PB86-128717
EVALUATION OF EMERGING TECHNOLOGIES FOR THE
DESTRUCTION OF HAZARDOUS WASTES
Department of Health Services
Sacramento, CA
Jun 85
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
NITS
-------
EPA/600/2-85/069
June 1985
Evaluation of Emerging Technologies
for the Destruction of Hazardous Wastes
Edited by
Jan Radimsky and Arvind Shah
Toxic Substances Control Division
Department of Health Services
• 714/744 P Street
Sacramento, CA 95814
Cooperative Agreement No. R-808908
Project Officer
Harry M. Freeman
Thermal Destruction Branch
Alternative Technologies Division
Hazardous Waste Engineering Laboratory
U. S. Environmental Protection Agency
Cincinnati, OH 45268
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
- TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA/600/2-85/069
3 RECIPIENT'S ACCESSIOI»NO.
4- TITLE AND SUBTITLE
Evaluation of Emerging Technologies for the
Destruction of Hazardous Wastes
5. REPORT DATE
June 1985
6. PERFORMING ORGANIZATION CODE
7. AUTMORISI
Jan Radimsky and Arvind Shah» Editors
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Health Services
714 P Street
Sacramento, California 95814
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
R308908
12. SPONSORING AGENCY NAME AND ADDRESS
Hazardous Waste Engineering Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/12
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The objective of this report is to provide detailed information regarding
four innovative alternative technologies demonstration projects for treating
and destroying hazardous wastes. Under a cooperative agreement between the
U.S. Environmental Protection Agency and the State of California, the Department
of Health Services (DHS) carried out a pilot scale test program on the following
promising technologies.
1. High Temperature Fluid-Wall
2. Evaluation of Emission Tests from SunOhio
Mobile Treatment Process
3. Wet Air Oxidation
4. Evaluation of Emission Tests from Wet Air
Oxidation Zimpro Process
Thagard Research
Air Resources Board
State of California
Zimpro
Air Resources Board
State of California
Discussions of the above processes include project descriptions, results,
conclusions, and recommendations.
- This report was submitted in partial fulfillment of Cooperative Agreement
Nn R-anaqrifi unrtPr gpnncnr«:hip nf tho II S FPfl anri tho State, nf ralifnrnia HHS
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATi Ficld'Crour
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
107
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (»-73)
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
11
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation of
solid and hazardous wastes. These materials, if improperly dealt with, can
threaten both public health and the environment. Abandoned waste sites and
accidental releases of toxic and hazardous substances to the environment also
have important environmental and public health implications. The Hazardous
Waste Engineering Research Laboratory assists in providing an authoritative
and defensible engineering basis for assessing and solving these problems.
Its products support the policies, programs and regulations of the Environmental
Protection Agency, the permitting and other responsibilities of State and local
governments and the needs of both large and small businesses in handling their
wastes responsibly and economically.
This report discusses the evaluation of several alternative technologies
for treating hazardous wastes. The information contained in this report will
be useful to those within the public and private sector responsible for the
management of hazardous wastes.
For further information, please contact the Alternative Technologies
Division of the Hazardous Waste Engineering Research Laboratory.
David G. Stephan
Director
Hazardous Waste Engineering Research Laboratory
-------
ABSTRACT
The objective of this report is to provide detailed information regarding four innovative
alternative technologies demonstration projects for treating and destroying hazardous wastes. Under
a cooperative agreement- between the U. S. EPA and the State of California, the Department of
Health Services (DHS) carried out a pilot scale test program on the following promising tech-
nologies.
1. High Temperature Fluid-Wall Thagard Research
2. Evaluation of Emission Tests from SunOhio Mobile PCS Air Resources Board
Treatment Process State of California
3. Wet Air Oxidation Zimpro
4. Evaluation of Emission Tests from Wet Air Oxidation Air Resources Board
Zimpro Process State of California
Discussions of the above processes include process descriptions, experimental procedures, test
methods, results, and discussions, conclusions, and recommendations.
This report was submitted in partial fulfillment of Cooperative Agreement No. R-808908
under sponsorship of the U. S. EPA and the State of California, DHS.
IV
-------
TABLE OF CONTENTS
Page
SECTION 1: Introduction 1
SECTION 2: High Temperature Fluid-Wall, Thagard Research Company 2
SECTION 3: Air Resources Board's (ARB) Evaluation to Determine
Emissions from SUNOHIO's Mobile PCS Treatment Process 24
SECTION 4: Commercial Demonstration of Wet Air Oxidation of
Hazardous Wastes 48
A. Phenolic and Organic Sulfur Waste Classes
B. Petroleum Refinery Spent Caustic Wastewater
C. General Organic Waste Classes
D. Cyanide Waste Class
E. Pesticide Waste Class
F. Solvent Still Bottoms Waste Class
SECTION 5: Air Resources Board's Evaluation Test Conducted
on a Wet Air Oxidation Process to Treat
Hazardous Wastes 83
REFERENCES 99
-------
FIGURES
Page
Figure 2-1: Vertical Cross Section of a Typical HTFW Reactor 5
Figure 2-2: Horizontal Cross Section of a HTFW reactor 6
Figure 2-3: Schematic Diagram of the HTFW Reactor System 19
Figure 3-1: SUNOHJO Process Schematic 26
Figure 3-2: Chevron U.S.A., El Segundo Refinery, Concentrations of Benzene,
Other Aromatics & Aliphatic Measured at the Inlet to the
Control Systems vs. Process Lapse Time • 44
Figure 3-3: PG&E, Union City, Concentrations of Benzene, Other Aromatics
& Aliphatic Measured at the Inlet to the Control Systems vs.
