EPA/540/R-98/500
April 1998
SIMULTANEOUS DESTRUCTION OF ORGANICS AND
STABILIZATION OF METALS IN SOILS
Emerging Technology Report
By
Center for Hazardous Materials Research
Pittsburgh, Pennsylvania 15238
Assistance Agreeement CR 819604-01-0
Project Officer
Randy A. Parker
Land Remediation & Pollution Control Division
National Risk Management Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
NATIONAL RISK MANAGEMENT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before comp'
1. REPORT NO.
EPA/54Q/R-98/5Q0
4. TITLE AND SUBTITLE
Simultaneous Destruction of Organics and
Stabilization of Metals in Soils
lllllllllllllllllllllllll
PB98-133150
5. REPORT DATE
April 1998
6, PERFORMING ORGANIZATION CODE
7. AUTHOR®
Dr. A. Bruce King,
Stephen A. Paff
8. PERFORMING ORGANIZATION REPORT NO.
Randy A. Parker,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Center for Hazardous Materials Research
320 William Pitt Way
Pittsburgh, PA 15238
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
CR 819604-1
12. SPONSORING AGENCY NAME AND ADDRESS
National Risk Management Research Lab
Office of Research and Development
U.S. Environmental Protection Agency"
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Project Report 9/93-9/96
14. SPONSORING AGENCY COOE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer; Randy A. Parker (513) 569-7271
Reproduced from
best available copy.
16. ABSTRACT
The Sulchem Process reacts the material being treated with elemental sulfur at elevated
temperatures in an' inert reactor system. Organic hydrocarbons react with the sulfur to
form an inert fine solid of carbon and sulfur, hydrogen sulfide gas, and modest amounts
of carbon disulfide. Heavy metals react to form sulfides or sulfide coated particles
which are less soluble.
Various types of batch reactors were evaluated in the laboratory test program to
establish process conditions and evaluate several reactor configurations. Tests were
employed^using a contaminated_soil sample from a manufactured gas plant (MGP) site.
Destruction and removal efficiencies for aromatic hydrocarbons from phenanthrene to
benzopyrene were all in excess of 99%. Cadmium, copper, lead, nickel and zinc were
found to provide significant reduction in EPA Toxicity Characteristic Leaching
Procedure (TCLP) values following treatment of the soil by the Sulchem process.
Process economics for remedial soil treatment were estimated to be in the range of
$105 to $181/ton depending on the size of the site and the processing rate.
17, KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Superfund
Soil
Hazardous Waste
Remediation
Organics
Metals
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report!
Unclassified
21. NO. OF PAGES
in. PRICE
20. SECURITY CLASS (This page)
Unclassified
EPA Form 2220-1 (Re*. 4-77) previous edition is obsolete
-------
NOTICE
The U.S. Environmental Protection Agency through its Office of Research and
Development funded the research described here under Assistance Agreement CR
819604-01-0 to the Center for Hazardous Materials Research. It has been subjected to
the Agency's peer and administrative review and has been approved for publication as
an EPA document.
This document has been prepared to report the results of a laboratory test program for
process treatment. No warranty, express or implied, is made with respect to the
methods and practices provided herein.
The Center for Hazardous Materials Research (CHMR) disclaims any and all liability
and responsibilities whatsoever in connection with any personal loss, injury including
death, property loss or damage, penalty imposed upon, or violation of any statute or
regulation, by, or in respect to any person or property, however caused, involving any
matter covered in this document.
Further, CHMR assumes no responsibility or liability for any errors or omissions in this
document.
PROTECTED UNDER INTERNATIONAL COPYRIGHT
ALL RIGHTS RESERVED.
NATIONAL TECHNICAL INFORMATION SERVICE
U.S. DEPARTMENT OF COMMERCE
ii
-------
FOREWORD
The U.S. Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions leading to a
compatible balance between human activities and the ability of natural systems to
support and nurture life. To meet this mandate, EPA's research program is providing
data and technical support for solving environmental problems today and building a
science knowledge base necessary to manage our ecological resources wisely,
understand how pollutants affect our health, and prevent or reduce environmental risks
in the future.
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks from
threats to human health and the environment. The focus of the Laboratory's research
program is on methods for the prevention and control of pollution to air, land, water and
subsurface resources; protection of water quality in public water systems; remediation
of contaminated sites and ground water; and prevention and control of indoor air
pollution. The goal of this research effort is to catalyze development and
implementation of innovative, cost-effective environmental technologies; develop
scientific and engineering information needed by EPA to support regulatory and policy
decisions; and provide technical support and information transfer to ensure effective
implementation of environmental regulations and strategies.
This publication has been produced as part of the Laboratory's strategic long-
term research plan. It is published and made available by EPA's Office of Research
and Development to assist the user community and to link researchers with their
clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
-------
ABSTRACT
The Center for Hazardous Materials Research (CHMR), through a Cooperative
Agreement with the U.S. Environmental Protection Agency's Risk Reduction
Engineering Laboratory (now the National Risk Management Research Laboratory),
conducted a laboratory evaluation of the Sulchem Process for treatment of soils
contaminated with organic hydrocarbons and heavy metals.
The Sulchem Process reacts the material being treated with elemental sulfur at
elevated temperatures in an inert reactor system. Organic hydrocarbons react with the
sulfur to form an inert fine solid of carbon and sulfur, hydrogen sulfide gas, and modest
amounts of carbon disulfide. Heavy metals react to form sulfides or sulfide coated
particles which are less soluble. The acid gases formed may be scrubbed or treated to
recover elemental sulfur using an auxiliary process unit.
Various types of batch reactors were evaluated in the laboratory test program to
establish process conditions and evaluate several reactor configurations. Processing
temperatures of 250° to 350°C are required to obtain sufficient conversion in
reasonable processing times. At these temperatures, hydrocarbons with boiling points
greater than 350 °C are virtually completely destroyed in the process. Hydrocarbons
with boiling points less than about 250° to 300°C desorb from the soils reactor before
reaction temperature is reached. Several alternatives were examined for treating the
lighter organics including passing the vapors through a second stage reactor with sulfur
vapor/liquid as well as collecting the organics in the condensate for alternative
treatment. The latter configuration is the process equivalent of a reactive desorber
which has a lower operating temperature.
Tests were employed using a contaminated soil sample from a manufactured gas plant
(MGP) site. Destruction and removal efficiencies for aromatic hydrocarbons from
phenanthrene to benzopyrene were all in excess of 99%.
Immobilization of heavy metals was determined by the concentration of the metals in
the leachate produced by the EPA Method 1311, Toxicity Characteristic Leaching
Procedure (TCLP), in which the metal concentration is compared to the EPA TCLP
regulatory limits. Cadmium, copper, lead, nickel and zinc were found to provide
significant reduction in the TCLP values following treatment of the soil by the Sulchem
process. Copper TCLP values were reduced most effectively by this treatment. Lead
TCLP values were reduced below regulatory targets when concentrations in the original
soil were below about 10,000 ppm. Cadmium TCLP values were reduced below TCLP
limits when the concentration in the original soil was below several thousand ppm.
Process economics for remedial soil treatment were estimated to be in the range of
$105 to $181/ton depending on the size of the site and the processing rate.
This report was submitted in fulfillment of Assistance Agreement CR 819604-01 by the
Center for Hazardous Materials Research under the sponsorship of the U.S.
Environmental Protection Agency. This report covers a period from October 1993 to
September 1996, and work was completed as of September 1997.
iv
-------
TABLE OF CONTENTS
Page
FOREWORD iii
ABSTRACT iv
FIGURES vii
TABLES viii
ABBREVIATIONS AND SYMBOLS ix
ACKNOWLEDGMENTS x
1.0 INTRODUCTION 1
1.1 SULCHEM PROCESS 1
1.2 PRIOR WORK 3
1.3 PROJECT OBJECTIVES 4
2.0 EQUIPMENT AND EXPERIMENTAL PROCEDURES 5
2.1 SULFUR PROPERTIES 5
2.2 BATCH REACTORS 5
2.2.1 Closed Mode Reactor 6
2.2.2 Vented Mode Reactor 6
2.2.3 Stirred Reactor 6
2.2.4 Batch Reactor Experimental Procedures 7
2.3 VAPOR REACTOR 9
2.4 ROTARY SOIL REACTOR 10
2.5 DATA AND SAMPLE COLLECTION . 12
2.5.1 Feed Characterization 12
2.5.2 Product Sampling 12
2.5.3 Data Collection 12
2.6 QUALITY ASSURANCE/QUALITY CONTROL 13
3.0 SOIL SAMPLES 15
3.1 SARM SAMPLES 15
3.2 SPECIAL SOIL BLENDS 15
3.3 MANUFACTURED GAS PLANT SAMPLE 16
4.0 ANALYSES ...18
4.1 ORGANIC ANALYSES 18
4.2 METALS ANALYSES 18
4.3 OTHER ANALYSES 18
5.0 ORGAN ICS DESTRUCTION 20
5.1 SOIL BATCH REACTOR TESTS 20
5.1.1 Test of Desorption versus Boiling Range 20
5.1.2 Effect of Process Parameters 24
5.2 VAPOR REACTOR 24
5.2.1 Test Results 24
5.2.2 Model for Vapor Reactor System 25
5.2.3 Residence Times and Reactor Sizes 27
v
-------
TABLE OF CONTENTS
(Continued)
Page
6.0 METALS STABILIZATION 31
6.1 TESTS WITH DIFFERENT METALS 31
6.2 TESTS OF PROCESS PARAMETERS 36
6.2.1 Stoiehiometry 36
6.2.2 Temperature 36
6.2.3 Reaction Time 36
6.3 OTHER TESTS 36
6.3.1 Two Stage Runs 40
6.3.2 Soluble Sulfides 40
6.3.3 Mineralogical Speciation 43
6.4 SUMMARY OF METALS TESTS 43
7.0 ROTARY SOIL REACTOR TESTS 44
7.1 DESCRIPTION OF REACTIONS 44
7.2 RESULTS 44
7.2.1 Overall GC Results 44
7.2.2 Quantitative Results 47
7.3 DISCUSSION OF SOIL REACTOR RESULTS 49
7.3.1 Sulfur/Soil Ratio 49
7.3.2 Residence Time 49
7.3.3 Temperature 49
7.4 VAPOR REACTOR PERFORMANCE 50
7.5 CONCLUSIONS FROM ROTARY REACTOR RUNS 51
8.0 DETAILED PROCESS DESCRIPTION 52
8.1 SOIL PROCESSING 52
8.2 SOIL REACTOR 52
8.2.1 Multiple Screw Conveyer System 53
8.2.2 Rotary Reactor 53
8.2.3 Stationary Furnace Reactor 53
8.2.4 Selected Reactor and Considerations 54
8.3 SULFUR/VAPOR REACTOR SYSTEM 55
8.4 OFF-GAS HANDLING SYSTEM 56
8.5 SOIL POST-PROCESSING .56
8.6 WASTEWATER TREATMENT 57
9.0 PROCESS ECONOMICS 58
10.0 CONCLUSIONS AND RECOMMENDATIONS 62
REFERENCES 63
APPENDIX A Results from Parametric Studies Conducted at 250°C
Using A Vented Stirred Reactor
APPENDIX B Quality Assurance/Quality Control
vi
-------
LIST OF FIGURES
Page
1. Sulchem Process 2
2. Autoclave Reactor 8
3. Rotary Soil Reactor 11
4. First Order Kinetics for Toluene Destruction 26
5. Arrhenius Plot of Toluene Destruction 29
6. Toluene Destruction - Calculated vs. Actual 29
7. Lead TCLP Values as a Function of Reactor Sulfur Content 38
8. Cadmium TCLP Values as a Function of Reactor Sulfur Content .......... 39
9. Comparison of MGP Soil and Product Fractions . 46
vii
-------
LIST OF TABLES
Page
1. Quantitative QA Objectives 14
2. Characterization of MGP Sample 16
3. Analytical Methods 19
4. Effect of Boiling Point on Destruction and Removal 22
5. Vapor Reactor Results 25
6. Calculated Residence Times 27
7. Vapor Stream 30
8. Calculated Vapor Reactor Volume 30
9. Tests of Various Metals vs. Reaction Temperature 32
10. Tests of SARM III Soil vs. Reaction Temperature 34
11. Cadmium, Lead, Nickel, Zinc, and Cobalt vs. Reaction Temperature 35
12. Results of Process Parameter Tests - Lead and Cadmium 37
13. Two Stage Run 41
14. Effect of Soluble Sulfide 42
15. Summary of Rotary Reactor Tests 45
16. Rotary Reactor Tests - Recoveries 48
17. Rotary Reactor Assumptions and Parameters 54
18. Summary of Process Economics 59
viii
-------
ABBREVIATION AND SYMBOLS
A pre-exponential Arrhenius reaction constant
BP boiling point
CHMR Center for Hazardous Materials Research
cP centipoise
Ea activation energy
FID flame ionization detector
GAG granular activated carbon
GC gas chromatograph
ICP inductively coupled plasma
ID internal diameter
k rate constant
MGP manufactured gas plant
mL milliliter
MS mass spectrometry
PCB polychlorinated biphenyls
R universal gas constant
SARM Standard Analytical Reference Matrix
SS stainless steel
SSM Synthetic Soil Mixture
t reaction or residence time
T temperature
TCLP Toxicity Characteristics Leaching Procedure
ix
-------
ACKNOWLEDGMENTS
The Center for Hazardous Materials Research (CHMR) would like to acknowledge the
cooperation received throughout the project from the U.S. Environmental Protection
Agency, Office of Research and Development. In particular, we would like to thank Mr.
Randy A. Parker, from the EPA's Superfund Technology Demonstration Division who
provided helpful guidance for this project.
CHMR would like to acknowledge the following individuals from the research team who
were responsible for important contributions to this project.
Center for Hazardous Materials Research
A. Bruce King, Ph.D.
Stephen W. Paff
Brian Bosilovich
William E. Benusa
Theodore Tabacchi
Project Manager
Engineering Manager
Engineer
Chemist
Technician
Sulchem. Inc.
Harold W. Adams
President
x
-------
1.0 INTRODUCTION
In November 1992, the U.S. Environmental Protection Agency (EPA) Risk
Reduction Engineering Laboratory (RREL) awarded the Center for Hazardous
Materia! Research (CHMR) an Assistance Agreement (CR 819604-01-0) to test and
evaluate the Sulchem Process. The Assistance Agreement was provided under the
Emerging Technologies Program, E05 Solicitation. This report documents the
results of this study. Prior reports are referenced [1,2,3,4].
1.1 SULCHEM PROCESS
The basis for the Sulchem technology is a family of interrelated processing
technologies described in a series of four issued patents (U.S. Patent Nos.
4,581,442, dated April 8, 1986; 4,921,936, dated May 1, 1990; 4,990,404,
dated February 5, 1991; and 5,347,072, dated July 13, 1994) and one patent
pending, (Harold W. Adams, inventor), collectively referred to here as the Sulchem
technology.
The Sulchem technology works by reacting elemental sulfur with waste materials
in an oxygen-free environment, in the original patents, the reactions of the sulfur
were with pure organic mixtures in either the liquid (typically at 250°C to 350°C)
or vapor phase {typically at 500°C). The products are primarily hydrogen sulfide
and an inert biack solid of approximate formula CS0 5e/57.
This research and development focussed on the application of the Sulchem Process
to contaminated soils at Superfund sites. The Sulchem Process was anticipated to
provide destruction of hazardous organics while simultaneously stabilizing metais in
contaminated soils. The Sulchem Process uses elemental sulfur, which reacts with
the carbon in organic materials at moderately elevated temperatures to form an
insoluble, inert carbon-sulfur amorphous solid (CS0 56) The contained heavy metals
are immobilized through formation of insoluble metal sulfides.
The Sulchem Process's main process components include;
• A pre-reaction mixer where the soil and sulfur are mixed;
• An externally heated rotating solids reactor;
• A vapor phase reactor where desorbed organics from the first reactor
ate further reacted with elemental sulfur;
• The off-gas handling system, which collects and treats condensable
by-products and scrubs acid gases from the effluent vapors; and
• A post-reaction treatment unit that recovers excess reagent and
prepares the treated product to comply with on-site disposal
requirements.
A general block flow diagram for the Sulchem Process is shown in Figure 1. The
Sulchem Process, as applied to treatment of soils and sludge, consists of two
reactors, one for treating the solid phase material and a second for treating the
gases emitted from the first reactor, including desorbed organic vapors.
1
-------
TREATED GAS
ABSORBER
NaOH
SALTS WATER
MAKEUP
SULFUR
SULFUR
FEED SOIL
REGENERATOR
GAS
TREATMENT
MIXER
OXIDIZER
POST-
TREATMENT
SECTION
SOLIDS SECTION
VAPOR SECTION
REACTOR
TREATED SOIL
-------
The second reactor is required because, for the more volatile organic compounds,
desorption effectively competes with reaction with liquid phase sulfur. The details
of the integration between these two reactor stages will be discussed in
subsequent sections.
The feed soil, possibly after some dewatering, is fed into a pre-reaction mixer
where elemental sulfur (and other reagents, if used) are added to the feed soil and
lightly mixed. The feed mixture (soil and reagents) is next fed to the first reactor
which consists of an indirectly heated, rotary reactor. A controlled atmosphere is
provided in the reactor with flow of inert gases over the tumbling solids excluding
oxygen from the reactor and removing off-gases released by the process. Heat up,
reaction, and cool down zones for the reactor will be employed as the solids move
through the reactor.
1.2 PRIOR WORK
The National Environmental Technology Applications Center (NETAC), an affiliate
of CHMR, previously constructed and operated a pilot-scale production unit of the
Sulchem technology [5J. The unit enabled the collection of steady-state samples
of the reaction products and by-products for subsequent analysis. The results
demonstrated the application of the Sulchem technology to destroy organic
compounds in a vapor phase reaction with sulfur. The organic compounds tested
in these vapor phase pilot tests included perchloroethylene (PCE), trichloroethylene
(TCE), trichlorobenzene, Freon 113, and Aroclor® 1242.
