filtrated from dust. Other-
wise dust will be melted and settled down in the incinerator.
High temperature gas filtration is necessary because cooling down
the polluted gas stream will condense the high boiling organic
pollutants as polyaromatic hydrocarbons.
Ceramic pipes filtrate the gas stream at temperature levels above
550 C. The filtered dust is clean of the organic pollutants if
the temperature has been high enough.
3.2 Equipment
3.2.0 Pilot plant
Before realising a full scale plant a pilot plant was built in the
first half of 1984. It had a capacity of 0.5 tons of soil/hour.
In the middle of 1984 a great number of test were carried out with
soil originating from former gasworks sites.
Figure 3 shows some characteristic results of testing at various
temperatures.
Sou Mfflo«r«tur*
1000
800
800 h - -
CN content of polluted soil before
and after cleaning as a function of
the furnace temperature.
400]
200i ,
1000
150
-------
NBM BODEMSANERING
The pilot plant tests demonstrated, that the selected system
yields to the values of the Dutch Government, that is to say near
the detection limit. Accordingly, this system formed the basis for
the concept developed for a production plant.
In general it must be stated that no plant can be so designed that
it is an optimal design for any type of soil and any sort of
pollution. The basic principles - evaporating in an indirect
heated rotary tube and oxydation of the pollutants in an incinera-
tor will always apply. However, the optimum design will show
differences, depending on the specific conditions for any given
proj ect.
Plant of Schiedam
The design was made for:
maximum 13 tons/hour
moisture content 25%
energy content 1200 kJ/kg.d.s
any type of soil
continuous operating, 7000 hours a year
3.2.1 Commercial plant
The commercial plant consists of the following items:
1. Soil input
The polluted soil, from which very coarse pieces, exceeding 100
nun, have been removed, is loaded into a feed unit.
2. Feed unit
The soil coming from the bin drops into a picking belt, where
coarse pieces and non-ferrous metals may be removed. Then the soil
is moved upwards by a conveyor belt to the dryer. A magnet mounted
above the belt removes ferrous metals as much as possible from the
soil.
151
-------
NBM BODEMSANERING
3. Dryer
The dryer comprises mainly a rotary drum having a diameter of 3 m
and a length of 21,5 m. The drum is heated externally, that is
indirectly, by means of hot gases. The heat sources for the dryer
are:
- burner gases exiting from the indirect heated rotary tube furna-
ce,
- the flue gases from the incinerator, and
- auxiliary burners.
4. Screen
When leaving the dryer, practically all of the free water has been
evaporated from the soil. At this stage the structure of the soil
is such that fragments of stone, rubble and pieces of wood may be
removed by screening. This is necessary because the transportation
and sealing means of the rotary tube furnace cannot cope with very
coarse pieces.
5. Transfer to the tube furnace
The fine fraction passing the screen, together with the dust
filtered from the exhaust gas from the dryer is passed through a
screw conveyor to an elevator. The elevator carries the soil into
an intermediate bin, where it is introduced into the tube furnace.
6. Rotary tube furnace
The dried and screened soil is heated in the tube furnace to a
temperature between 450 and 600 C. The achievable maximum tempe-
rature of the soil and the residence time of the soil in the drum
are, obviously, a function of the quantity of soil introduced per
hour. The tube furnace is provided with gas-sealing gaskets and
seals.
152
-------
NBM BODEMSANERING
7. Soil cooler
The cooler consist of a rotating drum. A plurality of pipes are
located inside the drum, through wich cooling air is blown. The
cleaned soil introduced into the cooler is cooled to approximately
150 C. The air leaving the cooler has a temperature of approxima-
tely 250 C and is used as preheated combustion air for the
burners of the tube furnace and the dryer.
8. Mixer humidifier
The soil is continuously humidified in a mixer. The soil leaving
the mixer has a temperature of approximately 50 C and a moisture
content that can be adjusted in advance. A conveyor belt carries
the soil to a bin.
9. Treatment of the gas leaving the dryer
The gases exhausted from the dryer include mainly steam, light
pollutants evaporated from the soil and air sucked in via the feed
mechanism and some of the seals. This gas contains a certain
quantity of dust which is removed from the gas stream by a filter.
The dust collected in the filter is fed back to the main stream of
soil.
Depending on the type of the soil and the type of the pollutants,
the gas stream from the dryer may be processed by either one of
two routes.
Steam condensation
By lowering the temperature of the gases exiting in the dryer the
major part of the steam is condensed in a condenser. The condensed
water is passed to a water treatment unit.
Direct exhaust to the incinerator
If it is not desirable to condense the steam from the dryer, the
dryer effluent passes directly from the filters to the incinera-
tor. The gas stream to the incinerator is now much more substanti-
al and so is the energy consumption.
153
-------
NBM BODEMSANERING
10. Treatment of the gas leaving the tube furnace
The gas stream from the tube furnace comprises the following
groups of substances:
- pollutants evaporated from the soil.
- moisture evaporated from the soil.
- products of pyrolysis of organic substances present in the
soil.
- inert gas entering into the gas stream through the flushing
device of the ceramic filter located on the line between the
tube furnace and the incinerator.
The temperature of the gases leaving the tube furnace is a functi-
on of the selected soil temperature. It has the same order of
magnitude as the temperature of the existing soil. The gaseous
mixture is then passed to the ceramic filter unit.
11. Ceramic filter unit
The ceramic filter unit comprises the following parts:
a. A settling chamber where the coarser dust particles contained
in the gas stream and entering the chamber are pre-separated by
settling.
b. Two parallel filter units, each provided with valves and an
exhaust fan. The gases pass from the settling chamber through
one or both of the filters and go on to the incinererator.
12. Incinerator
All the gases having been in contact with the polluted soil are
burnt in the incinerator. If they are burnt in the incinerator for
a sufficient period of time at an adequately high temperature and
with an adequate oxygen percentage, all pollutants will be conver-
ted into non-toxic compounds such as H-0, C02 and N2«
The composition of the gas stream fed to tne incinerator depend
greatly on the composition of the soil, its moisture content and
the type and quantity of the pollutants.
154
-------
NBM BODEMSANERING
A number of variables are known:
a. The plant is operated with or without the condensers.
The amount of steam, and thus the fuel consumption, is very
much dependent on this variable.
b. Type and quantity of the pollutants.
All pollutants in the soil will, through the gas phase, end up
in the incinerator. The energy contents of the substances will
be more or less positive.
c. The quantity of organic matter present in the soil (besides the
pollutants).
Under the process conditions in the tube furnace, organic
matter will be broken down to carbon, gases and tar-like
materials, which will be exhausted to the incinerator. The
energy contents of these materials may be considerable.
d. The final temperature of the soil in the furnace.
The pyrolysis process described above will also be influenced
by the process temperature.
Analysis of the above parameters shows that the gas stream ente-
ring the incinerator is subject to considerable variations. If
requirements are established for the residence time and/or the end
temperature of the gases in the incinerator, the quantity of soil
introduced into the installation must be adjusted as a function of
the process conditions.
13. Exhaust
When the installation is in normal operation, practically all of
the gases produced in the incinerator, together with the gases
from the tube furnace, will be exhausted through the heating
jacket of the dryer.
An exhaust fan removes the gases from the system and ensure a
sub-atmospheric pressure in the entire system.
If the gas stream from the incinerator is too great to be handled
through the dryer, this stream (or a part of it) will be removed
from the installation by another exhaust fan.
155
-------
NBM BODEMSANERING
4. Operational costs and production factors
The industrial plant in Schiedam starts cleaning in April 1987.
The production hours and tonnages of polluted soils are given
below.
Production indirect thermal cleaning plant Schiedam
1987
1988
1989
Production
hours
4.400
6.080
6.800
Tons of polluted soils
cleaned
50.200
76.000
82.000
The cleaning results proves to be independent of the input concen-
tration. In the following table the average value for the most
analysed pollutants in the cleaned soil are given.
Cleaning results
Input Output
CN total
Aliphatic hydrocarbons
Aromates
Naptalene
PAH's (EPA)
0
0
0
0
0
- 10,000
- 20,000
- 5,000
- 10,000
- 20,000
< 5
< 50
0.05 -
1
1
0.1
156
-------
NBM BODEMSANERING
The operational costs are principally energy, labor, maintenance
and investment. The relative costs of the selling price are shown
below.
Breakdown of cleaning price
Item
Risk/benefit
Energy
Labor
Maintenance
Capital cost
Analyses
%
15
20
17
15
30
3
The operational costs greatly depend on the production capacity.
Only the energy cost is related to the production. The production
capacity depends on:
. type of soil
. moisture content
. energy content
. type of pollutants
The total selling price under Dutch conditions and at an exchange
rate of 1 $ - fl. 2.00 is $ 100 to $ 125 per ton.
157
-------
NBM BODEMSANERING
5. Legislation and licenses
In the case of the indirect thermal soil cleaning plant the
following licences were necessary:
. water outlet licences ("WVO vergunning"),
. general environmental protection ("hinderwetvergunning") valid
to August 1988,
. waste handling licences ("afvalstoffenwet") from August 1988.
The principal limits were put on wateroutlet emissions, noise
emission and stack emission. The actual licences permits the
following values concerning stack emissions.
Indirect thermal soil cleaning
Stack emission values of the actuel licences
Pollutants
Fluoride
Sulphur-
dioxide
Chloride
acid
Cyanides
total
Total
Aroma tes
Total
PAH's
Cadmium
Mercury
Dust
Limit value of waste
licence
5 mg/mo3
460 mg/mo3
75 mg/mo3
1 mg/mo3
5 mg/mo3
2 ug/mo3
0.1 mg/mo3
0.1 mg/mo3
75 mg/mo3
Measured value
( mean )
0.12
200
9.15
0.05
0.28
3.5
0.29
0.001
12
158
-------
NBM BODEMSANERING
6. Development
The results of the laboratory studies proved that the cleaning of
chlorinated hydrocarbons is possible at the NBM indirect heated
thermal cleaning plant.
The central and local authorities recognized this conclusion but
asked for trials on the full scale plant.
After three years of discussion the administration agreed with the
trials proposed, total cost a million Dutch guilders.
The trials are carried out in two phases.
6.1 Trial on incinerator
The original maximum value of products of incomplete combustion
(PIC's) was 1 ng out of 1 gram chlorinated hydrocarbons. It does
not appear to be feasible to determine this rest formation during
trials while the time fluctuations were too important.
It was decided to run tests on the incinerator using principal
organic hazardous constituers (POHC's) during normal production.
The following limits were set up:
1) stable temperatures in the incinerator in time and in space,
and stable gas velocities
2) combustion efficiency of more then 99,9%
3) destruction efficiency of POHC's of more then 99,99%
A schematic presentation of the incinerator is shown in figure
(4). During the test the temperature and gas velocity proves to be
stable indicating that the gas flow conditions were good and the
residence time distribution small enough.
The gas velocities are shown in figure (5).
Figure 4
umpllng
Figure 5
Gas velocity distribution in incinerator
Schematic Incinerator NBM Indirect thermal cleaning plant
40 ta 76 »»
Distance ( cm )
159
-------
NBM BODEMSANERING
The residence time varies between 1,8 to 2,1 seconds depending on
temperature and gas-flow.
The combustion efficiency was so good that the results mostly meet
the limits of 99,9%
The destruction efficiency was determined under the following
conditions:
POCH's trichloro ethylene
tetrachlore methane
Temperatures : 1100 C and 1200 C
Residence time: 1,8 —> 2,1 sec.
Oxygen content: +_ 10%
C1CH-CC1.
CCL,
The calculation of the destruction efficiency takes account of:
1) Values below detection limit, are counted at a minimum - 0,
maximum - detection limit
2) thrichloro ethylene concentration in cooling sampling nitrogen
gas of 115 ng/m3
3) tetrachloro methane concentration in cooling sampling nitrogen
gas of 57 ng/m3
4) sampling in triplo
The results are shown in the table below:
Destruction efficiency of trichloro ethlynene
sampling
1
2
3
Temperature 1100° C DE%
min . max .
99.9999 99.9998
99.9999 99.9999
99.9999 99.9999
Temperature 1200° C DE%
min. max.
99.9999 99.9999
99.9999 99.9999
99.9999 99.9999
160
-------
NBM BODEMSANERING
Destruction efficiency of tetra chloro methane
sampling
1
2
3
Temperature 1100° C DE%
min . max .
99.9999 99.9999
99.9970 99.9970
99.9900 99.9900
Temperature 1200° C DE%
min. max.
99.9999 99.9999
99.9999 99.9999
99.9999 99.9999
All the results meet the limits mentioned above. The administrati-
on, according to these results, allows phase 2.
6.2 Cleaning a batch of 200 tons HCH polluted soils and 500 tons
drins polluted soils
The following limits were set up:
1) cleaning results soil
"A"-value: HCH < 0.01 mg/kg d.s.
Drins < 0.01 mg/kg d.s.
2) Stack emission
Equivalent dioxines < 0.1 ng/nm3
The soils have been cleaned from 3 to 5 April 1990.
The cleaning results were as follows:
Pollutant
HCH's som
Drins som
Input value
mg/kg d.s.
5 - 50
500 - 1500
Output value
mg/kg d.s.
< detect lim 0.01 mg/kg
< detect lim 0.05 mg/kg
The stack emission values are not known at the moment of writing
this paper.
161
-------
_
LI
lf
-------
REVOLVING FLUIDIZED BED TECHNOLOGY
for the TREATMENT OF HAZARDOUS MATERIALS
BY: Geoff W. Boraston
Superburn Systems Ltd.
a Div. of Consolidated Environmental Technologies
Vancouver, B.C.
FOR PRESENTATION AT
SECOND FORUM ON INNOVATIVE HAZARDOUS
WASTE TREATMENT TECHNOLOGIES
DOMESTIC OR INTERNATIONAL
May 15 -17,1990
163
-------
Introduction
The 1990's is the decade of the environment and it is not surprising that technology
has advanced to meet the demands of our current crisis of waste disposal.
Incineration has been used for over a hundred years for waste disposal and because
of stricter environmental requirements it too has advanced to become recognized as a
safe and effective method of waste disposal.
Fluidized beds are also a combustion technique which has been used for decades
and have always been recognized as an efficient system. They are now being used
for waste disposal with excellent results. The Revolving Fluidized bed is a unique
type of fluidized bed which was developed particularly for waste material.
Background to Fluidized Bed Combustion
Fluidized beds have been around for a long time and date back to the early 1920's.
Because of certain unique features such as very high mass and heat transfer rates
between solids and gases/liquids, fluidized beds have been useful in a variety of
process applications including catalytic processes, drying and combustion. A fluid
bed is simply a bed of granular particles through which a flow of gas or liquid passes
upwards. The particles are suspended in the upwardly flowing stream and have the
appearance of a boiling liquid.
Feed Stock
Red Hot Bed
OfSand
Air Distributor
Plenum
AIR
Figure 1. Basic Principal of Fluidized
Bed Combustion.
With fluidized bed combustion fuel or waste is
burned in a turbulent bed of red hot inert
particles, normally sand, which is fluidized by
the combustion air. The basic fluid bed
combustor is shown in Figure 1. There are four
major components (1) plenum, or windbox
which first receives the combustion air from
the forced draft fan (2) the air distribution
plate which transmits the air from the windbox
to the fluidized bed (3) the fluidized itself
(which is usually sand, ash, limestone or a
mixture of these) into which the feed is injected
and (4) the freeboard area above the bed.
Fluid beds are designed with different bed
depth which can be anywhere between
164
-------
6 in. and 4 ft. deep. The bed of particles is suspended by the upcoming air and
expands to 1.1 to 1.8 times its static height, depending on air velocities, bed depths
and particle characteristics.
Prior to introducing the feed stock the bed is heated from the start-up temperature
to operating temperatures with the light up burners. Once the bed is hot, say
between 1400 F to 1800 F, the feed is introduced into the bed. The feedstock then
burns in the turbulent mixture of red hot solids and combustion air. Because of
the high heat transfer area the granular particles present and the mechanical action
in the bed the combustion is very rapid and very thorough and occurs in a controlled,
near ideal environment. At any given time, only a small portion of fuel is in the bed
and the bed material acts as a huge thermal flywheel which smooths out process
variations.
Most of the combustion occurs in the bed; however, the freeboard supplies additional
gas residence time for complete burn out of the flue gases. Secondary air or
recirculated flue gases are often added in the freeboard. This bed/freeboard
combination is really an integral combustor/after burner combination. The gas
residence time in the freeboard varies between 3 to 8 sec.
Ash is usually reduced to a sufficiently small size that it exits the combustor with
the flue gases. Larger objects which are introduced to the bed sink to the distributor
plate and are eventually discharged through a bed cleaning port.
Types of Fluidized Beds
For combustion applications there is generally two types of fluid beds, atmospheric
fluid beds and pressurized fluid beds. Atmospheric fluid beds operate at near
atmospheric pressures and usually at a slight vacuum produced by using both an
induced draft and forced draft fan. Pressurized fluid bed operate at high pressures
and are used in combined cycle power generation. Although pressurized fluid bed
seem promising as a clean coal burning technology with high efficiencies they have
not yet received commercial popularity. Atmospheric fluid beds are generally used
for power generation and waste disposal application.
There are different types and variations of atmospheric fluid bed, but historically
they have fallen into two categories, bubble beds and circulating fluidized beds.
Superburn Systems Ltd. uses a third type called the revolving fluidized bed which
is an enhancement of the bubble bed technique.
165
-------
A conventional bubble bed is shown in Figure 2. Bubble beds may operate with
fluidizing velocities between 4 and 8 ftVsec. These velocities are sufficient to keep
the bed fluidized and bubbling but are not high enough to carry over the bed
material into the flue
gases.
A circulating fluidized bed
is shown in Figure 2.
These beds operate at
much higher gas
velocities, than the bubble
beds, up to 30 ft/sec. As a
result a large portion of
the bed is carried over
with the flue gases. A
cyclone is used to remove
the bed material from the
flue gases so that it can be
returned to the bed.
Bubble bed
Revolving Fluidized
Bed
Circulating Fluidizes
Bed
Figure 2. Types of Fluidized Beds.
Revolving Fluidized Bed
The revolving fluidized bed (RFB) shown in Figure 3 is a major development of the
bubbling bed technique. Its enhanced mixing both laterally and vertically and inbed
circulation of material lends the revolving fluidized bed to applications which have
been more difficult for conventional fluidized beds. It is closer in characteristics to a
bubble bed than a circulating fluidized bed, but it is absolutely distinct in the way the
bed is caused to revolve in an elliptical pattern.
The characteristic features of the RFB are an inclined air distributor plate,
differential air velocities across the bed and an inwardly inclined deflector plate
above the bed of sand. The gas velocities on the low side of the distributor plate are
much higher than on the high side. As a result the sand rises from the high velocity
side and is deflected across the bed towards the low velocity side. This combination of
a rising and sinking bed along with the sloped distributor plate causes the bed to
revolve (circulate) in an elliptical pattern. The bed action is actually a churning
action where waves of sand move at regular intervals across the bed.
166
-------
The flow pattern of the bed material is such that the waste material introduced
onto the sinking bed region is engulfed by the red hot sand and mixed throughout
the bed volume. This ensures the most rapid heat transfer between the sand and
the waste material and an accelerated rate of combustion. Consistent agitation of
the bed material and burning feed stock ensures that no boundary conditions exists
and there is a high degree of combustion. The improved lateral mixing distributes
the feed stock evenly throughout the reactor volume and the inbed circulation allows
non-combustible objects to be easily discharged from the bed.
Light debris does not float on the surface of the bed but is embedded in the sand and
covered by the circulated sand. This ensures that a high portion of material is burned
within the bed where the most ideal combustion conditions exist.
As combustibles descend with the sinking portion of the bed, they undergo drying,
gasification and combustion. Volatiles are released fairly rapidly; however, the
majority of combustion still occurs within the bed due to the engulfing of the waste
material and the settling of sand on the sinking region.
AIR
Non-Combustibales
Figure 3. Single Revolving Fluidized Bed
Non-combustibles which could be anything from
tin cans to rocks travel with the moving bed of
sand as shown in Figure 3. They sink through
the low velocity zone and are carried
horizontally above the air distributor plate.
Once the non-combustibles reach the high
velocity zone they are not fluidized because of
their weight but are discharged through the ash
outlet. The residence time of the non-
combustibles in the bed is sufficient to ensure
complete burnout of any combustible portion
and can be controlled by the rate of ash
discharge and the circulation rate of the sand.
The inherent ability of the revolving fluidized
bed to easily discharge non-combustibles from
the bed distinctly separates it from the other
fluidized bed. The accumulation of rocks and
metal objects in fluidized beds can cause serious
problems and result in defluidization and
agglomeration of material.
167
-------
Foreign material can pile up along the air distributor resulting in deuuidization and
impairment of the combustion process. The revolving fluidized bed was originally
developed to overcome these weaknesses, and as a result, the system became popular
for municipal waste which contain a high percentage of non-combustibles.
The enhanced lateral mixing of the revolving fluidized bed improves the overall
quality of combustion. Bed temperatures are extremely even and the fuel is
distributed throughout the reactor volume. The enhanced mixing eliminates hot
and cold spots and greatly reduce the possibility of bed agglomeration. Any
agglomerated material which may form is
quickly broken down by the increased
turbulence or discharged from the furnace.
With the high degree of turbulence in the
bed, the revolving fluidized bed can handle
solids, liquid and sludges, while retaining
the majority of the combustion in the sand
bed. This leaves the freeboard volume for
the final stages of combustion in order to
ensure adequate destruction of organic
pollutants.
The single revolving fluidized bed shown in
Figure 3 can be placed face to face in order
to achieve the twin revolving fluidized bed
as shown in Figure 4. The performance of
these two bed configurations is identical.
The twin revolving fluidized bed is built
twice as wide as a single revolving fluidized
bed. For small applications, a single bed is
used and for larger applications, a twin bed
is used.
t t 1
AIR AIR AIR
Non-combustables
Figure 4. Twin Revolving Fluidized Bed.
Advantages of the Revolving Fluidized Bed Combustion
1. Efficient Combustion
The intimate mixing in the fluid bed results in high oxygen availability and hence
combustion efficiency. The revolving action of the bed ensures that the feedstock is
168
-------
carried away from feed location and mixed throughout the combustion chamber. The
high heat transfer area on the inert bed material and large thermal mass results in
rapid combustion at even temperatures. The constant collision of particles in the bed
reduces the ash layer which may be formed on organic particles exposing additional
combustible surface.
2. Feed Stock Variation
The revolving fluidized bed can handle a wide variety of feed stocks including
hazardous waste, industrial waste, municipal waste, sewage sludge and
contaminated soil. In addition the system is very tolerate to fluctuations in the
quality of the feedstocks. It can be designed to burn subautogenous or
superautogenous fuel or shift from one to the other.
Because the revolving fluidized bed contains such a large thermal flywheel changes
in the calorific valve or moisture content do not immediately upset process
temperatures. Changes in fuel consistency can be monitored by the (^reading and
drifts in operating temperatures. The process control system can detect slight
changes and make adjustments without upsetting the process.
3. Easily controlled
With the excellent mixing uniform temperatures and combustion conditions exist in
the fluid bed. This allows for simple and accurate process monitoring and simplified
autommation. Because the combustion is so rapid process response to changes in feed
rate and air supply is immediate.
4. Shift Operation and Shut-down
The combustion is rapid and very little fuel is held in the bed at any time so that shut-
down can be achieved almost instantly. This has advantages for both shift operation
and emergency shut-down.
The bed of sand loses its heat very slowly during shut down conditions,
approximately 40 F per hour, so start-up time can be very short.
5. Emissions
As with any combustion system air pollution control can be accomplished with a
169
-------
variety of air pollution control devices and the design of a complete system will
depend on the emission levels required by regulation. By nature of the fluid bed
many pollutants can be minimized at the combustion source.
In-bed control of sulphur gas emissions by direct lime addition to the bed is a well
known technique in fluid bed combustion. The calcium in the limestone reacts with
the SQs. to form gypsum, similarly, halide capture as calcium halide may also be
effected, reducing acid gas production. Sulphur removal depends on the sulphur in
the feedstock and the limestone characteristics, however 80% sulphur removal can
normally be achieved quite easily.
Fluid beds operate at relatively low temperatures so the formation of thermal oxides
of nitrogen from nitrogen in air is minimized. Fuel bound nitrogen primarily
contributes to the formation of NOx.
The fluid bed is unique in that the majority of ash is carried over as flyash and there
is little or no bottom ash. The amount of bottom ash will depend on the larger non-
combustible objects in the fuel. The flyash will be removed using either a baghouse or
electrostatic precipitator.
6. Maintenance
There are no moving parts in the combustion chamber itself since all the mixing and
material movement is achieved by air. This reduces the maintenance costs.
The highest maintenance item is the repair and replacement of refractory but with
the elimination of thermal shock resulting from the large thermal flywheel and
reasonable care during start-up and shut down, damage to the refractory is
minimized. Many fluid beds operate for over 10 years with only minor refractory
repairs.
7. Capital Cost
The fluid bed is a fairly simple system with relatively few easy to construct
components, so provided an elaborate waste treatment system is not required the
capital cost will be 60% to 85% that of a rotary kiln of equal capacity.
170
-------
8. Specific Advantages of the Revolving Fluidized Bed over other Fluidized Bed.
The revolving fluidized bed was developed in order to overcome specific weakness of
conventional fluidized bed, namely to improve lateral mixing and also to improve the
movement of large non- combustible objects out of the bed. This ability to easily
discharge inert lumps allows the revolving fluidized bed to burn waste material
which contains rocks, metal etc. Also because of the revolving action the feedstock is
drawn into the bed where the best combustion conditions occur. It is because of these
strengths that the revolving fluidized bed has found acceptance for feedstocks such a
sludge or soil containing rocks and municipal waste containing a variety of non-
combustible objects.
Development Status of the Revolving Fluidized Bed
Fluid bed combustion is a well known technique and has been used for decades. The
revolving fluidized bed was developed in the late 1960's particularly for waste
material and has been used commercially since the early 1970's. There are over 60
incineration and steam generation plants existing using the revolving bed technique.
Existing Installations: Number Installations:
Great Britain Coal and Industrial Waste 8
Coal and Hospital Waste 2
West Germany Soil Remediation 1 (under const.)
Wood 1 (under const.)
Municipal Waste 1 (under const.)
Japan Industrial Waste and Sewage Sludge 9
Municipal Waste 31
9 (under const.)
Tunisia Industrial Waste 1
Canada Various Waste 1
Hazardous Waste 1 (under const.)
Oil Refinery Waste 1 (under const.)
Lumber Industry Waste 1 (under const.)
171
-------
CASE STUDY
SYDNEY TAR PONDS CLEAN-UP PROJECT
Background
The Sydney Tar Ponds in Sydney, Nova Scotia, are one of the most hazardous
chemical waste sites in eastern Canada, and the largest chemical waste site in the
country. More than 700,000 Tons of coke oven residue was discharged over 80
years by the local steel mill into a creek bed. The results are oily black mud flats.
The coke oven residue contains a high concentration of hazardous organic
compounds, principally polyaromatic hydrocarbons and heterocyclic nitrogen
compounds. It has the appearance of wet tarry, muddy sand and has a high ash
content and a moisture content between 30 and 70%.
The Provincial and Federal Governments have sponsored a program to remediate
the site by excavating the waste product over 7 year period and treating it in an
incineration plant. The principal objective is to destroy the hazardous organics
compounds with an efficiency of 99.99%. As a bonus, the incineration process will
generate up to 10 MW of electrical energy which will be used by the steel mills new
electric arc furnaces.
The incineration plant currently under construction, is a revolving fluidized bed
supplied by Superburn Contractors and designed by Superburn Systems Ltd.
Superburn Contractors was selected because a series of test burns confirmed the
effectiveness of the system and it had the lowest construction costs, lowest
operating costs and best delivery schedule.
Test Burns
During the summer of 1988, Superburn Systems Ltd. conducted a series of trial
burns on the tar pond sludge in a 6 MBtu per hour, demonstration plant located in
Vancouver, B.C. The purpose of the tests were to demonstrate the contaminated
sludge could sustain combustion under steady state conditions and that 99.99%
destruction of the principal hazardous organic compounds could be achieved.
Approximately 29 tons of tar pond sludge were shipped to Vancouver from Sydney
under the hazardous waste transportation regulations. The sludge delivered
172
-------
represented a characteristic range of material consistency.
The test program concluded that the sludge could be burned under stable conditions
and that between 99.99% and 99.9999% destruction efficiency could be achieved on
the PAH's with an operating temperature of 1750 F and 1.7 sec gas residence time.
See Table 1, Appendix 1.
Sydney Tar Pond Project
Revolving Fluidized Bed Incineration Plant
Location:
Rated Throughput:
Steam Production:
Power Production:
Major Equipment:
Capital Cost:
Sydney, Nova Scotia
31,000 Ib/hr Tar Pond Sludge
100,000 Ib/hr
at 450 psi and 750 F
Up to 10 MW
Three day sludge storage lagoon
Two piston pumps
Two furnaces with F.D & Sec. air fans
One Boiler
One Baghouse
One ID. Fan
Stack
$16.25 million Canadian
Emission Criteria:
1. Stack emissions:
DRE
Sulphur Removal
Particulate
Carbon Monoxide
99.99%
80%
0.02 gr/scf
80 ppm
173
-------
2. Ground Level Concentration
Maximum Ground Level concentration at point of impingement as
calculated by the Ontario Ministry of Environment Dispersion Model.
Heavy Metals
Lead
Arsenic
Cadnium
Chromium
Copper
Zinc
(ug/m3)
10
1
5
5
100
100
PCDD's andPCDFs
JL x
450 (450X50)
Where X is the ground level concentration of PCDD's in pg/m3 and Y is the ground
level concentrations of PCDF's in pg/m3.
Plant Description (see Figure 5)
The primary purpose of this plant is to effectively dispose of the tar pond sludge in
an environmentally sound manner. Power will be produced, but the plant will
operate in a turbine follow mode which will ensure that the plant will always
operate under the most ideal conditions for sludge destruction.
The sludge is delivered to a storage lagoon divided into three cells, each cell
having one day storage capacity. There are two piston pumps, one for each
furnace which are fed from the storage cells by a gantry crane. The sludge is fed
directly to the furnaces and spread over the fluid bed by a steam injection nozzle.
A large freeboard section provides a 3.5 sec gas residence time in order to ensure
complete burn-out of the organic pollutants. Secondary air and flue gas is
introduced into the freeboard which both controls the temperature and provides
turbulence to aid complete burn-out. The flue gas from the two furnaces are then
ducted into a boiler.
The boiler used is an existing piece of equipment which is being refurbished. The
174
-------
incineration plant is built adjacent to the power house which contains the boiler
and the hot gas duct penetrates the building wall to enter the boiler. The steam
from the boiler is fed to an existing turbine.
The cool gas from the boiler is ducted out of the power house to a baghouse. The
treated gas then passes through a final heat recovery unit before being emitted to
atmosphere by the stack.
Steam Turbine
Stack
Induced
Draft
Fan
*
Ash
Collection
Figure 5. Flow Chart of Sydney Tar Pond Incineration Plant.
Lime is fed to the fluid bed in order to capture the sulphur and convert it to gypsum.
A fine lime powder will be carried over to the baghouse where it will form a fine
coating on the bags. This process will ensure maximum effectiveness of the lime.
Coal is used for plant start-up and also as a sweet fuel in the event the calorific
value of the tar pond sludge drops too low for good operation.
There are three ash stream, bottom ash for the furnace, economizer ash and baghouse
ash. The bottom ash from the fluid beds is extracted with water cooled screws. This
ash consists of the rocks contained in the sludge and bed material. The bed material
is separated and returned to the furnace. The economizer and baghouse ash is the
fine ash and lime which will make up approximately 95% of the total ash.
The control system is a Bailey Network 90 DCS and is design for fully automatic
operation with all the necessary safe guards to ensure ideal combustion condition
and safe plant shutdown. The overall plant is designed to operate 24 hours per day
310 days per year and dispose of the tarpond sludge over a 7 year period.
175
-------
APPENDIX I
TABLE 1 SYDNEY TAR PONDS TEST BURN
ORGANIC EMISSION RESULTS
Parameter
Date
Test Time
Duration
PAH DUE %
Moisture (Vol%)
C02 (Vol% Dry)
02 (Vol% Dry)
PCDD (ng/Nm3)
PCDF (ng/Nm3)
Gas Residence Time (sec)
Operating Temp. (oF)
Run 1
Furnace Outlet
July 23/88
1828-1924
56.0
99.996
10.6
14.0
9.7
-
-
1.78
1886 °F
Run 2
Furnace Outlet
July 24/88
1358-1454
56.0
99.996
11.7
7.5
14.0
.
-
1.57
1870 °F
Run 3
Furnace Outlet
July 25/88
123-1622
200
>99.99
9.1
10.0
10.7
-
-
1.68
1560 °F
Run 3
Stack
July 25/88
1223-1622
200
99.99
31.6
4.3
15.7
N.D.
N.D.
1.68
1560 °F
Note:All tests were taken at the furnace outlet before the air pollution control
equipment, except for Run 3 where a stack test was taken as well.
N.D. - Not Detected
176
-------
APPENDKII
TAR POND SEDIMENT CHEMICAL AND THERMAL CHARACTERISTICS
A. PAH CONTENT
The primary POHC's (Principal Organic Hazardous Constituents) to be destructed in
the tar ponds sediments are PAH's (Polynuclear Aromatic Hydrocarbons). A review
of the analytical data for the tar pond sediment PAH content indicated that average
PAH content of the sediment decreases with distance downstream of Coke Oven
Brook. This parallels the overall decrease in grain size in the downstream direction.
However, considerable fluctuation occurs in any given area. The average total PAH's
at the 0.6m depth by location, are summarized below:
Area Average total PAH (mg/kg)
Wash Brook Arm 3,700
Remainder of South Pond 12,000
North Pond to Narrows 8,000
North Pond beyond Narrows 5,500
Reviewing trends with depth, the data (as summarized below) show little variation
over the top two depth intervals and with a noticeable reduction below 1.2m. Even
taking into account the fact that some of the samples have a small portion of the
underlying clean materials incorporated in with the tar pond sediments, total PAH
generally declines with depth.
Depth Average total PAH (mg/kg)
0.00-0.61 9,000
0.61-1.22 8,200
1.22-1.83 2,300
1.83-2.44 6,100*
2.44-3.05 1,400
* = one sample enormously high
177
-------
The general order of PAHs in terms of their relative abundance is as follows:
1. Naphthalene 2. Phenanthrene
3. Benzo (a) anthracene 4. Pyrene
5. Fluoranthene 6. Anthracene
Naphthalene and phenanthrene account for about 60 to 80% of the total PAH content.
These two compounds are also the two most variable.
The minimum, maximum and average values for the PAH's analyzed are tabulated in
Table 1 included in the Appendix to this section.
B. HNC CONTENT
The concentration of total HNC's is less variable than total PAH's with no clearly
discernable trend except to note that HNC's generally correlate with PAH's.
The general order of HNC's in terms of their relative abundance is as follows:
1. Benzacridines 2. Benzocarbazole
3. Acridine 4. 7,8 Benzoquinoline
S.Methylated Benzoacridines 6. Methylated Benzocarbazoles
The first three HNC compounds comprise 60-70% of the total HNC content. It is
therefore to be expected that these compounds will be named as POHC'c for which a
99.99% DRE must be demonstrated during trial burns.
The minimum, maximum and average concentration of selected HNC's obtained from
13 analyses is as follows:
HNC's (mg/kg)
14
Maximum 1,510
Average 169
178
-------
The selected UNO's were composed of:
Methylated Benzocarbazoles
2, 4 Dimethylquinoline
Acridine
Benzacridines
Dimethylquinoline
Methylated Benzoacridines
Carbazole
Quinoline
Benzocarbazol
A more detailed breakdown of tar pond sediment HNC content is given in Table 2 of
this section.
C. HEAVY METALS CONTENT
The most distinctive trend for heavy metal concentrations in the tar ponds is for zinc,
lead and to a lesser extent, copper to increase in concentration near the mouth of
Muggah Creek. This suggests an association with the finer-grained sediments (silts
and clays), which in turn may reflect adsorption due to charge and increased surface
area with smaller sized sediment particles. The concentrations of arsenic and
chromium are lower and more variable over the Tar Ponds.
Zinc is generally the most abundant heavy metal followed by lead, copper or arsenic,
andthen chromium.
The minimum, maximum, and average concentration for the following heavy metals
is as tabulated below:
Metal
Arsenic
Chromium
Copper
Lead
Zinc
Average
Concentration
(mg/kg)
11
6
48.4
131
218
Average
Concentration
(mg/kg)
197
58
321.8
741
2,112
Average
Concentration
(mg/kg)
102.4
15.7
118.7
309
560.5
179
-------
D. HEATING VALUE AND CHEMICAL ANALYSIS
The following tar properties represent the results of some sample analyses conducted
on tar pond sediment samples obtained from the tar pond sampling program.
Moisture content: The air dried moisture content of the samples
range from 10.64 to 69.2% on a weight basis.
Average content was 35.3%. this moisture
content represents the weight percentage of
the original sample that was evaporated
when the sample was air dried.
Proximate Analysis: (% w/w, air-dried)
Range Average
Moisture 1.06 - 9.57% 2.81
Ash 19.33 - 70.68 40.16
Volatiles 13.42 - 33.20 25.84
Fixed Carbon 8.61 - 48.27 31.19
100
Ultimate Analysis (% w/w, air-dried)
Range Average
Moisture 1.06 - 9.57*
Ash 19.33 - 70.68
Sulphur /1.00 - 5.66
Carbon /21.1Q • 70.78
Hydrogen 1.49-4.47
Nitrogen 0.25 - 3.02
Oxygen
H.H.V.kj/kg air-dried 7,073 - 2,9510.0 22,036
(Btu/lb) (3,140 - 12,687) ( 9,474)
180
-------
"This moisture represents that quantity of moisture which remained after air drying,
but which was subsequently removed when the samples were dried in accordance
withASTMD3173.
Note that the proximate analysis is based on the first 13 samples analyzed while the
ultimate analysis is based on some 63 tar pond sediment samples that were analyzed.
Additional Chemical Analysis:
Several chemical analyses were performed on the six samples noted above under
Laboratory Determined Values, Tar Ponds Sediments and Physical Characteristics
The results of these additional chemical analyses are as follows:
Chlorine content: 0.25 to 0.66% by weight
Sodium content: 0.22 to 0.90% by weight
Vanadium content: 29 to 30 ppm
Tar Pond Sediment Ash Characteristics:
a) Physical
Ash Fusion Temperatures
Range Average
Initial Deformation temperatures 1160 -1216 1187
Softening Temperature 1232 -1338 1276
Hemispherical Temperature 1266 -1360 1320
Fluid Temperature 1349 -1449 1390
181
-------
TOXIC CHEMICAL
SPECIALISTS
ECO LOGIC 000
Thermal Gas Phase Reduction of
Organic Hazardous Wastes
in Aqueous Matrices
by
Douglas J. Hallett, Ph.D.,
President
Kelvin R. Campbell, P.Eng.
General Manager
Wayland R. Swain, Ph.D.,
U.S. Vice-President
ELI Eco Technologies Inc.
prepared for
EPA Second Forum on Innovative Hazardous
Waste Treatment Technologies:
Domestic and International
Philadelphia, PA
May 15-17, 1990
182
-------
Eco Logic International, through its subsidiary ELI Eco Technologies, will
be demonstrating a new type of hazardous waste processor for Canada's Department
of National Defence as part of the Defence Industrial Research Program. The test
program will be taking place during the summer and fall of 1990. Of direct
importance to the military worldwide is the ability to clean up and process PCBs,
chlorinated solvents, chlorinated dioxins, etc. which may be contained in harbour
sediments, soil, hydraulic fluids, cleaning fluids, defoliants, mustard, and
residual chemical warfare agents.
The thermo-chemical reaction that makes the Eco Logic process possible is
the ability of hydrogen to dechlorinate organic molecules at elevated
temperatures. Bench-scale tests have shown that a well-mixed combination of
hydrogen and chlorinated organic waste subjected to 850°C or higher for a period
of 1 second will result in 99.9999% destruction or better. Consistent results
have been obtained for over 100 tests during the last 16 months. The end
products of the reaction are HC1 and dechlorinated organics themselves broken
down by hydrogenation to methane and ethylene.
The Eco Logic process is not an incineration technology. Destruction of
chlorinated organic waste using incineration and pyrolytic processes is
accomplished by breaking contaminant molecules apart with high temperatures and
combining them with oxygen, usually from air. PCBs first fragment into
chlorobenzenes. which sometimes combine with oxygen to form dioxins and furans,
which are more toxic than the original PCBs. The Eco Logic process uses hydrogen
at elevated temperatures to reduce, rather than oxidize, chlorinated organics.
Since there is no free oxygen in the reducing atmosphere, no dioxin or furan
formation is possible. As well, since combustion air is not required, there is
no nitrogen to use up reactor volume and heating, resulting in the reactor being
much smaller than an incinerator handling the same throughput.
The chemical reaction is shown on Figure 1. Hydrogen and a PCB molecule
with 4 chlorine atoms attached are reacted above 850°C to form HC1 and benzene.
In the second reaction, which occurs at the same time, benzene and hydrogen react
to produce methane (and some ethylene). In the third reaction, a non-chlorinated
alkane hydrocarbon reacts with hydrogen to form methane. The presence of water
enhances all of these reactions, as does an excess of hydrogen, and an increased
residence time in the reactor.
Figure 2 shows a schematic of the reactor module designed to accommodate
the thenr.o-cherr.ical reaction. A mixture of preheated waste and hydrogen is
183
-------
injected by nozzles mounted tangentially near the top of the reactor. The
mixture swirls around a central ceramic tube past glo-bar heaters and is heated
to 850°C by the time it passes through the ports at the bottom of the ceramic
tube. Particulate matter up to 3 mm diameter not entrained in the gas stream
will impact the hot ceramic walls of the reactor, thereby volatilizing any
organic material associated with the particulate. That particulate will exit
out the reactor bottom to a quench tank, while finer particulate entrained in
the gas stream will flow up the ceramic tube into the exit elbow and through the
retention zone. The reduction reaction takes place from the bottom of the ceramic
tube onwards, and the organic waste is completely dechlorinated and reduced to
methane, ethylene, and hydrogen chloride. Depending on the water content of the
waste, carbon monoxide and hydrogen may also be formed from the reaction of water
with methane.
The reactor module is six feet in diameter and nine feet tall without the
exit elbow. Photos of the reactor steel shell under construction are shown
following Figure 2. Figure 3 shows the process equipment layout for mounting
on the two drop-deck transport trailers. Photographs of the trailers follow
Figure 3. The layout shows the reactor (R) with the attached exit elbow (E) and
extension (EXT) leading to an HC1 and particulate scrubbing system (SCRUB). In
the scrubber, the gases are first cooled by direct injection of water spray from
a later condensation step. Removal of HC1 and particulate is accomplished using
a wet contact scrubber and lime scrubber water.
An amount of water approximately equal to the water in the waste stream
is carried through the scrubber as humidity in the hydrogen and hydrocarbon gas
stream bv controlling the scrubber temperature. This prevents rapid accumulation
of scrubber water waste. Most of the water carried through is then condensed
out using a plate heat exchanger, leaving relatively dry excess hydrogen,
methane, ethylene and carbon monoxide. Approximately 90% of this dry gas is
recycled into the reactor in order to use up excess hydrogen, and the remaining
10% is sent to the propane boiler as supplementary fuel. The boiler produces
steam for the heat exchanger that preheats the waste stream prior to injection
into the reactor.
The only gas emissions are from the boiler, which will require a small
stack, but since the fuel going into the boiler is very clean, and contains no
chlorine, the emissions will be insignificant. Effluent from the process will
be stored and analvsed in batches before release. It will consist of grits from
the bottom exit of the reactor, calcium chloride and particulate sludge from the
184
-------
HC1 scrubber, and condensed water. Some of that water will be recycled into the
process for primary cooling in the scrubber. Any effluent not conforming to
emission standards can be processed through the system again.
The process control and emission monitoring systems for the waste processor
are interlocked to provide an extra measure of safety, and to continuously
document proper destruction efficiencies. The V&F CIMS-500 chemical ionization
mass spectrometer is capable of measuring PCBs and their breakdown products on
a continuous basis, so that any potential process upsets result in an immediate
automatic halt to waste processing. As well, when the process is operating
normally, a continuous readout of destruction efficiency is obtained. During
the demonstration program, regulatory test methods using Modified Method 5 stack
testing trains will be used to confirm destruction efficiencies.
Net Benefits of the Eco Logic Process
Mobility/size - Incineration systems are carried on a number of large
transport vehicles, require weeks of set-up time, occupy large areas when set
up (football fields) and as a consequence of such overheads, must be used with
high volume, lengthy burns. Eco Logic's waste processor requires two standard
tractor trailers, is completely mobile (compared to transportable), requires only
davs to set up, and occupies little more area than the vehicles. Minimum runs
may be less than a single unit's daily capacity.
Scale of job - The continuous throughput process is well-suited to high
volume long run jobs, compared to sodium and potassium stripping methods, which
are typically small batch processes. Throughput capacity of the Eco Logic
process can be increased by ganging reactor units on a single ancillary support
and control system, allowing flexibility of operation, and redundancy of design.
Aqueous Content - Some of the largest and most serious contaminant clean-
up requirements are soils and sediments having a high water content.
Incineration technologies consume very large amounts of energy to heat up the
water component to the incineration temperature. As well, because they use air
(79% nitrogen) for combustion, and must combust all the organic material, they
require approximately 10 times the volume for the same residence time of
reaction. Sodium and potassium processes are precluded from treating water-
bearing wastes because of explosive reaction potential. Eco Logic's destructor
185
-------
must also heat water, but to temperatures which are 30% lower, and the water
component actually enhances the chemical reduction process. As well, the use
of hvdrogen (no excess nitrogen) in the reaction reduces the gas volume of the
products and therefore the size of the reactor required.
No Dioxin/Furan Emissions - Since oxygen is not used in the reaction,
formation of chlorinated dioxin or furan molecules is precluded. Furthermore,
if a process upset does occur, the CIMS-500 continuous organic emission monitor
will automatically divert the sidestream gas flow to the boiler back into the
reactor, so that no air emissions occur. The waste stream would be shut off
automatically and the gas stream recirculated until the problem was rectified.
A charcoal filtration unit prior to a divert stack is available if the reactor
had to be purged.
Cost - The combination of hardware requirements and process characteristics
suggest lower capital cost for the Eco Logic system, by a factor of 5 to 10
tir.es, compared to incineration processes. Operating economies are predicted
to be frorc 3 to 5 times lower than incineration technologies of comparable
capacities.
186
-------
FIGURE 1
THERMD-CHEMICAL REDUCTION REACTIDNS
Cl
/
Cl
5 H
Cl Cl
+ 4 HCl
9 H
6 CH4
-l> H
CH
187
-------
FIGURE 2
THERMQ-CHEMICAL REDUCTIDN REACTOR
/
REFRACTORY
GLDBAR HEATER
CERAMIC TUBE
188
-------
-------
IO
o
FIGURE 3 - PROCESS EQUIPMENT LAYOUT
-------
I6T
-------
X*TRAX™ - TRANSPORTABLE THERMAL SEPARATOR
for
ORGANIC CONTAMINATED SOLIDS
Carl Swanstrom
Carl Palmer
Chemical Waste Management, Inc.
Geneva Research Center
1950 South Batavia Avenue
Geneva, Illinois 60134
Presented at:
SECOND FORUM ON INNOVATIVE HAZARDOUS
WASTE TREATMENT TECHNOLOGIES:
DOMESTIC AND INTERNATIONAL
Philadelphia, Pennsylvania
May 15-17, 1990
ABSTRACT
Chemical Waste Management (CWM) has developed a patented system, X*TRAX™, that thermally
separates organics from solids, such as soils, pond sludges and filter cakes, in an indirectly heated
rotary dryer. Vaporized organics and water are transported with a nitrogen carrier gas to a gas
treatment system where they are condensed and collected as a liquid. The gas is then reheated and
recycled to the dryer. To control oxygen, a small portion of the carrier gas is vented to atmosphere
through carbon adsorbers. CWM has constructed a full scale transportable X*TRAX system that has
been contracted for a PCB soils cleanup at a mid-size Superfund site in the Eastern U.S. The system
is expected to mobilize in mid to late 1990 on that site. In an innovative combination, the condensed
organic liquid will be chemically dechlorinated on site prior to off site disposal. A SITE demonstration
test will be conducted at this site during the performance of the cleanup. Since early 1988, CWM has
operated both a 5 ton/day pilot demonstration system, and a 2-4 Ib/hour laboratory system for
treatability studies. The pilot system has recently completed ten tests using TSCA regulated PCB soils
and is currently in preparation for an extensive testing program using RCRA regulated materials. In
addition to a number of surrogate wastes, the laboratory system has processed over 19 RCRA and
TSCA regulated waste materials.
192
-------
X*TRAX™ - TRANSPORTABLE THERMAL SEPARATOR
for
ORGANIC CONTAMINATED SOLIDS
1. INTRODUCTION
The widespread problem of soils and solids that are contaminated with organic chemicals, coupled with
the EPA's increased restrictions on organics in landfills has resulted in the unavoidable fact that millions
of cubic yards of soil and solids will have to be treated to reduce or eliminate the organics.
Historically, the most likely treatment alternative has been high temperature incineration, which is
costly, difficult to permit, and requires lengthy mobilization periods for system installation and trial
burns. Chemical Waste Management (CWM) believes that many of these waste streams can be
treated using a thermal separation system; in essence, by drying them. Wastes such as contaminated
soils, pond or process sludges, filter cakes and others are likely candidates. Laboratory and pilot
testing by CWM has shown that at low temperatures (500-800 °F) many organic compounds including
high boiling compounds (PCBs) can be successfully separated from solids such as soil, sand, etc.
Thermal separation is now a treatment option with significant advantages in all of the above mentioned
areas for a broad class of waste materials that have relatively low organic concentrations - typically
less than 10%.
Contaminated solids are heated in an indirectly fired rotary dryer to volatilize the organics. The vapors
are carried to a gas handling system with an inert gas where they are scrubbed for particulate solids
and cooled to condense the organics. The carrier gas is reheated and recycled to the dryer. The
recovered organics can be reclaimed, used on- or off-site as supplemental fuel or destroyed by
incineration. Organic contaminants can range from high boiling, semi-volatile compounds such as
PCBs, to low boiling, volatile compounds such as RCRA regulated solvents.1-2 This X'TRAX process
has been granted U.S. Patent No. 4,864,942.
CWM has been actively developing the X'TRAX process since late 1986. This paper describes the
X*TRAX process and each of the three X'TRAX systems that have been constructed: laboratory, pilot
and full scale. Test data are presented for both the lab and pilot systems. The first full scale unit has
been functionally tested and will be moved to a Superfund site in mid to late 1990, depending on
receipt of approvals from EPA.
2. SYSTEM DESCRIPTION
The X'TRAX system is a separation process to remove volatile or semi-volatile compounds from a solid
matrix. Thermal energy is the driving force used to affect the separation. The process flow diagram
is presented in Figure 1.
Feed material, which can be either solid or pumpable sludge, is fed into the dryer. The dryer is an
externally fired rotary kiln. It is essentially a sealed rotating cylinder with the feed material tumbling
inside and the heat source (propane burners) on the outside. Since the dryer is externally fired, the
combustion products do not contact the waste material (feed) being processed. The use of an
externally fired dryer has two distinct advantages. First, and most important, is that the combustion
gases do not pass through the associated air pollution control (APC) devices. Propane is a readily
available clean burning fuel. Air permits for vent stacks from propane combustors are easily obtained,
usually without any required APC devices. This allows the APC devices for X'TRAX to be one tenth
to one hundredth the size of that for an equivalent capacity incinerator. In addition, the small volume
of nitrogen carrier gas discharged makes cleaning it to very high standards quite inexpensive. The
second advantage of external firing is that it makes the X'TRAX system a separation process, not an
incinerator because no organic combustion occurs. It is usually much easier to permit a separation
193
-------
X*
TRAX: PROCESS FLOW DIAGRAM VENT^^ CARBON
nir I DRUMS 1
^ffo ^—L-^nuwj ^-i—^
1 *~ ' '
PRIMARY VWMMRY f
»rJ~i r OIWWR CONDENSER m TFR fv7!
EDUCTOR -™' ™ &7^ !,
SCRUI3BER LOW T ^
r . • REHEAT
i, '
UkGANICS — pHASE J CONDENSATE
rrn — ^TDRAlT r?
SLUDGE J±^ MAKEUP 1JSJ «?
\
HIGH T
REHEAT
t I
ROTARY DRYFR ~ ^^r™
— — -- -— •
1
UKY KKUUUU — J
Figure 1
process than a waste incinerator.
The heated solids are discharged from the dryer as a powdered or granular dry material. For most
applications, water will be mixed with the exiting solids to cool them and to prevent dusting. By
adding reagents at this point, metal containing wastes can be stabilized with the reagent cost being
the only additional expense. The water will normally be condensate from the gas treatment portion
of X*TRAX.
The carrier gas first passes through a liquid scrubber where entrained solid particles are removed and
the gas stream is cooled to its saturation temperature. The scrubber also removes a portion of the
volatilized organics. The recirculated scrubber water continuously passes through a phase separator.
The phase separator collects any condensed light organic from the liquid surface and continuously
discharges a bottom sludge containing solids, water and organics. The sludge is dewatered using a
filter press. The dewatered solids are either returned to the feed stream or disposed of.
The scrubbed gas passes to the first heat exchanger where it is cooled to 10°F above ambient
temperature. This heat exchanger will produce the bulk of the liquid condensate. The carrier gas now
goes to a second heat exchanger where it is cooled to 40°F. The liquid condensates from both heat
exchangers are mixed and allowed to gravity separate. Organics are removed for disposal. The
condensed water is used to cool and dedust the treated solids exiting the dryer.
The 40 °F carrier gas now contains some residual moisture and organics that were present in the feed
at levels equal to or less than their equilibrium saturation concentration at 40 °F. The carrier gas then
passes to a recirculation blower. After the blower, 5 to 10% of the carrier gas is vented, and the
remainder is heated to 400-700°F before returning it to the dryer.
The process vent gas stream passes through a particulate filter (2 micron) and then through a carbon
adsorber, where at least 80% of the remaining organics will be removed. Actual practice has shown
removal efficiencies by the carbon ranging from 89 to 98%. This gas is then vented to the
194
-------
atmosphere. A 100 ton per day X*TRAX system would release anywhere from 0.25 to 5 pounds per
day of VOCs which is considerably lower than most regulatory constraints.
2.1 Laboratory
Since January, 1988, CWM has operated a laboratory X*TRAX system at its Riverdale Technical
Center in Riverdale, Illinois. This system typically processes 2 to 4 Ib/hour. It consists of a 4-inch
diameter, 48-inch long electrically heated tube furnace coupled to a small scale gas treatment system
that closely simulates the pilot and full scale systems.
This unit is used for treatability studies and to screen materials for pilot testing and commercial
operations. To date, 23 separate test runs have been performed, with 19 being on actual RCRA and
TSCA waste materials. The laboratory system was operated under CWM's TSCA R&D permit for the
Riverdale Center, as well as CWM's Illinois authorization for RCRA treatability studies.
In September of 1989 the system was transferred to Chem-Nuclear Systems Inc. (CNSI) in Barnwell,
SC. CNSI is using the system to evaluate the applicability of X'TRAX to treating mixed
(radioactive/hazardous) wastes, having conducted two tests to date. A second laboratory X*TRAX
system has been constructed and is operational at the new CWM R&D facility located in Geneva, IL.
A new TSCA R&D permit was granted on January 30, 1990.
2.2 Pilot System
The pilot X*TRAX system is a mobile unit mounted on two semi trailers; one containing the dryer and
another containing the gas treatment system. The dryer is 24-inch diameter, 20 feet heated length,
with 10 propane burners. The pilot system has a nominal capacity of 5 tons per day for a feed
material containing 30% moisture. Figure 2 is an artist's rendering of the pilot X'TRAX system.
The pilot system was used to provide design data on capacity, material handling, and gas system
performance for the full scale system. It has been and continues to be used to provide treatability and
emissions data on candidate waste streams, and is available for the performance of demonstrations.
The pilot system became operational in January, 1988. Since then it has been used to test over 87
tons of materials, including: 59 tons of simulated RCRA wastes, 5 tons of mixed radioactive/hazardous
waste and 20 tons of TSCA regulated PCB soils.
The pilot system is presently installed at CWM's Kettleman Hills Facility in central California. The
system is operated under a variety of permits at Kettleman. The most basic of these is an operating
permit from Kings County, allowing CWM to have an air emission source. CWM also has a variance
from the California Department of Health Services to treat non-RCRA wastes such as California special
wastes. The testing on PCB materials was conducted under a three month R&D permit from the EPA's
TSCA branch, which expired October 4, 1989. A 90 day extension was granted starting November
1, 1989. CWM has also filed for a RCRA RD&D permit to allow for testing on RCRA regulated
materials. This permit request is currently under review, and is expected to be granted during the third
quarter of 1990. CWM will perform one year of testing under the RD&D permit, investigating all types
of RCRA wastes but focusing on those that have or will be restricted from land disposal ("land
banned").
2.3 X*TRAX Model 200 Full Scale Production System
The X*TRAX Model 200 is a full scale production system that was constructed for onsite cleanup of
contaminated soil. The system is capable of treating 125 tons per day of contaminated soil with a
moisture content of 20%. Like the pilot system, the Model 200 has a rotary dryer and a gas treatment
195
-------
0>
00
i-t
n
I
T)
-------
system; however, they are much larger, requiring the use of modular construction techniques. The
Model 200 is fully transportable, consisting of three semi trailers, one control room trailer, eight
equipment skids and various pieces of removable equipment. Figure 3 is an artist's rendering of the
Model 200 system. The area required for the equipment measures 120 ft. by 120 ft.