Process Lapse Time 45
Figure 3-4: Maxwell Laboratory, San Diego, Benzene Concentration Measured
at the Inlet to the Control system vs. Process Lapse Time 46
Figure 5-1: Schematic of the Wet-Air Oxidation Process and the Air Pollution
Control System ..' . 85
Figure 5-2: ZIMPRO's Wisconsin Laboratory Method for Screening Potential
Wastes for Wet-Air Oxidation 87
Figure 5-3: Z1MPRO/CASMALIA Resources Process Sampling Locations
and Sample Analysis 88
Figure 5-4: Sampling Locations & Sample Parameters 92
Figure 5-5: Sampling Train for Gas Phase Cyanide, Sulfide, & Phenols 93
Figure 5-6: Sampling Train for Gas Phase Organics 94
VI
-------
TABLES
Page
Table 2-1: Critical Factors and Results in Volatile Chlorinated Hydrocarbon
Treatment 7
Table 2-2: Typical Test Matrix for Volatile Chlorinated Hydrocarbons 13
Table 3-1: Summary of Sampling Methods 32
Table 3-2: Summary of ARB Test Results (Chevron USA, El Segundo,
& PG&E, Union City) 33
Table 3-3: Summary of SCAQMD Test Results (Chevron USA, El Segundo
Refinery) 34
Table 3-4: Summary of TRW Incorporated Test Results (Chevron USA,
El Segundo Refinery) -. 35
Table 3-5: Summary of BAAQMD Test ResuIts'{PG&E, Union City) 36
table 3-6: Summary of ARB Test Results (Maxwell Laboratory, San Diego) 38
Table 3-7: Quantitative Analysis of Compounds Present in the Gas Stream
Entering and Leaving the Activated Carbon Cannister and
Oil Burner 39
Table 3-8: Concentration of Benzene, "Other Aromatics," and Aliphatics
Measured at the Inlet to the Vacuum Degasser's Control
System With Respect to Process Lapse Time 41
table 3-9: Mass Rate of Benzene, Toluene, & Aliphatic Hydrocarbons
from the Vacuum Degasser (ARB Data Only) 43
Table 4-1 A: Wet Air Oxidation Demonstration, Petroleum Refinery Spent
Caustic Wastewater 56
Table 4-2A: Casmalia Spent Caustic Feed Volatile Reaction 57
Table 4-3A: Casmalia Spent Caustic Oxidized Volatile Fraction ; 57
Table 4-4A: Casmalia Spent Caustic Feed Acid Fraction 58
Table 4-5A: Casmalia Spent Caustic Oxidized Acid Fraction 58
Table 4-6A: Casmalia Spent Caustic Feed Base/Neutral Fraction 59
Table 4-7A: Casmalia Spent Caustic Oxidized Base/Neutral Fraction 59
Table 4-8A: Wet Air Oxidation Demonstration, Petroleum Refinery Spent
Caustic Wastewater 60
Table 4-1B: Petroleum Refinery Spent Caustic Wastewater, Waste Code
No. 5003 64
Table 4-1C: Wet Air Oxidation Demonstration, General Organic Wastewater
Class 67
Table 4-2C: Wet Air Oxidation Demonstration, General Organic Wastewater
Class 68
Table 4-1D: Wet Air Oxidation Demonstration, Cyanide Wastewater Class 71
Table 4-2D: Wet Air Oxidation Demonstration, Cyanide Wastewater Class 72
Table 4-1E: Wet Air Oxidation Destruction of Pesticides Added to Acidic
Distillate Waste 75
Table 4-2E: Wet Air Oxidation Demonstration, Pesticide Waste Class 76
Table 4-3E: Wet Air Oxidation Demonstration, Pesticide Waste Class 77
Table 4-1F: Wet Air Oxidation Demonstration, Solvent Still Bottoms
Waste Class ., 81
Table 4-2F: Wet Air Oxidation Demonstration, Solvent Still Bottoms
Waste Class 82
Table 5-1: Summary of ARB Test Results 95
Table 5-2: Percent Reduction in Concentration Across the Reaction Vessel 96
Table 5-3: Percent Reduction in Concentration Across the Scrubber and
Carbon Bed 97
Table 5-4: Noncondensible Hydrocarbons Detected in the Gaseous Effluents
From the Separator 93
vii
-------
SECTION 1
INTRODUCTION
On January 20, 1981, the State of California entered into a cooperative agreement with the
Office of Research and Development, U. S. EPA, to evaluate selected promising technologies for
destroying hazardous waste using pilot scale test programs. This report summarizes detailed infor-
mation such as process descriptions, results, and conclusions for each process studied. The
mention of a process in the report should in no way be considered an endorsement of the process
-by either State of California or the U. S. EPA.
1 -
-------
SECTION 2
HIGH TEMPERATURE FLUID-WALL, THAGARD RESEARCH COMPANY
SUMMARY
Four volatile chlorinated hydrocarbons, dichloromethane; 1,1,1-trichloroethane; carbon
tetrachloride; and Freon-12 (dichlorodifluoromethane) and one nonvolatile chlorinated hydro-
carbon (hexachlorobenzene) have been decomposed in the High Temperature Fluid-Wall (HTFW)
Reactor in bench-scale tests to assess the applicability of the device for efficient destruction of these
particular compounds.