Exploratory screening tests were conducted using simulated contaminated soil to
test the reactivity of sulfur with some of the typical heavy metals encountered in
hazardous waste. These tests were designed to determine if the sulfur reaction
with heavy metals immobilized the metals sufficiently to pass the TCLP {Toxicity
Characteristics Leaching Procedure) test. A small-scale batch test was run on an
EPA Synthetic Soil Matrix (SSM) spiked with arsenic, chromate, lead, and ferric
ferrocyanide (as a complex cyanide). The results demonstrated reduction of TCLP
leac'nate values from 6 and 35 mg/L for arsenic and lead,respectively, to below
RCRA regulatory levels (1 and 5 mg/L, respectively) [6J.
Based on these separate demonstrations of the ability of the Sulchem Process to
destroy organics in the vapor phase, and initial tests to immobilize metals in soils
using liquid phase sulfur, CHMR prepared the proposal to the EPA Emerging
Technologies program. The proposal, was to evaluate the process conditions
necessary to both destroy organics and immobilize metals in one process and to
establish the limits for this process on these different contaminants. That
objective thus became the basis for the project being reported here.
Concurrent with this project, a privately funded treatability test was made of the
Sulchem Process in both closed and vented mode batch reactors [10]. The
samples tested were two variations on a soil blend (prepared according to Section
3.2) containing 1.70% cadmium, 1.90% mercury, 1.47% zinc and 750 ppm of
Aroclor® 1260. PCB destruction efficiency varied from 99.0 to 99.95% for runs
ranging from 275 to 350°C. TCLP metals met RCRA standards for mercury and
indicated 95% reduction for TCLP cadmium and zinc TCLP reduction of 42 to
99.4% depending on soil type and process conditions.
3
-------
1.3 PROJECT OBJECTIVES
The purposes of this project were to determine the technical and economic
feasibility of using the Sulchem Process for site remediation. Therefore, the
specific objectives for this project were as follows:
• Establish that both organic compound destruction and heavy metals
immobilization can be carried out simultaneously;
• Establish the limits of applicability for destruction of different organic
compounds in processing soils, sludge and sediments;
• Establish the limits of applicability for different heavy metals in
processing soils, sludge, and sediments for heavy metal stabilization;
• Establish processing conditions to achieve organic compound
destruction and heavy metal stabilization;
• Determine the chemical/mineralogical mechanism for the stabilization
of heavy metals in soil matrices;
• Establish process requirements for post-reaction treatment of the raw
reactor product {solids} as well as the off gases/condensate produced;
• Develop process economics for the Sulchem Process for treatment of
soils, sludge, and sediments; and
• Make recommendations concerning further development of the
Sulchem Process.
4
-------
2.0 EQUIPMENT AND EXPERIMENTAL PROCEDURES
Three types of reactor systems were used in the experimental program. The initial
screening tests of soils were performed in three variations of small batch reactors.
Subsequently, a vapor phase reactor was used to evaluate a second stage reactor
in the process whereby unreacted organic vapors emerging from the soils reactor
was simulated as feed to a vapor reactor. Finally, an integrated reactor system
was used to evaluate the overall process performance in a larger batch reactor.
The following section includes a brief summary of sulfur properties at typical
reaction temperatures. Subsequent sections describes the reactor systems used
and the general operating procedures employed in the test runs with each of these
reactor systems.
2.1 SULFUR PROPERTIES
Elemental sulfur is used as the chemical reagent in the Sulchem Process. Sulfur is
an oxidizer, with properties similar to that of oxygen, particularly at temperatures
above 250°C. Key parameters which are important to understand the behavior of
sulfur and its use in the Sulchem Process include its melting point, liquid viscosity,
boiling point, autoignition temperature, and the temperature at which it exhibits
oxidative behavior in the presence of organics.
The melting point for sulfur is approximately 120°C. Sulfur melts to form a
viscous liquid, with viscosity of approximately 120 cP. The viscosity drops with
the temperature as the temperature rises to 160°C jj = 67 cP). Then the viscosity
anomalously increases to approximately 93,000 cP, forming a practicaliy
unflowable liquid. Above 187°C, the viscosity decreases again with temperature,
dropping to 15,000 cP at 250°C, and to below 1,000 cP at about 300°C [7J. The
"viscous region" for liquid sulfur presents problems when designing or building a
sulfur heating system, as the highly viscous sulfur resists heat transfer.
The boiling point for liquid sulfur is 444°C, which provides an upper bound for
temperature for a liquid sulfur reactor system. The autoignition temperature of
sulfur is approximately 230-260°C although open cup flash points of 187-207°C
have been reported [81. The literature indicates that sulfur begins to undergo
relatively fast reactions with hydrocarbons, causing the evolution of hydrogen
sulfide gas, at temperatures as low as 180°C [8], although previous Sulchem
work suggested that a significant sulfur/hydrocarbon reaction does not begin until
temperatures Qf approximately 220-240°C [5],
When sulfur reacts with a hydrocarbon, two principal products are formed:
hydrogen sulfide gas (H2S), and an amorphous carbon-sulfur solid, having the
approximate formula of CS0 56. Previous work showed that the reaction proceeds
both in the vapor and liquid phases.
2.2 BATCH REACTORS
Three different small batch test reactors were employed for the initial screening
studies. Each was sized to heat batches of approximately 200 g of soil in an inert
atmosphere from ambient temperature to nominal reaction temperatures ranging
from 250° to 450°C.
5
-------
For the initial tests, two unstirred reactors were employed, One of these was
designated as the closed mode reactor (high pressure) and the other the vented
mode reactor (low pressure). These two reactors were heated by immersion in a
heated fluidized sand bath. Subsequently, an autoclave was modified to provide
an auger mixed reactor. The autoclave was heated with a cylindrical furnace
jacket which could be lowered after the run for rapid cooling of the reactor. The
following briefly describes these three reactor systems and the general
experimental procedures followed.
2.2.1 Closed Mode Reactor
The closed mode reactor consisted of a high pressure bomb approximately 10"
long and 1-3/8" l.D. (nominal internal volume approximately 270 mL) made of 304
SS. The bomb was fitted with an internal thermowell along the center axis and
was equipped with a pressure gauge and shut-off valve. After filling with the
reaction mixture, the bomb was leak tested and inserted into a heated fluidized
sand bath.
After cooling, the excess gas was bled through a bubbler flask of 10N NaOH to
absorb the H2S and into a gas burette to measure the volume of non-condensable
gases. The quantity of H2S was measured by weight gain of the caustic flask.
After the gases were removed and measured, the bomb was opened and the solid
residue removed and weighed.
2.2.2 Vented Mode Reactor
The vented mode reactor was made from a 2" l.D. 316 SS pipe and was operated
at ambient pressure. The reactor was outfitted with an upper condenser section
which permits sulfur and other higher boiling components to condense and reflux
back to the reactor. The upper section was maintained between the melting point
and viscosity limit for molten sulfur (i.e., 120° to 1 50°C).
The reactor section was approximately 7 Yz" long and welded to a 2" flange. The
flange was attached to an upper section of the reactor approximately 12" long and
containing an internal coil of 14" SS tubing. The reactor was heated by immersing
the lower section and the flange in the fluidized sand bath. The upper section was
heated by both external heating tape and by circulation of hot oil through the coil
to keep this zone at 120°C to 150°C. Off-gas from the condenser section was
passed through a.heated line to an ice trap and then a 10N NaOH bubbler.
The initial experiments involved shakedown and redesign of the vented mode
reactor system. These experiments included testing use of hot water for the
condenser section, which was subsequently replaced with a hot oil ( ~ 150°C)
circulating system.
2.2.3 Stirred Reactor
Initial tests with the above two reactors, which lacked any means of mechanical
stirring, demonstrated that the segregation of the reactants occurred during heating
due to the sulfur melting and collecting at the bottom of the reactor. Therefore,
there was the possibility that for runs below the boiling point of sulfur the soil in
the upper part of the reactor might not have sufficient opportunity to react with
either sulfur vapor or sulfur liquid. A stirred batch reactor would better simulate
the expected continuous reactors employed on larger scale than an unstirred
6
-------
reactor. Secondly, there was a concern in using the closed mode reactor that
there might be pressure effects on either the reaction (due to the high molecularity
of gaseous products) or on the formation of the stabilized treated soil product.
Consequently, a stirred, vented mode reactor was employed for the subsequent
experimental runs.
The reactor was a standard Autoclave Engineers stirred 1-liter autoclave made of
316 SS. The autoclave is rated for 9000 psi although it was used at ambient
pressures for these experiments. A 5-inch high auger was installed on the stirrer
shaft, which left approximately 1A" space next to the wall that was not stirred.
The auger was slowly turned during the experiment with an air drive. A
thermowell was located between the wall and auger near the bottom of the
reactor. A second thermowell was located in the vapor space of the reactor. A
cooling coil was located in the upper portion of the reactor zone, but was not used
during the experiments. The autoclave reactor is illustrated schematically in
Figure 2.
The vent line from the autoclave was attached to a small stainless steel dip-tube
trap. The connecting line was heated with heating tape and the trap was cooled in
ice water. In later experiments when recovery of unreacted organic compounds
was being determined, this single trap was replaced with two traps in series for
more efficient trapping.
The effluent gas from the trap was connected to a 10N NaOH bubbler as before.
The non-condensable gases were vented to the hood. For some experiments, the
volume of non-condensable off-gas produced over measured time intervals was
determined by collection over water in a gas burette. No analyses of these gases
were made during this project although previous work in this laboratory [5] has
shown the minor amount of non-condensable gaseous product formed was
primarily hydrogen, in addition to the nitrogen used for purge gas.
2.2.4 Batch Reactor Experimental Procedures
The following outlines the general operating procedures for the batch reactor runs.
Weighed quantities of soil blend material used for each experiment were
determined as well as the weights of sulfur and other reagents, if any. The initial
tare weights of the caustic scrubber (including solution) was also determined prior
to each run.
At the conclusion of each experimental run and after cool down to room
temperature, the caustic scrubber was removed and weighed. The weight gain
was attributed primarily to neutralized hydrogen sulfide. The ice trap was weighed
and the collected liquid removed. Following rinsing of the trap with solvent, and
then acetone and drying, the tare weight of the trap was determined. The weight
of condensate was calculated by difference.
The reactor was then disassembled and the raw reactor solids (generally a dark
gray powder) were removed from the reactor. A clean spatula and a clean small
wire brush was used to remove residual solids (principally sulfur) from the auger
and condenser coil of the reactor. For the lower temperature runs, sulfur crystals
occasionally had to be chipped from the reactor with a small chisel. After all of the
collected solids were removed, a total weight was determined and the mixture
manually homogenized before sampling. The reactor equipment, including auger
7
-------
Figure 2. Autoclave Reactor
Air driven
auger
Pressure
gage
Air
* Gas Samples
Thermocouples
Removable
Digital
display
cylinder
Temperature
Controller
is not to scale
Reactor
Furnace
8
-------
and coil, as well as the tools used (spatula, brush, and chisel) were cleaned with
water and detergent, rinsed with acetone and dried before reuse on the next run.
For those runs in which recovered organic compounds were determined, the
reaction system hardware were separately rinsed with methylene chloride to
remove and recover any residual organic compounds. Several separate rinses were
used for the reactor including auger and coil which were saved for use in the
Soxhlet extraction of the raw reactor solids sample. A volumetric fraction of these
rinses, calculated from the weight ratio of the Soxhlet sample to the weight of
total reactor solids, was used in the Soxhlet boiling flask. In this way, the Soxhlet
extract was based on a representative sample of the raw reactor solids.
In addition, for the runs with organic compounds, all rinses of the connecting
tubing and the ice trap were used in the liquid-liquid extraction of the ice trap
condensate. Similarly, all of the rinses of the connecting tubing and caustic
bubbler flask were also used in the liquid-liquid extraction of the caustic solution.
2.3 VAPOR REACTOR
CHMR designed and constructed a vapor reactor test system. The reactor was a
316 SS vertical pipe (24" x 3" diameter) partially filled with molten sulfur and with
gas fed through a sparger in the bottom of the reactor. The system was tested
using toluene vapors as the organic feed. Toluene was chosen because it
represents a volatile organic, anticipated to be vaporized in the soils reactor; it is
an aromatic and expected to be less reactive than aliphatic compounds; and, it is
well characterized and relatively easy to work with.
Nitrogen gas from a high pressure cylinder was split into three streams, each of
which had the flow rate separately monitored. One stream was fed through a
sparge tank, in which it became saturated or nearly saturated in toluene vapors.
The second fraction bypassed the sparge tank, but was mixed with the effluent
from the sparge system before being fed to the reactor. By controlling the ratio of
the flows between these two streams, the toluene influent concentration was
controlled. The third stream, typically on the order of 50 mL/min, was a nitrogen
purge stream fed directly to the top of the reactor vessel. It was designed to
reduce the likelihood of sulfur plugging in the vent to the reactor. The vapor
residence time in the reactor was calculated by dividing the overall free volume in
the reactor (reactor volume less the volume of liquid sulfur) by the total gas flow
rate into the reactor, which is the sum of the flow rates of the three nitrogen inlet
streams.
The toluene/nitrogen mixture was passed into the reactor at the bottom of a bath
of molten sulfur. The mixture passed into the sulfur, initially through a perforated
plate sparger, which was later replaced with a 3/8" fine mesh screen, which CHMR
believes produced smaller bubbles. Passing through the perforated plate or screen,
bubbles form, and become saturated with sulfur as they pass through the molten
sulfur. The organic compounds inside the bubbles reacted with sulfur in the vapor
phase as the bubbles rose. Further reactions with the vapor would occur as the
vapors pass through the top of the reactor.
The feed stream passed through the molten sulfur and out the top of the reactor,
A splash guard near the reactor top protected against escape of liquid sulfur. The
gases passed through two one-half inch stainless steel pipe traps, maintained at
9
-------
-20°C by periodically adding dry ice to ethylene glycol, Care was taken to avoid
lower temperatures where H2S would condense (BP -42°C). These traps removed
water vapor, sulfur, and residual organic vapors. The gases were next passed
through a ION NaOH scrubber to remove hydrogen sulfide before discharge
through a vent.
The reactor was heated using Kelrod heaters, and maintained at the appropriate
temperature using a combination of variacs on the heater units and an automatic
temperature controller. The reactor was packed in glass wool insulation to reduce
heat loss and help maintain constant and consistent skin temperatures. The off-
gas system was heated with heat tape to the first ice trap to reduce sulfur
condensation, which may clog the piping.
To determine the toluene feed to the system, the toluene sparge vessel was
weighed before and after each experiment. To determine overall toluene
destruction, the materials collected in the traps were weighed, extracted with
methylene chloride, and analyzed for toluene after each experiment by GC-FID
using a recovery standard. The weight change (gain) in the scrubber system was
also noted.
2.4 ROTARY SOIL REACTOR
To test the full reactor system concept, CHMR designed and constructed a pilot-
scale rotary soil reactor. The soil reactor is shown in Fig ure 3. It consisted of two
31 6 SS cylindrical externally heated chambers, each of which was 8-inches in
diameter and 12-inches long. The two chambers were separated by a metal flange
with a two-inch opening in the center. The system was designed to rotate at 0 to
10 revolutions per minute (rpm). The reactor was heated using approximately
9,000 watts electrical heater (208V, 3-phase), controlled via a temperature
controller which received as its input the internal soil temperature. The entire
reactor system was maintained under anoxic conditions through the use of a
nitrogen purge. The rotating reactor was connected on either side to a 2-inch
piping which in turn was connected to a rotating union, which enabled the reactor
to turn while the downstream and upstream piping and connections did not.
Soil mixed with sulfur was placed to the first chamber. Sulfur was added to the
second. The first chamber was baffled along the horizontal axis with 2-inch
baffles, to pick up and mix the soil /sulfur mixture. The second chamber was
baffled vertically with 6-inch baffles, to hold up gases passing through it and to
provide surface area for contact between the gas stream and the molten sulfur so
as to facilitate reactions between the sulfur and organics in the vapor phase. The
second stage reactor was designed to provide between three and five minutes
residence time.
Nitrogen gas was fed to the first chamber, at between 1 and 2-L/min flow rate to
maintain an oxygen-free chamber. The nitrogen purge gas picked up water vapor,
volatile organics, hydrogen sulfide, elemental sulfur vapor and other volatile
compounds. The purge stream passed through the soil reactor and into the sulfur
reactor. Inside the sulfur reactor, which was heated to the same temperature as
the soil reactor, sulfur reacted with some of the organic compounds volatilized
from the soil. The purge stream passed through the sulfur reactor and out via a
heated line to a series of traps, which were designed to condense the water vapor
10
-------
FIGURE 3. ROTARY SOIL REACTOR.
FLANGE BOLTS NOT DRAWN (OVERALL LENGTH -29")
. JACKING
THREADS
11/2" PIPE,
THREADED
ON END
SOIL SIDE
(INLET)
SULFUR SIDE
(OUTLET)
RODS CONNECT
ALL BAFFLES TO
OUTER FLANGE
316 SS FLANGE
(~ 11" O.D.)
FOUR BAFFLES,
EVENLY SPACED
BAFFLES
(1/8" THICK)
HIGH
TEMPERATURE
GASKETS IN ALL
FLANGES
CROSS SECTION
8" 316 SS PIPE, SCHEDULE
40 (FLANGES NOT DRAWN)
6" TO
BOTTOM OF
BAFFLE
-------
and the volatile organics present in the gas stream. The first trap consisted of a
6-inch diameter closed 316 SS pipe, with a dip leg reaching to within 6 inches
from the bottom. The trap was maintained in an ice bath at approximately 2°C.
This trap collected the majority of the volatile compounds, particularly water vapor.
The second trap was maintained using dry ice at -20°C. This trap collected the
remaining water vapor, organics, and carbon disulfide.