All of the equipment has been designed for over the road transport anywhere in the U.S. or Canada.
The dryer is the largest of its kind that can be transported over the road. The components are
mobilized to the project site and assembled using a relatively small 15 ton crane. Approximately three
to four weeks are required to completely install the equipment. Site preparation involves grading the
site level and providing a firm base such as compacted gravel. Concrete footings are not usually
required; however concrete housekeeping pads may be required. All skids or trailers that normally
contain liquids have integral liquid containment curbs for spill control.
The system requires three phase, 460 volt electric power, propane storage tanks, and a liquid nitrogen
storage tank. The electric motors are sized such that the system can be operated from a commercially
available diesel generator if electric power is not available at the site.
Construction of the Model 200 system has been completed and the unit has been functionally tested.
CWM has a series of performance tests planned during which the unit will be operated on non-
regulated feed materials.
3. SITE DEMONSTRATION PLANS
In 1988 EPA contracted with CWM to perform a demonstration of X*TRAX under the Superfund
Innovative Technology Evaluation (SITE) program. Under this contract, EPA will perform sampling and
prepare the various reports and CWM will provide the equipment, waste streams, and permits, as well
as operate the equipment. CWM was to use the five ton/day pilot X*TRAX system on two or three
different PCB contaminated soils during the TSCA phase of the pilot testing that CWM conducted at
its Kettleman Hills Facility. Because of the difficulty of coordinating the schedule for development of
a SITE demonstration test plan with the short duration permits that CWM received for the TSCA
testing, EPA elected not to perform the demonstration as planned.
However, as a result of pilot work done by CWM during the TSCA testing at Kettleman, the company
was awarded a contract to clean up a PCB contaminated site using the X*TRAX Model 200, full scale
system. CWM and EPA have agreed to perform a SITE demonstration during the active part of the
remediation. This is currently scheduled for late 1990 or early 1991. A demonstration using the
Model 200 will fully evaluate the operational capabilities of the X*TRAX technology, as well as provide
treatability data.
The site of the remediation is primarily contaminated with PCB's, with some contamination from
chlorinated solvents such as trichloroethene and perchloroethene as well. The site has a sandy soil
underlain by clay. A marshy wetland area of the site with high humic content will also require
treatment. The X*TRAX Model 200 will be used to clean the soil, which will then be returned to the
site as backfill. A cleanup standard of 25 ppm was established in the Record of Decision (ROD) for
soil returned to the non-wetland portion of the site.
An interesting requirement of the ROD was that the PCB's be dechlorinated on site as part of the
treatment. In order to perform a chemical dechlorination cost effectively, CWM elected to treat the soil
using X*TRAX, and to dechlorinate only the condensed, recovered oil. The ROD had originally been
developed assuming that the dechlorination would be performed on bulk contaminated soil. By
dechlorinating the condensed oil from X*TRAX, CWM has greatly simplified this step. Also, a much
smaller reactor and less excess chemicals can be used. Furthermore, no chemicals are added to the
soil, avoiding the need to establish a secondary treatment standard for residual reagents in the soil
returned to the site.
197
-------
4. TEST RESULTS
CWM has performed extensive testing as part of the X'TRAX development program. Both simulated
waste materials and actual regulated materials have been tested in the lab and pilot systems. The
results of most of these tests have been previously reported1'3. The key results to date are summarized
here. All values for feed and product concentrations are on a dry basis, as is the EPA's convention.
Although the testing program has focused on PCB's to date, this should not be interpreted as a
restriction on the X*TRAX process from other organic materials. To the contrary, CWM has pursued
PCB testing as a means to validate the process on one of the most difficult organics to thermally
desorb (because of its very high boiling point). Processing conditions that are successful with PCB's
will almost certainly work for other volatile and semi-volatile organics. Also, a large percentage of
Superfund sites that have organics as the primary contaminants, have PCB's as one of the POHCs.
4.1 Lab Testing Program
The principal use of the laboratory system has been in performing treatability studies for waste streams
or remediation sites that are candidates for the X'TRAX process. Treatability studies are performed
on a 50-100 Ib sample, and yield results much quicker and at less expense than pilot testing. In all,
23 different test runs have been performed to date, 19 on actual RCRA or TSCA regulated materials.
As a benchmark of the system's capability, a simulated Superfund soil mixture prepared for the EPA
was tested. Table 1 presents the results. This material was originally referred to by EPA as the
Synthetic Analytical Reference Matrix, or SARM. It is now called Synthetic Soil Matrix, or SSM. SSM-
1 had high organics concentration and low metals concentration. For both the volatile and semi-volatile
organics, better than 99% removal was achieved.
Table 1
LABORATORY X"TRAX™
SSM-I
Compound
Feed Cone
(ppm)
VOLATILES:
Acetone
Total Xylene
Ethylbenzene
Styrene
Tetrachloroethylene
Chlorobenzene
1,2-Dichloroethane
3,200
2,900
1,900
240
180
130
46
Product Cone
(ppm)
Removal
(%)
16.0
9.50
5.20
< 0.005
0.094
0.180
0.062
99.5
99.7
99.7
> 99.99
99.95
99.86
99.87
SEMI-VOLATILES:
Anthracene
Bis(2-Ethylhexyl)Phthalate
Pentachlorophenol
3,100
3,020
397
12.0
< 0.33
2.8
99.6
> 99.99
99.3
198
-------
To date, eleven lab tests have been performed using TSCA regulated, PCB contaminated soils. Many
of these tests were developmental, testing different gas system configurations. The results of this
work have been summarized in Table 2.
Table 2
LABORATORY X'TRAX™
PCB Contaminated Soils
Run #
RS0829
GR0524
GB0504
Matrix
Sandy
Silty Clay
Topsoil
Peed
(ppm)
5,100
962
172
Product
(ppm)
9.7
21
2.8
Removal
(%)
99.3
97.8
98.4
All of these materials had other minor organic contaminants, all of which were generally reduced to
less than detection limits in the treated product (typically in the ppb range).
The most recent test was a series of four tests on soil, pond sludge and mixtures thereof. These
materials were all from the same site, which is a large remediation project estimated to have in excess
of 500,000 cubic yards of contaminated material. The organic contamination was a complex mixture
of chlorinated semi-volatile organics, aromatics and organic solvents. A summary of the test results
is presented in Table 3. Treatment standards for this remediation have not yet been developed. If
required treatment levels are set on a risk basis, it is very probable that the X'TRAX treated product
would be acceptable.
Presently eight treatability study samples are scheduled for testing in the lab system for this summer.
The agenda is quite full, and continued activity in the area of thermal separation indicates that lab
testing will continue to be in high demand.
4.2 Pilot Testing Program
CWM has performed 57 pilot test runs over the last 2-1/2 years using a variety of simulated, mixed
and TSCA wastes. This has resulted in the generation of a wealth of treatability data as well as
confirmed the operational reliability of the basic process equipment in the X'TRAX system.
Early in the development of the process, a large number of simulated waste streams were tested using
a variety of organic chemicals to spike the feed. These results are summarized in Table 4 as an
indication of system performance on various organic chemicals.
199
-------
Table 3
LABORATORY X»TRAX™
Non-PCB Soil, Sludges and Mixture
(Cone = mg/kg)
Run No.
DB0627
Clay Soil
DB0629
Soil/Sludge
DB0706
Sludge
DB0710
Sludge
Parameter
Total Solids (%)
Azobenzene
3,3'-Dichlorobenzidine
Benzidine
2-Chloroanaline
Nitrobenzene
Total Solids (%)
3,3'-Dichlorobenzidine
Azobenzene
Benzidine
Total Solids (%}
Azobenzene
Toluene
3,3'-Dichlorobenzidine
2-Chloroaniline
Benzene
Benzidine
Aniline
Total Solids (%)
3,3'-Dichlorobenzidine
Azobenzene
Concentration
Feed
94.1
3,190
1,820
842
828
45.6
73.1
958
61.0
17.8
52.4
47,900
4,470
3,590
2,100
1,870
1,010
267
47.0
1,070
35.7
Product
100
4.9
<0.66
ND
ND
<0.33
100
<0.66
ND
ND
100
327
<0.42
18.4
47.5
<0.21
3.7
43.3
100
<0.66
ND
Removal (%)
N/A
99.8
>99.96
—
—
>98.6
N/A
>99.0
-
~
N/A
99.3
>99.99
99.5
97.7
>99.99
99.6
83.8
N/A
>99.94
—
Table 4
PILOT X*TRAX™
Surrogate Feed Materials
Compound
Methyl Ethyl Ketone
Tetrachloroethylene
Chlorobenzene
Xylene
1 ,4-Dichlorobenzene
1 ,2-Dichlorobenzene
Hexachlorobenzene
Feed Cone
(ppb)
100,900
91,000
61,810
56,365
78,400
537,000
79,200
Product Cone
(ppb)
< 100.0
15.6
6.5
2.8
1.4
74.1
300.0
Removal
(%)
> 99.90
99.98
99.98
99.99
99.99
99.99
99.62
200
-------
The pilot system has recently completed a series of tests on ten PCB contaminated soils under a TSCA
R&O permit at CWM's Kettleman Hills Facility. The last test was completed on January 26, 1990.
Approximately 20 tons of material was tested.
The results of this testing are summarized in Table 5. Nine of the ten materials tested were reduced
to below 25 ppm PCB in the treated product and four were reduced to below 10 ppm. This confirms
at a relatively large scale that X*TRAX can separate PCBs from soil and produce a treated product with
a very low residual PCB concentration. As with the lab testing, other organic chemicals were present
at lower levels. These were generally reduced to detection limits (typically the ppb range). Oil and
grease (and TPH) were reduced to near or below the 30 ppm detection limit, even when the feed had
45,000 ppm O&G.
Table 5
PCB CONTAMINATED SOILS
PILOT X«TRAX™
Run #
0919
0810
1003
0727
0929
Matrix
Clay
Silty Clay
Clay
Sandy
Clay
Feed
(ppm)
5,000
2,800
1,600
1,480
630
Product
(ppm)
24
19
4.8
8.7
17
Removal
(%)
99.5
99.3
99.7
99.1
97.3
During the TSCA testing, the process vent was continuously monitored for total hydrocarbon
emissions. Samples were also taken and analyzed for PCBs. Table 6 summarizes these data. The
average release rate for hydrocarbons was very low and the PCBs were non-detectable.
Table 6
PILOT X"TRAX™
TSCA Testing • Vent Emissions
Run No.
0914
0919
0921
0926
0929
Total Hydrocarbons
(ppm-V)
Before
Carbon
1,320
1,031
530
2,950
2,100
After
Carbon
57
72
35
170
180
Removal
(%)
95.6
93.0
93.3
94.2
91.4
VOC
(Ib/day)
0.02
0.03
0.01
0.07
0.08
PCS'
(mg/m3)
< 0.00056
< 0.00055
< 0.00051
< 0.00058
< 0.00052
'Note: OSHA permits 0.50 mg/m3 PCB (1254) for 8-hr exposure
201
-------
It should be pointed out that the solids feed system has presented the most problems during operation
of the pilot X'TRAX. Soils with high sand concentrations presented few problems. Soils with high
clay content proved very difficult to convey at a constant rate, and sometimes at any rate. Over the
last two years four different feed systems have been tried. Each of the first three were modified
several times before progressing to the next design. The current feed system has proven itself capable
of metering and conveying anything from dry sand to damp clay that had to be picked out of an
inverted drum. This design has been incorporated into the Model 200.
5.
DISCUSSION OF RESULTS
In general, results to date have been quite good for the X*TRAX process. The basic process
equipment has been reliable, and performance has exceeded expectations, especially with regard to
the system's capacity. The area that deserves special attention is that of comparing the residual
organic levels in the product with various applicable treatment standards.
Most of the test data to date on regulated material has been on PCB contaminated soils. Discussions
with EPA have centered around meeting a treatment standard of 2 ppm. This is the analytical method's
quantitation limit for RGB's when un-modified procedures are used. As can be seen from the data,
X*TRAX has not achieved this standard. However, there is little or no basis for this numerical
standard present in the regulatory citations. Rather, it has been implemented as a specific permit
condition on a case by case basis. The PCB disposal regulations (40 CFR 761.60) simply state that
non-liquid PCB's with concentrations over 50 ppm and less than 500 ppm have to be disposed in an
approved landfill and that solids over 500 ppm have to be incinerated. The TSCA spill regulations (40
CFR 761.125) give cleanup requirements for soil as either 25 or 10 ppm, depending on the level of
restricted access to the site.
Recently, the EPA has issued Superfund LDR #6A4 as guidance to allowable treatment standards for
Superfund soil and debris with respect to the RCRA land disposal restrictions. This document gives
recommended numerical treatment standards to apply to Superfund soil and debris when the land
disposal restrictions cannot be met. For PCBs the treatment standard is given as 0.1 -10 ppm for soils
containing less than 100 ppm, and 90 - 99.9% removal for soils greater than 100 ppm. Similar
standards are given for other classes of organic compounds.
Table 7 is the same treatability data presented in Table 5, with the percent removal column replaced
with the maximum residual level allowable under LDR guide #6A. Clearly the X*TRAX process meets
the residual levels that could be allowed under these treatability guidelines. It would seem appropriate
to strike a compromise position, somewhere between the LDR Guide's allowable values and the 2 ppm
defacto standard.
Table 7
PCB CONTAMINATED SOILS
PILOT X*TRAX™
Comparison With LDR Guide #6A
Run #
0919
0810
1003
0727
0929
Matrix
Clay
Silty Clay
Clay
Sandy
Clay
Feed
(ppm)
5,000
2,800
1,600
1,480
630
Product
(ppm)
24
19
4.8
8.7
17
Required Lever
(max)
500
280
160
148
63
•Based on 90% removal
202
-------
An interesting observation about the residual PCB levels from the X'TRAX process is that the product
concentration is relatively insensitive to the feed concentration. The product values are generally in
the 5 - 25 ppm range, even when the feed is 5,000 ppm or higher. The residual levels are however,
strongly dependent on the solids temperature and to a lesser extent the soil matrix. This is a source
of comfort when working on a site where there are "hot spots." If the required treatment level is being
achieved with the average feed material, the treatment of a hot spot should not exceed the standard.
Two of the materials tested in the pilot system were also tested in the lab system. Table 8 is a
comparison of the results of these tests. Clearly, the residual levels in the product show good
agreement, indicating that the lab system closely simulates the pilot system.
Table 8
COMPARISON OF LAB AND PILOT X'TRAX™
PCB Contaminated Soils
Matrix
Sandy
Silty Clay
System
Lab
Pilot
Lab
Pilot
Run ID
RS0829
RS0727
GR0524
GR0810
Amount
(Ib)
19
4,958
31
4,584
Feed
(ppm)
5,100
1,480
962
2,800
Product
(ppm)
9.7
8.7
21
19
6. COSTS
As is typically the case with waste treatment processes, the operating economics of X'TRAX are
highly dependent on the waste type, specific contaminants and other project specific factors such
as disposal requirements and restrictions, and the amount of material to be processed. All of this
said, the treatment price for a material that meets the broad scope of acceptable feed for X*TRAX
will usually fall in the $150 - $250 per ton of feed. This treatment price is in the appropriate range
to make X'TRAX very competitive with both traditional disposal methods and also alternative
treatment processes.
As with any on-site method, *• mobilization charges to bring equipment on site need to be spread
over the volume of material re _ jiring treatment. To keep these mobilization charges within reason,
a lower limit of about 5,000 cubic yards is appropriate for the X'TRAX Model 200.
7.
SUMMARY AND CONCLUSIONS
CWM has progressed to a point in the development of the X'TRAX thermal separation process at
which the following statements can be made:
• the X'TRAX thermal separation process has been demonstrated as effective on actual
waste materials at both the lab and pilot scale
• pilot data has shown the process to have very low atmospheric emissions, both for PCBs
and total organics
203
-------
very low residual contamination levels have been achieved for the treated product, typically
in the 5-25 ppm range for PCBs and at the ppb level for most other organics, however, a
2 ppm treatment standard for PCBs cannot presently be met
the process is tolerant of wide variation of organic levels in the feed
the commercial system is easily modified to add stabilization for metal containing wastes
the process is economically viable, with treatment price in the $150 - $250 per ton; it is
suitable for remediations >5,000 yd3
References
1. Swanstrom, C., Palmer, C., X*TRA)(rM Transportable Thermal Separator for Solids
Contaminated with Organics, Air & Waste Management Association International Symposium
on Hazardous Waste Treatment: Treatment of Contaminated Soils, Cincinnati, Ohio, February
5-8, 1990.
2. Daley, P.S., Cleaning up Sites with On-Site Process Plants, Environmental Science &
Technology, August, 1989.
3. Palmer, C.R., Hollenbeck, P.E., Sludge Detoxification Demonstration, Incineration Conference.
May 1-5, 1989.
4. USEPA - Office of Solid Waste and Emergency Response, SuperfundLDR Guide #6A: Obtaining
a Soil and Debris Treatability Variance for Remedial Actions, Directive 9347.3-06FS, July
1989.
204
-------
ANAEROBIC PYROLYSIS FOR TREATMENT OF ORGANIC
CONTAMINANTS IN SOLIDS WASTES AND SLUDGES -
THE AOSTRA TACIUK PROCESS SYSTEM
THERMAL TREATMENT
AUTHOR:
Robert M. Ritcey, P.Eng.,
Manager, Demonstration Operations
UMATAC Industrial Proceses
A Division of UMA Engineering Ltd.
Calgary, Alberta
PRESENTATION TO:
Second Forum on Innovative Treatment Technologies:
Domestic and International
Philadelphia, PA
May 15-17, 1990
205
-------
ANAEROBIC PYROLYSIS FOR TREATMENT OF ORGANIC
CONTAMINANTS IN SOLID WASTES AND SLUDGES -
THE AOSTRA TACIUK PROCESS SYSTEM
ABSTRACT
The AOSTRA Taciuk Process (ATP) is a thermal treatment technology which has been
developed and proven effective for separating organic contaminants such as oils and
petrochemicals from solid wastes, soils and sludges. The Process was developed jointly by the
Alberta Oil Sands Technology and Research Authority and UMATAC Industrial Processes for
producing bitumen from Alberta's oil sands deposits as distillate oil product, and for producing
oil from oil shales. The extensive test work in these fields, and recent work directly with waste
materials are the background to the recent commercial application of the ATP System in waste
treatment.
The ATP technology is a continuous flow pyrolysis system which achieves the separation of
organic contaminants from host solids in the wastes, and offers the advantages of minimum air
emissions, ability to recycle the organics, rapid treatment of waste volumes, and low cost. The
ATP System has been extensively tested and its capability demonstrated for separating water
and organic contaminants from soils and sludges.
The plant is a recycling facility for those constituents in the waste which are re-usable. It
renders the solids free of the organic contaminant(s), the liquid products (water and oil) for re-
use, secondary treatment or disposal, and the air emissions meet regulatory criteria applicable
to thermal remediation treatment of wastes.
The first commercial ATP treatment plant has a design capacity of 10 tph and was ready for
service in September, 1989. It will be employed in waste treatment, initially on a PCB
separation soil treatment Superfund project in the U.S.A. The plant is portable and will be
used on numerous job sites in North America. Other remediation projects and applications for
the ATP technology and portable treatment plants of various sizes are being developed.
This paper describes the ATP System, its scope of applicability in waste treatment, and
considerations for specific applications of portable ATP treatment plants. Performance results
from demonstration treatment programs and treatment cost data are discussed.
2nd FORUM Mty 15-17,1990
206
-------
ANAEROBIC PYROLYSIS FOR TREATMENT OF ORGANIC
CONTAMINANTS IN SOLID WASTES AND SLUDGES -
THE AOSTRA TACIUK PROCESS SYSTEM
The AOSTRA Taciuk Process (ATP) is a continuous flow pyrolysis treatment system which is
applicable to the remediation of various waste soils and sludges containing organic
contaminants, such as oils or other hydrocarbon materials and products. This capability
evolved from the extensive development of the ATP technology as a process for producing oil
from oil sands and oil shales which has been conducted since 1976 by UMATAC Industrial
Processes division of UMA Engineering Ltd, under agreement with the Alberta Oil Sands
Technology and Research Authority (AOSTRA).
Over the past five years, the development work has included physical and production testing of
small and large quantities of oily soils from many locations in North America. The work
included tests on contaminated soils and sludges from the oil and petrochemical industries,
wastes from municipal and urban sites such as old industrial sites or dumps, and from
orphaned or Superfund sites. Contaminants found in these sites include organic and inorganic
materials, and the ATP technology was tested and demonstrated for separating and recovering
the organics, such as oils, PCBs, coal tars and acid chemicals.
UMATAC offers the ATP technology both in the form of technical expertise and treatment
plant facilities which are currently available for commercial application. The treatment plants
can be designed and provided in a wide range of sizes and configurations to meet the needs of
various situations and clients. The technology is sub-licensed for commercial use in the United
States to SoilTech, Inc., which is an operating company jointly held by UMATAC and
Canonic Environmental Services Corp.
In this remediation work on sites and materials, the need is usually for a treatment facility
which is portable and easily applied to the job, maintains performance capability when treating
wastes of differing characteristics, and can achieve the treatment at reasonable cost. The ATP
System offers many features to meet these requirements, such as:
plants in sizes up to 25 tph of feed capacity, or greater as needed. Larger plants are of
modularized construction to facilitate moving. Most inquiries are for small, portable
plants of 5 tph capacity, or less.
continuous, stable, safe and efficient operation with widely varying feed materials.
effective separation of organic contaminants from host soils or other inorganic
materials, rendering the soils free of the contaminants (oil, PCB, PAH, etc.).
recycling of the contaminants which are re-usable, such as oils and petrochemicals.
ability to operate within, and meet permitting criteria.
The ATP System
The AOSTRA Taciuk Processor and the associated processes of the ATP System are shown in
Figures 1 and 2. A technical description of the System is briefly summarized; more complete
technical descriptions are available in the references.
The Processor (Figure 1) is a horizontal rotating vessel which can receive feed wastes in
physical form varying from solids to liquids, and which contain varying concentrations of .pa
2nd FORUM Miy 15-17, 1990
207
-------
RUE MS
STEAM
AUOMff
CLEAN DRY*
TA1UNGS SAND
__ _ — _ MUJ^^^pHHl
COOUNG ZONE | COMBUSTION ZONE |**NER
ATP PROCESSOR
INTERNAL PROCESS FLOWS
FIGURE 1
{COMBUSTION
AJR
STACK
L
WATER
FLUE GAS
TREATMENT
L
Oil
t
FUEL
VAPOUR
TREATMENT
OFF CAS
ATP
PROCESSOR
OIL FREE
TAILINGS
SOURl WATER
WATER
1
1 * I
p VAPOUR "
OIL
RECOVERY
PRODUCT
(XL l
AOSTRA TACIUK PROCESS (ATP)
FIGURE 2
2nd FORUM M«y 15-17, 1990
208
-------
water, solids and organic contaminants. The usual requirement for treating the waste is to
separate the contaminants from the inorganic solids and render the solids non-toxic so they can
be re-used or disposed directly to landfill.
The Processor effects separation of the solids, water and organics (liquids and solids) by
heating in the first zone (preheat) where the feed is received, and pyrolysis in the reaction
zone. Pyrolysis temperature is usually between 500 and 600°C, and the operating pressures of
the zones are near atmospheric (-2 to -3 mm).
Heat supply for the Process is generated in the same vessel in a combustion zone which is
situated around the pyrolysis reaction zone as an annular section. The Processor makes use of
non-vaporizable organic material (coke) which passes with the solids to the combustion zone
from the reaction zone, and non-condensible hydrocarbon vapor products of the pyrolysis for
process heat fuels. These fuels and the solids are de-contaminated in the reaction zone before
they enter the combustion zone. Auxiliary industrial fuels such as gas or oil, must be available
for stan up, control and supply to augment any inadequacies in energy for process heat from
the internally produced fuels.
An important and unique feature of the Processor is the recovery and use of heat from the
solids and combustion gases in the cooling zone, before these products are emitted from the
vessel. This heat is usually sufficient for the requirements of the preheat zone; i.e., to heat the
incoming feed to about 200°C to vaporize the water in the feed and prepare the oily solids for
the higher temperature retort zone.
The auxiliary process systems are shown in Figure 2 in block flow form, integrated with the
Processor. These systems handle the materials (feed and products), condense the hydrocarbon
vapors and steam to liquid oil and water products, and treat the combustion gases for an
acceptable stack emission.
The products of the ATP System are oil-free solids which are usually suitable for direct
disposal to landfill, and water and oils. The water usually requires secondary treatment such as
steam stripping and biological treatment to remove the organics and reduce the oxygen
demand. Oil from oil industry-type wastes is usually suitable for addition to refinery
feedstocks; hazardous type oil products can be separately disposed as a relatively small
quantity for incineration by an ATP plant combustor, or by an off site facility. In most cases, it
is important to also establish the levels of non-organic contaminants in the feed, such as heavy
metals or other hazardous materials, in order to provide the necessary, secondary treatment for
these in addition to the organics.
The ATP System thus consists of distinct process functions - feed supply, the Processor for
separation of phases, vapor handling for liquids recovery, solids product handling, flue gas
treatment and plant control. In the range of process capacities normally required for waste
treatment plants, the System equipment readily lends itself to the modular design needed for
highly mobile, portable plants.
Use in Treating Contaminated Soils and Sludges
The Process has been tested on a wide range of wastes in addition to the oil sands and
shales4*5. The testing during the development and proving phase for the process and mechanical
equipment was conducted on approximately 14,000 tonnes of oil sands and shales. Test work
on wastes began in 1986 and has included about 500 tonnes of oily soils and sludges from the
production, refining and marketing sectors of the oil industry, waste rubber and plastics
2nd FORUM M«y 15-17, 1990
209
-------
products, and PCB contaminated soils. In 1989, about 1000 tonnes of oily materials were
treated with the 10 tph commercial plant in the course of plant testing and demonstration.
In addition to these tests on bulk samples using the ATP plants, about 500 small scale batch
tests have been run on the various wastes. Batch tests simulate the vaporization and pyrolysis
reactions of the ATP System and are used for initial evaluation of the ATP for treating a
waste.
The Processor can treat any mixture of solids and liquids, to an upper limit on solids size of
six to eight inches, so large objects must be separated from the feed by sorting or screening.
Depending on the treatment situation, sludges can be fed directly as liquid, or pre-mixed with
solids (waste or carrier sand) to enable handling by conventional loader-hopper-conveyor
equipment.
The work and results from four (4) different test programs on oil industry wastes illustrate the
capability and consistency of the ATP System to treat these wastes. Data of analysis
determinations for some of the key contaminants in these wastes are shown in Tables 1 through
5. The program to test PCB contaminated soil is described as the fifth reference.
The wastes treated in these programs were:
1) Refinery waste stream sludges - EPA classifications K051, K048, and K049. These are
sludges from API Separators, DAF and slop oil emulsion, respectively.
2) Oily soil from a decommissioned refinery, labelled "A".
3) Oily soil from a decpmmissioned refinery, labelled "B".
4) Drilling cuttings which contain diesel oil, from use of invert type drilling mud.
5) PCB contaminated soils.
1) Refinery waste stream sludges
UMATAC participated in the 1987 program of the American Petroleum Institute (API) Task
Force which studied technologies applicable and ready for use in treating the "K" waste
streams of refinery operations in the United States. The API wished to determine treatment
options prior to the imposition by the EPA of the ban on disposal and treatment of these wastes
in land fills.
The products of the ATP tests, which are oil, water and combusted solids, were analysed by
commercial laboratories for characterization of organic and inorganic constituents. This
procedure is also followed with most of the feeds and products of tests of this type on other
wastes to assess the environmental characteristics of the materials. The results showed that the
treated solids were oil free, and that most of the metals and salts in the feed remain with the
solids.
The API published a report of its study of the technologies, and reported that pyrolysis (the
AOSTRA Taciuk Process) is the "most efficient in removing" organic contaminants from the
solids and in achieving low leachability of organic compounds from the product solids. The
reference is API Report #4465, published in May, 19887.
2,3) Refinery wastes - from Sites "A" and "B"
A 35 tonne sample of material from the clean-up of a decommissioned refinery ("A") was
received for treatability tests and plant test operation in 1986. The sample was a blend of a
number of oily sludges and soils at the site. It was processed in the 4.5 tph pilot plant. A
2nd FORUM May 15-17, 1990
210
-------
second major sample from another site ("B") was received in 1989. It was about 150 tonnes,
and was processed in the commercial, 10 tph plant.
The treated solids from these tests contained little or no oil and grease. Leachate data on solids
products showed oil and grease levels of 0.7 mg/1 or less, with metals concentrations being low
or less than detectable. Additional analyses showed monocyclic and polycyclic aromatic
hydrocarbons at or below detection limits in the leachate from the solids residue.
The flue gas paniculate discharge was less than required levels, with metals emissions being as
much as 1000 times less than the criteria. Typical emissions data are presented in Table 5. In
the "B" tests, flue gases were sampled using seven sampling trains. Each train complied with
the approved methods of appropriate regulatory agencies, which were the sampling and
analysis protocols of EPA MM5, or equal.
The results for particulates, HC1, SO,, and oxides of nitrogen are well within criteria. There
is a general lack of information in industry on emissions criteria for PAH, MAH, and phenols
in the stack gases, but total concentrations of these compounds were 1.5, 130 and 70 ug/m3.
The heavy metals concentrations are generally much less than the emissions criteria. Carbon
monoxide reported a maximum of 190 ppm, to a low of 130 ppm, which meets criteria in some
jurisdictions.
Generally all the water products were contaminated at constituent levels above the emission
criteria, and would require further treatment for reuse or disposal. Possible treatments would
include steam stripping, air flotation and activated carbon beds. The sour water, being a
product of the hydrocarbons cracking, contained high quantities of phenols, and detectable oil
and grease. The preheat waters are not exposed to pyrolysis conditions and are not expected to
contain phenols.
4) Invert drilling mud cuttings
Invert drilling muds are comprised of large amounts of diesel oil. Most of the mud, with oil,
is recovered from the cuttings and recycled in the drilling operation. The separated cuttings
usually still contain a significant amount of oil, averaging 15% by weight. Nearly 120 tonnes
of these cuttings were obtained for testing at the AOSTRA-UMATAC plant in Calgary.
The cuttings, as received, were quite sloppy and had to be contained by a berm. A full range
of batch, pilot and commercial unit treatment operations in 1988 and 1989 proved oil removal
capabilities, with more than 60 % of the oil being recovered as butane and heavier liquids. The
treated solids were oil free and suitable for reuse or disposal as landfill.
The flue gas paniculate emission was less than half the allowable criteria, as presented in
Table 2. The solids effluent leachate data showed low levels of chlorides, sulphates and metals
as presented in Table 3.
5) PCB Contaminated Soil
Soils contaminated with PCB Arochlor 1242 were treated in bulk sample tests using the 5 tph
Processor unit in 1988. The PCB concentrations ranged from 200 to 20,000 ppm, and 27 tons
of soil were treated in 2 test runs. The results showed the soil to be cleaned; the PCBs are
vaporized in the reaction zone and removed as a vapor. No PCBs are subjected to combustion,
and furans and dioxins were not produced. The DRE of the PCB from the feed met the six
nines requirement.
2nd FORUM May 15-17, 1990
211
-------
This test work demonstrated the ATP System for acceptance of use at the \\fcukegan Harbour
Superfund Site remediation. Soil treatment operations at this site are scheduled for 1991.
Costs and Sensitivities
The capital supply cost of ATP plants ranges from $ 3 million to $ 8 million for plant sizes of
3 to 20 tph, respectively. Table 6 shows approximate costs and operating data for these plant
sizes in applications on non-hazardous waste such as soils and product streams from refineries.
Higher costs are expected when treating more complex wastes which may require secondary
treatment or costly disposal of separated constituents. Specific data must be prepared for each
usage situation. The costs are reasonably indicative for a commercial facility, but must be
defined much more accurately by specific examination for each requirement.
If the waste quantities per plant set up are less, the unit costs increase because most costs are
fixed, such as a permanent operating staff and capital recovery costs, and the production is
reduced while the plant is moved or waits for feed. The annual operating costs for all plants
larger than 5 tph to 20 tph are much the same for labor because the cost of the operating staff
is a large portion of the total. The cost for most operating supplies, especially consumables,
varies with the throughput.
Besides the cost of plant operations, the 2 main factors which influence the total unit cost of
treatment are plant supply cost and utilization. The plant usage is a function of the relocation
frequency and the quantities of waste for treatment. The costs shown are for the plant supply
and basic operation, and are only generally indicative of the economics of the ATP System.
They do not include other costs, such as site development, permitting, waste supply and
preparation, products disposal, local requirements affecting labor costs or special monitoring,
analysis and reporting of products and emissions. These items are all peculiar to each waste
situation.
Other factors which can affect the costs of treatment by the ATP System are:
contents of moisture and oil in the waste.
particle size distribution of the solids
acid content of the wastes
vaporization of the organic materials
handling or disposing the condensed hydrocarbon products
mobility of metals in the solids products.
In summary, the total costs of pyrolysis treatment of oily wastes with the ATP System are
generally in the range of $100 to $200 per tonne of feed, but can range higher or lower
depending on the situation. Costs of direct plant operation only range from $40 to $150 per
tonne. The objective is to correctly size the plant for continuous operations, which will enable
minimum cost of treatment. Costs of thermal treatment may not be as low as other methods
such as land treatment, but the remediation considerations must include environmental
acceptability of the method and products, and the time requirements.
Alternatively, the costs and economics can be considered in a completely different manner in
the case of a large production plant (oil, petrochemical, etc,) into which an ATP plant is
incorporated to treat waste products or accumulations, and which is operated as part of the
large plant. For this situation, the mobilization would occur only once, and the plant supply
cost could be amortized over a longer period, say 15 years, and the plant staff cost could be
2nd FORUM May 15-17, 1990
212
-------
lower as the crew is part of a larger work force.
ATP SYSTEM ADVANTAGES
The ATP System pyrplysis technology offers a number of unique features for consideration of
its use in treating various oily, organic wastes. They include the following:
the Process has the intermediate step of extracting the contaminants from the solids,
which presents them as available for re-use or refining, rather than destruction.
treatment plants can be supplied for large throughput capacity, to 20 tph or greater.
low unit costs of plant operation, particularly for the larger capacity plants.
all of the treatment process steps for extraction of the contaminants are confined to a
single process vessel.
the Process operates at low to intermediate temperatures (to 600°C), which enhances
safety and indicates low energy requirement.
water and oil contaminants are recovered as separate products, which makes them
available for secondary treatment, re-use or separate disposal.
solids products are hydrocarbon-free, and can usually meet leachate criteria for direct
disposal.
heat make up uses Process by-products as fuel.
combustion conditions include low gas velocity, long residence time for the fuels, and
extensive turbulence and mixing in the combustion zone. This assists combustion
completion and control of the gas emissions
the System has been extensively tested and shows capability to operate with stability
and safety on widely varying wastes.
Much interest has been expressed by industry and regulatory representatives in the AOSTRA
Taciuk Process System, particularly since the first commercial waste treatment plant became
available and was purchased by SoilTech, Inc. The ATP pyrolysis technology can address a
broad field of applications for treating hazardous and non-hazardous industrial wastes. It offers
fast clean up capability combined with the recovery and reuse of the contaminants, while
operating within the stringent emissions standards required in today's modern world.
References
1. R.M. Ritcey, "Anaerobic Pyrolysis of Solid Wastes and Sludges - The AOSTRA
Taciuk Process System", HAZTECH Canada Conference, Edmonton, Alberta, October
1989.
2. L.R. Turner, "Treatment of Oilsands and Heavy Oil Production Wastes Using the
AOSTRA Taciuk Process", Conference on Oil Field Production Wastes. May 1989,
Calgary, Alberta.
3. American Petroleum Institute, Health and Environmental Sciences Department,
Washington, D.C., Publication No. 4465. "Evaluation of Treatment Technologies for
Listed Petroleum Refinery Wastes", May 1988.
4. W. Taciuk, R.M. Ritcey, "Taciuk Processor for Treatment of Contaminated Wastes",
AOSTRA Conference, Advances in Petroleum Recovery and Upgrading Technology
1987. Edmonton, Alberta.
2nd FORUM May 15-17, 1990
213
-------
5. M. Vorum, A.H. Montgomery, "The Taciuk Process Technology For Anaerobic
Pyrolysis Of Solid Wastes And Sludges", 1989, Canonie Environmental Services
Corporation. Englewood, Colorado.
6. SoilTech, Inc., "The Taciuk Process Technology: Thermal Remediation of Solid
Wastes and Sludges". Technical Information Manual.
7. American Petroleum Institute, "Evaluation of Treatment Technologies for Listed
Petroleum Refinery Wastes". API Publication No. 4465, Health and Environmental
Sciences Department, May 1988. API, 1220 L Street, Washington, D.C. 20005.
2nd FORUM May 15-17, 1990
214
-------
TESTING UNIT
YEAR TESTED
TABLE 1 - FEED CHARACTERISTICS
TEST RESULTS DATA
AOSTRA TACIUK PROCESSOR TESTS ON OIL CONTAMINATED SOILS
Refinery Refinery Drill
API Hastes Hastes Mud CRITERIA
PROGRAM A" B" Cuttings ENVIRON-
- PLANT BATCH Pilot 10 tph Pilot MENT
1987 1986 1989 1989 CANADA
PROCESSOR FEED:
Oil /Grease Ht %
Hater Ht %
Solids Ht %
HASTE ONLY CONSTITUENTS:
Oil/Grease Ht %
Hater Ht %
Solids Ht %
2.1
17.7
80.2
20 1.8
52 30.0
28 68.2
0.3
10.9
88.8
0.5
14.0
85.5
15.3
7.3
77.4
15.3
7.3
77.4
TABLE 2 - FLUE GAS CHARACTERISTICS
PARTICULATES
METALS
mg/m3
mg/m3
Arsenic
Barium
Cadmium
Chromium
Cobalt
Copper
Lead
Mercury
Nickel
Selenium
Vanadium
Zinc
979* 17.655
18.9
0.040
30
-0.0001
0.105
0.025
0.425
0.430
0.021
0.150
0.004
0.450
0.045
-0.0031
1.3104
0.0035
0.0071
0.0192
0.0017
0.0070
0.0008
0.9638
-0.0025
0.0037
0.7603
50
.2
1
1
5
5
.2
1
1
5
5
* - NO BAGHOUSE IN THE FLUE GAS TREATMENT; HET SCRUBBER ONLY.
2nd FORUM M»y 15-17, 1990
215
-------
TESTING UNIT
YEAR TESTED
TABLE 2 - FLUE GAS CHARACTERISTICS (cont'd)
TEST RESULTS DATA
AOSTRA TACIUK PROCESSOR TESTS ON OIL CONTAMINATED SOILS
Refinery Refinery Drill
API Wastes Wastes Mud CRITERIA
PROGRAM A" B" Cuttings ENVIRON-
- PLANT BATCH Pilot 10 tph Pilot MENT
1987 1986 1989 1989 CANADA
HCL ug/m3
MAH's ug/m3
Benzene
Ethyl Benzene
Toluene
Styrene
Xylene
PAH's ug/m3
acenaphthene
acenaphthylene
anthracene
benzo ( a ) anthracene
benzo ( a ) pyrene
benzo (b) f luoranthene
benzo (k) f luoranthene
benzo(g,h, i)perylene
chrysene
dibenzo ( a , h ) anthracene
f luorene
fluoranthene
indeno ( 1 , 2 , 3 , c , d ) pyrene
naphthalene
phenanthrene
pyrene
3516
70
8067
2404
7.720
1.520
0.745
0.245
-0.120
0.065
-0.120
0.000
0.125
0.000
0.930
2.595
0.000
6.105
4.570
3.275
6 75000
69.0
2.6
13.1
0.6
18.1
-0.091
0.090
0.110
-0.091
-0.091
-0.091
-0.091
-0.091
-0.091
-0.091
-0.091
0.160
-0.091
0.596
0.537
0.110
NOTE: (-nn) indicates 'value less than detection limit of nn'
{-) indicates 'NOT DETECTED'
{ ) indicates 'not tested for1
(NA) indicates 'Not Applicable'
2nd FORUM Miy 15-17, 1990
216
-------
TABLE 3 - LEACHATE CHARACTERISTICS OF TREATED SOLIDS
TEST RESULTS DATA
AOSTRA TACIUK PROCESSOR TESTS ON OIL CONTAMINATED SOILS
TESTING UNIT
YEAR TESTED
- PLANT
Refinery Refinery Drill
API Wastes Wastes Mud CRITERIA
PROGRAM A" B" Cuttings ENVIRON-
BATCH Pilot 10 tph Pilot MENT
1987 1986 1989 1989 CANADA
Oil/Grease
Specific Cond,
MAJOR CATIONS
METALS
mg/1
umho/cm
mg/1
Chloride
Sulphate
mg/1
Arsenic
Barium
Boron
Cadmium
Chromium
Lead
Mercury
Selenium
Sodium
Vanadium
Zinc
1.0
-0.1
-0.1
NONE
520
2720
0.026
0.14
0.59
-0.001
-0.001
0.02
0.0003
0.011
0.53
0.064
0.7
40.1
824
180
165
109
0.03
0.12
-0.0008
5
100
500
0.5
5
5
0.1
1
3.5
0.07
2nd FORUM May 15-17, 1990
217
-------
TABLE 3 - LEACHATE CHARACTERISTICS OF TREATED SOLIDS (cont'd)
TEST RESULTS DATA
AOSTRA TACIUK PROCESSOR TESTS ON OIL CONTAMINATED SOILS
TESTING UNIT - PLANT
YEAR TESTED
ORGAN I CS:
MAH's mg/1
Benzene
Ethyl Benzene
Toluene
Styrene
Xylene
PAH's ug/1
acenaphthene
acenaphthylene
anthracene
benzo ( a ) anthracene
benzo ( a ) pyrene
benzo ( b ) f luoranthene
benzo(k) f luoranthene
benzo (g/h, i)perylene
chrysene
dibenzo ( a , h ) anthracene
f luorene
f luoranthene
indeno ( 1 , 2 , 3 , c , d ) pyrene
naphthalene
phenanthrene
pyrene
NOTE: (~nn) indicates 'value
Refinery
API Wastes
PROGRAM A"
BATCH Pilot
1987 1986
XXX XXX
{-) -0.2
-0.2
(-) -0.2
(-) -0.2
XX
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
less than detection
Refinery Drill
Wastes Mud CRITERIA
B" Cuttings ENVIRON-
10 tph Pilot MENT
1989 1989 CANADA
X
-0.002
-0.002
0.004
-0.002
0.006
X
-1.0
-1.0
-1.0
-1.0
-1.0
-1.0
-1.0
-2.0
-1.0
-1.0
-1.0
-1.0
-2.0
-1.0
-1.0
-1.0
limit of nn'
(-) indicates 'NOT DETECTED'
( ) indicates 'not tested for1
x - ANALYSED BY GC/MS
xx - ANALYSED BY HPLC (units are ppm)
xxx - ANALYSED BY GC/FID
2nd FORUM May 15-17, 1990
218
-------
TABLE 4 - CHARACTERISTICS OF TREATED SOLIDS
TEST RESULTS DATA
AOSTRA TACIUK PROCESSOR TESTS ON OIL CONTAMINATED SOILS
TESTING UNIT
YEAR TESTED
- PLANT
Refinery Refinery
API Hastes Wastes
PROGRAM A" B"
BATCH Pilot 10 tph
1987 1986 1989
Drill
Mud CRITERIA
Cuttings ENVIRON-
Pilot MENT
1989 CANADA
Oil/Grease nig/I
Specific Cond. umho/cm
MAJOR CATIONS
METALS
mg/1
Chloride
Sulphate
mg/1
Aluminum
Arsenic
Barium
Boron
Cadmium
Chromium
Copper
Lead
Nickel
Selenium
Vanadium
Zinc
NONE
500
2800
50
88
422
17
4230
3290
2180
13500
5.1
86.5
33.4
-0.1
26.3
195
0.65
0.3
26.00
31.00
38
39.00
122
142
70.00
2nd FORUM M«y 15-17, 1990
219
-------
TABLE 4 - CHARACTERISTICS OF TREATED SOLIDS - ORGANICS
TEST RESULTS DATA
AOSTRA TACXUK PROCESSOR TESTS ON OIL CONTAMINATED SOILS
TESTING UNIT
YEAR TESTED
- PLANT
Refinery Refinery
API Wastes Wastes
PROGRAM A" B"
BATCH Pilot 10 tph
1987 1986 1989
Drill
Mud CRITERIA
Cuttings ENVIRON-
Pilot MENT
1989 CANADA
ORGANICS:
MAH's
mg/1
Benzene
Ethyl Benzene
Toluene
Styrene
Xylene
PAH's
ug/1
acenaphthene
acenaphthylene
anthracene
benzo(a)anthracene
benzo(a)pyrene
benzo(b)fluoranthene
benzo(k)fluoranthene
benzo(g,h,i)perylene
chrysene
dibenzo(a,h)anthracene
fluorene
fluoranthene
indeno(1,2,3,c,d)pyrene
naphthalene
phenanthrene
pyrene
XXX
-0.1
3
0.8
XXX
XX
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
-0.01
X
-0.1
-0.1
-0.1
-0.1
X
-0.05
-0.05
-0.05
-0.05
-0.05
-0.05
-0.05
-0.1
-0.05
-0.05
-0.05
-0.1
-0.05
-0.05
-0.05
NOTE: (-nn) indicates 'value less than detection limit of nn1
(-) indicates 'NOT DETECTED'
( ) indicates 'not tested for'
x - ANALYSED BY GC/MS
xx - ANALYSED BY HPLC (units are ppm)
xxx - ANALYSED BY GC/FID
2nd FORUM May 15-17, 1990
220
-------
TABLE 5
FLUE GAS EMISSIONS
AVERAGES FOR THREE TESTS - MB" SAMPLE
Particulates
Hydrogen Chloride
Total Specified Metals
Sulphur Dioxide
Oxides of Nitrogen
Total Reduced Sulphur Compounds
Emissions
Criteria
mg/m *
20a
75a
200a
300a
Concentration
mg/m *
17.655
0.006
3.360
1.400
108.6
<3.0
Concentration
ug/m *
Emission
mg/h
0.090
<0.001
0.017
0.007
0.6
<0.016
Emission
mg/h
Total Poly-Aromatic Hydrocarbons
Total Non-Chlorinated Phenols
Total Mono-Aromatic Hydrocarbons
1.330 6.812
69.246 344.853
127.6 645.9
At 25°C, 760 mm Hg, dry basis, O2 cone. 11.8 % (avg)
Environment Canada Criteria for Haz. Waste Incinerators and Thermal Treatment
Facilities, O2 cone. 11 %, 25°C, 760 mmHg.
Mole Percentages of Total Emission:
Carbon Monoxide
Carbon Dioxide
Hydrogen Sulphide
0.016
7.82
<0.0001
2nd FORUM May 15-17, 1990
221
-------
TABLE 6
BASIC COSTS OF ATP PYROLYSIS TREATMENT (1)
ITEM
PLANT SIZE AND OPERATING ASSUMPTION
ASSUMPTIONS:
PLANT SIZE - TPH
ANNUAL THROUGHPUT (T)
UTILIZATION - TIME %
PLANT CAPITAL ($ x 1000)
3
21,000
80
3,000
10
70,000
80
5,000
20
140,000
80
8,000
UNIT COSTS - $/T:
PLANT OPERATION
125 to 175 75 to 125 40 to 90
(1) Costs are for the supply and operation of the ATP only. See text for exclusions.
222
2nd FORUM May 15-17, 1990
-------
ALGASORB®: A NEW TECHNOLOGY FOR REMOVAL AND RECOVERY OF
METAL IONS FROM GROUNDWATERS
By: Dennis W. Darnall, Sandy Svec and Maria Alvarez
Bio-Recovery Systems, Inc.
P. O. Box 3982, UPB
Las Cruces, NM 88003
ABSTRACT
A new sorption process for removing toxic metal ions from water has been
developed. This process is based upon the natural, very strong affinity of
biological materials, such as the cell walls of plants and microorganisms, for
heavy metal ions. Biological materials, primarily algae, have been immobilized in
a polymer to produce a "biological" ion exchange resin, called AlgaSORB®. The
material has a remarkable affinity for heavy metal ions and is capable of
concentrating these ions by a factor of many thousand-fold. Additionally, the
bound metals can be stripped and recovered from the algal material in a manner
similar to conventional resins.
This new technology has been demonstrated to be an effective method for
removing toxic metals from groundwaters. Metal concentrations can be reduced to
low pans per billion (ppb) levels. An important characteristic of the binding
material is that high concentrations of common ions such as calcium, magnesium,
sodium, potassium, chloride and sulfate do not interfere with the binding of heavy
metals. Waters containing a total dissolved solids (TDS) content of several thousand
and a hardness of several hundred parts per million (ppm) can be successfully
treated to remove and recover heavy metals. The process has been demonstrated
under the SITE Emerging Technology program for the effective removal of
mercury from a contaminated groundwater.
This paper has been reviewed in accordance with
the U. S. Environmental Protection Agency's peer
and administrative review policies and approved
for presentation and publication.
223
-------
INTRODUCTION
In recent years there has been increased attention focused on pollution of
water supplies by heavy metal ions. These metals are toxic in rather low
concentrations and can lead to acute and chronic illness in humans and other
animals. Past waste disposal practices have resulted in serious contamination of
the environment. The major sources of heavy metals are leachates from legal and
illegal landfills and drainage from old mines. Several of the sites on the National
Priorities List are landfill facilities that have problems caused by improper
placement of heavy metal wastes below ground.
Chemical treatment using lime or caustic precipitation has been the most
common method of removing dissolved metals from leachates, mine drainages, and
contaminated aquifers. Other treatments include reverse osmosis, electrodialysis,
and carbon adsorption. The energy and chemical costs of these various methods
become major operating expenses during the use of these treatments. A serious
limitation of many of the current treatment technologies is the difficulty and/or
expense of treating waters for removal of heavy metals to allowable drinking
water levels. Alternatives for economical recovery of dissolved metals from
contaminated waters are limited. The development of rapid, widely applicable, low
cost methods for the removal, recovery and recycling of heavy metal ions from
contaminated waters at Superfund sites is a high priority.
For a number of years Bio-Recovery Systems has been working with a new
sorption process for removing heavy metal ions from water (1-4). This sorption
process is based upon the natural, very strong affinity of the cell walls of algae for
heavy metal ions. Algal cells have been immobilized in a silica gel polymer
(AlgaSORB®) and used much as an ion-exchange resin. The algae are killed in the
immobilization process indicating that sorption does not require a living
organism, and hence the algal matrix can be exposed, with little or no ill effects, to
solution conditions which would normally kill living cells. The pores of the
polymer are apparently large enough to allow free diffusion of ions to the algal
cells, since similar concentrations of metal ions are bound by free and immobilized
cells.
AlgaSORB® functions as a "biological" ion exchange resin and binds both
metallic cations and metallic oxoanions. Anions such as chloride or sulfate are
only weakly bound or not bound at all. Like ion-exchange resins, the algae-silica
system can be recycled. Metal ions have been sorbed and stripped over as many as
75 cycles with no noticeable loss in efficiency. In contrast to current ion
exchange technology, however, another real advantage of the algae-silica matrix
is that the components of hard water (Ca+2, Mg+2) or monovalent cations (Na*, K+)
do not significantly interfere with the binding of toxic, heavy metal ions (2, 3).
The binding of Ca+2 and Mg42 to ion-exchange resins often limits its usefulness
224
-------
since these ions are frequently present in high concentrations and compete for
heavy metal ion binding. Frequent regeneration of ion-exchange resins is
necessary in order to remove heavy metal ions from solution. Another advantage
of AlgaSORB® is that it has a high affinity for metal ions. Effluents from AlgaSORB®
columns frequently show that metal ions can be reduced to the part per billion
level. AlgaSORB* can be used to remove aluminum, cadmium, chromium, cobalt,
copper, gold, iron, lead, maganese, mercury, molybdenum, nickel, platinum, silver,
uranium, vanadium, and zinc. Many of these metals are hazardous wastes at
Superfund sites.
A major advantage of AlgaSORB® is that the efficiency of heavy metal ion
removal is not diminished by the presence of organic compounds in the water. For
example, waters containing copper ion and high concentrations of organics such
as butyl cellosolve, alcohols, other etchers and halogenated hydrocarbons have
been successfully treated (4). Humic and fulvic acids seem to have little or no
effect on metal ion binding to AlgaSORB®. The presence of organic compounds
limits the usefulness of ion exchange resins since organics often will bind to these
resins and decrease their metal binding capacity.
In 1988, Bio-Recovery Systems was selected to participate in the Superfund
Innovative Technology Evaluation (SITE) Emerging Technologies Program. The
goal of the project was to test AlgaSORB® for the removal of mercury from a
contaminated groundwater.
DESCRIPTION OF GROUNDWATERS
A number of years ago an industrial process using mercury resulted in soil
contamination with elemental mercury. The mercury subsequently percolated
through the soils and contaminated groundwater. At some point the mercury was
oxidized to the bivalent oxidation state and was found at various concentrations in
the groundwaters depending upon the monitoring site. Currently, the
groundwaters are extracted from an upper perched groundwater table via a
drainage gallery. A facility has been constructed to treat extracted groundwaters
by the use of precipitation with dithiocarbamates, followed by polishing with
activated carbon and a specialty ion exchange resin. The water is pumped from
the gallery at mercury concentrations of 0.1-3.0 ppm and is currently treated to
allowable discharge limits of 10 ppb mercury.
Wells monitoring the ground water during the late 1980's showed seasonal
variations in the mercury concentrations. It appears that mercury levels decrease
in the dry seasons compared to the rainy seasons. Chemical speciation of the
mercury in the groundwaters was not rigorously determined, but speciation
studies on soils overlying the groundwater indicated the predominant species was
oxidized inorganic mercury. The composition of other elements in the
groundwater seems to change with the seasons as well, but an average composition
is given in Table 1. Variations in mercury content over a four year monitoring
period in waters from two wells, which are well separated from one another, are
shown in Table 2.
225
-------
TABLE 1. AVERAGE COMPOSITION OF MERCURY-CONTAINING
GROUNDWATERS
Constituent
Concentrations (mg/1)
Chloride
Sodium
Calcium
Magnesium
Total Dissolved Solids
pH
5,800
2,900
460
440
11,000
8.0
TABLE 2. SEASONAL VARIATION OF MERCURY CONCENTRATION IN
MONITORING WELLS
Month/Yr
Oct/1
Nov/1
Dec/1
Jan/2
Mar/2
Apr/2
May/2
Sep/2
Dec/2
Feb/3
Sep/3
Dec/3
Apr/4
May/4
Jun/4
Aug/4
Sep/4
Oct/4
Well 1
(mg/1)
9.60
3.35
0.29
5.50
3.80
10.00
4.20
7.70
6.10
6.20
8.50
2.70
4.00
4.00
4.40
5.80
7.70
13.00
Well 2
(mg/1)
0.370
0.293
0.426
0.230
0.390
0.200
0.300
0.370
0.510
0.500
0.240
0.140
-
0.260
0.170
0.180
0.086
0.240
EXPERIMENTAL PROCEDURES
Mercury analyses were performed using the EPA Method 245.1 of cold vapor
atomic absorption spectroscopy (5). A Perkin Elmer Model 3030B AAS instrument
was calibrated daily for mercury, and a calibration verification record was
maintained using data collected by the analysis of EPA certified check standards.
Preparation of standards for mercury analysis were performed in accordance with
226
-------
the specifications in Methods for the Chemical Analysis of Water and Wastes (5).
Spiked samples were analyzed with each batch of samples to determine if matrix
interference existed, and frequent blanks were run to ensure there was no
mercury carry over during analysis.
Mercury concentrations in groundwaters, column effluents and
regenerating solutions were determined by linear regression calibration curves
generated from four point standard calibration analysis (5).
Samples collected in the field pilot studies were split and sent to Woodward-
Clyde Consultants, EER Technologies and Bio-Recovery Systems for mercury
analysis.
Laboratory tests on the efficiency of mercury adsorption on AlgaSORB®
were determined using small glass columns (1.5 cm i.d. x 20 cm) which contained
the sorbent. Mercury-containing groundwaters were pumped through the
column at flow rates which varied from 6-20 bed volumes per hour. Effluents
from the columns were collected using a fraction collector and mercury content
was determined. Once the columns became saturated or leaked mercury above
discharge limits (10 ppb), the column was stripped with 10-bed volumes of a
selected stripping reagent followed by 10-bed volumes of deionized water.
Analyses of stripping effluents were performed to verify stripping.
Pilot studies were conducted with a small portable effluent treatment system
which has two AlgaSORB®-containing columns in series and which is capable of
treating flows of up to 0.5 gpm. On-site pilot testing was conducted on August 30 to
September 1, 1989 and November 6 to December 1, 1989.
RESULTS AND DISCUSSION
Samples of groundwater were collected at various times during 1989. All
samples were acidified to pH 2 with nitric acid in the field prior to transport for
laboratory studies. Once the samples were received at Bio-Recovery Systems, the
solutions were neutralized to the original pH with dilute sodium hydroxide.
Laboratory and field studies were complicated by the fact that over a 10 month
period, mercury concentrations changed by an order of magnitude. Table 3 shows
mercury concentration variation over the sampling period.
Different species of algae can be immobilized to produce different
AlgaSORB® resins. Since different bioploymers comprise the cell walls of different
algae, some species of algae behave differently from others with respect to metal
ion binding. Thus, different AlgaSORBs containing different algal species were
tested for mercury removal from the groundwaters. However, these experiments
were complicated by the fact that consistent mercury removal performance was
not observed using a single immobilized alga on waters collected at different times.
For example, Table 4 shows mercury contents in effluents from columns
containing AlgaSORB 602. The results shown in Table 4 can not simply be
explained by variation in mercury content in the influent waters. For example,
227
-------
Column C shows lower leakage levels of mercury than column B in the first 40 bed
volumes of effluent even though the influent concentration of mercury was
nearly three times higher in column C than column B. This suggests that a
variation in the chemical species of mercury may occur with time.