The hexachlorobenzene, loaded onto a solid radiation target, exhibited high «99.9999 per-
cent) destruction efficiency, while the vapors (which could be heated only by secondary thermal
conduction from solid radiation target) exhibited destruction efficiencies related inversely to the
compound heats of formation, indicating that vapor-phase reaction temperatures were lower than
the solid reaction. Destruction efficiencies ranged from 99.999965 percent for dichloromethane to
84.99 percent for Freon-12. Heat transfer analysis indicated that vapor heating is dependent on the
solid particle density, and that efficient heating (and destruction) of vapors can be achieved simply
by increasing the particle density.
-2-
-------
INTRODUCTION
j
HTFW Reactor
The HTFW Reactor was developed originally for the continuous dissociation of methane into
carbon fines and hydrogen. This particular process required the generation of stable temperatures
above 1,700"C and the prevention of precipitate formation on the reactor walls. This fact has
particular relevance to the project herein.
To achieve both goals simultaneously, the reacting stream is kept out of physical contact with
the reactor wall by means of a gaseous blanket formed by flowing an inert gas radially inward
through the porous reactor tube (or core). Both high temperatures and high rates of heat transfer
are achieved by heating the porous carbon core to incandescense so that the predominant mode of
heat transfer is by radioactive coupling from the core to the stream.
A partial vertical cross section pf a typical HTFW Reactor is shown in Figure 2—1, and a
horizontal cross section is shown in Figure 2—2.
The reactor is heated electrically with six carbon resistance heaters. Because of the extreme
temperatures encountered in operation of the device, the insulation package consists not of
refractory brick but of a radiation shield made of multiple layers of graphite paper backed up with
carbon felt
The short residence time associated with the reactor which demands the pulverization of solid
feeds also lends itself to a compact system where a reasonably high throughput is seen for a small
installation, and portability becomes an attainable design feature. Thus, it is suggested that the
HTFW Reactor can provide a practical solution to the problem of disposing of many multi-
component toxic and hazardous wastes. Work done with the reactor prior to this effort on a variety
of wastes, in addition to a wide range of chemical processing, has indicated that there was a
potential for this technology in waste disposal.
The Experimental Program
In the subject program, volatile chlorinated hydrocarbons of low molecular weight were
examined to identify and perhaps quantify the applicability of the reactor to this class of waste.
These tests then represent an extension of previous work where solid and liquid chlorinated hydro-
carbons loaded onto solid carriers were treated in the reactor with high destruction efficiencies.
• 3-
-------
The treatment of radiation-transparent vapors and gases in the HTFW Reactor involves conductive
heating of the vapors in physical contact with a target solid which is being heated radiatively by the
reactor, so the extremely rapid surface heating which is seen in destruction of materials loaded onto
solids is not encountered and the operating mechanisms differ in kind rather than just degree.
A detailed computer analysis of the governing parameters of secondary heating of gases by
radiatively-heated solids has been performed. The analysis gives rise to a conclusion that only a
small gas "bubble" surrounding the particle is actually conductively heated to a high temperature
and thus the governing parameter is actually the number of particles introduced into the reactor.
This analysis had not been completed at the time of the actual experimentation, so the design of the
experiments unfortunately could not take these factors into account Thus, while the results are
encouraging, they do not represent what a final process may accomplish.
-4-
-------
7 EXPANSION BELLOWS
2. POWER FEEDTHROUGH
COOLING MANIFOLD
4. POWER ^-^
F£EOrHflOL/GH
-------
HEAT SHIELD
INSULATOR
COSE RADIOMETER
PORT
RADIATION HEAT
\ X SHIELD
VESSEL WALL
COOLING
AATSR
•i SLECTHOOE
"RADIOMETER PORT
\
FIGURE 2-2: HORIZONTAL CROSS SECTION OF A HTFW REACTOR
-6-
-------
RESULTS AND CONCLUSIONS
Chlorinated Hydrocarbon Vapor Samples
The most important conclusion we draw from the work completed is that chlorinated hydro-
carbons introduced into the reactor in the vapor form are much more difficult to destroy than
similar materials loaded onto solids, given identical residence times and reactor temperatures.
While the governing factor in the case of the solids was the direct absorption of radiation by
the solid surfaces and consequent extremely rapid heating, the governing factor in the case of the
vapors was the conduction of the heat from the particle surface into the vapor—a much slower
process limited in rate by the thermal conductivity of the vapor itself.
Thus, as might be anticipated, the temperature levels achieved in the vapors for a given
residence time will not, in general, be as high as the temperatures achieved on the solids. The
experimental result of this behavior will be that minimum temperatures and minimum residence
times for complete destruction of the vapor-phase substances will not be achieved, and that the
observed destruction levels will now be critically dependent on the heat of formation of the
substance being investigated.