The gas stream passed through the second trap and into a 4-L caustic scrubber,
containing 10N NaOH. The scrubber removed hydrogen sulfide gas. The gas
stream passed through this scrubber and was vented to the atmosphere-via a
hood.
The reactor was operated in batch mode, similarly to the full-scale unit envisioned
under this configuration. However, for a full-scale unit, the sulfur and soil sides
would likely be separated, to enable independent control of the heating systems {to
maintain the sulfur temperature! and to allow for isolation of either side should it
be necessary.
2.5 DATA AND SAMPLE COLLECTION
All data and sample collection was performed in conformance with the approved
project Quality Assurance Project Plan (QAPP), unless noted below.
2.5.1 Feed Characterization
Feed materials for the early runs consisted of spiked samples, made with reagent
grade chemicals, to form standard soil mixtures. Larger scale tests employed field
samples from a manufactured gas plant site, as described in Section 3.3. Samples
of this material were analyzed using ultrasonic extraction (Method 3550) and
GC/FID analysis (Method 8100). Duplicate samples of representative materials
were analyzed to determine the feed soil concentrations. Duplicate FID runs were
performed to determine the precision of the analytical procedure, which was found
to be within the QAPP standard of ± 15% (see Appendix B). 1,2,4-trichloro-
benzene was used throughout the experiments as an internal recovery standard.
2.5.2 Product Sampling
Products were collected from the traps, scrubber and reactor. All the material
present in the traps was used in the extraction. If the scrubber contained less than
100 mL of material, then the entire contents of the scrubber were used in the
analysis. If the scrubber contained more than 100 mL, then a 100 mL sample of
the scrubber water was taken for analysis. For the small-scale tests, reactor
samples included the entire reactor product. For the larger-scale tests,
representative samples were by homogenizing the reactor contents in a Vee-
blender, then removing 5 mL from three different locations within the blended
material.
2.5.3 Data Collection
Data such as temperature and pressure were read directly from the thermocouples
or pressure gauges which were part of the equipment. Gas flow rates were
determined using calibrated rotameters, and are reported at standard temperature
and pressures (25°C and 1 atm) unless otherwise noted.
12
-------
2.6 QUALITY ASSURANCE/QUALITY CONTROL
As noted, the project was conducted in conformance with an approved QAPP [2J.
Critical measurements made during the process runs include the TCLP of the feed
and products for runs involving heavy metal contaminated feeds, for runs
involving organic contaminants, critical measurements included semivolatiles
content of the feed, reactor product, and off-gases. The off-gas semi-volatile
content was typically measured by summing the contributions from each
component of the off-gas collection system, including the ice trap(s) and scrubber.
Quantitative QA objectives for critical analyses for the project are noted in Table 1.
The two major QA objectives for the project were (1) to determine the input
concentration and destruction efficiency of various semi-volatile organic
compounds during process tests and (2) measure the TCLP leachability of the
untreated and treated soil to determine whether or not the treated soil passed the
TCLP leachability tests.
As specified in the QAPP, CHMR performed duplicate analyses of 10% of the
critical analyses for samples taken using the rotary reactor. Duplicate analyses of
the screening tests (performed using the vented stirred reactor) were not
performed unless inter-experiment anomalies were noted. Results from select
duplicate analysis performed as part of the quality assurance procedures are given
in Appendix B.
Data reduction procedures primarily included summing the various components of
product streams to determine destruction efficiency. Specific reduction methods
are discussed in more detail in Section 5.
13
-------
Table 1. Quantitative OA Objectives
CRITICAL
MEASUREMENT
MATRIX
TCLP
REG. LIM.
METHOD
DET. LIM
(mg/L)
PREC.
(%)
ACCU.
<%>
COMPLE
TENESS
Toxicity
Characteristics
Leaching Proced.
S
Part 268,
Appendix 1;
Method 1313
Arsenic
S
5 mg/L
SW 7060
EPA 7000
.005
±15
±20
95
Barium
S
100 mg/L
SW 6010
EPA 7000
.01
±15
±20
95
Cadmium
S
1 mg/L
SW 6010
EPA 7000
.01
±15
±20
95
Chromium
s
5 mg/L
SW 6010
EPA 7000
.02
±15
±20
95
Lead
s
5 mg/L
SW 6010
EPA 7000
.05
±15
± 20
95
Mercury
s
.2 mg/L
SW 7470
EPA 7000
.005
±15
±20
95
Nicke!
s
5 mg/L
SW 6010
EPA 7000
.02
±15
±20
95
Semivolatile
Organic
Compounds
(examples):
L:SW3510
S:3540/3550
Product:3660
Naphthalene
S/L
NA
SW8100
.3 mg/kg
±15
±20
90
Anthracene
S/L
NA
SW8100
.3 mg/kg
±15
±20
90
Trichlorobenzene
S/L
NA
SW8100
.3 mg/kg
±15
±20
90
Volatile Organic
Compounds
(examples):
L:SW3510
S:3540/3550
Product:3660
Toluene
S/L
NA
SW8010/
8020
5 ug/kg
±15
±20
90
Xylene
S/L
NA
SW8010/
8020
5 ug/kg
±15
±20
90
Tetra-
chloroethylene
S/L
NA
SW8010/
8020
5 ug/kg
±15
±20
90
-SW refers to "Test Methods for Evaluating Solid Waste." SW-846, Third Edition.
14
-------
3.0 SOIL SAMPLES
The initial screening tests employed several standard soil mixtures which were
used to prepared various spiked mixtures for use in individual laboratory tests. A
field sample was obtained for the larger scale process tests. The characterization
of these various test mixtures and samples used in the project are discussed in this
section.
3.1 SARM SAMPLES
One set of test soil blends used in this project were the SARM-II and SARM-III
samples prepared by EPA and Enviresponse, Inc. The SARM samples are based on
the same SSM soil blend (discussed below) that has had a number of heavy metal
and organic compounds (both volatiles and semivolatiles) added. The SARM
samples had been prepared by addition of: arsenic trioxide (As203), cadmium
sulfate (CdS04), chromium nitrate (Cr(N03)3), copper sulfate (CuS04), lead sulfate
(PbS04), nickel nitrate (Ni(N03)2), and zinc oxide (ZnO).
For the purpose of the initial screening experiments on metal stabilization, the
SARM-III sample was used only to study the stabilization of the contained heavy
metals. No analyses of the contained organic compounds were made on the
products from the SARM-III runs.
3.2 SPECIAL SOIL BLENDS
Most of the experiments were performed on synthetically prepared soil blends
containing weighed quantities of heavy metals or semivolatile organic compounds
specifically added for each series of experiments. Two different soil blends were
used in this experimental program as the starting materials from which the heavy
metal and organic spiked samples were prepared.
One of the soil blends used in this work was the EPA Synthetic Soil Matrix (SSM)
prepared and distributed by Foster Wheeler Enviresponse, Inc. of Edison, NJ. This
material is the unspiked blend used to prepare the SARM and SSM samples. This
material was supplied in a 55 gal drum. Fresh samples were prepared by spreading
approximately 1 to 2 kg in a shallow pan to air dry overnight. After drying, the soil
was screened with a 9 mesh screen to remove oversized material. The screened
sample was stored in screw top glass jars.
Combustion tests on the SSM material, after acidifying to remove carbonate
carbon, indicated a total organic carbon content (TOC) of 0.5% or less. It is well
known that the capacity of soils to adsorb semivolatile organic compounds relates
to the organic carbon content. Accordingly, a second soil blend was prepared and
used for most of the test program. The second soil blend consisted of 75% by
weight of the SSM soil described above and 25% by weight of horticultural
topsoil. Each of these two materials were dried overnight by spreading the
material in a shallow pan. Each dried soil was separately screened to remove
oversize material with a 9 mesh screen and stored in screw top glass jars.
One kilogram samples were prepared by weighing out the dry soil blend
components, as well as the added spike materials, and placing them in a screw top
quart glass jar. The mixture was vigorously blended by shaking and rotation of the
closed jar. In other cases, larger samples were prepared using a commercial
laboratory-size Vee blender.
15
-------
3.3 MANUFACTURED GAS PLANT SAMPLE
A manufactured gas plant (MGP) sample was obtained from a utility in upstate
New York. The sample was obtained from a storage pile and was selected by the
utilities's contractor to be a heavily contaminated sample. Approximately 15
gallons of the material were obtained from the site. The soil was dark brown, with
a strong "coal tar" odor, and obvious small chunks of tar.
A MGP sample was selected for the larger scale rotary soil reactor runs since these
sites typically have high levels of higher boiling aromatics hydrocarbons but very
little VOCs. There are also over 1500 manufactured gas plant sites across the
country. It was therefore felt to be an appropriate test feed for demonstrating
performance of organics destruction in the Sulchem Process.
Table 2 summarizes the characterization of the MGP sample:
Table 2. Characterization of MGP Sample
Moisture
20.1%
Total Extractable Organics
2.8%
Particle Size Distribution
<4 mesh: 51.1%
4-10 mesh: 20.7%
10-20 mesh: 14.7%
20-60 mesh: 11.2%
>60 mesh: 2.2%
Oroanic Compound Tvoes:
naphthalene, and Cr, C2-, C3- substituted
dibenzothiophene, and Cr, C2-, Ca- substituted
fluorene, and C,-, Cz-, C3- substituted
phenanthrene, and Cr, C2-< C3- substituted
pyrene, and Cr and C2- substituted
chrysene, and C,- and C2- substituted
dibenzoanthracene
benzopyrenes and benzofluoranthenes
pristane
phytane
1253 //g/g
423 //g/g
623 //g/g
1626 //g/g
1343 //g/g
605 //g/g
30 //g/g
543 //g/g
366 //g/g
256 //g/g
The organic hydrocarbons present in the MGP soil sample were determined by
extraction and GC/MS employing a variation of Method 8270 used in this
laboratory for quantifying petroleum degradation found in bioremediation product
evaluation 19]. This method determines alkanes from C10 to C35 and various
polycyciic aromatic hydrocarbons including totals for mono-, di-, and tri-methyl
derivatives based on the sensitivity of the unsubstituted homolog.
The hydrocarbons found ranged in boiling point from methyl naphthalenes to
benzopyrenes. The major compounds were methyl naphthalenes (mono, di and trig
phenanthrene, pyrene and other similar polycyciic aromatic hydrocarbons including
the alkylated homologs. The only other major hydrocarbons observed were
pristane, phytane and several analogs (C18 to C21). Minor amounts of higher
n-alkanes were also present. The major compound classes found in this sample
16
-------
and quantified by the GC/MS method used, for which standards were available, are
summarized in Table 2.
Several samples of the MGP soil were also oven dried and air dried at room
temperature for two of the test runs to evaluate the effect of moisture level.
Moisture contents of these separate feed samples were also determined.
17
-------
4.0 ANALYSES
This Section references the standard methods used for the analyses performed for
the test runs for these projects. These include measurement of the organics in the
test blends, as well as the MGP test samples, and comparison with the organics
recovered form the reactor and associated traps to determine organics destruction.
In addition, metal stabilization was determined and compared to the metal spiked
feed test soils and compared with the raw reactor soiids product.
4.1 ORGANIC ANALYSES
In the experiments where destruction and removal of organic compounds were
studied, separate extractions were made of the raw reactor solids, of the ice trap
condensate, and of the caustic scrubber solution. These extractions, along with
rinsings of the appropriate equipment, were made in order to determine the
quantity of any unreacted, or desorbed/distilled, organic compounds from the
reaction products or equipment hardware.
Most of the hydrocarbon and chlorohydrocarbons used for these tests were
semivolatile organics and were analyzed by gas chromatography outfitted with
flame ionization detector (GC-FID) and chemical assignments confirmed by gas
chromatography-mass spectrometry (GC/MS). The standard EPA analytical
methods employed for these tests are listed in Table 3.
4.2 METALS ANALYSES
For those experiments in which metals stabilization was being tested, samples of
the soil blend used, as well as the raw reactor solids product were analyzed by the
Toxicity Characteristics Leaching Procedure [(TCLP) - EPA Method 1311]. (see
Table 3) Following acid digestion (Method 3005), the leachate was analyzed by
Inductively Coupled Plasma (ICP) Atomic Emission Spectroscopy (Method 6010)
for the test metals which included: arsenic, barium, cadmium, chromium, cobalt,
copper, lead, nickel, and zinc.
4.3 OTHER ANALYSES
Mineralogical testing and' speciation was employed using optical microscopy. X-ray
diffraction, and scanning electron microscopy with energy selection detection.
18
-------
Table 3. Analytical Methods
Type of Analysis
Methods
Samples
ORGANICS DESTRUCTION
Extraction
Soxhlet (3540)
Sonication (3350)
soil feed
raw reactor solids
Liquid-Liquid (3510)
condensate
scrubber solution
Clean-up
Sulfur removal (3660)
raw reactor solids
condensate
scrubber solution
Florisil column
(3620)
soil feed
raw reactor solids
Gas Chromatography
FID (8100)
GC/MS (8270)
soil feed
raw reactor solids
condensate
scrubber solution
METALS IMMOBILIZATION
TCLP
Method 1311
soil feed
raw reactor solids
ICP Atomic Emission
Spectroscopy
Method 3005
Method 6010
soil feed
raw reactor solids
Moisture
Oven drying
soil feed
Particle size
distribution
Microtrac analysis
soil feed
Total extractable
organics
Method 3550
soil feed
reactor feed
19
-------
5.0 ORGANICS DESTRUCTION
The Sulchem Process is conceived to be a two-stage process in which the soil is
reacted with sulfur in the first stage. Unreacted organics desorbed from the first
stage react with sulfur in the vapor phase in the second stage. In this Section,
preliminary tests of organic destruction in the soil reactor and in the vapor reactor
are reported.
The purpose of these tests were: (1) to establish the soil treatment temperatures
and other process conditions necessary to achieve organic compound destruction
in the soil reactor; (2) to estimate the boiling range of volatile organics that will be
desorbed during heat up of the soil reactor before reaction temperatures are
reached; and (3) to evaluate process conditions for a second stage vapor reactor to
treat the volatile organics desorbed from the soil reactor.
5.1 SOIL BATCH REACTOR TESTS
In batch reactor tests of the reaction mixture, the soil and sulfur are heated, while
mixing, from ambient temperature to the desired reaction temperature and then
held at the run temperature for the desired time interval. The temperature at which
significant reaction first occurs was estimated, based on small batch tests with
mineral oils in the closed and vented mode reactors, to begin in the range of
200°C to 250°C for the more reactive saturated hydrocarbons. Therefore, the
initial screening tests were run at temperatures of 250°C and higher.
A comparison of the boiling range of the organics in the feed soil with the boiling
range of the desorbed vapors collected in downstream traps indicated the
approximate temperature range over which desorption and reaction compete in the
solids reactor. For example, those compounds with high recoveries in the
overhead represent the boiling range where desorption takes place before any
reaction occurs. Those compounds not found in the overhead represent the boiling
range above the reaction temperature regime. An intermediate boiling point
regime, where only partial recovery is found, represents the boiling range where
reaction and desorption processes compete.
Therefore several scoping tests were performed in the stirred reactor using a series
of organic compounds of successively higher boiling points to: 1} establish the
boiling range where desorption occurred before reaction (thereby establishing the
approximate threshold reaction temperature); and 2) provide an initial screening of
the level destruction, and effect of process variables, for the higher boiling
components where desorption is not important.
5.1.1 Test of Desorption versus Boiling Range
Several initial screening tests were done with topped crude oil mixed with soil to
establish the temperature limits for desorption on heat up of the soil. The crude oil
was used because it contained compounds with boiling points over a wide
temperature range and provided a qualitative estimate of the temperature at which
the reaction predominated over desorption. In these tests, GC scans of the
recovered condensate, and extracts of the residual oil and the feed material, were
compared for the carbon numbers of the n-paraffin peaks in each fraction.
Running at 250°C, the transition from boiling range of the overhead to the residue
(in a run without sulfur) corresponded closely to the reactor operating temperature
(e.g., n-C14, BP 254°C). A comparison experiment in the presence of sulfur did
20
-------
not show significant decrease in the boiling range of the overhead suggesting that
these hydrocarbons did not react significantly below 250°C.
Next, a soil biend was prepared that was spiked with nine aromatic compounds
(2000 ppm each, except pyrene at 1000 ppm) with boiling points from 165°C to
393°C mixed into the 75/25 SSM top soil blend. Several series of three sets of
runs were carried out at 250°C, 350°C, and 440°C for 2 hours with 200 g soil
and 30 g sulfur. Analysis of the recovered samples for the spiked organic
compounds was compared with a sample of the original spiked soil. The results
are presented in Table 4. Surrogate recoveries for all recovered fractions were in
the 60 to 90% range.
In the experiment, a slow stream of nitrogen flowed through the reactor into an ice
trap and then through a caustic scrubber before release to the hood. As described
in Section 2.1.4, after the run, the reactor system is disassembled and three
sample fractions, including equipment rinsings, are removed for analysis. These
three product fractions are the raw reactor solids, the condensate in the ice trap,
and any organic compounds extracted from the caustic scrubber. As expected,
the ice trap condensate represents most of the desorbed material recovered.
The combined recovery from the three product fractions were compared with the
analytical results from the extraction of the feed soil {based on a 200 g charge
which nominally contains 400 mg of each compound except pyrene and 200 mg of
pyrene) to arrive at a total recovery for each compound. This procedure corrects
for losses in the experiment, sampling, and analysis. It should also be noted that
absolute recoveries of mesitylene and durene in both the soil charge and product
analyses were about 70% and below the values of 85 to 90% observed for the
higher boiling compounds spiked in the soil. These compounds are technically not
semivoiatile organics (defined as BP > 200°C) and are therefore outside the
method range for the Soxhlet extraction (Method 3540) used for the sample work
up and therefore recoveries were lower.