TABLE 3. MERCURY CONCENTRATIONS IN GROUNDWATERS
Sample Number
103-13089
176-42089
177-42089
265-070589
343-090189
368-100489
369-100489
PH
8.5
8.0
8.0
7.9
7.8
7.9
7.9
Mercury
Concentration
(Hg/D
150
435
144
1120
620
1550
1550
Date
Collected
01-30-89
04-20-89
04-20-89
07-05-89
08-31-89
10-04-89
10-04-89
TABLE 4. TEST OF AlgaSORB602 ON MERCURY-CONTAMINATED
GROUNDWATERS t
Bed Volumes
of Effluent
1-4
17-20
37-40
54-60
69-72
73-76
93-96
113-116
133-136
149-152
A*
Effluent Hg
(Ug/0
0.5
0.8
1.3
4.0
.
2.2
2.3
3.0
5.0
6.5
B*
Effluent Hg
(Jig/1)
9.9
10.1
21.8
14.8
.
31.0
C*
Effluent Hg
(Hg/0
1.3
3.4
8.1
27.0
72.5
t Groundwalers collected at various times were pumped at a flow rate of 6 bed-volumes per hour
through identical columns containing AlgaSORB 602. Effluents from each column were collected
in four bed-volume fractions and analyzed for mercury content. Influent flow rate was six bed
volumes per hour.
* Column A influent water was collected January 30. 1989, and had an influent mercury
concentration of 150 ng/1. Column B influent water was collected April 20, 1989, and had an
influent mercury concentration of 435 ng/1. Column C influent water was collected July 5, 1989,
and had an influent mercury concentration of 1120 ng/1.
228
-------
Over 99 percent of the mercury was stripped from the columns in Table 4 by
the passage of 10 bed volumes of 1.0 M sodium thiosulfate through the column (data
not shown).
After examining several different AlgaSORB® preparations and noting
similar types of variations as shown in Table 4, it was decided to settle on two
different AlgaSORB® resins for final testing. While these two resins could have
been blended into a single column, they were placed in two columns which were
connected in series and from which effluents samples could be taken from each
column for mercury analysis. Table 5 shows results of these experiments.
Data in Table 5 shows that the two columns arranged in series were effective
in mercury removal to below one ppb through passage of 250 bed volumes of
mercury contaminated waters which contained 1550 ppb mercury.
On site pilot testing of the AlgaSORB® resins was performed from November
6, 1989, to December 1, 1989. Two columns (2.54 cm i.d. x 81 cm) were separately
filled with AlgaSORB 624 and AlgaSORB 640. The columns each had a bed volume of
400 ml and were connected in series. Mercury-contaminated waters were pumped
through the two columns and two bed-volume fractions (800 ml) were collected,
split and sent to HER Technologies, Woodward-Clyde and Bio-Recovery Systems for
analysis. Results of on-site pilot testing are shown in Table 6.
TABLE 5. TEST OF AlgaSORB 624 AND AlgaSORB 640 ON MERCURY-
CONTAMINATED GROUNDWATERS*
Bed Volumes of Effluent Effluent Hg (u.g/1)
0-12
12-24
24-36
48-60
60-72
84-96
108-112
132-144
168-180
192-204
252-264
288-300
312-324
324-336
0.3
0.2
0.2
0.3
0.5
0.7
0.8
0.9
0.8
0.9
0.6
0.6
2.0
1.9
* Two columns (1.0 cm i.d. x 37 cm) coupled in series were used. Groundwaters collected October 4,
1989, and containing 1550 ng/1 mercury were passed through the columns at a rate of six bed-
volumes per minute. Ten bed volume fractions were collected and analyzed for mercury. Data
shown above are mercury concentrations in effluents from the second column.
229
-------
By the time the on-site testing bad begun in November the mercury
concentration in the ground waters had changed from about 1500 ppb (in October)
to about 700 ppb (see Table 6). With the exception of the first fraction collected,
the data in Table 6 shows that over 500 bed volumes of mercury-contaminated
waters were treated before mercury in effluents approached the 10 ppb discharge
limit.
TABLE 6. ON-SITE PILOT TESTING FOR MERCURY REMOVAL FROM
GROUNDWATERS *
Mercury Concentration fue/1)
Bed Volumes of Bio-Recovery Woodward Clyde EER Technologies
Effluent Influent Analysis Analysis Analysis
7-8
85-86
163-64
229-230
289-290
313-314
343-344
379-380
415-416
449-450
467-468
503-504
533-534
587-588
A'.
f.(
9.5
5.3
2.1
1.4
1.8
1.9
5.5
2.0
1.8
4.9
4.0
5.8
7.7
10.5
l&* fifiO
) \> OOU
)\j / /U
14.2
8.0
3.6
1.4
2.6
2.4
9.3
3.1
3.2
7.8
7.2
9.6
10.3
13.0
780
/ ou
11
<10
<10
<10
<10
<10
10.0
<10
<10
10.0
<10
<10
<10
15
AOO
790
/ 4*\)
' A portable water treatment system was equipped with two columns connected in series. The first
column was filled with AlgaSORB 624 and the second was filled with AlgaSORB 640. Groundwaters
were pumped through the system at a flow rate of 6 bed-volumes per hour. Effluent samples were
collected and sent to Woodward-Clyde Consultants, EPA (EER Technologies Corporation) and Bio-
Recovery Systems for analysis.
* Influent samples were collected for analysis just prior to 436 and 600 bed volume fractions of
effluent were collected.
CONCLUSIONS
Initial laboratory testing of AlgaSORB9 resins clearly showed promise for
mercury recovery from contaminated groundwaters. Once the mercury was loaded
on the resin, it could be stripped with sodium thiosulfate. On-site pilot testing
confirmed the laboratory studies and showed that AlgaSORB* resins treated over
500 bed volumes of groundwaters before mercury levels in effluents exceeded
discharge limits.
230
-------
JCES
1. Darnall, D.W., Greene, B., Henzl, M., Hosea, M., McPherson, R., Sneddon, J.
and Alexander, M.D., Binding and Recovery of Gold(III) and Other Metal
Ions from Aqueous Solution by Algal Biomass, Environ. Sci. Technol. 20:
206, 1986.
2. Darnall, D.W., Greene, B., Hosea, M., McPherson, R., Henzl, M., and
Alexander, M.D., Recovery of Heavy Metal Ions by Immobilized Algae, 1m
R. Thompson (ed.), Trace Metal Removal From Aqueous Solution, The Royal
Society of Chemistry, London, 1986, p. 1.
3. Greene, B., McPherson, R. and Darnall, D.W., Algal Sorbents for Selective
Metal Ion Recovery, In: J.W. Patterson and R. Passimo (eds.), Metals
Speciaiion, Separation and Recovery. Lewis Publishers, Chicago, Illinois,
1987, p. 315.
4. Darnall, D.W. and Gabel, A., A New Biotechnology for Recovery Heavy
Metal Ions from Wastewater, Jn; Proceedings of the Third National
Conference on New Frontiers for Hazardous Waste Management, U.S.
Environmental Protection Agency, EPA-600/9-89-072, 1989, p. 217.
5. Methods for the Chemical Analysis of Water and Wastes. EPA-600/4-79-020,
U.S. Environmental Protection Agency, Revised March 1983 and
subsequent EPA-600/4 Technical Additions Thereto, Cincinnati, Ohio, 1983.
231
-------
BIOSORFTIOH - A POTENTIAL MECHANISM FOR THE REMOVAL OF
RADIOHUCLIDES FROM NUCLEAR EFFLUENT STREAMS
Peter Barratt Ph.D
Biotreatment Limited
5 Chiltern Close, Cardiff CF4 5DL, UK
Abstract
Fifteen fungi and bacteria were isolated from both metal-
contaminated and uncontaminated environments, and tested for
their tolerance to strontium, ruthenium and cobalt in agar media.
Three fungal isolates, Trichoderma viride, Penicillium expansum
and Aspergillus niger, were examined for their ability to remove
these metal ions from solution. T. viride and P. expansum
removed 100% of cobalt and strontium from solution within 2 days,
wai
-1
at concentrations of lOmgl" , although biosorption of metals was
much less effective at concentrations between, 30 and 200mgl
Composition of the liquid medium was found to affect biosorption
of the metals studied. Additional studies were undertaken on the
biosorption of these metals by immobilised biomass of
Saccharomyces cerevisiae and A. niger in laboratory filter bed
systems. Removal of >50% of ruthenium in a model effluent stream
was measured in a system containing A. niger.
Introduction
Of the various stages involved in processing and reprocessing in
the nuclear fuel industry, many produce low level wastes in the
form of aqueous effluents which contain both radioactive and non-
active contaminants that can be considered as hazardous to the
232
-------
environment (Schumate et al, 1978). These reprocessing effluent
streams vary widely in composition. Some of the most significant
radioactive isotopes in these streams are fission products, such
as 137Cs, 144Cs, 89Sr, 90Sr and 106Ru, but large quantities of
non-radioactive effluent components, notably nitrates, also occur
during fuel reprocessing.
It is now well established that many microorganisms have the
capacity to accumulate metallic cations from the environment, via
a process generally referred to as biosorption. In the microbial
cell this phenomenon is usually considered to comprise two
distinct phases, those of active uptake, dependent upon the
metabolism of the cell, and metabolism-independent cell surface
binding (Gadd, 1986; Norris and Kelly, 1976).
i
Cation accumulation by microbial cells is a complex process. The
degree of concentration inside the cell (absorption) or on the
cell surface (adsorption) are dependent upon the cationic species
and upon the properties of the cell in terms of such factors as
cellular charge, metal requirement and tolerance, and competition
for active binding sites amongst different cations, as well as
environmental factors such as pH, temperature and the presence of
other chemicals in solution (Gadd, 1986; Ross and Townsley,
1986). However, where conditions are conducive it has been found
that there is a rapid accumulation of cations from solution,
often in the first minutes following the addition of microbial
biomass (Kiff and Little, 1986)
The uptake of radionuclides by microbial biomass and biomass cell
wall constituents has been observed in previous studies, and the
biosorption of Co , Sr , Cs and Ru has been demonstrated on
a number of occasions (Gadd, 1986; Norris and Kelly, 1977; Ross
and Townsley, 1986). On the basis of these observations there
233
-------
appears to be some potential for the use of microbial biomass as
a means of removing radionuclides from contaminated effluent
streams.
Many of the treatment processes currently in use for the removal
of active contamination from effluent streams demonstrate high
efficiency where substantial concentrations of radionuclides are
present. However, where effluent metal concentrations lie in the
lower range of 1 to lOOmgl" , and amelioration is still required
before discharge, some of the classical treatments become less
economical and even ineffective. It is in this area that novel
microbial-based techniques may be able to offer considerable
benefits in the future (Schumate et al, 1978).
3. Experimental methods
3.1. Biomass screening and selection
A number of microorganisms were isolated from the natural
environment using standard agar plate isolation techniques, and
maintained in axenic culture. These organisms included fungi,
bacteria and yeasts from a variety of sources, including metal-
contaminated soil from a former waste lagoon (Luton,
Bedfordshire, UK), contaminated leaf surfaces near a smelting
works (Avonmouth, UK), and uncontaminated agricultural sources,
notably silage and hay.
As an initial indication of the organisms' tolerance to elevated
metal concentrations, each isolate was cultured on agar plates
containing a range of concentrations of the non-radioactive metal
species ruthenium (101Ru), strontium (88Sr), cesium ( Cs) and
cobalt (59Co). These metals were selected as being
representative of four important cationic constituents
effluent streams from the UK nuclear reprocessing industry.
-------
Agar plates were prepared containing 0, 5, 10, 50, 100, 200, 400
and 500mgl~ of each metal. Fungal and yeast isolates were
streaked onto malt extract agar (Oxoid, UK), and bacteria onto
nutrient agar (Oxoid, UK), each containing the appropriate
concentration of metals in the form of ammoniated ruthenium
oxychloride (CltH.-N-.O-Ru.,), and the chloride salts of
strontium, cesium and cobalt. All plates were incubated over a
period of 5 days at 25°C, after which the presence or absence of
microbial growth was recorded.
Some of the microorganisms obtained were later identified to
species level.
3.2. Biosorption studies in liquid culture
The adsorption of cobalt and strontium by living cell cultures of
four of the microorganisms isolated in 3.1 was tested in liquid
culture.
Three fungal and one bacterial isolate were tested for their
growth and metal uptake in pure culture. Fungal isolates A3, A6
and A8 were grown in two liquid media; malt extract (ME) broth
(Oxoid, UK) to simulate an organic-rich environment, and Czapeks-
Dox (CD) broth (Oxoid, UK), a synthetic inorganic medium. The
bacterial isolate, B4, was cultured in nutrient broth (NB; Oxoid,
UK).
Conical flasks containing 100ml of the selected sterile media
were prepared. A range of concentrations of strontium and cobalt
had been established in these liquid preparations, by the
addition of SrCl- and CoCl0, to give concentrations of each metal
-l
as follows: 0, 10, 20, 50, 100 and 200mgl . Flasks were
inoculated with the relevant cell culture to give the following
treatments at each metal concentration:
235
-------
A3 in ME A3 in CD
A6 in ME A6 in CD
A8 in ME A8 in CD
B4 in MRD
Each treatment flask was duplicated to give a total of 84 liquid
samples. After inoculation of the flasks under sterile
conditions, each was plugged and incubated at 25°C on a rotary
shaker (ISOrpm) for 48h.
Following incubation cell culture media were filtered through a
packed glass bead column and centrifuged at 6000rpm for 15min.
The supernatant from each treatment was collected for cobalt and
strontium analysis by atomic absorption spectrophotometry (AAS).
Uninoculated control samples for each liquid medium at each metal
concentration were also analysed for the two metals.
3.3. Metal analysis
All metal analyses were performed using a Perkin Elmer AAS (model
2380), with the appropriate cathode lamp for each element.
Cobalt was analysed at a wavelength of 240.7nm, ruthenium at
349.9nm, and strontium at 460.7nm. All three metals were ionised
in a nitrogen/acetylene flame. Lanathanuro chloride was used to
enhance the sensitivity of the machine to ruthenium and
strontium.
Metals in the samples were measured against standard analytical
solutions.
236
-------
3.4. Biosorption studies in a model effluent stream
Using information from various sources in the UK nuclear fuels
industry, a model 'low level' effluent stream was prepared in the
laboratory. This comprised the following non-radioactive
components:
Ruthenium (as RuCl.s) 5mg
Cesium (as CsCl2) 5mg
Strontium (as SrCl2) 5mg
Cobalt (as CoCl2) 5mg
Sodium nitrate 20000mg
Sodium hydroxide SOOmg
pH 8
Deionised water 1000ml
Model filter beds comprising approximately 350g of a mixture of
washed sand and gravel (3:5 wt/wt), supported on a woven polymer
mat, were constructed in Buchner funnels in the laboratory. A
header tank (1 litre capacity) with a perforated base, of the
same diameter as the Buchner funnel, was suspended above each
filter bed, to allow even distribution of liquid from the tank
across the surface of the bed as it percolated through. Control
beds contained sand and gravel only, and biomass treatments
incorporated lOg dry weight of either S. cerevisiae or A. niger,
evenly distributed amongs- the gravel base.
Biomass had been collected as waste products from the brewing
industry (Bass Brewing, UK), and a citric acid fermentation plant
(Rhone-Poulenc, UK) respectively, and had been air-dried prior to
incorporation into the gravel filters. Each treatment system was
duplicated, giving a total of six filter systems, and each system
was thoroughly rinsed through with distilled/deionised water, and
allowed to drain.
237
-------
Effluent stream (1 litre) was pumped into each header tank at a
known rate, via a series of peristaltic pumps, and as the stream
passed through the filter bed, samples were collected for
analysis. For each filtration system, 4 x 200ml treated samples
were collected at 200ml intervals, so that at the end of the
experiment, each system had collected one sample of the
following: 0-200, 200-400, 400-600 and 600-800ml.
Collected samples were analysed for strontium, cobalt and
ruthenium by AAS.
4. Results and Discussion
4.1. Metal tolerance screening
Table 1 gives the results of the metal tolerance tests. In all,
15 organisms were screened in this way, and tolerance appeared to
vary widely between isolates. Seven organisms were identified to
species, and these were: Saccharomyces cerevijsiae, Penicillium
spinulosum, Penicillium expansum, Aspergillus niger, Trichoderma
viride, Zoogloea ramigera and Bacillus subtilis. The results
show that only 3 of these (Al, A3 and A6) were able to grow on
agar media containing the highest concentration of the four
metals (SOOmgl ). These three organisms were all identified as
fungi or yeasts. Isolate Al was S. cerevisiae, A3 P. expansum,
and A6 T. viride. The bacterial isolate, B4, which showed
limited growth at a concentration of 400mgl~ , was B. subtilis.
These four microbial isolates were selected for further work. In
addition, fungal isolate A8, (A. niger), was selected.
The concentrations of metals in the media used in this stage of
the study were for the purpose of organism selection only. Due
to the metal binding properties of some organic molecules it is
unlikely that the concentrations specified represented the
238
-------
available metal ion concentrations present in the agar media.
The interaction of metals with non-biomass organic materials
should always be borne in mind during experiments of this kind.
4.2. Metal uptake from liquid media
Figures 1-4 illustrate the results of microbial treatment of Co2+
ions in metal-supplemented liquid media at concentrations of 200,
3 On
.2+
100, 50 and 30mgl~ respectively. Figures 5-8 give similar
results for Sr
Cobalt concentrations in CD medium treated with P. expansum were
considerably less than any other treatment after incubation of
the 200mgl~ solutions. This treatment demonstrated a 22%
decrease when compared with the control samples with no inoculum.
A similar pattern was detected in media containing an initial
concentration of lOOmgCol" , where the treatment inoculated with
P. expansum reduced Co concentrations by 26% after 48h
incubation (Fig. 2).
At 200mgl , substantial Sr accumulation in biomass occurred in
both P. expansum and T. viride cultures (Fig. 5). Differences
between Sr uptake from the two media were not well defined at
100 and 200mgSrl~ , although Co uptake by P. expansum was
clearly greater in CD. This could have been due to the binding
of metal ions to organic molecules in ' ME broth at higher
concentrations, which may consequently neutralise the ionic
charge and reduce the overall affinity of the metal for cell
surfaces (adsorption). The results for Sr at 100 and
200mgSrl show no such difference between biosorption in the two
mycological media, which suggests that Co*""1" may have a greater
affinity than Sr for components of the ME broth.
239
-------
The pattern of increased biomass affinity for Co in CD medium
is consistent for P. expansum at all metal concentrations
•
At lOmgl P. expansum and T. viride both demonstrated 100%
removal of Co from CD broth (Fig. 9). The data for Sr2"*"
accumulation in these two organisms were very similar to that of
Co , with complete removal of lOmgSrl taking place from both
media after incubation with P. expansum, and 85% reduction
occurring in the T.viride treatment in CD broth (Fig. 10).
Removal of Co and Sr by T. viride, although almost completely
effective in CD, was poor in ME broth.
In general, incubation in the presence of A. niger had little
effect upon the concentration of metal ions in solution. At
_i
lOmgl only minimal reduction in Co occurred, and the maximum
reduction observed in Sr was only 17% of the untreated control.
4.3. Metal uptake from effluent stream
Figures 11 and 12 illustrate the concentrations of metals in the
effluent stream after filtering through the sand and gravel beds
containing A. niger and S. cerevisiae respectively.
Both biomass treatments resulted in a marked decrease in the
metal loading of the stream in the first 200ml treated when
compared with the non-biomass controls. Following this initial
decrease in the concentrations, metal levels rose by
approximately 50% in all treatments between 200 and 400ml, and
generally these levels, all markedly lower than the controls,
were maintained between 200 and 800ml. Differences between the
three samples taken from 200-800ml were minimal for all metals in
either treatment.
240
-------
Although the trends discussed occur in nearly all the data sets,
the results for ruthenium in the effluent treated with A. niger
differ, in that, between the samples taken from 0 and 800ml,
there were no significant differences, and ruthenium was
maintained at <50% of that in either the influent or the
controls. In this way Ru adsorption by A. niger appears to be
the most efficient of the treatments applied to the filter bed
systems.
Results from the biomass filter systems demonstrated some
biosorption of Co, Sr and Ru from an acid effluent by A. niger
and S. cerevisiae. The positive results obtained for dried
A. niger biomass, compared with those from the cell culture
experiments, indicates the influence of both the form of the
bioroass and the environmental conditions prevalent within the
liquid on cation accumulation by microbial cells.
Conclusions
The biosorption of Sr and Co from axenic liquid cultures of
PeniciIlium expansum and Trichoderma viride has been
demonstrated. Complete biosorption of both Sr and Co occurred
where the initial concentration was 10mgl~ . At concentrations
above 30mgl~ accumulation of metal ions was less efficient
during 48h incubation. There was no clear pattern of biosorption
at metal concentrations increasing between 30 and 200mgl .
The uptake of Sr and Co by a culture of Aspergillus niger was
generally poor in liquid media, even at low initial
concentrations (10mgl~ ).
There appear to be significant effects of the chemical
environment on biosorption by fungal biomass in a liquid system.
Other organic components in the liquid phase may inhibit
241
-------
biosorption of Co and Sr. It seems likely that there are complex
interactions in operation during the biosorption process which
influence its effectiveness. Such factors may include
competition between cation species, metal sequestration with
organic molecules in solution, pH and the physical form of the
biosorbant matrix.
The use of living microbial cells in continuous culture systems
for the removal of metal ions from nuclear effluent streams may
be a feasible proposition where metal concentrations are low, and
effluents can be held in treatment tanks or holding ponds prior
to treatment, although further verification of the processes is
required both in stable and radioactive streams. Subsequent
extraction of bioaccumulated metals into a small volume of an
inorganic matrix is also likely to be a prerequisite for a viable
treatment process, as biosorption treatment has the potential to
give rise to higher level waste from low level streams, and such
waste is likely to require encapsulation. Encapsulation of
organic wastes is not a favoured practice, in the nuclear
industry, and although the elution of active metal species prior
to disposal is clearly an option for such a treatment process,
this may be more difficult where ions are accumulated inside as
well as outside the cell (Gams, 1986).
References
Gadd. G. M. (1986) The uptake of heavy metals by fungi and
yeasts: The chemistry and physiology of the process and
applications for biotechnology. In: Immobilization of ions by
bio-sorption (eds. H. Eccles and S. Hunt). Published for the
Soc. Chem. Ind. by Ellis Horwood, U.K.
242
-------
Kiff R. J. and D. R. Little. (1986) Biosorption of heavy metals
by immobilised fungal biomass. In Immobilization of ions by
bio-sorption (eds. H. Eccles and S. Hunt). Published for the
Soc. Chem. Ind. by Ellis Horwood, U.K.
Norris P. R. and D. P. Kelly. (1977) Accumulation of cadmium and
cobalt by Saccharomyces cerevisiae. J. Gen. Micro. 99:317-324.
Ross I. S. and C. C. Townsley (1986) The uptake of heavy metals
by filamentous fungi. In Immobilization of ions by bio-sorption
(eds. H. Eccles and S. Hunt). Published for the Soc. Chem. Ind.
by Ellis Horwood, U.K.
Schumate II S. E., C. W. Rancher, G. W. Strandberg and C. D.
Scott. (1978) Biological processes for environmental control of
effluent streams in the nuclear fuel cycle. In Waste Management
and Fuel Cycles 1978. Proc. Symp. Waste Management, Tucson,
Arizona.
Acknowledgeinents
Grateful thanks to Deirdre O'Toole for her technical support
during the course of this project, and to the Department of the
Environment (UK) for funding the work.
243
-------
TABLE 1
ORGANISM VIABILITY AT 25°C
TN
ORGANISM
YEASTS AND FUNGI
Sacc. cerevisiae (Al)
Pen. spinulosum (A2)
Pen. expansum (A3)
Fungus A 4
Fungus A 5
Tr. viride (A6)
Fungus A7
Asp. niger (A8)
BACTERIA
Bacterium Bl
Bacterium B2
Bacterium B3
Bacillus subtilis (B4)
Bacterium B5
Bacterium B6
Zoo. ramigera (B7)
THE PRESENCE OF VARYING
CONCENTRATIONS OF Ru.
Sr, Cs and Co
4- : growth
- : no growth
na : no result available
METALS CONCENTRATION (mg.l'1;
0 5 10 50 100 200 400
+ na + + + 4- 4-
+ na 4- +
4- na 4- 4- 4- 4- +
+ 4- 4- 4-
4- 4- 4- + 4- 4-
+ + + 4- 4- 4- 4-
+ + + 4- +' - -
+ + 4- 4- + - -
+ + + + 4-
+ + + 4- + 4- -
+ 4- 4-
+ + + + + + +
+ + 4-
+ + 4-
+ + + ----
500
+
+
-
4-
-
-
-
-
-
-
—
244
-------
COBALT CONCENTRATION (mg/1)
COBALT CONCENTRATION (mg/I)
w
o
6
o
ro
-P>
en
i
z
O H
E *
w B
M "
Ol
o
I
(O
§
M
Ol
O
O
O
to
003
II
oo
z«
i
1
31
m
O
w
0
§
w
>
&
m
OB
>-
wO
o<
o>
3r
oO
b«
°U
-i
op
c>
3ir
^
°5
is
§>
v
2«
FROM
CUBATI
oo
Z?
^
M
O
c
»
w
1— *
-------
COBALT CONCENTRATION (mg/1)
o
j
Ol
o
1
o
I
o
I
9
O
o
o
03
H
t s
§<
3r
-CO
zo>
§5
a
H
31
m
COBALT CONCENTRATION (mg/l)
M
•0
M
O
B
r
o
3r
§S
§?
m
w
LJ
-------
FIGURE 5
STRONTIUM REMOVAL BY FUNGAL BIOMA88 FROM CULTURE
MEDIA (200 mg.Sr/1) DURNG 48h INCUBATION
860
20O-
160-
100-
60-
12 CONTROL
B CZAPBKS-DOX (B MALT EXTRACT
A.NK3EH
T.VRCE
FUNGAL SPECIES
P.EXPAN8UM
FIGURE 6
QTHONTIUM REMOVAL BY FUNGAL BIOMASS FROM CULTURE
STRONTMEDIA (100 mo-Sr/l) DURING 48h INCUBATION
1
o
8
260
20O-
16O-
100-
S CZAPEKS-DOX B MALT EXTRACT
60-
-------
^>
?
FIGURE 7
STRONTIUM REMOVAL BY FUNGAL BIO MA S3 FROM CULTURE
MEDIA (60 mg.Sr/l) DURING 48h INCUBATION
80-
70-
eo-
60-
40-
30-
20-
10-
0-
C3 CONTROL
A.NK3ER
12 CZAPEKS-DOX EB MALT EXTRACT
T.VRDE
FUNGAL SPECIES
P.EXPAN8UM
o
g
E-
FIGURE 8
STRONTIUM REMOVAL BY FUNGAL BIOMASS FROM CULTURE
MEDIA (30 mg.Sr/l) DURING 48h INCUBATION
80-
7O-
60-
60-
40-
30-
20-
10-
(2 CONTROL
A.NUER
12 CZAPEKS-DOX B MALT EXTRACT
T.VRDE
FUNGAL SPECIBS
P.EXPAN8UM
248
-------
STRONTIUM CONCENTRATION (mg/1)
COBALT CONCENTRATION (mg/l)
ro
c
z
o
M
•D
O
a-
o
m
I I I I I I I I
« o 2
I I I
(0
H
3)
O
cn
c
»
M
is
11
30 Q
OoQ
do
02
20
5
33
m
c
o
•o
M
o
9
a
o
m
O
O
ID
8
*
m
§
VŁ>
DS
03 3D
51
ip
H
V
m
-------
FIGURE 11
METAL CONCENTRATIONS IN EFFLUENT STREAM
AFTER TREATMENT WITH ASPERQILLUS NIGER
m CONTROL
E3 300-400ml
600-800ml
ED 0-200ml
OH 400-600ml
COBALT
STRONTIUM
RUTHENIUM
METAL SPECIES
FIGURE 12
METAL CONCENTRATIONS IN EFFLUENT STREAM AFTER
TREATMENT WITH SACCHAROMYCES CEREVISIAE
CONTROL
200-4OOml
600-aOOml
S3 0-200ml
IE 400-600ml
63
a
COBALT
8TRONTWM
METAL SPECIES
RUTHENIUM
250
-------
BlOTRCX...
• ... for environmental solutions, naturally!
Biological Treatment of Wastewaters
by
Thomas J. Chresand
Dennis D. Chilcote
BioTrol, Inc.
Presented at
U.S. Environmental Protection Agency
Second Forum on Innovative Hazardous Waste Treatment
Technologies: Domestic and International
Philadelphia, Pennsylvania
May 15-17, 1990
ABSTRACT
The BioTrol Aqueous Treatment System (BATS) was demonstrated
as part of the Superfund Innovative Technology Evaluation (SITE)
program for treatment of groundwater contaminated with
pentachlorophenol (PCP). The system employs indigenous
microorganisms; however, it is also amended with a specific PCP-
degrading bacterium. A mobile trailer-mounted system was used for
the demonstration. Three flow rates were tested, corresponding to
residence times of 9, 3 and 1.8 hours. PCP removal ranged from
97.6 to 99.8 percent, with average effluent concentrations as low
as 0.13 ppm. It was shown that biological degradation was the
predominant removal mechanism while air stripping and
bioaccumulation were negligible. Acute biomonitoring with minnows
and water fleas showed complete removal of toxicity by the
treatment system.
The BATS has also been shown to be highly effective for
treatment of a variety of wastewaters, including process and lagoon
waters, and for. a wide range of contaminants.. Results are
presented on treatment of gasoline, phenolic, and solvent-
contaminated wastewaters.
BioTrol, Inc -11 Peavey Road • Chaska, Minnesota 55318 • 612/448-2515 • FAX 612/448-6050
251
-------
TECHNOLOGY DESCRIPTION
Reactor Design
Reactor design is a critical aspect to successful
implementation of this technology. BioTrol employs a multi-stage,
submerged, aerated, fixed-film reactor that provides a high biomass
concentration and, therefore, reduced reactor volume. The reactor
and a schematic of the trailer-mounted system are shown together in
Figure 1. The multiple stages create a dispersed plug flow of
wastewater through the reactor. The plug flow configuration is
important in that contaminant concentrations typically fall in the
range of first order removal kinetics. That is, the rate of
removal is proportional to the concentration of contaminant. A
completely mixed tank system operating at a low effluent
concentration will experience low removal rates. A system with
plug flow characteristics, on the other hand, will experience the
same low rates in the effluent sections; however, the upstream
sections will operate at very high rates, thus yielding higher
overall removal rates.
The use of a fixed-film system allows for a long cell
retention time and, therefore, lowered production of excess
sloughed biomass. Moreover, the fixed-film system eliminates the
often problematic biomass separation step which is crucial to
successful operation of an activated sludge system.
Bacterial Amendment
Many of the priority pollutants can be degraded by
microorganisms indigenous to a given wastewater. For these
compounds, treatment can be accomplished by simply adding the
appropriate inorganic nutrients and allowing time for acclimation.
However, in cases where a highly toxic or recalcitrant compound is
to be treated, the appropriate microorganisms may not be present.
In these cases, treatment can be accomplished by adding organisms
with the appropriate degradative capabilities. This technique,
called microbial amendment, is finding increasing use as
microbiologists continue to isolate organisms with novel metabolic
pathways.
As an example of microbial amendment, a Flavobacterium species
is used by BioTrol for treatment of pentachlorophenol-contaminated
wastewaters. This microorganism can perform rapid mineralization
of pentachlorophenol at concentrations up to 200 ppm.
252
-------
PREVIOUS WORK
In the fall of 1986, a nine-month pilot study was performed to
demonstrate the feasibility of groundwater treatment at a wood
preserving facility. The study was funded by a grant from the
United States Geological Survey. A single-stage packed bed reactor
(30 gal) was activated with indigenous microflora and later amended
with inoculations of a Flavobacterium species. The target compound
in this study was PCP, present in the groundwater at concentrations
from 60 to 100 ppm. A typical monthly performance chart is shown
in Figure 2. As shown in Figure 2, the flow rate for most of the
month was 25 gph, corresponding to a residence time of 1.2 hours.
The study proved the effectiveness of amending a microbial
consortium with a unique organism to treat a highly toxic
groundwater. The specific rate of PCP degradation demonstrated was
as high as 70 mg PGP/liter reactor volume/hour - or 105 Ib PCP/1000
cf/day - a loading rate which would be high even for treatment of
the BOD of municipal sewage. Moreover, the nine-month study
demonstrated the stability of such a microbial system.
BATS Performance
Figure 2
253
-------
RESULTS OF SITE DEMONSTRATION
In July and August of 1989 a demonstration of the BATS was carried
out under the U.S. EPA SITE program. A wood treating site in
Minnesota was -chosen as the location for the demonstration.
Groundwater at the site was contaminated with PCP, creosote, and
heavy metals (copper, chromium, and arsenic). The goals of the
demonstration were to:
1) Reliably measure the removal of contaminants across the
reactor at a series of different flow rates
2) Carefully monitor the fate of PCP to determine the
predominant removal mechanisms. In particular, to determine
complete or partial degradation and the extent of air
stripping, and/or bioaccumulation
3) Determine the decrease in toxicity of the groundwater after
biological treatment
4) Determine operating costs for treatment of similar
wastewaters
A mobile trailer-mounted system with a 5 gpm capacity was used for
the demonstration. The test was carried out over a six-week
period, and consisted of treatment at three flow rates for 2 weeks
each. The flow rates investigated were 1, 3, and 5 gpm. Since the
packed volume of the reactor was 540 gallons, these flow rates
corresponded to residence times of 9, 3 and 1.8 hours respectively.
PCP Degradation
Table 1 shows the results of PCP removal by the BATS. The
data are presented as averages over the two-week period of each
flow rate. Two important points are illustrated by the data in
Table 1. First, that treatment of PCP was effective up to 42 ppm,
and second, that even at the highest loading rate (5 gpm) effluent
concentrations below 1 ppm were achieved.
To determine the mechanism of PCP removal, analyses of
chloride ions in the influent and effluent were performed. The
complete destruction, or mineralization, of PCP will yield chloride
ions in a 5:1 molar ratio (C1~:PCP). Thus, if PCP is disappearing
based on chromatographic analysis, and Cl~ is not appearing in a
stoichiometric ratio, it is likely that the PCP is being either
partially degraded or transformed. Likewise, analyses of air
exhaust samples were performed to determine whether air stripping
was a significant removal mechanism. The results of these
measurements are given in Table 2.
254
-------
The data in Table 2 show that more Cl~ was produced than would
be expected by stoichiometric degradation of PCP alone. It is not
clear why the amount is higher than expected; however, the
hypothesis of PCP mineralization is nonetheless supported.
Likewise, the air stripping results indicate that biodegradation
was the predominant removal mechanism.
Concentrations of the various PAHs in the incoming well water
were lower than expected. In many cases the concentrations were
below detection limits. Thus, no conclusions on biodegradation of
PAHs were drawn from the study. BioTrol has measured significant
removals of 2 to 4 ring PAHs in other studies. Low levels of
arsenic and various other heavy metals were found in the system.
There was little or no change in concentration upon passage through
the system.
Bioassavs
It was anticipated that the groundwater would be toxic to
aquatic organisms since LC50 values for PCP tend to be quite low.
Thus, biomonitoring was performed to determine the system's ability
to detoxify the groundwater. Fathead minnows (P. promelas) and
water fleas fD. Macma) were used for the bioassays. Table 3 shows
the results of these bioassays.
%
Table 3
Bioassay Results
of wastewater in test
Minnow
Week Influent Effluent
1 0.35
2 0.84
3 0.26
4 0.54
5 0.61
6 0.66
>100
>100
>100
>100
>100
>100
water
Water
Influent
0.3
1.1
0.43
0.3
0.2
0.2
Flea
Effluent
>100
>100
>100
35
>100
>100
The values in the table indicate the percentage of influent or
255
-------
TREATMENT OF OTHER CONTAMINANTS
The technology demonstrated in the SITE program has vide
applicability for the treatment of wastewaters contaminated with
various toxic organics. In fact, treatment of PCP demonstrated a
"worst case" in that PCP is one the most highly toxic of the common
priority pollutants. BioTrol has had success treating a variety of
other wastewaters with the BATS approach. These waters range from
very high strength process water with high concentrations of
substituted phenols to low strength groundwater contaminated with
trace concentrations of benzene.
Figure 3 shows the performance of a full-scale BATS unit
treating gasoline-contaminated groundwater (benzene as primary
contaminant) at 15 gpm. This system has consistently performed >99
percent removal since installation, and the effluent gualifies for
discharge without a polishing step.
BATS Performance
Benzene Treatment
5000
4000
3000
2000
1000 -
influent, ppb
Effluent, ppb
40
30
20
10
0
10
15
20
25
Days
256
-------
BATS Performance
Treatment of Solvent-Contaminated
Process Water
Influent Effluent
(mg/L) (mg/L) % Removal
Methylethylketone 43.0 <0.005 >99.9
Total BTEX 1.3 <0.01 >99
Tetrahydrofuran 5.7 0.014 >99.7
Total Unknown Peaks 5.0 <0.05 >99
In summary, the BATS technology demonstrated under the SITE
program has wide applicability beyond treatment of PCP-contaminated
groundwaters. Many contaminants are amenable to biodegradation by
indigenous microorganisms, and high removal efficiencies can be
achieved at relatively high loading rates. Contaminants amenable
to treatment include gasoline components, solvents, PAHs, and
substituted phenols.
257
-------
A-S Biotekniskjordrens
BIOTECHNICAL SOIL PURIFICATION OF SOIL POLLUTED
BY OIL/CHEMICALS
Presented at:
"Second Forum on innoyative Harzadous
Waste Treatment Technologies -
Domestic and International"
Hay 15 - 17, 1990
Philadelphia, Pennsylvania.
SUSANNE SGHI0TZ HANSEN
H.Sc.
Administration &Anlaeg Telefon Telefax Giro
Maglehojvej 10 53504668 53504490 8465800
4400 Kalundborg 33156872 A/SReg.nr. 151387
258
-------
BIOTECNICAL SOIL PURIFICATION OF SOIL POLLUTED BY OIL/
By Maniging director H.sc. Susanna S. Hansen,
Biotecnical Soilclean Ltd.
SUMMARY;
Soil polluted by light oils and by some easily decomposable
cemicals such as esters, katones and alcohols may be cleaned
mikrobioally in about 2-8 months. Heavier pollution will
require a processing time of 8 - 16 months.
The processing is one of oxidizing, watering, and adding
nutrients and bacteria to the soil! ««x"y
Soil polluted by organic solvents may be cleaned by gas
extraction, with subsequent cleaning of the air in a biofilter
and/or coal filter, in 1 - 4 weeks (MIBK, chlorbenzene, toluene,
6tC•J•
BACKGROUND - MICROBIOLOGY
Oil and many chemicals are organic substances - and may thus be
decomposed microbially.
60% of a surface garden soil contents of fugi and bacteria being
oil decomposing.
Descreibed for the first time in 1895 when Miyoshi deschreibed
how the mould penetrate parafine.
More than 100 species of.microorganisms belonging to bacteria,
mould fungi and yeast fungi are said to be able to decompose oil.
DEGREPATION - CONDITIONS
First and foremost OXYGEN but also soil moisture (appr.60%),
supplementary nutrients, and right pH (7-9 ) are elements
basic to how the oil/chem. decomposing bacteria thrive.
The decomposition works well down to 8 - 5 degrees of C, but a
higher temp, will promote the decomposition.
SLUDGE FARMING
The pricipeles of "using11 natures own forces has been applied for
some years to get ridof oil sludge from refineries where this
sludge has been scattered on agricultural areas and mixed with
the surface soil by harrowing.
Small guatities of oil (1%) would seem not to harm the plants
(but it may be a danger to the groyundwater). On the contrary,
one may notice that grass the year after will grow better on the
oilspot than outside.
259
-------
SOME PROBLEMS -
ray grass
•' stress"
that AROMATE
It has been shown that at 2% radishes and
growth inhibition. The radish root s show
way of short, corkscrew curled roots.
Flourescence microscope photos indicate
penetrate.
The oil are able to percolate to the groundwater( few dops
1000 1 of drinkingwater).
Heavevy degredable parts of the oil and heavy metals
accumulate.
will
CONTROLED "SLUDGE FARMING"
A place where the oil soil could be taken in; Created in a
controled way; without any waste to the enwiromental; and put back
into use; seemed to be a useful solution to the problems related
to sludge farming but using the positive elements such as:
* it is a cheap cleaning method
* it is not energy consuming
** S!sSmeŁhoSywS!lUno? gfne^an^Iecondary pollution (smoke
or polluted water:
The drawbacks of this method are:
* it is fairly slow
i it doe^not dTSnySing about the inorganic pollution elements
such as heavy metals:
BIOTECNICAL SOILCLEAN LTD.
the plants which has been in function for three years; are build
after a system; which has been arranged in giving high priority
to the safty to the envinment; as would appear from the plant
sketch below: two sets of drainage systems (nos: 2 and 4)
diaphragm (3). Fibertex protect the diaphragm. The top-layer is
jointed SF-brlcks. A specially close-vibrated concrete brick.
Rainwater is recycled from the percolate bassin, which is airedso
it may be sprinkled back over the stacks.
Thus generating propagation of bacteria.
Nutrients and tentatively bacteria are added to the bassin.
RAMPART KITH
PLANTATION
Fig. i.
Outline of plant.
Watercanal
for surface
outflow
to reservoir
Pile
.6 cm SF-stones
.Stable gravel
•Primary drainUI)
Protective matta)
Membrane tl)
Secondary
260
Aeration _inotallnt ion
Reservoir for oer-
colator-and surface-
water
-------
OPERATION AND OPERATIONS EXPERIENCE
SOIL DESCRIPTION:
The way this place is run requires rather strict control
soil received: we need to be sure that the soil may be re1
at no risk, and without running too much of finanslal ris)
should be able to clean the soil within the time allowed
financially, the ideal thing would, of course, be if the soil had
been through analysed before arriving - but this is rather
unrealistic.
Analyses provide answers merely to what was asked - and the
answers being greatly dependent on the analysis method, too.
The method seemed by us to be the most acceptable is: interwevs
of the landlords and GC/FID (capillary column) to chek the
content of extractable contents in the soil.
Fig. 2. Gascromatogrammes showing the oilcontents in the
same soil shown as funktion of the time.
On the site, appr. 100.000 tons of soil has been treated, divided
on appr. 630 cases.
The major part of this soil volume has been polluted by oil
products (diesel, light fuel oil, turpentine, petrol, heavy fuel
oil) minor quantities( 10%) having been polluted by MIBK,
acetone ect).
Degredation times ranged: for the light fractions, it takes
two to eight , and for the heavy fractions, from eight
sixteen months.
from
to
261
-------
Fig. 3. Airphoto of the plant.
The acceptance criteration for soil for recycling is 100 ppm so
far. For the light fractions, we normally do not recycle the soil
until the content has come down to appr. 50 ppm.
As most of the polluted soil is raw-soil, there is certain
recycling limitations. It will generally be used as filler soil.
ORGANIC SOLVENTS
It is not acceptable, due to the working enviroment and for
enviromental reasons, to clean soil polluted by organic solvents
by this method. Due to this fact, Biotecnical Soilclean Ltd.
cleans the soil for this type of pollution by gas extraction with
subsequent cleaning of the air in a biofilter and/or coal filter.
262
-------
Description of Gas-extraction Tent.
For the tine being, we have two, one stationary at the processing
site at Kalundborg, and one mobile which may easily be moved from
one site to an other, thus conducting "on site" purification.
The basic idea of the two is the same, that of
1 a dense bottom
2 one layer of specially
developed concrete brick,
being perneable (MULTI BLOCK
3 a conpreaaed air-unit.
4 an air - tight tent of an
organic solvent proof diejhrag
diapragn
5 a blover, providing sub-
pressure in the tent
6 a blofilter
7 a coal filter
§ test tubes
Fig. 4. GAS-EXTRACTION TENT.
The mobile tent has been built up over a slightly enlarged open-
top container, enabeling the whole system to be pulled up on a
truck and moved.
TESTS.
In two rounds, Institute of Tecnology, Denmark, has conducted
measurings in order to evaluate how much would be absorbed by the
biofilter, and how much would have to be absorbed by the coal filter,
We were aiming at having as little as possible land up in the
coal filter.
1. Measuring series (table 1) showed in the first round
that 30 - 90 % of what was blown out (toluene) was
absorbed by the biofilter.
263
-------
TABLE It
MEASUREMENTS OF THE REDUCTION OF TOLUEN IN THE MR WHEN CONDUCTED
THROUGH A COMPOST FILTER AS WELL AS A CHARCOAL FILTER.
POINTS OF
SAMPLING:
j>
C
C
I:
C
(ppm)
8.6
1
0
1.4
0.7
0
0.9
0.6
0
TOLUEN
(mg/m*)
33.2
3.8
0.1
5.1
2.0
0
3.5
2.3
0
REDUCTION
B9
100
45
100
34
100
0.3
1.3
62
Airsaroples points:
A: Compost filter inAtake.
B: Compost filter exit.
C: Charcoal filter exit.
Method:
Airsamples in charcoal tubes.
Analysis on G.C.
Comments.
The reasons why the figures were no higher nay be that the
bacteria on the graftet filter had been starved for some
months before we got the tests going, an that the
concentration of the rather large air volumes blown trough
were too low to keep a dense bacterial strain alive.
Various tests have been run in the permanent site tent.
As these tests were required to be realistic, the series of tests
was controlled by what emerged in the way of current cases.
I.e., the pollution cases cropping up should also be
*
*
*
*
of a suitable size, i.e. it should be possible to
?rocess it in max two rounds in the tent,
t would have to be a case polluted by organic solvents
only, and not, for instance, a nixed pollution with oil,
they were to await our notifying the county of the test
and our recieving their reply,
and the pollution were not to include chlorinated
compounds which are not handled by us on the site.
264
-------
To achive, in a single test, concentrations appropriately h,
the air stage to give clear measuringresults, we designed an'
accident by pouring five liters of toluene out in the soil in the
tent.
This will undoubedly produce the best blow-uot as the toluene
will not have time to penetrate into the micropores of the soil.
The measuring is just taking place.
With these tests, for which no mass balance have been made
(beyond the artificially generated pollution), it may prove hard
to 5udge haw much pollution has been blown off, and how much has
been decomposed by pumping oxygen into the soil.
Measuring results.
TABLE 2;
SUMMARY OF THE VARIOUS SOILS CLEANSED IN THE BLOW-OUT TENT.
CASE
NUMBER
470/89
515/89
Mixture
507/89
(104)
POLLUTION
ether
dibutyl-
phtahalat
toluen
clorbenzeen
CONSEHTRATION
(ppm in soil)
START FINISH
loot
26
1500
170
to
0. 5t *
(n.d.)
n.d.
n.d.
n.d.
/•»•*
TIME
(days)
9
36
10
49
2e
2O tummtr
This eerie has been measured as reduction of the total area
of the gascromatogramrae.
CASE.
The Mobile Tent case has been working since December on a
pollution caused by chloride benzene.
The pollution had been excavated and wal left in containers.
In this case the biofilter was left out, as chlor benzene is
toxic forbacteria, thus the filter would have been useless (apart
from the fact that some amount of absorbtion would probably have
taken place).
This cleaning was made under the worst imaginable conditions.
* In each container, there was about 25% more soil than
there should be in the tent. For economical reasons, a
whole container was emptied at a time.
* The soil was clayey and wet.
* The cleaning was effected in the coldest months of
December, January and February.
* The blow-out could only take place between 1500 and 1700
hours, as demanded by the building workers on the site.
* For periods of time, site electricity was switched off.
Requirements were made by the County jof Frederiksborg:
* The soil was to be cleaned down to 0.1 ppm, for later
deposit on a local dump!
365
-------
ppm
50
40
30
20
10
I
\
\
\
v..._
****f
3 10 20
DAYS
30
Fig. 5. REMOVAL OF CHLORBENZEZE FROM SOIL BY GASEXTRACTION.
A less expensive way of heating the air would undoubtedly
improve upon this method during the winter half of the year.
266
-------
In Situ Physical and Biological Treatment
of Volatile Organic Contamination:
A Case Study Through Closure
presented to:
United States Environmental Protection
Agency
SECOND FORUM ON INNOVATIVE
HAZARDOUS WASTE TREATMENT
TECHNOLOGIES:
DOMESTIC AND INTERNATIONAL
May 15-17,1990
Wyndham Franklin Plaza
Philadelphia, PA
presented by:
Richard Brown, Ph.D.
Groundwater Technology Canada, Inc.
Richard Tribe, P.E.
Ultramar Canada, Inc.
GROUNDWATER
TECHNOLOGY, INC.
267
-------
Introduction
Even with aggressive policies of preventing product losses from occurring or from
having substantial impact, incidents can and will happen. Additionally with the transfer of
properties, problems can be inherited, either knowingly or unknowingly. In this context,
it is important to have a systematic approach to addressing environmental incidents.
Ultramar Canada, Inc., a refining and marketing company operating in Eastern
Canada, has been a proponent and user of a systematic approach to addressing the loss
of petroleum products. The recently completed clean-up of an Ultramar site in a suburb
of Montreal illustrates the application of this approach.
There are three basic elements that Ultramar addresses in systematically
responding to any leak or spill. The first is the safety of the general public, employees
and facilities. Once a site has been secured, and all immediate danger has been
addressed, the second element becomes central to an effective response. This element
is an assessment of the situation to define the impact of the problem - short and long
term, the context of the problem - its geological, regulatory and public settings, and the
extent of the problem - the amount(s) and distribution of the product loss. When the
situation is defined, an appropriate remedial action, the third element of response, can be
adopted. Selection of a remedial action involves setting the clean-up criteria and then
choosing the technology (s) best suited to achieving those criteria.
The case study that will be reviewed in this paper involves a service station in
suburban Montreal (Figure 1). The product loss was the result of leakage over a long
period. Eventually, the product migrated across a busy intersection to a commercial
facility that served both as a retail store and a warehouse/distribution facility. Petroleum
vapors accumulated in the subgrade warehouse facility resulting in a potentially explosive
environment. The remediation project that will be discussed in this paper, was initiated
in response to safety issues created by the vapor accumulation.
•Copyright Groundwater Technology, Inc. 1990
All Rights Reserved.
268
GROUNDWATER TECHNOLOGY, INC.
-------
Safety
When a hazardous situation exists or may potentially result from a product loss, the
number one priority must be safety. This means protecting the immediate and long term
well being of the general public, employees at the impacted facility(s), and the facilities
themselves.
The most immediate hazard which can result from a product loss is the migration
and accumulation of vapors emanating from contaminated soils and/or free phase
products. These vapors may represent an acute hazard in the form of explosive
environments. Vapors can accumulate in basements, utility lines and buildings with
subgrade construction.
A second, longer term hazard due to a product loss is the potential health impact.
This may be due to the inhalation of vapors or the ingestion of contaminated groundwater.
Where a safety issue exists, the immediate response must be to secure the
impacted facility. This means stopping the source and removing the acute hazards:
Once the situation is secured the safety focus must be to address the longer term safety
issue such as health impacts.
The safety issues that were created at the study site were a result of the long term
product loss, and were the accumulation of vapors in the basement of the commercial
building across the street from the site. The vapors entered the basement through small
cracks in the southern wall and floor (facing the service station). At times, these vapors
approached explosive levels. For the most part, the vapors were present at
concentrations well below explosive levels (Figure 2). They were however, detectable by
smell and created an odor nuisance and potential health risk for workers in the basement
warehouse.
Ultramar's response to the safety concerns was to install a vapor abatement
system, consisting of a vacuum extraction and air intake system keyed into the crushed
269
JDD GROUNDWATER TECHNOLOGY, INC.
-------
stone fill below the floor. This system had an immediate impact on the vapor levels
(Figure 3).
Ultramar's second response was to install an interceptor well between the impacted
commercial building and the service station. This well served to intercept any potential
dissolved or free phase product migrating from the station (Figure 4).
It should be noted that Ultramar undertook these steps without it being fully
established that a significant loss had occurred at Ultramar's facility or that the loss was
the source of the vapors in the commercial building. There were other underground tanks
to the west and south of the impacted building which could also represent potential
sources. However, being the closest service station and because of the immediate safety
issues, Ultramar dealt with the safety problems at hand.
With the safety issues in hand, the next phase of response was to assess the
source and extent of the problem.
Assessment
Assessment involves three factors. The first is identifying the sources of the
contamination - i.e. line leaks, tank leaks, overfills etc. The purpose of this identification
is to stop the source. The second factor involved in assessment is to identify the
migration route(s) of the contamination. The purpose of identifying routes is to assess the
potential for impact and thus the near and long term liability. The third factor addressed
in an assessment is to determine the extent of the contamination both areally and degree.
Knowing the extent is important to assessing, again, the potential impact and/or also
important to deciding on the appropriate course of action.
During an assessment, the geology, hydrogeology and contaminant distribution
(adsorbed, dissolved, and free phase) need to be determined. This is accomplished by
the installation of monitoring wells and by obtaining soil borings.
270
GROUNDWATER TECHNOLOGY, INC.
-------
At the study site, a series of monitoring wells were placed around the service
station and on adjacent properties by Qroundwater Technology, Inc. The wells were
gauged and sampled. As shown in Figure 5, the Ultramar station did appear to be a
source of contamination and did appear to be impacting the commercial building. Wells
in the vicinity of the tank pits had several inches of free phase product indicating that the
probable source was a tank leak. The adjacent station also appeared to be a potential
source of contamination, based on elevated dissolved VOC levels at that station.
The groundwater contours from the site show that product loss at the station could
be drawn into the store basement, Figure 6. As can be seen, there is a natural
groundwater low at the store. This low draws groundwater radially into the area of the
store.
The well logs were used to develop two geological cross sections of the site
(Figures 7 & 8). The reason for this groundwater low is evident from the geological cross
section. The basement of the store was blasted into the bedrock underlying the general
area. The bottom of the basement, being below the water table necessitated the
operation of a french drain and surnp to keep the basement dry. This operation caused
a draining of the groundwater from the area into the basement sump. The drainage
pattern is basically radial. Any contamination in the area would be drawn to the store
basement. This is what appears to have happened with the Ultramar site. The
contamination resulting from long term losses at the Ultramar site was drawn across the
roadway by the continual operation of the basement sump.
One of the potential complications from this site geology is the ability to access the
adsorbed phase contamination. In fractured bedrock, it is not unusual to have small
pockets, "hot spots", of contamination trapped in small fractures. These fractures are
often difficult to remediate and their presence can result in the long term persistence of
low levels of dissolved VOC.
A second focus of a site assessment is to define the phase distribution of the
petroleum product that has resulted from the loss. There are basically four phases of
271
in GROUNDWATER TECHNOLOGY, INC.
-------
contamination - free phase, adsorbed, dissolved, and vapors. The actual distribution is
a function of the site geology, the amount lost, and the type of product. For gasoline, the
relative percentages are as follow: free phase 20-70%, adsorbed 20-70%, dissolved 0.5-
3.0%, and vapors 0.1-1.0%. In fractured bedrock, the amount of adsorbed contamination
can be elevated due to physical retention in the fractures.
These contaminant phases are in equilibrium with each other. Because of this
equilibrium, addressing one phase often necessitates addressing other phases,
particularly when the focus is either vapors or dissolved contaminants. At the study site
the initial concern was the appearance of vapors in the store basement and the drawing
of dissolved contamination into the basement sump. As the source of the vapors is the
free phase and/or adsorbed contamination in the bedrock and fill. Solving these
problems necessitates addressing the presence of adsorbed/free phase at the station and
in the fill underlying the basement. Figure 9 shows the approximate extent of
contamination at the site as determined during the assessment. A small amount of free
phase product appeared in monitoring wells downgradient of the tank pit. Adsorbed
phase contamination extended across the street and into the basement fill. The adsorbed
phase extends through the fill into the surface of the bedrock.
Remedial Action
Remediation of a product loss involves not just the successful application of
technology, it also entails identifying and satisfying the stakeholders, those that have an
interest in the problem, and determining and meeting the criteria that will be used to
regulate the cleanup effort. For a remediation effort to be successful, it must reasonably
satisfy the stakeholders and meet the predetermined cleanup criteria. Thus the
stakeholders and the cleanup criteria become the context for evaluating and choosing a
remedial plan.
The stakeholders are persons who are affected by the loss and/or by the operation
• of the remedial system, and persons who have to give the necessary approvals for any
cleanup program. The list of stakeholders can be quite long. Basically there are four
272
inn GROUNDWATER TECHNOLOGY, INC.
-------
groups. Company personnel are an important group of stakeholders. These include not
only the operators of the impacted facility but also the insurance, marketing, environmental
and legal departments. A second group are the regulators that must be dealt with. These
include both local and provincial agencies. Local officials who may be involved are
municipal officials such as the health department, the fire department and police. The
provincial agency was the ministry of the environment. A third group of stakeholders is
the general public, both those immediately impacted as well as those in the immediate
area who may be "concerned" about activities on the site. The final group of stakeholders
are the contractors hired to address the problem.
With the stakeholders there are two items of concern. The first is the protection
of their well being, that is the health and safety of impacted parties and the economic well
being of the company. The second is the understanding of stakeholders of the
effectiveness/appropriateness of the remedial action(s) chosen. If either concern is not
addressed there may be significant resistance to the remedial program. Education is,
therefore, key to obtaining the necessary approvals. Education entails both an accurate
assessment of the problem and documentation of remedial options.
The cleanup criteria chosen for the site must address the interests of the different
stakeholder groups. Obviously the priorities and relative weights of these interests need
to be considered. At the study site the cleanup criteria chosen were as follow:
o Restoration and maintenance of satisfactory working conditions in the
warehouse structure.
o Minimum disruption of operations in warehouse facility and Ultramar Station.
o Work be carried out in a reasonable time frame.