This was graphically demonstrated in the tests performed as indicated in Table 2-1:
TABLE 2-1: CRITICAL FACTORS AND RESULTS IN VOLATILE
CHLORINATED HYDROCARBON TREATMENT
MATERIAL
Dichloromethane, CH2C12
1,1,1-Trichloroethane, CC13CH3
Carbon Tetrachloride, CC14
Freon-12, Dichloridifluoromethane, Cd2p2
Hff KCAL/MOLE
21.0
32 (est)
25.5
94.0
DESTRUCTION
EFFICIENCY
99.999965%
99.99278%
99.9756%
84.99%
Chlorinated Hydrocarbons Solid Sample
The results for hexachlorobenzene (HCB) in soil are in agreement with HCB results on carbon,
thus duplicating both the destruction and analytical methods and further substantiating the
cpnclusion that the solids are rapidly heated and the toxic material effectively decomposed.
-7-
-------
Despite some data problems, the fraction of HCB remaining on the solids and in the effluent
gases was approximately 1QT6, corresponding to a destruction efficiency of approximately 99.9999
percent
The demonstration of gettering of chlorine (from decomposition of HCB) with calcined lime
(CaO) mixed with the soil produced no identifiable results. Since gettering of halogens and sulfur
with lime and subsequent fixing into a vitreous slag has been previously demonstrated on numerous
occasions with other materials we must conclude that some deficiency in the experimental
procedure was introduced (inadequate mixing of the lime in the soil, too low a CaO/Cl ratio for
this particular application, etc.). Subsequent experimental work will seek to quantify this
parameter which is an important system consideration of on-site disposal.
Discussion of Analytical Results
Comparison of the analytical results from the principal laboratory (West Coast Technical
Service [WCTS]) and the back-up laboratory (Rockwell) indicated poor agreement between the
two sets. Therefore, we have based our conclusions (on destruction efficiencies) primarily on the
WCTS results because of the greater number of samples analyzed by them and a much stronger
indication of internal consistency for this.particular set of experiments.
RECOMMENDATIONS
The bench-scale tests performed under this program utilized a general purpose three-inch
HTFW Reactor and -100 mesh carbon granules as the radiation target material to provide heating of
the chlorinated hydrocarbon vapors.
The results of destruction levels of these vapors are considerably poorer than results obtained
by loading the hydrocarbon directly onto solids. However, the critical operational parameters
relating to this behavior have been identified, and we believe that the destruction of vapors as such
without formation of degradation products can be accomplished.
The first three recommendations below are directed at the problem of incomplete degradation
of vapors.
The fourth recommendation is addressed at optimization of halogen gettering, while the fifth
calls for a demonstration of a high level of destruction of chlorinated dioxins and dibenzofurans.
-8-
-------
The last recommendation addresses the practical problem of mechanically introducing samples
as completely dispersed powders.
1. Keeping the ratio of carbon to vapor a constant, the destruction efficiency should be measured
as a function of the size of the carbon particles down to one micron, our prediction being that
the smallest sizes, corresponding to higher particle density, will result in higher destruction
efficiency.
2. Using the optimum particle size determined in No. 1 above, destruction efficiency should be
measured as a function of the ratio of the weight of carbon to that of the vapor.
3. The optimum particle size and concentration (relative to the vapor) should then be tested with
a variety of materials to assess the potential of the HTFW Reactor for direct destruction of
vapors. Trichloroethane is suggested to test for formation of degradation by-products: Freon-
12 would represent a difficult fully-halogenated material.
4. Gettering should be demonstrated and optimized using, for example, HCB on soil. The para-
meters to be varied will include particle size and stoichiometry.
5. A demonstration should be made of the efficient destruction of chlorinated dioxins and
dibenzofurans. If possible, less toxic isomers can be used as surrogates to reduce hazards.
Such a demonstration will be useful not only to qualify the HTFW Reactor for such
specialized materials, but also demonstrate that they are not created in the reactor.
6. In addition to determining optimum particle size it is necessary to optimize the methods of
introduction without allowing agglomeration. The present system is known to be far from
optimum, although^ critical design parameters are identified and much existing machinery from
the mining and food processing industries can be brought to bear on the problem.
BACKGROUND AND RELATED WORK
As mentioned in the introduction, the most generally acceptable methods for treatment of
many hazardous materials involve high temperatures. The use of high temperatures to destroy
chemical substances and to produce glasses has been standard practice for many years; however,
practical process equipment operating much above 1200°C has generally encountered limitations in
materials of construction. Most refractory liners either degrade mechanically or dissolve in the slags
found in many waste streams. Another limitation occurs because of the inefficiency arising when
-9-
-------
the process temperature approaches the effective flame temperature._ Maintenance of pyrolysis or
incineration equipment with moving parts in the high temperature region presents a potentially
severe problem when uncharacterized feed materials with fusible species are introduced or
precipitable products are formed.
In theory, the high temperature with the HTFW Reactor core should reduce samples to their
simple molecular forms. The high rate of cooling prevent reformation of potentially hazardous
recombination products, and the high rate of heating allows relatively large sampling rates at short
residence times. Hydrocarbons are readily pyrolyzed to C +1^; oxygen-bearing hydrocarbonaceous
materials are gasified to CO + H2 + C; halogenated organic compounds are decomposed to
C +• C12 + HC1. In the last case, the addition of calcium hydroxide (or calcium oxide) results in the
"gettering" of the chlorine to form CaC^. Inorganic compounds such as clay which melt below the
reactor core temperature are fused into vitreous solids. Most metal salts, such as CaC^, are soluble
in such molten glasses leaving vitrified beads with the salts locked up in solid solution.
Organic Compounds
In the absence of added secondary materials, organic compounds will be pyrolyzed to produce
elemental or simple molecular forms of the starting materials, e.g., C, h^HCI, or C12 for
chlorinated hydrocarbons.