These tests measured the performance of two processes which are in competition:
desorption and chemical destruction. The relative importance of these two
processes vary with the boiling point of the constituent. The five lower boiling
compounds show relative recoveries of about 85-90% for all three runs. Based on
the relatively constant recovery factor observed for these compounds, it is
concluded that these represent thermal desorption only with negligible effects of
chemical destruction. The decreased total recoveries for the four higher boiling
compounds sfjow the effect of the competition of chemical destruction with
desorption. Bibenzyl and hexachlorobenzene still yield significant recoveries in the
overhead fractions with destruction representing about 20% of the total compound
fed. As the boiling point is further increased, the fraction of feed compound that is
destroyed increases markedly.
Table 4 shows three different measures of process performance: recovery from the
overhead, destruction, and destruction and removal efficiency (DRE) of the treated
soil. These measures are discussed further.
21
-------
Table 4 Effect of Boiling Point on Destruction and Removal
4a. Reactor run at 260 C and 2 Hour Residence Time
Destruction &
Compound
BP
Charge
Reactor
Recovery In
Recovery in
Percent
Percent
Removal
(nig)
Residual
tee Trap
Scrubber
Total
recovered
Destroyed
Efficiency
mesitytene
165
276.9
0.15
225.6
0.35
226
81.6%
18,4%
99.8%
durene
197
276,2
0.15
249.3
0.08
249.53
90.3%
9.7%
99.9%
naphthalene
218
343.7
0.65
287.9
0.17
288.72
83.8%
16,0%
99.8%
2 Me Naphthalene
241
353
1.07
276.9
0.06
278.03
78.6%
21.2%
39.7%
biphenyl
264
368.6
3.11
295.8
0,06
298.97
80.3%
18.9%
99.2%
btbenzyl
285
356. B
3.87
229,2
0.00
233,07
64,2%
34.7%
98.9%
hexachlorobenzene
322
327.5
8,60
204.4
0.36
213.36
62.5%
34.9%
97,4%
anthracene
340
306.1
7.32
64.0
0.00
71.32
20,9%
76,7%
97,6%
pyrene
383
176
2.56
6.9
0,00
9.46
3.9%
94.6%
98.5%
4b Reactor Run at 3B0 C and 2 Hours Residence Time
Destruction
Compound
BP
Charge
Reactor
Recovery in
Recovery In
Percent
Percent
Removal
!mg)
Residual
Ice Trap
Scrubber
Total
recovered
Destroyed
Efficiency
mesitylene
165
276.9
1.13
220.3
2.98
224.41
80.6%
19.0%
99.6%
durene
197
276.2
0.27
245.6
5.09
250.96
90,6%
9,1%
89.9%
naphthalene
218
343.7
0.2
286.3
8.26
304.75
88,6%
11.3%
99.9%
2-Me Naphthalene
241
353
0.15
281.8
7.24
289.19
81,9%
18.1%
>99,9%
biphenyl
254
368.5
0.32
238.4
7.68
306.4
83,1%
16.9%
89.8%
bi benzyl
265
356.8
0.28
221.5
4,67
226.45
63,4%
36,5%
99.9%
hexachlorobenzene
322
327.5
17.66
96.5
2.11
118.27
30.1%
64,5%
94.6%
anthracene
340
306.1
9.15
23.3
0,59
33.04
7.8%
89.2%
97,0%
pyrene
393
175
2.71
2.3
0.04
5.05
1.3%
97.1 %
88.5%
4c. Reactor run at 440 C and 2
Hour Residence Time
Destruction
Compound
BP
Charge
Reactor
Recovery In
Recovery In
Percent
Percent
Removal
(mg|
Residual
loe Trap
Scrubber
Total
recovered
Destroyed
Efficiency
mesitylene
165
276.9
0.27
243,8
4.24
248.31
88.6%
10.3%
99.9%
durene
197
270.2
0.08
245.8
10.15
256.03
92,7%
7.3%
>99.9%
naphthalene
218
343.7
0.03
280.2
12.46
292.77
85,2%
14.8%
>99.8%
2-Me Naphthalene
241
353
0.08
269.2
8.22
277,5
78.0%
21.4%
>99.9%
biphenyl
264
368. E
0.15
282.5
7.64
280.29
78.7%
21.2%
>99.0%
bibenzyl
285
366.8
0.08
228.8
4.87
233,76
65.5%
34.5%
>99.9%
hexachloro benzene
322
327.6
0.49
205,4
3.56
209.45
63.8%
36.0%
99.9%
anthracene
340
306.1
0.35
69,0
1.22
70.57
22,9%
76.9%
99.9%
pyrene
393
175
0.67
12.7
0.14
13.51
7,3%
82,3%
99.6%
22
-------
Recovery from Overhead
The percentage recovery from overhead is a measure of the amount of a
compound which is desorbed from the soil rather than reacted. It is
calculated as the ratio between the total amount of the compound recovered
in the ice traps and scrubber to the amount originally present in the soil.
Destruction
The percentage destroyed is a measure of the effectiveness of the sulfur in
the reactor. The percentage destroyed was calculated simply as the
difference between the amount originally present in the soil and the amount
recovered both in the overhead and soil fractions, divided by the amount
originally present in the soil.
Destruction and Removal Efficiency (DRE)
The DRE is a measure of overall removal of each compound from the soil.
The DRE is calculated by subtracting the ratio of the amount of residual
compound left in the soil to the amount originally present from one. For the
lower boiling point compounds, it reflects primarily the effect of desorption.
For the higher boiling compounds, it reflects reaction efficiency.
The differences in the recovery between hexachlorobenzene and anthracene, which
have similar boiling points, are attributed to the expected much slower reaction
rate for hexachlorobenzene with sulfur due to lack of hydrogen. Other research in
this laboratory on reaction of sulfur with various chlorohydrocarbons has suggested
the thermodynamics for hexachlorobenzene reaction with sulfur would be much
less favorable [10].
Reaction temperature is not seen to greatly affect the recoveries as the desorption
step removes the compound from further opportunity to react in the simple batch
reactor system. The residual content of the treated soil is nonetheless greatly
reduced, corresponding to DRE values of better than 99% at the higher reaction
temperatures.
These same observations are noted whether the reaction temperature is 250°C,
350°C, or 440°C. The reduction in amount remaining in the reactor at 440°C for
the higher boiling compounds, e.g., pyrene, suggests that reaction temperatures
above 350°C jnay be needed to obtain complete destruction of the non-volatilized
compounds in the soil.
These results suggest reaction begins about 250°C, or slightly less (there may be
a slight amount of destruction for methyl naphthalene and biphenyl). The lower
temperature limit for reaction can not be determined without an additional
experiment at a lower temperature, since the time for volatilization during the ramp
up to 250°C is much shorter than the two hour reaction interval at 250°C. Thus,
desorption can occur before there is opportunity for reaction to occur at the lowest
possible reaction temperature.
As discussed in the QAPP, duplicate analyses of some of the analyses were
performed. Results of these analyses are presented in Appendix B.
23
-------
5.1.2 Effect of Process Parameters
Preliminary tests were made of other process variables in additional runs at 250°C.
These tests used only the four higher boiling aromatics listed in Table 4. The
results are given in Table A-1 in Appendix A. These runs evaluated reaction times
of 0.5 and 1 hour as well as sulfur contents of 4.8 % and 9.1 % for comparison
with the run in Table 4 (2 hours and 13% sulfur). It was found that decreasing
reaction time had a greater effect than decreasing sulfur/feed ratio, as might be
expected, by increasing the residual organics in the reactor solids and decreasing
the quantity carried over. Reaction times less than 2 hours may require
temperatures higher than 250°C.
5.2 VAPOR REACTOR
Results are presented for destruction of toluene in the vapor reactor in order to
establish conversion efficiency as a function of process variables. From these
results, a kinetic model is developed in order to predict process conditions
necessary to achieve a satisfactory destruction efficiency.
5.2.1 Test Results
Twelve experiments were conducted using the vapor reactor system. The major
parameters varied included toluene inlet vapor concentrations, residence times, and
operating temperatures. In addition, the height of the liquid sulfur in the system
was varied (affecting primarily the residence time), as well as the type of sparging
inside the reactor, which was found to have little overall effect on the reaction.
Several preliminary experiments were performed to shakedown the experimental
equipment and procedures. During these preliminary experiments, only weight gain
to the trap was measured and no gas chromatograph {GO analyses were
performed. Therefore, accurate outlet toluene concentrations were not obtained.
The data from experiments in which GO analyses were performed are presented in
Table 5. The data in Table 5 is sorted by temperature, and gives the run number,
residence time, inlet and outlet toluene concentrations in the overall feed stream in
parts per million (molar ratio), and percentage toluene destroyed.
The inlet toluene concentrations were determined by finding the mass difference in
the container of toluene feed before and after each run. This was converted to a
molar amount.and divided by the total moles of nitrogen fed to the system
(calculated assuming the ideal gas law) to determine the toluene concentration.
The outlet toluene concentrations were determined by GC analyses of the material
found in the ice traps. Again, the toluene effluent was divided by the overall gas
flow rate to determine concentration.
24
-------
Table 5. Vapor Reactor Results
Reactor
Temp{°C)
Run
#
Resid.
Time
(sec)
Inlet
Toluene
(ppm)
Outlet
Toluene
(ppm)
Destruction
Efficiency
(%)
300
8
149
5847
1494
74
300
12
150
8233
3345
41
300
15
262
4711
1675
64
300
17
543
5648
1303
77
350
10
138
9262
5455
41
350
11
138
16983
11488
32
350
13
138
8902
4220
53
350
14
118
7719
2676
65
350
16
241
4502
1152
74
350
18
500
5105
760
85
350
19
667
5217
217
96
400
20
216
3449
1943
44
400
9
127
7719
2161
72
5.2.2 Model for Vapor Reactor System
It was anticipated that the vapor reactor would operate with first order kinetics,
which would imply a zero-intercept linear relationship between the logarithm of the
output/input concentration ratios and the residence time, as shown below:
In (C/C0) = -kt
C is the concentration of toluene in the vapor stream as a function of time, and C0
is the initial concentration of the toluene fed to the reactor, t is residence time (in
seconds), and k is the rate constant.
The rate constant k is a function of temperature according to the Arrhenius
relationship:
k = A exp (-Ea/RT)
in which A is a reaction constant, Ea is the activation energy for the reaction {in
cal/gmol), R is the universal gas constant, 1.987 cal/gmoI-°K, and T is the reaction
temperature (in °K).
Figure 4 graphically depicts the first order data which were used to determine the
rate constants at each reactor temperature. The figures show the function
ln(C/C0) plotted against residence time. Figures 4a and 4b show the data obtained
25
-------
4a, Least-Squares Fit: ln(C/Q) vs Res Time; 3Q0°C
o "l
—2
-15
soo, see *rac
residence time (sec)
predicted In (C/CO) = -let;
„ InC/CO 300 C
k = 0.0030/sec
4b. Least-Sgaares Fit: In (C/C,) vs. ResUme; 35(y*C
-i
o
o
100
600
TOO
H InC/CO 350 C Predicted In (C/C0) = -kt;
k « 0.dQ456/sec
4c, Average k vs Residence Time; 400°C
o
0-2
y
O
jj s
-4
300 . «8 , .500
residence time (sec)
m
300
0
90S
B InC/CO 400 C predicted In (C/CO) = -kt;
k - 0.0063 /sec
FIGURE 4. FIRST ORDER KINETICS
FOR TOLUENE DESTRUCTION
26
-------
at 300 and 350°C, respectively. The data were fitted to a line, using a standard
least-squares fit, with a zero-intercept. The data from run #8, which can be seen
to be quite distant from the line in Figure 4a, was not used in the least-squares
algorithm. This was the first run performed during the experiment, and the sulfur
vapor reactor was inadvertently not purged prior to the test, which may have led
to erroneous results. The rate constants calculated at 300 and 350°, were
0.0030/sec and 0.0046/sec, and the R-squared coefficients for the lines were
0.79 and 0.89, respectively. Because only two points were obtained at 400°C, a
least-squares fit was impossible. However, k-values could be calculated for both
runs, and these were averaged to obtain an overall k at 400°C of 0.0063/sec.
This value was used to plot a line on Figure 4c, which falls approximately halfway
between the two data points at that temperature.
The Arrhenius equation can be rearranged to yield a linear relationship between in k
and the inverse of temperature;
In k = -Ea/R * 1/T + in A
The natural logarithms of the k values obtained at the three reaction temperatures
were plotted against the inverse of that temperature in Figure 5. The points were
remarkably close to linear, and a least-squares fit yielded the following values:
In k = -2926/T - 0.71
or, Eu = 5814 cal/gmol and k = 0.49 exp (-2926/T), with T in °K and k in sec"1,
which is the required result. The R-squared for this fit was 0.99, indicating that
the three points are collinear as predicted. (The R-squared value was coincidentally
higher than anticipated given the uncertainties in the data.) Figure 6 shows the
empirical fit for the data, calculated versus actual destruction, with the calculated
destruction determined using the Arrhenius relationship determined above.
5.2.3 Residence Times and Reactor Sizes
Based on the first order model and extrapolations from data obtained at lower
temperatures and destruction efficiencies, the residence time required for 99%
destruction of toluene at the three tested reactor temperatures, and one additional
higher temperature, were determined;
Table 6. Calculated Residence Times
Temperature (°C)
Residence Time (sec)
300
1535
350
1010
400
730
500 (extrapolated!
419
27
-------
These residence times were approximations only, but were calculated to determine
the rough order of magnitude for the reactor volume and residence times. An
approximate breakdown of the effluent gases can be predicted assuming a 10,000
kg/hour unit, which processes soil containing 10% {by weight) moisture and 0.5%
by weight organics (MW= 100), 75% of which is converted in the soil and 25%
of which are volatilized. If minimal nitrogen purge is assumed, the waste stream
will consist of:
28
-------
Arrhenius Plot of Toluene Destruction
-4.5
'5
-5.5
6
-1
i,
1.4
1,9
H based on linear regressions from experimental runs
Predicted In k = — E/RT + A
E = -5800 cal/gmol; A = -0.71; R = 1.987 caf/gmoI-K
Figure 6
Toluene Destruction — Calculated vs. Actual
12
1«
I
ffl
4
2
0
^ Thousands
actual destruction, ppm
o
2
4
1#
12
B Experimental Data
Predicted 1st order fit In (aC0)=-kt;lc=0.49 exp(-2926/r)
29
-------
Table 7. Vapor Stream
Constituent
Molar Row
Water
50 kg-moles/hr
Hydrogen sulfide
3 kg-moles/hr
Organics
0.2 kg-moles/hr
Nitrogen (purge)
6 kg-moles hr
TOTAL
60 kg-moles/hr
The vapor reactors will need to be sized to handle the above flow rates. The
following table shows the required size of the vapor reactors containing the whole
flow, or the flow without the water vapor (assuming it is condensed out of the
stream), at various operating temperatures. These estimates are developed
assuming that the gases are ideal, at one atmosphere pressure.
Table 8. Calculated Vapor Reactor Volume
Temperature (°C)
Volume (m3) assuming all
gases in reactor
Volume (m3) assuming
water is removed
300
1200
200
350
860
170
400
670
135
500
440
90
The key result from this analysis was that the required reactor volumes are
extremely large for transportable units, even assuming that the water vapor is
removed. Obviously, the reactor volumes could be decreased by increasing the
pressure at which the reaction is conducted, but then a pressure vessel,
compressors, and other processing equipment would be required- This option
would have to be economically and technically evaluated to find the optimal
feasible pressure.
Based on these results, the vapor section (freeboard volume of the second section)
of the rotary soil reactor (Section 2.4) was designed to provide a nominal vapor
residence time of between three and five minutes. This would allow for
destruction of organic compounds in the range of 40 to 85%, for temperatures of
300 to 400°C respectively, while not being too large as to be unwieldy for the test
application.
30
-------
6.0 METALS STABILIZATION
in this section, processing tests of various metal-spiked soils are discussed to
determine the specific heavy metals, the applicable concentration range, and the
process conditions, for which the Sulchem Process will immobilize the metals.
6.1 TESTS WITH DIFFERENT METALS
A series of batch screening runs were made on several soil blends containing
various heavy metals. The purpose of these initial tests were to find how well
each of the various heavy metals responded to the sulfur treatment of the Sulchem
Process. A priori one might expect that the stabilization mechanism might be the
formation of metal sulfides as an insoluble coating. If this is the case, then the
heavy metals whose sulfides are soluble in acid (i.e., chromium, cobalt, iron,
nickel, and zinc) might not be rendered immobile as much as other metals (e.g.,
lead) with the Sulchem treatment since the TCLP test leaches the sample with a
buffered acetic acid solution. It was for the purpose of examining this premise that
the initial metal screening tests were made.
More detail on the effect of process conditions on TCLP reduction are described in
the next section. For the screening studies to compare the behavior of different
heavy metals reported here, the process parameters were sulfur to soil ratio,
reaction temperature, and reaction time. Generally one sulfur to soil ratio was
used (typically 0.15), a range of reaction temperatures were used (e.g., 250°C,
300°C, and 350°C), and the reaction time was typically one-half hour.
The initial metal screening tests were done on soil samples of SARM-III, as well as
prepared blends of metals spiked in either SSM soil or a 75/25 blend of SSM and
horticultural topsoil. The SARM samples had been prepared with arsenic trioxide
(As203), cadmium sulfate (CdS04), chromium nitrate (Cr(N03)3), copper sulfate
(CuS04), lead sulfate (PbS04), nickel nitrate (Ni(N03)2), and zinc oxide (ZnO).
Screening studies of various metals were made with a series of closed mode
reaction runs made on five separate metal/SSM mixtures. These individual blends
were made using lead oxide (PbO), cadmium oxide (CdO), arsenic oxide (As203),
chromium (111) oxide (Cr203), and nickel hydroxide (Ni(OH)2) to contain 1000 ppm of
each of the metals. TCLP analyses were made of the soil blend as well as the
three raw reactor products from each processing temperature. Two hour reaction
runs were used in these tests which were conducted in the closed mode unstirred
reactor. These data are listed in Table 9. It was concluded that both lead and
cadmium responded to the treatment, the latter more at elevated temperatures.