273
LQ GROUNDWATER TECHNOLOGY, INC.
-------
o The solution is permanent.
o The solution is cost effective.
Within this context of identified stakeholders and cleanup criteria, a number of different
cleanup options were evaluated. This evaluation was conducted in two stages. In the first
stage all possible options were listed in a "brain-storming" session without regard to cost
or practicality. In the second stage the options were evaluated in relation to the
established cleanup criteria.
The treatment options that were considered consisted of the following:
Purchasing impacted business
Excavating soil under basement
Pump and treat
In situ bioreclamation.
The first three were rejected. Purchasing the building was neither a permanent solution
nor cost effective. Excavating the basement did not address contamination outside the
building, would be extremely disruptive and very costly. Pump and treat would not
effectively restore working conditions and is not effective against adsorbed phase
contaminants.
In situ bioreclamation was chosen as the treatment option that most fully satisfied
the cleanup criteria. Bioreclamation addresses all phases of contamination. It destroys
the contaminant, so it is a permanent solution. It can be completed in a reasonable time,
2-3 years, and is, after installation, not disruptive of routine operations. Having a system
which could operate without inconveniencing the workers in the warehouse was an
important factor in choosing bioreclamation. While expensive compared to non-
permanent solutions such as pump and treat, bioreclamation was a cost effective solution
for this particular location in that it brought a permanent resolution to the contamination.
274
nm GROUNDWATER TECHNOLOGY, INC.
-------
Even with the selection of a remedial operation, there were still several tasks that
needed to be accomplished before the system could be installed and operated.
One such task was to select a contractor. The most important criteria for choosing
a contractor is a proven track record. Documented case histories and operating sites that
can be visited are invaluable in obtaining a first hand evaluation. Based on these criteria,
Ultramar chose Groundwater Technology, Inc. as the remedial contractor. Groundwater
Technology had successfully completed and was currently operating a number of
bioreclamation projects in the United States.
The second pre-installation task was to gain regulatory approval. Many of the
regulatory questions about impact migration, disruption to the business, safety,
performance were addressed in a report. Other questions were handled in a face-to-
face meeting. Key to Ultramar's gaining regulatory approval was the "homework" that
Ultramar undertook before meeting with the ministry. Ultramar visited bioreclamation sites
in the United States to see the operations and to talk with effected parties. This enabled
Ultramar to more knowledgeably answer the ministries' questions. Having satisfied the
ministries' concerns, Ultramar obtained approval to proceed.
Principles of Bioreclamation
Bioreclamation, in situ aerobic biological treatment, is one of the most versatile
remediation processes, dealing with a wide range of organic compounds in a number of
different hydrogeological conditions. In situ enhanced bioreclamation is a proven method
for remediating groundwater aquifers contaminated with petroleum hydrocarbons and
many organic chemicals. It is simply the use of common aerobic soil bacteria to degrade
organic contaminants. It involves the stimulation of indigenous bacterial through the
addition of essential nutrients. The bacteria used in the process are already there.
275
m GROUNDWATER TECHNOLOGY, INC.
-------
The process can be simply viewed as a step-wise degradation as pictured below:
Biodegradation of Organic Chemicals:
Step-wise Metabolism
CONTAMINANT CELL MATERIAL MINERALIZATION
Bacteria Bacteria
(C,H) (C,H,N,P,0)
Bacteria which can feed directly on the contaminant use it to grow, producing primarily
cell material. Looking at this first step one can see what is required for stimulation. The
contaminant is primarily carbon and hydrogen; cell consists of carbon, hydrogen,
nitrogen, phosphorus, and oxygen (with minute amounts of minerals). The key to
accelerating this natural degradation process is to add sufficient nitrogen, phosphorous,
and oxygen (N,P,O) to balance the available carbon and hydrogen (the contaminant). On
a contaminated site, biodegradation is already occurring but is very slow because the
bacteria quickly expend the naturally occurring nutrients and oxygen.
Once the specific degraders convert the contaminant to cell material they die and serve
as a readily utilizable food source for other bacteria. This secondary metabolism also
requires continued addition of oxygen and nutrients to sustain high biological activity. The
end result of the overall process is the conversion of the contaminant into carbon dioxide
and water.
The key to in situ bioreclamation is getting the nutrients and oxygen to the
contaminated area. The first stage in this process is to hydrogeologically create a
reaction vessel by pumping and injecting groundwater. This accomplishes two things.
First, the treatment area is confined; second, the contaminant and nutrients (N,P,O), the
reactants, are mixed. Since normal groundwater flow ranges from 10 to 200 feet per year,
it could take years for nutrients and oxygen under natural conditions to traverse even a
276
JLD GROUNDWATER TECHNOLOGY, INC.
-------
small site. By increasing the gradient through pumping (draw-down) and injection
(mounding), transport times are reduced from years to months or days. Since the
availability of nutrients and oxygen limit the process, the more quickly they can be added,
the faster the remediation. Compared to the rate of nutrient transport, the biological
processes are essentially instantaneous.
Once the groundwater system is properly installed, maintaining a proper nutrient (N,P)
balance is relatively straight forward. The nutrients are generally soluble ammonium and
phosphate salts. They can be added at quite high levels and are eventually recycled by
the biological process. Thus nutrient supply is easily managed.
With nutrients, the key management principle is maintaining a threshold concentration.
Beyond this threshold level additional nutrients do not benefit the biodegradation process.
The threshold level is a function of two factors: the basic requirements of the metabolic
cycle, and the degree of sorbtion and retention of the nutrients by the soil matrix. The first
factor, the metabolic requirements is quite small. The ideal metabolic ratio of carbon to
nitrogen is 10:1, and carbon to phosphorus, 30:1. Soil retention of nutrients, however,
can be quite high on the order of 10's to 100's of ppm. It is this retention factor that drives
the nutrient concentration in applying bioreclamation. For example, the minimum nutrient
concentrations in an in situ bioreclamation system are typically 100 ppm versus 1-2 ppm
in a typical aqueous bioreactor. The difference is the impact of soil retention.
With oxygen, however, the supply/demand situation is quite different. Maintaining a
high oxygen tension is critical to biodegradation. First, considerably more oxygen than
nitrogen or phosphorus is required for biodegradation. Each kilogram of hydrocarbon
that is metabolized requires approximately 3.5 kilograms of oxygen to convert it to
and water:
Reaction: (Chy +1.5 Ob — > COb +
Weights: 14Kg 48Kg 44Kg 18kg
IDE 277
~TJJ GROUNDWATER TECHNOLOGY, INC.
-------
In unsaturated soils (above the water table) oxygen supply is relatively easy, requiring
only active aeration. Each 1000 cubic meters of air supplies approximately 250 kg of
oxygen. A soil vent system consisting of vacuum blowers, such as was employed under
the commercial building, is an effective means of aeration.
In saturated soils (below the water table) oxygen supply is much more difficult.
Natural dissolved oxygen levels in groundwater are 4-5 ppm; the maximum is 8 ppm.
Thus, if oxygen in a saturated system is supplied by air diffusers, a common method of
aerating water, approximately 125,000 liters of water would be needed to supply one
kilogram of oxygen. A more effective means of oxygen supply in a saturated system is
to use a chemical oxygen source, hydrogen peroxide.
Hydrogen peroxide decomposes by a variety of mechanisms to give oxygen and
water:
2HA —> 2^0 + Oj
During decomposition, each part of hydrogen peroxide supplies one half part of oxygen.
Decomposition is mediated by heavy metal catalysis, surface area effects, and biological
catalysis. The two most common decomposition catalysts are iron and enzymes such
as catalase. Most common aerobic bacteria contain catalase and therefore are able to
"use" hydrogen peroxide as an oxygen source.
An advantage of using hydrogen peroxide as an oxygen source is that it is miscible
in water and may, therefore, be theoretically added in any concentration. However,
hydrogen peroxide is also a known biocide, as evidenced by its use as a topical antiseptic
in many homes. Thus, it must be added at levels that are not biotoxic but are still able
to maintain high oxygen availability. The practical limit in peroxide concentration is
-2,000 ppm. At this level, 1,000 liters of peroxide amended water supplies close to one
kilogram of oxygen. This is greater than a two order of magnitude increase in oxygen
availability over air diffusers. Because of its ability to supply significantly more oxygen,
and because the low permeable soils limited water circulation, hydrogen peroxide was a
critical factor in the use of bioreclamation at the Ultramar site.
,-H-~- 278
mi:
ID GROUNDWATER TECHNOLOGY, INC.
-------
In choosing Groundwater Technology, Inc. (GTI) as the remediation contractor, one
of the factors that Ultramar considered in particular was GTI's experience with the use of
hydrogen peroxide. GTI has been actively employing hydrogen peroxide in bioreclamation
systems since 1983.
System Installation and Operation
The bioreclamation system was installed at both the Ultramar service station and
the commercial building. In both cases, the system was designed so that oxygen and
nutrients could be swept through the areas of contamination effectively stimulating
biological activity.
The service station system is pictured in Figures 10 and 11. Figure 10 shows the
injection recovery system, which consisted of two recharge galleries and an interceptor
trench keyed into a recovery well (RW-2). The first gallery was placed between monitoring
wells MW-1 and MW-2 to treat the area around the southern most pump island. The
second gallery was positioned to sweep the front of the station and treat the
contamination resulting from the tank-pit. The interceptor trench stretched across the full
front of the station and was positioned to prevent the migration of contamination towards
the commercial building. The process equipment for the service station system is pictured
in Figure 11. Extracted groundwater was air stripped to remove dissolved hydrocarbons.
A dilute peroxide solution (10%) was continuously metered into the stripped groundwater
at about 2,000 mg/L. On a batch basis, a nutrient solution consisting of ammonium and
phosphate salts was also added to the groundwater. This was accomplished by diverting
groundwater to a 300 gallon mix tank to make a nutrient concentrate. The concentrate
was then metered into the water stream over a 8-10 hour period. The amended
groundwater was then discharged into the recharge galleries. The injected water flow rate
was adjusted to maintain hydrogeological balance. Excess water was disposed of to the
municipal sewer. Total groundwater flow ranged 3-5 gpm. At this rate the system
supplied approximately 22 Kg of oxygen a day.
279
Em GROUNDWATER TECHNOLOGY, INC.
-------
Figures 12 and 13 show the bioreclamation system as installed and operated in the
commercial building. The injection/recovery system is pictured in Figure 12. The injection
system consisted of six injection points that were drilled through the basement floor and
screened in the gravel fill. Recovery of groundwater was done through recovery well RW-
1 and the french drain sump. Figure 13 shows the process equipment for the
bioreclamation system. Groundwater from the sump and RW-1 was amended with
hydrogen peroxide at -2,000 ppm and nutrients (on a batch basis) and reinjected into the
injection ports. A hydrogeological water balance was maintained by disposing of excess
water to the municipal sewer. Total groundwater flow rate was 3-5 gpm.
Results
The implementation and performance of the bioreclamation system varied by area
on the site. As shown in Figure 14, there were four focus areas in the site remediation
program.
Area "A", the area surrounding the southern most pump island generally evidenced
low initial levels of contamination. It was the first area treated and rapidly responded.
Area "B" was the area on the station around the tank pit (the source of the
problem). It was the most heavily contaminated with both free phase and adsorbed
phase contamination being present. This area was the second area treated and was the
area treated for the longest time.
Area "C" was the area between the service station and the commercial building.
While having some contamination present, this area was not specifically treated because
of accessibility. (Area "C" was under a heavily traveled roadway.) While not specifically
treated, contamination in Area "C", being downgradient, was none-the-less impacted by
the treatment systems in Areas A & B.
The final area, Area "D", was the basement warehouse. This area had both
dissolved and adsorbed phase contamination resulting in a vapor contamination problem.
DTK 28°
LED GROUNDWATER TECHNOLOGY, INC.
-------
Treatment consisted of venting and bioreclamation. This area was the last area to be
bioreclaimed. Response to bioreclamation was faster than observed in Area "B" due to
the lower level of contamination. The Area "D" system was terminated before the Area "B"
system.
The specific results for the bioreclamation systems are pictured in Figures 15-20.
Results are given for the individual focus areas.
Figure 15 shows the bacterial response to the addition of oxygen and nutrients for
the bioreclamation system on the service station and in the warehouse. Prior to remedial
activity the counts of hydrocarbon utilizing bacteria were 10* colony forming units
(CFU)/ml in both areas. The first phase of remediation was the circulation of treated
(aerated) water. This resulted in slightly higher counts in the treated wells relative to
background wells. Addition of nutrients with the aerated water resulted in a 2-3 fold
increase in hydrocarbon utilizing bacteria. Administering hydrogen peroxide to boost
oxygen availability resulted in an order of magnitude increase in hydrocarbon utilizing
bacteria. These results show that biodegradation requires the supply of nutrients and
high levels of oxygen.
The results for Area "A" are depicted in Figure 16. Prior to remediation the
maximum dissolved level was 15,000 ppb total VOC (volatile organic chemicals). With
simple pump and treat, this level was dropped to ~ 1,300 ppb at the beginning of the
bioreclamation phase. Continued reduction in dissolved VOC's was rapid with the
subsequent injection of nutrients and oxygen in recharge gallery #1. The VOC level
dropped from the 1,300 ppb to < 100 ppb in six months of treatment. At this point
injection through recharge gallery #1 was terminated. The post treatment levels of VOC
varied between 100 and 250 ppb over the next two years indicating that treatment had
been effective.
In Area "B", the results of treatment do not show the simple response as did those
for Area "A". This is due to the fact that the degree of contamination was much greater,
and the source was different, being a tank leak as opposed to surface spillage (during
281
ED GROUNDWATER TECHNOLOGY, INC.
-------
fueling of autos). Figure 17 shows that the initial pump and treat operation (recovery
through RW-2) had a substantial impact on the contamination reducing the dissolved VOC
level from the appearance of free phase to VOC levels < 10 ppm. With the implementation
of the bioreclamation system the VOC levels generally remained low, with the exception
of two major concentration spikes at 16-18 months and 28-30 months of total remedial
operation. The reason for these spikes is evident in the overlying plot of depth to water.
As can be seen, the increase in VOC levels coincides with a drop in the water table.
There are pockets of deep contamination which are normally "locked" in the bedrock.
When the water table drops significantly, these are "exposed" and result in an increase
in dissolved VOC. It should be noted that the second spike was much less than the first,
indicating that bioreclamation had effectively addressed even the deep contamination.
The residual contamination in Area "B" is now confined to an isolated hot spot where there
is still minor bedrock contamination. Because the groundwater VOC levels in this formerly
highly contaminated area have stabilized at <10 ppm, the bioreclamation system was
terminated.
Figure 18 shows the impact of the bioreclamation process on Area "C". The results
parallel those for Area "B". Generally, there was a reduction in dissolved VOC levels.
There is, however, a depth-dependent spike in VOC level. This, again, indicates that
there are pockets of bedrock contamination that impact dissolved levels at low water.
What is interesting in the results for Area "C" is that the concentration lags the depth to
water. This is due to the fact that Area "C" is downgradient. The lag time is due to the
transport time for the dissolved VOCs to migrate from Area "B" to Area "C". The results
for Area "C" show that general site contamination has been reduced. With the second low
water event (months 28-30) there was not a corresponding increase in VOC levels as was
seen after the low water event at months 18-20. This indicates that the bioreclamation
system is also addressing the trapped contamination in the bedrock.
Figures 19 and 20 depict the results for Area "D". The initial phase of operation in
Area "D" was the vapor extraction and interceptor well system. During this period the
dissolved VOC levels were reduced from 15,000 ppb to 3,200 ppb indicating that the
contaminant load was being reduced. With the implementation of the bioreclamation
282
GROUNDWATER TECHNOLOGY, INC.
-------
system the levels were reduced to <200 ppb in 10 months of treatment. During treatment
there were considerable fluctuations in VOC level due to changes in water table and the
activity of the bacteria. However, the trend was generally downward. At the end of
treatment the residual contamination was confined to an isolated area at the front of the
store (near RW-1). In this area the subfloor consisted of a tight clay fill which trapped the
contamination. The impact of this hot spot was negligible on the final results as shown
in Figure 20. At the end of treatment, the general air VOC concentration was under 500
ppb and the dissolved level <200 ppb. These were well below the target cleanup levels.
The bioreclamation/venting system has been removed.
Costs
Total project costs for the site have been approximately $850,000 direct costs. The
breakdown of costs are as follow:
Emergency response (vent system & interceptor well) $ 50K
Site investigation $ 80K
System design $ 50K
System installation $100K
Capital equipment $200K
Operations (3 years) $330K
Post closure monitoring $ 40K
The equipment was designed and installed as self contained, trailer mounted units
so that it would be reusable. With the termination of the bioreclamation program, the units
from this site have been removed and installed at other Ultramar sites. Thus the capital
expenditures were, in fact, a long term investment for Ultramar.
Conclusion
The loss of petroleum products from retail operations is a situation that should be
prevented to the fullest extent possible. However, such events can and do occur.
283
PZR
GROUNDWATER TECHNOLOGY, INC.
-------
When product losses do occur having a systematic approach, makes managing
the costs, liability and exposure a controllable process. The approach that Ultramar has
developed and effectively employs is an orderly process consisting of:
o Addressing the immediate safety concerns
o Defining the extent and impact of the loss
o Identifying driving forces for cleanup, i.e., stakeholders, regulatory factors,
and cleanup criteria
o Choosing appropriate remedial action (s)
o Installing and operating the remedial system
o Attaining closure.
The above general process can be used to deal with a wide range of loss
situations. When site specific factors necessitate aggressive action, the use of innovative
techniques such as bioreclamation can be used as part of this process to significantly
reduce contaminant levels and achieve a permanent solution. This paper documents the
successful application of Ultramar's response system to a site involving long term losses
of product resulting in the vapor contamination of an adjacent commercial building.
Bioreclamation was chosen, in this case, to eliminate persistent adsorbed phase
contamination.
Activity at the site was initiated by a safety concern - the accumulation of vapors
in a basement warehouse. Ultramars response was to address these immediate concerns
and then assess the extent of the problem and its long term implication. Because of the
present and potential for continued off-site impact, Ultramar chose an aggressive remedial
approach, bioreclamation, which could bring the site to closure.
HTW 284
:TD GROUNDWATER TECHNOLOGY, INC.
-------
The application of the bioreclamation system was conducted in two parts. One part
dealt with the contamination at the Ultramar facility. This was to treat the upgradient
source of the problem. The second part dealt with contamination specifically in and under
the commercial building. This was to restore the impacted building.
The total remediation program was conducted over a three year period. The
bioreclamation systems were operated over two and one half years. The implementation
of the bioreclamation was a phased program with different areas being treated for different
periods of time. The overall remedial program successfully attained its cleanup goals and
the site is currently being monitored.
In summary, dealing with product losses is a manageable process. It entails
developing a systematic process and using the appropriate tools. Diligence to both
technical and political (regulatory, social) issues makes the system work smoothly. A part
of the success of the remedial program at this site was the attention to this non-technical
issue. The bioreclamation system caused minimum inconvenience to the workers and
operations of the warehouse and therefore was not an added "burden". Second, the
professionalism of the contractor and the Ultramar personnel involved in the remediation
prevented any conflicts between environmental activities and normal site operations, and
kept all the stakeholders informed and satisfied.
285
GROUNDWATER TECHNOLOGY, INC.
-------
BIOGRAPHIC SKETCH
RICHARD A. BROWN
Dr. Richard Brown is currently the Director of Chemical Technology at Groundwater
Technology, Inc. His responsibilities include project management, system design, and
business development for in situ and on-site treatment of hazardous wastes. He has
developed and implemented remediation systems using biological and vapor extraction
technologies. Dr. Brown is responsible for development of new physical/chemical
treatment processes including advanced oxidation and metal recovery.
Previously Dr. Brown was the Director of Business Development of Bioremediation
Systems, a Division of Cambridge Analytical Associates, (CAA). Prior to joining CAA, Dr.
Brown founded FMC Aquifer Remediation Systems Division.
Dr. Brown holds a B.A. degree in Chemistry from Harvard University in Boston, a
M.A. degree and a Ph.D. in Inorganic Chemistry from Cornell University in Ithaca, New
York. He has authored 15 patents and is involved with many professional scientific
societies such as The American Chemical Society of Petroleum Engineers, The Water
Pollution Control Federation, and The American Chemical Society.
For more information on groundwater and soil remediation using bioremediation
technology, Dr. Brown may be contacted at Groundwater Technology, Inc. located at 100
Youngs Road, Mercerville, NJ 08619, phone number (609) 587-0300.
286
GROUNDWATER TECHNOLOGY, INC.
-------
J
©
•/
J>
MM
-\s
— •, — ArmOKHlATC UXATnUOF
\ UNOCHCHCMIO TANk
f
_J
-l—j
-K>
k >
r
i ••*
RMNC« UHDtRCROUNO TANKS— <^~^-ZIJl'
IISS
IIT»
MW-9
s.v * y /
Ss ~^\\
rvntn ISLANDS ^_^
^°
A-l
r~!' '
1 I1 1
1 l< '
— L.J'.-1_
MW-9
• i l^
* — RfCM**»CC 6*
*
'
/
,< .
LLCHV NO 1 -^
ft
J
nkitKiw;
ONUM-.
— i rtuci . 1
CMCIOSUNC v ' ]
""" ^^
r- Iti
OHAUST moit
BASCMCNT VCNTIlATKn
MW-S
•
CW-4
0
MW-IZ
• CW-3
0
MW-4 • A RW-I
^
•M
FIGURE 1
SITE MAP
PROJECT:
LOCATION
DRAWING NO.'
MONITORING DATE'
A RECOVERY WELL
• MONITORING WELL
• UNDERGROUND TANKS
weLl lOtNTWlCAtlON
ex • HW- i - veil, to
- w»it« T«gie f icv.ON rtm
© CORE WELL
0 «O
SCAII IN net
Mte "*< w-«r
» II
MW-U
•
' ,- INTCICCrTOO '»eNCM
amr^ MWH(<5 ^X^ IRCCHMCC SAueitT »OJI
Y"^"^"*"' _ ~~L— -^^ — Pours m.*NOS
! jj ~jj |! 1 ^ »eCHAH«€ CAlXtHT HO.*
n^^ "«°
1 r sowici
U STATION
U0H.W-I4 onua^ri-,
Wi-J^«' J
*1
-------
CO
BASEMENT LEVEL 1135
FIGURE 2
VAPOR CONCENTRATIONS
IN BASEMENT
(Total VOC ppm)
PROJECf
LOCATON'
DRAWING NO.' 19-8167 2
O VAPOUR PONT
• AIR SAMPLING LOCATION
© INJECTION PORT
• CORE WELL
OATC ••*.«.(.?
LCGEND
• MONITORING WELL
MW-4 • "ECOVERY WELL
VUVC f OK
VACUUM cannot.
GROUNOWATER (.
OL RECOVERY
SYSTEMS. WC
-------
ro
CO
FMMIMC a
VENTILATION SYSTEM
WITH VAPOR CONTOURS
BASEMENT LEVEL 1135
O VAPOU* POUT
• AIK SAMPLING LOCATIOM
0 INJECTION POUT
• CORC WELL
I I
GROUNOWATER&
OL RECOVERY
SYSTEMS. WC
-------
PO
VO
o
HGUME 4
GROUNOWATER CONTROL WITH
INTERCEPTOR WELL
PROJECT
LOCATION
DRAWING NO.' 15-8167-23
MONITORING DATE MARCH 11-67
A RECOVERY WELL
• MONITORING WELL
!_ ___ j UNDERGROUND TANKS
•III IDCNTiriCATION
in MW I -WC II 10
9O.ll - •»?!• 1»»U (1C VOX FCC n
-------
ro
vo
©
J<
MW
•
S>
- •^X""' *"*WlllilT€ IOCATWW OF
* UNOC ffdiouMO TANK
i
h ~"-i
j
-»
L ,
A
__ r 1
j i
romMft tmoCKiiouNO T«N«$ «C-i-"^_~ " " '
II 5»
IITf
XMW-9
V /**&
- /_ — ^
//-x
/* / 0
( /T~i
l &
KHtn taunt »>_ I O J
^V^ / « MW-7 /
•\ 1 " •30-°00/
^ 1 \ y
\r_.l\ir-| X
« !' f**-*
V i* !i i ^
Nj- • i _-/
., - /^-^
\> • >
*"
u
/
x^"
— ,.<
J
MmiMC
^— (IIUtnT (BOM
\ (AMUCNT VHHItHlOM
\ MW-6
\«
1131
MW-5
• >r ^ -.
/ STORE \
/X ^\ \
/7' \ \
FIGURE 5
PETROLEUM DISTRIR'JTION MAP
PROJECT.
LOCATION'
DRAWING NO • K- 0167 - 4
MONITORING DATE' FEB. 9-87
A RECOVERY WELL
• MONITORING WELL
1 J UNDERGROUND TANKS
Will IOCNlirK>T>ON
ft «»•! - WILL 10
- »»1IH HBll HCVIINfCCfl
O «O
SCAll W Ff < t
OATt «M ;o-«*
MOIt CONTOU* U«C IMOICAKS MtOKAClC
OTCHD OF VITHOLCUM COMtlMINMHM
/ MW-4 • A RW-I 1 1
' Z4.000 > 1
\ / 1 _»I«»-H
l«^ H«0 |
|t^«._^ tt«wet I CAiwet
LJ Simon 1
.--1 — s-±
MW 2 * '
* l>lo°
-------
ro
(O
ro
FIGURE 6
GROUNDWATER CONTOURS
PRO JET-
ENOROIT
DC SSI N NO 'IS B063- 3
DATE D'OBSERVATIONi M-ZS-86
• PUITS OBSERVATION
A PUITS OE RECUPERATION
ANOCNS IKSCWWIIS CNfOUIS
MfHT
DC LA VI NTH ATION
CHI >OUt-5at
KM NTiriCATtON DUfUITS
II MW-I IOCNT OUPUITS
> NM>n PWWAtKMX
RESERVOIRS ENFOUIS
CROUNOWATER FLOW
DIRECTION
GROUNOWATER
OIL RECOVERY
SYSTEMS, INC
-------
ro
VO
CO
FIGURC 7
GEOLOGICAL
CROSS SECTIONS
PROJECT-
LOCATION
DRAWING HO ' 9«l -130 -8167-3
MONITORING DATE
A RECOVERY WELL
• MONITORING WELL
'_ j UNDERGROUND TANKS
WELL IDC)*TI*(CATION
rx • MW i - we it 10
- WATCH TABie ClfVIINfffT)
© CORE WELL
0
IttTCKCCPTO* TftCMCH
IKCCHAKCt CALLCRT HO it
ram ISIAMOS
HCCMAMf CALICUT HO I
-------
\o
Figure 8 GEOLOGICAL CROSS SECTIONS
A
MW-2
SERVICE
STATION
B'
MW-I
I
MW-3 RW 2
' '
MW-II MW-4 MW-12
' ' . '
ORGANIC FILL
SILTY SAND GRAVEL
WEATHERED LIMESTONE
i 1,
LIMESTONE BEDROCK
MW-5
1
A1
MW-6
J
STORE BASEMENT
B
MW-2
MW-7
I
MW-8
I
SILTY SAND GRAVEL
WEATHERED BEDROCK
—
LIMESTONE BEDROCK
•50-
Horizontal Seal*
Vertical Seal*
LA COMPACNIE G. T. CANADA
CROUNUWATER TecH.vH.oct INC
-------
ro
vo
tn
FIGURE 9
PHASE DISTRIBUTION
• UNMKCROUM) 1«M«S
PROJECT-
LOCATION'
DRAWING NO.'94I-I3O-8I«T-S
MONITORING DATE-
A RECOVERY WELL
• MONITORING WELL
UNDERGROUND TANKS
OHAUST FROM
BASCMCHT VINTIIMION
wfLL IMMtlf ICtTIOK
[»,. UW-I -WJlt 10
- W<7€» T«*lf (ICV (IN F«T>
ADSORBED
INItHCtfTOK TIICNCH
IBtCM««Ct CAUCKT NO SI
rum niiuios
IWCNJIIWC •Au.or no t
n«o
scnvict
SIAtKW
ItlltlMHAHM HI N« llltWMX (JIMS l*«
-------
IV)
ID
CT>
VJ
1
v_x
Jf
^;
C^
V-iowwts tm
- «*TI(I OISCHA««C LINC
ricunc 10
BIORECLAMATION SYSTEM
SERVICE STATION
PHOJCCT UlTRAMAN CANADA INC
ioc« r ION-i UNGur utL.oueecc
ONAWING NO I5-07O2-4
A «ICOVt«Y WtLL
• MONITORING WCLL
f) MONITORING PORT
SC*l I t«i n f T
•«OTf U« O«G'. «^ 87O7 5 * C f
O< ICCTIOM^ A « TO I f
-------
rvi
10
SERVICE STATION
NUTRIENT INJECTION SYSTEM
-------
10
00
BASEMENT LEVEL
OXYGENATED WATER LINE 2*« PVC
ricuftc 12
STORE BIOREMEDIATOR
SYSTEM
PROJECT
LOCATON'
DRAWING NO.' I5-BI67 2
O VAPOUR POUT
• MX SAMPLING LOCATION
© INJECTION PORT
• COM WCLL
o >o*
sc»ic i
GROUNOWATEft t.
OIL RECOVERY
SYSTEMS. WC
-------
STORE
NUTRIENT INJECTION SYSTEM
— (XJ— THREE- WAT VALVE
SOLENOID VALVE
Ji
Tc.1
l L—'
I |l f^l CIHM MiWIIIHl UN KlltRIW MSTIHS I
L " ™^ I ««« \K 14 HlllllHMkK
-------
v:
o
o
MW-IO
UMXMIIOUia TANK
mm ISLAMOS
I t<
, 1..1I-
'Mw-e
V
MW-9
• y
^
— t
MW-7
l l
ij
i
INTERCEPTOR
TRENCH ._
J ' 1
mtRKING
ENCLOSUHt -^ 71
MW-« W i
EHHAUST EROM 1
DASEMENT VENTILATION j
II S5
WMW"S STORE
AREA'O"
1
CW-4 j
©
MW 12 ;
• CW-3 I
0 1
FIGURE 11
TREATMENT FOCUS AREAS
PROJECT
LOCATION-
DRAWING NO = 941 -ISO -8167-3
MONITORING DATE'
A RECOVERY WELL
• MONITORING WELL
'• ' '• UNDERGROUND TANKS
WELL lOENTiriCATION
f> M* 1 -*ELL ID
- WATER I««LEElEV(mFEET)
© CORE WELL
0 «0
SCALE W FEET
DATE AM 2O-C
MW-4 • ARW-I
I— -f '"""
^ 1
AREA-C"
MW-ISl
• f
^ AREA'B"
RW-Z"! MW-<6 f
\^T^]^^
INTERCEPTOR TRENCH
IRECHARCE CALIERT no.5)
_, PUN'S ISLANDS
- RCCMAMC CAILERT NO t
Ir^ JERVCt I CARME
LJ STATION .
* AREAVMTi"^ ~H'^«"^!H" |
u^tt-2.^- r' 1 '
r«" — 1 t'ttf-l^
s '
«*CCMM>OC CALICMV NO t -
-------
co
o
E
v.
3
c
UJ
m
lo'-,
io9-
I04-
IOS-
Figure: 15 BACTERIA COUNTS
1 — 1
A
I
—
B
\
—
C
!
n
A) CHANGE IN HYDROCARBON UTILIZING BACTERIA COUNTS
DURING BIORECLAMATION - STATION SIDE AREA A8B
o I0 1
z
N
^ 10'-
o IO4-
CD
OC
O
g io5-
o
X
V
c V/
B
A
\
\
$
\
1
7-
y,
\
\
u*
A
B
C
B) CHANGE IN HYDROCARBON UTILIZING BACTERIA COUNTS
DURING BIORECLAMATION-STORE SIDE AREA D
BACKGROUND WELLS
TREATED WELLS
(Nutrient* 8 oxygen added)
AERATED WATER
AERATED WATER + NUTRIENTS
H202+ NUTRIENTS
e
T CANADA
GROUNUWATCR TECHNOLOGY INC
-------
zoi;
§
s
o
a
8
5-
o
S SH
SH
8
§0
w
TOTAL DISSOLVED HYDROCARBON (PPB)
- BEGIN BIORECLAMATION-
-• STOP BIORECLAMATION —
m
m
«•
•i
>
o
TI
CD
O
O
>
-------
(FREE PHASE)
I2O
Figure 17 AREAHBM IMPACT OF BIORECLAMATION
CO
o
CO
AREA'S'
DEPTH TO
WATER
30 32 34 36
MONTHS OF OPERATION
LA CONPACNie G. T. CANADA
GROUNOWATF.R TECHNOLOGY INC
-------
111
60-
Figure 18 AREA IMPACT OF BIORECLAMATION
50-
00
o
12 * 16 18 20 22 24
MONTHS OF OPERATION
2»
90
8.0
LA COMPACNIC C.T CANADA
GROUNUWATF.R Tr.aiNot.ocv INC
-------
26OO-I
Figure 19 AREA"D" IMPACT OF BIORECLAMATION
2OOO-
1400 •
co
o
tn
eoo-
MONTHS OF OPERATION
96
LA COMPAGNII: G. T. CANADA
GROUNIJWATKR TECHNOLOGY
-------
Co
\
\
m«>i«^«»* »•»
BASEMENT LEVEL
HOT SPOT ARE*
(tfiMolvwt <2n>m)
IhrMhold)
<200ppb DISSOLVED HYDROCARBON
FIGURE 20
POST CLOSURE STATUS
PROJECT
LOCATONt
DRAWING NO.- 13-8167-2
O VAPOUR PONT
• AIM SAMPLING LOCATION
& INJECTION PORT
• CORE WELL
0 10'
SCALC IN rCCT
D«1C MAY tMf-
GROUNtMMTER 6
ON. RECOVERY
SYSTEMS, me
-------
GEODUR PAPER
FOR
SECOND FORUM ON INNOVATIVE
HAZARDOUS WASTE
TREATMENT TECHNOLOGIES:
DOMESTIC AND INTERNATIONAL
UNITED STATES
ENVIRONMENTAL
PROTECTION
AGENCY
May 15-17, 1990
Philadelphia, Pennsylvania
USA
307
-------
Lecture by Svend Mortensen GEODUR A/S
and introduction to The GEODUR System
THE HISTORY OF GEODUR:
GEODUR is a product made of chemicals in the molecular stabilization technology. The product
is the result of concentrated scientific research and practical experiments.
The chemical process as a result from using GEODUR makes it possible to stabilize loose organic
matters and inorganic materials to a product which is very much alike concrete.
These loose organic and inorganic materials are e.g. mould, desert sand, fly-ash and mud, as
well as slag from combustion
plants and power stations, paper, waste, etc.
WHAT IS GEODUR:
The name "GEODUR" gradually has become several different things. The following list explains
the word:
The GEODUR Concentrate: A product (The GEODUR Product) produced of chemicals.
GEODUR A/S: The company which is producing and marketing the GEODUR
Concentrate.
The GEODUR Solution: A compound of the GEODUR Concentrate and water mixed in
the ratio 1:100 (1 kg of concentrate for 100 Itr. of water).
The GEODUR Process: The chemical process, which makes stabilization possible.
The GEODUR System: The method of stabilizing which binds heavy metals in soil, fly-
ash or other waste products, and the method of producing soil
concrete.
The GEODUR Soil
Concrete: A type of concrete made of soil, cement and the GEODUR Solution.
The GEODUR Treatment: The treatment itself by means of the GEODUR Process
MORE ABOUT GEODUR:
When producing GEODUR Soil Concrete almost all kind of available local materials are being
used as aggregates instead of the ordinarily known aggregates. These available local aggregates
possibly can be the soil excavated e.g. the top soil such as mould, clay, sand etc.
The GEODUR Solution and a small amount of cement (typically 4-8%) are to be mixed into
the soil.
The quite new aspect of making soil concrete is the fact that the GEODUR Process makes it
possible to bind the organic matters
which are in almost all kind of soils and in unsorted and unwashed types of sand.
308
-------
It has always been impossible to produce a type of concrete of ordinary soil or of polluting waste
products.
The GEODUR Process makes these things possible.
oo 00 oo
The GEODUR Process makes it possible to bind heavy metals in different materials, so they
cannot be set free again. Typically it is heavy metals in polluted soil, fly-ash etc.
Various tests and analysis have been made, e.g. an analysis of
leaching has shown that the leaching of heavy metals is so poor that the heavy metals cannot
be measured in the samples stabilized by GEODUR.
The following kinds of soils and solid wastes have until now been solidified:
Garden topsoil, clay and clayey humus, sand and sandy humus, silt, desert sand (salty),
coast sand (salty), laterite, volcanic lava, mud (sediments with blue clay), municipal sludge
(heavy metals), hazardous chemicals and oil contaminated
soils, fly ashes, stabilized and raw sulphite powder.
Among the test results the following should be mentioned:
FLY ASH
Variable Content in untreated Content leached
sample out of sample
mg/l mg/l
Arsenic
Barium
Cadmiumb
Chromium
Copper
Mercury
Lead
Zinc
(As)
(Ba)
(Cd)
(Cr)
(Cu)
(Hg)
(Pb)
(Zn)
48
449
<5
56
135
0.11
145
249
<0.02
<1
<0.02
0.04
<0.02
<0.001
<0.02
<0.00
Apart from the teachability tests of contaminants also the chemical and ecotoxicological properties
of the leachate are studied together with the mechanical strength and durability of the solidified
materials. Also petrographic analysis have been carried out.
It should be mentioned that a special ecotoxicological fouling test is carried out on different
solidified materials exposed in-situ in receiving waters.
The GEODUR Concentrate is a non-poisonous product (see certificate from the Danish Toxicology
Centre). Therefore it does not harm neither our environment nor the persons being using
GEODUR.
309
-------
FIELDS OF APPLICATION:
1. GEODUR and the Environment.
GEODUR can be used to improve our environment, both
a) preventively for infrastructure purposes in the fight against further pollution,
and
b) in the treatment of soil already polluted.
oo OO oo
a) GEODUR together with e.g. fly-ash is being used for infrastructure purposes,
e.g. for road construction. The fly-ash being stabilized by mixing it into cement
and the GEODUR Solution is applied on to the ground section excavated
as the stabilizing base under asphalt. (See Road Construction Project Report).
b) Polluted soil is being stabilized by mixing it into the GEODUR Solution and
cement (only a few %). Then the GEODUR Process makes the soil "concrete
like" to be casted in any form required easy to be kept on site, or to be re-
used e.g. as embankments, coast protection, sound banks, roads, etc.
Furthermore GEODUR is being tested to solve storage problems of nuclear waste
materials to prevent these materials from being spread by dust formation.
2. GEODUR and Construction.
As mentioned it is possible to make soil concrete by means of the GEODUR Process
by replacing the ordinary aggregates by the top soil of the site.
Thus the purchase and the transportation costs of the aggregates are being eliminated.
By applying the very special concrete "ANTI-FIRE" on to the GEODUR Soil Concrete
the strength is being increased and the concrete appears as ordinary concrete.
GEODUR Soil Concrete is Composed of:
Soil from the site.
GEODUR Concentrate.
Portland Cement.
Water.
310
-------
How to Produce GEODUR Soil Concrete on the Site:
1-2% GEODUR solution (Calculated as percentage of the
soilweight).
Dry mixing the soil.
Dry mixing the cement required (typically 4-8% of the soilweight
which is far less than the amount of cement when producing ordinary
concrete)
After repeated mixing, water is to be added until the compound
reaches a proper consistency for compressing (roads) or for casting
(building materials).
The GEODUR Process activates chemically all the natural binding properties of the soil
by changing the surface tension. This process of agglomeration produces an ultimate
stabilizing effect on the soil with high density and strength.
The PH-factor of the soil does not have any influence on the final strength of the
GEODUR Concrete, because it is chemically controlled by the GEODUR Solution.
Different Types of Soil Can be Used for GEODUR Soil Concrete:
Black vegetable mould (garden top soil).
Clay and clayey mould.
Sand and sandy mould.
Mud.
Desert sand (salty).
Coast sand (salty).
Volcanic lava.
TESTS AND ANALYSISES:
Following tests and analysises have been carried out:
1. Ecotoxicological Fouling Test in Sea Wate by:
The Danish Water Quality Institute VKI
2. Leaching Test by:
Steins Laboratory, Denmark and
RIS0 The Danish Atomic Research Center
3. Petrographical Analysis by:
The Danish Technogical Institute
4. Mechanical Strength and Durability Test by:
ELKRAFT, The Danish Power Station Association.
5. Full Scale Test of Road Construction by:
ELKRAFT, The Danish Power Station Association.
6. Permeability Test by:
Geological Institute, Denmark
7. Certificate by:
Danish Toxicology Centre
311
-------
TEST IN SEA WATER
The field tests are designed on the basis of the results of a preliminary investigation in 1989. With
the objective to get an indication of the stability and leaching properties of different GEODUR stabilized
waste materials and concrete, test blocks were exposed in the sea during a period of four months.
Colonization of fouling organisms has been followed with regular intervals, and the accumulation of
selected trace elements in the biota was examined at the end of the exposure period.
MATERIALS AND METHODS
Five GEODUR stabilized materials have been exposed in the sea together with blocks of commercially
manufactured concrete as a reference material (table 1).
Table 1 Materials exposed in the sea
NO. I CODE I DESCRIPTION
1
2
3
A
5
6
106
98
136
U9
110
Fly ash with stones i
i
Polluted soil from Frederikssund i
Sewage sludge from Clostrup Purification Plant
i
Incinerator ash from Refshalec
i
Unpolluted soil from Canlcse !
Concrete (reference)
Test blocks sized 10 cm x 10 cm x 10 cm were delivered by GEODUR. At the end of July six blocks
of each type were suspended one metre below the surface on a raft in the Innerbroad of the Isefjord,
Zealand. The Isefjord is a moderately eutrophicated area with a salinity around 20 o/oo.
After 5, 10, 12, and 16 weeks of exposure one or two blocks of each material were collected and
photographed. The species composition and the biomass (60°C constant weight) were described.
At the end of the experiment samples of the dominating fouling organisms (hydroids) were collected
from material Nos. 3,4, and 6, freeze-dried and analysed for cadmium and lead. Samples have been
digested in a teflon bomb system using quartz destined nitric acid. Analysis has been performed by
heated graphite furnace AAS. Background correction and standard addition technique was used.
The blocks were weighed before the exposure in the sea and after 2 months at room temperature.
RESULTS AND DISCUSSION
Physical Stability
The loss of weight during the exposure in the sea was approximately 1% in concrete (table 2). The
greatest loss of weight was found in incinerator ash with values between 9-16%. This is 3-4 times
higher than the other stabilized materials, where the mean toss of weight varied between 3.5-4.1%.
There was no relationship between the loss of weight and the length of the period of exposure.
312
-------
Table 2. - Loss of weight In percentage after exposure in the sea.
NO.
1
2
3
4
5
6
MATERIALS
riy «h
Polluted foil
Sewage sludge
Incinerator ash
Unpolluted coll
Concrete
PERIOD OF EXPOSURE
5 weeks
5.6-5.7
4.2-4.S
3.9-4.6
12.1-16.0
2.5-2.8
0.6-1.6
10 week*
3.2-4.2*
0-3.4
2.8-3.7
5.2-15.4
2.5-2.6
0.9-0.9
12 weeks
7.7
4.2
11.7
4.4
0.6
16 weeke
5.2
4.2
9.2
6.3
1.1
* drifted ashore
Colonization, Species Composition and Biomass
After 5 weeks of exposure there was only a slight 'hairy' fouling on the blocks and quite a numerous
fauna of free living crustaceans, especially amphipods. After 10 weeks the blocks were covered by
a dense growth of hydroids (Laomedea longissima) (figure 1.).
A further development of the fouling community was seen after 16 weeks (table 3).
Table 3. - Species composition of the fouling community after 16 weeks of exposure.
COELENTERATES
CRUSTACEANS
Microdeutopus gryllotalpa
Corophium Insidiosun
Idotea baltica
Dexanine spinosa
Phtisica marina
Ostracods
POLYCKAETES
Nereis sp.
Harmothoe sp.
Blue mussels
Snails
Sea squirts
Number of species
MATERIAL NO.
2
444
44
44
4
444
4
4
9
3
+ 44
4-4
4
4
+ +
*
e
4
444
44
+
4
4
4
4
9
5
+ 44
+
4
4
4
+
e
6
+++
++
++
+
++
+
+
9
< 5 individuals
5-25 individuals
> 25 Individuals
During the period of exposure there was a slight increase in the richness of the species, and the same
species and the same number of species developed on the different sorts of stabilized materials and
on concrete.
The biomass of the fouling organisms (totally dominated by hydroids) increased rapidly after 10 weeks
of exposure (figure 2 and table 4).
The biomass developed in the same way on blocks of concrete, sewage sludge, and unpolluted soil.
A higher biomass was found in incinerator ash, and until 12 weeks also on blocks of polluted soil.
313
-------
However, during the last four weeks of the experiment there was a staanation in the biomass on the
blocks of polluted soil.
BIOMASS
.. 4
16
Figure 2 - Fouling biomass on test blocks after 10,12, and 16 weeks of exposure in the sea.
Table 4 - Fouling biomass on test blocks after 10,12, and 16 weeks of exposure in the sea.
NO.
2
3
4
5
6
MATERIAL
Polluted soil
Sewage sludge
Incinerator «»h
Unpolluted soil
Concrete
10 WEEKS
g DW
6.72
2.36
3.86
2.30
2.65
*
253
89
145
87
100
12 WEEKS
g DW
8.87
3.59
9.43
5.17
3.27
*
271
110
288
158
100
16 WEEKS
g DW
8.66
11.18
13.58
10.64
10.84
%
BO
103
125
98
100
Trace Elements
The concentration of cadmium and lead is the same in hydroids grown on blocks of sewage sludge
as on blocks of concrete and the level is almost the same as in blue mussels from unpolluted areas
(table 5). In contrast, slightly elevated levels of cadmium and very high values of lead are measured
in hydroids grown on blocks of incinerator ash. Compared to sewage sludge, which has a higher content
of lead, the stabilization and binding of lead in incinerator ash has been unsuccesfull. The blocks of
incinerator ash were porous and had the greatest loss of weight (table 1).
In spite of the different physical and chemical properties the highest biomass was found on blocks
of incinerator ash (table 4).
314
-------
Table 5 - The concentration of cadmium and lead in hydroids grown on stabilized test blocks
of sewage sludge, incinerator ash and concrete and the concentrations in unstabllized
material. All values in mg/kg DW.
NO.
3
4
6
MATERIAL
Sewage sludge
Incinerator aah
Concrete
HYDROIDS
Cadmium Lead
0.32
0.46
0.35
3.1
79
2.7
UNSTABILIZED
MATERIAL
Cadmium Lead
51
70
6410
2500
CONCLUSIONS
Within an exposure period of four months in the sea test blocks of concrete had a loss of weight around
1%. The mean loss of weight of fly ash, polluted soil, sewage sludge and unpolluted soil varied between
3.5 - 4.1%. The loss of weight in incinerator ash varied between 9-16% with a mean value of 12%.
During the period of exposure a fouling community, quantitatively dominated by hydrojds (Laomedea
longissima) developed on all materials, and there was no differences in species composition and species
richness. Compared to biomass on blocks of concrete, the biomass was the same on blocks of sewage
sludge and unpolluted soil and higher on blocks of incinerator ash and polluted soil (up to 12 weeks).
The concentrations of cadmium and lead is the same in hydroids grown on blocks of sewage sludge
and concrete. The level in hydroids is comparable to blue mussels from unpolluted areas.
In hydroids grown on blocks of incinerator ash the concentration of cadmium is almost the same in
hydroids grown on concrete, but the concentration of lead is about 30 times higher.
It is therefore recommended to improve the stabilization of incinerator ash. Renewed fouling tests and
physical and chemical laboratory tests are planned in 1990 on incinerator ash and soil polluted by
trace elements.
315
-------
THE SITE PROGRAM:
HEAVY METAL FIXATION IN SOIL
Prepared by:
Philip N. Baldwin, Jr.
Chemfix Environmental Services
2424 Edenborn Avenue, Suite 620
Metairie, LA 70001
CHEMFIX TECHNOLOGIES, INC. (CTI) 1s a publicly owned company based in the
New Orleans, Louisiana area with historical roots going back to CHEMFIX, INC.
founded in Pennsylvania in the early 1970's.
CTI specializes in the stabilization of hazardous, industrial and municipal
wastes. CTI operates fixed facilities on the West, Gulf and East Coasts as well
as a couple to be operational in the near future in the Midwest region of the
country. CTI also has an active licensee in the far East. In addition to
stabilization activities, CTI owns two EPA contract analytical laboratory
companies: One in Louisiana (E.I.R.A., Inc.) and one in California (BTC
Associates).
CTI has historically based its stabilization business on the patented
CHEMFIX* Process which involves the use of calcium based setting agents and
liquid anionic silicates. In the late 1980's, the basic CHEMFIX® Process evolved,
when treating hazardous wastes, into the CHEMSET~/CHEMFIX* Process. This
modification has been brought about by the requirements of the landban
regulations and Best Demonstrated Available Technology (BOAT) treatment
standards. The CHEMSET~ Process modifications to the historical CHEMFIX* Process
were also brought about by the need to treat wastes at a higher solids and higher
316
-------
mobile analytes content than had been typical in the 1970's and early to middle
1980's. These CHEMSET™ modifications are many and they add substantially to the
depth of stabilization capability established by the CHEMFIX* patented Process.
The addition of CHEMSET1" molecular bonding enhancement reagents to the
CHEMFIX* Process allows improved interactions with analyte anions and cations as
well as non-ionic organic molecules. The results are decreased molecular and
physical diffusivities of the analytes of concern. Fixation is affected by
several types of chemical bonds: (1) Ionic, (2) Co-valent and (3) Electro-
static.
MOLECULAR BONDING ENHANCEMENT
Types of Bondings
Ionic Na:Cl Sodium chloride (Table salt)
Co-valent Si::0 Silicon Dioxide (sand)
0*
Electrostatic H-0 Mater and their hydrates
The traditional CHEMFIX* Process, as it has been applied to industrial metal
contaminated wastes, was geared toward the reduction in the mobility of cation
analytes. The CHEMSET™ addition to the CHEMFIX* Process substantially broadens
the capability of the fixation technology.
In 1988 when CTI and the EPA agreed to a joint participation in the 1989
SITE PROGRAM, CTI was looking for a site that could utilize some of the newest
expressions of stabilization technology. The plan was to tackle a very high
level of a heavy metal in a high solids matrix. It should be noted that since
317
-------
late 1988, the concept of CHEMSET"1 has moved forward substantially and though
the chemistries used at Clakamus, Oregon in March 1989 were successful overall,
they would not be repeated today on similar wastes without using the updated
refinements. Chemical fixation has become like the microchip business: Good
today, better tomorrow. Regardless of the more advanced stage of the technology
today in 1990 vs. late 1988, the results of the CTI/EPA 1989 SITE PROGRAM stand
on their own as impressive indication that the family of CHEMFIX Process
technologies are viable when applied to very high concentrations of heavy metal
in a high sol Ids waste matrix.
The high solids processing equipment taken to the Clakamus site has
demonstrated an ability not only to successfully treat high solids, but
demonstrated the capacity to treat an average of 400 tons of contaminated soil
per one shift day. Peak production rates exceeded 600 tons per day during a late
1989 commercial job.
During the actual SITE program, four individual areas designated as A, C,
E, and F by the EPA, were to be treated. Only a total of 40 cubic yards were
treated in the field, which posed an unexpected control problem at such a low
volume output. The Pre-Production laboratory data for the key analytes is shown
below.
318
-------
PRE-PRODUCTION TOTALS ANALYSIS
RAW WASTE
TOTALS
mg/kg
mean
AREAS: AC E F
LEAD: 21,000 117,000 88,000 47,000
COPPER: 18,000 8,500 110,000 48,000
PRE-PRODUCTION TCLP ANALYSIS
AREAS:
LEAD:
COPPER:
RAW WASTE
mg/1
mean
A
650
78
C
590
23
E
680
4.5
F
220
110
NATURFIL*
mg/1
mean
A C E F
0.12 0.003 0.23 0.03
1.2 1.8 0.60 0.93
319
-------
*»****»*»**************«**»»*******************************»************«*»>»*
PRODUCTION TOTALS ANALYSIS
RAW WASTE
AREAS:
TOTALS
mg/kg
mean
LEAD: 21,000 140,000 92.000 11,000
COPPER: 18,000 18,000 74,000 33,000
******************************************************************************
PRODUCTION TCLP ANALYSIS
AREAS:
LEAD:
RAW WASTE
mg/1
mean
A C E F
610 880 740 390
COPPER: 45 12 81 120
NATURFIL*
mg/1
minimum/mean
A C E F
.05/.05 .05/2.5 .05/47 .05/.10
.3S/.57 .43/.S4 .16/.65 .49/.60
% Deviation of Liquid -8.0% -8.2% -23.0% -17.4%
****************************************************************
320
-------
This data Indicates that a very high metal content in a high solids waste,
both total and mobile, can be handled by the CHEMSET~/CHEMFIX* technology.
MURPHY, however, came along with us on our Clakamus project. A malfunction
occurred in our computer control panel and for a series of reasons we had to
complete the brief demonstration with only semi-automatic control conditions.
This situation led to a greater than a +/- 5% reagent variation and as a result
we were not able to completely reproduce all of the Pre-Product ion treatment
results. In spite of this deficiency, the mean TCLP treated waste values for
Area E still represented a 94% reduction in lead mobility and a 99% reduction
for Copper. The results from Area C represented a 99% TCLP mobility reduction
for lead and 95% for Copper. Less contaminated Areas A & F reached well over
99% in mobility reduction for both metals.
What was re-learned from the problems and successes CTI had at the Clakamus
site, is that the chemistries designed into the fixation technologies by CTI
are not random but are waste specific. Any particular chemistry design can
address a wide variety of cations, anions, mixtures of each and more recently
non-ionic organic compounds, but the chemistries must be monitored and mixed
properly to get the consistent and dramatic TCLP mobility reductions that we have
seen are possible. There is a great deal yet to learn but the CHEMSETYCHEMFIX*
Processes are an excellent format for unlocking waste stabilization problems.
CTI also continues to work on improved chemical delivery and mixing systems.
In closing I would like to make a few final overview comments about our
SITE PROGRAM experience.
The PC8 levels at the SITE were too low to make any hard decisions
and hence I have left any of that discussion out of this
presentation.
321
-------
The leachate tests applied by the EPA to the stabilized products were
the TCLP, Multiple Extraction Procedure (MEP), ANS 16.1 and the Batch
Extraction Test. Through field dellstlngs for AMOCO, GENERAL MOTORS
and US STEEL, CTI has a track record of achieving excellent MEP
results. The MEP data produced from the Clakamus site was only
performed on location C. Although the chemical deliveries were only
slightly off of prediction, site C had the highest mean lead
concentration- nearly 14% lead. These two factors together produced
erratic test results.
The resultant product of the CHEMFIX* and/or CHEMFIX*/CHEMSET™
processes (NATURFIL*) is a clay like material, not a monolith. The
American Nuclear Society 16.1 extraction test was mistakenly applied
to the NATURFIL* produced at the SITE. This test is to be applied to
monoliths, not to clay like materials. Any material that might
physically sluff off of a clay like substance would be picked up as
leached product thereby improperly slanting the results. In spite
of this problem, the NATURFIL* still produced ANS 16.1 test
results that EXCEEDED the NCR criteria by 6 orders of magnitude.
CTI was pleased to participate in the SITE PROGRAM, but with the experience
of having participated, we have made several suggestions to the EPA that might
help future programs:
1. There should be better co-ordination of the types of and difficulties
of wastes treated during a SITE PROGRAM YEAR. CTI treated soils with
up to 14% lead and 11% copper with combined TCLP mobilities of over
322
-------
600 ppm while most other contractors dealt with mobile metal
concentration less than 50 with totals less than 1.0%. We chose this
challenge but in so doing made any real comparisons between
technologies more difficult for the public.
2. Closer oversight is needed by the EPA R&D department regarding the
type of tests that are applied to the treated wastes.
3. More interface is needed between the EPA and the treatment firms
regarding the amount of waste that should be treated as that relates
to the equipment available from the treatment firm.
4. Treatment is both engineering capability, equipment and chemistry.
The EPA only evaluated chemistry AFTER the fact. This shortchanges
the concept of waste treatment.
CTI believes that the days of stabilization as an art are passing quickly
and that the days of stabilization as a science are upon us. CTI is pleased to
be apart of that science!