In the presence of oxygen (air) they will be gasified to produce CO, V\2 and HC1.
In the presence of oxygen, h^O and Ca(OH>2, gasification will be accompanied by gettering of
the chlorine. In the presence of soils or sediments, or if additional clay is added, the CaC"^ can be
locked up in an insoluble vitreous mass.
Hexachjorobenzene
A final reporttitled "A Feasibility Study of a High Temperature Fluid Wail Reactor (HTFWR)"
was submitted to the Industrial Environmental Research Laboratory of EPA in Edison, New Jersey,
in February 1981.* The primary objective of this feasibility program was to demonstrate that the
surrogate material HC8 is effectively degraded during passage through the HTFW Reactor operating
Work carried out under contract with Mason and Hanger.
10-
-------
under normal conditions. The reactor used was the six-inch prototype, designed, and operated by
Thagard Research Corporation at South Gate, California. HCB was selected as a surrogate both to
avoid the difficulties of experimenting with hazardous materials such as polychlorinated biphenyls
(RGBs) or Kepone, and to provide a substitute which is similar but which is know to be even more
difficult to decompose. ^
Measurements were carried out on October23, October30, and Novembers, 1980. The
experiments indicated that in excess of 99.9999 percent of the original material was destroyed for a
core temperature of 2200°C (4000°F) and a residence time of approximately 0.1 second. Since
HCB is known to have greater thermal stability than PCBs or Kepone, it is expected that the
degradation efficiency for these hazardous materials would be at least as high as indicated for HCB.
PCB-Contaminated Sediment
In a recently concluded experimental program,** the HTFW Reactor was used to demonstrate
destruction of PCBs in solids. Samples consisted of approximately 1 percent PCS as Aroclor 1242
in either fine soil for -100 mesh carbon. The conclusions reached were:
1. An overall reduction in concentration of PCS by a factor of 2.3 x 10"° was achieved with a
standard deviation of 0.8x10" . This corresponds to a destruction efficiency of
99.99972 percent
2. Destruction was essentially similar for carbon and soil substrates, with slightly higher
efficiencies for soil.
3. Destruction efficiency was essentially independent of feedrate in the 50—100g/m range for
•the 3-inch reactor at 2343°C.
4. A 2:1 reduction in HCI was observed for a 2x stoichiometric addition of lime. Thegettering
factor was limited by the unexpected presence of sulfur in the soil with the resultant competi-
tion to form CaS.
'• The program was carried out by Baird Corporation under agreement No. N2560S5ZKX with Rockwell International Corpora-
tion. Environmental and Monitoring Services. Rockwell, in turn, was operating under contract No. 68-03-3014 with the
U. S. EPA.
1 R. A. Games "Thermal Degradation of Kepone" in the News of Environmental Research in Cincinnati, MERL, February 15,
1977.
-11 -
-------
While a high efficiency of destruction was demonstrated, it was not as high as attained in
earlier work in HCB. This is attributed to problems with sample introduction in the three-inch
reactor, compared to the six-inch reactor used previously.
Limitations of the HTFW Reactor
Two requirements for efficient operation of the HTFW Reactor may be regarded as limitations
for degradation of hazardous materials. First, for efficient radiation heating to occur, the sample
must possess reasonable optical absorption in the near infrared. If the substance is optically
transparent, an auxiliary absorbing material (such as .carbon) must be added. The volatile
chlorinated hydrocarbons considered in this program were introduced as vapors with low optical
absorption. The transfer of energy to the vapor via fine carbon particles involves collisions between
the carbon and vapor molecules which is inherently less efficient and slower than direct optical
absorption. In the case of hazardous materials on solids or sediments, the requirement will often be
satisfied automatically by the absorption charteristics of the soil or sediment. The other require-
ment is that the sample be finely divided. While the limiting particle size depends on optical
absorption, experience has demonstrated that sample size should usually be -100 mesh or finer.
This precludes handling of waste directly in many instances and will often demand preprocessing.
On the other hand, many soils and sediments have the required fineness for direct injection, and up
to 50 percent water may be tolerated in certain cases.
EXPERIMENTS
Design of Experiments
The experimental program on the thermal destruction of chlorinated hydrocarbons (CLHC)
was based on our experience in conducting preliminary studies, the destruction of HC8 and PCS, as
discussed earlier. All experiments were scheduled to be carried out over consecutive days to ensure
that no other materials entered the reactor during the test period.
Statement of Work
As outlined in the proposal, the project was conducted under eight tasks as follows:
Task A: Preparation of a detailed program plan including development of quality assurance
methods to guarantee validity of analytical results, and preparation of safety procedures
to assure personnel safeguards during the experimental phase of the program.
-12-
-------
Task B: Performance of thermal destruction tests on dichloromethane at 2200°C.
Task C: Performance of thermal destruction tests on carbon tetrachloride at 2200°C.
Task D: Performance of thermal destruction tests on Freon-12 at 2200°C.
Task E: Performance of thermal destruction tests on 1,1,1-inch I oroethane at 2200°C.
Task F: Performance" of thermal destruction tests on hexachlorobenzene at 2200°C, 2000°C,
1700°C, and 1400°C; also, with and without the addition of calcium oxide.
Task G: Correlation and interpretation of all data and incorporation in a final report.
Test Matrices
The test and effluent samples to be collected for a typical CLHC are given in Table 2-2.