For example, cadmium had a TCLP of 38.9 mg/L. This was reduced to 8.3 mg/L
at the mildest conditions and at temperatures of 350°C reduced to below the
regulatory limit of 1.0 mg/L. The results on the arsenic and nickel were
inconclusive.
Next, SARM-III was run to evaluate the performance of the different contained
heavy metals. These runs were carried out in the stirred reactor autoclave to
ensure that adequate mixing was used. Since the objective was to determine
metals stabilization, no analyses were made on these runs for the contained
organic compounds also present in the SARM-III.
31
-------
Table 9. Tests of Various Metals vs. Reaction Temperature
Temp
Time
Pb
PbTCLP
Cd
CdTCLP
As
AsTCLP
Cr
CrTCLP
Ni
NiTCLP
°_c
Hrs
Sol
S/soil
ppm
mg/L
ppm
mg/L
ppm
mg/L
ppm
mg/L
ppm
mfl/L
Untreated blend
1000
14.4
1000
38.9
1000
3.07
1000
<0.05
1000
1.51
250
2
SSM
0.25
1000
0.69
1000
8.34
1000
5.97
1000
<0.05
1000
0.89
300
2
SSM
0.25
1000
0.58
1000
4.43
1000
8.23
1000
<0.05
1000
1.21
350
2
SSM
0.25
1000
0.20
1000
0.62
1000
11.7
1000
<0.05
1000
1.16
-------
Table 10 shows the results for the SARM-III feed. For the main metals present in
the SARM-III TCLP leachate {i.e., all but arsenic and chromium}, copper responds
the most effectively to the Sulchem Process decreasing the TCLP value 100-fold at
the mildest conditions and to the detection limit at the highest temperature,
presumably because of insoluble sulfide formation. In Table 10 a continuous
reduction of TCLP nickel leachate levels for the treated soils is shown at
successively higher temperatures, from 22,2 mg/L to 17.3 mg/L at 250°C to 0.4
mg/L at 440°C. The results for cadmium, lead, and zinc also demonstrate some
temperature effect, but not as extensive as observed for nickel.
Based on these results, additional test blends were prepared using the oxides or
hydroxides of lead, cadmium, nickel, and zinc to further evaluate the process
conditions necessary to improve the TCLP results on the treated soils for these
four metals. It was felt important to evaluate the stabilization of these metals in
soils with a higher organic carbon content than those employed above. Therefore,
the next series of tests was based on using the 75/25 SSM top soil blend, which
was used in all of the subsequent tests. These experiments were all run in the
stirred reactor.
The first four data sets in Table 11 shows runs on various blends of Cd, Pb, Ni,
and Zn as a function of temperature. The first three test blends were at relatively
high loading of metals to evaluate possible process limits for reduction of the
leachate to TCLP limits. The cadmium results demonstrate the previously observed
effect of process temperature on the reduction of TCLP, although the effect of
process temperature is much less for lead in this case. Substituting the more
soluble nitrate at the same lead loading demonstrates how the more soluble form
can prevent reaching the TCLP regulatory limit of 5 mg/L in this case. Although
somewhat higher TCLP levels are observed for the starting soil blend using more
soluble salts, the response of the soil to the Sulchem treatment shows comparable
reduction in TCLP values. The nickel and zinc spiked SSM topsoil blend did not
demonstrate as great an effect of process temperature effect previously noted for
these metals in the SARM blend (Table 10).
Based on the results from these three tests with high levels of added metals in
75/25 SSM/soil blend, as welt as the results from the SARM-III tests (Table 10), a
fourth blend was prepared at intermediate concentrations of Cd, Pb, Ni, and Zn and
run at various temperatures. This blend was prepared in order to more fully
challenge the process at the highest concentrations possible and yet still achieve a
passing TCLP for the treated soils.
The results for the fourth blend in Table 11 however turned out to not adequately
challenge the process for all of the metals. These experiments demonstrate
significant reduction of TCLP values for the treated soils for lead and, at elevated
temperatures, for cadmium. Zinc was not greatly reduced and the nickel results
were inconclusive due to too low a TCLP value for the starting soil. Tests of
additional blends with higher metal contents are needed to arrive at a more suitable
demonstration test mixture.
Finally in Table 11, a soil blend was then spiked to 10,000 ppm by weight of
cobaltous (II) oxide and run at several temperatures. In contrast with the other
metals, lower treatment temperatures actually enhance the leachability of the
cobalt from the added cobaltous oxide. This may be due to formation of a more
acid soluble surface on the particles.
33
-------
Table 10.
Test of SARM III Soil vs. Reaction Temperature
Untreated Soil Reaction Products
Run No.
27-77
27-78
27-79
27-84
27-92
Temp °C
250
300
350
400
440
Time hr
0.5
0.5
0.5
0.5
0.5
Sulfur %
13.0%
13.0%
13.0%
13.0%
13.0%
Contained
Metal
TCLP
TCLP
TCLP
TCLP
TCLP
TCLP
Metal
mq/ka
mcj/L
ma/L
mq/L
ma/L
mq/L
mg/L
As
500
0.21
0.16
0.23
0.37
0.21
<0.05
Cd
1,000
36.8
22.5
15.1
12.4
6.02
3.66
Cr
1,500
<0.05
0.07
0.12
<0.05
<0.05
<0.05
Pb
14,000
35,5
25.5
22.7
21.2
16.1
12.2
Ni
1,000
22.2
17.3
6.71
3.72
0.72
0.4
Cu
9,500
153
1.13
0.05
0.03
0.03
<0,01
Zn
22,500
791
628
361
162
58.1
32.4
34
-------
Table 11. Cd, Pb, Ni, Zn, and Go vs. Reaction Temperature
75/25 SSM/Top Soil Blend
Untreated Soil
Reaction Products
Run No,
Temp°C
27-69
250
27-71
300
27-70
350
27-76
400
27-91
440
Time hr
Sulfur %
0.5
13.0%
0.5
13.0%
0.5
13.0%
0.5
13.0%
0.5
13.0%
Metal
Salt added
Contained
Metal
mg/kg
TCLP
mg/L
TCLP
mg/L
TCLP
mg/L
TCLP
mg/L
TCLP
mg/L.
TCLP
mg/L
Cd
Pb
CdO
PbO
5,000
10,000
144
61.7
28.3
4,08
42.0
1.22
22.4
1.21
4,41
2.21
0,54
2.03
Run No.
Temp °C
Time hr
Sulfur %
33-40
250
0.5
13.0%
33-41
440
0.5
13,0%
Metal
Salt added
Contained
Metal
mg/kg
TCLP
mg/L
TCLP
mg/L
TCLP
mg/L
Cd
Pb
CdCI2
Pb(N03)2
1,000
10,000
30.1
96.4
4.98
29.2
0.31
7.61
Run No. 27—95 27-96 27-97
Temp °C 250 300 350
Time hr 0.5 0,5 0,5
Sulfur % 13.0% 13.0% 13.0%
Contained
Metal
TCLP
TCLP
TCLP
TCLP
Metal
Salt added
ma/kg
mg/L
ma A.
mg/L
mg/L
Ni
Ni(OH}2
2,000
0.54
1.95
2.5
4.7
Zn
ZnO
2,000
36.3
25.6
22.60
24,80
Run No.
33-12
33-13
33-14
33-15
Temp °C
250
300
350
400
Time hr
0.5
0.5
0.5
0.5
Sulfur %
13.0%
13,0%
13,0%
13.0%
Contained
Metal
TCLP
TCLP
TO LP
TCLP
TCLP
Metal
Salt added
mg/kg
mg/L
mg/L
mg/L
mg/L
mg/L
Cd
CdO
200
2.37
0.33
0.17
<0,01
<0,01
Pb
PbO
2,000
2.40
0,38
0.41
<0.10
<0.10
Ni
Ni(OH)2
500
0.097
0,20
0.35
0.08
<0.01
Zn
ZnO
500
6.76
3,49
4.60
0.80
0.11
Run No,
Temp "C
Time hr
Sulfur %
23-40
350
0.5
13.0%
23-42
440
0.5
13.0%
Metal Salt added
Contained
Metal
mg/kg
TCLP
mg/L
TCLP
mg/L
TCLP
mg/L
Co
CoO
10,000
15.0
65.8
11.3
35
-------
6.2 TESTS OF PROCESS PARAMETERS
The tests reported in the previous section include some process parameter
comparisons at different reaction temperatures. Additional process parameters
which were also studied included reaction time, process stoichiometry (sulfur to
soil ratio), and metal content. Based on the results of the metal screening runs
above, it was decided to concentrate these process tests for lead and cadmium in
the 75/25 SSM topsoil blend. A number of these test results are presented in
Table 12 along with several runs on a field sample containing lead. All
experiments were run in the stirred autoclave reactor.
6.2.1 Stoichiometry
The effect of reaction stoichiometry, as indicated by sulfur content of the feed,
was evaluated. From the data set in Table 12 all run at the same temperature and
reaction time (250°C and 0.5 hr), Figures 7 and 8 present a plot of the TCLP of
the treated soils for lead and cadmium as a function of sulfur content of the
reactor feed. A clear asymptotic dependence is noted as a function of sulfur
content of the reactor feed. The curves are based on a fit of TCLP results to
inverse sulfur content. The main point is that the asymptotic TCLP value is finite.
That is, increasing sulfur content beyond some point does not further reduce the
TCLP value of the treated soil. A sulfur content in the range of 15 to 20 percent
apparently is sufficient to reduce the TCLP of the treated soil as far as it will go.
The asymptotic TCLP value at higher sulfur contents however varies with the
heavy metal and the metal content. Thus for the more concentrated blend tested
in Table 12, a passing TCLP for cadmium (< 1 mg/L) on the treated soil could not
be achieved, at these processing conditions, whereas lead could be made to pass
(< 5 mg/L) with sufficient sulfur addition.
6.2.2 Temperature
Elevated temperatures further reduce the TCLP value somewhat although the
temperature effect for cadmium and nickel as an example (Tables 10, 11, and 12}
is greater for temperatures near the boiling point of sulfur (445 °C) whereas the
temperature effect is much less pronounced for lead.
6.2.3 Reaction Time
A limited number of runs were made at longer reaction times (see Table 12).
Some further reduction in the TCLP of the treated soil was observed on these runs,
but the level of reduction achievable is not as great as that observed by increasing
the temperature or the sulfur/soil stoichiometry.
6.3 OTHER TESTS
In this section several sets of experiments are described, which were performed to
evaluate additional process approaches or modifications for consideration.
36
-------
Table 12. Results
of Process Parameter Tests — Pb and Cd
Run
Temp
Time
S/soil
- Soil Mixture -
Pb
Pb TCLP
Cd
Cd TCLP
IB 027
!C
hrs
SSM
Top Soil
ppm
mq/L
ppm
mq/L
42
Untreateti
0
75
25
1000
0.79
44
250
2
0.05
75
25
1000
0.24
43
250
2
0.1
75
25
1000
0.22
52
250
2
0.15
75
25
1000
0.14
46
250
2
0.1
75
25
3000
0.34
47
250
2
0.1
75
25
5000
0.47
Untreated
0
Pedricktown Soil
512.0
58
250
2
0.10
Pedricktown Soil
77.1
56
250
2
0.25
Pedricktown Soil
82.4
67
250
2
1.00
Pedricktown Soil
30.3
57
350
2
0.25
Pedricktown Soil
78.5
53
Untreated
0
75
25
1000
13.8
55
250
2
0.05
75
25
1000
5.46
54
250
2
0.1
75
25
1000
4.37
63
Untreated
0
75
25
10000
61.7
5000
144
66
250
0.5
0.05
75
25
10000
7.53
5000
88.6
65
250
0.5
0.10
75
25
10000
5.95
5000
70.3
69
250
0.5
0.15
75
25
10000
5.46
5000
28.3
64
250
0.5
0.25
75
25
10000
3.39
5000
33.8
72
250
0.5
0.50
75
25
10000
3.36
5000
28.9
73
250
0.5
1.00
75
25
10000
4.88
5000
24.5
71
300
0.5
0.15
75
25
10000
1.22
5000
42.0
70
350
0.5
0.15
75
25
10000
1.21
5000
22.4
76
400
0.5
0.15
75
25
10000
2.21
5000
4.4
74
250
1.0
0.15
75
25
10000
2.84
5000
38.3
75
250
2.0
0.15
75
25
10000
1.24
5000
27.5
-------
8
6
4
2
0
0.5
Sufrur contend
0.4
0
0.1
FIGURE 7
LEAD TCLP VALUES AS A FUNCTION OF REACTOR SULFUR CONTENT
75/25 SSM Top Soil, 250°C, 0.5 hours
-------
150
100
0.5
SuMr contend3
0.4
0.1
0
FIGURE 8
CADMIUM TCLP VALUES AS A FUNCTION OF REACTOR SULFUR CONTENT
75/25 SSM Top Soil, 250°C, 0.5 hours
-------
6.3.1 Two Stage Runs
The reaction time process parameter discussed above for these small batch
experiments is not well defined due to heat up and cool down times contributing to
the total reaction time interval. Evidence for some further reduction in the TCLP of
the metals in the treated soils as a function of run time suggests additional reaction
time will enhance the TCLP reduction. Experiments were done to examine
successive batch treatment in which the product from the first run was used as
feed for the second run.
The runs were carried out at 250°C, 0.5 hour, and S/soil ratio of 0.15 (i.e., total
sulfur content 13.0 percent). Two duplicate runs were made at these conditions
to prepare enough material, after sampling for analysis of the first stage product,
to feed to the second stage. An additional 13 percent of fresh sulfur was added
for the second stage run. Thus, in comparing the results from the second stage
run, one needs to consider the added reaction time, the additional dilution of the
metal and the additional sulfur used.
The results are displayed in Table 13 along with comparable runs with longer run
times and higher sulfur loading. The contained metal content of the reactor feed
includes the dilution of the added sulfur. The percent leached by the TCLP test of
the raw reactor product (based on 100 g of soil in 2 liters of leachate) is also
displayed as well the metal concentration in the final leachate.
There are, however, no comparable runs with both added reaction time and higher
sulfur loading, but the results for both lead and cadmium seem consistent with the
earlier single stage tests (Table 12) if both reaction time and sulfur loading are
considered. Thus, where further processing is justified, added stages of
processing may accomplish it by extending the reaction time.
6.3.2 Soluble Sulfides
It is well known that addition of soluble sulfide in water-based solidification/
stabilization media are able to decrease TCLP leachate values for contained heavy
metals by precipitation or formation of a layer of insoluble metal sulfide. In the
Sulchem Process, excess elemental sulfur is heated up with the soil and presumed
to also react to form insoluble sulfides. This will occur however only if there is a
reducing agent present to reduce the elemental sulfur to sulfide ion. Thus, even
with the large excess of sulfur, there may still not be sufficient soluble sulfide
present to fully convert the metals to an insoluble sulfide form.
Accordingly, several screening tests were carried out on different heavy metal
spiked soils with sodium sulfide nonahydrate added to the reaction mixture. Table
14 summarizes the comparison runs at both 250°C and 440°C for the different
metal spiked soils with added sodium sulfide compared with runs (reported above)
at similar reaction conditions without the added soluble sulfide.
A striking reduction in the TCLP values with added sulfide are noted. Extending
the reaction temperature to 440 °C and adding a soluble sulfide to the process
reagents provides a passing TCLP of the treated soil for metals such as cadmium,
nickel, cobalt and zinc. The results with lead are mixed.
40
-------
Table 13.
Temp
Run °C
First Stage
81,82 250
Second Stage
83 250
Total reaction time
Comparison Runs
64 250
74 250
Time Soil Sulfur
hrs _£ g
0.5 200 30
0.5 178.2 51.8
1.0
0.5 200 50
1.0 200 30
Two Stage Run
s
Content
Lead
Reactor
feed
Pb
ppm
TCLP
Pb
mq/L
Leached
%
Cadmium
Reactor
feed
Cd
ppm
TCLP
Cd
mg/L
Leached
%
13.0%
8696
4.08 0.94%
4348
45.6 20.98%
22.5%
7748
1.56 0.40%
3874
25.4 13.11%
20.0% 8000 3.39 0.85%
13.0% 8696 2.84 0.65%
4000 33.8 16.90%
4348 38.3 17.62%
-------
Table 14.
Effect of Soluble Sulfide
Run Without
Run With
Run Without
Run With
Untreated Sol
Sulfide
Sulfide
Sulfide
Sulfide
75/25 SSM/soil +
Cd/Pb
Run No,
27-69
27-80
Temp "C
250
250
Time hr
0.5
0.5
Sulfur %
13.0%
12.8%
Sulfide %
0.0%
2.1%
Contained
Metal
TCLP
TCLP
TCLP
Metal
ma/ka
mg/L
mg/L
mg/L
Cd
5,000
144
28.3
0.56
Pb
10,000
61.7
4.08
0.19
SARM lil
Run No.
27-77
27-34
27-92
33-19
Temp °C
250
250
440
440
Time hr
0.5
0.5
0.5
0.5
Sulfur %
13.0%
12.8%
13.0%
12.8%
Sulfide %
0.0%
2.1%
0.0%
2.1%
Contained
Metal
TCLP
TCLP
TCLP
TCLP
TCLP
Metal
mg/kg
*ng/L
mg/L
mg/L
mg/L
mg/L
As
500
0.21
0.16
0.24
<0.05
0,33
Cd
1,000
36.8
22.5
7.9
3.66
0.52
Cr
1,500
<0.05
0.07
0,2
<0.05
<0.05
Pb
14,000
35.5
25,5
43,3
12.2
7.75
Ni
1,000
22.2
17.3
12
0,4
0.54
Cu
9,500
153
1.13
0.01
<0.01
0.08
Zn
22,500
791
628
482
32.4
9.93
75/25 SSM/soil + Ni/Zn
Run No.