323
-------
Evaluation of Three
Leading Technologies
For Second Forum on Innovative Hazardous
Waste Treatment Technologies
By
Armand A. Balasco
Case Study on Innovative Technology
Development for Rocky Mountain Arsenal
Contracting Agency: U.S. Army
Toxic and Hazardous Materials
Agency (USATHAMA)
Performance Period: July 1086-1988
ArthirD Little
-------
DOD Hazardous Waste Sites on
Super-fund National Priority List
ArtlurD Little
Rocky Mountain Arsenal
(RMA) profile
• Located northeast of Denver, Colorado
• Occupies more than 17,000 acres
• Began operations in 1942
ArthirD Little
325
-------
Rocky Mountain Arsenal
(RMA) profile
• On-site operations (1942-1982):
- Chemical/incendiary manufacture
- Chemical munitions demilitarization
- Insecticide/pesticide manufacture
ArthirD Little
General Map
of RMA
North Boundvy
IrondaJe
Ran CtaMifkaUon
and
Warehouse Are*-
StopMonM'l
AJrport
ArthirD Little
N
I
326
-------
RMA Basin F profile
• Asphalt-lined evaporation basin
• Period in use: 1956-1982
• Average depth: 10 feet
• Surface area: 90 acres
• Capacity: 243 million gallons
ArtfurD Little
RMA Basin F profile
• Current amount of hazardous material:
- Soils (> 400,000 cubic yards)
- Liquids (4 million gallons)
• Major contaminants:
- Organics
- Inorganics
- Metals
AtthirD Little
327
-------
RMA Basin F profile
Basin F material
contaminant
Aldrin
Isodrin
Endrin
Sulfoxide
Sulfone
Mercury
Arsenic
Contaminant
cone, (ppm)
100-1,100
50-1,300
65-180
130-300
300-700
<1
20
ArthirD Little
Situation at RMA
• Industrial operations generated variety
of wastewaters
• Wastewaters stored in on-site lagoons
• Lagoons pose threat to environment
• Environmentally acceptable and reasonable-cost
clean-up needed
ArthirD Little
328
-------
RMA case study objectives
• Review treatment technology databases
• Evaluate/rank most applicable technologies
• Bench-scale test most promising technologies
• Develop full-scale designs/cost estimates
• Select technology for on-site (RMA) pilot-
scale testing
ArthirD Little
Technologies to be evaluated/ranked
• Encapsulation
• Fluidized/circulating bed combustion
• Classification
• In-situ glassification
• Soil washing
• Solvent extraction
• Wet-air oxidation
ArthirD Little
329
-------
Evaluation criteria employed
(Level of importance)
Highest
Safety
Treatment of organ ics
Throughput rate
Vendor test facilities
- System safety
- Permitting
- Availability
Average
Treatment of metals
Treatment of inorganics
Capital cost
Operating cost
Reliability
Proprietary status
Vendor test facilities
- Emissions monitoring
Lowest
Associated equipment
Equipment complexity
Maintenance
Ease of operation
Barriers
ArthirD Little
Relative evaluation of leading
technologies for Basin F
Technology
Rotary kiln
(base case)
Encapsulation
Fluidized/circutating
bed combustion
Glassification
In-situ vitrification
Soil washing
Engineering
design/ Capital Operating System
status cost cost complexity
Perceived effectiveness
Organics Inorganics
ArthirD Little
330
-------
Technologies selected for laboratory/
bench-scale testing
• Fluidized/circulating bed combustion
• Glassification
• Soil washing
ArthirD Little
Fluidized/circulating bed combustion
• Vendor test facility permitting problems
• USATHAMA schedule constraints
• Elimination of technology from
further consideration
ArthirD Little
331
-------
Classification
; Basin F
Material
• Glass >•
Product
.Scrubber
Effluent
ArthirD Little
Cleaned
GLASSFICATION BENCH-SCALE
TEST SYSTEM
OH-Gas
Samping Port Pitot
Tube
Jouto-Heated
Ceramic-Uned MeRer
Stack
Blower
Scrubber
Sampbtg Port Samping Port
^Continuous
Analyzers
ArthirD Little
332
-------
Glassification equipment/
test conditions
• Glass melter volume: 8.2 liters
• Glass melter operating temperature: 2,200°F
• Glass melter operating pressure: -3 to -5 in.
ArthirD Little
Classification equipment/
test conditions
Glass melter off-gas treatment system:
- Venturi scrubber
- Heat exchanger
- Demister
- HEPA filter
- Activated carbon bed
ArthirD Little
333
-------
Classification equipment/
test conditions
• Glass melter feed system:
- Submerged drop tube (Basin F material)
- Above glass melt surface (glass formers)
• Basin F/glass formers
- Basin F material
- Sodium carbonate
- Sodium borate
ArthirD Little
Classification equipment/
test conditions
• Basin F feed rate: 3.3 Ibs/hr
• Glass formers feed rate: 0.5 Ibs every
15 minutes
• Glass draw-off rate: 9.9 Ibs every 2 hrs
ArthirD Little
334
-------
Classification equipment/
test conditions
• Gas sampling performed:
- Modified Method 5 (6.2 hrs)
- Method 101A (2.2 hrs)
• Continuous gas analyzer:
- CO - SO2
- C02 - NOx
-O2
ArthirD Little
Classification test results
Advantages
• Organochlorine compounds: > 99.99% reduction
• Organophosphorus compounds: > 99.76%
reduction
• Organosulfur compounds: > 99.60% reduction
ArthirD Little
335
-------
Classification test results
Advantages
(continued)
• TCLP leachate from glass product
- No target contaminants detected
- Endrin, arsenic, and mercury levels below
regulatory limits
JirthirD Little
Classification test results
Disadvantages
• Volatile metals (e.g., mercury, arsenic)
in off-gas
• Acid gases (e.g., HCI, H2SO4, H3PO4)
in off-gas
• High dust carryout in off-gas
ArthirD Little
336
-------
Air
Emissions
Soil Washing
ArthirD Little
SOIL WASHING LABORATORY-SCALE
TEST SYSTEM
Add
Add
and
FbccUant Rocotant
•I Organic
Decant
6th Organic Froth
Decant Slurry
Decant 1st Add 2nd Add
Wash Wash
3rd Add Washed
Wash Material
ArthirD Little
337
-------
Soil washing equipment/
test conditions
Organic washing (five stages):
- 700 grams of basin F material
- Toluene/kerosene/octanol (TKO) addition
- Heating (120°F)
- Mixing
- Organic layer decanting
- Slurry removal to flotation stage
flrthirD Little
Soil washing equipment/
test conditions
Froth flotation (1,000 gram cell):
- Slurry from organic washing
- Reagents (caustic, sodium silicate
and surfactant) addition
- Heating (120°F)
- Aeration
- Mixing
A rthirD Little
338
-------
Soil washing equipment/
test conditions
• Froth flotation (continued)
- Froth removal
- Flocculant addition
- Clean slurry setting
- Water decanting
- Settled solids removed to acid washing
ArthirD Little
Soil washing equipment/
test conditions
Acid washing (three stages):
- Settled solids from froth flotation
- Acid (HCI) addition
- Clean solids settling
- Water decanting
- Clean solids filtering
- Filtrate removal
- Clean solids removal for analysis
ArthirD Little
339
-------
Soil washing test results
Advantages
• Organochlorine compounds: >98.9% reduction
• Organosulfur compounds: > 78% reduction
• Metals: < 55% reduction
ArtlurD Little
Soil washing test results
Advantages
(continued)
• TCLP leachate from washed material
- Some target contaminants detected
- Endrin, arsenic, and mercury levels below
regulatory limits
• Acid washing may not be required
ArtlurD Little
340
-------
Soil washing test results
Disadvantages
• Organic extractants required
• Minimum three stages organic washing required
• Necessity for stream recycle
ArthirD Little
Soil washing test results
Disadvantages
(continued)
• Non-pesticide organic contaminants do not
follow pesticides in spent organic extractant
- Carbon adsorption required
• Technology requires many operational steps
ArthirD Little
341
-------
Results - RMA
• Fluidized bed combustion, glassification, and
soil washing selected for testing
• Fluidized bed combustion not tested; permits
could not be obtained
• Glassification and soil washing both
technically attractive
• Glassification selected for pilot testing
at RMA
ArthirD Little
342
-------
Soil Vapor Extraction and Treatment of VOCs
At a Superfund Site in Michigan
Joseph P. Danko, Michael J. McCann, and William O. Byers
CH2M HILL
Corvallis, Oregon
Abstract
The Verona Well Field supplies potable water to the City
of Battle Creek, Michigan, three townships, and another
small city. The combined service area equates to
approximately 50,000 people. In 1981, the well field and
surrounding area were found to be contaminated with
chlorinated solvents and other volatile organic compounds
(VOCs). The contaminant plume extended throughout an
area of approximately 1 mile by one-half mile. Two facilities
operated by a local solvent wholesaler/distributor were
identified as the primary sources of the contamination.
VOC concentrations as high as 1,000 parts per million
(ppm) were found in the groundwater and soil at the
wholesaler's primary facility.
EPA selected enhanced volatilization of VOCs using soil
vapor extraction (SVE) to clean up the contaminated soils
at the primary contaminant source area. A pilot SVE system
was installed in late 1987 and operated for about 69 hours,
removing approximately 3,000 pounds of VOCs from the
vadose zone. Using the results of the pilot phase, a full-
scale SVE system was designed and installed. It began
operation in March 1988 and has since removed more than
40,000pounds of VOCs.
From system startup until January 1990, extracted vapors
containing VOCs were passed through vapor-phase activated
carbon for treatment prior to discharge. In January 1990,
the carbon system was replaced by a catalytic oxidation
system capable ofonsite destruction of VOCs. The catalytic
oxidation system is expected to provide more continuous
operation and a cost savings over the life of the project.
Introduction
Soil vapor extraction (SVE) is rapidly becoming a
common and preferred treatment method for removing
volatile organic compounds (VOCs) from contaminated
soil. This paper presents the case history of an operating
SVE system at the primary contamination source area
in the Verona Well Field Superfund site in Battle Creek,
Michigan. The paper includes a site description, a
description of the SVE system and the associated offgas
treatment system, and a summary of SVE and offgas
treatment performance to date.
Background
The Verona Well Field (Figure 1) provides potable
water to approximately 50,000 residents in and around
Battle Creek, Michigan. In August 1981, VOC
contamination was detected in approximately one-
third of the city's production wells and in numerous
private wells outside the well field.
Initial investigations of the site by the U.S. EPA
Technical Assistance Team identified three primary
potential source areas and a contaminant plume with
VOC concentrations of up to 346 micrograms per liter
(Hg/l) in the well field. These three areas are also
identified in Figure 1.
FIGURE 1
VICINITY MAP
A solvent distribution facility run by the Thomas
Solvent Company was found to be the most
contaminated of the source areas. This facility, referred
to as Thomas Solvent Raymond Road (TSRR), was used
for the storage, transfer, and packaging of chlorinated
and nonchlorinated solvents from the mid-1960s to
1984. Contamination of the soil and groundwater at
TSRR is thought to have resulted from underground
tank leakage and surface spills in the tank truck loading/
unloading area and the warehouse (now demolished).
Leak tests showed nine of 21 tanks on the site to be
leaking (Figure 2). Vadose zone contamination covered
more than an acre.
Remedial Measures
In May 1984, EPA signed a Record of Decision
(ROD) to implement an Initial Remedial Measure (IRM)
at the Verona Well Field site in order to protect the
Ob-
343
-------
city's water supply. The IRM included converting a
number of the contaminated production wells to
blocking wells to prevent further migration of the
contaminant plume. An air stripping system was
installed to treat the extracted groundwater. Three
additional production wells were also installed as part
of the IRM to supplement the city's existing water
supply system.
A OROUNONOEREXnvieiKINWaL
FIGURE 2
LOCATION OF
UNDERGROUND TANKS
In the fall of 1985, EPA signed a second ROD that
addressed the TSRR facility. The ROD specified a
corrective action that included a network of
groundwater extraction wells to remove contaminated
groundwater, the treatment of extracted groundwater
by air stripping, and an SVE system to remove VOCs
from the unsaturated zone.
Groundwater Extraction
The installation of nine groundwater extraction
wells at the TSRR facility began in the fall of 1986. The
groundwater extraction system started up in March
1987. Since that time, more than 11,000 pounds of
VOCs have been extracted with the groundwater.
Groundwater concentrations initially were as high as
19,000 Jig/l total VOCs; as of October 1989, the
concentrations had dropped to about 1,500 ng/l.
Soil Vapor Extraction
Contract documents for the construction, operation,
and maintenance of an SVE system at the TSRR facility
were issued for bid in June 1987. The contract called
for the following performance objectives:
• None of the soil samples analyzed could
have VOC concentrations above 10
mg/kg of dry soil.
• No more than 15 percent of the soil
samples analyzed could have VOC
concentrations above 1 mg/kg of dry soil.
The contract also included the removal of the 21
underground storage tanks. The underground storage
tanks were surrounded by heavily contaminated soil,
and direct excavation and removal would significantly
violate the State of Michigan's air quality criteria. This
problem has been avoided by using the SVE system to
remove the majority of VOCs before removing the
underground tanks.
The contract was awarded to Terra Vac, Inc. of San
Juan, Puerto Rico in October 1987.
Site Characteristics
The lithology at the site consists of fine- to coarse-
grained sand with trace day, silt, and pebbles. The
water table is approximately 25 feet belowgrade and
the hydraulic gradient is to the northwest.
In addition to the vadose zone contamination, there
was also a floating nonaqueous phase liquid (NAPL)
layer in the vidnity of Extraction Well 8, whose location
is shown in Figure 2. This well was installed as a
product recovery well, combining groundwater
extraction with intermittent removal of the NAPL
product as it accumulated.
Soil Vapor Extraction (SVE) System
The SVE system includes vapor extraction wells, an
air/water separator, offgas treatment equipment, and a
blower. Figure 3 presents a simplified schematic of the
system. The system was installed following a
preconstruction investigation designed to help place
SVE wells and to determine the pre-operation
magnitude of contamination.
The extraction network consists of 23 2- and 4-
inch-diameter PVC wells with slotted screens from
approximately 5 feet belowgrade to 3 feet below the
water table. The wells are packed with silica sand,
sealed at the screen/casing interface with bentonite,
and grouted to the existing grade to prevent short
circuiting. The wells are connected by a surface
collection manifold. Each well head has a throttling
valve, sample port, and vacuum pressure gauge. The
surface manifold is connected to a centrifugal air/
water separator which is in turn connected to a blower
and offgas treatment system.
Offgas Treatment Systems
The original SVE system at the TSRR site used a
carbon adsorption system to remove contaminants
from the vapor stream. That system was replaced in
January 1990 with a catalytic oxidation unit for offgas
treatment (both systems are described below). Following
344
-------
TYnCM.MI.WOfl
emuenoHWEu.
FIGURES
SCHEMATIC OF SOIL VAPOR
EXTRACTION SYSTEM
offgas treatment, the air is discharged to the atmosphere
through a 30-foot stack.
Carbon Adsorption System
The carbon adsorption system consisted of two sets
of four stainless steel carbon vessels connected in series.
A schematic of the system is presented in Figure 4.
The first (primary) set of vessels was used for the
majority of VOC adsorption, while the second set
acted as backup in the event of contaminant
breakthrough in the primary set. The system was
installed under negative pressure to prevent the leaking
of VOCs. Each vessel held 1,000 pounds of vapor-
phase carbon. The vessels were connected to header
piping with flexible hoses and quick disconnect
couplings. Sample ports, vacuum pressure gauges, and
temperature probes were installed upstream, between,
and downstream of the two sets of vessels. A carbon
monoxide meter was also installed between the sets
of carbon vessels for early detection of combustion in
the primary carbon units.
During system operation, an in-line organic vapor
analyzer (HNu) between the primary and secondary
carbon units was used to monitor contaminant
breakthrough. When breakthrough occurred, the
primary carbon unit was changed out and the
secondary carbon was placed In the primary position.
VACUUM EXTRACTION UNIT
PRESSURE INDICATOR
TEMPERATURE INDICATOR
(S) SAMPLING PORT
Q FLOWMETER
FIGURE 4
SCHEMATIC OF CARBON
ADSORPTION SYSTEM
345
-------
This minimized the chances of breakthrough of the
secondary system.
Catalytic Oxidation System
The catalytic oxidation system, designed to destroy
contaminants onsite, is essentially a thermal reactor
used for the destruction of chlorinated and
nonchlorinated VOCs. A schematic of the system is
presented in Figure 5. The SVE air stream flows from
the positive displacement blower through a heat
exchanger, where it is preheated to approximately
430°F. The preheated gas then enters the burner, where
it is heated to approximately 800°F by a natural gas
burner. The air stream is then passed through a catalyst
bed, which causes VOC oxidation. The catalyst, without
being altered, accelerates the oxidation reaction by
adsorbing the reactants (oxygen and VOC) on
catalytically active sites1. This greatly accelerates the
reaction rate converting chlorinated VOCs to carbon
dioxide, water vapor, and hydrochloric acid gas. The
catalyst also allows the oxidation reaction to occur at
much lower temperatures than would conventional
thermal incineration.
The gas stream exits the reactor at approximately
820°F and enters the shell side of the heat exchanger.
In the heat exchanger, the heat produced from the
natural gas burner and from combustion of VOCs in
the catalyst bed is used to heat the inlet gas and reduce
the fuel requirements of the preheat burner. In the
heat exchanger, exhaust gas is cooled to about 550°F
before being vented to the atmosphere through a 12-
inch-diameter stack.
Because the catalytic oxidation system uses natural
gas as fuel and is designed for remote operation, the
following automatic shut-down controls have been
included:
• Cutoffs for low and high temperatures
(750°F and 1,300°F)
• Ultraviolet flame sensor cutoff (if flame
goes out)
• Blower default cutoff (if the SVE blower
fails)
The catalytic oxidation system was selected to
replace the vapor-phase carbon system at the TSRR site
because:
• The catalytic oxidation system does not
require the intermittent downtimes needed
by the carbon adsorption system. Site
cleanup could therefore proceed faster.
• VOCs are destroyed onsite with the
catalytic oxidation system rather than
transferred onto carbon and hauled several
hundred miles for thermal destruction. The
National Contingency Plan favors onsite
destruction of wastes.
• Use of the catalytic oxidation system is
expected to result in an overall cost savings
for the project based on the estimated
remaining mass of contaminants.
Performance of the SVE and
Offgas Treatment Systems
Pilot Phase SVE
Pilot-phase operation of the SVE system was initiated
in November 1987. Figure 6 shows the location of the
SVE wells and the piping layout at the TSRR site. Prior
to complete system operation, individual wells were
operated to determine their radius of influence, flow
rate, and initial extraction rate. All gas stream analyses
were done using an onsite gas chromatograph.
PARTKULATE
ALTER
EXHAUST TO
ATMOSPHERE
HE AT EXCHANGER ^- BURNER
CATALYST BED
EXHAUST FROM SOIL
VAPOR EXTRACTION
SYSTEM
FIGURES
SCHEMATIC OF CATALYTIC OXIDATION
SYSTEM WITH HEAT RECOVERY
•Hardlson, LC. and Ej. Dowd. August 1977. Emission Control Via Fluidized Bed Oxidation. Central Engineeriny Progress.
Radian Corp. July 1984. Control of Industrial VOC Emissions by Catalytic Incineration. Prepared for the Industrial Environmental Research Lab.
346
-------
Total VOC concentrations ranged from a low of 2
mg/1 at VE-6 to 204 mg/1 at VE-2. The radius of influence
of individual wells was measured by recording the
vacuum in nearby vapor extraction wells and in
vacuum piezometers. A 1-1/4-inch water vacuum was
recorded 60 feet from an extraction well. Vacuum
piezometers located between tanks to investigate tank
shielding showed at least a 2-inch water vacuum.
Pilot-phase operation continued intermittently over
IS days with a total operation time of 69 hours.
Approximately 3,000 pounds of contaminants were
removed during the pilot test, as determined by gas
stream and carbon loading analyses. The average VOC
stack cc centration was 0.0666 mg/1. At an average
flow rate of 500 cubic feet per minute (cfm),
approximately 4.6 pounds of VOCs were released
during the pilot phase, indicating an approximate
removal efficiency of 99.8 percent.
A
EW2
A
BM
eo
aoLwpon EXTRACTION (SKI w^i.
A QROUNDWATER EXTRACTION WCU.
FIGURE 6
SOIL VAPOR EXTRACTION SYSTEM AT
THOMAS SOLVENT RAYMOND ROAD SITE
Full-Scale SVE
The full-scale SVE system began operation in March
1988 with 23 vapor extraction wells. Operation has
generally been limited to under 14 wells at any one
time, with well head concentrations determining the
wells selected for operation. This is done to maximize
contaminant loading to the offgas system. The typical
flow rate from individual wells is 100 scfm, with the
combined flow between 1,400 and 1,600 scfm.
Wellhead pressure in the extraction wells is generally 2
to 3 inches of mercury.
As of May 1,1990, approximately 43,000 pounds of
VOCs have been removed by the system in 252 days of
operation. The total VOC loading rate has dropped
from an initial high of approximately 1,080 pounds of
VOCs per day to under 100 pounds per day. The NAPL
layer originally observed under the site has not been
present since October 1988.
Offgas Treatment Systems
Carbon Adsorption
Between November 1987 and January 1990, VOC
contaminants were removed from the vapor stream
using activated carbon adsorption. During that time,
approximately 266,400 pounds of carbon were loaded
and shipped offsite for regeneration. The average carbon
loading rate during this period was 16 percent.
Catalytic Oxidation
Since January 1990, VOCs have been destroyed
onsite using the catalytic oxidation system. The
catalytic oxidation system was performance tested
before being put into operation. Process operating
data from the performance tests are shown in Table 1.
The primary difference between the four runs was
reactor operating temperature, which was varied to
determine the effect of operating temperature on
destruction removal efficiencies.
Table 2 presents the results from discharge stack air
sampling during the performance tests. All
contaminants except tetrachloroethene were
significantly less than the allowable state discharge
limits at each of the operating temperatures.
Tetrachloroethene at 780°F had an average discharge
concentration of 204 ppb, approximately 60 percent
of the discharge limit. At 860°F, the stack concentration
of tetrachloroethene decreased to 35 ppb, 10 percent
of the allowable limit. The highest detected
concentration of hydrochloric acid, a byproduct of the
oxidation, was 21 ppm, well below the discharge limit
of 132 ppm.
Since the performance test, the catalytic oxidation
system has operated at 800°F without any operating
difficulties.
Acknowledgements
The soil vapor extraction system is part of a U.S.
EPA Superfund Operable Unit Remedial Action under
Contract No. 68-01-7251. This paper has not been
subjected to the Agency's peer and administrative
review. It therefore does not necessarily reflect the
views of the Agency, and no official endorsement
should be inferred. Similarly, any use of specific names
in the paper should not be viewed as an endorsement.
The authors would like to thank Margaret Guerriero
of the U.S. EPA and Terra Vac, Inc., the SVE contractor,
for their cooperation in making this paper possible.
347
-------
Table 1
Catalytic Oxidation System
Performance Testing Operation Data
Parameter
Temperatures
Heat Exchanger (in)
Heat Exchanger (out)
Catalyst (in)
Catalyst (out)
Stack
SVE Process Gas
Pressure drop (flow rate)
Temperature
Pressure
Flow rate
Natural Gas
Flow rate
Mass rate
Miscellaneous Data
Stack velocity
Stack flow rate
NG valve position
DA valve position
Units
oF
°F
op
oF
oF
in. Hp
op
in.Hg
scfm
scfm
pph
fpm
scfm
%
%
Run A
(1/9/90)
98
430
780
792
468
3.4
41
4.2
1,579
13.4
37.0
3,740
1,636
43
13
RunB
(1/10/90)
93
446
820
838
482
3.4
38
4.4
1,577
_
—
3,290
1,626
46.7
13
RunC
(1/10/90)
95
470
860
875
510
3.4
37
4.3
1,582
— —
—
3,321
1,418
49.5
13
RunD
(1/30/90)
91
431
800
822
—
3.4
39
4.5
1,572
13.7
37.8
—
—
—
—
Note: — means not measured.
Table 2
Discharge Stack Air Sampling
Catalytic Oxidation System Performance Tests
Component
Chloromethane1
Tetrachloroethene1
Tetrachloroethene
Trichloroethene
Carbon Tetrachloride
1,1,1-trlchloroethane
Chloroform
Methylene Chloride
Benzene
Vinyl Chloride
Acetone2
2-Butanone2
Toluene
Hydrogen Chloride2
IlniK
Ulllld
ppbv
ppbv
ppbv
ppbv
pptv
ppbv
pptv
ppbv
ppbv
ppbv
ppbv
ppbv
ppbv
ppmv
MDNR
Limits
• -•••HT"*j
_4
347
347
1,337
249,000
_«
161,000
11,464
1,752
6,220
_4
_4
_4
132
Run A
(1/9/90)
Reactor
Temp.-780°F
A_1
~i
41
202
178
<0.04
352
0.08
249
58.5
2.4
<0.04
NA
NA
1.3
--
A .2
mi
36
221
193
1.7
35
0.04
60
17.6
<0.1
<0.04
NA
NA
0.46
11.8
A-3
31
188
180
1.6
59
0.03
198
18.0
<0.1
<0.04
S5.7
5.1
1.5
~"
RunB
(1/10/90)
Reactor
Temp.«820°F
R 1
0*1.
29
87
83
0.9
2
<0.02
8J
6.1
<0.1
<0.04
NA
NA
<0.1
"*
R-2
D~m*
30
85
83
0.8
<2
<0.02
<20
5.0
<0.1
<0.04
NA
NA
<0.1
20.8
R3
mjtj
28
88
91
0.9
22
<0.02
46
10.1
<0.1
<0.04
31.6
2.5
<0.1
™"
RunC
(1/10/9Q)
Reactor
Temp.
-860°F
C-l
^* M.
22
35
34
0.3
<2
-------
SLUDGE AND SOIL
TREATABILITY STUDIES AT A
LARGE, COMPLEX SUPERFUND SITE
BOFORS-NOBEL SUPERFUND SITE
MUSKEGON, MICHIGAN
BY:
Susan Roberts Shultz*
Wendy Oresik
Doug Graham
Larry Kiener
Sarah Levin
John Trynoski
David Shultz
DONOHUE & ASSOCIATES, INC.
PRESENTED AT
EPA SECOND FORUM ON INNOVATIVE
HAZARDOUS WASTE
TREATMENT TECHNOLOGIES: DOMESTIC AND INTERNATIONAL
349
-------
1.0 INTRODUCTION
1.1 INTRODUCTION
1.1.1 Site Description
The Bofors-Nobel (Bofors) Superfund Site is located six miles east of downtown
Muskegon in Egelston Township, Muskegon County, Michigan. This site is
currently under the management of the Michigan Department of Natural Resources
(MDNR), Grand Rapids District, in cooperation with the U.S. Environmental
Protection Agency, Region V (EPA). This 85-acre parcel of land is irregularly
shaped and includes an operating specialty chemical production facility, a
biological/carbon (PACT™) treatment plant, an unused landfill, 10 abandoned
sludge lagoons used for disposal of wastes from the production facility, and
associated soils. The southern portion of the site is bounded by Big Black
Creek. Figure 1 presents a map showing these and other important site
features.
1.1.2 Site History
Many investigations and studies of this site have been performed during the
last 20 years. A description of the site history is briefly summarized below.
Lakeway Chemicals began producing industrial chemicals at the site in 1960.
Throughout the 1960s and early 1970s, the State of Michigan placed various
restrictions on wastewater disposal from the site because of surface water and
potential groundwater contamination from wastewater discharge to Big Black
Creek. In 1976, wastewater from the plant was accepted at the Muskegon County
Wastewater Treatment Plant, and purge wells were installed at the site to
extract contaminated groundwater. Bofors Industries, Inc., merged with
Lakeway in 1977 and with Nobel Industries in 1981. In 1985, Bofors-Nobel
(Bofors) filed for bankruptcy for a variety of reasons, including reported
environmental expenditures in excess of $60 million. As a result of legal
action in bankruptcy court, Bofors was allowed to sell the operating chemical
plant to Lomac, Inc. (Lomac). As part of the sale, agreements were reached
between Lomac, MDNR, and EPA that Lomac could not be liable for cleanup of
contamination existing prior to the sale of the plant area property. These
agreements allowed Lomac to operate the plant independently of previous site
activities. The site was then nominated for the National Priorities List
(NPL) and was placed on the NPL in March, 1989.
1.1.3 Site Operations
The development and operation of the industrial chemical facility and
associated disposal activities at the Bofors Site was investigated during the
RI by examining aerial photographs of the area dating from 1955 to 1987,
interviewing past workers and owners of the facility, and reviewing past
investigations and studies of the facility and adjoining property.
350
-------
SOILS AROUND
LAGOONS
BOFORS
SITE BOUNDARY
NUMBERS:SOURCE AREAS
Figure I
Site Map
351
-------
Before 1970, the plant produced alcohol-base detergents, saccharin, and dye
intermediates such as 3,3'-dichlorobenzidine (DCB), benzidine, and azobenzene.
Raw materials included fatty alcohols, fatty ether alcohol, sulfur dioxide,
aqua ammonia, muriatic acid, sulfuric acid, nitrobenzene, o-nitrobenzene,
methanol, benzene, sodium chloride, and zinc. During the 1970s, Lakeway pro-
duced a lauryl alcohol base detergent, dye intermediates, pesticides, and
herbicides. The lagoons were used for plant wastewater and sludge disposal
until 1976.
Lagoon 1, the iron pond, was used for the disposal of iron sludge or iron
scale from the production of DCB. Lagoon 2, the sodium formate pond, was a
synthetically lined lagoon and was abandoned after attempts to line the lagoon
were unsuccessful. Lagoon 3 was used as a settling pond for DCB waste, zinc
oxide waste, and waste from cleanup operations and spillage. Lagoon 4 was
used for the disposal of non-contact cooling water. Lagoons 5, 6, 7, 8,
and 10 received calcium sulfate sludge pumped to the lagoons as a slurry.
Lagoon 9 was used for discharge of calcium sulfate liquid and detergent waste.
As indicated by plant personnel, Lagoon 9 berm failures discharged sludge
southeast to the Big Black Creek floodplain in late 1974 and early 1975. Berm
failure from Lagoon 10 also occurred into the floodplain in early 1975. The
sludge from the berm failures was removed from the floodplain and disposed of
in the southern portion of Lagoon 9.
1.1.4 Treatabilitv Studies
Identification and screening of remedial technologies and process options
during the Feasibility Study for the Bofors Site resulted in three innovative
process options determined to be potentially effective, implementable, and
cost-effective for the remediation of the contaminated sludges and soils at
the Bofors Site: soil washing, low temperature thermal desorption (LTTD), and
solidification/stabilization. Because of the innovative nature of the tech-
nologies, particularly as applied to the unique, heterogeneous nature of the
wastes at the site, treatability studies were performed to evaluate the
effectiveness of these process options and to obtain additional data necessary
to evaluate remedial alternatives during the detailed analysis phase of the
Feasibility Study (FS). The data generated from the treatability studies was
used in the FS to identify materials handling issues and operating parameters,
reduce cost estimate and performance uncertainties, and evaluate post-
treatment requirements. Treatability and analytical testing was performed
during the bench scale treatability studies by vendors and laboratories desig-
nated by Donohue & Associates, Inc., (as part of a joint venture with
Goldberg, Zoino, & Associates) in accordance with the Treatability Study Work
Plan (May, 1989), and QAPP/SAP Addendum (GZA/Donohue, May 1989) to the
approved Work Plan for the Bofors Site (GZA/Donohue March 1988). The treat-
ability studies were completed between May 1989 and March 1990. Delays in
receiving analytical laboratory data and vender reports caused the extended
time frame. Bench scale treatability studies were performed to evaluate the
general ability of the technologies to treat the wastes at the site. Pilot
scale treatability studies are beyond the scope of this RI/FS.
352
-------
The lagoon sludges and underlying soils were initially characterized from
samples collected in 1988 during Phase I RI activities. Samples were analyzed
for volatile organic compounds (VOCs), metals, cyanide, semivolatile organic
compounds, polychlorinated biphenyls (PCBs), pesticides, and special semi-
volatile organic compound analytes. No PCBs or pesticides were detected above
the analytical detection limit. Compounds of concern for the contaminated
soil and sludge of the L.O.U. were selected based on the results of the base-
line risk assessment and are listed below:
' Aniline
0 Azobenzene
0 Benzene
* Benzidine
0 3,3'-Dichlorobenzidine (DCB)
Samples from Lagoons 3, 8, 9, and 10 were selected for the bench scale treat-
ability studies. Phase I RI data indicate Lagoon 3 contains the highest
detected concentrations of organic compounds in the soil beneath the lagoons,
and Lagoon 9 contains the highest concentrations of organic compounds in the
sludge, possibly representing the worst case soil and sludge. Lagoons 8
and 10 comprise the largest volume of contaminated sludge and soil in the
L.O.U. and therefore may represent average concentrations of contaminants for
both soil and sludge. Approximately 62 percent of the lagoon sludge are con-
tained in Lagoons 8 and 10.
Sludge and soil samples were collected for chemical and geotechnical charac-
terization. Samples were provided to vendors performing the treatability
studies. Four samples were collected from the four lagoons to represent
various conditions, as follows:
1. Lagoon 3 composite soil sample to represent the worst-case
contaminated soil.
2. Lagoon 9 composite sludge sample to represent the worst-case
contaminated sludge.
3. Lagoons 8 and 10 composite sludge sample to represent the average
contaminated sludge.
4. Lagoons 8 and 10 composite soil sample to represent the average
contaminated soil.
Request for proposals were issued to nine vendors capable of performing treat-
ability studies for soil washing, LTTD, and solidification/stabilization.
Three vendors were selected and contracted to perform the treatability studies
based upon their qualifications. BioTrol, Inc. (BioTrol), of Chaska,
Minnesota, performed the soil washing treatability study on the worst case and
average soil samples. Chemical Waste Management, Inc. (CWM), of Riverdale,
Illinois, performed the LTTD treatability study on the worst case soil, worst
case sludge, and the average sludge samples. Enreco, Inc. (Enreco), of
Amarillo, Texas, performed the solidification/stabilization treatability study
on the worst case soil, worst case sludge, and average sludge samples.
353
-------
Results from the treatability studies were used to address the following
questions:
1. Are the treatment processes able to achieve adequate performances for
the compounds of concern for the lagoon soil and sludge tested?
2. Do any residuals require further treatment prior to disposal?
Treatability experimental procedures, experimental results, and conclusions
for each process option are discussed in the following sections. Experimental
results from the treatability studies are evaluated according to performance
and the identification and characterization of residual waste streams.
Performance was evaluated either by reviewing attainment of cleanup standards
(soil washing and LTTD), or by reviewing potential reductions in contaminant
mobility (solidification/stabilization). Attainment of cleanup standards was
addressed by comparing the concentrations of the compounds of concern in the
treated material with 10"^ and 10"" risk-based cleanup standards developed in
the baseline risk assessment. Because the treated material will be covered or
capped after replacement, groundwater ingestion is the most significant
exposure route. Therefore, cleanup standards used in these studies are based
on remediation of sludges and soils to prevent contamination of the ground-
water. Potential reductions in contaminant mobility are evaluated by com-
paring TCLP extract concentrations in the prestabilized and stabilized
samples.
Data tables showing detected compound concentrations are summarized in this
chapter. Analytical holding times were exceeded by the laboratory on some
samples; these samples are noted in the tables. Some samples were not
analyzed due to extreme exceedances on analytical holding times by the labora-
tory. Because of these suspect or non-existent data, CWM-generated pre-treat-
ment data are sometimes used for the other treatability studies. In addition,
because of the suspect data, the highest concentrations of each post-treatment
compound detected is used where multiple analyses of the same type of sample
have been performed.
Quantitation limits are often high for the Bofors sludges and soils due to
interferences in the samples. Therefore, the risk-based cleanup standards are
sometimes below the quantitation limit on the treated samples; thus, the per-
formance of the process options cannot always be confirmed for each compound
of concern. In these cases, the results of the TCLP for the treated sample
are evaluated to gain additional information on the potential concentrations
in the treated samples. Since the TCLP data are generated from an extraction
procedure, the concentration in the actual treated sample is most likely sig-
nificantly greater than in the TCLP extract. It is assumed in this study that
if the TCLP extract concentration of a compound of concern is above the
quantitation limit, then the actual concentration in the treated sample
exceeds the cleanup standard. This is a conservative assumption; however,
since no information on the relationship between TCLP and actual concen-
trations is known, this assumption may be justified. If the compound is not
present in the TCLP extract above the quantitation limit, then removal of that
354
-------
compound is assumed to be adequate to achieve the cleanup standard. In some
cases, the TCLP quantitation limit may also be higher than the cleanup stan-
dard. In these cases, the limits of currently available analytical techniques
have been reached. Since the analytical techniques used in these studies are
the best available accepted technology, it is assumed that if TCLP concen-
trations are below the quantitation limits, the technology has reached an
acceptable degree of cleanup.
1.2 SOIL WASHING
The soil washing process extracts contaminants from soil using water or an
aqueous solution which may be composed of chelating agents, surfactants,
acids, or bases. Chelating agents are organic compounds in which atoms form
coordinate bonds with metals. Surfactants are soluble compounds that reduce
the surface tension of liquids or the interfacial tension between two liquids
or between a liquid and a soil. These compounds help create phase changes and
disperse contaminants to facilitate contaminant removal. The primary function
of this technology is to reduce the volume of contaminated fine silt, clay,
and colloidal fractions from coarse sand and gravel components. Fine silts
and clays absorb organic contaminants due to the surface chemistry and high
surface area of the natural mineral and organic material and, therefore, the
majority of contamination is associated with these size fractions. Using an
intensive scrubbing action, soil aggregates are broken up and dispersed,
freeing highly contaminated fine particles from the coarser sand and gravel.
The surfaces of the coarser particles are cleaned by abrasive processes.
A typical soil washing removal treatment process is a multi-step process which
begins with excavating and screening the contaminated soil to remove oversized
material and debris. The contaminated soil is then slurried with water and
screened at 14 mesh, or approximately 0.05 inches in diameter. Material
larger than 14 mesh can be washed further if necessary. Material smaller than
14 mesh enters a froth flotation circuit where hydrophobic components
(compounds with a low water solubility) are removed. The flotation underflow
enters an intensive, multi-stage, countercurrent scrubbing system. Fine
particles leave the process and undergo residuals treatment or disposal. The
final washed product is dewatered and may be replaced on the site. The con-
taminated water from the scrubbing system enters a treatment system and can be
recycled to the soil washing system. Soil washing technology was evaluated in
the U.S. EPA Superfund Innovative Technology Evaluation (SITE) Program.
1.2.1 Treatability Experimental Procedures
As discussed previously, chemical analytical results from soil and sludge
samples collected during the Phase I RI were used to select samples for the
Phase II treatability study. Lagoon 3 soil (worst case), and Lagoon 8 and 10
soil (average) samples were sent to BioTrol. Sludge samples were not tested
in the soil washing treatability study because soil washing is not a suitable
process option for sludge remediation.
355
-------
The soil washing treatability study involved the following processes:
o Sample preparation
o Soil scrubbing
o Soil leaching
o Gravity separation
o Flotation
Each of these processes can be used as individual processes for specific
applications or can be included as parts of a treatment process train. These
processes have all been used in the hydrometallurgical industry and are being
combined here to form a soil washing remediation technology designed to remove
contamination from soil. Each process is described below.
1.2.1.1 Sample Preparation
Samples were provided by in one- to five-gallon containers. These samples
were reduced in size using a riffle splitter. This device became coated with
a heavy grease-like substance (tar) when used once to produce the smaller
samples. This device was not used again, and smaller samples were obtained
using a trowel.
1.2.1.2 Soil Scrubbing
In this process each soil sample was washed multiple times with the following
scrubber solutions:
Water
* Orvus K liquid (Proctor & Gamble Corporation)
0 Citrikleen (Penetone Corporation)
* Industrial Tide (Proctor & Gamble Corporation)
Triton X-100 (Rohm and Haas Corporation)
The additives are emulsifiers used to disperse hydrophobia compounds that are
not soluble in water.
Samples of Lagoon 3 soil and Lagoon 8 and 10 soil were preweighed
(1.0 kilogram) and combined with water and additive to produce a 50 percent
solids slurry in the scrubbing unit. The slurries were thoroughly agitated
with an impeller for 5 minutes and then diluted to 2 liters and allowed to
settle. The supernatant was decanted and stored for future chemical analysis.
A 50 percent slurry was made with the remaining solids. The procedure was
then repeated. After completing four or five washing stages, the solids were
removed from the washing unit, dewatered, and chemically analyzed. The
supernatant from each washing stage was composited and submitted for analysis.
1.2.1.3 Soil Leaching
To develop an understanding of how the Bofors soil samples react in an aqueous
environment, a series of 72-hour leaching tests were conducted. Flasks were
prepared separately for Lagoon 3 soil and Lagoon 8 and 10 soil samples.
356
-------
Flasks containing 20 to 50 percent solids for each soil type were placed on a
shaker table which continuously mixed the samples for 72 hours. The samples
were then filtered and air-dried.
1.2.1.4 Gravity Separation
Gravity separation was used to separate materials in an aqueous environment
using dense fluid. Contaminants are less dense than the soils, therefore the
contaminants should float as the "clean" soil sinks. Calcium chloride was
used to separate contaminants from Lagoon soil samples because of its availa-
bility and non-solvent properties.
1.2.1.5 Flotation
In a flotation cell, water containing the fine particles and chemicals is
brought in contact with very small air bubbles. Chemical additives are used
to separate particulate solids by causing one group to float. Differences in
surface chemical properties of the particles allow separation since some par-
ticles are entirely wetted by water, and others are not. The bubbles attach
to the physically- and chemically-altered surfaces of the particles and rise
to the top of the flotation cell where they are skimmed. This procedure was
tested with washed soil to evaluate the potential of a flotation process to
act as a final polishing step in the remediation of the soil.
1.2.2 Experimental Results
1.2.2.1 Laeoon 3 Soil
Lagoon 3 soil was sampled to represent worst-case contaminated soil in the
L.O.U. Although soil characteristics indicated that soil washing would be an
excellent technology, the soil contained a heavy, grease-like substance (or
tar) which, when slurried with water and agitated, coated metallic equipment
surfaces in contact with the slurry.
Leaching tests on Lagoon 3 soil indicated that low intensity agitation could
be used to agglomerate the tar. During leaching, the tar formed spherical tar
balls or tar granules that were generally larger than the sand particles in
the soil. The spherical tar balls appeared when the solids content was
greater than 30 percent in the testing. Following leaching, this material was
removed by screening. Analysis showed that the leached soil was similar to
the soil subjected only to high-intensity attrition scrubbing. The concen-
tration of contaminants in the tar was extremely high, indicating that small
amounts of tar in clean soil would render the soil "contaminated." Approxi-
mately 80 percent of the DCB were contained in less than 10 percent of the
soil.
Leached soil was then subjected to high-intensity agitation, which sig-
nificantly reduced the concentration of contaminants. By combining a tar-
removal step with soil washing, the Lagoon 3 soil was cleaned to approximately
the same level as Lagoons 8 and 10 soil.
357
-------
Washed soil consisted of predominantly sand and small particles of tar. The
tar was much less dense than sand. Therefore, gravity separation using
saturated solutions of calcium chloride was used to separate the contaminated
tar particles and improve cleanup of the coarse material. However, this
process did not achieve separation of the small tar particles from soil. A
flotation procedure was also implemented to remove the residual tar from soil.
The froth from this procedure appeared to carry the tar particles out of the
slurry. However, no significant differences resulted in the concentration of
the contaminants in leached, washed soil before and after flotation.
Performance Evaluation
The cleanup standards for the Lagoon 3 soil and the pre-treatment and post-
treatment concentrations for the compounds of concern are listed in Table 1.
Aniline was detected in the pre-treatment sample below the 10"^ but greater
than the 10"^ risk-based cleanup standard. Aniline was not detected in the
post-treatment sample above the quantitation limit. The quantitation limit is
greater than the 10*6 cleanup standard but less than the 10"^ cleanup
standard. However, the TCLP extract concentration is below its quantitation
limit. Therefore, as discussed previously, it is assumed that an acceptable
removal level has been achieved for aniline. Azobenzene and benzene were
detected in the pre-treatment sample at less than the 10*^ and the 10"^ risk-
based cleanup standard; therefore, no removal is needed for these compounds.
Benzidine was detected in the pre-treatment sample above the 10"^ and 10"^
cleanup standards. Benzidine was not detected above the quantitation limit in
the post-treatment sample. The quantitation limit is above the 10"^ and 10'^
standards. However, the TCLP extract concentration is below its quantitation
limit. Therefore, the attainment of cleanup standards is assumed to be
acceptable for benzidine. DCB was detected in the pre-treatment sample above
the 10"^ and 10"^ cleanup standards. The post-treatment DCB concentration
exceeds the 10*^ cleanup standard; therefore, an acceptable removal has not
been achieved for a 10"^ remedy, but appears to be acceptable for a 10"^
remedy.
In summary, soil washing for Lagoon 3 soil appears to be successful in this
test for the removal of aniline, azobenzene, benzene, and benzidine to 10"^
and 10"" cleanup standards. However, soil washing did not effectively remove
DCB to 10~6 levels, but may be appropriate for a 10"^ level remedy.
Process Residuals Evaluation
The treatment residual wastes for soil washing are the fines and the aqueous
phase (wash water). Chemical analyses of the fines and aqueous phase are
presented in Table 2. Pre-treatment compound concentrations are also shown
to indicate compounds originally present in the sample. Based on comparing
pre-treatment soil and post-treatment fines concentrations, it appears that
benzine, azobenzene, and DCB have been concentrated in the fines phase.
Aniline and benzidine are not present in the fines above the quantitation
limit. The five compounds of concern for Lagoon 3 soil are also present in
the aqueous phase in significant concentrations. These results indicate that
soil washing of the sample has caused phase transfer and volume reduction of
contaminants of concern.
358
-------
TABLE 1
SOIL WASHING
ATTAINMENT OF CLEANUP STANDARDS
FOR COMPOUNDS OF CONCERN FOR LAGOON 3 SOIL
Botors Site
Muskegon, Michigan
Cleanup Standards*
Compound
10"4 Risk
(ug/kg)
10* Risk
(ug/kg)
Pre-Treatment*
(Coarse)
(ug/kg)
Post-Treatment*
(Coarse)
(ug/kg)
Ouantitation
Limit
fug/kg)
Post-Treatment
TCLP
(ug/l)
Aniline
Azobenzene
Benzene
Benzidine
3,3'- Dichlorobenzidine
97,000
42,000,000
180,000
110
85,000
970
420,000
1,800
1.1
850
9,200
150,0000
130 J
590,000 D
1,100,0000
16,000 J
3J
28,000 J
20,000
96,000
LEGEND:
D : Value reported from a diluted sample aliqout in order to stay within the linear calibration of GC/MS.
J : Estimated value due to minor OC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
— : Compound was not detected.
Cleanup standards based on remediation for the groundwater ingestion exposure route.
TCLP: Toxicity characteristic leaching procedure.
NA: Not applicable to this sample.
*: Soil concentrations and cleanup standards are on a dry weight basis.
359
-------
TABLE 2
SOIL WASHING
CHEMICAL CHARACTERIZATION OF PROCESS RESIDUALS
FOR LAGOON 3 SOIL
Bofors Site
Muskegon, Michigan
Compound
Volatile Organic Compounds
Acetone
Benzene
2-Butanone
Ethylbenzene
Methylene Chloride
2-Methylpropane
Toluene
Total Unknown Hydrocarbons
Xylene (Total)
Semi-Volatile Organic Compounds
Aniline
Azobenzene
Azoxybenzene
Benzidine
2- Chloraniline
3,3' - Oichlorobenzidlne
Dichlorobenzidine Isomer
Methanone, (3-Chlorophenyl) (4-Chlorophenyl)
Nitrobenzene
Total Alkyl Benzenes
Total Unknowns
Metals
Aluminum
Arsenic
Barium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Nickel
Potassium
Selenium
Sodium
Vanadium
Zinc
Pro-Treatment*
(ug/kg)
130 J
15 BJ
3J
490 BEJ
18,343 J
4J
9,200
150,0000
460,000 J
590,000 D
17,0000
1,100,0000
1,300 J
450,000 J
13,000 EJ
1,080,000 J
829,000 J
185,000
610 B
4,800 B
461,0008
3,000
790 B
2,0006
533,000
6,400
74,9006
3,1006
51,5008
431,0006
5308
324,000
Post-Treatment Residuals
Fine*
(ug/kg)
4,000
22,000
3,800 B
75 B
4,600 BJ
57,000
3,300 BJ
3,200,000
13,000,000
84,000 J
9,700,000
12,000,000
785,000 J
1,4008
6,300 B
7506
1,580,000
14,400
16,600
1,710,000
96,800
383,0006
16,100
2,4006
200,0008
1,400
212,000 B
4,3008
1,920,000
Aqueous
(ug/l)
78 J
11,000
190 J
85 BJ
19,000
2,000
190 J
13,0000
26,000 J
30,000 0
4,300
55,000 DJ
3,900 J
5,500 J
250
7,400 J
156,000 J
6,300
5.2 B
378
2.9 B
141,000
1,130
7.1 B
228
6,470
360 J
38,100
132
753
8,630
72,000
8.3 B
8,540
Legend:
B:
D:
E:
J:
Compound was also detected in the associated laboratory blank.
Value reported from a diluted sample aliquot in order to stay within the linear calibration of GC/MS.
Concentration exceeded the linear range of GC/MS calibration.
Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
Compound was below the quantifiable limit.
Concentrations reported on a dry weight basis.
360
-------
The fines and aqueous phase contained several VOCs (acetone, methylene
chloride, and 2-methylpropane) and metals (cadmium, nickel', and selenium)
which were not detected in the original sample. This may be due to insuf-
ficient equipment decontamination, laboratory contamination, or differences in
compound quantitation limits.
Based on the results shown in the table, the fines and the aqueous phase will
require further treatment prior to disposal and discharge. Metal concen-
trations in the aqueous phase are below literature values for inhibition of an
activated sludge process and should not present a toxicity problem if a
biological process is employed (Wastewater Treatment Plant Design and Treat-
ment, Water Pollution Control Federation, 1977). However, the relatively high
concentrations of calcium and magnesium may result in equipment scaling
problems.
1.2.2.2 Lagoon 8 and 10 Soil
A soil composite from Lagoons 8 and 10 was used in the treatability study to
represent the average contaminated soil at the Bofors site. Lagoon 8 and 10
soil was washed with the following wash solutions:
Water
8 Orvus K liquid and hot water
0 Citrikleen and hot water.
There did not appear to be a significant difference in performance between the
various wash solutions. Leaching tests were also performed on Lagoons 8 and
10 soil. The results of these tests were similar to those presented for
Lagoon 3 soil.
A second series of soil washing was performed with hot water and Citrikleen,
Industrial Tide, and Triton X-100 at 1 percent doses. Following this removal
step, the concentrations of the compounds of concern were below quantitation
limits except for DCB. Both the Triton X-100 and the Industrial Tide formula-
tions were able to destabilize 'the colloidal dispersion of fines and give
clear (but colored) aqueous decants. The Triton X-100 material generated much
less foam than the Tide and was chosen as the additive for other iterations.
Performance Evaluation
The 10"^ and 10"^ risk-based cleanup standards for Lagoon 8 and 10 soil and
the pre-treatment and post-treatment concentrations for the compounds of con-
cern are listed in Table 3. Aniline and benzene were detected in the pre-
treatment sample below 10'^ and 10"6 risk-based cleanup standards; therefore,
no removal is needed for these compounds. Azobenzene was detected in the pre-
treatment sample below the 10"^ cleanup standard but above the 10*° cleanup
standard. Azobenzene was not detected in the post-treatment sample above the
quantitation limit. The quantitation limit is greater than the 10*^ cleanup
standard. As discussed previously, the possible concentration in the treated
sample was evaluated using the TCLP extract concentration. Azobenzene was
present in the post-treatment TCLP extract (3 ppb). Therefore, the actual
361
-------
TABLE 3
SOIL WASHING
ATTAINMENT OF CLEANUP STANDARDS
FOR COMPOUNDS OF CONCERN FOR LAGOON 8 AND 10 SOIL
Bofors Site
Muskegon, Michigan
Cleanup Standards*
Compound
1CT Risk
(ug/kg)
10"6 Risk
(ug/kg)
Pre-Treatment* Post-Treatment* Quantitation Post-Treat-
(Coarse) (Coarse) Limit ment TCLP
(ug/kg) (ug/kg) (ug/kg) (ug/l)
Aniline 360,000 3,600 900
Azobenzene 46,000 460 1,800
Benzene 42,000 420 —
Benzidine 9.2 0.092 7,500
3,3'- Dichlorobenzidine 140 1.4 120,0000
660
660
8
3,200
3J
6,500
LEGEND:
J:
TCLP:
Estimated value due to minor OC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
Compound was not detected.
Toxicity characteristic leaching procedure.
Soil concentrations and cleanup standards are on a dry weight basis.
362
-------
concentration in the treated sample is assumed to be above the cleanup
standard. Benzidine was detected in the pre-treatment sample above the 10~4
and the 10"^ cleanup standards, but was not detected in the post-treatment
sample above the quantitation limit. This value is greater than the 10*4 and
10'° risk-based cleanup standards; however, the TCLP extract concentration is
below its quantitation limit. Therefore, removal is assumed to be acceptable.
DCB detected in the post-treatment sample exceeded the cleanup standards by at
least one order of magnitude for 10'4 risk and three orders of magnitude for
10'6 risk indicating that DCB was not effectively removed. In summary, soil
washing does not appear to be an effective process option for this soil,
specifically for azobenzene and DCB.
Process Residuals Evaluation
The concentration of organic compounds and metals for the pre-treatment sample
and residual waste streams of Lagoon 8 and 10 soil are shown in Table 4. Pre-
treatment compound concentrations are shown to indicate compounds originally
present in the sample. As mentioned previously, only semivolatile organic
compound analyses are available for the fines for Lagoon 8 and 10 soil.
Based on comparison of pre-treatment soil and post-treatment fines concentra-
tions, only DCB appears to have been concentrated slightly in the fines phase.
Azobenzene and benzidine are not present in the fines above the quantitation
limit. DCB is also the only compound of concern detected in the aqueous phase
at a concentration significantly above the quantitation limit. These results
indicate a phase transfer and volume reduction of DCB contaminated material
but it is uncertain whether this has occurred for azobenzene and benzidine.
The results also indicate that the aqueous waste stream will most likely
require treatment for organic compounds and metals prior to discharge.
However, the concentration of calcium and magnesium in the aqueous waste
stream may cause equipment scaling problems and require preventive measures.
1.2.3 Conclusions
Results of the soil washing treatability study indicate that this process was
not effective in cleaning worst-case or average soil in the L.O.U. to risk-
based cleanup standards, specifically in attaining cleanup standards for DCB
and azobenzene. This may be explained by the high affinity of DCB and
azobenzene for soil compared to the other compounds. Soil partition
coefficients (Kd) were developed during the RI for the compounds of concern
using site soil information (GZA/Donohue 1990), and are summarized as follows:
aniline; Kd-0.027 cm3/g
azobenzene; Kd-14.9 cm-yg
benzene; Kd-0.389 cm3/g
benzidine; Kd-0.071 cm-yg
DCB; Kd-7.634 cm3/g
363
-------
TABLE 4
SOIL WASHING
CHEMICAL CHARACTERIZATION OF PROCESS RESIDUALS
FOR LAGOON 8 AND 10 SOIL
Bofors Site
Muskegon, Michigan
Analyte
Volatile Organic Compounds
Acetone
Benzene
Chloroform
Methylene Chloride
Styrene
Toluene
1,1,1 - Trichloroethane
Xylene (Total)
Total Unknowns
Semi-Volatile Organic Compounds
Aniline
Azobenzene
Azoxybenzene
Benzidine
3,3'-Dichlorobenzldine
Dichlorobenzidine Isomer
Methanone, (3-Chlorphenyl)(4-Chlorophenyl)
Phenol, 4-(2,2,3,3-Tetranethylbutyl)
1-Propene, 2-Methyl, Tetramer
1-Tetradecanol
Total Unknown Alkene Derivatives
Total Alkyl Benzenes
Total Unknowns
Metals
Aluminum
Barium
Calcium
Chromium
Copper
Iron
Magnesium
Manganese
Nickel
Potassium
Sodium
Znc
Pre-Treatment*
(ug/kg)
11J
7
49 B
1 BJ
gj
900
1,800
7,500
120,0000
6,500 J
20,000 J
19,700 J
130,600 J
1,150,000
7,8008
7,140,000
6,100
1.400B
1,990,000
892,000 B
24,400
177,0006
103,0008
311,000
Post-Treatment Residuals
Fine*
(ug/kg)
Legend:
B:
D:
J:
NA:
NA
NA
NA
NA
NA
NA
NA
NA
NA
380,000
390,000 J
18,000 J
15,000 J
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Aqueous
(ug/l)
8B
1 J
4J
SB
2J
10
15B
5
30 J
340 J
23,000 DJ
12,000 J
570 J
800 J
2,000 J
340 J
2,290 J
1.280J
20.150J
417
19.3B
213,000
27.8
17.1 B
406
37,000
45.8
27.7 B
6,710
31,100
192
Compound was also detected in the associated laboratory blank.
Value reported from a diluted sample aliquot in order to stay within the linear calibration of GC/MS.
Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
Compound was below the quantifiable limit.
Compound was not analyzed.
Concentrations reported on a dry weight basis.
364
-------
The K(j estimates ,the affinity of a compound to soil; as K^ increases, the
affinity increases. Azobenzene and DCS have the highest K^ of these
compounds. They are higher by one to three orders of magnitude than the other
compounds.
Questions that remain for this technology are effective tar removal before
soil washing and identification of the final polishing step to remove residual
tar from the washed sand. The agglomeration of tar and its subsequent removal
by screening are important issues. The leaching tests were performed for
72 hours and have not been optimized. Time constraints on the original study
precluded more than a cursory assessment of gravity separation and flotation
as final process steps. The effectiveness of gravity separation versus flota-
tion as a final polishing step could be examined in greater detail.
The aqueous phase and fines carry a relatively high level of chemical oxygen
demand (COD) and will require treatment before being reused in the process or
before discharge. Treatment of these residuals, which make up approximately
5 to 10 percent of the feed material, must be addressed.
Further bench scale testing and a pilot study would also be necessary to
further evaluate attainment of cleanup standards. However, this is not
recommended because of the inability of this soil washing technology to
achieve cleanup standards for DCB at the bench scale.
1.3 LOW TEMPERATURE THERMAL DESORPTION
Low temperature thermal desorption (LTTD) removes contaminants from a solid
matrix by applying heat to the material. The heat allows the matrix to
release contaminants by several mechanisms, including direct volatilization
of the organic compounds whose boiling points have been reached, steam
stripping of organic compounds by water within the solid matrix, and evapora-
tion of organic compounds due to their increased vapor pressures at elevated
temperatures.
1.3.1 Treatability Experimental Procedure
The following samples were collected and sent to CWM for LTTD treatability
testing: Lagoon 3 soil (worst-case soil), Lagoon 9 sludge (worst-case
sludge), and Lagoon 8 and 10 composite sludge (average sludge).