Both the soil background and the CLHC samples were run in triplicate to provide additional
precision and accuracy. For trichloromethane, three additional tests were performed at successively
lower core temperatures (1000°C, 1700°C, and 14008C) to assess the importance of core
temperature.
TABLE 2-2: TYPICAL TEST MATRIX FOR VOLATILE CHLORINATED HYDROCARBONS
TEST
NO.
1
2
3
4
5
6
MATERIAL
Carbon
Carbon + CLHC
Carbon
Carbon + CLHC
Carbon
Carbon + CLHC
FEED
X
X
PAN
X
X
X
X
X
X
CARTRIDGE
X
X
X
X
X
X
HEAT
EXCHANGED
all at
2200° C
X
-13-
-------
For HCB, the above typical matrix was augmented to illustrate the "gettering" of chlorine by
lime. In order to demonstrate gettering and production of a nonleaching product, soil was used
with HCB. Therefore, the matrix of tests given above was modified by substituting "soil" for
"carbon" and including additional tests with lime. The extra tests were:
TEST
NO.
6
7
8
MATERIAL
Soil -r HCB
Soil + Lime
Soil +-Lime
+ HCB
FEED
X
X
PAN
X
' X
X
CARTRIDGE
X
X
X
HEAT
EXCHANGER
X
BUBBLERS
X
X
Bubbler samples were taken at the end of Test 6 document, the chlorine present without the
addition of lime; those taken at the end of Test 8 were taken to document the chlorine remaining
after "gettering" by lime.
In the earlier plans, the variation in core temperatures was assigned to tests on HCB; however,
they were performed in tests on trichloroethane as indicated above. Also, the earlier plans called
for measurements in vinyl chloride. Because-of unavailability of vinyl chloride, Freon-12 was
substituted.
Sample Size
Sample size was chosen to allow determination of destruction of efficiency to 99.9999 percent
or better. In a typical experiment, the feedstock consisted of approximately 2kg solids with a
loading of 20g CLHCs (1 percent) which was introduced at a rate of approximately 100g/min.
for 20 minutes. If ail of the effluent was analyzed, 20g of sample had to be detected to realize the
desired sensitivity. In the case of vapor samples, approximately 1/12 of the total vapor was
sampled, requiring a detection sensitivity of 1.7g. Since detection sensitivities with GC and electron
capture are usually at least on order of magnitude better than this, the required sensitivity was
realized.
-14-
-------
Reactor Modifications
Prior to actual experimentation, the reactor was fitted with a new qore to ensure the absence
of cross contamination by any previous tests. The entire reactor from just below the hot zone to
the final exit to the scrubber (see Figure 2-3) was disassembled and scrubbed with water and deter-
gent. After a water rinse, the stainless steel parts were treated with sulfuric acid and then passivated
with dilute nitric acid and finally rinsed again with water.
A gas/liquid injection system was added to allow co-introduction of gas/liquid samples with
the solid substrate, as will be described in the next sections.
Feedstock Preparation
All of the feedstocks except HCB are volatile haloge'nated hydrocarbons, and all of these
except Freon-12 are liquids under normal temperature and pressure. Liquid samples are most easily
introduced into the reactor by causing them to be absorbed onto a solid (such as fine carbon or
clay) at a known concentration, and introducing the spiked solid with a Vibra-Screw Feeder.
Because of the volatility of these samples there was a question about the integrity of such samples
as they were handled, poured into the hopper, etc., and exposed to air. Therefore, it was decided to
introduce the volatile samples as vapors, mixed with the sweep nitrogen. The absorbing fine carbon
was introduced separately with*the Vibra-Screw Feeder. Hence, all feedstocks except HCB were
used as delivered.
Preliminary mixing of HCB and soil was performed in a hood with a high flow rate, thus
eliminating any exposure to HCB vapor or soil with HCB. The mixed materials were then placed in
clean sealed metal cans and carried to the V-blender. After careful addition to the blender, the
sample was blended for a minimum of four hours after which it was again placed into the cans and
sealed. During blending and transfer of material, personnel wore disposable gloves, coveralls, and
dust masks. Final distribution and weighing into labeled cans was performed under the hood. All
exposed surfaces were cleansed with hexane and the rags used were sealed in a can for disposal.
Safety Procedures
The flasks to be used for vapor entrainment were loaded with CLHCs in the laboratory under a
hood, using standard laboratory practice.
15-
-------
Safety procedures for mixing HCB and soil have been detailed above under Feedstock
Preparation.
During the Reactor operation, each crew member had specified tasks. The person responsible
for loading samples into the hopper wore not only disposable gloves and coveralls, but a mask
designed to trap organic vapors as well as dust. Other crew members wore coveralls, gloves, and
dust masks. A plastic tent structure had been erected to protect personnel against potential release
of sample material du-ing Reactor loading. Fresh garments were available as needed. Visitors were
kept at a distance from .'ie Reactor and were supplied with coveralls, gloves, and masks as needed.