27 -95
27-98
NA
33-20
Temp °C
250
250
440
Time hr
0.5
0,5
0.5
Sulfur %
13.0%
12.8%
12.8%
Sulfide %
0.0%
2.1%
2.1%
Contained
Metal
TCLP
TCLP
TCLP
TCLP
Metal
' mg/kg
mg/L
mg/L
mg/L
mg/L
Ni
2,000
0.54
1.95
0.72
<0.01
Zn
2,000
36,3
25.6
6.79
0.02
75/25 SSM/soil + Co
Run No.
23-42
23-43
Temp °C
440
440
Time hr
0.5
0,5
Sulfur %
13,0%
12,8%
Sulfide %
0.0%
2,1%
Contained
Metal
TCLP
TCLP
TCLP
Metal
ma/kg
mg/L
mg/L
mg/L
Co
10,000
15.0
11.3
<0.10
42
-------
6.3.3 Mineralogies! Speeiation
The original work plan expected to obtain mineralogical speeiation of selected
samples of processed treated soils in order to identify the chemical mineralogical
mechanism for heavy metal stabilization using the Sulchem Process. No specific
tests were done as part of this project, since a parallel effort on a separate project
had evaluated the product from similar batch tests in the unstirred vented mode
reactor of the SSM top soil blend to which salts of cadmium, mercury, and zinc
were added. Standard optical microscopy, X-ray diffraction, and scanning electron
microscope analyses were employed. These results are reported here to provide
mineralogical information on a typical product.
The main mineralogical form found in the raw reactor product from this test was a
siag-like material which was identified as a sulfur-rich cement. The heavy metals
were found as sulfides as well as a complex cementing matrix of elemental sulfur,
calcium oxides, and soil products. The predominant heavy metal phases present
were: mercury-cadmium-sulfides with traces of zinc; zinc sulfide with traces of
mercury, cadmium, and iron; and trace amounts of mercury and zinc associated
with sulfur bearing soil silicate minerals. The heavy metal sulfides exhibit a wide
range of particle sizes with the majority as fine particles in the 5 to 30 micron
range.
6.4 SUMMARY OF METALS TESTS
Preliminary results of the screening tests on different spiked soil mixtures provides
the opportunity to assess the response of the different heavy metals to
stabilization by the Sulchem Process. Each metal responds differently, but in
general there will be a maximum metal content which can be processed to achieve
passing TCLP leachate values. The precise maximum metal content which can be
processed to passing TCLP values will vary somewhat based on the reaction
temperature, reaction time, sulfur stoichiometry, or organic content of the soil. In
general, however, a batch screening test at 250°C, sulfur content of the order of
10 to 15 percent and reaction time of 1/2 hour will define an approximate upper
limit of the content of each metal which can be processed. Higher levels of metals
can be processed to give passing TCLP by increasing the temperature and/or
adding soluble sulfide, particularly for cadmium, nicked cobalt, and zinc.
Increasing the stoichiometry or reaction time provides only a marginal
improvement.
Recognizing that actual soil composition will affect the results, treatability studies
are required to more precisely define the metal concentration envelope, which can
be processed for a particular soil to give acceptable TCLP leachate values. Based
on the very limited tests to date, lead limits of approximately 10,000 ppm and
cadmium of several thousand ppm would seem to be generally feasible. Copper
responds very well and nickel and zinc appear to be processed as well although the
results are mixed. No definitive information on arsenic or chromium could be
developed from the SARM-III tests.
43
-------
7.0 ROTARY SOIL REACTOR TESTS
CHMR conducted six tests for organic compound destruction using the rotary soil
reactor described in Section 2.4. These tests were conducted using manufactured
gas plant (MGP) site soils described in Section 3.3. The objectives of the tests
were to demonstrate organic destruction at a larger scale using real Superfund soil
and the Sulchem configuration present in the rotary reactor, and to determine the
appropriate process conditions for optima! destruction. As designed, the rotary
reactor was intended to present two sections to destroy the organic compounds: a
soils reactor in which soil and sulfur are heated under inert atmospheres, and a
vapor reactor in which the unreacted organic vapors desorbed in the first reactor
are to react with molten sulfur.
Tests were not conducted using heavy metal contamination, because the MGP site
soils did not contain appreciable quantities of heavy metal contamination, and
because the stabilization of metals had been adequately demonstrated in previous
experiments on a smaller scale.
7.1 DESCRIPTION OF REACTIONS
Table 15 summarizes the reaction parameters and the results of the experiments.
Runs were made at 300° and 350°C at reaction times of 0.5 and 1.0 hours.
CHMR initially planned to conduct runs at temperatures below 300°C, but the
results of initial runs indicated that the DRE would be too low at such
temperatures. A one hour residence time at reaction temperature was used
initially because it was thought to provide sufficient time for the reaction to proceed
completely. A half hour residence time was used for two of the six runs to
determine whether a half hour was sufficient for the reaction to proceed.
The sulfur/soil ratio was generally 10%, but runs were also conducted at 6 and
20%. Four of the runs were conducted with the soil as it was obtained from the
site. Before conducting two of the runs, the soil was dried to reduce the moisture
content and thereby reducing the vapor flow rate.
7.2 RESULTS
Analyses were conducted using EPA Methods 3550, 3660, and 8100 which involve
extraction with methylene chloride followed by analysis using gas chromatographic
methods. The GC/FID results were quantitated by calibrating for four major
constituents (2-methyl naphthalene, acenaphthene, phenanthrene, and pyrene)
which covered the boiling point range for the contaminants in the soil. In addition
to quantifying the recovery of these compounds, semi-quantitative recoveries for
other constituents in the GC (which had been identified by GC/MS) could also be
determined from the ratio of GC/FID peak areas for both the starting soil and the
product fractions based on their relative quantities and dilutions as were done for
the four compounds that were quantitated. These other constituents showed
similar behavior of recovery as a function of boiling point.
7.2.1 Overall GC Results
Figure 9 shows three typical chromatograms - one from the untreated soil, the
second from the condensate trap, and the third from the treated soil. The figure
shows qualitatively how the products are separated between the reactor and the
44
-------
TABLE 15. SUMMARY OF ROTARY REACTOR TESTS
RUN CONDITIONS
Run Number
Feed Soil
46-6
46-10
46-14
46-18
46-22
46-26
Temperature, °C
350
350
300
350
300
350
Time, hrs
1
1
1
1
0.5
0.5
CHARGE
Weight soil charged, g
3688
2000
2000
1607
2000
2000
Moisture content
20.1%
20.1%
20.1%
20.1%
0.0%
13.5%
20.1%
Soil charged, mf, g
2946.7
1598.0
1598.0
1607.0
1730.6
1598.0
Sulfur charged, g
220
400
200
160
200
200
Percent Sulfur, mf, %
6.9%
20.0%
11.1%
9.1%
10.4%
11.1%
PRODUCTS
Weight treated soil, g
2742
1366
2079.7
1670
1632.2
1632.0
First trap, g
581.4
1429.6
373.0
27.7
282.4
349.1
Second trap, g
69.8
2.6
3.3
1.9
2.4
5.4
Caustic trap, g
3309.5
4662.2
4546.8
4177.4
4537.4
4268.6
Sulfur content
15.2%
NA
NA
3.6%
2.8%
3.5%
calculated H2S, g
532.8
—
—
159.8
135.0
158.7
weight increase, g
NA
172.9
70.8
86.7
68.4
151.4
ORGANICS (mg/kg mf soil)
BP °C
2-methyl naphthalene
241
59
55.96
25.26
72.12
20.69
118.38
109.87
acenaphthene
278
258
20.61
9.84
22.50
12.76
<0.05
17.07
phenanthrene
340
380
51.48
13.88
63.94
58.08
62.97
. 173.94
pyrene
393
634
3.98
0.08
0.58
1.99
<0.05
7.08
chrysene
448
323 *
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
benzopyrene/benzofluoranthene
>450
680 *
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
NA = Not available or not determined
* = GC/MS analysis
-------
Approximate
Amount
Present
Approximate
Amount
Present .
Approximate
.Amount
Present
j. 4J
|lw
Original MGP Sol
SO
¦4. O
GO
to
"UJl
Condensata, Run 46-10
HT Grease
o
•*4- O
© o
Treated SghI, Run 46-10
T Grease
o
SO 40
Elution Time
GO
Figure 9. Comparison of MGP Soil and Product Fractions
46
-------
condensate. The scales of each have been adjusted so that the peak heights are
approximately proportional to the amount of material present. (The scales are not
identical since each extraction involved different sample size and dilution factor.)
In the initial soil chromatogram, the retention time for the compounds are shown to
range between 16 and 55 minutes, with the bulk of the compounds lying between
19 and 48 minutes. In the treated soil chromatogram, the relative proportion of
most of the compounds is significantly diminished. Also, virtually all of the higher
boiling compounds (with elution times greater than 35 minutes) have been
eliminated. (The product fractions show contamination from the high temperature
grease used in the rotating unions giving rise to a series of n-paraffin peaks (C22-
C30), an antioxidant (methylene-bis ethyl, t-butyl phenol), and sitosterol which were
identified by GC/MS. These are not considered in the analysis.) In the
chromatogram of the condensate trap, no higher boiling point compounds are
found above pyrene. However, significant quantities of lower boiling point
compounds, which evidently desorbed from the soil, were present in the trap.
7.2.2 Quantitative Results
Table 16 shows the run conditions and quantitative results from the experimental
runs including individual product fractions. Several of the caustic traps were also
extracted for organics, but very low quantities were found. Therefore the summary
table only lists the reactor solids, condensate trap (trap #1), and the ice trap (trap
#2). The percent recovered from the overhead, percent destroyed, and DRE for
six compounds are given (refer to Section 5.1.1 for definition of terms). The
results for the two highest boiling compounds are based on initial concentrations
measured by GC/MS analysis.
The lowest boiling compounds (methyl naphthalenes) showed very little destruction
although the ratio of 1-methyl naphthalene to 2-methyl naphthalene decreased by
a factor of two to three. In addition, naphthalene, which was not found in the
original soil, was produced in the process, presumably by partial reaction of higher
homologs. Differences in reaction rates are also observed for the intermediate
boiling aromatics (dimethyl naphthalenes, acenaphthene, fluorene, phenanthrene,
in the boiling point range between 260 and 340°C.) These compounds show
destructions ranging from about 50 to over 90% destroyed whereas some of the
saturated hydrocarbons in the same boiling range (pristane and phytane) are
generally present in the products at about 50% of the feed content.
Higher boiling aromatic hydrocarbons (pyrene, chrysene, benzopyrene, etc. with
BP >340°C) are nearly completely destroyed with only very low levels, or non-
detect levels, observed in any of the product fractions. This indicates that the
process works well for the high boiling point compounds, even at temperatures
below their boiling points.
Thus in Table 16 the recoveries of the four compounds that were quantified by the
analytical method are representative of the yields observed semi-quantitatively for
the other hydrocarbons with similar boiling ranges in the test soil. The effect of
boiling range on the fate of the hydrocarbon contaminant in the soil, whether
desorbed into the overhead, chemically destroyed, or left as trace residuals on the
treated soil are similar to the initial screening studies in Section 5 on the effect of
boiling range on the fate of contaminants in the process.
47
-------
TABLE 16. ROTARY REACTOR TESTS - RECOVERIES
Run 4S-6: S50°C, 1 hr, 6.9% S
2-methyl naphthalene
acenaphthalene
phona.ithrer.e
pyrene
chrysene
bsnzopyrene/benzafluoranthene
2-methyt naphthalene
acenaphthalene
phenanthrene
pyrene
chrysene
bsnzopyene/fcenzofluoranthene
2-methyl naphthalene
acenaphthalene
phenanttirene
pyrene
chrysene
benzopyrene/benzofluoranthene
2-methyl naphthalene
acenaphthalene ,
phenanthrene
pyrene
chrysene
benzopyrene/benzofluoranthene
2-methyt naphthalene *
acenaphthalene
phenanthrene
pyrene
chrysens
benzopyrene/benzofluoranthene
2-methyl naphthalene
acenaphthalene
phenanthrene
pyrens
chrysene
benzopyrene/benzofluoranthene
mg/kg mf
mg recovered/kg mf feed
Overhead
Destruction
DRE
BP *C
feed
reactor
trap 1
trap 2
Recovery
%
%
241
58.9
0.09
- 49.03
6.84
94.8%
5.0%
99.8%
278
257.9
0.00
19.87
0.74
8.0%
92.0%
>99.9
340
380.0
0.09
46.09
5.30
13.5%
86.5%
>99.9
393
633.6
0.08
3.90
0.00
0.6%
99.4%
>99.9
448
323
0.00
0.00
0.00
0.0%
>99.9
>99.9
>450
680
0.00
0.00
0.00
0.0%
>99.9
>99.9
Run 46-10: 350°C, 1 hr, 20% S
mg/kg mf
mg recovered/kg mf feed
Overhead
Destruction
DRE
BP°C
feed
reactor
trap 1
trap 2
Recovery
%
%
241
58.9
0.00
24.96
0.30
42.9%
57.1%
>99.9
278
257.9
0.00
9.79
0.05
3.8%
96.2%
>99.9
340
380.0
0.00
13.54
0.34
3.7%
96.3%
>99.9
393
633.6
0.08
0.00
O.QO
0.0%
>99.9
>99.9
448
323
0.00
0,00
0.00
0.0%
>99.9
>99.9
>450
680
0.00
0.00
0.00
0.0%
>99.9
>99.9
Run 46-14: 300*0,1 hr, 11% S
mg/kg mf
mg recovered/kg mf feed
Overhead
Destruction
DRE
BP *C
feed
reactor
trap 1
trap 2
Recovery
%
%
241
58.9
0.00
71.08
1.04
122.4%
-22.4%
>99.9
278
257.9
0.44
21.91
0,15
8.6%
91.3%
99.8%
340
380.0
1.07
60.50
2.37
16,5%
83.2%
99.7%
393
633.6
0.58
0.00
O.QO
0.0%
>99.9
>99.9
448
323
0.00
0,00
0.00
0.0%
>99.9
>99.9
>450
680
0.00
0.00
0.00
0.0%
>99.9
>99.9
Run 46-18: 350"C, 1 hr, 3.1% S
mg/kg mf
mg recovered/kg mf feed
Overhead
Destruction
DRE
BP *C
feed
reactor
trap 1
trap 2
Recovery
%
%
241
58.9
1.90
18.00
0.79
31.9%
64.9%
96.8%
278
257.9
1.56
11.20
0,00
4.3%
95.1%
99.4%
340
380.0
5.74
50.22
2.12
13.8%
84.7%
98.5%
393
633.6
1.99
0.00
0.00
0.0%
99.7%
99.7%
448
323
0.00
0.00
0.00
0.0%
>99.9
>99,9
>450
680
0.00
0.00
0.00
0.0%
>99.9
>99.9
Run 46-22:
300*C, 0.5 hr, 10.4% S
mg/kg mf
mg recovered/kg mf feed
Overhead
Destruction
DRE
BP *C
feed
reactor
trap 1
trap 2
Recovery
%
%
241
58.9
0.88
115.57
1.93
199.4%
-100.9%
98.5%
278
257.9
0.00
0.00
0,00
0.0%
>99,9
>99.9
340
380.0
0.89
59.09
2.99
16.3%
83.4%
99.8%
393
633.6
0.00
0.00
0.00
0.0%
>99.9
>99.9
448
323
0.00
0.00
0.00
0.0%
>99.9
>99.9
>450
680
0.00
0.00
0.00
0.0%
>99.9
>99,9
Run 46-26:
350°C, 0.5 hr, 11.1% S
mg/kg mf
mg raeovered/kg mffeed
Overhead
Destruction
DRE
BP *C
feed
reactor
trap 1
trap 2
Recovery
%
%
241
58.9
0.69
106.71
2.47
185.3%
-86.4%
98.8%
278
257.9
0.38
16.69
0.00
6.5%
93.4%
99,9%
340
380.0
0.96
165.34
7.64
45.5%
54.2%
99.7%
393
633.S
0.55
6.53
0.00
1.0%
98.9%
>99.9
448
323
0,00
0.00
0.00
0.0%
>99.9
>99.9
>450
680
0.00
0.00
0.00
0.0%
>99.9
>99.9
48
-------
The weight gain in the caustic trap is presumed to be mainly hydrogen sulfide.
Sulfur analyses of the caustic trap permit an independent calculation of the weight
of hydrogen sulfide and higher quantities are inferred from the sulfur analyses.
These differences may be due to other sources of sulfur products in the caustic
trap such as polysulfides and carbon disulfide.
7.3 DISCUSSION OF SOIL REACTOR RESULTS
The soil reactor showed nearly complete conversion for chrysene and benzopyrene,
two compounds with boiling points above 400 °C, and 99% conversion for pyrene.
For lower boiling compounds, as anticipated, the soil reactor showed the effects of
competition between reactions with sulfur, and desorption for the organic
compounds. This section discusses the effects of varying the sulfur/soil ratio, the
temperature and the residence time in the reactor.
7.2.1 Sulfur/Soil Ratio
The initial experiment was conducted using only 6% sulfur, which was below the
estimated stoichiometric requirement of sulfur of 10%. During this experiment,
virtually all of the methyl naphthalene was recovered in the overhead products.
When 20% sulfur as added, the percentage of methyl naphthalene recovered in the
overhead product decreased to 35%, indicating significant reaction with the excess
sulfur. When 10% sulfur was added, with the temperature maintained at 350°C,
the results were mixed: one run showed 65% methyl naphthalene conversion,
while a second showed negative conversions.
7.3.2 Residence Time
Four of the six runs were conducted at one hour soil residence time. Two runs
were conducted at 300°C, both using 10% sulfur in soil, with one hour and one-
half hour residence times. Both runs showed overall DRE's above 99.5%
The half-hour run showed a greater amount of material in the traps, particularly
acenaphthene and phenanthrene. This result appeared to be anomalous, except
that it was duplicated when dual one hour and half hour runs were conducted at
350°C. One explanation was that most of the reaction occurred for the two
compounds in the vapor phase, and the two compounds desorbed relatively
slowly. It is possible that a significant amount of acenaphthene and phenanthrene
desorbed after the reactor started to cool down, and therefore passed through the
sulfur without .reacting.