1.3.1.1 Sample Preparation
Whenever visual inspection of the feed material indicated that oversize
materials were present, the batch of feed material was passed through a
1/4-inch sieve. The screened feed was then thoroughly mixed with a drum
roller or hand tools.
After the feed was screened and mixed, a composite sample was taken by grab
sampling at several locations throughout the container. A separate cleaned
volatile organic analysis jar was used for the VOC portion of the composite
sample. This portion of the sample was refrigerated before analysis to pre-
365
-------
serve its integrity. The feed sample was then taken from the remaining
material. The resulting removal efficiencies represent the material that was
actually processed, and no credit was taken by the LTTD system for losses of
volatile organic compounds due to feed preparation activities.
1.3.1.2 Test Procedures
The bench scale system delivered feed material to a dryer using a volumetric
screw feeder. Approximately 1,600 to 2,000 grams of material were fed into a
4-inch diameter cylindrical furnace (kiln) for each sample processed. The
kiln was operated at maximum temperature (800 degrees Fahrenheit) and maximum
residence time (100 minutes). These operation parameters were considered
economically viable by CWM. The dried solids were collected in a hopper and
emptied through a valve.
Heated nitrogen gas entered and exited the dryer in a flow parallel to the
solids. This nitrogen gas stream carried water vapor and the volatilized feed
constituents to the spray tower where particulate carry-over solids were
removed and the gas cooled to its saturation temperature (typically from
110 to 150 degrees Fahrenheit). Some contaminants with high boiling points
were condensed in the spray tower. The spray tower scrubbing fluid then
flowed to the phase separator where the solids and immiscible organic com-
pounds were separated from the scrubbing water.
The gas stream continued to the primary condenser where it was cooled to 80 to
120 degrees Fahrenheit. The gas then entered the secondary condenser where it
was further cooled by a mechanical refrigeration unit to a temperature less
than 40 degrees Fahrenheit. Finally, the gas stream was passed through a
carbon adsorption unit before being purged to the atmosphere.
The treated solids were collected after the unit reached steady-state tempera-
ture. After mixing, the test sample was divided into five portions. Four of
these subsamples were analyzed by laboratories selected by GZA/Donohue and the
fifth was analyzed by CWM.
The phase separator, which supplies the spray tower fluid, was filled with
17 liters of tap water at the start of each test run. At the end of the run,
it contained 12 to 16 liters. A representative sample of this liquid was
taken for analysis.
Liquid retained during the steady state portion of the test period was com-
bined at the end of each run. The combined sample from each test run was
mixed and put in appropriate containers for analysis.
1.3.2 Experimental Results
1.3.2.1 Lagoon 3 Soil
Lagoon 3 soil was bench scale tested to evaluate the upper limit of effective-
ness of LTTD for soil, and to identify potential limitations of this treatment
process at the Bofors Site.
366
-------
Performance Evaluation
The 10"^ and 10"^ cleanup standards and the pre- and post-treatment concentra-
tions for the compounds of concern are listed in Table 5. Azobenzene and
benzene were detected in the pre-treatment sample below 10*^ and 10~*> cleanup
standards. Therefore, no removal is needed for these compounds.
Aniline was detected in the pre-treatment sample below 10*^ but greater than
10*6 cleanup standards. Aniline was not detected in the post-treatment sample
above the quantitation limit. The quantitation limit is below the 10"6 clean-
up standard; therefore, an acceptable removal level has been achieved for
aniline, based on these results. Benzidine was detected in the pre-treatment
sample above the 10"^ and 10*^ cleanup standards, but was not detected in the
treated soil above the quantitation limit. The quantitation limit is above
the cleanup standards. However, the benzidine is not detected in the TCLP
extract above the quantitation limit; therefore, it is assumed that an accept-
able removal level has been achieved for benzidine. DCB was also detected in
the pre-treatment sample above the cleanup standards and below the quantita-
tion limit in the treated sample. Since the quantitation limit is below both
cleanup standards, an acceptable removal level has been achieved for DCB,
based on these results for Lagoon 3 soil. In summary, LTTD appears to be
effective in removing the contaminants of concern in worst case soil, Lagoon 3
soil.
Process Residuals Evaluation
Analytical results for post-treatment residual phase separator and condensate
liquids are shown in Table 6. Pre-treatment compound concentrations are shown
to indicate compounds present in the pre-treatment soil. The phase separator
is designed to contain solids and immiscible compounds removed from the con-
taminated soil. The condensate is designed to contain volatile compounds
condensed out of the gas phase. As shown in Table 6, benzene, a VOC, was only
detected in the condensate, not in the phase separator liquid. However,
aniline, a semi-volatile compound, has a higher concentration in the conden-
sate than in the phase separator liquid. Azobenzene, benzidine, and DCB, all
semi-volatile compounds, have greater concentrations in the phase separator
liquid, as expected. Therefore, the compounds of concern generally par-
titioned into their expected phases.
The phase separator and condensate residuals contained several VOCs which were
not detected in the original Lagoon 3 sample. These compounds have several
possible sources. They may be a result of insufficient equipment decontamina-
tion or laboratory cross-contamination. Differences in compound detection
limits from one analysis to another may also have occurred (i.e., the detec-
tion limit for the pretreated material may have been an order of magnitude
higher than the post-treatment detection limit).
Because high concentrations of contaminants were collected in both waste
streams, it is likely that both condensate and phase separator liquid will
require further treatment before disposal or discharge. Calcium and magnesium
may require treatment to prevent equipment scaling problems.
367
-------
TABLE 5
LOW TEMPERATURE THERMAL DESORPTION
ATTAINMENT OF CLEANUP STANDARDS
FOR COMPOUNDS OF CONCERN FOR LAGOON 3 SOIL
Bofors Site
Muskegon, Michigan
Cleanup Standards*
Compound
10"4 Risk
(ug/kg)
10"6 Risk
(ug/kg)
Pre-Treatment*
(ug/kg)
Post-Treatment*
(ug/kg)
Quantitation
Limit
(ug/kg)
Post-Treatment
TCLP
(us/l)
Aniline
Azobenzene
Benzene
Benzidine
3,3'- Dichlorobenzidine
97,000
42,000,000
180,000
110
85,000
970
420,000
1,800
1.1
850
9,200
150,0000
130 J
590,0000
1,100,0000
4,920
7
330
1,600
650
LEGEND:
D : Value reported from a diluted sample aliquot in order to stay within the linear calibration of GC/MS.
J : Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
— : Compound was not detected.
Cleanup standards are based on remediation for the groundwater ingestion exposure route.
Cleanup standards and soil concentrations are on a dry weight basis.
368
-------
TABLE 6
LOW TEMPERATURE THERMAL DESORPTION
CHEMICAL CHARACTERIZATION OF PROCESS RESIDUALS
FOR LAGOON 3 SOIL
Bofors Site
Muskegon, Michigan
Post-Treatment Residuals
Pre-Treatment* Condensate
Analyte
Volatile Organic Compounds
Acetone
Benzene
2-Butanone
2-Hexanone
Toluene
1,1,1 - Trichloroethane
Semi-Volatile Organic Compounds
Aniline
Azobenzene
Azoxybenzene
Benzene, Isocyano
Benzidine
Benzole Acid
2-Chloroaniline
4-Chloroaniline
2-Chlorophenol
3,3'-Dichlorobenz!dine
Methanone, (3-Chlorphenyl)(4-Chlorophenyl)
2-Methylphenol
4-Methylphenol
Nitrobenzene
Phenol
1-Tetradecanol
Total Alkyl Benzenes
Total Unknown Trichlorobiphenyls
Total Unknowns
Metals
Aluminum
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Potassium
Sodium
Vanadium
Zinc
Legend:
B : Compound was also detected in associated
(ug/kg)
_
130 J
15 BJ
—
490 EBJ
—
9,200
150,0000
460,000 J
—
590,000 D
—
170,000 D
820
—
1,100,0000
450,000 J
—
_
13,000 EJ
—
290,000 J
1 ,080,000 J
—
829,000 J
185,000
610 B
4,800 B
—
—
461, 000 B
3,000
2,000 B
533,000
6,400
74,9008
3,100 B
—
—
51,5006
431, 000 B
530 B
324,000
laboratory blank.
D : Value reported from a diluted sample aliquot in order to stay within
(ua/l)
89
30
72
46
65
10
60,000
6,000
3,600 J
760 J
6,700
—
28,000 EJ
390 J
3,100
4,800 J
—
480
—
1,300 J
17,400 E
—
760 J
2,570
—
34.8 B
R
3.6 B
1.8 B
—
2,640 B
—
17,400
437 J
52.6
685 B
53.2
3.9 J
23.4 B
—
1.170B
—
934
linear calibration of GC/MS.
Phase Separator
Liquid
(ua/l)
35
__
—
—
46
—
20,640
220,000 J
610,000 J
—
19,100
710 E
5,900 J
—
200
21 0,000 J
490,000
130
160
—
1 7,000 J
73,000 J
200,000 J
—
151.000J
1,870
2B
73.6 B
—
4B
70,000
135
430
3,650
131
19,400
53.2
1.1
126
2,860 B
14,900
7B
4,290
E: Concentration exceeded the linear range of GC/MS calibration.
J : Estimated value due to minor OC deviations or for tentatively identified compound (no standard
spectra indicates compound present below contract detection limit,
R: Data unusable due to major QC deviations.
— : Compound was below the quantifiable limit
but greater than zero.
available) or mass
*: Concentrations reported on a dry weight basis.
369
-------
1.3.2.2 Lagoon 9 Sludge
Sludge from Lagoon 9 was tested to determine LTTD performance on worst-case
sludge and to identify potential limitations of the treatment process for the
L.O.U.
Performance Evaluation
Table 7 presents the cleanup standards and pre- and post-treatment concentra-
tions for the compounds of concern for Lagoon 9 sludge. Aniline was detected
in the pre-treatment and post-treatment samples below the 10*^ cleanup
standards but above the 10"*> cleanup standard. Therefore, sufficient removal
occurred only to attain a 10"^ risk cleanup. Azobenzene was also detected in
the pre-treatment samples below the 10"^ but above the 10**> cleanup standards.
The post-treatment concentration was less than both cleanup standards. There-
fore, an acceptable level of removal has been achieved. Benzene concentra-
tions in the pre-treatment sample were greater than both cleanup standards and
were less than the cleanup standards in the post-treatment sample, indicating
an adequate degree of removal. For benzidine, the pre-treatment and post-
treatment concentrations were above the cleanup standards; therefore, an
acceptable removal level was not achieved for benzidine during this test. DCB
pre-treatment concentrations exceeded both cleanup standards; the post-treat-
ment concentration was below the standards. Therefore, acceptable removal was
achieved for DCB. In summary, an acceptable removal level was attained in the
worst case sludge for azobenzene, benzene, and DCB. However, LTTD only
achieved a 10'^ level removal for aniline and did not achieve an acceptable
removal level for DCB. Therefore, LTTD is not an effective process option for
remediating Lagoon 9 sludge.
Process Residuals Evaluation
The organic compound and metal analyses for post-treatment residuals of
Lagoon 9 sludge are shown in Table 8. Pre-treatment compounds are shown to
indicate compounds originally present in the sample. Benzene is distributed
evenly between the condensate and the phase separator liquid. Overall, the
semivolatile organic compounds appeared to be collected in the phase separator
liquid. However, for the compounds of concern, aniline and azobenzene are
present in higher concentrations in the condensate. Benzidine and DCB are
more concentrated in the phase separator liquid, as expected. Both waste
streams contained very high levels of VOCs and semi-volatile compounds, indi-
cating the need for further treatment. Metals may also need treatment prior
to discharge. Calcium and magnesium treatment may be required to prevent
scaling of equipment.
1.3.2.3 Lagoon 8 and 10 Sludge
A composite of sludge from Lagoons 8 and 10 was tested to evaluate the
effectiveness of LTTD to remediate the average contaminated sludge that may be
encountered during remediation of the Bofors Site.
370
-------
Units: ug/kg
TABLE 7
LOW TEMPERATURE THERMAL DESORPTION
ATTAINMENT OF CLEANUP STANDARDS
FOR COMPOUNDS OF CONCERN FOR LAGOON 9 SLUDGE
Bofors Site
Muskegon, Michigan
Cleanup Standards*
Compound
1CT Risk
10"6 Risk
Pre-Treatment*
Post-Treatment*
Aniline
Azobenzene
Benzene
Benzidine
3,3'- Dichlorobenzidine
1,100,000
920,000,000
230,000
220
4,600,000
11,000
9,200,000
2,300
2.2
46,000
52,000 J
37,000,000 DJ
980,000
1,000,000 J
11,000,000 DJ
43,300
326,700
330
3,700
18,400
LEGEND:
D : Value reported from a diluted sample aliquot in order to stay within the linear calibration of GC/MS.
J : Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
Cleanup standards are based on remediation for the groundwater ingestion exposure route.
*: Cleanup standards and sludge concentrations are on a dry weight basis.
371
-------
TABLE 8
LOW TEMPERATURE THERMAL DESORPTION
CHEMICAL CHARACTERIZATION OF PROCESS RESIDUALS
FOR LAGOON 9 SLUDGE
Bofors Site
Muskegon, Michigan
Post-Treatment Residuals
Analyte
Volatile Organic Compounds
Acetaldehyde
Acetone
Acetonitrile
Benzene
2-Butanone
Chlorobenzene
1,4-Dioxane
Ethylbenzene
2-Hexanone
Methylene Chloride
Styrene
Xylene (Total)
Vinyl Chloride
Unknown
Semi-Volatile Organic Compounds
Aniline
Azobenzene
Benzidine
Benzole Acid
2-Chloroaniline
2-Chlorophenol
3,3'-Dichlorobenzldine
2-Methylphenol
4-Methylphenol
Phenol
Metals
Aluminum
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Nickel
Pot8",sium
Silver
Sodium
Vanadium
Zinc
Pre-Treatment*
(ug/kg)
MA
NA
980,000
NA
6,000
52,000
37,000,000 DJ
1,000,000 J
600,000 J
11,000,000 DJ
2,720,000
4,1006
25,600 B
12,400
45,700,000 J
28,700 J
120,000 J
11,100,OOOJ
170,000 J
4,910,000
158,000
16,900
271,0008
3,570,000
9,7008
83,300,000
Condensate
(ug/l)
6,600 J
9,300 E
370 BJ
738
2,162
50
188
14 BJ
62
50
45 J
6,900,000
320,000
1,200,000
460
3,3106
10.5
7,350
174
6
4166
7.26
27.6
7766
254
Phase Separator
Liquid
(ug/l)
3,500 J
3,890 E
1.400BJ
730
710 J
6
160 J
55
59 BJ
53
20
36 BJ
1,070,000
210,000
64,000
21,000
444,000 J
4,000
20,700
8,900
25,300
84,000 E
2,230
4.8 B
56.2 B
2.76
8.4 B
206,000
17.4
123
7,170
119
35,900
191
29.6 B
2,420 B
42,400
14.3B
65,000
Legend:
B:
D:
E:
J:
NA:
Compound was also detected in associated laboratory blank.
Value reported from a diluted sample aliquot in order to stay within linear calibration of GC/MS.
Concentration exceeded the linea- range of GC/MS calibration.
Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
Compound was below the quantifiable limit.
Compound was not analyzed for.
Concentrations reported on a dry weight basis.
372
-------
Performance Evaluation
Cleanup standards and pre- and post-treatment concentrations for the compounds
of concern are presented in Table 9. Aniline and azobenzene concentrations in
the pre-treatment sample are below the 10*^ and the 10*6 cleanup standards;
therefore, no removal of these contaminants is necessary. Benzene was not
detected above the quantitation limit in the pre-treatment sample; however,
benzene was detected in the post-treatment sample at a concentration higher
than the pre-treatment quantitation limit. The presence of benzene is most
likely a result of cross-contamination during the treatability study testing.
Therefore, it is assumed that no removal of benzene is needed. Benzidine and
DCB are detected in the pre-treatment sample above the 10'^ and the 10"^
cleanup standards, but are not detected in the treated sample above the
quantitation limit. The quantitation limit is greater than the cleanup
standards. However, neither benzidine nor DCB are detected in the TCLP
extract above the quantitation limit; therefore, it is assumed that an accept-
able removal level has been achieved for these compounds. In summary, LTTD
appears to be effective in removing the contaminants of concern in Lagoon 8
and 10 sludge.
Process Residuals Evaluation
Analysis of post-treatment residuals is shown in Table 10. Pre-treatment
compounds are shown to indicate compounds originally in the sample. Overall
the VOCs were predominantly in the condensate, and the semivolatile organic
compounds were predominantly in the phase separator liquid. However, aniline
and azobenzene were detected at greater concentrations in the condensate than
the phase separator liquid; while benzidine and DCB were only detected in the
phase separator liquid. Several volatile organic compounds which were not
detected in the pre-treatment sample were found in the waste streams. This
may be indicative of different detection limits, or problems associated with
cross-contamination. Both condensate and the phase separator residuals will
require further treatment for organic compounds prior to discharge.
Generally, the metals which were detected in the residual waste streams par-
titioned to the phase separator liquids, with the exception of copper. Treat-
ment may be required for calcium and magnesium to prevent scaling of
equipment.
1.3.3 Conclusions
The bench-scale testing of LTTD for these samples indicates that LTTD may be
effective for treating Lagoon 3 soil and Lagoon 8 and 10 sludge. The testing
indicates that LTTD is not effective for Lagoon 9 sludge.
VOCs generally appeared to partition into the condensate phase, as expected.
However, the semi-volatile compounds, including aniline and azobenzene, often
were concentrated in the condensate. This may be due to the relative vola-
tility of these compounds compared to the other semi-volatile compounds.
Benzidine and DCB consistently concentrated into the phase separator liquid,
as expected.
373
-------
TABLE 9
LOW TEMPERATURE THERMAL DESORPTION
ATTAINMENT OF CLEANUP STANDARDS
FOR COMPOUNDS OF CONCERN FOR LAGOON 8 AND 10 SLUDGE
Bofors Site
Muskegon, Michigan
Cleanup Standards*
Compound
10"4 Risk
(ug/kg)
10"6 Risk
(ug/kg)
Pre-Treatment*
(ug/kg)
Post-Treatment*
(ug/kg)
Quantitation
Limit
(ug/kg)
Post-Treatment
TCLP
(ug/l)
Aniline
Azobenzene
Benzene
Benzidine
3,3'- Dichlorobenzidine
250,000 2,500 1,700 230 BJ
2,800,000 28,000 5,700 420 J
47,000 470 — 13,000 ESJ**
5.4 0.054 13,000 —
440,000 4,400 2,200,0000 —
3,300
1,300
LEGEND:
B:
D:
E:
J:
S:
Compound was also detected in the associated laboratory blank.
Value reported from a diluted sample aliquot in order to stay within the linear calibration of GC/MS.
Concentration exceeded the linear range of GC/MS calibration.
Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass
spectra indicates compound present below contract detection limit, but greater than zero.
Instrument response was saturated, result represents an estimated minimum concentration present.
— : Compound was not detected.
**: Represents probable cross-contamination during treatability testing.
Cleanup standards are based on remediation for the groundwater exposure route.
TCLP: Toxicity characteristic leaching procedure.
*: Cleanup standards and sludge concentrations are on a dry weight basis.
374
-------
TABLE 10
LOW TEMPERATURE THERMAL DESORPTION
CHEMICAL CHARACTERIZATION OF PROCESS RESIDUALS
FOR LAGOON 8 AND 10 SLUDGE
Bofors Site
Muskegon, Michigan
Post-Treatment Residuals
Analyte
Volatile Organic Compounds
Acetaldehyde
Acetone
Acetonitrile
Benzene
2-Butanone
Hexane
Methylcyclopentane
Methylene Chloride
Toluene
Semi-Volatile Organic Compounds
Aniline
Azobenzene
Benzidine
2-Chloroaniiine
4-Chloroaniline
2-Chlorophenol
1 ,2-Dichlorobenzene
3,3'-Dichlorobenzidine
Dichlorobenzidine Isomer
Methanone, (3-Chlorophenyl) (4-Chlorophenyl)
2-Methylphenol
4-Methylphenol
Phenol
Total Unknowns
Metals
Aluminum
Arsenic
Barium
Beryllium
Cadmium
Calcium
Copper
Iron
Lead
Magnesium
Manganese
Nickel
Potassium
Sodium
Vanadium
Zinc
Legend:
B : Compound was also detected in associated
Pre-Treatment
(ug/kg)
_
—
—
—
30 BJ
—
—
210 BJ
4BJ
1,700
5,700
13,000
9,400
—
—
—
2,200,000 D
—
88,000 J
—
—
—
216,000 J
3,930,000
4,500
53,200 B
—
1,6006
241,000,000
18,200
1,330,000
11,200
2,470,000
32,900
—
79,600 B
284,000 B
4,0006
1,190,000
laboratory blank.
D : Value reported from a diluted sample aliquot in order to stay within
* Condensate
(Ufl/l)
1,500 J
6.557 E
1,300 J
20
249 E
—
—
22 BJ
19
62,000
28,100
—
85,600 J
2,600
20
—
—
—
—
—
30
210
—
_
—
—
—
—
2,0108
25,000
82.9 B
2.6 B
269 B
—
—
—
669 B
—
384
linear calibration of GC/MS.
Phase Separator
Liquid
(us/l)
20 J
71 J
98 J
11
—
20 J
8J
2BJ
66
8,000
15,800
32,000
3,700 J
670
—
30
21,000
7,000 J
16,000 J
30
—
240
22,900 J
4,650
3.2 B
80.86
1.8B
7.7
329,000
553
3,530
51
16,700
68.6
25 B
2,060 B
9,520
8.76
14,700
E: Concentration exceeded the linear range of GC/MS calibration.
J : Estimated value due to minor QC deviations or for tentatively identified compound (no standard
spectra indicates compound present below contract detection limit
— : Compound was below the quantifiable limit
, but greater than zero.
available) or mass
*: Concentrations reported on a dry weight basis.
375
-------
1.4 SOLIDIFICATION/STABILIZATION
On-site solidification/stabilization processes involve mixing excavated con-
taminated materials with proportional amounts of treatment reagents that
combine physically and/or chemically with contaminated materials to decrease
the mobility of waste constituents. Depending upon the amount and type of
reagent added, the end product may be a standing'monolithic solid or may have
a crumbly consistency. Mixing of wastes with treatment reagents is usually
performed in batch plants. Mixtures may be conveyed to concrete mixing tanks,
where hydration water is added if needed, and thoroughly blended. Then,
treated materials are generally placed in confining pits on-site for curing.
Portland cement and pozzolanic materials such as fly ash, ground blast furnace
slag, and cement kiln dust are widely used as immobilization reagents because
of their availability and effectiveness in binding contaminants to minimize
leaching. These materials may be used singly or in combination. A number of
proprietary additives, including polymers and absorbents, have been developed
for use with cement and pozzolanic materials to improve the physical charac-
teristics and decrease the leaching losses from the resulting solidified
material.
1.4.1 Treatability Experimental Procedure
The following samples were sent to Enreco for use in the solidification/
stabilization treatability study: Lagoon 3 soil (worst-case soil); Lagoon 9
sludge (worst-case sludge); and Lagoon 8 and 10 composite sludge (average
sludge).
Enreco performed a literature search to identify reagents which could be used
to solidify/stabilize contaminated soil and/or sludge from the Bofors Site at
this site. Based on findings of the literature search and previous
experience, three solidification/stabilization processes were identified:
1. Solidification and stabilization using Portland cement, fluidized bed
material and Quickline;
2. Adsorption, solidification, and stabilization using organophyllic
clays, high carbon coke fines, carbide lime and combinations of
reagents;
3. Polymerization, solidification, and stabilization using monomers
polymerized by catalysts.
The third process was eliminated during the literature search because it was
determined to be uneconomical.
The testing program involved a three-round iterative process to evaluate each
of the two remaining processes to determine the optimal formulation to achieve
cleanup standards. Samples containing 200 grams of contaminated material were
mixed with varying ratios of the fixation reagents and were then cured in
airtight containers. Reagents that produced favorable results were selected
for further testing in succeeding rounds. Each additional round of testing
was used to refine the mix designs.
376
-------
Round 1 mixing and testing consisted of screening techniques with the
reagents. This round was used to evaluate the physical and chemical
properties of the mixtures. The samples were tested for unconfined com-
pressive strength using a pocket penetrometer, and for density and volume
expansion.
Round 2 mixing and testing was performed with the preferred reagents
determined in Round 1. The mixtures were prepared to refine the reagent
formulations and optimize material usage. Some additional mixtures were
prepared to narrow the field of potential reagents. After a review of the
results and an economic analysis, mix designs were selected for the final
round.
Round 3 used the ratios determined in Round 2 to prepare sufficient volume of
material which was then refrigerated and cured for seven days. The stabilized
material was then distributed to laboratories to undergo TCLP extraction.
1.4.2 Experimental Results
As discussed previously, the analytical results from the solidification/
stabilization treatability study were evaluated differently than the soil
washing and LTTD treatability studies. Soil washing and LTTD are removal
options that extract the contaminants from the soil and/or sludge. Solidifi-
cation/stabilization does not extract the contaminants from soil and/or
sludge, but entrains the contaminants in the stabilized product so that leach-
ability is decreased. Therefore, concentrations in the TCLP extracts of the
pre- and stabilized soil and/or sludge samples were compared to evaluate the
reduction in mobility. If the compounds of concern are detected in the
stabilized material TCLP extracts, this may indicate that the stabilizing
reagents are not reducing the mobility of these compounds.
1.4.2.1 Lagoon 3 Soil
The recommended stabilization reagent mix for final treatment of Lagoon 3 soil
to reduce mobility of the compounds of concern and be cost effective was
60 percent fluidized bed material and 10 percent carbide lime.
Performance Evaluation
Unstabilized and stabilized TCLP extract concentrations and percent reduction
in concentration are presented in Table 11. Aniline was not detected in the
unstabilized extract but was detected in the stabilized extract. This result
may be due to laboratory contamination or differences in quantitation limits.
There was a greater than 90 percent reduction in TCLP concentration for
azobenzene between the unstabilized and stabilized samples. However, there
was only a 57 percent reduction for benzidine and a 79 percent reduction for
DCB. In summary, azobenzene, benzidine and DCS were detected in the
stabilized TCLP extract. Concentration reductions were achieved; however,
significant amounts of these compounds remained in the stabilized TCLP
extract; indicating that the mobility of azobenzene, benzidine, and DCB was
not sufficiently reduced by stabilization.
377
-------
co
•sj
00
TABLE 11
SOLIDIFICATION/STABILIZATION
COMPARISON OF TCLP EXTRACT CONCENTRATIONS
Bofors Site
Muskegon, Michigan
Units: ug/l
Lagoon 3 Soil Lagoon
Pre- Post- Reduction in Pre- Po
Treatment Treatment Concentra- Treatment Treat
Compound TCLP8 TCLP8 tion(%) TCLP8 TC
Aniline — 110 * 2,200 €
Azobenzene 9.400 8800 90.64 1,500 1,1
Benzene 5J 24 * NA 9,C
Benzidine 21,000 9,000 D 57.14 16,000 5,7
3, 3'-Dichlorobenzidine 14,000 2,9000 79.29 2,300 1.E
LEGEND:
a : Concentration in TCLP extract from stabilized sample.
9 Sludge Lagoon 8 & 10 Sludge
st- Reduction in Pre- Post- Reduction in
ment Concentra- Treatment Treatment Concentra-
LP" tion(%) TCLPa TCLP8 tion(%)
>30 71.36 53 J — - *
00 26.67 160J 28 J 82.50
IOO * — 3J *
rOO 64 44 .... _.„ *
(00 J 34.78 3,500 3,000 14.29
D : Value reported from diluted sample aliquot in order to stay within the linear calibration of GC/MS.
J : Estimated value due to minor QC deviations or for tentatively identified compound (no standard available) or mass spectra indicates compound present below contract
detection limit, but greater than zero.
B : Compound was also detected in associated laboratory blank.
TCLP : Toxicity characteristic leaching procedure.
— : Compound was below quantifiable limit or no associated risk.
*: Reduction in concentration cannot be calculated or is not applicable.
NA: Compound not analyzed.
-------
1.4.2.2 Lagoon 9 Sludge
The recommended stabilization reagent mix for Lagoon 9 sludge was 60 percent
fluidized bed material and 2 percent organophillic clay.
Performance Evaluation
As shown in Table 11, aniline, azobenzene, benzidine, and DCB were detected in
both the unstabilized and stabilized TCLP samples. Benzene was not analyzed
in the unstabilized TCLP sample due to loss of sample. Benzene is present in
the stabilized TCLP sample. Percent reductions in concentration range from
34 percent to 71 percent; however, significant amounts of all six compounds of
concern were still detected in the stabilized TCLP extract. This indicates
that the mobility of the compounds was not sufficiently reduced by
stabilization.
1.4.2.3 Lagoon 8 and 10 Sludge
The recommended formulation of stabilization reagents for Lagoon 8 and 10
sludge was 32 percent fluidized bed material and 5 percent carbide lime.
Performance Evaluation
As shown in Table 11, aniline was detected in the unstabilized TCLP extract
but was not detected in the stabilized TCLP extract, indicating sufficient
reduction in mobility. Benzene was not detected in the unstabilized sludge
but was detected in the stabilized sludge at a low concentration, possibly
indicating laboratory contamination. Benzidine was not detected in either
sample. Azobenzene and DCB achieved some reduction in concentration; however,
significant amounts of these compounds remained in the stabilized TCLP
extract, indicating that their mobility was not sufficiently reduced.
1.4.3 Conclusions
Different formulations were developed to combine with sample materials from
Lagoon 3 soil, Lagoon 9 sludge, and Lagoon 8 and 10 sludge. The reagent
common to all three formulations was fluidized bed material.
A comparison of unstabilized and stabilized TCLP extract concentrations showed
that, while some compounds achieved sufficient concentration reductions, most
compounds were still present in the stabilized TCLP extract in significant
concentrations. This indicates that stabilization is not an effective tech-
nology for use on any of the samples tested.
1.5 UNCERTAINTY OF TREATABILITY STUDIES
Treatability studies were conducted to obtain additional data necessary to
adequately evaluate remediation alternatives during the RI/FS. The results
will be used for costing and determining performance uncertainties, as well as
to eliminate alternatives which are not technically feasible for site cleanup.
Bench scale tests may simulate the full scale processes, but the results of
these studies should be used with discretion to develop cost estimates and
design parameters.
379
-------
Soil washing is an innovative technology with respect to Superfund site
remediation. The soil washing process utilizes existing technology which is
common to the metallurgical and mining industries. The soil washing unit
operations are standard, yet the process itself is unique to the site and for
the treatment objectives. Unit operations are typically much larger for
industrial uses than those recommended for the L.O.U. Additional problems and
costs may be incurred in designing a full scale process because of the use of
smaller, specialty equipment in the bench scale studies. Another concern with
the soil washing process involves the volume of water used to clean the soil.
The accuracy of this estimate from bench scale tests is dependent upon the
mixing intensity, chemicals used, and contaminant concentrations. If the
mixing intensity is greater for the bench scale process than for full scale,
the cleaning action will be greater and less wash water will be needed. This
uncertainty also relates to the type of detergents used to scrub the soil and
their effectiveness under lower agitation.
For LTTD, the initial concentration of semivolatile organic compounds will
drastically affect the other process streams associated with the system. The
volumes and contents of the condensate and phase separator liquids are
directly related to the initial moisture content and concentrations of the
volatiles and semi-volatiles. The bench scale test uses only a kilogram of
material; therefore, these waste streams may not be accurately simulated due
to their low volumes. The bench scale results may not accurately predict
problems which may occur in the condensate and phase separator treatment
process. The treated soil fraction should represent a reasonable estimate of
the final product in the full scale process since removals are primarily
dependent upon kiln temperature and retention time for a given initial concen-
tration of contaminants.
Variations in contaminant concentrations will also affect the results of the
solidification/stabilization bench scale tests, but the major area of
uncertainty is the ability of the bench scale tests to simulate the addition
of the reagents during the full scale process and whether TCLP is an accurate
predictor of long-term performance and reduction in mobility.
RP/EPASSTS/AA1
380
-------
BIBLIOGRAPHY
Water Pollution Control Federation, Wastewater Treatment Plant Design, Manual
of Practice No. 8, 1977.
Remedial Investigation Report for the Bofors Site, MDNR, February 1990.
Draft Phase II Remedial Investigation Report for the Bofors Site, MDNR,
March 1990.
Draft Feasibility Study Report for the Lagoon Operable Unit at the Bofors
Site, MDNR, May 1990.
RP/EPASSTS/AA1
381
-------
CONCEPTUAL COST EVALUATION
OF VOLATILE ORGANIC COMPOUND
TREATMENT BY ADVANCED OXIDATION
GLENN J. MAYER, P.E.
CH2M HILL, Inc.
2510 Red Hill Avenue, Suite A
Santa Ana, California 92705
WILLIAM D. BELLAMY
CH2M HILL, Inc.
P.O. Box 22508
Denver, Colorado 80222
NEIL ZIEMBA/LEE A. OTIS
U.S. EPA Region DC
1235 Mission Street
San Francisco, California 94103
Presented at: Second Forum on Innovative Hazardous Waste Treatment
Technologies: Domestic and International, May 15-17, 1990,
Philadelphia, Pennsylvania
LAO22033U31 007.51
382
-------
SUMMARY AND CONCLUSION
Cost estimates based on conceptual designs of groundwater treatment facilities for
volatile organic compounds (VOCs) using three different treatment technologies have
been developed for a local area of the San Gabriel Basin. These technologies are:
1) packed tower air stripping with vapor phase carbon adsorption, 2) liquid phase
carbon adsorption, and 3) ozone/peroxide advanced oxidation. Traditional design data
and costs are used in evaluating the first two technologies. Bench-scale testing on site-
specific groundwater is used for the development of conceptual treatment plant design
and cost estimates for the advanced oxidation process.
Using the same plant sizing, influent water quality, and target treatment levels for each
treatment technology, a cost comparison has been prepared. Costs for two influent
conditions are evaluated: the probable design case, representing the conditions most
likely to be encountered at the treatment plant given data currently available; and, the
maximum credible deviation case representing the worst conditions that may credibly
be encountered during the next 30 years. The results of the cost comparison are as
follows:
Treatment Technology
Air Stripping
Carbon Adsorption
Advanced Oxidation
Cost in $/l,000
gallons
Probable
Design
Case
0.19
0.43
0.25
Maximum
Credible
Deviation
Case
0.35
__a
0.25
*Liquid phase carbon adsorption is not eco-
nomically feasible in the maximum credible
deviation case due to high carbon loading.
Considering the costs of treatment alone, advanced oxidation appears to be more
economical in the maximum credible deviation case. However, a number of other
criteria, including reliability, ease of operation and maintenance, expected lifetime, and
presence of treatment process hazards, must also be considered in evaluating treatment
options.
383
LAO22033\231 008.51
-------
BACKGROUND
The 170-square-mile San Gabriel Basin is located in the northeastern portion of Los
Angeles County. The water-bearing formations of the basin consist of gravel, boulders,
and coarse sediments ranging to a thickness of over 4,000 feet or more. Groundwater
movement is generally from the perimeter of the basin toward the pumping centers in
the basin and then toward the Whittier Narrows, which is the only identified subsurface
discharge from the basin. Figure 1 shows the location of the San Gabriel Basin.
In recent years, over 200,000 acre-feet of water have been extracted annually from the
basin by more than 45 water purveyors. Groundwater provides water supply for
approximately 90 percent of the basin population of over 1 million people.
In December 1979, VOC contamination of the groundwater was discovered. Further
sampling by the Los Angeles Regional Water Quality Control Board and the California
Department of Health Services confirmed contamination over a wide area. EPA pro-
posed in September 1983, and subsequently listed in May 1984, that four general areas
of groundwater contamination be included on the National Priority List (NPL). In
February 1986, EPA initiated remedial investigation/feasibility study (RI/FS) activities in
a combined effort for the entire San Gabriel Basin.
A primary objective of the RI activities was to develop appropriate treatment technolo-
gies for contaminated groundwater. During the review and development of techno-
logies for treatment of synthetic organics, especially VOCs, it was determined that
activated carbon, air stripping, and oxidation were all possible treatment techniques.
Carbon adsorption and air stripping have been accepted and are proven technologies
for many of the organic constituents present in the basin. Oxidation, on the other
hand, has only recently been proposed as a cost-effective treatment technology for
drinking water.
Since this technology is relatively new and unproven and it has not yet been extensively
used for large municipal water treatment systems, it was decided that additional inves-
tigation into the technology was appropriate. A literature review was conducted, and
recent work in the field, most notably by Dr. William Glaze of the University of North
Carolina, suggests that some VOCs (e.g., trichloroethene [TCE] and tetrachloroethene
[PCE]) can be oxidized at an equivalent cost to air stripping with off-gas carbon
adsorption. Oxidation can result in destruction of the VOCs with minimal or no
recognized by-products in the water and no significant VOCs released to the air.
Reaction kinetics for traditional ozonation have not been considered favorable for the
oxidation of most chlorinated organics in potable water systems. The inefficient oxida-
tion of most VOCs by ozone (O3), relatively low contaminant concentrations, and large
water volumes associated with VOCs and potable water treatment create excessive
capital and operating costs for a conventional ozone system.
384
LAO22033N231 008.51
-------
CO
CO
en
RAYMOND
BASIN
SAN GABRIEL MOUNTAINS
SAN GABRIEL BASIN
GROUNDWATEH
FLOW
DIRECTION
WHITTIER
NARROWS
CENTRAL
ORANGE COUNTY
GROUNDWATER
BASIN
NOT TO SCALE
FIGURE 1
LOCATION MAP
-------
Developments in oxidation technology for aqueous organics, known as "advanced
oxidation processes" (AOPs), appear more feasible for the treatment of VOC
contaminated potable water sources. AOPs involve the generation of highly reactive
free radical intermediates from the decomposition of ozone or hydrogen peroxide
(H2O2). Organics are oxidized much faster by ozone decomposition by-products (e.g.,
free radicals) than by O3. Consequently, AOPs are designed to promote the formation
of free radicals, thereby enhancing oxidation of organics to levels practicable for
drinking water treatment.
Formation of free radicals by decomposition of ozone can be initiated by hydroxide
ions (OH "), hydrogen peroxide, ultraviolet (UV) light, and some transition metal ions.
The AOP systems currently receiving attention include Oj/high pH, O3/H2O2, O^V
light, and H2O2/UV light. On the basis of the literature review, the O3/H2O2 system
appears to be the most practical for potable water treatment when the organics
oxidized do not react with UV light. Hence, it was decided to conduct a bench-scale
evaluation of the O3/H2O2 process for oxidizing VOCs in groundwater from the San
Gabriel Basin.
AOPs can effectively oxidize VOCs. This is a potential advantage over treatment
methods currently designated as best available technologies (BATs) for VOC removal
(packed tower aeration or granular activated carbon [GAC] adsorption; EPA, 1979).
The BATs transfer the contaminants from water to another phase or medium which, in
turn, has to be treated or disposed of. Air stripping followed by gas phase oxidation
has the potential to destroy the VOCs, resulting in the same benefits as the AOPs. Gas
phase oxidation was not evaluated in this study because, at the time of initiation of the
oxidation study, gas phase oxidation had not been demonstrated adequately.
Ozone/peroxide bench-scale testing was undertaken as part of the remedial investiga-
tion process of the San Gabriel Basin RI/FS. The purpose of the bench-scale tests was
to evaluate the applicability of the ozone/peroxide process to the treatment of
contaminated groundwater from the basin. Results of the tests were used to predict
effectiveness of the technology and treatment costs. This paper briefly highlights the
advanced oxidation process, describes the bench-scale testing methods and results of
the testing, and presents a conceptual cost comparison of the ozone/peroxide technol-
ogy with packed tower air stripping/vapor phase granular activated carbon adsorption
(PTAS/VGAC) and liquid phase granular activated carbon adsorption (LGAC). Cost
estimates are expected to be accurate within plus 50 percent to minus 30 percent. Cost
is only one of nine factors considered in the selection of a remedial action at an NPL
site.
386
LAO22033\231 008.51
-------
ADVANCED OXIDATION PROCESS
Aqueous ozone can oxidize dissolved organics by two mechanisms: 1) direct reaction
or 2) decomposition of the ozone to intermediates (e.g., radicals), which in turn reacts
directly or indirectly with the organics. The direct ozone reactions are highly
substrate-specific and relatively slow, occurring on the order of minutes. Hoigne and
Bader (1983) reported second order rate constants for the direct oxidation of organics
in the range or 1 to 1,000 per molar per second. In contrast, reactions between ozone
decomposition intermediates and solutes are nonselective and fast, occurring in micro-
seconds. For example, rate constants for oxidation of organics by the hydroxyl radical
(OH •) are typically 10* to 1010 per molar per second (Farhataziz and Ross, 1977).
Figure 2 illustrates the complex cyclic reactions involved in the aqueous chemistry of
ozone. The primary free radicals formed by ozone decomposition may: 1) enter into
chain reactions with other ozone molecules, perpetuating (autocatalyzing) ozone
decomposition; 2) react with organic solutes to form oxidized carbon constituents and
secondary radicals that also participate in chain reactions with ozone; or 3) react with
radical-scavenging solutes (carbonate, bicarbonate, certain organics), forming inefficient
oxidation species and effectively inhibiting radical chain reactions.
The observed oxidation of solutes in any ozonated system is a combination of direct
and radical reactions. The relative concentrations of the two types of reactants and the
importance of the two types of reactions are highly dependent on the chemistry of the
water being treated. This is due to the presence of naturally occurring constituents that
promote or inhibit ozone decomposition and radical chain reactions. Since the overall
rate at which organics are oxidized is a function of both direct and radical reactions,
bench- or pilot-scale testing is currently required to evaluate the effect of source water
chemistry on the efficiency of an advanced oxidation system.
BENCH-SCALE TEST PROGRAM
Bench-scale testing of groundwater collected from the San Gabriel Basin was under-
taken to determine the rate of destruction of select VOCs. The design of the bench-
scale testing system was based on the apparatus described by Glaze and Kang (1988).
A 75-liter, stainless steel, continuously mixed batch reactor (CMBR) was used for the
study. The reactor was equipped with a dual-impeller mixing system; a 20-cm-diameter
ceramic ozone sparger; and sampling and viewing ports. A laboratory ozone generator
was used to generate ozone from bottled air. Hydrogen peroxide solution was fed to
the reactor through teflon tubing by a positive-displacement chemical feed pump.
Figure 3 is a schematic of the bench-scale system.
387
-------
OH*
Direct Oxidation of Substrate
Slow, Selective
Cyclic Decomposition
Products
ofOjby Secondary Radicals
Cyclic Decomposition
/" of O, by OH"
Radical Formation ~.i.» Radical Oxidation
• Utl •
COjandHCOj
Fast, Nonselective
Secondary
Radicals
Products
Radical Consumption
CO* and HCO*
Figure 2
Cyclical Reactions of Ozone
(from Ai eta etal., 1988)
388
-------
Ozone
Monitor
Vent -*•
CO
00
IO
Oa Purge
-« '
*
Flow
Controller
Vent
i
Ozone
Generator
CD
Vent
Mixer
Drive
#
% W^A,
Batch VCx
Reactor.
>:
-Vent
a
Flow
Controller
4
4
4
k—t**—-O
Peroxide
Feed
Pump
Peroxide
Reservoir
Vent
ft\ Needle Valve
if I 3 - Way Valve
Figure 3
Schematic of Bench-Scale System
-------
Groundwaters for the AOP bench-scale experiments were collected from wells near the
Whittier Narrows (Southwest Suburban Water System) and near the town of Azusa
(Azusa Valley Water Company) in the San Gabriel Basin and shipped by air freight in
55-gallon polyethylene barrels to the laboratory.
Four target VOCs were used to evaluate the oxidative efficacy of the ozone/peroxide
process:
• Trichloroethene
• Tetrachloroethene
• Trans- 1,2-dichloroethene (t-l,2-DCE)
• Carbon tetrachloride (CTC)
Because groundwater concentrations of these VOCs in the wells sampled were well
below the desired testing concentration range, the groundwaters were spiked using a
semisaturated stock solution of the four test organics. Initial concentrations of the four
target compounds were either 50 or 500 micrograms per liter (ng/1).
A total of 15 laboratory tests were carried out under varying conditions of initial con-
centration, ozone mass feed rate, and ratio of peroxide to ozone molar dose. Table 1
summarizes the bench-scale test program.
RESULTS OF BENCH-SCALE TESTING
The results of bench-scale testing of the ozone/peroxide AOP for San Gabriel Basin
groundwaters are detailed in a summary technical memorandum prepared for EPA
(CH2M HILL, 1989).
Reaction rate constants for removal of target VOCs in this study are expressed in terms
of time as k, (sec ~1). In addition, empirical stoichiometric constants dependent on dose
have been calculated and are expressed as kD (liter/milligram [1/mg] ozone). Results
support the following conclusions:
• kt values for the removal of TCE and PCE in test runs utilizing an
H2O2/O3 molar ratio of 0.50 equalled or exceeded rate constants from
runs using other ratios (i.e., 0.25 or 1.00). In addition, lower residual O3
and H2O2 concentrations resulting from use of the 0.5 molar ratio suggest
that it provides the most efficient use of oridants.
390
LAO22Q33\231_008.51
-------
Table 1
Conditions for VOC Oxidation Runs
Test Number
2
3,14
4
5
6
11
8
7,15
9
10
13
12
Water Source
Suburban
Suburban
Suburban
Suburban
Suburban
Suburban
Azusa
Azusa
Azusa
Azusa
Azusa
Azusa
Target VOC
Concentrations
(ng/i)
500
500
500
500
50
50
500
500
500
500
50
50
Ozone Mass
Feed Rate
(mg/l-min)
0.205
0.205
0.205
0.051
0.205
0.051
0.205
0.205
0.205
0.051
0.205
0.051
HjOj/Oj
Dose Ratio
(Molar)
0.25
0.50
1.00
0.50
0.50
0.50
0.25
0.50
1.00
0.50
0.50
0.50
Note: Test 1 was a dry run using tap water.
LAO22033U31 001.51
391
-------
As expected from process theory, a higher ozone mass feed rate yielded
higher k, constants for PCE and TCE than a lower ozone mass feed rate.
Groundwater from the two wells studied were very similar with respect to
pH, alkalinity, hardness, and total dissolved solids (TDS). As a result, k,
values were not noticeably different for the two waters.
k, values were essentially independent of initial VOC concentration.
For conditions yielding the highest reaction rates (O3 mass feed rate =
0.205 mg/L-min; H2O^/O3 molar ratio = 0.50), k, values were:
PCE: 15-30 x 10" sec'1
TCE: 32-46 x 10-4 sec"1
kj values for t-l,2-DCE generally were too high to be quantified by the
sampling protocol employed (i.e., concentration was below detection limits
in the first sample taken after initiation of oxidant feed). Based on semi-
quantitative data, Iq values for t-l,2-DCE appeared to be two to three
times higher than those for TCE.
k, and kD values for TCE and PCE are consistent with those reported in
similar studies.
Control experiments conducted to compare kt values for the H2O2/O3
process with rate constants for gas stripping and oxidation by ozone alone
suggest the following:
PCE: Less than 10 percent of observed ozone/peroxide removal is
attributable to gas stripping.
TCE: Approximately 5 percent of observed ozone/peroxide remo-
val is attributable to gas stripping.
t-l,2-DCE: Gas stripping is negligible; the rate of oxidation by O3
alone is on the same order of magnitude as that of oxidation by
ozone/peroxide.
CTC: Virtually all observed ozone/peroxide removal is attributable
to gas stripping, not oxidation.
392
LAO22033U31 008.51
-------
Note: The relative contribution of gas stripping to observed
removals of VOCs in this study is applicable only to a batch system;
stripping would be expected to have less effect in a properly
designed full-scale, flow-through system.
The test results indicated that a stoichiometric constant, kD, of 0.61/mg O3
would be appropriate for estimates of ozone dose requirements. The fol-
lowing exponential equation would apply:
ln(C/C0) = kD [03]
where:
C0 is initial contaminant concentration, C is the final contaminant
concentration, kD is as defined above, and [O3] is the required
ozone dose in mg/1.
CONCEPTUAL COST EVALUATION
The above observations are used in the evaluation of the ozone/peroxide AOP for treat-
ment of groundwater in the Whittier Narrows area of the San Gabriel Basin. The
Whittier Narrows Operable Unit Feasibility Study (OUFS) has been prepared to
evaluate the appropriate action to be taken to mitigate contaminant migration out of the
San Gabriel Basin and into the Central Basin, an adjacent aquifer.
Preparation of a feasibility study for an EPA NPL (Superfund) site requires consideration
of the following nine criteria for each remedial alternative:
• Overall protection of human health and the environment
• Compliance with applicable or relevant and appropriate regulations
(ARARs)
• Long-term effectiveness and permanence
• Reduction in mobility, toxicity, and volume
• Short-term protectiveness
• Implementability
• Cost
393
LAO22033\231 008.51
-------
State acceptance
• Community acceptance
A number of alternatives are presented in the Whittier Narrows OUFS, encompassing
not only treatment technologies, but also plume containment measures (i.e., where and
how much to pump) and treated-water distribution measures. The combined alternatives
are evaluated with respect to the above nine criteria. The remainder of this paper
presents only the evaluation of the cost effectiveness of the following three treatment
technologies considered in the Whittier Narrows OUFS:
PTAC/VGAC
LGAC
• Ozone/peroxide advanced oxidation
BASIS OF THE COST EVALUATION
A common set of input conditions are required to accurately compare the costs
associated with each of the treatment technologies. Conditions that are anticipated for
remedial action at the Whittier Narrows area of the San Gabriel Basin are used in this
cost evaluation.
The Whittier Narrows OUFS has been prepared using a new approach to feasibility
studies at EPA Superfund sites. This approach, referred to as the "observational"
approach, requires the evaluation of two conditions for each remedial alternative: 1) the
probable design case, and 2) the maximum credible deviation case. The probable design
case represents the conditions most likely to be encountered during remedial action,
given information available at the time of FS preparation. The maximum credible devia-
tion case represents conditions that may credibly occur within the next 30 years. An
evaluation of the two conditions allows the remedial alternatives flexibility to handle
changing conditions. In the preparation of the Whittier Narrows OUFS, treatment costs
are evaluated for both of these cases.
The following groundwater flow rates are used in the development of costs for the treat-
ment portion of the remedial alternatives.
Flow Rate
Peak (gpm)
Average (gpm)
Probable
Design
Case
14,600
7,200
Maximum
Credible
Deviation
Case
17,700
11,800
394
LA022033\231 008.51
-------
The concentration of contaminants in groundwater requiring treatment as part of a
Whittier Narrows remedial action is based on observed concentrations in both the
Whittier Narrows and upgradient in the San Gabriel Basin. Probable design
concentrations and maximum credible deviations from the probable concentrations were
estimated using regional hydrogeological modeling. Table 2 presents the influent
concentrations used in developing the costs for treatment of Whittier Narrows
groundwater, as well as the target treatment levels (state and federal maximum
contaminant levels [MCLs]).
Using the estimated influent concentrations and groundwater flow rates, costs are esti-
mated for the treatment of the groundwater extracted in the Whittier Narrows area.
Additional assumptions and input data needed to develop costs for each specific
technology are presented below. Capital and operation and maintenance (O&M) cost
estimates are expressed in terms of dollars per 1,000 gallons of treated water.
Capital cost estimates are based on the equipment required to meet the peak flow.
Average flow rates are used to calculate annual O&M costs. The annualized cost of
capital is calculated assuming a 30-year facility life and 5 percent interest rate. The
annual cost of the capital investment is added to the annual O&M cost to determine the
total annual treatment cost for the facility. This value was divided by the total volume
of water treated by the facility, assuming average flow conditions of 24 hours per day,
365 days per year.
COST OF AIR STRIPPING
A conceptual design layout is required prior to estimating costs of treating groundwater
by air stripping. Because of the location of the project in the South Coast Air Basin of
Southern California, off-gas treatment is a required process for both probable design and
maximum credible deviation conditions.
Due to the large flow rate on which the sizing of a groundwater treatment plant is based,
a number of air strippers operated in parallel is required. Using computer modeling of
the air stripper, single units capable of meeting the target treatment levels were
conceptually designed. The specifications for these units are outlined below.
395
LAO22033\231 008.51
-------
Table 2
Influent Water Quality and Treatment Criteria
(All Values in ppb)
Treatment Plant Influent
Contaminants
PCE
TCE
Carbon Tetrachloride
1,1,1-TCA
1,1-DCA
1,1-DCE
cjs-U-DCE
trans-l,2-DCE
U-DCA
UA2-TCA
Acetone
Benzene
Ethylbenzene
Methylene Chloride
Toluene
Total Xylenes
Vinyl Chloride
MEK
Bromoform
Dibromochloromethane
Chloroform
Freon 113
Probable
Design
Condition*
75
15
1.1
3 ntn
.08 ntn
3 ntn
2.7 ntn
20 ntn
0.6
0.25 ntn
3.3 ntn
1 ntn
1 ntn
3 ntn
1 ntn
1 ntn
0.4 ntn
5.5 ntn
2 ntn
2 ntn
6 ntn
3 ntn
Credible
Deviation
Condition1*
240
170
2.5
135 ntn
16 ntn
10
22 ntn
65 ntn
0.9
1
250
3.4
5.6 ntn
2,800
40 ntn
55 ntn
1
10 ntn
8 ntn
2.6 ntn
6.5 ntn
4 ntn
Treatment Level
5
5
0.5
200
.20
6
0.5
1
na
1
680
40
1,750
0.5
200
18,000
Source of
Regulation
Cal MCL
Cal MCL
Cal MCL
Cal MCL
CalAL
Cal MCL
Cal MCL
Cal MCL
Cal MCL
Cal MCL
CalAL
Cal MCL
Cal MCL
HEA ADV
CalAL
'Probable design conditions are based on migration of contaminants found in production wells
within the 30-year groundwater travel zone. Vertical and areal dilution have also been accounted
for.
''Maximum credible deviation conditions are the same as (a) with the exception that concentra-
tions used were from source investigation sampling within the 30-year travel zone.
ntn - no treatment necessary (the concentration is below the target treatment level)
na - no standard was available
Cal AL - California Action Levels
MCL - Maximum Contaminant Level
HEA ADV - Federal Health Advisories
396
LAO22033\231 002J1
-------
Diameter (ft)
Packed Depth (ft)
Water Flow Rate (gpm)
Air: Water Ratio
Air Flow Rate (cfm)
GAC Off-Gas
Treatment
Probable
Design
Case
12
20
3,000
20:1
8,000
Nonregenerable
Maximum
Credible
Deviation
Case
12
20
3,000
75:1
30,000
Regenerable
Table 3 summarizes the cost estimate for packed tower air stripping with VGAC, as
estimated for the Whittier Narrows OUFS. These costs are estimates based on
preliminary design data and neglect such factors a facility location, delivery system
requirements, etc. The cost estimates are intended to represent order-of-magnitude
(+50 and -30 percent) estimates. These costs are expected to change during detailed
remedial design. Development of these conceptual costs is based on the following
additional assumptions:
• Regenerable off-gas carbon units in the maximum credible deviation case
are Hastelloy-constructed units.
• Condensate will be shipped to a local treatment facility at a total cost of
$1.00 per gallon in the maximum credible deviation case.
• Each 3,000-gpm stripper operating at full capacity requires
1,090,000 pounds of steam per year to regenerate the vapor-phase carbon
in the maximum credible deviation case.
• The regenerable carbon will require replacement every 3 years, and the
annual replacement cost is based on replacing one-third of the full charge
each year.
• Carbon replacement for the probable design case is based on an average
150 ng/1 VOC loading and a 0.044 Ib VOC/lb GAC available capacity.
397
LAO22033\231 008.51
-------
Table 3
Treatment Costs Using Packed Tower Air Stripping*
CAPITAL COST ESTIMATES
Air Stripper (3,000 gpm/unit)
Air Supply fans
Tower Recirculation Pumps
Off-Gas Prebeaters
Off-Gas Carbon Units
Off-Gas Fan
Water Feed Pump
Equipment Cost Subtotal
Bid Contingencies @ 15%
Scope Contingencies @ 25%
Construction Total
Service During Construction @
10%
Land Acquisition
Total Implementation Cost
Engineering, l^gai, and
Administrative Costs @ 22%
TOTAL CAPITAL COST
Unit Cost (VUnit)
105,750
4,200 (Design)
11,800 (CredDev.)
4,100
1,600
93,750 (Design)
152,000 (Cred.Dev)
7,100 (Design)
11300 (Cred.Dev.)
13,800
_
_
_
_
__
300,000
M
_
-
Units Required
Probable Design
Sea
6ea
Sea
6ea
6ea
6ea
6ea
_
_
-
_
„
2 acres
_
_
-
Maximum
Credible
Deviation
6ea
7ea
Sea
18 ea
24 ea
7ea
7ea
—
_
-
—
_
2 acres
_
_
-
Treatment Facility
Cost(s)
Probable
Design
S 528,800
25,200
12300
9,600
562^00
42,600
82,800
$1,2453,800
189,600
316,000
$1,769,400
176,900
600,000
$2^46300
560,200
$3,106Ł00
Maximum
Credible
Deviation
$ 634400
82,600
12300
28300
3,648,000
82,600
96,600
$4,585,400
687,800
1,146,400
$6,419,600
642,000
600,000
$7,661,600
1,685,600
$9347,100
LA022033\231_003.51-16
398
-------
Table 3
(Continued)
O&M COST ESTIMATES
Electrical Power
Carbon Replacement
Natural Gas
Boiler Fuel for Steam in Carbon
Regeneration
Water Sample Analysis
Air Sample Analysis
Condensate Disposal
Operating Labor
Maintenance @ 2% of Equipment
Costs
TOTAL O&M COSTS
Annualized Cost of Capital
Total Annul Costs
UNIT COST FOR TREATMENT
(V1.000 gal)a
Unit Cost (VUnit)
0.07
Z50
3.00
0.45
110
250
1.00
35
-
_
-
_
-
Unite Required
Probable Design
1,913,700 kW-hr
110,100 Ib
8,400 MM Btu
Ogal
72 ea
72 ea
Ogal
730 hr
-
_
_
_
-
jLffjMrliiingn
Credible
Deviation
7,633^00 kW-hr
82,400 Ib
37,700 MM Btu
4,400
84 ea
216 ea
523,400 gal
730 hr
-
_
_
_
-
Treatment Facility
Cost(s)
Probable
Design
$134,000
275,200
25,200
0
7,900
18,000
0
25,600
25300
$511,200
202300
$713,400
0.19
Mttdmnm
Credible
Deviation
$534300
206,000
113,100
2,000
9,200
54,000
523,400
25,600
91,700
$1^59300
608,500
$2,167,800
OJ5
aUnit treatment costs are estimates based on preliminary design data and neglects factors such as facility location, distribution
system requirements, etc. The cost estimates are intended to represent order-of-magnitude (+50 and -30 percent) estimates.