Tests of December 8—13,1982
Most of the tests were carried out during the period of December 8—11, 1982. On Friday
afternoon, December 11, the Reactor became clogged during the second test involving HCB in soil,
and tests had to be discontinued. At 'that time, it was felt that the remaining tests would not be
completed prior to dismantling of the Reactor and moving to a new location. However, the Reactor
was successfully repaired over the weekend and the last six tests on HCB/soil were carried out
Feedstock Introduction
Feedstock vapor was introduced directly into the nitrogen sweep gas. The liquid in a gas-
washing bottle was maintained at 0°C in an ice bath, and the rate of nitrogen flow bubbling through
the bottle was adjusted to vaporize approximately 1 g/min. of feedstock. Assuming that the
nitrogen flowrate was low enough to permit the vapor to reach equilibrium, the rate of nitrogen
flow at standard temperature and pressure to produce 1 g/min. of feed is given by:
F = 760 „ 22.4
VP MW
N, . • . •
N = desired nitrogen flow, 1/min.
VP = vapor pressure of the feed gas at 0°C, mm Hg.
MW = molecular weight of feed gas.
16-
-------
For the three liquid samples the required nitrogen flow rates are estimated in the following
table:
MW
VP@0°C(mmHg)
2C12
85
130
1.54
CC14
154
33
3.35
CH3CC13
133
36
3.56
These flows were easily realized using standard flow meters. The calculation'was only used to
produce an approximate sample flow of 1 g/min. The actual feed during a 20-minute test was
determined by loss of weight of the gas-washing bottles.
In a typical test the gas-washing bottle.containing a halogenated hydrocarbon liquid gas placed
in the ice bath until it had equilibrated. Just before the test, it was removed from the bath, dried
quickly, and weighed. It was then returned to the bath, attached to the nitrogen line and to the
sweep line into the Reactor, and the test was begun. After the test, the bottle was again removed,
dried, and weighed again. Between tests, the intet and outlet were sealed with parafilm to prevent
water condensation or contamination.
The procedure with dichlorodifluoromethane (Freon-12) was slightly different because it is a
gas at room temperature and pressure. The Freon was contained in a standard 14 oz. pressurized
•>
refrigerant can with valve. The can was not cooled and the desired flow rate for 1 g/min. was
estimated by:
0.189 1/min.
The actual weight of Freon used during a test was determined by the loss of weight of the
container.
For all 4 of the vapor samples discussed above, fine carbon (-100 mesh) was co-introduced into
the Reactor via the Vibra-Screw Feeder at the rate of approximately 100 g/min. The HCB-in-
carbon sample was introduced in the same fashion.
17-
-------
Effluent Collection
Vapor samples (with entrained fine particulates) were withdrawn continuously from a point in
the gas discharge line directly before the cyclone and baghouse, using a calibrated pump. Since the
gas flowing through the Reactor is controlled (i.e., the Reactor is a closed system with no leaks),
a metered sample flow may be related to total throughput The vapor samples passed through a
water-cooled heat exchanger prior to entering a cartridge where any undegraded CLHCs were
adsorbed. For HCB, the cartridge contained a polyurethane plug, as in previous experiments: For
the other CLHCs, the cartridge contained activated charcoal, following NIOSH methods for "purge-
able halocarbons." Sampling points are indicated in Figure 2-3.
Vapor sampling began before introduction of CLHC had begun and was terminated after
CLHC feed had terminated. (For a nominal 20 g input of CLHC, sampling time was approximately
20 minutes at a rate of 1 g/min.). After a.test, the cartridge was removed and refrigerated prior to
shipment to an analytical laboratory.
Solids fell into a pan at the bottom of the Reactor. A fresh one-gallon steel can was positioned
under a stainless steel funnel to receive the sample automatically. At the termination of a test the
can was immediately capped preparatory to shipment to a laboratory for analysis.
The heat exchanger was necessary to ensure that the temperature of the vapor effluent was low
enough to be trapped efficiently by the cartridges. Previous work had shown that the material
trapped in the heat exchanger was small but not negligible. Therefore, the entire heat exchanger
glassware was shipped for analysis; however, the heat exchanger was used over an entire series of
tests, so any remanent source material would be apportioned to the various tests according to the
inputs.
Effluent Analysis
Two EPA-approved laboratories were selected to perform analyses. WCTS of Cerritos,
California, was the primary laboratory, with confirming work performed by the Environmental
Monitoring and Services Center (EMSC) of Rockwell International in Newbury Park, California. In
the standard matrix of six tests the middle pair of carbon and carbon/CLHC effluent samples was
shipped to Rockwell, while the others were analyzed by WCTS.
Analysis of the outside sample pairs by a single analytical laboratory was expected to establish
average effluent levels and determine whether tests had proceeded uniformly. Analysis of the
-18-
-------
SEALED SCREW
FEEDER
GAS SAMPLING
(CARTRIDGE)
HEAT
EXCHANGER
NITROGEN BLANKET
INLET
OUST COLLECTION
(8AGHOUSE)
3-INCH
REACTOR
GAS SAMPLING
SYSTEM
(BUBBLERS)
SEALED SOLIDS
COLLECTION CHAMBER
(PAN)
PARTICIPATE
COLLECTION
(CYCLONE)
POST REACTOR
HEAT EXCHANGER
VAPORS
(VENTURI
SCRUBBER)
FIGURE 2-3: SCHEMATIC DIAGRAM OF THE HTFW REACTOR SYSTEM
-19-
-------
central pair of samples by a second laboratory was expected to be a basic check on the first labora-
tory using samples expected to be most typical for a set of tests.