7.3.3 Temperature
Runs were conducted at 300 and 350°C. The plans initially included runs at lower
temperatures, to try to determine the lower limit of effectiveness for the process,
because the evidence from gas flow and scrubber water discoloration indicated
significant production of H2S {and therefore, the initiation of the Sulchem reaction)
beginning about 190°C. However, analysis of the results at 300°C indicated that
little could be gained from decreasing the temperature below that point. Elevating
the reaction temperature above 350°C was not considered practical, since it would
likely require reactor skin temperatures well in excess of sulfur's boiling point of
440°C.
49
-------
7.4 VAPOR REACTOR PERFORMANCE
Most of the results obtained related to the performance of the overall reactor with
respect to the original starting soil, which is appropriate since that is the purpose
of the reactor system. However, the reactor system was essentially two joined
reactors: a soil reactor and a vapor reactor. The soil reactor was intended to react
contaminants in the soil with sulfur, and desorb the remaining organic
contaminants. The vapor reactor was intended to destroy the desorbed organics
by forcing them through a heated, baffled cylinder containing molten sulfur.
The vapor reactor was sized to provide a nominal residence time of between three
and five minutes for the vapors emanating from the soil reactor. This nominal
residence time was calculated based on estimated nitrogen purge flow rates and
hydrogen sulfide production. Vapor flow through the vapor reactor consisted
almost entirely of nitrogen while the reactor was cool. Then water vapor would
flow through the reactor as the temperature of the soil rose to 100°C. The water
vapor flow rate was anticipated to be relatively high, as the soil contained up to
20% moisture. The water was not anticipated to carry over many organics with it,
since the organics volatilized at much higher temperatures. After the water vapor
was forced off the soil, the flow rate would tend to decrease again. It would
increase once more at or near the reactor temperatures, as some organics
volatilized and others began to react to form hydrogen sulfide. Nominally, this
would have produced conversions in the vapor reactor in the range of 40 to 75%
(based on toluene vapor rate data in Section 5), at temperatures ranging between
300 and 350°C.
Unfortunately, in order to maintain sufficient pressure throughout the system and
in the scrubber, the nitrogen purge gas flow rate had to be maintained at nearly 2
L/min (based on ambient temperatures). This reduced the residence time in the
vapor reactor to approximately 2.5 minutes. Further reductions in residence time
occurred because the MGP soil used in the experiment had a relatively high level of
organics present, and therefore produced more hydrogen sulfide than originally
anticipated, and because most of the hydrogen sulfide tended to be produced over
an approximately 15 minute time span, rather than the half hour or hour residence
time. Estimated residence times for the organic vapors in the vapor reactor tended
to range between 40 and 100 seconds, therefore, with theoretical conversions of
between only 11 and 35%.
Based on the results for acenaphthene and 2-methyl-naphthalene, both of which
volatilize below 300°C and therefore would not be expected to show much
conversion while within the soil, it appears that a significant fraction of the organic
compounds were destroyed in the vapor reactor. The destruction rates were
higher for reactions run at higher temperatures. This may be due to the rate of
heating allowed by the reactor temperature control system, which tended to ramp
up more quickly when the temperature was set higher, and therefore would be
anticipated to raise the temperature of the reactor and the sulfur more while the
compound were volatilizing.
A limitation on the vapor reactor was that it was heated at the same rate as the
soil reactor. Therefore, if compounds volatilized without reacting significantly in
the soil reactor because it was not sufficiently hot, then the compounds would be
anticipated to flow through the vapor reactor without reacting as well. Thus, for
example, any compounds which were steam stripped from the soil while the water
was boiling off could conceivable pass through the vapor reactor while the sulfur,
50
-------
which melts at 120°C, was still solid and unreactive. During the final run (Run No.
46-26), CHMR attempted to compensate for this by heating the vapor reactor side
of the rotary reactor more quickly than the soil side. The results from the run were
disappointing, however, and showed no affect of this operational change.
In order to compensate for the deficiencies in the design and increase the
conversion in the vapor reactor, CHMR recommends two changes: a larger reactor
volume to increase the vapor residence time; and a decoupled soil/vapor reactor
system which will enable the vapor reactor to be maintained at reaction
temperatures even as the soil is being heated, so that desorbed organics can react.
However, based on the rotary reactor results and the previous vapor reactor study
described in Section 5, CHMR believes that the vapor reactor should not be
envisioned as the sole means of treating or destroying organics with boiling points
below about 200°C. A secondary activated carbon or condensation system is
likely to be necessary to destroy or remove the organic compounds from the vapor
stream.
7.5 CONCLUSIONS FROM ROTARY REACTOR RUNS
Based on the results of the rotary reactor runs, CHMR concluded the following
regarding the feasibility of the Sulchem Process for the destruction of organic
compounds.
• The rotary reactor completely or nearly completely destroyed
compounds with boiling points above 350°C when operated at 300 or
350°C.
• The rotary reactor produced efficient destruction when operated with
roughly the stoichiometric quantity of sulfur.
• The rotary reactor showed efficient desorption of compounds with
boiling points below 300°C. A portion of these compounds were
destroyed in the attached vapor reactor.
• The vapor reactor as configured could destroy a significant percentage
of the desorbed organics. To increase destruction efficiency, CHMR
recommends a larger vapor reactor (higher residence time) which is
coupled with the soil reactor to better maintain it at reaction
temperature.
• The soil reactor, coupled with the vapor reactor and a trap to
condense water vapor and semi-volatile organics (boiling point range
100-200°C), is a technically effective means of treating soils
contaminated with organic compounds. A more sophisticated
trap/condensation system, coupled perhaps with activated carbon
adsorption may be necessary to treat soils contaminated with volatile
compounds.
51
-------
8.0 DETAILED PROCESS DESCRIPTION
The Sulchem Process is designed to provide destruction of hazardous organics
while simultaneously stabilizing metals in contaminated soils. The Sulchem
Process uses elemental sulfur, which reacts with the carbon in organic materials at
moderately elevated temperatures to form an insoluble, inert carbon-sulfur
amorphous solid. The contained heavy metals are immobilized through formation
of insoluble metal sulfides.
The Sulchem Process's main process components include:
• A pre-reaction mixer where the soil and sulfur are mixed;
• A reactor in which the soil/sulfur mixture is heated sufficiently to react
the organics with sulfur;
• A vapor phase reactor or reactor zone in which desorbed organics
from the first reactor are further reacted with elemental sulfur;
• The off-gas handling system, which collects and treats condensable
by-products and scrubs acid gases from the effluent vapors;
• A water treatment system which removes accumulated organic
chemicals from the wastewater; and,
• A post-reaction treatment unit that recovers excess reagent and
prepares the treated product to comply with on-site disposal
requirements.
The system in its most basic form is shown schematically in Figure 1. The focus
of the research was on the development of the system in order to better detail the
process components and configurations.
This section provides a discussion of two key process subsystems which were the
focus of the research: the soils reactor and the vapor reactor systems. Then the
overall process configuration is discussed.
8.1 SOIL PROCESSING
Soil pre-processing will include screening to remove oversized particles {probably
those greater than 1-inch diameter), dewatering in applications in which it is
necessary, and mixing of the soil with sulfur. Screening is anticipated to occur in a
trommel screen, or other standard soil screening system. Dewatering may occur in
a filter press, if required. The soil is anticipated to be mixed with sulfur in a soil
scrubber, or other grinding type mixer. The soil will pass through the scrubber
(which will also break up clumps) and into the reactor. For some applications, the
soil may be scrubbed and mixed with sulfur before it is sieved. This will help break
up clods, if it is required.
8.2 SOIL REACTOR
CHMR considered three alternative soil reactor designs:
• Multiple screw conveyor system;
52
-------
• Rotary soil reactor; and,
* Stationary furnace type reactor.
Each of these systems would be assumed to operate at approximately the same
temperature, atmospheric pressure, and with downstream vapor processing of off-
gases and desorbed organics. The advantages and disadvantages of each are
discussed below.
8.2.1 Multiple Screw Conveyor System
Under this system, a multiple Holoflite screw system or porcupine processor would
be used to mix and agitate the soil in a chamber, while being heated internally with
hot oil and/or molten salts. A vapor space would be provided above the reactor,
and off-gases would leave the reactor through ports located above the soil.
The advantages to this system are that it provides good mixing of the soil and
sulfur, and potentially good heat transfer through the Holoflite screws. The
disadvantages include difficulties sealing the system, the potential for mechanical
problems with the inter-meshing screw systems, difficulties heating a system using
molten salts (both in terms of seal problems into the Holoflite system as well as in
heating and maintaining the salts in the first place), and difficulties in maintaining
seals at the soil inlet and outlet.
8.2.2 Rotary Reactor
Under this system, an externally fired rotary reactor would be used as a soil
reactor. The reactor would essentially be a long cylinder on a slight angle, which
rotates at 1 to 5 rpm, which is heated either by natural gas, or by electric inductive
coils to the appropriate reaction temperature. Soils would enter through one end
of the reactor, and slowly pass down through the reactor and out the far end. The
reactor would be sized to provide residence times on the order of 1 hour. The soil
would enter the system through an interlocking dual door mechanism, in which a
holding chamber is filled with soil, then sealed, and then a door opens to allow that
soil to pass into the rotary reactor. {Such systems are used in rotary kiln
incinerators to maintain a sealed system as solids are entered or removed.)
Likewise, a similar interlocking system would be employed to remove the solids
from the lower end of the rotary reactor. Vapors would leave the system through
a rotating joint at the lower end of the reactor.
8.2.3 Stationary Furnace Reactor
This reactor would be similar in design to the furnace type reactors used in coal
coking operations. A soil/sulfur charge would be added to one of a series of
stationary furnaces, the off-gases collected and treated, and the furnaces would be
heated by indirect gas firing between each furnace. Soil would then be pushed out
of the furnace and quenched, as it is in coking operations. This design presented
several challenges, including collection of gases when the soil leaves the furnace
and lack of mixing in the furnace. It may be appropriate for high residence time
operations (i.e., eight or more hours), but was not deemed appropriate for the
Sulchem process.
53
-------
8.2.4 Selected Reactor arid Considerations
After evaluating the advantages and disadvantages of the three alternative basic
reactor designs, CHMR selected the rotary reactor as the most promising, because
it appeared to be among the simpler mechanically, while offering great flexibility
during actual operations.
The major parameters which needed to be considered for the reactor were then;
• Reactor size, determined primarily by; individual reactor feed
capacity, residence time, transportability requirements, and reactor
freeboard (vapor space above soil);
• Method of heating, which was determined by; required reaction
temperature, soil capacity and moisture level, and individual charge
feed rates (if feed was semi-continuous);
• Equipment used for conveying soil into and out of reactor. Two basic
alternatives existed: a screw type conveyance system and a door
mechanism. Considerations included: seal requirements, nitrogen
blanketing, soil type, and whether the feed could be semi-continuous.
CHMR based the overall system economics on two base cases: a 10-ton per hour
unit and a 20-ton per hour unit. The other design parameters and assumptions
made regarding the reactor are summarized in Table 17. The overall reactor size
requirement was determined assuming a maximum one-hour residence time after
the soil was heated to reaction temperature, with a nominal 150 second gas
stream residence time.
Table 17. Rotary Reactor Assumptions and Parameters
Parameter
10-ton/hour Reactor
20-ton/hour Reactor
Residence Time
1 hour after soil heat up
1 hour after soil heat up
Vapor Residence Time
150 seconds
150 seconds
Transportability
Truck mounted system
Truck mounted system
Diameter Limitation:
7 feet
7 feet
Number of modular
units:
3
5
Inlet Soil Conveyance
Continuous; screw type
from hopper to reactor
center-line
Continuous; screw type
from hopper to reactor
center-line
Outlet Soil Conveyance
Semi-continuous; sealed
door mechanism
Semi-continuous; sealed
door mechanism
Maximum Reaction
Temperature
400°C
400°C
Method of Heating
Propane/natural gas
Propane/natural gas
54
-------
8.3 SULFUR/VAPOR REACTOR SYSTEM
An originally envisioned, the Sulchem system would include first a soil reactor to
destroy the organics and stabilize the metals in soil, then a sulfur/vapor reactor to
destroy desorbed organics in the off-gas stream. However, four basic options exist
for the vapor reactor:
• No vapor reactor at all -- condense the desorbed organic chemicals
and treat them with the wastewater stream.
• Sulfur/vapor reactor incorporated into soil reactor -- allow excess
freeboard space in soil reactor to try to react some of the desorbed
organics with sulfur, then condense/treat the remainder. The
desorbed organics would be assumed to react with the elemental
sulfur present in the vapor space from volatilization from the soil.
• Separate liquid sulfur/vapor reactor -- Use a sulfur/vapor reactor
system similar to that described and tested in Section 1, which reacts
the desorbed organics with liquid sulfur in a slowly rotating,
approximately plug-flow reactor. Then condense any unreacted
organics with the water stream and treat.
• Separate sulfur vapor/organic vapor reactor - Based on previous
Sulchem testing, construct a system which reacts the desorbed
organics with sulfur vapor. It is in theory possible to construct a
system capable of reacting nearly 100% of the desorbed organics,
thereby leaving only a trace of organics in the wastewater stream.
However, high vapor residence times, high temperature and extensive
sulfur vapor handling systems would be required.
Based on the results of the experimental work, in which the vapor reactor was
shown to require relatively high residence times for very efficient destruction (on
the order of 400 to 1,000 seconds residence time, depending on the temperature
for 99% toluene destruction), it was decided that it would not be practical to use
any of the first three options to target destruction of all of the desorbed organics.
Therefore, it is assumed that condensation of the organics with the water or from
the gas stream, followed by further treatment and/or disposal, will be required.
The first three options, may qualitatively be anticipated to yield different
destruction efficiencies: the first will yield negligible destruction in the vapor
phase; the second may be designed for 30 to 60% destruction of the desorbed
organic vapors, and the third for 60 to 90% destruction. The actual efficiencies
for the process depend on the vapor residence time, which is a function both of
the size of the space available for contact between the volatilized organics and
sulfur, and the overall flow rate of the vapor stream. This latter parameter, as was
seen in Section 5.2.3 {c.f., Table 7), is mainly determined by the moisture content
of the soil. Because all three of the liquid sulfur/vapor reaction options are
anticipated to require some downstream removal and processing of organics, it
was decided that the second option, which allowed for some reduction in organic
flow without requiring an entire separate reactor system to accomplish it, would be
the most reasonable alternative of the three.
The fourth alternative, a sulfur vapor/organic vapor reactor has the potential of
destroying virtually all of the organic compounds, provided that sufficient
55
-------
contacting time was maintained. CHMR does not have data concerning the reactor
rates at temperatures above 500°C, but based on extrapolation of data from lower
temperatures, residence times on the order of several hundred seconds would be
required for complete reaction. If this is the case, then the complexity of a sulfur
vapor reactor, and the required sulfur vapor handling system, could not be Justified.
Therefore, CHMR based the overall system design on the use of a sulfur/organic
vapor reactor incorporated into the design of the soil reactor. For this, CHMR
assumed that it would ensure a minimum vapor residence time of 150 seconds,
(which would yield a 50% conversion at 350°C, based on the toluene data), based
on the maximum moisture content assumed for the soil.
8.4 OFF-GAS HANDLING SYSTEM
The off-gas handling system is assumed to consist of two key units: a
cooling/condenser system to cool the gas and condense the water vapor and
desorbed organics; and a scrubbing system to remove the hydrogen sulfide. The
hydrogen sulfide removed in the scrubber may be converted back to sulfur for
reuse in the process using LO-CAT IP or other sulfur process. The off-gases may
be filtered prior to the condenser using a cloth filter to remove large particles, if
this is necessary based on the soil behavior in the reactor.
If cooling water is available on the site, CHMR anticipates using a water-cooled
condenser system to cool the off-gases and condense the water vapor. If such
water is unavailable on the site, CHMR may use an air-cooled system instead.
Both systems will have to be designed to allow for the condensed liquids to be
collected, and to minimize the possibility for sulfur vapors to condense and solidify
on small heat exchanger tubes.
The off-gas would pass through the first condenser, in which its temperature
would be reduced from 300° to approximately 80°C, to a second condenser,
which will operate under refrigeration and will cool the off-gases to the range of
10°C, effectively condensing nearly all the water vapor and any desorbed
organics.
The remaining gas, consisting primarily of nitrogen purge gas and hydrogen sulfide,
but with low concentrations of water vapor and a trace concentration of carbon
disulfide, will be fed through a scrubber associated with the patented LO-CAT II™
process. This process can be licensed from ARI Technologies of Palatine, Illinois,
LO-CAT IT removes hydrogen sulfide from gas streams using a chelated iron
system, then oxides the sulfide to elemental sulfur, which is separated. The
advantages to using a LO-CAT II™ unit is that it will enable the sulfur in the
hydrogen sulfide to be recycled and reused in the process, reducing sulfur costs
and eliminating a stream requiring disposal.
8.5 SOIL POST-PROCESSING
Based on the results from the current work, minimal post-treatment of the resulting
soil will be necessary. Virtually all the organic compounds originally present in the
soil are anticipated to be removed or destroyed by the process. The hot soil will be
quenched using water condensed from the gas stream. The only additional
components present in the soil are anticipated to be a small amount of an inert
solid, CSO50, and unreacted elemental sulfur. Elemental sulfur is already a natural
56
-------
component of soils, and therefore is not anticipated to present an environmental
concern.
8.6 WATER TREATMENT
As envisioned, the process will generate no wastewater streams; all the water
originally driven off the soil will be used to cool and re-moisten the soil in the
quench system. The condensed water from the primary reactor is anticipated to
contain some residual organic compounds. These compounds will need to be
removed and/or destroyed before the water is used in the quench system.