These costs will change during detailed design.
399
LAO22033\231 003.51-17
-------
As shown in Table 3, the unit treatment cost is estimated at $0.19 per 1,000 gallons for
the probable design case and $0.35 per 1,000 gallons for the maximum credible deviation
case. Treatment costs for the maximum credible deviation case are higher because of
higher capital and operating costs. Capital costs are higher because corrosion-resistant
Hastelloy carbon vessels are needed for the regenerable GAC system. Operating costs
are higher primarily because of higher power costs and condensate disposal costs. Power
costs are higher because of the increase in airrwater ratio from 20:1 to 75:1.
COST OF LGAC
A conceptual design layout for LGAC was developed prior to estimating costs for
treatment.
Based on the probable design and maximum credible deviation influent concentrations
and the required removal efficiencies to achieve target treatment levels, the following
design criteria were developed.
Diameter (ft)
GAC depth (ft)
Empty bed contact time (minutes)
Liquid Loading (gpm/ft2)
Probable
Design
Case
12
8
6
10
Maximum
Credible
Deviation
Case
—
~
—
—
The use of LGAC in the maximum credible deviation case is unrealistic because of the
very large carbon replacement requirements. In this case, air stripping was assumed.
The design criteria and costs for air stripping were discussed above.
A liquid loading rate of 10 gpm per square foot of GAC and a bed diameter of 12 feet
results in a maximum hydraulic capacity of 1,130 gpm per carbon bed. To achieve the
required treatment capacity of 14,600 gpm, a total of 14 carbon beds are required. The
cost estimate for carbon treatment assumes a system of two carbon beds in series to
avoid breakthrough. Additionally, one extra set of beds is included to accommodate
shutdowns. Therefore, treatment of Whittier Narrows groundwater with LGAC would
require a total of 30 carbon beds.
400
LA022033\231 008.51
-------
Table 4 summarizes the cost estimate for LGAC, as estimated for the Whittier Narrows
OUFS. These costs are estimates based on preliminary design data and neglect such
factors as facility location, delivery system requirements, etc. The cost estimates are
intended to represent order-of-magnitude (+50 and -30 percent) estimates. These costs
are expected to change during detailed remedial design. As shown in Table 4, the cost
of treating Whittier Narrows groundwater by LGAC in the probable design case is
estimated at $0.43 per 1,000 gallons treated.
COST OF O3/H202 ADVANCED OXIDATION
Development of costs for ozone/peroxide advanced oxidation treatment of the Whittier
Narrows groundwater is more difficult than the traditional treatment technologies
described above because of the lack of data on implementation costs. Significant
assumptions were necessary to develop a cost estimate for the OUFS. The following
conditions were used to develop the costs:
• The reactor configuration is a flow-through system constructed of
concrete. Dimensions of the reactor were based on a residence time
within the reactor of 5.5 minutes. This residence time is an assumed value
which must be verified during pilot testing. The reactor was designed to
consist of four individual cells with ozone and peroxide feed in the first
three cells.
• Three reactors operating at 50 percent capacity were assumed to allow for
operating redundancy and flexibility in operating with the new technology.
• Ozone and peroxide chemical usage were determined using conditions
determined during the bench-scale testing described earlier. The following
conditions were used:
peroxide/ozone molar ratio = 0.5
empirical stoichiometric constant = 0.61/mg O3
• Electrical usage for generation of O3 from air = 12 kW-hr/pound Oj, cost
of electricity = $0.07/kW-hr, and cost of peroxide = $1.10/pound.
401
LAO22033V231 008.51
-------
Table 4
Treatment Costs Using Liquid-Phase Carbon Adsorption"'1'
CAPITAL COST ESTIMATES
Carton Adsorption Vessels
Chlorine Treatment System
Inlet Water Pumps
Air Strippers
Air Supply Fans
Retirculation Pumps
OS-Gas Reheaters
Off-Gas Carbon Vessels
Off-Gas Fan
Water Feed Pump
Equipment Cost Subtotal
Bid Contingencies @ 15%
Scope Contingencies @ 25%
Construction Total
Service During Construct 10;
@10%
Land Acquisition
Total Implementation Cost
Engineering, Legal, and
Administrative Costs @ 22%
TOTAL CAPITAL COST
Unit Cost (VUnit)
167,000
15,000
11,000
105,760
11300
4,100
1,600
152,000
11,800
13,800
—
_
_
—
-
300,000
_
-
-
Units Required
Probabk Design
30 ea
lea
Sea
0
0
0
0
0
0
0
H
_
_
H
-
2 acres
H
-
-
Maximum
Credible
Deviation
0
0
0
6ea
7ea
3ea
18 ea
24 ea
7ea
7ea
_
—
—
_
-
2 acres
_
-
-
Treatment Facility Cost(s)
Probabk
Design
$5,010,000
15,000
55,000
0
0
0
0
0
0
0
$5,0*0,000
762,000
1,270,000
$7,112,000
711,200
600,000
$8,423400
1353,100
$10,276^300
Maximum
Credible
Deviation
0
0
0
634,500
82,600
12300
28,800
3,648,000
82,600
96,600
$4,585,400
687,800
1,146,400
$ 6,419,600
642,000
600,000
$7,661,600
1,685,600
$ 9^47,100
LA022033\231 004.51-20
402
-------
Table 4
(Continued)
O&M COST ESTIMATES
Electrical Power
Carbon Replacement
Condensate Disposal
Natural Gas
Boiler Fuel
Water Sample Analysis
Air Sample Analysis
Operating Labor
Maintenance @ 2% of Capital
Costs
TOTAL O&M COSTS
Annualized Cost of Capital
Total Annual Costs
UNIT COST FOR TREATMENT
(Vl.000 gal)«
Unit Cost (S/Unit)
0.07
130 (Design)
Z50 (Cred.Dev.)
1.00
3
0.45
no
250
35
-
_
—
_
-
Units Required
Probable Design
693,800 kW-hr
600,000 Ib
0
0
Ogal
31 ea
0
730 hr
-
w
_
H
-
p^yicfni^m^
Credible
Deviation
7,633300 kW-hr
81,400 Ib
523,400 gal
37,700 MM Btu
4,400 gal
84 ea
216 ea
730 hr
-
M
_
H
-
Treatment Facility Coct(s)
Probable
Design
$48,600
780,000
0
0
0
3,400
0
25,600
101,600
$959,200
669,000
$1,628,200
0.43
Maximum
Credible
Deviation
$534300
206,000
523,400
113,100
2,000
9,200
54,000
25,600
91,700
$1,559300
608,500
$2,167,800
035"
'Unit treatment costs are estimates based on preliminary design data and neglects factors such as facility location, delivery system, etc.
The cost estimates are intended to represent order-of-magnitude (+50 and -30 percent) estimates. These costs will change during
detailed design.
Tjquid phase carbon adsorption is not economically feasible in the marimum credible deviation case due to high carbon loading.
Costs shown are for air stripping.
403
LAO22033\231 004.51-21
-------
Using the above conditions and additional data for the capital cost of ozone generation
equipment, a preliminary design for ozone/peroxide treatment of Whittier Narrows
groundwater was developed. The following conditions were used to develop treatment
costs:
Number of reaction chambers
Volume of reaction chambers (gal)
Outer dimension (ft, including concrete thickness)
Ozone demand (Ib/day)
Total installed ozone generating capacity (Ib/day)
Peroxide demand (Ib/day)
Probable
Design
Case
3
40,000
15 x 43 x 15
534
1,300
190
Maximum
Credible
Deviation
Case
3
49,000
15 x 49 x 15
1,150
2,300
407
Note that the ozone and peroxide usages per day are based on an average flow rate.
Usage would be higher during peak demand.
Table 5 summarizes the cost estimate for ozone/peroxide advanced oxidation as
estimated for the Whittier Narrows OUFS. These costs are estimates based on pre-
liminary design data and reglect such factors as facility location, delivery system require-
ments, etc. The cost estimates are intended to represent order-of-magnitude (+50 and -
30 percent) estimates. These costs are expected to change during detailed remedial
design.
As shown in Table 5, the cost of treating Whittier Narrows groundwater by advanced
oxidation in the probable design case is estimated at $0.25 per 1,000 gallons treated and
$0.25 per 1,000 gallons for the maximum credible deviation case. As mentioned
previously, these costs are based on bench-scale data collected using a batch-type
reactor. Actual pilot evaluation of a flow-through system is required to more accurately
quantify the costs of treating groundwater.
The advanced oxidation process has not yet been applied at the large scale required for
the Whittier Narrows OUFS (e.g., 21 to 25.5 million gallons per day [MOD]). Although
the cost estimates are based on sound engineering principles, the actual costs
experienced during startup of such a large system are likely to be higher than predicted
using accepted cost estimating techniques. Such costs were not accounted for in this
study because they are as yet unquantifiable.
404
LAO22033\231 008.51
-------
Table 5
Treatment Costs Using Advanced Oxidation
CAPITAL COST ESTIMATES
Reactors
Ozone Generation System
Peroxide System
Building
Electrical and Instrumentation
Equipment Cost Subtotal
Bid Contingencies @ 15%
Scope Contingencies @ 25%
Construction Total
Service During Construction
@10%
Land Acquisition
Total Implementation Cost
Engineering, Legal, and
Administrative Costs @ 22%
TOTAL CAPITAL COST
Unit Cost (VTJnlt)
339,000 (Design)
401,500 (CredXtev.)
2,193,500 (Design)
3,193,000 (Cred.Dev.)
30,000 (Design)
42,000 (Cred.Dev.)
120
579,800 (Design)
859,500 (Cred.Dev.)
_
-
_
_
-
300,000
_
-
-
Units Required
Probable Design
lea
1 ea
lea
5300ft2
lea
H
_
„
_
-
2 acres
_
-
-
Maximum
Credible
Deviation
lea
lea
lea
S^OOft2
lea
..
_
_
»»
-
2 acres
H
-
-
Treatment Facility Coct(s)
Probable
Design
$339,000
2,193,500
30,000
636,000
579,800
$3,778300
566,700
944,600
$5,289,600
529,000
600,000
$6,418,600
1,412,100
$7,830,700
MBtam
Credible
Deviation
$ 401,500
3,193,000
42,000
996,000
859,500
$5,490,000
823,500
1372400
$7,686,000
768,600
600,000
$ 9,054,600
1,992,000
$ 11,046,600
LAO22033\231 00541-23
405
-------
Table 5
(Continued)
O&M COST ESTIMATES
Electrical
Peroxide
Water Sample Analysis
Air Sample Analysis
Operating Labor
Maintenance @ 2% of Capital
Costs
TOTAL O&M COSTS
Annualized Cost of Capital
Total Annul Costs
UNIT COST FOR TREATMENT
(S/1,000 gal)"
Unit Cost ($/Unlt)
0.07
1.10
110
250
35
-
H
_
H
-
Units Required
Probable Design
1,543,400 kW-hr
69,000 Ib
48 ea
36 ea
4,400
-
H
_
_
-
Maximum
Credible
Deviation
5,261,900 kW-hr
148,700 Ib
48 ea
36 ea
4,400
-
_
_
—
-
Treatment Facility Coat(s)
Probable
Design
$108,000
75,900
5300
9,000
154,000
75,600
$427,800
509,800
$937,600
0.25
Maximum
Credible
Deviation
$368400
163,600
5,300
9,000
154,000
109,800
$ 810,000
719,100
$1,529,100
0.25
aUnit treatment costs are estimates based on preliminary design data and neglects factors such as facility location, distribution system
requirements, etc. The cost estimates are intended to represent order-of-magnitude (+50 and -30 percent) estimates. These costs will
change during detailed design.
406
LAO22033\231_OOS.Sl-24
-------
In addition, it is assumed that additional treatment (e.g., filtration) is not necessary to
control by-products such as assimilable organic carbons (AOC) or other ozone
byproducts.
Finally, during the San Gabriel bench-scale testing, the advanced oxidation process
proved infeasible for the treatment of carbon tetrachloride due to the observed stripping
effect prior to chemical oxidation. This effect may only be a function of the con-
figuration of the bench-scale equipment. However, if pilot- or full-scale equipment
exhibit a similar phenomena, pre- or postoxidation treatment may be necessary to control
the CTC. This will result in increased treatment costs.
COMPARISON OF TREATMENT COSTS
Table 6 summarizes the capital, O&M, and unit costs for treatment of groundwater using
each of the three technologies discussed earlier. Based on review of this summary table,
air stripping with vapor phase GAC may be most economical for the probable design
case. However, advanced oxidation may be most economical of each of the treatment
technology options under maximum credible deviation conditions.
Table 6
Summary Comparison of Treatment Costs8
Cost in $1,000 Gallons
Air Stripping
Carbon Adsorption
Advanced Oxidation
Probable Design
Case
0.19
0.43
0.25
Maximum Credible
Deviation Case
0.35
__b
0.25
"These costs are estimates based on preliminary design data and neglect such
factors as facility location, delivery system requirements, etc. The cost
estimates are intended to represent order-of-magnitude (+50 and -30 percent)
estimates. These costs are expected to change during detailed remedial design.
bLiquid phase carbon adsorption is not economically feasible in the maximum
credible deviation case due to high carbon loading.
As mentioned above, eight additional criteria must be considered in choosing a remedial
alternative for the preparation of a feasibility study for an EPA NPL site. Some of these
criteria include:
• Reliability
• Ease of operation
LAO22033\231 008.51
407
-------
• Expected lifetime
• Ease of maintenance
• Presence of process hazards
When these criteria are evaluated alongside costs, the selection of a remedial treatment
technology becomes more subjective. For example, although advanced oxidation appears
less costly (for the maximum credible deviation case), each of the above criteria are
much more easily met by air stripping and carbon adsorption. Additionally, the presence
of gaseous ozone and liquid hydrogen peroxide onsite creates additional safety concerns.
All of these concerns must be carefully weighed prior to selection of the appropriate
technology for groundwater treatment.
408
LA022033\231 008.51
-------
REFERENCES
U.S. Environmental Protection Agency. National Interim Primary Drinking Water
Regulations; Control of Trihalomethanes in Drinking Water. Federal Register
44(231):68624 November 29, 1979.
Hoigne, J. and H. Bader. Rates Constants of Reactions of Ozone with Organic and
Inorganic Compounds in Water. I. Non-Disassociating Organic Compounds. Water
Resources 17 (1983): 173.
Hoigne, J. and H. Bader. Rates Constants of Reactions of Ozone with Organic and
Inorganic Compounds in Water. II. Disassociating Organic Compounds. Water
Resources 17 (1983): 185.
Farhataziz, P. C. and A. B. Ross. Selective Specific Rates of Reactions of Transients in
Water and Aqueous Solutions. III. Hydroxyl Radical and Perhydroxyl Radical and
Their Radical Ions. National Standard Reference Data Service 59, U.S. National
Bureau of Standards. 1977.
Glaze, W. H. and J. W. Kang. Advanced Oxidation Processes for Treating Ground-
water Contaminated with TCE and PCE: Laboratory Studies. JAWWA 80 (1988):51.
U.S. Environmental Protection Agency. Agency Review Draft Technical Memorandum,
San Gabriel Basin, Results of Ozone/Peroxide Bench-Scale Treatability Test. Prepared
by CH2M HILL January 10, 1989.
409
LAO22033\231 009.51
-------
RECOVERY OF METALS FROM WATER
USING ION EXCHANGE
A CASE STUDY
PREPARED FOR:
SECOND FORUM ON
INNOVATIVE HAZARDOUS WASTE
TREATMENT TECHNOLOGIES
PREPARED BY:
B&V WASTE SCIENCE AND TECHNOLOGY CORP.
MAY 1990
410
-------
RECOVERY OF METALS FROM WATER
USING ION EXCHANGE
A CASE STUDY
By
Thomas A. Hickey, B&V Waste Science and Technology Corp.
David K. Stevens, Pritchard Corporation
Since 1984, B&V Waste Science and Technology Corp. has been assisting an
industrial client in the remediation of a wood treating facility in
northern California. The wood treating facility consists of a 15-acre
paved wood treating and storage yard, the wood treating facilities
themselves, and a storm water retention pond which was recently closed.
Because of past operating practices, wood treating chemicals have entered
into the soil and ground water at the facility. Wood treating chemicals
have also entered storm water by rinsing from freshly treated wood that is
stored on the yard.
The wood treating chemicals used by the client are inorganic metals,
chromium, copper, and arsenic. When wood treatment operations began at the
site in 1966, a chromated copper arsenate, or CCA, solution was used, which
was a blend of sodium dichromate, copper sulfate, and arsenic acid. The
ratio of the metals in the solution was about 60 percent chromium, 25
percent copper, and 15 percent arsenic. In 1982, the solution was modified
to use only sodium dichromate and copper sulfate in a solution called acid
411
-------
copper chromate, or ACC. The metal ratio in this solution is about 60
percent chromium and 40 percent copper.
The wood treating chemicals each have different characteristics. The
Safe Drinking Water Act criteria limit the total chromium in potable water
to 50 parts per billion (ppb), although recent recommendations call for
revising the maximum to 100 ppb. At the wood treating facility, chromium
is found
in two valence states: as trivalent chromium and hexavalent chromium.
Trivalent chromium generally acts as a cation; hexavalent chromium is
present as chromate8 or dichromates and acts as an anion. The solubility of
trivalent chromium species is very low at a pH greater than 5. Hexavalent
chromium species are quite soluble in the normal pH range of ground water.
Arsenic commonly occurs as the anion arsenate. The maximum contaminant
level (MCL) established by the the U.S. Environmental Protection Agency for
arsenic in potable water is 50 ppb.
Copper is only found as a cation. The oxidized state, cupric, is copper's
more stable form and has a valence of +2. A secondary standard in the Safe
Drinking Water Act limits copper to 1.0 ppm.
Site investigations revealed that the soil and storm water runoff contained
elevated concentrations of all three metals, but only hexavalent chromium
had migrated in detectable quantities to the ground water. These findings
412
-------
are consistent with the chemical characteristics of the metals and the site
soil. A number of previous studies have documented that common clays, such
as montmorillonite and bentonite, act as strong adsorbants for arsenic,
copper, and trivalent chromium compounds. The highly soluble hexavalent
chromium compounds are very mobile in soil, however, and are readily
transported through soils with normal water penetration.
Hexavalent chromium concentrations in the ground water vary across the site
from not detectable in the outer regions of the affected area to greater
than 50 ppm directly beneath the wood treating facilities. Ground water
extracted from a variety of locations in the affected area and blended
exhibits relatively consistent characteristics. The ground water does not
contain any of the other wood treating chemicals, and it has a low
concentration of suspended solids. We have estimated that an extraction
rate in excess of 300 gallons per minute (gpm) or 160 million gallons per
year is needed to control contaminant migration and accomplish ground water
remediation in a reasonable time frame.
Approximately six million gallons of precipitation fall annually on the
wood treating facility's paved yard. All storm water runoff is contained
on the yard and treated. Storm water runoff accounts for approximately
five percent of the potential water volume treated through the treatment
plant. The characteristics of storm water are much less consistent than
those of ground water. The concentrations of all three wood treating
chemicals vary with the runoff patterns of each storm as well as the timing
of the storm during the winter, or rainy, season. The first rains of the
413
-------
season carry much higher loads of the contaminants than later storms.
Suspended solids and oil and grease loads are also present in storm water
runoff and vary significantly. A treatment rate of approximately 200 gpm
is needed to limit containment of storm water on the yard after major
storms to a period of 2 or 3 days. During that period, the storm water is
treated and discharged.
In 1987, the client began operating an ion exchange water treatment plant
that our firm designed. We selected ion exchange technology for the water
treatment plant because it allows removing of the wood treating chemicals
from storm water and extracted ground water in a manner that enables
recovery and reuse of the chemicals in the wood treating operations. The
ion exchange water treatment plant allowed the client to fulfill the basic
intentions of RCRA: to conserve and recover resources.
Other major design objectives included:
o Process and equipment flexibility to treat both storm water and
ground water influents.
o Process performance to meet stringent effluent requirements.
o Cost optimization and operator time minimization.
414
-------
Several factors allowed implementing the ion exchange technology for this
application, including:
o The client's philosophy encouraged waste minimization and
resource recovery.
o The client understood the technology and had experience with
similar programmable controller-driven equipment.
o Ongoing wood treating operations at the facility presented the
opportunity.
o The characteristics of wood treating chemicals in the waters were
suitable for the technology.
Our remedial design work began with an evaluation of the technologies
available for removing metals from contaminated water. The technology
evaluation focused on hexavalent chromium removal processes for ground
water treatment with secondary emphasis on removing the other metals from
storm water.
The chromium removal processes that were considered included chemical
reduction and precipitation, proprietary processes using electrochemical
co-precipitation, and co-precipitation with iron. These processes produce
a sludge that is unsuitable for reuse in the wood treating process and
requires landfill disposal.
415
-------
Ion exchange was selected as the most desirable technology to meet our
client's objectives. We looked for the following characteristics when
selecting resins for the application:
o High selectivity for the chromium species present.
o High capacity for the chromate radical at normal pH.
o Efficient regeneration to achieve the highest possible
concentration of metals for reuse in the wood treating facility.
o Regeneration in a medium suitable for reuse in wood treatment.
o Physical stability for long bed life.
A considerable amount of research was conducted in the 1970's on the
removal of chromates from cooling tower discharges, which may contain up to
20 ppm of chromium. This work showed that styrene resins are highly
selective for the chromate radical compared to acrylic resins.
Weak base resins exhibited the other desirable properties listed. To
establish the effectiveness and economics of the process, bench-scale and
pilot tests were conducted. Various weak base styrene resins were tested
for the following parameters:
o Chromate capacity measured in bed volumes treated against
416
-------
hexavalent chromium leakage concentration in the effluent.
o Concentration of chromium in the regeneration solution by
determination of the elution curve.
o Effect of inlet pH on the capacity of the resin.
These studies confirmed that satisfactory performance was achieved with a
weak base anion resin in the sulfate form. The optimum inlet pH to
maintain the sulfate resin form was found to be 4.5, and the most efficient
regeneration rate for chromium concentration was achieved with 2 to 2-1/2
bed volumes of 3 percent caustic soda solution.
After elution with the caustic solution, the resin is returned to the
sulfate form with a sulfuric acid conditioning step. In this step, the
hydroxyl ions supplied by the caustic are in turn replaced by sulfate ions.
The sulfate radical is readily displaced by anionic chromate.
A small amount of hexavalent chromium was found to reduce to trivalent
state by oxidation of the resin. For this reason, as well as to remove the
lower concentrations of trivalent chromium and copper present in storm
water, a cation bed was installed after the anion beds. Through a series
of steps similar to those described in the selection of the anion resin, a
weak acid cation resin was selected. Figure 1 illustrates the system
selected. Regeneration of the cation bed is accomplished with one bed
volume of 5 percent sulfuric acid.
417
-------
WATER TREATMENT PLANT
ION EXCHANGE SERVICE
Water from _
Pre-lreatment
System
_H2SO4
Pri
Anion
NaOH
Sec
Anion
Cation
To Discharge
NaOH
-------
The final effluent water from the cation exchanger is adjusted to near
neutral pH with caustic soda for discharge. The regeneration streams from
the anion and cation exchange columns are combined in the recovery tank.
The resulting mixed metal concentrate is returned to the wood treatment
process by incorporating it into the treating solution, as illustrated in
Figure 2.
The ion exchange system resins selected are more suitable for treating
waters with low contaminant load, and cannot tolerate significant organic
or solids loadings. To protect the resins, we installed a multifunctional
pre-treatment system, as shown in Figure 3. The pre-treatment system
consisted of chemical feed, coagulation and flocculation chambers, settling
basin, flow equalization basin, strainer, and banks of filters. Because
storm water has the greatest variety of extraneous materials, it is treated
through all of the pre-treatment components, which allows for removing
suspended solids, some organics, and arsenic. We found that the arsenate
radical was not adsorbed by the anion resin, so separate treatment of the
surface water containing arsenic was required. Ferric chloride addition is
a successful flocculant for arsenic removal. Ground water is processed
only through the strainer and filters before ion exchange. Solids from the
settler, strainer, and backwashable filters is held in sludge drying beds
until they are disposed of offsite.
The completed ion exchange water treatment plant has been in operation for
three years and has produced water in compliance with the NPDES permit
conditions even though effluent limits were significantly reduced after
419
-------
WATER TREATMENT PLANT
ION EXCHANGE REGENERATION
ro
o
Pri
Anion
Cr'
Recovery Tank
NaOH
Sec
Anion
H2SO4
Cation
Cu + Cr+3
To Wood Treating
Chemical Storage Tanks
FIGURE 2
-------
WATER TREATMENT PLANT
PRE-TREATMENT SYSTEM
Groundwater
Stormwater
r\>
Pre-Treated Water
to Ion Exchange
Train
1
Flow Equalization
Lamella Separator
Sludge to Drying Beds
Cartridge Filters
Bag Filters
Automatic
Strainer
FIGURE 3
-------
plant design and construction were completed. For example, the original
hexavalent chromium effluent limit was set at SO ppb. As part of NPDES
permit revisions, the limit was reduced to 11 ppb. The effects of this
revision are discussed in detail later in this paper. Other major findings
during the past three years of operation are summarized as follows. Each
of these problems is described in detail in following sections.
o The resins have been fouled by iron and organics despite
pretreatment and other precautions.
o The breakthrough of hexavalent chromium from the secondary anion
bed has limited treatment plant run lengths between regeneration
cycles.
o Normally, anion and cation exchange resins have been fully
regenerated simultaneously, and the use of regenerant streams in
the wood treating process has been successful.
o The balance between the supply of recovered chemicals and the
demand for the chemicals and makeup water in the wood
treating facility is critical to successful operation.
Figure 4 illustrates the behavior of the water treatment plant in terms of
effluent hexavalent chromium as a function of gallons of water treated.
The pH of the influent water has been found to affect the chromium capacity
of the anion resins and the duration of the runs before breakthrough.
422
-------
WATER TREATMENT PLANT
PERFORMANCE DATA
ro
ca
Effluent
Concentration
(ppb)
/• urigmai urn
Pri Anion Effluent
Sec Anion Effluent
Reduced Limit
Gallons Ireated (x106)
Primary Anion Inlet pH @ 4.5
Primary Anion Inlet pH @ 7.5
FIGURE 4
-------
Under the design condition of influent pH control to 4.5, chromium
breakthrough occurs with a gradually increasing effluent concentration, or
a "slow leak". This operation method maximized throughput before the
original effluent limit of 50 ppb was achieved. Under normal pH or
uncontrolled pH, chromium breakthrough concentrations are very low
initially and rise sharply at the end of the run. This behavior allows for
maximizing throughput against the revised limit of 11 ppb. Sengupta [5]
has reported similar results and has shown that the effect of pH is a
result of the hexavalent chromium equilibrium characteristics, not the
result of mass transfer effects.
We have been able to combine the effects of pH control to optimize run
lengths by operating the beds under "hybrid pH control". In this operating
mode, the primary anion bed operates under acidic conditions and secondary
anion bed operates under near neutral conditions. The result allows the
primary anion to generate a "slow leak" of hexavalent chromium and the
secondary bed to remove the "slow leak" under favorable equilibrium
conditions. The total hexavalent chromium retention capacity is greater in
this mode of operation than it would be for the current effluent limit, as
indicated in Figure 5.
Resin fouling has primarily resulted from iron carry-over from the
pre-treatment facility and organic fouling from materials in storm waters.
Resin fouling has been evidenced by excessive pressure losses through resin
beds that have not been reduced by backwashing and by low capacity for
wood treating chemicals on the resin. The following steps have been taken
424
-------
WATER TREATMENT PLANT
PERFORMANCE DATA
Original Limit
+6
Effluent Cr
Concentration
(ppb)
40-
3d-i
20-
10-
r
Pri Anion Effluent
Sec Anion Effluent
w
t / Reduced Limit
0.5
I
1.0
i
1.5
6
Gallons Treated (xlO )
Primary Anion Inlet pH @ 4.5
Primary Anion Inlet pH @ 7.5
FIGURE 5
-------
to alleviate the problems associated with resin fouling:
o Periodic resin samples have been taken from the resin beds and
submitted to the resin manufacturer for analysis.
o The pretreatment systems have been operated properly to minimize
the potential of iron carry-over and to maximize removal of other
deleterious materials in the waters.
o Periodic surfactant and brine-caustic cleanings have restored
resin activity and performance.
The balance between the "supply" and the "demand* for recovered chemicals
has, at times, limited the effectiveness of resource recovery. When the
demand is low, during periods of limited wood treatment, water treatment
has been curtailed. Obviously, this seriously affects the effectiveness
of ground water extraction for site remediation. To counter this effect,
we have optimized the recovery process by increasing water treatment
plant runs between regeneration cycles and by minimizing the amount of
regenerant produced during each regeneration.
The major methods employed to maximize treatment plant runs are those
described previously: pH control for hybrid operation and resin
maintenance for maximum capacity. We have also explored utilizing detailed
aquatic toxicity testing to modify the effluent limits currently imposed on
the plant's performance. Preliminary investigations indicate that effluent
426
-------
limits could be raised without increasing aquatic toxicity to the most
sensitive organisms.
Several methods have been used to reduce the amount of regenerant produced
during each regeneration. Careful timing and control of the regenerant
steps have produced marginal reductions in regenerant volumes. Recycling
of certain chemical rinse steps has been implemented to reuse lightly
contaminated caustic solutions as the first elution step in the subsequent
regeneration cycle.
In summary, hexavalent chromium and other soluble metals can be
successfully recovered in a condition suitable for reuse. We have seen
that this recovery is effective and successful when the wastewater
characteristics are suitable for the resins selected, the operator commits
to operation and maintenance practices to keep the resin and system in
optimum condition and balance, and the "demand" equals or exceeds the
"supply" of the recovered chemicals. The ion exchange process was applied
to produce a usable regenerant concentrate, and it was economical when
concentrating to 50 times the extracted ground water contamination level.
427
-------
REFERENCES
1. Bartlett, R. J. and Kimble, J. M., "Behavior of Chromium in Soils",
Journal of Environmental Quality: Volume 5, Ho. 4, 1976 and Volume 8,
No. 1, 1979.
2. Lang's Handbook of Chemistry, Ed. VII.
3. U.S. Environmental Protection Agency: "Hazardous Waste Land
Treatment", Office of Solid Waste, SW-274, April 1983.
4. Sorg, Thomas J. and Logsdon, Gary S., "Treatment Technology to Meet the
Interim Primary Drinking Water Regulations for Inorganics", Journal of
the American Water Works Association: July 1978.
5. Sengupta, A. and Clifford, D., "Some Unique Characteristics of Chromate
Ion Exchange", Reactive Polymers, 4 (1986) 113-130, Elsevier
Scientific Publishers; Amsterdam.
6. Kumin, Robert, "New Technology for the Recovery of Chromates from
Cooling Tower Slowdown", Amber Hi-Lites, No. 151, Rohm & Haas Co.
publication, May 1976.
428
-------
SECOND FORUM ON
INNOVATIVE HAZARDOUS WASTE
TREATMENT TECHNOLOGIES:
DOMESTIC AND INTERNATIONAL
CRITICAL FLUID SOLVENT EXTRACTION
Presented May 17,1990
Authors:
Cynthia Kaleri, EPA, Region VI, Dallas, Texas
Louis Rogers, Texas Water Commission, Austin, Texas
Calvin Spencer, Roy F. Weston, Houston, Texas
429
-------
TABLE OF CONTENTS
SECTION DESCRIPTION PAGE
1.0 INTRODUCTION 1-1
2.0 SITE DESCRIPTION AND HISTORY 2-1
3.0 TECHNOLOGY DESCRIPTION
3.1 Technology Theory 3-1
3.2 Simplified Critical Fluid
Solvent Extraction 3-1
3.3 Propane Solvent Extraction Process 3-4
3.4 Treatment 3-7
3.5 Characterization 3-7
3.6 Sample Preparation 3-8
4.0 RESULTS
4.1 Treatment 4-1
4.2 Material Handling Problems 4-1
5.0 TREATABILITY CONSIDERATIONS 5-1
430
-------
SECTION 1.0
INTRODUCTION
As mandated by the Superfund Amendments and Reauthorization Act of
1986 (SARA) and in accordance with the National Contingency Plan
(NCP), EPA must consider several factors in selecting remedial
action for a superfund site. Overall protection of human health
and the environment and consistency with other environmental laws
are primary criteria for selection of a cost effective response
action. However, the statutory preference for treatment in
permanently and significantly reducing toxicity, mobility, or
volume must also be considered and balanced accordingly. In
addition, state and community acceptance influence the selection
of remediation at a Superfund site.
EPA realizes the need for less costly, effective treatment
technologies to manage hazardous waste sites. In response to this
need, EPA established the Superfund Innovative Technology Evalua-
tion (SITE) Program in 1986 to promote the development and use of
new or innovative treatment technologies to clean up Superfund
sites across the country. The objective of the demonstration
portion of the SITE program is to develop reliable performance and
cost information of the technologies selected so that they can be
adequately considered in the Superfund decision making process.
Although the SITE program identifies the feasibility of an
innovative technology for particular types of waste, the need to
evaluate and refine a technology for site specific application is
a practical necessity for selection of the most appropriate remedy
for a site. EPA has encouraged the use of treatability studies
during the early phases of the Superfund investigative process in
order to effectively evaluate the use of innovative technologies
at individual Superfund sites.
This paper discusses the results of a treatability study performed
for the United Creosoting Superfund site and the evaluation of
those results in relation to the remedy selection process.
Although Critical Fluid Extraction is an innovative technology
under the SITE program, the demonstration of the pilot scale unit
at this site was carried out through a separate contract between
the vendor and Roy F. Weston, Incorporated (WESTON). WESTON was
selected by the Texas Water Commission [TWC] to conduct the
treatability studies and incorporate the results into an Amended
Feasibility Study Report.1
1 An abbreviated site history from the 1989 Record of
Decision [ROD] is presented prior to the technology
description and results discussion. The 1989 ROD should
be consulted for in-depth discussions in all areas to be
presented in this paper.
admin:crflpapr.425 431
-------
SECTION 2.0
BITE DESCRIPTION AND HISTORY
The United Creosoting site is located 40 miles north of Houston in
the City of Conroe, Montgomery County, Texas [Figure 2-1]. Bound
on the west and south by Alligator Creek, on the north by Dolores
Street, and on the east by the Missouri-Pacific rail lines, the
property is approximately one hundred acres in size.
The United Creosoting Company operated as a wood preserving
facility from 1946 through the summer of 1972. With the exception
of the process building where timber was debarked and cut to the
desired product, the process areas became scarred by an accumula-
tion of the black oily chemicals used for treating the lumber.
Historical aerial photographs and analytical data obtained to date
have been utilized to describe the process areas as they existed
during active operations.
Formed lumber, such as telephone poles and railroad ties, were
treated in a two-step process by the pressurized addition of
pentachlorophenol [PCP] and creosote. The pressure cylinders were
rinsed and the wastewater routed to one of the two process waste
ponds located onsite. Segregation of the two waste streams allowed
possible reclamation and reuse. The larger pond held mainly the
creosote waste and the smaller pond the PCP process waste.
No evidence exists that PCP was produced onsite. However, PCP was
stored in one or more of the storage tanks onsite. Creosote was
produced via a coal tar distillation unit onsite and stored in
lined pits just east of the process waste ponds. Creosote and
other distillate fractions of coal tar included polycyclic aromatic
hydrocarbons [PAHs] of varying molecular weights. Coal tar pitch,
a dark brown to black amorphous residue, was an unusable byproduct
which was apparently disposed of in the larger process waste pond.
The physical characteristics of the site have been altered by
redevelopment of the property, which has resulted in residential
and light industrial structures typical of suburban settings.
Other residential areas surround the site to the immediate north,
west and south, while industrial and commercial land uses are
evident to the east. Approximately 13,000 people currently live
within a two-mile radius of the site.
The Texas Water Commission [TWC] submitted the United Creosoting
site as a candidate for cleanup under the Superfund program in
August 1982. The immediate concern at that time was contaminated
surface water runoff flowing from the former waste ponds area into
Tanglewood East Subdivision, the residential portion of the site.
TWC collected additional soil, water, and air samples from the site
and in September 1983, the United Creosoting site was included on
the proposed National Priorities List by EPA and thus became
eligible for remedial funding.
432
-------
433
-------
In early December 1983, EPA initiated an immediate response action
at United Creosoting. Surficial soils samples taken in the
vicinity of the former waste ponds and within the Tanglewood East
subdivision revealed PCP contamination and chlorinated dioxins and
dibenzofurans, trace byproducts of commercial grade PCP. It was
suspected that the source of the contamination might be storm water
runoff from former waste pond areas located on the commercial
portion of the site. Therefore, exposed sections of contaminated
soils in the former waste ponds area were regraded and covered with
a synthetic membrane and at least 6 inches of compacted clay.
Access to the cap area was restricted by fencing the area and
constructing drainage ditches to channel cap area runoff away from
the subdivision.
The Remedial Investigation/Feasibility Study [RI/FS] was completed
in May 1986. Although incineration was considered a feasible
alternative, no commercial facilities were available at that time
to accept dioxin contaminated soils; the residents opposed incin-
eration so close to their homes. In August 1986, EPA formally
proposed natural attenuation of the groundwater and a temporary
consolidation and capping remedy for contaminated soils at the
site. EPA further specified that innovative technologies would be
evaluated in the next 5 years for permanent soils remediation in
order to mitigate further groundwater contamination. Following
formal public comment, EPA selected the proposed remedial action.
In March 1987, two treatability studies were initiated to evaluate
innovative technologies as possible remedies for soils at the site.
These treatability studies involved biological treatment and
critical fluid extraction. This paper will focus on the critical
fluid extraction technology treatability results and the evaluation
of these results in relation to the remedy selection criteria in
place at the signing of the September 1989 Record of Decision.
admin:erfIpapr.425
-------
SECTION 3.0
TECHNOLOGY DESCRIPTION
3.1 Technology Theory
Solvent extraction, at its simplest, is the separation of the
constituents of a liquid solution by contact with another insoluble
liquid. If the constituents in the original solution distribute
differently between the two liquid phases, a certain degree of
separation will result. In such an operation, the original
solution is called the feed, and the liquid with which the feed is
contacted is called the solvent. The solvent-rich product of the
operation is called the extract and the residual liquid from which
the solute is removed is called the raffinate.
In the case of Critical Fluid Solvent Extraction, the solvent is
a pure liquid whose temperature and pressure are controlled at or
near the materials' critical point or state. At the critical point
(Point C on Figure 3-1), distinctions between the materials' liquid
or vapor phase disappears and the properties of the liquid, such
as density, viscosity, refractive index, etc., are identical to
those of the vapors.
Figure 3-1 is a simplified vapor-pressure curve which shows phase
differences (solid, liquid, gas), the critical point (C) with
corresponding critical pressure and critical temperature, boiling
point temperature, and the triple point (T). The triple point is
that physical state at which all three phases coexist.
For the United Creosoting site, a technology using propane as the
solvent was selected. Propane is an excellent solvent because it
exhibits an excellent solubility of hydrocarbons, is inexpensive,
is readily compressed, and if released, does not pose a significant
environmental hazard.
Although a relatively new technology for hazardous waste treatment,
propane solvent extraction was used by the petroleum refining
industry in the late 1940's and early 1950's for the de-asphalting
of lube oil feed stocks. Since the early 1970's fundamental
research in critical fluid extraction has been ongoing. The
technology presented below is a product of this ongoing research.
3.2 Simplified Critical Fluid Solvent Extraction
As shown in Figure 3-2, a liquid or slurried feed such as an
organic-containing hazardous waste is admitted to an extractor
along with the solvent at or near the solvent critical point and
the organics in the waste dissolve into the solvent. Extracted
organics are removed with solvent from the top of the vessel, while
clean water and solids exit through the bottom. The extract then
435
-------
Pcrlt
GAS
tnbp Wit
TEMPERATURE
FIGURE 3-1
VAPOR PRESSURE OF
A PURE UQUID
436
-------
FEED
^^—
EXTRACT T
SE
CC
EXTRACTOR
4 i
SOLVENT
:PARATOR
)LUMN
r
mt^m
*\
^.
SOLVENT
MAKEUP
COMPRESSOR
EXTRACT
RAFFINATE
FIGURE 3-2
SIMPLIFIED FLOW CHART
OENERMjGeNERU7.TV i-J-SO 1:1
437
-------
goes to a second vessel, where the temperature and pressure are
decreased, causing the organic to separate from the solvent. Clean
solvent is recycled to the extractor, and the concentrated organics
are recovered from the bottom of the separator.
3.3 Propane Solvent Extraction Process
The treatment process for the extraction of organics using propane
from soils and sludges is outlined in Figures 3-3 and 3-4.
Although this represents a commercial process, the pilot unit used
at United Creosoting was operated to simulate this process. Figure
3-3 illustrates the extraction of hydrocarbons from soil and
sludges and Figure 3-4 outlines the solvent recovery system.
Extraction
The Extraction Unit operates on a continuous basis in a countercur-
rent manner. Feed is processed in a series of extractors/separa-
tors. Maximum extraction efficiency is achieved by feeding the
fresh solvent to the last extraction stage and operating with
solvent flowing in a countercurrent mode. The following is a
description of a three stage extraction unit. The number of
extraction stages is decreased or increased to meet the require-
ments of the specific application.
Slurry from the feed surge drum (D-101) and the solvent/organics
steam from the second-stage decanter (D-103) are combined in the
first-state extractor (XT-101). After thorough mixing, the
resulting stream is sent to the first-stage decanter (D-102).
Substantial separation of the solvent/organics solution from the
water/solids residue occurs in this decanter, and a concentrated
solvent/organics solution exits overhead to the decanter/coalescer
(D-105) in the solvent recovery section. The water/solids residue
from the first decanter (D-102), which contains some residual
organics, and the solvent/organics solution from the third-stage
decanter (D-104) are combined and mixed in the second-stage
extractor (XT-102).
Separation of the water/solids residue and the solvent/organics
solution from the second-stage extractor (XT-102) occurs in the
second-stage decanter (D-103); the solvent/organics solution is
pumped to the first-stage extractor (XT-101) and the water/solids
residue is sent to the third-stage extractor (XT-103). The
water/solids residue from the second-stage contains water and
solids with some residual organics. Fresh solvent and the
water/solids residue from the second-stage decanter (D-103) are
combined in the third-stage extractor (XT-103). Final separation
of solvent/organics solution and water/solids residue occurs in the
third-stage separator (D-104).
Water/solids residue from the third-stage decanter (D-104),
consisting of water and solids, together with water from the
decanter/coalescer (D-105), are sent to the dewatering system. If
needed, a portion of the water recovered in the dewatering system
438
-------
CO
XT-101
FIRST STAGE
EXTRACTOR
XT-102
SECOND STAGE
EXTRACTOR
XT-103
THIRD STAGE
EXTRACTOR
D-101
FEED SURGE DRUM
D-102
FIRST STAGE
DECANTER
D-103
SECOND STAGE
DECANTER
D-104
THIRD STAGE
DECANTER
EXTRACT TO D-105
« CLEAN SOLVENT FROM P-104
SLURRIED
FEED
P-101
P-101
MAIN FEED PUMP
P-102
FIRST INTERSTAGE
PRESSURE BOOSTER
PUMP
P-103
E-103
SECOND INTERSTAGE
PRESSURE BOOSTER
PUMP
« WATER FROM
D-105
WATER
#• SOLIDS
FIGURE 3-3
PROCESS FLOW DIAGRAM
EXTRACTION SECTION
CENBMUCENEIU38.JP1V B-2-W t:t
-------
D-105 T-101 C-101 C-102 T-102
DEACANTER/ PRIMARY SOLVENT PRIMARY SECONDARY SECONDARY SOLVENT
COALESCER RECOVERY STILL COMPRESSOR COMPRESSOR RECOVERY STILL
E-101
PRIMARY STILL
REBOILER HEAT * "~ " ' " > '
EXCHANGER
D-105 TT-IOI I
. . . . s^~-\ r-*. ^
f X N 1
TYTPAPT ... . , M , l~ ....- .-*
/r-ooki ri •i/T'A v - - -• ' ^ ? \^«^ •***.
^rKUM U— IVZ) ^<^ "^y ^^
C-101 1 C-102
WATER TO n^>E-101
nrwATrmisir < \\^ j *>
SYSTEM
^
<• _. k
T-102
^
D-106 ^^
SOLVENT ^^
Em^
IUj
i
^TO XT-103) ( > *
r~\
P-104- P-105
J " E-102
., k .., ^ nit
-y
P-105
E-102
SECONDARY STILL SECONDARY STILL
CIRCULATING PUMP HEAT EXCHANGER
— IU4 U— lUu L— 1UO
SOLVENT RECYCLE SOLVENT SURGE SOLVENT
PUMP VESSEL SUBCOOLER/
CONDENSER
RGURE 3-4
MATERIAL BALANCE
SAMPLE 1
-------
is recycled as slurrying liquid. Effluent water is sewered or sent
to an on-site wastewater treatment facility.
Solvent Recovery
The decanter/coalescer (D-105) removes entrained water from the
solvent/organics solution stream prior to the primary solvent-
recovery still (T-101) . The recovered water is sent to the
dewatering system.
The water-free extract stream is sent to the primary solvent-
recovery still (T-101) where the solvent is stripped from the
organics. Vaporized solvent goes overhead while a highly con-
centrated organics-solvent mixture is withdrawn from the bottom of
the still and sent to the secondary solvent-recovery still (T-
102). The remaining solvent is stripped from the organics in the
secondary still, and the recovered organics are pumped from the
still bottom to the Client's storage facilities (outside CFS1
battery limits).
Vaporized solvent from the secondary still is compressed to
primary-still pressure, combined with the vapor from the primary
still, and further compressed to extractor pressure. The high-
pressure vapor is then used as the heat source for the primary-
still reboiler (E-101), where it is partially condensed. Complete
condensation is achieved in the condenser/subcooler (E-103), and
the condensed high pressure solvent is stored in the solvent surge
drum (D-106) prior to recycle to the extraction system. Thus,
thermal energy requirements of the solvent recovery system are
minimized by using the hot compressed solvent vapor from the main
compressor, to supply the heat required to vaporize most of the
solvent from the organic in the primary still.
3.4 Treatment
Treatment may be simply stated as the process by which compound x
is removed from compound y. The theory of the critical fluid
solvent extraction system has been described in detail, but the
theory is only a portion of the removal process. The treatment
process must pay close attention to the physical and chemical
characteristics of the material to be treated. Removal of the PAHs
and dioxins from the parent material to be treated. Removal of the
PAHs and dioxins from the parent material studied at the United
Creosoting site requires as close a scrutiny of the physics and
chemistry as is cost effective.
3.5 Sample Characterization
For treatment, the contaminated material was mixed with potable
water from the City of Conroe; the chemical makeup of which is
known. However, budgetary constraints did not allow chemical
characterization of the parent material. The physical mixture of
the contaminated material consisted of clayey sand, clay, sand,
rocks and woody material in the form of bark, sticks and old
boards. This wide range of compositions and sizes presented a
441
-------
materials handling problem which may best be solved by the many
vendors who produce machinery such as that used in the mining or
road building industry.
3.6 Sample Preparation
This treatment demonstration consisted of treating two batches of
contaminated soil from separate locations at the site, the former
waste pond and the former operations area. The sample for the
first batch was collected in the former waste pond and weighed 238
pounds. Contaminated soil material was prepared for treatment by
sifting the material through a 1/2" sieve and discarding all
material which did not pass the sieve. The solids which passed
through the 1/2" screen were combined with potable water from the
City of Conroe and mixed with an electric mixer for 20 hours. The
resulting slurry was passed through a 1/8" screen and all material
not passing the sieve was also discarded. The slurry passing the
1/8" screen was very homogeneous and showed marked tendencies for
the particulate to stay in suspension during the time the mixture
was being processed. Figure 3-5 shows a schematic of the feed
preparation/treatment process with the weight of the various
material streams added or lost.
Additional water was added during the actual treatment process to
rinse soil solids from the walls of the feed preparation tank and
the treated slurry collection tank. Losses of both water and soil
occurred by accumulating in the pilot units pipes, filters, pumps,
extractor vessels, and overflows into a second raffinate tank from
the first tank. Some loss occurred due to a clogged pump and
circulation piping which required dismantling in order to return
the system to operation. The additions and losses identified in
Figure 1 for both water and soil resulted from conditions which are
more characteristic of a pilot scale operation than a full scale
unit. Proper design and selection of mechanical equipment for full
scale operation will eliminate such inefficiencies.
The sample for the second batch to be treated was collected from
the former operations area and weighed 280 pounds. The sample to
be treated in the pilot unit was prepared in a manner similar to
that of the first batch. However, this soil exhibited considerably
different working characteristics than the material from the pond
area. The second batch contained heavier sand particles and what
appeared to be black flakes of tar. When mixing ceased, the sand
would immediately drop out of suspension in the mixing tank. The
tar flakes, even with thorough mixing, would not dissolve; pre-
senting potential problems with treatment. A decision was made to
treat only a portion of the material which had been prepared and
dilute it with more water in an attempt to reduce the heavy sand
load and provide for better suspension of the remaining particles.
Figure 3-6 shows a schematic of the feed preparation/treatment
process with the actual weight of the material treated. The
material wasted may have contributed to the organic loading, but
the PAH concentration of the slurry treated was attributed entirely
to the soil mass within the slurry.
admin:erfIpapr.425
-------
oo
SAMPLE
4-1/2 EXTRACTIONS *
TREATMENT
1/2" <1/2" 1/8"
•^-» 205# + 183# H20 -§SVE_ 346# ^55L
20 HRS. MIXING
238#
I
>1/2"33#
>1/8"42#
TREATED &
RECOVERED
I
93# SOIL
177# H20
1# EXTRACT
i
TO ANAYLSIS
* AVERAGE SOLVENT TO FEED RATIO = 2.2
ADDITION TO PROCESS
75# H20
-*• LOST IN SYSTEM
69# H20
81 # SOIL
RGURE 3-5
MATERIAL BALANCE
SAMPLE 1
F: \MAWMGS\GENERM.\GENERMO.JP S-2-90 4:1
-------
SAMPLE
1/2"
280#
267# + 281 # H20
18 HRS. MIXING
>1/2"13#
* AVERAGE SOLVENT TO FEED RATIO = 3.2
3-1/2 EXTRACTIONS *
TREATMENT
1/8"
448# SLURRY >
>1/8"15#
HEAVY
SAND
85#
TREATED &
RECOVERED
59# SOIL
341# H20
2# EXTRACT
TO ANAYLSIS
ADDITION TO PROCESS
95# H20
LOST IN SYSTEM
34# H20
107# SOIL
RGURE 3-6
MATERIAL BALANCE
SAMPLE 2
F:\fMA«HWS\GEMERM.\aENERMI.JP S-2-W 1:1
-------
SECTION 4.0
RESULTS
4.1 Treatment Results
Removal efficiency may be calculated by dividing the difference
between the initial reference and the final reference with the
initial reference point. However, the reference point (mass versus
concentration versus toxicity) influences the interpretation of
removal efficiency. For example, a mass basis may provide high
removal efficiencies for PAHs; from the two tests completed at
United Creosoting, removal efficiencies of 94% and 98% were
achieved for total mass PAHs. In comparison, using concentration
as the indicator for removal efficiency - - expressed in terms of
toxicity equivalents or Benzo(a)Pyrene Equivalents [BAPE] - -
removal efficiencies are slightly lower for each test, 86% and 91%
for total BAPE concentrations.
Table 4-1 presents the laboratory results for the two tests
conducted at the United Creosoting site in terms of mass of
individual and total PAHs, PCP, and individual and total dioxins
and furans. An interesting consideration in the variance among in-
dividual PAHs is the different structures of individual PAHs.
Figure 4-1 demonstrates that those similar structures correspond
in the percent of mass removal. More polar compounds have simi-
larly high removal efficiencies, as anticipated since the solvent
propane is polar.
Table 4-2 presents the concentrations of the chemicals of concern
in terms of individual carcinogenic PAHs, naphthalene, total PAHs
versus total BAPE, PCP, and dioxins/furans versus total TCDDE
(toxicity equivalent similar to BAPE). Again, those chemicals
similar in structure have similar removal efficiencies in terms of
toxicity and/or concentrations (noncarcinogens).
4.2 Material Handling Problems
The problems encountered in handling and treating the waste pond
soil were very different than those encountered during treatment
of the residential surface soil, demonstrating that variability in
soil composition across the site will present dynamic soils
handling problems. The variability in soil will require that the
full scale unit be equipped and operated to respond effectively to
meet the changing influent conditions. Problems encountered in the
pilot study should provide valuable insight into design and
operating parameters that must be incorporated into full scale
treatment.
The 1989 ROD has an in-depth discussion on the use of these
equivalents and the determination of action levels for the
United Creosoting site.
445
-------
TABLE 4-1
tENOVAL EFFICIENCY
TEST ONE
TEST TWO
COMPOUND
ACENAPTHENE
ACENAPHTHYLENE
ANTHRACENE
BENZO(A)ANTHRACENE
BENZO(A)PYRENE
BENZO(B)FLUORANTHENE
BENZO(GHI)PERYLENE
BENZO(K)FLUC«ANTHENE
CHRYSENE
DI BENZOCA, H)ANTHRACENE
FLUORANTHENE
FLUORENE
INDENO(1,2,3-CD)PYRENE
NAPHTHALENE
PKENANTHRENE
PYRENE
TOTAL MASS OF PAH (HG)
MASS IN
UNTREATED
SOIL (HG)
26,640
1.110 J
24,420
7,400
3,552
3,774
1,480 J
3,700
8,140
ND
26,640
28,120
1,406 J
10.360
43,660
26,640
213,046
MASS IN
TREATED
SOIL (HG)
143 J
126
373.8
331.8
504
407.4
504
714
382.2
181 J
462
160 J
462
63 J
546
462
5,149
REMOVAL
EFFICIENCY (X)
NA
NA
98X
96X
86X
89X
NA
81X
95X
NA
98X
NA
NA
NA
99X
98X
98X
MASS IN
UNTREATED
SOIL (MG)
ND
675 J
2,957 J
1,441 J
24,413
20,319
120,020
136,245
44,733
54,892
34,573
40,714
28,508
12,813
2,047 J
12,207
529,437
MASS IN
TREATED
SOIL (HG)
ND
246 J
ND
ND
670 J
616 J
5,089
5,357
2,679
3,482
4,018
2,170
2,679
2,009
964 J
2,089
29,572
REMOVAL
EFFICIENCY (X)
NA
NA
NA
NA
NA
NA
96X
96X
94X
94X
88X
95X
91X
84X
NA
83X
94X
COMPOUND
PENTACHLOROPHENOL
TOTAL TCOD
TOTAL PeCDD
TOTAL HxCDD
TOTAL HpCDD
TOTAL OCDD
TOTAL TCOF
TOTAL PeCDF
TOTAL HxCDF
'OTAL HpCDF
•OTAL OCOF
MASS IN
UNTREATED,
SOIL (UG)
28,120 *
ND
ND
1,184
26,640
96,200
ND
74
2.220
11,840
11,840
TEST ONE
MASS IN
TREATED .
SOIL (UG)
2,436*
ND
ND
202
7,560
28,980
6
109
756
3,150
3,654
PCP AND COO/CDF
RESULTS
TEST TWO
MASS IN MASS IN
REMOVAL UNTREATED TREATED REMOVAL
EFFICIENCY (X) SOIL (MG) SOIL (MG) EFFICIENCY (X)
91X 3,563 J 455 J NA
NA
NA
83X
72X
70X
NA
NA
66X
73X
69X
XA « Not Applicable (because results are below detection limits or not detected)
•
• (UG) « micrograms, pentachlorophenol results in mi IKgrans
lote: Dioxin analysis was not performed on soil from drum 2.