Therefore, a water treatment system is required. There is a tradeoff between the
requirements for the water treatment system and the sulfur/vapor reactor system:
the more efficient the sulfur/vapor system, the fewer organic will be present in the
wastewater for treatment, and vice versa. In order to evaluate this tradeoff for the
purposes of process economic calculations, CHMR chose a relatively simple,
straightforward water treatment system which relied on granular activated carbon
(GAC) to remove most of the organic compounds. CHMR did not evaluate other
potential alternatives, such as UV/oxidation technology or air flotation
technologies, which may ultimately be less expensive than GAC.
The wastewater treatment system is envisioned to include two major process
units: a filter to remove entrained particles from the water, and a GAC column to
remove any remaining organics from the water stream. The filter solids are
anticipated to include primarily unreacted sulfur, CS0 B6, some small soil particles
and perhaps trace organics. Most likely, the solids could be re-mixed back into the
soil and re-processed. The activated carbon will require periodic regeneration. The
treated water will be used to quench the hot soil exiting the soil reactor. No
wastewater disposal or discharge to a stream or POTW is anticipated from the
process.
57
-------
9.0 PROCESS ECONOMICS
Process economics were estimated based on the system designed above using
several alternative scenarios. Major site and process parameters which were varied
include:
• Site size;
• Treatment system overall capacity;
• Individual reactor capacity (number of parallel reactors); and,
• Soil moisture content.
Table 18 summarizes the economic calculation as a function of these variables. In
each case, the site was assumed to contain soil with 2.5% total organic content
and heavy metals concentrations below 1000 mg/kg. The process was operated
at 350°C. The distribution of organic compounds was assumed to allow for 90%
destruction of the organic compounds while they were within the soil. The vapor
space residence time was allowed to vary between 90 and 140 seconds to size the
reactors so they would fit on a flatbed trailer. All components which came in
direct contact with the process were assumed to be constructed of 316 stainless
steel, at a total cost of $10.00 per lb fabricated. Sulfur was assumed to cost
$0.05 per lb ($0.11/kg). The LO-CAT II™ System was assumed to be leased to the
process, at a cost of $350/ton ($385/1000 kg) of sulfur recovered. The process
was assumed to be operated 80 hours per week. Finally, the granular activated
carbon (GAC) used in wastewater treatment was assumed to cost $2.00 per lb
($4.40/kg), including the cost of regeneration.
Capital costs were taken as the sum of the costs of the following major units:
• Site preparation;
« Soil pre-treatment;
• Reactor, including the reactor units, plus the associated heater, rotary
mechanism, insulation, and inlet and outlet system;
• Heat exchangers/condensers;
• Soil post-processing;
• Water treatment (assumed to use activated carbon).
The capital costs associated with the LO-CAT™ system were assumed to be
included in the lease costs for the system, which were based on the amount of
sulfur recovered.
Operating costs included the following major cost items:
• Personnel costs, including 3 personnel associated with excavation, 2
plant operators and 1 supervisor or safety officer at the plant during
operating hours (80 hours per week), and 2 laboratory technicians
each working 40 hours per week;
• Equipment leasing costs associated with the LO-CAT™ system and
excavation equipment;
• Consumables including sulfur (10% by weight in the soil, but 50% of
which was assumed to be recovered by LO-CAT II™), activated carbon
in the wastewater treatment plant, and analytical supplies;
58
-------
TABLE 18. Summary of Process Economics
Parameter
Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
Site Size
(tons)
20,000
20,000
10,000
40,000
40,000
40,000
System
Capacity
(tons/hr)
10
10
10
20
20
20
Moisture
Content (%)
14
7
14
14
14
7
Number of
Reactors
3
3 -
3
6
4
3
Amount of
GAC req'd
(lb/ton soil)
13.4
12.8
14.6
14.0
14.7
14.0
Total Capital
Cost
$2.6
million
$2.5
million
$2.5
million
$4.4
million
$3.2
million
$2.6
million
Operating
Cost/ton
$104
$90
$113
$85
$85
$83
Capital
Cost/ton
$43
$43
$68
$37
$27
$22
Total
Treatment
Cost/ton
$147
$133
$181
$122
$112
$105
• Utilities, including gas or propane to heat the system at $6/milIion btu,
water (for cooling), and electricity;
• Maintenance, at 5% of the capital cost per year; and,
• Start-up costs, at $50,000 per site.
Other costs which were factored into capital costs included the cost of
transportation and assembly of the equipment.
As can be seen in the table, total costs vary between $105 and 181 per ton of soil
($115 and $199/1000 kg), depending on the site conditions and process
parameters. A discussion of some of the key process cost factors is given below.
Site Size
As anticipated, there are returns to scale with the process, in both the operating
and capital cost domains that yield a decrease in the cost per ton with increasing
site size. In addition, the cost is strongly a function of contaminant concentration.
Case 1 costs decrease to $ 130/ton ($143/1000 kg) when the contaminant
concentration is reduced to 1 % rather than 2.5%. Most of this cost is associated
with a reduction of the amount of GAC needed from 13.4 lbs/ton (6.7 kg/1000 kg
of soil) to approximately 5 lbs/ton (2.5 kg/1000 kg).
59
-------
System Capacity
Increasing the system capacity (tons per hour) tends to increase the capital cost
(since bigger equipment is needed) while lowering the operating costs (since the
same number of operators can handle more material). For sites of 20,000 cubic
yards or less, a plant capacity of 10 tons per hour seemed to be the most
reasonable. For sites larger than 20,000 tons, a system capable of treating 20
tons per hour {18,000 kg/hr) yields a significantly lower cost.
Moisture Content
Moisture in the inlet soil is volatilized in the reactor, and must later be condensed
out of the gas stream. Thus, moisture causes an increase in both the heating and
cooling requirements for the reactor system. In addition, and perhaps more
importantly, moisture increases the volume of gas which must be handled in the
system, proportionally increasing the reactor size for the same vapor residence
time. Because the reactor lengths were assumed to be constrained to 30 to 35
feet maximum (which could fit on a truck), the vapor residence times were
adjusted downward for soils with high moisture levels. This decreased reactor
capital costs, but increased operating costs since the organic vapor destruction
efficiency in the reactor decreased, thereby requiring the water treatment system
to remove more organics using expensive activated carbon. (Again, it should be
noted that the use of GAC as the primary means of separating the organics from
the condensed water was not necessarily the optimal choice. Rather it was done
to show the expected tradeoffs between destroying the organic compounds in the
sulfur vapor reactor before they condensed with the water vapor and destroying
them after they condensed with the water vapor.)
System costs at two moisture levels were compared in the table; 7% and 14%. A
decrease in moisture from 14 to 7% was shown to cause a decrease in overall
costs of approximately 8%. Although this decrease in estimated costs was
significant, it was not large enough to justify the addition of dewatering equipment
to drive the moisture level of the inlet soil to the single digit ranger. The only case
in which the use of such equipment may be justified is when the moisture level is
above 20%.
Number of Reactors
This parameter was a function of both the unit reactor capacity and the specified
vapor residence time. It was constrained by overall reactor size limitations, based
on the requirement that the reactor fit onto a truck for transportability. In general,
the analysis showed that costs were reduced by reducing the number of reactors.
Cases 4 and 5 are identical except that six reactors were used under Case 4 and
four were used under Case 5. Both sets of reactors were approximately 1,600
cubic feet (45 m3) in size. Case 5 was estimated to cost approximately 10% less
than Case 4.
One major tradeoff which was seen repeatedly in the economic analysis was the
tradeoff between reactor size, which affected overall destruction efficiency for the
organic compounds, and the costs of downstream wastewater processing. The
activated carbon costs under many of the cases represented 25 to 40% of the
overall system operating costs. Thus, if there were a more economical means of
removing the organic compounds from the wastewater stream, then the overall
system cost could be decreased significantly.
60
-------
One additional case was run, in which it was assumed that a large, fixed
installation was built in which the entire reaction occurred in one soil reactor with a
capacity of 10 tons/hr (9000 kg/hr). This scenario was otherwise identical to Case
1. Assuming a non-transportable plant with a scrap value of 40% of its original
value after one year, the cost per ton of soil rose from $147 to approximately
$160 per ton ($176/1000 kg). Thus, no advantage was seen in dropping the
constraints required by transportability and building a fixed installation.
61
-------
10.0 CONCLUSIONS AND RECOMMENDATIONS
The Sulchem process was shown to destroy certain polynuclear aromatic
compounds in soil {particularly higher boiling compounds such as pyrene, chrysene
and benzopyrene) i n a reactor when operated at temperatures between 300° and
350°C. However, a reactor configuration capable of efficient destruction of a
broader range of compounds was not obtained. This limitation may have been
more due to limitations of the laboratory study, rather than inherent limitations of
the technology. Specific conclusions from this laboratory study of the Sulchem
Process are as follows:
• Destruction within the soil reactor was strongly correlated with
compound boiling point;
organic compounds with boiling points above 350°C are
essentially completely destroyed in the process (destruction >
99.5%);
organic compounds with boiling points in the range of 250 to
350° are partially destroyed. However, a significant quantity
volatilize before destruction occurs;
organic compounds with boiling points below about 250° C
primarily volatilize from the soil reactor before reaction can
occur,
• A second stage sulfur/vapor reactor was shown to destroy a
significant percentage of the organics desorbed from the soil reactor,
thereby requiring subsequent treatment of the condensate produced;
• Metal stabilization in the treated soil (as measured by TCLP) is
achievable for certain metals (particularly lead, cadmium, zinc, copper,
and nickel) due to sulfide formation, with performance limits
depending on the chemical form and concentration (e.g., typically lead
below 10,000 mg/kg, cadmium below 1000 mg/kg);
• Remediation costs employing the Sulchem Process are estimated at
$105 to $183/ton based on site size, reactor configuration, and
processing rate.
Additional testing is recommended to demonstrate integration of the process
components. Only very limited testing of reactor configurations or techniques to
destroy volatilized organics were employed. CHMR recommends additional testing
of vapor-phase organic reactors at higher temperatures (400°C or higher) and
longer residence times. From this, the destruction efficiency (and its limits) need
to be determined for an integrated soil/vapor reactor system.
62
-------
REFERENCES
1. Center for Hazardous Materials Research. Simultaneous Destruction of
Organics arid Stabilization of Metals and Metal ions in Soils. Quality
Assurance Project Plan and Workplan. February 1993.
2. Center for Hazardous Materials Research. Simultaneous Destruction of
Organics and Stabilization of Metals and Metal ions in Soils. Quality
Assurance Project Plan, Revision 2. March 1994.
3. Center for Hazardous Materials Research. Simultaneous Destruction of
Organics and Stabilization of Metals and Metal ions in Soils. First Year
Report No. 300-3146-016. December 1993.
4. Center for Hazardous Materials Research. Simultaneous Destruction of
Organics and Stabilization of Metals and Metal ions in Soils. Second Year -
Interim Report, January 1, 1994 to August 31, 1994. No. 300-3146-016.
September 1994.
5. National Environmental Technology Applications Center and Center for
Hazardous Materials Research. Continuous Pilot-Scale Production Unit;
Demonstration of the Adams Process. No. C40-578/C1-123-2. August,
1991.
6. National Environmental Technology Applications Center. Adams Process
Development Tests. No. 7-1134-000. December 1991.
7. Freeport Sulfur Company, W. N. Tuller, ed., The Sulohur Data Book.
McGraw Hill, NY, 1954).
8. Texas Gulf Sulfur Bulletin, Handling and Storage of Molten Sulfur.
9. Federal Register, 52 (178), pp. 47464-47473, September 15, 1994.
10. Center for Hazardous Materials Research. Treatability Tests of the Suichem
Process. No. 300-3010-000. April 1993.
63
-------
APPENDIX A
RESULTS FROM PARAMETRIC STUDIES CONDUCTED AT 250°C
USING A VENTED STIRRED REACTOR
-------
TABLE A1 EFFECTS OF REACTOR PARAMETERS ON DESTRUCTION EFFICIENCY IN VENTED STIRRED REACTOR
Effect of Residence Time
Run #1
Temp C
250
Run #2 TempC
250
Run #3
Temp C
250
S%
0.130435
S%
0.130435
S%
0.130435
Time hrs
0.5
Time hrs
1
Time hrs
2
Boiling
Reactor
RECOVERIES (in mg)
Percent
RECOVERIES (in mg)
Percent
RECOVERIES (in mg)
Percent
Compound
Point (C)
Feed (mg)
Reactor
Ice Trap
Scrubber
Total ug Destroyed
Reactor Ice Trap
Scrubber
Total ug Destroyed
Rkr 30c
33-30-A1
Scr 30b
Total ug Destroyed
bibenzyl
285
368.5
42.25
46.17
1.39
89.81
75.6%
59.13 73.06
0.98
133.17
63.9%
3.87
229.18
0.00
233.05
36.8%
hexachlorobenzene
322
327.5
6.24
11.31
0.54
18.09
95.1%
10.52 23.35
0.37
34.24
90.7%
8.60
204.42
0.36
213.37
42.1%
anthracene
340
306.1
13.37
2.10
0.13
15.60
95.8%
13.91 5.14
0.08
19.13
94.8%
7.32
63.99
0.00
71.32
80.6%
pyrene
393
175
17.76
19.16
0.00
36.93
90.0%
12.04 14.22
0.00
26.27
92.9%
2.56
6.88
0.00
9.44
97.4%
Effect of Sulfur Loading
Run #1
Temp C
250
Run #2 Temp C
250
Run #3
Temp C
250
S%
13.0%
S%
13.0%
S%
13.0%
Time hrs
0.5
Time hrs
1.0
Time hrs
2.0
Boiling
Reactor
RECOVERIES (in mg)
Percent
RECOVERIES (in mg)
Percent
RECOVERIES (in mg)
Percent
Compound
Point (O
Feed fmal
Reactor
Ice Trap
Scrubber
Total uo Destroved
Reactor Ice T rap
Scrubber
Total ua Destroyed
Rkr 30c
33-30-A1
Scr 30b
bibenzyl
285
368.5
56.34
76.10
2.88
135.32
44.3%
23.26 63.22
0.68
87.16
28.5%
3.87
229.18
0.00
233.05
65.31%
hexachlorobenzene
322
327.5
12.88
20.13
0.89
33.90
12.8%
8.19 22.69
0.16
31.04
11.8%
8.60
204.42
0.36
213.37
65.15%
anthracene
340
306.1
15.52
5.00
0.20
20.72
14.3%
7.88 5.69
0.05
13.62
9.4%
7.32
63.99
0.00
71.32
23.30%
pyrene
393
175
7.17
6.61
0.10
13.87
12.5%
2.26 2.92
0.04
5.23
4.7%
2.56
6.88
0.00
9.44
5.39%
-------
APPENDIX B
QUALITY ASSURANCE/QUALITY CONTROL
*
-------
APPENDIX B
QUALITY ASSURANCE/QUALITY CONTROL
This appendix presents select quality assurance/quality control results from
duplicate experimental runs and analyses. Table B-1 shows analytical results from
duplicate analyses of samples from the heavy metals experiments in the vented
stirred reactor for experiment 33-13. The results show relative percent differences
below 15% for three of the four analytes.
TABLE B-1
DUPLICATE HEAVY METALS RESULTS FOR 33-13
Metal Analyte
Method
TCLP Result #1
TCLP Result #2
RPD
Cadmium
EPA 6010
0.17
0.31
58%
Lead
EPA 6010
0.41
0.46
11%
Nickel
EPA 6010
0.35
0.31
12%
Zinc
EPA 6010
4.60
4.85
5%
Results from TCLP analyses for lead conducted on split samples for which the
TCLP extractions were performed at two separate laboratories (Microbac
Laboratories and NETAC) are given in Table B-2. The results show good
agreement for two of the three analyses.
TABLE B-2
DUPLICATE TCLP EXTRACTION RESULTS FOR LEAD ANALYSES
Experimental
Reference #
Microbac
Extraction
Result (mg/L)
NETAC
Extraction
Result (mg/L)
RPD
Action
56
82.4
75.8
8%
—
58
77.1
57.9
28%
Reanalysis
(TCLP = 78.5)
67
30.3
32.0
5%
—
Table B-3 shows results from three sets of duplicate organics analysis. The table
shows relatively good agreement (RPD's generally below 15%) for the duplicate
analyses for experiments 30 and 31. The RPD's were all below the threshold limit
of 15%, except for some of the higher boiling compounds, which were found in
relatively low concentrations. The RPD's for Run #32 were relatively high.
Subsequent review of the analyses for Run 32 indicated a possible error in the
extraction for the second analysis, which may have biased the results. Therefore
the first analysis was accepted. Corrective steps were taken to ensure that the
error had not occurred previously and would not recur in subsequent analyses.
B-1
-------
TABLE B-3 RESULTS OF DUPLICATE ANALYSES OF SCRUBBER WATER FROM VENTED STIRRED REACTOR RUNS
*
Analyte
Run #30
Analysis 1
mg
Analysis 2
mg RPD
Run #31
Analysis 1
mg
Analysis 2
mg
RPD
Run #32
Analysis 1
mg
Analysis 2
mg RPD
mesitylene
225.5
224.0
1%
220.3
232.4
5%
199.6
243.8
20%
durene
249.3
254.6
2%
245.6
261.7
6%
184.8
245.8
28%
naphthalene
287.9
317.9
10%
296.3
313.4
6%
212.9
280.2
27%
2-Me naphthalene
276.9
302.7
9%
281.8
301.7
7%
190.6
269.2
34%
biphenyl
295.8
323.8
9%
298.4
323.5
8%
186.2
282.5
41%
bibenzyl
229.2
244.2
6%
221.5
254.6
14%
128.3
228.8
56%
hexachlorobenzene
204.4
99.3
69%
96.5
136.2
34%
84.1
205.4
84%
anthracene
64.0
27.7
79%
23.3
75.6
106%
29.3
69.0
81%
pyrene
6.9
2.6
91%
2.3
8.2
113%
4.5
12.7
95%
------- |