446
-------
NAPHTHALENE
ACENAPHTHALENE
CHRYSENE
BENZ(A)ANTHRACENE
ACENAPHTHENE
BENZO(B)FLUORANTHENE
FLUORANTHENE
BENZO(K)FLUORANTHENE
PHENANTHRENE
BENZO(A)PYRENE
ANTHRACENE
1NDENO(1,2,3-CD)PYRENE
FLUORENE
DIBENZO(A,H)ANTHRACENE
PYRENE
BENZO(G,H,F)PERYLŁNE
FIGURE 4-1
STRUCTURE OF MAJOR
COMPONENTS OF
CREOSOTE
OENOWjflOeM4S.1V 5-7-90 1:1
447
-------
TABLE 4-2
CRITICAL FLUID EXTRACTION TREATABILITY STUDY
Chemicals of Concern
PAHs:
Benzo(a)Anthracene
Benzo(a)Pyrene
Benzo(b)Fluoranthene
Benzo(k)Fluoranthene
Benzo(g,h,i)Perylene
Chrysene
Dibenzo(a,h)Anthracene
Indeno(1,2,3-cd)Pyrene
Naphthalene
Test 1
78 /
36 /
38 /
39 /
16 /
85 /
194 /
15 /
105 /
8
12
11
14
12
9
4
11
Test 2
19 /
322 /
268 /
1797 /
1583 /
590 /
724 /
376 /
nd
25
23
200
190
100
130
100
169 / 75
Total PAHS
Total BAPE*
2502 / 123
175 / 16
6983 / 1104
875 / 121
Pentachlorophenol
Total TCDDE
295 / 55
5.3 / 2.7
47 /
17
Chlorinated Dioxins:
Tetra-, Total
Penta-, Total
Hexa -, Total
Hepta-, Total
Octa -, Total
Chlorinated Furans:
Tetra-, Total
Penta- , Total
Hexa -, Total
Hepta-, Total
Octa -, Total
nd
nd
16
360
1300
nd
1
30
160
160
/ nd
/ nd
/ 5
/ 180
/ 690
/ nd
/ 3
/ 18
/ 75
/ 87
C0::Concentration Before
Cx::Concentration After Extraction
nd::Not Detected; 1/2 detection limit used for those compounds
known to be present at the site.
*::Concentrations for PAHs in Test 1 are averages based upon
duplicate samples (both C0 and Cx) available for analysis.
**::Benzo(a)Pyrene equivalency, only carcinogenic compounds are
included in equivalencies; i.e., those PAHs listed, less
Naphthalene; the 2,3,7,8-disubstituted isomers of dioxin and
furan.
448
-------
Soils handling problems that were encountered due to limitations
in the design of the pilot unit should not be considered along with
legitimate concerns for full scale design. In other words, some
of the problems encountered in the pilot study were due to the
generic design of the pilot unit. The pumps, pipe sizes, cooling
capacity, etc., of the pilot unit were designed for the wide range
of operating conditions to which a pilot unit is subjected.
Proper selection of pumps and pipe sizes for a full scale unit
would eliminate several pilot scale materials handling problems.
Also, if accumulator tanks were designed to be emptied through a
hopper-type mechanism instead of through valves, plugging problems
encountered during blowdown procedures could be eliminated in full
scale treatment.
Another intrinsic pilot scale problem involved temperature/
pressure capabilities. The pilot unit is designed to operate under
limited pressure conditions. Cooling of the unit and complete
condensation of the propane is provided by tap water circulated
through a condenser. When ambient temperature at the site
increased, there was some trouble maintaining a complete liquid
phase of propane. This problem is eliminated on the full scale
unit by two design modifications: 1) higher operating pressure and
2) inclusion of a chiller to provide colder cooling water.
The proper choice of pumps in full scale design is very important.
Pumps that are designed to handle slurries with varying solids
composition should be chosen. Variable speed pumps are required
since control of the flow rate through the unit can effectively
eliminate several handling problems associated with lighter soil
particles that are pumped too fast.
Careful consideration should be given to design of dewatering
equipment. The differences in settling properties of the two types
of soil treated in the pilot study are anticipated to cause
variable dewatering conditions in full scale treatment. The light,
black tar mat particles in the residential soil tended to stay
suspended in water, even after centrifugation.
Proper feed preparation equipment will also be crucial to effici-
ent pretreatment. There was a significant difference between the
two soils regarding the material that was dry sieved through the
1/2 inch screen. The > 1/2 inch material of the waste pond soil
consisted mainly of large gravel and pieces of wood. The >l/2 inch
material from the residential surface soil consisted mainly of
organic material (leaves, grass, small gravel). Proper crushers
or shredders to handle this material must be included in full scale
design. The >l/8 inch material that was sieved after the soil was
slurred also differed between soil types. The >l/8 inch waste pond
material was mainly composed of uniformly sized gravel. The >l/8
inch material from the residential soil contained more organic
matter than rock. Therefore, it is important that pretreatment
equipment be selected that is capable of providing effective
handling of a variety of soil types. If the screened material is
to be crushed and treated, these factors should also be considered
in pump selection.
449
-------
Another important design consideration deals with the configuration
of the vessels, mixers, and piping. Although complete mixing is
not possible in any vessel, some of the handling problems in the
pilot unit could be rectified by changing the process flow in full
scale treatment. The loss of material during treatment is mainly
attributed to the fact
that the slurry flows out of the top of the extractor into the
decanter. Heavier solids that do not stay suspended will not
overflow into the decanter, but will settle in the extractor.
Process flow should be designed so that maximum mixing and easy
flow-through of material can be achieved to minimize losses due to
settling and sticking. Efficient mixers should be included in the
extractor design.
General maintenance and mechanical problems on the full scale unit
will be similar to those of the pilot unit. However, some of the
mechanical problems with the pilot unit were due to lack of
preventive maintenance and wear and tear from transportation to
many sites.
From a solids handling standpoint, if the problems that were
encountered in the pilot study are given serious consideration in
the design of a full scale unit, full scale remediation using
solvent extraction technology appears feasible. Particular
attention should be given to feed preparation equipment, pumps,
dewatering equipment, pipe size, efficient process flow, and vessel
blowdown procedures.
admin:crflpapr.425 459
-------
SECTION 5.0
TREATABILITY CONSIDERATIONS
Some of the questions which were asked in the case of evaluating
the feasibility of using the Critical Fluid Extraction Technology
at United Creosoting included:
• Can contaminants be extracted from the soil down to
health based levels? Levels associated with ARARs (i.e.,
LDRs)?
• Is the process realistically controlled through various
parameters in order to increase removal efficiency?
• Once treated, can the soils be replaced without concern
of residuals from the treatment process?
• Does implementation of the process present short-term
risks to human health or the environment? Air emissions,
etc.?
• How can the extract be handled (i.e., further treatment,
off-site disposal)?
Health based remediation levels for soils at United Creosoting were
determined from an EPA in-house risk assessment [documentation,
1989 ROD]. In addition, the soils at United Creosoting were
determined to be contaminated with a listed hazardous waste under
the Resource Conservation and Recovery Act [RCRA], K001 Wood
Preserving Waste [1989 ROD, 40 CFR 261.32]. Treatment standards
for nonwastewaters of this category, invoked under the Land
Disposal Restrictions [LDRs] of RCRA, included chemicals separate
of the risk-based contaminants assessed for human health and
environmental impacts.
Results of the treatability tests were evaluated as favorable since
the toxicity of PAHs, dioxins, and PCP had been decreased if not
to, then close to, the desired concentration levels in soils; those
contaminants of concern under LDR were similarly reduced. A full
scale unit, with the process designed specifically for the soils
at United Creosoting, could very likely meet, if not better, those
levels necessary for allowing soils to be placed back onsite.
Under Superfund, keeping treated soils onsite would be
preferred to transporting such a large quantity off-site
to a commercial facility.
451
-------
Most of the remaining questions were easily answered due to the
properties inherent of the process, as previously discussed. For
example, the use of propane as the solvent near the critical state
should not leave significant residue in the treated soil. Several
factors influence the efficiency of the process and these factors
can be modified to increase the efficiency over the pilot unit's
performance. Although removal efficiency may be based upon the
mass transfer of contaminants from the soil, the concentrations of
contaminants remaining in the soil is of most consequence; the
efficiency of the process to decrease toxicity of the soils can be
improved. Finally, the process is conducted in an enclosed unit,
and space is available on the commercial portion of the site to
manage the treatment of soils with minimal risk to workers and the
surrounding community.
Treatment alternatives for a Superfund site must satisfy such
fundamental requirements in order to contend for remedy selection
under the Superfund program. In many cases, various treatment
technologies may be combined to form alternatives which adequately
address individual site circumstances. In the case of United
Creosoting, application of the Critical Fluid Extraction Technology
would mean that once contaminants were removed from the soils, the
resulting concentrated organics must also be dealt with in order
for the remedy to be complete.
In consideration of the small volume of extract and the charac-
teristics of the concentrated contaminants, and due to the close
proximity of the residential area, off-site management of the
concentrate was selected. Incineration (the Best Demonstrated
Available Technology [BDAT] for dioxins - total destruction) was
selected as the most effective means of permanent remediation for
site contaminants.4
As a separate alternative, incineration of the soils
presents an interesting point of comparison for the
effective use of more than one treatment technology.
For example, off-site incineration of soils would
necessitate off-site transport of a large volume of
contaminated material and ultimately a large volume of
ash for disposal (off-site). Onsite incineration of
soils would necessitate a temporary incinerator being set
up fairly close to the residential area; a similar plan
opposed in past discussions with the residents at United
Creosoting.
In comparison, onsite critical fluid extraction coupled
with off-site incineration of the extract offers certain
advantages: a tremendous reduction in the volume of
contaminated material to be shipped off-site for
destruction, with minimal ash anticipated from the liquid
concentrate. In addition, the critical fluid extraction
process is entirely enclosed and is perceived more
favorably by the public than an onsite incinerator — so
close to the residential area.
admin:erfIpapr.425
452
-------
SECTION 6.0
REFERENCES
Nelson, W. L., 1969. Petroleum Refinery Engineering, McGraw-Hill
Book Company, Fourth Edition, 1969.
Texas Water Commission, 1989. United Creosoting Superfund Site,
Feasibility Study Amendment - Preferred Alternatives Analysis, July
1989.
Treybal, R. E., 1980. Mass-Transfer Operations, McGraw-Hill Book
Company, Third Edition, 1980.
U.S. EPA, 1989. Record of Decision for the United Creosoting
Superfund Site, Conroe, Montgomery County, Texas, September 1989.
admin:erfIpapr.425
453
-------
C«« Study
Bench-Scale Solvent Extraction
May 1990
CKMHILL Joseph A. Sandrin
Milwaukee, Wisconsin ^___^_i^^ John Flelssner
454
-------
Case Study:
Bench-Scale Solvent Extraction
Treatability Testing of Contaminated Soils and Sludges
from the Arrowhead Refinery Superfund Site, Minnesota
Joseph A. Sandrin
John Fleissner
CH2M HILL, Milwaukee, Wisconsin
ABSTRACT
Solvent extraction is a separation process that has emerged as an effective hazardous waste
treatment technology. It has been applied successfully to industrial wastewaters, soils, and sludges
contaminated with hydrocarbons, petroleum products, and heavy organic compounds.
Solvent extraction was evaluated as an alternative treatment technology at the Arrowhead
Refinery Superfund Site in Hermantown, Minnesota, the location of a former waste oil recycling facility.
Highly acidic, metal laden sludge bottoms and oil saturated clay filter cake were disposed of in a 2-acre
lagoon. The peat layer underlying the lagoon and the surrounding soils are contaminated with oil,
metals, and numerous organic compounds. Under contract to CH2M HILL, Resources Conservation
Company (RCC) conducted bench-scale tests of its Basic Extractive Solvent Technology (B.E.S.T.™).
The results of bench-scale testing and a discussion of the applicability of the process to the wastes at this
site are presented.
INTRODUCTION
Site History
The Arrowhead Refinery Site occupies about 10 acres in northeast Minnesota near Duluth
(Figure 1). According to the Minnesota Pollution Control Agency (MPCA), milk cans were retinned at
the site before 1945. From 1945 to February 1977 the site was used as a waste oil recycling facility.
During oil processing, waste oil was treated with sulfuric acid to deemulsify the oil/water mixture. The
wastewater recovered from this process was discharged to the wastewater ditch, and the waste oil was
filtered through a clay/sand filter. The sludge from the deemulsification process and the filter cake were
disposed of in an unlined 2-acre lagoon in a wetland on the site. The filter cake was also used as fill in
the process area adjacent to the sludge lagoon (Figure 2).
Nature and Extent of Contamination
The U.S. Environmental Protection Agency and MPCA investigated the environmental effects of
onsite waste disposal from 1979 through 1984. The results of their investigations indicate that various
organic and inorganic hazardous substances are present at the site in the subsurface soil, sediment,
surface water, groundwater, and sludge lagoon. The two major contaminant sources defined during the
remedial investigation (RI) were the sludge and filter cake disposed of in the sludge lagoon and the
contaminated soils in the process area.
The surface soils consist of gravelly sand, silt, and fill that was deposited during site operations.
Much of the soil is visibly stained and saturated with waste oil. The lagoon contains a viscous, black oily
liquid sludge, and a black filter cake that consists of an oily clay and a silty sand and gravel fill layer.
The entire lagoon is underlain by peat that appears to persist across the site and to be highly
contaminated. Hazardous substances detected at the site included polycyclic aromatic hydrocarbons
(PAHs), volatile organic compounds, lead, zinc, and small quantities of polychlorinated biphenyls
(PCBs). Estimated contaminated waste volumes as known at the time of the treatability study are:
455
-------
ARROWHEAD
REFINERY
FIGURE 1 Kw.
VICINITY MAP
456
-------
LEGEND
*-—••• EPA DITCH
........ SITE BOUNDARY
NOTE: Arrows Indicate direction of (low.
rr
FIGURE 2 r*:ri:nii
SITE MAP •••
-------
Soil 19,000 yd3
Peat 13,000 yd3
Sludge 4,600 yd3
As documented in its Record of Decision (ROD) for the site, the EPA's selected remedial
action was thermal treatment. The EPA and MPCA were both interested in the application of
alternative treatment technologies that might achieve similar levels of treatment more economically than
thermal treatment. As a result, the EPA agreed to fund a treatability study using a solvent extraction
treatment process.
CH2M HILL had already performed a remedial investigation and feasibility study of the site for
the EPA. Under contract to the EPA, CH2M HILL subcontracted the treatability study to Resources
Conservation Company (RCC) for testing of its Basic Extractive Solvent Technology (B.E.S.T.™).
SOLVENT EXTRACTION
Background
Solvent extraction technology has been used for many years as solvent leaching to recovery
valuable minerals from ores, to remove unwanted materials from coal processing operations, and to de-
oil quench waters in refinery processing operations. More recently it has been used to treat sediments
and soils contaminated with PCBs, chemical manufacturing wastes, and oily, hazardous and toxic wastes.
Organic solvent extraction is particularly suited for treatment of oily wastes because the process
separates wastes into product oil, solids, and water fractions. Solvent extraction can effectively extract
the oil fraction of a waste including such contaminants as PCBs. The remaining solids can sometimes be
disposed of as non-hazardous wastes and the water discharged to a wastewater treatment plant.
The successful application of solvent extraction depends on solvent selection, process
configuration, the nature of the waste and the contaminants, the economic value of the potentially
recoverable compounds, and the treatment or remedial goals. Solvent extraction can be an especially
attractive treatment alternative where (1) valuable products can be recovered from the waste; (2) the
process can yield nonhazardous or uncontaminated residual products; and (3) inordinate wastewater
disposal or air emissions problems are not encountered.
B.E.S.T.™ Solvent Extraction Technology
RCC is the owner of the B.E.S.T. solvent extraction technology, a patented process that takes
advantage of the peculiar solubility behavior of certain aliphatic amines. The process was originally
developed for NASA as a waste treatment process for space travel, where resource recycling is critical.
Triethylamine (TEA) has chemical and physical properties that make it a good candidate for use
in solvent extraction. At temperatures below 65°F, TEA is completely miscible with water arid is also a
good solvent for many organic compounds such as PCBs, PAHs, and petroleum products. At 65°F the
soluble organic and water components of a waste can be separated from the solid components.
When TEA is heated, the solubility of water in TEA drops to less than 2%, separating the water
fraction from the soluble organic fraction. The TEA is then removed to yield an organic fraction. The
B.E.S.T. process separates the waste into three waste product streams or fractions: a solid with soluble
organic contaminants removed; a water that may require treatment before discharge; and an oil product
that can be recycled for energy recovery or incinerated (Figure 3).
TEA is a chemically basic compound that reacts with acids in the waste to yield ammonium
salts. Excessive reaction will result in the loss of expensive solvent. To minimize solvent loss the
B.E.S.T. process includes the addition of caustic to increase the pH of the waste to greater than 11. The
high pH minimizes solvent loss and has the side benefit of precipitating low concentrations of metals
into the product solids and possibly decreasing the teachability of metals in the toxicity characteristic
leaching procedure (TCLP).
458
-------
0.065569 TS Rl SM* FIG 1 5-10 «0
Contaminated Sludge,
Peat, or Soil
-Pa
cn
B.E.S.T.^ Process
I Oversized
I Contaminated
I Material
^
Product Oil
Product Solids
Product Water
® FIGURE 3
B.E.S.T. BENCH-SCALE SOLVENT
EXTRACTION PROCESS STREAMS
-------
RCC conducts bench-scale treatability studies at its laboratory to evaluate the ability of the
process to treat a given waste. In the treatability tests, 1-kilogram batches of waste are subjected to the
same unit processes as a full-scale facility to simulate full-scale operation. Performance is evaluated at
each step, and samples of the products and process intermediates are analyzed to determine if the wastes
are being treated effectively.
TREATABILITY TESTING
CH2M HILL developed a sampling and analysis plan to evaluate the treatability study. Besides
the wastes processed to meet RCC's treatability protocol, additional samples were processed to generate
additional quantities of treated solids and treated oil product. Samples of the three raw wastes and
treated solids and treated oil from each raw waste were analyzed through the EPA's Contract Laboratory
Program (CLP) to provide an independent assessment of process performance. Samples were analyzed
for metals, PAHs, and PCBs. Because of cost limits, it was not possible to process sufficient waste to
provide treated water for CLP analysis.
Lead and PAHs are the primary contaminants at the site. Low levels of PCBs are also present.
The overall purpose of the treatability study was to assess whether a solvent extraction process could
effectively treat the Arrowhead Refinery site wastes and to assess whether it could be a cost-effective
alternative to incineration for site remediation.
Samples of contaminated soil, sludge, and peat wastes were sent to RCC in May 1989 for bench-
scale treatability testing. An observer from CH2M HILL was present at RCC's laboratory throughout
the bench-scale test. RCC conducted its tests according to the protocol presented in Figure 4.
Sample Preparation
Before bench-scale testing began, the raw waste feed composition (Table 1) and various feed
preparation steps were performed. Samples of the soil, peat, and sludge were screened through a 1A*
screen. Samples were then mixed with caustic to determine the amount of caustic required to raise the
pH to 11 as required before the addition of TEA. The sludge and peat had low pH and substantial
caustic addition was required.
Table 1
RAW WASTE FEED COMPOSITION
RCC Analysis
Waste
Type
Sludge
Peat
Soil
Oil
(wt%)
42
22
5
Water
(wt%)
43
53
12
Solids
(wt%)
15
25
83
Ash (550'F)
(wt%)
6.3
9.1
79
After determination of caustic requirements, samples were mixed with required amounts of 50%
sodium hydroxide and the TEA was added. Each waste type was extracted three times with TEA and the
extracts were combined. The first extraction was performed with cold TEA (40°F) to maximize water
solubility. Subsequent extractions were performed with heated TEA (140°F) to maximize extraction of
organic constituents. Solids were separated from the TEA/water solution by centrifugation.
The extracts were combined and heated to 140°F to decrease water solubility. Under ideal
conditions the water should have separated cleanly from the TEA, and the two layers would be separated
460
-------
Raw Waste Feed
(sludge / peat / soil)
Contaminated
Rajact
^ Material
Screened
Waste
Feed
NaOH
TEA
H20
TEA
40°F
1st Extraction
i
Centrifugation I
TEA
iJ
Solids
2nd Extraction
«^^^^J^»»~^
Centrifugation I
TEA
11
140°F
Solids
3rd Extraction
I
I Centrifugation
Solids I
JL
TEA
NaOH (recycle)
I t
Decantatlon
7 140°F
Water /
TEA
Liquid
(water / oil /
TEA) ^
Evaporation
Liquid
(oil / TEA)
Liquid
(oil/TEA)
Drying
PRODUCT
SOUPS
PRODUCT
WATER
TEA
(recycle)
oil /TEA
t
Distillation
PRODUCT
OIL
•a
e
a FIGURE 4
B.E.S.T.® GLASSWARE
BENCH-SCALE PROCESS FLOW
461
-------
in a separatory funnel. Separation was not good for any of the wastes, and it was necessary to recover
the water from the mixture by evaporation.
The TEA was then removed by distillation and steam stripping, leaving the treated oil fraction.
The treated water fraction was recovered from the evaporation, and the treated solids were recovered
from the centrifuge. Samples of each treated product and the raw wastes were analyzed for PCBs, lead,
and PAHs. Mass balances were performed to compare actual recoveries of treated products to predicted
values from the initial waste composition analysis (Table 2). RCC also performed a mass balance for
lead comparing the lead content of the raw wastes to the total lead content of the treated products.
Table 2
PHASE FRACTION MASS BALANCE
RCC Analysis
Waste Type
Sludge
Oil
Water
Solids
Peat
Oil
Water
Solids
Soil
Oil
Water
Solids
Compositional Assay
(wt%)
42
43
15
22
53
25
5
12
83
Bench-Scale Test
% Recovery (wt %)
43
14
13
19
16
22
5
7
82
Analytical Results
Mass balance recoveries for solids and oil were acceptable compared to the initial compositional
analysis. However, recovery of water was poor and RCC was unable to account for the losses (Table 2).
Lead recoveries were 80% from the soil, 180% from the peat, and 200% from the sludge. The high lead
recoveries may be due to the oily matrix of the raw wastes that interfered with lead analysis. The solvent
extraction process separated and concentrated much of the lead in the treated solids product. Based on
the high lead recoveries, it is possible that the actual lead concentrations in the raw wastes are
substantially higher than indicated by the raw waste analyses.
The treated product water had no detectable contaminants except for elevated pH due to the
TEA distilled with the water in evaporation. The low level of contaminants should be expected since
the water was recovered by evaporation.
Solvent extraction concentrated the inorganic contaminants in the treated solids while
concentrating the organic compounds in the oil fraction separated from the TEA. The treated solids
were also analyzed using the EP toxicity procedure. The solids from the peat and the sludge were at or
exceeded the allowable lead concentration of 5 mg/1 in the extract and would be considered hazardous
wastes (Table 3).
462
-------
Waste
Type
Sludge
Peat
Soil
Waste
(mg/kg)
12,700
13,000
2,600
Table 3
LEAD Analysis
CLP Analysis
Product Oil
(mg/kg)
6,500
6,000
1,700
Product
Solids
(mg/kg)
32,000
38,000
2,900
Product Solids
EP Toxicity
(mg/l)
7.9
5
1.9
The organic contaminants were concentrated in the treated oil fraction. However, substantial
concentrations of lead were also detected in the treated oil fraction. The concentrations were high
enough that it might be necessary to dispose of the oil as a hazardous waste by incineration.
PERFORMANCE EVALUATION
The B.E.S.T. solvent extraction process was able to successfully separate the refinery wastes
(sludge, peat, soil) into three treated product fractions (water, solids, oil).
The treated product water could possibly be discharged to a publicly owned treatment works
(POTW) with little or no additional treatment following pH adjustment, since it was recovered by
distillation. However, water mass balance showed poor recovery for water, and RCC was not able to
explain this satisfactorily.
The treated product oil was high in total lead content and might have to be disposed of as a
hazardous waste. The changing nature of the waste oil market would not allow final disposal
requirements to be determined until the oil was available.
Treated product solids would probably have to be disposed of as a hazardous waste because of
the high total lead content in all three treated solids products. The solids from the peat and sludge were
at or above the allowable limit for the EP toxicity extraction procedure, which designated them
hazardous wastes at the time of the study. The evolving RCRA land ban regulations and the recent
adoption of the TCLP test to replace the EP toxicity test increase the likelihood that the solids would
have to be disposed of as hazardous wastes.
The estimated hopper-to-hopper treatment cost for solvent extraction in this case was
comparable to incineration. The estimated cost for solvent extraction provided by RCC was
approximately $289/ton as compared to $300/cubic yard for incineration. The requirement to use
distillation to recover the water because of poor decantation in the extraction step contributes
significantly to the treatment cost for solvent extraction, as does the high caustic requirement for pH
adjustment of the acidic waste materials.
The cost estimate for solvent extraction is from the vendor and is based on minimal
demonstrated field experience, whereas the estimate for incineration is based on actual costs from
remedial actions at other sites. The uncertainty of the solvent extraction cost estimate is, therefore,
likely to be greater than that developed for incineration.
Materials handling considerations would also affect the implementability of a solvent extraction
remedy at this site. Solvent extraction processes typically require that the waste be delivered to the
treatment system at sizes less than 1 inch, as opposed to a range of 4 to 6 inches for rotary kiln
incinerators. At a site like Arrowhead, with the large debris, tree stumps, and other oversize wastes,
considerable waste preparation would be required for incineration, and even more for solvent extraction.
463
-------
The poor water recovery demonstrated in the bench-scale testing, and the materials handling
questions would require that a pilot-scale study be performed to better evaluate performance and
estimate costs for solvent extraction.
The bench-scale tests did not monitor or evaluate the fate of volatile contaminants in the
treatment system. The site is known to contain substantial volatile contamination. The fate and control
of volatile emissions would have to be addressed in any selected remedy at the site.
It is difficult for new and innovative technologies to demonstrate a substantial savings based on
the uncertainties inherent in the fact that they are new and innovative. If new or innovative
technologies are to be selected and used they must therefore offer other advantages to overcome the
uncertainties. The recovery of a potentially recyclable product and a reduction of the volume of waste
that must be handled as a hazardous waste are examples of such advantages. Unfortunately, at this site,
the difficult nature of the wastes did not provide demonstration of such process advantages.
CONCLUSION
The complex nature of the wastes at this site, the low pH, and the high lead content make the
wastes at the Arrowhead Refinery site difficult to treat using any technology. Solvent extraction did not
present an advantage over incineration in performance or cost. It was, however, able to separate the
water, solids, and oil fractions from the waste. At properly selected sites with a less difficult matrix,
solvent extraction may offer more of an advantage as a remediation technology. It should continue to be
examined and evaluated at sites where it is appropriate.
GLT977/022.51
464
-------
FIELD DEMONSTRATION OF A CIRCULATING BED COMBUSTOR (CBC)
OPERATED BY OGDEN ENVIRONMENTAL SERVICES OF
SAN DIEGO, CALIFORNIA
Prepared for Presentation at:
EPA's Second Forum on
Innovative Hazardous Waste Treatment Technologies:
Domestic and International
May 15-17, 1990
Prepared by
Nicholas Pangaro
Charles W. Young
Douglas R. Roeck
ALLIANCE
JWt TK,-,:-.; .:
465
-------
FIELD DEMONSTRATION OF A CIRCULATING BED COMBUSTOR (CBC)
OPERATED BY OGDEN ENVIRONMENTAL SERVICES OF
SAN DIEGO, CALIFORNIA
This paper describes several support aspects of a TSCA Demonstration Test Burn and site
remediation program conducted by Alliance Technologies Corporation and Ogden Environmental
Services (OES) at the Swanson River Oil Field on the Kenai Peninsula, along the south Coast
of Alaska. The intent is to summarize the planning requirements associated with conducting
testing of this nature at a remote site, to describe problems encountered and problem resolution,
and to summarize the results of the Demonstration Test Burn. Because of the site location,
Alliance focused strongly on project planning to anticipate and minimize technical and logistical
problems that otherwise could have caused significant time delays and additional costs. Problems
that did arise were successfully addressed. The merits of careful, conservative planning should
become clear through reference to this case study.
HISTORY OF THE PROBLEM1
PCBs were used at the Swanson River Field between 1962 and 1972 as a coolant material in
electrical transformers and in heat transfer oils. A compressor explosion occurred in 1972,
resulting in the release of an unknown quantity of PCB laden oil. Later, in 1983 and 1984, PCB
contaminated sand and gravel from the explosion area were used for dust suppression along two
miles of roads within the Swanson River site. Soil sampling conducted during and after the
summer of 1984 demonstrated PCB contamination throughout the area. Based on extensive site
characterization studies, most of the identified contaminated soil areas were shown to have PCB
levels of less than 50 ppm. A cleanup target level of 12 ppm PCB was established with a
relaxed target of 24 ppm applicable to areas with difficult access. About 75,000 tons of soil are
estimated to be contaminated, requiring treatment. PCB mixtures of concern are primarily
Aroclor 1242/1248.
PROBLEM RESOLUTION
A feasibility study was conducted to identify alternative cleanup methodologies. This study
resulted in a list of five possible thermal technologies and six possible non-thermal technologies.
Based on cost, feasibility, permitting and policy issues, cleanup capability, future liability, and
other related considerations, the list of candidate technologies was narrowed to the five thermal
techniques or landfilling. Landfilling was ruled out due to concerns about future liability.
Based on the results of the feasibility study, the facility owner prepared a remediation
specification and released it for open competition. The solicitation process resulted in the
selection of Ogden Environmental Services (OES) of LaJolla, California to conduct the
remediation program. OES owns and operates transportable incineration units which are based
on the OES circulating fluidized bed combustion technology.
ALLIANCE
466
-------
THE OGDEN CBC2
The Circulating Bed Combustor (CBC) is an advanced fluidized bed system that employs high
velocity combustion air (14-20 ft/sec) to circulate solids through a highly turbulent loop. The
system's high combustion efficiency is well suited to burning materials with relatively low
heating value, such as contaminated soils. Residence times range from 2 seconds for gases to
30 minutes for circulating solids. Waste solids and limestone are combined and added to the bed
at an inlet located between the cyclone and the combustion chamber. Upon calcination,
limestone reacts with acid gases such as HC1 and SO2 to form calcium chloride and calcium
sulfate. The feed materials are rapidly heated after introduction to the unit. A uniform
temperature (over the operating range of 1450-1800°F, ± 50°F) is maintained in the system by
virtue of the highly turbulent well-mixed combustion zone, the circulating soil and limestone, and
use of auxiliary fuels. Ash is periodically removed and cooled. Hot off gases pass through a
flue gas cooler/feed water heater and then to a baghouse for removal of fine particles, prior to
exiting the stack. A schematic diagram of the CBC is shown in Figure 1.
OPERATING REQUIREMENTS
In order to obtain the requisite permits for conducting on-site destruction of PCB (>500 ppm) or
PCB contaminated (50-500 ppm) material, EPA's Office of Toxic Substances (EPA-OTS)
imposed a number of requirements for data gathering and monitoring. To comply with
regulations promulgated under the Toxic Substances Control Act (TSCA), OES was required to
demonstrate:
99.9999% destruction and removal efficiencies for PCBs;
• 99% removal of hydrochloric acid; and
paniculate emission rates of less than 0.08 gr/dscf.
Due to the relatively low levels of potential HC1 precursors present in the soils being incinerated
(600-700 ppm total PCB), the less stringent RCRA standard for HC1 removal (total emissions of
less than 4 pounds per hour) was applied to this case in lieu of demonstrating 99% removal
efficiency. However, in addition to the regulatory requirements, OES was required to:
• monitor for a number of potential products of incomplete combustion (PICs)
during the Demonstration Test Burn, including a demonstration that the stack
exhaust contained less than 10 ng/m3 of polychlorinated dibenzo dioxins and
furans; and
• continuously monitor levels of PCBs in the solid effluents from the incinerator
(bottom and baghouse ash) to assure that materials contaminated with PCBs above
2 ppm were not redeposited at the site.
ALLIANCE
467
-------
STACK
COMBUSTOFK.
COOLING
WATEn
Figure 1. Schematic of the Ogden Circulating Bed Combustor
ALLIANCE
46R
-------
ALLIANCE'S ROLE
OES secured the services of Alliance Technologies Corporation of Bedford, Massachusetts to
assist in securing proper operating permits. Alliance's role was to address test requirements to
be used during the test burn and to implement the test burn. In addition, Alliance assisted in the
set up and initial operation of an on-site laboratory where routine analysis and monitoring of
incinerator effluents and feeds would be conducted.
ON-SITE LABORATORY OPERATION
During routine operations, it was necessary to verify decontamination as batches of soils were
treated by the CBC. Since the storage capacity for post treatment soils at the site was limited,
the need for rapid turn-around analysis for PCBs was obvious. In addition, analysis of the feed
materials was required to provide information on the amount of limestone needed to maintain
CBC operation at approximately 10% above stoichiometric conditions, to record the amounts of
PCB material treated, and to verify compliance with permit requirements on concentration and
homogeneity limitations. Considering the remoteness of the site, the ability to obtain rapid turn-
around analysis at an off-site location was not possible, leading to the establishment of an on-site
laboratory. This on-site laboratory, housed in a 40'xl2' skid mounted transportable building, was
equipped with preparatory equipment and a Varian Model 3400 Gas Chromatograph with an
Electron Capture Detector (GC/ECD).
Although the methods used for these analyses were based on standard EPA procedures, certain
adaptations were made to allow for the limitations of working in a field laboratory and to provide
data in a rapid (overnight) manner. Standard Operating Procedures (SOPs) were prepared to:
• Supply personnel at the on-site laboratory with guidance on analyzing both pre-
and post-treated soils for PCB contamination levels;
• Define the application of the standard methods to this project and modifications
made to allow for the limitations mentioned;
• Define acceptance/rejection criteria for evaluation of the analytical data.
Since the PCB material found in the contaminated soils were a mixture of several individual PCB
isomers, it was likely that residual contamination would not retain its identity as an Aroclor
mixture in the post-treated soils. The analytical scheme made allowances for this by using two
sets of parameters for analysis. Samples from the pre-treatment streams (feeds) were analyzed
for the predominant PCB mixtures on the site, Aroclor 1242/1248. Samples from post-treatment
solid streams (treated soils and ash) were analyzed for PCB homologue groups (mono- through
deca-chlorinated biphenyls) to verify treatment to a level below the limit applied for landfill
disposal, i.e.,total PCB material <2.0 ppm. In order to make the analysis as routine as possible,
Alliance developed several computerized calculation schemes, based in a LOTUS-123
spreadsheet, to allow fast reduction of raw analytical data.
ALLIANCE
-------
Since the period required to obtain and ship replacement parts to the South Coast of Alaska made
potential down time unacceptable, the laboratory was equipped with an overabundance of backup
materials, including a complete set of expendable and breakable items for the GC/ECD.
Additional problems were encountered in supplying power to the laboratory as the analytical
instrumentation and computers proved to be sensitive to the industrial grade power available at
the site. This problem was finally overcome by supplying power to the laboratory from a
dedicated generator.
DEMONSTRATION TEST BURN PROGRAM
Logistical Considerations
Due to the location of the site, planning for the Demonstration Test Burn required much more
attention than is usually associated with such efforts. Although Alliance had previously
conducted Trial Burns at locations which are quite distant from its Bedford, Massachusetts base
(for instance, Houston, Texas, McDowell, Missouri, and San Diego, California), the difficulty
involved in procuring additional equipment or spare parts within a short time frame required
packing redundant materials, supplies, and instrumentation to be used on the program. Secondly,
transportation of equipment and personnel to a remote site in Alaska provided a challenge in
logistics and scheduling.
Two Alliance employees provided overland transportation of equipment to the Port of Seattle,
Washington in a company truck, a five-day trip. The truck was then loaded on a barge for the
three-day journey to the Port of Anchorage, Alaska. Since the barges leave Seattle only once
every three days, scheduling was critical to allow cost-effective use of the traveling employee's
time. After a flight to Anchorage, the drivers continued the overland trip to the site,
approximately 160 driving miles south of Anchorage. In all, the equipment was hauled
approximately 4,700 one-way miles to the site, taking nine days at a cost of approximately
$10,000.
Redundancy in equipment and material was considered a necessity. Commercial shipping of
spare parts between Bedford, Massachusetts and the site would have required at least 3 days in
the event of breakage or a mishap. All equipment and material, from metering boxes to duct tape
to aluminum foil, was backed up. Materials lists, formulated based on Alliance's past experience,
were checked, rechecked, and all amounts were increased. Other materials were brought as
contingencies (for instance, several additional solvents were packed despite the low probability
of use).
Permitting Issues Associated with the Demonstration Test Burn
In order to satisfy the regulatory agency's needs for adequate data from the test bum, Alliance
had to meet OTS' strict Quality Assurance specifications. Alliance prepared a Demonstration
Test Burn Quality Assurance Project Plan (QAPP), which underwent two revisions prior to final
approval by OTS. Alliance also had to accommodate OTS laboratory audits, and coordinate with
OTS on scheduling the burn to allow EPA observation of the field testing.
470 /** ALLIANCE
/iftft V.-.:•:; -:-•-•
-------
EPA-OTS guidelines on performing laboratory audits resulted in the expenditure of considerable
time and effort to meet the requirements. Previous exper'ence with laboratory audits had
consisted of one-day site visits with review of laboratory procedures and capabilities through
review of records and interviews and personnel. In many cases, receipt of audit check samples,
with results due some time after the visit, is included in the audit OTS, however, required the
analysis of check samples during the time of the audit, with the auditor observing all aspects of
the analysis. To maintain confidentiality of other client's data during the auditor's visit, the
laboratory did not accept other work during this time frame. As a result, OES compensated the
laboratory for time required to conduct the visit.
The audit lasted for approximately three and one-half days. All results from the audit were
determined to be well within acceptable ranges and laboratory data submitted in conjunction with
the Demonstration Test Burn were accepted by the agency.
RESULTS OF THE DEMONSTRATION TEST BURN PROGRAM
During the September 24-27, 1988 time period, Alliance completed three replicate runs at each
of two process operating conditions. Results of the program are summarized in Table 1. The
major change in the process for the second operating condition as compared to the first was a
7 percent increase in the soil feed rate and an increase in the CBC exit temperature from 1615°F
to 1700°F. Along with these changes, the CBC exit velocity increased from 20 to 22.5 ft/sec.
As shown in Table 1, all applicable emission limits as imposed by OTS were met. DREs were
essentially the same for both test conditions despite the low average PCB concentrations in the
soil feed. HC1 emissions were well below the 4.0 Ib/hr limit and dioxin/furan concentrations in
the post-treated ash were well below the imposed limit of 2 ppb. Paniculate emissions doubled
from Condition 1 to Condition 2 but remained well below the standard of 0.08 gr/dscf @
7 percent O2 standard.
During the test burn, few delays were encountered due to operation of the CBC unit itself. Some
process delays were experienced, due mainly to jamming of the rotary valve in the soil feed line
and/or the feed auger.
The final report for this project was submitted in December 1988. On June 23, 1989, Ogden
announced that they had been issued a nationwide federal permit for use of the CBC system in
remediation of PCB contaminated soils. At the time of this announcement, the Swanson River
project represented the world's first major remediation program using CBC technology and the
largest PCB/soil cleanup to date.
In conclusion, the successful completion of any field sampling program can be traced back to
adequate planning prior to initiation of the program. Without adequate planning, factors involved
in the Swanson River Demonstration Test Burn, such as the distance and remoteness of the site,
may have combined to cause major problems, delays, and cost increases in the program. Because
of the approach followed, the project team was able to readily address most problems ultimately
encountered.
ALLIANCE
471
-------
TABLE 1. SWANSON RIVER TEST RESULTS SUMMARY
Parameter
Soil Feed Rate (Ib/hr)
Sorbent Feed Rate (Ib/hr)
CBC Exit Temp (°F)
CBC Exit Velocity (ft./sec.)
Ash Rate (Ib/hr)
Ash (% of feed + sorbent)
Soil PCB Concentration (ppm)
Soil RCl Concentration (ppm)
Ash PCB Concentration (ppb)
Ash PCDD/PCDF Concentration (ppt)
PM (gr/dscf @ 7% 02)
HCl Emissions (Ib/hr)
Avg PCB ORE (%)
PCDDs/PCDFs Concentration (ng/dscm)
Total Organic Chloride (Ib/hr)
Condition 1
8,500
170
1,615
20
8,233
95
683
162
< 9
< 175
0.0077
1.3
> 99.99994
<1.5
< 7.3E-5
Condition 2
9,100
170
1,700
22.5
8,790
95
507
229
< 16
< 195
0.016
1.4
> 99.99994
<2.0
< 2.4E-4
472
ALLIANCE
-------
REFERENCES
1. Ives, J.A. (ARCO Alaska, Inc.) and D.T. Young (OES). PCB Remediation in Alaska.
Prepared for Presentation at the PCB Forum in Houston, Texas. August, 1989.
2. Anderson, B.M., and R.G. Wilbourn (OES). Contaminated Soil Remediation by
Circulating Bed Combustion - Demonstration Test Results. Prepared for Presentation at
Superfund '89 in Washington, D.C. November 1989.
ALLIANCE
14 >:v,;.:,;••;:. :••;: "
473
-------
LOW TEMPERATURE THERMAL TREATMENT (LT3) OF SOILS
CONTAMINATED WITH AVIATION FUEL AND CHLORINATED SOLVENTS
Roger K. Nielson
Weston Services, Inc.
West Chester, PA 19380
and
Craig A. Myler
U.S. Army Toxic and Hazardous Materials Agency
Aberdeen Proving Ground, MD 21010
ABSTRACT
Successful pilot study of Low Temperature Thermal Treatment
(LTJ) (Patent No. 4,738,206) of soils contaminated with
volatile organic compounds led to a full scale demonstration
to evaluate the use of this technology at DOD installations.
The LT process developed by Weston Services, Inc. under
contract with the U.S. Army Toxic and Hazardous Materials
Agency (USATHAMA) was used in a test of remediating soils
contaminated with aviation fuel (JP4) and trichloroethylene
(TCE). Tinker Air Force Base, Oklahoma was selected as the
demonstration site. An abandoned landfill area had high
concentrations of contamination in a clay soil which was
determined to be ideal for this test. The LT process
performed better than expected and achieved cleanup levels
at lower temperatures, higher processing rates and shorter
residence times than previously thought possible. This
directly translates into cost savings for future remediation
sites. Test results indicitive of the performance and
potential application of this technology will be presented.
INTRODUCTION
Thermal processes for decontaminating hazardous waste
laden soils have focused on incineration. Raising the soil
to temperatures at which the contaminants present undergo
decomposition is an effective means of cleaning the soil.
Effective that is in terms of contaminant removal, not cost.
In raising the soil to the elevated temperature, energy is
expended needlessly, causing increased operating costs. In
addition, contact heating between the combustion gases and
the contaminated soils produces a large volume of
contaminant laden vapor which must be processed through
pollution abatement equipment. The larger the volume of
this vapor phase, the larger (and costlier) the down stream
equipment.
To decrease energy requirements and simultaneously
reduce downstream pollution abatement equipment and
operations costs, a low temperature thermal treatment system
474
-------
has been developed. The rationale for this system centers
around those drawbacks to incineration listed above. First,
by heating the contaminated soil only enough to volatilize
the organic contaminants, energy is saved. Second, using
indirect heating of the soil, downstream pollution abatement
equipment cost is reduced, both in capital and operating
costs. -
Development of the LT process began in August, 1985
with the operation of a pilot scale thermal processor. Use
was made of an indirect heat exchanger which operated by
contacting the soils with heated augers which both
transported and heated the soil. Volatile contaminants were
released from the soil and processed through an afterburner.
The pilot system schematic is shown in Figure 1. The
success of this pilot scale system prompted the performance
of an economic evaluation and bench scale tests that
resulted in a full scale LT system.
The bench scale system was operated to determine if
results from the pilot scale could be effectively reproduced
on a smaller scale, thereby providing a low cost means to
pre-screen potential soils and contaminants. The bench
scale system was also used to determine the effects of
temperature and residence times on different soils. Of
particular note during this study was the processing of
soils containing semi-volatile contaminants from diesel fuel
and JP4. Results indicated the potential for bench scale
data to be used in scale up. This would greatly simplify
pre-screening of soils by reducing the costs of sampling as
well as the amount of soil required. A central laboratory
was established at Weston to perform pre-screening tests
rather than constructing a pilot unit at each site.
On April 19, 1988 U.S. Patent Number 4,738,206 was
issued for a Low Temperature Thermal Treatment Process for
removing volatile and semi-volatile organic contaminants
from soil. A full scale LT system was designed and
constructed by Roy F. Weston, Inc. under this patent based
on the results of the pilot and bench scale systems.
Initially, the full scale system was used to treat petroleum
contaminated soils at a site in Illinois. The first test of
this equipment in the remediation of hazardous wastes was
performed at Tinker Air Force Base, Oklahoma. The results
of the test program at Tinker will be described.
PROCESS DESCRIPTION
The full scale low temperature thermal treatment system
used two, thermal processors operated in series. The first
processor is mounted on top the second. Four intermeshed
screws in the first processor convey the soil through a
jacketed trough where it falls by gravity into the second
processor. The second is identical to the first except the
direction of travel of material is reversed. In this way,
soil enters the unit on one side and exits the unit on the
opposite side thereby reducing the space requirements
475
-------
Air to
Atmosphere
Hot Oil
Reservoir
Air Containing
Stripped VOC's
After-
burner
Oil Heating
System
Combustion Air
Blower
Air In
Air
Preheater
Figure 1: Schematic Diagram of the Pilot Scale LT System
476
-------
necessary for operations as well as reducing the flow of
contaminated soil outside of the processors.
A 7.6 million KJ/hr hot oil system provides a heat
transfer fluid, Dowtherm HT, to the hollow screws and
jackets of the thermal processors. Temperature controls
allow the oil to be maintained at temperatures up to 343 C.
A portion of the combustion gases from the oil heater are
passed through the thermal processor at about 375 C. This
is done for two reasons. First, waste heat recovered from
this stream maintains the volatile stream exiting the
thermal processor at a temperature high enough to avoid
condensation of the contaminants on the walls of the
exhaust ductwork. Second, the exhaust acts as an inert
atmosphere to maintain the gases in the thermal processor
below their lower explosive limit.
The vapor stream from the thermal processor consists of
the contaminants being removed, water vapor from the soil
and exhaust gases from the hot oil heater. This stream
exits at approximately 150 C (maximum) and flows through a
fabric filter, condenser, afterburner and caustic scrubber
system. The fabric filter removes particulate carried over
from the processor. The vapor stream then passes through an
air cooled condenser which reduces the temperature to
approximately 52 C. Water and organics condensed reduce the
load on the afterburner. The afterburner is a 3.7 million
KJ/hr gas fired, vertical, fume incinerator operating at 982
C. The afterburner is operated at a minimum 3% excess
oxygen and the exhaust is continuously monitored for 0~/ CO
and total hydrocarbons. The 982 C exhaust from the
afterburner is quenched to approximately 82 C. It then
passes through a packed bed absorber where acid gases
produced in the afterburner are neutralized with a caustic
solution.
A liquid stream is produced by the condenser which is
water rich but does contain some hydrocarbons. The aqueous
phase is separated from the organic phase in an oil-water
separator. The aqueous phase is processed through a water
treatment system consisting of fabric filters followed by
granular activated carbon. This water is then used as
makeup water for the scrubber and for dust control on
processed soil. The organic phase from the separator is
either drummed for off site disposal or injected into the
afterburner.
A block diagram of the system is shown in Figure 2 and a
schematic layout of the system is shown in Figure 3. The
system described here is mobile and can be transported on 5
trailers. Utilities required for operation are propane or
natural gas, electricity and process water.
discharges from the system include the scrubber stack
exhaust, the processed soil, the granular activated carbon
and filter cake, and, if not injected into the afterburner,
477
-------
00
Contaminated
soil
storage
Classifier
Oversize
Sweep gas
To
Drag (light
conveyor
Surge
hopper
Feed soil
Dust
Fuel/
combustion
air
Hot
on
atinuauhsie
f Hoti
Hot on
system
Hotel burner off-gases
— Fuel/combustion air
Cool
oil
Treated soil
Thermal
Processor
rested«
Discharge
conveyor
Truck feed
conveyor
Processed
soil storage
bin
Fabric fror
off-gases
Condonsor
ONAMtor
Organtes
Condensate 1 Water
55-gtlon
drum
LD.fan
Carbon
adsorption
unit
Attorbumar
Water
tank
Scrubber
Carbon
UMMfctL
sorpo
unit
To stack
Figure 2: Block Diagram of the Full Scale LT System Used
at Tinker Air Force Base, Oklahoma
-------
1
10
w
O
_ J JL
70"
A Thermal Processors
B Baghousa
C. Motor Control Center
D Thermal Processor Drive Units
E Hot CM System
F Condenser
G Induced Draft Fan
H Afterburner
I Scrubber ID Fan
J Scrubber System
K Exhaust Slack
L Caustic Storage Tank
M. Frvsn Watsr Systtni
N Slowdown Carbon Adsorptnn Units
Oa*«—. *niii« Tii*a»m
oiwoown oynvin
P Recycle Water Tank & Pump
0 Oil/Water Saparator
R Walar Carbon Adsorption Units
S. day Shredder
T Drag Conveyor
U. Discharge Conveyor
V. Dump Truck
W Continuous Emissions Monitoring Trailer (CEM)
-W-
Figure 3: Schematic Diagram Showing Equipment Layout of the
Full Scale LT System Used at Tinker Air Force
Base, Oklahoma
-------
the organic phase from the oil-water separator. The maximum
utility load for the unit is as follows:
Liquid Propane 2.27 m^/day
Process Water 40.88 m /day
Electricity 600 amp/460 V (3 phase)
Operation requires 8 personnel for continuous operations
including a site manager and an instrumentation technician.
SITE DESCRIPTION
To test the capabilities of the full scale LT system a
site contaminated with compounds of specific Army concern
was desired. Objectives in site selection were to find a
location which had a clay soil matrix contaminated with JP4.
JP4 was selected as the contaminant of choice as spills of
this fuel on Army bases may require remediation and diesel
fuel contaminated soil had already been tested in this
unit.D
Coordination with the U.S. Air Force resulted in Tinker
Air Force Base, Oklahoma being selected as the test site.
Remedial investigations on this active Air Force support
base turned up an abandoned sludge dump in one of the bases
landfill areas. Survey and analysis of the area estimated
600 m of sludge contaminated clay soil. Analysis of soil
and water from this site indicated JP4 and trichloroethylene
in high concentrations along with other potential
contaminants in lesser concentrations, probably constituents
of diesel, kerosene and other solvents. The land fill was
no longer in use and outside of everyday operational
traffic. The Air Force personnel at Tinker Air Force Base
as well as the state and federal regulators were in favor of
the test. The mixed wastes which included TCE, a listed
hazardous waste, provided an opportune site for
demonstration of the LT technology on a full scale. The
site layout is shown in Figure 4.
Due to the presence of listed hazardous wastes, a permit
was required for the test. An application for a Research,
Development and Demonstration (RD&D) permit was filed in
accordance with the requirements of RCRA. An approved RD&D
permit was issued by Region VI, EPA 120 days after
application. This permit authorized the conduct of the test
plan for the LT process.
TEST PLAN AND RESULTS
The test plan for the full scale LT system was designed
to meet the following objectives:
- determination of effectiveness of the full scale LT
system at removing JP4 and chlorinated solvents from
soil.
480
-------
00
i SCALE
4 «TACKC* 6tir CotvMoR
« o*-
t. SCtuaBEft 5VSTEM
7. C*u5TiC STelTASt SVSTSrt
6, PCW«« T*A«SFcCMe*
H MCON T«AlLM
1 I
*?
.?.
1
i
M*
|
*Mta1
«»r
,
/
/
f
/
\ b ~
i M«r
••••!g«3"-
m
I
r
/
X
/
n
GD n;
i — i ^
1 ^* 1 **
if
iwr
VZc
1
l—ii-l
» 1
1
1
r >
Sl^
nn
L««nLL «OA.
iTScBT^l
!
"2
ft
••
SDK
•*
m
n
Wt
«TN
*TI
!•»
tk»*
'
U
'S'J'" ^"W* ASP-AIT
.MMn* M.
"•-«"•«" ""
o«!<*m»
Ml >U
ft rD
55 ar
r •*
Ct»t
-------
Table 2: Results of First Test Run
Contaminant Feed Soil Processed Soil GCL
Concentration Concentration (Ppb)
(ug/kg) (ug/kg)
TCE 37,250 5.4 70
1f2-Dichlorobenzene 35,000 BDL 125
1,3-Dichlorobenzene 3,500 ND NS
1,4-Dichlorobenzene 8,700 ND 10,800
Toluene 8,700 ND 330
Napthalene 4,300 BDL NS
Total Xylene 5,900 ND 150
ND - not detected
BDL - below detection limit
NS - not specified
Table 3: Results of Second Test Run
Contaminant Feed Soil Processed Soil GCL
Concentration Concentration
(ug/kg) (ug/kg)
TCE 111,000 5 70
1,2-Dichlorobenzene 15,000 BDL 125
1,3-Dichlorobenzene BDL BDL NS
1,4-Dichlorobenzene ND ND 10,800
Toluene 8,300 ND 330
Napthalene 5,000 BDL NS
Total Xylene 11,400 ND 150
ND - not detected
BDL - below detection limit
NS - not specified
Testing at higher production rates (less residence time)
and lower temperatures resulted in mechanical failure of the
process equipment. The system was overloaded. The maximum
soil processing rate achieved was 60% above the design rate
at 11,364 kg/hr. Although this rate was not demonstrated
for a sustained period, a sustained rate of 9091 kg/hr was
established. A final test was conducted at a higher
production rate but at a previously tested intermediate
temperature (204 C) to confirm the data of the 3 previous
tests. The results are presented in Table 4.
At this time in the test program, analysis of the soil
indicated the presence of Arachlor 1260, a polychlorinated
Biphenyl (PCB). The system was never intended to process
wastes of this nature nor were they expected. This
discovery halted all operations and investigation eventually
led to the cancellation of the remainder of the test. A
482
-------
- determination of impact of system parameters on
effectiveness.
- evaluation of the use of a stripping agent to enhance
contaminant removal.
- determination of optimum operating conditions.
- determination of stack emissions.
Goal cleanup levels (6CL) were established for the test
based on previous requirements or established standards.
These goals were accepted by the EPA regional office in the
RD&D permit. Table 1 lists the GCL for contaminants found
above the detection level in soil processed.
Table 1: Partial List of Goal Cleanup Levels
Contaminant GCL (ppb)
Trichloroethylene (TCE) 70
1,2-Dichlorobenzene 125
1,4-Dichlorobenzene 10,800
Toluene 330
Total Xylenes 150
The first test run was conducted at the high end of the
hot oil temperature to be tested, 315 C. This run
established that the process is capable of remediating soils
contaminated with JP4 and TCE. Results are presented in
Table 2. Following this run, the system was operated at a
reduced hot oil temperature (204 C) in an attempt to
establish a lower operating limit for hot oil temperature
from which an optimum would be determined. The results of
this run are presented in Table 3. Comparison of the
results shown in table 2 and 3 indicate no significant
difference between the two operating conditions in terms of
contaminant removal.
At this point, an attempt was made to reduce the
residence time to establish a maximum processing rate. The
temperature was also reduced to 150 C to determine the
lower effective operating temperature. The results of this
test are not conclusive as the test was aborted due to
mechanical failure caused by overloading the processors.
483
-------
complete description of the events following the PCB
discovery and the actions taken will be reported at a later
time.
Table 4: Results of Fourth Test Run
Contaminant Feed Soil Processed Soil 6CL
Concentration Concentration (ppb)
(ug/kg) (ug/kg)
TCE 10,575 23.4 70
1,2-Dichlorobenzene 53,000 BDL 125
1,3-Dichlorobenzene BDL ND NS
1,4-Dichlorobenzene 14,750 BDL 10,800
Tolune BDL BDL 330
Napthalene BDL BDL NS
Total Xylene ND ND 150
ND - not detected
BDL - below detection limit
NS - not specified
CONCLUSIONS
Although the test plan was cancelled prior to
completion, the results of the test at Tinker Air Force Base
are positive. Three of the five objectives of the test were
met. It was determined that low temperature thermal
treatment technology is effective at removing JP4 and TCE
from soil. Optimum operating conditions for the full scale
system were established as the mechanical limits of the
processing equipment. At the levels of contamination in the
soils tested, the range of operating conditions had no
discernible effect on removal efficiency.
Two objectives were not met in the test plan due to the
premature cancellation of the test. First, the use of
stripping agents to improve contaminant removal was not
investigated. The results of tests performed showed removal
of contamination to near or below detectable limits in all
cases. The use of a stripping agent would not have improved
this result at a detectable level. The second objective,
determination of emissions, was not met as stack testing was
not performed for the test runs made. Stack sampling was to
be conducted only at the optimum operating conditions. As
the PCB discovery halted testing during the determination of
the optimum conditions, no stack sampling was performed.
This is not felt to be a significant impediment to fielding
the LT technology. The afterburner and scrubber system are
standard unit operations used in similar treatment systems
for contaminant laden streams. Although no field data was
collected on this system, its standard configuration is
expected to meet or exceed the required discharge criteria.
484
-------
The LT full scale system tested was capable of
remediation of soils contaminated with volatile and
semi-volatile compounds. Processing rates obtained
indicated a cost for remediation between $90 and $100 per
metric ton of soil fed to the system. The unit was
determined capable of processing at a rate of 9090 kg/hr.
Removal of the contaminants present was consistantly below
the criteria established for cleanup.
References Cited
1. Pilot Investigation of Low Tennperature Thermal
Stripping of Volatile Organic Compounds (VOC's) From Soil,
U.S. Army Toxic and Hazardous Materials Agency Report No.
AMXTH-TE-CR-86074, APG, MD, June 1986.
2. Economic Evaluation of Low Temperature Thermal Stripping
of Volatile Organic Compounds from Soil, U.S. Army Toxic and
Hazardous Materials Agency Report No. AMXTH-TE-CR-86085,
APG, MD, August 1986.
3. Bench-Scale Investigation of Low Temperature Thermal
Stripping of Volatile Organic Compounds (VOC's) From Various
Soil Types, U.S. Army Toxic and Hazardous Materials Agency
Report No. AMXTH-TE-CR-87124, APG, MD, November 1987.
4. United States Patent Number 4,738,206, Apparatus and
Method for Low Temperature Thermal Stripping of Volatile
Organic Compounds from Soil, Inventor: John W. Noland,
Issued April 19, 1988.
5. Nielson, 5-K. and Cosmos, M.G., "Low Temperature Thermal
Treatment (LT ) of Volatile Organic Compounds from Soil: A
Technology Demonstrated", Presented at the 1988 Summer
National Meeting of the AICHE, Denver, Colorado, 21-24
August, 1988.
485
tt US GOVERNMENT PHNTNG OFFCE 1990-748-159/20442
------- |