Sampling and analysis methodologies were selected with consultation of WCTS to meet the
special nature of the effluents (e.g., a vapor with relatively large quantities of fine carbon) and also
provide the necessary sensitivity. For HCB the selected sampling and analytical methodology was
based on the work of IERL in Research Triangle Park. The adsorbing cartridge contained a poly-
urethane foam (PUF) plug with a glass filter to trap the fine carbon. Analysis, after extraction
of sample from both PUF and pan solids, was essentially that of EPA Method 612 for Chlorinated
Hydrocarbons.
For the remaining four vapor samples a new, larger cross section cartridge, containing activated
charcoal and a glass filter for carbon was designed, based on NIOSH Method P&CAM127 for
Organic Solvents in Air. After extraction.with selected solvents, analysis was essentially that of
EPA Method 601 for Purgeable Halocarbons.
All analysis involved solvent extraction followed by GC with electron capture detection. The
methodology involved solvent extraction followed by GC with electron capture detection. The
methodology is summarized in the following table:
Sample
1. Methylene Chloride
Z Carbon Tetrachloride
3. 1,1,1-Trichloroethane
4. Freon-12
5. Hexach I oro benzene
Extraction
Solvent
I so octane
I so octane
I so octane
Pentane
Ether/hexane
GC Column
Temperature
60°C
120°C
120°C
50°C
160°C
Related EPA
Method
601
601
601
601
612
Charcoal Cartridges
For analysis, the activated charcoal was poured into a 7 cm x 45 cm column and desorbed with
250 ml of solvent The entire column was shaken continuously for 3 to 5 minutes. The eluant was
collected and analyzed immediately for the compound of interest by direct injection into the GC.
• 20
-------
The eluant was concentrated and brought to 10ml for analysis of nonvolatile chlorinated hydro-
carbons. The extracts were analyzed on a gas chromatograph equipped with an electron capture
detector.
The gas chromatograph conditions for the chlorinated solvents were as follows:
Gas Chromatograph: Varian 3700
Column:
Detector:
6' x 2mm glass column packed with 1% SP-1000 on Carbonpack B
Electron Capture
Detector Temperature: 300°C
Injector Temperature: 200°C
Column Temperature: See Table
Extraction Solvent; See Table
PUP Cartridges
The PUF filter samples were extracted in a Soxhlet extractor which had been precleaned by
extracting continously with 10 percent ether in hexane for 20 hours. The ash on the screen, the
glass filter, and the PUF filter were placed in the extractor and the sample was extracted
continuously with 10 percent ether in hexane for 24 hours. The extract was concentrated, brought
to 10 ml and analyzed by direct injection into the GC. The gas chromatographic conditions were as
follows:
Gas Chromatograph:
Column:
Detector:
Detector Temperature:
Injector Temperature:
Column Temperature:
Varian 3700
6' x 2mm glass column packed with 1.95% SP-2401 and 1.56 SP-2250
on 150/120 mesh Supelcoport
Electron Caputre
300°C
250°C
160°C
-21
-------
Pan Samples
The pan samples were approximately 2000 grams each. The sample size made it necessary to
use a method of extraction that would include agitation of the sample. Also, the sample size was
much too large for extraction in a normal size laboratory soxhlet extractor..
Each sample was divided into three aliquots. Each aliquot was placed into a two-liter
Erlenmeyer flask. A sufficient amount of methylene chloride was added to each sample aliquot tc
achieve saturation. An additional 300 milliliters was then added to the first flask, and the flask was
agitated for approximately ten minutes.
The excess methylene chloride from the first flask was decanted into the second flask, and the
second flask was agitated for approximately ten minutes. The excess methylene chloride was
decanted into the third flask, the third flask was agitated for approximately ten minutes. The
excess methylene chloride from the third flask was decanted into a Kuderna-Danish apparatus, and
the methylene chloride was evaporated.
This process was repeated one more time.
The final step was to add 300 ml of fresh methylene chloride to each two-liter flask, and
agitate each flask for 10 minutes. The excess methylene chloride was decanted into the Kuderna-
Danish appartus, and the methylene chloride in each flask was removed by filtration through a
Buchner funnel. The total pan ash of each sample was recovered from the filtration step and stored.
All methylene chloride fractions from each extraction step were collected and concentrated in
the same Kuderna-Danish apparatus.
The total extract volume was reduced to less than 1.0ml, exchanged into hexane, and brought
to 1 .Oml for analysis.
The extract was analyzed by direct injection into the GC. The analysis conditions were as
follows:0
Gas Chromatograph: Varian 3700
Column: 6'x2mm glass column packed with 1.95% SP-2401 and 156%
SP-2250 on 100/120 mesh Supelcoport.
Detector: Electron Capture
-22-
-------
Detector Temperature: 280°C
Injector Temperature: 250°C
Column Temperature: 180°C
Miscellaneous Samples
The heat of exchanger samples were extracted using the same procedure as the PUF filter
samples. The soil feedstock was extracted in parallel with the pan ash samples. The analysis
conditions for the heat exchange samples were identical to the above described conditions for the
pan samples.
The analytical methodology described above was used by WCTS, the principal analytical
laboratory. Rockwell's methods were similar.
While analysis was principally for the specific hazardous material placed in the feedstock,
the GC traces were examined for evidences of other materials which might be products of
incomplete pyrolysis, with special attention given to either tetrachlorodibenzodioxin or tetra-
chlorodibenzofuran.
No evidence was seen of these last two compounds in this particular test series.
•23-
-------
SECTION 3
AIR RESOURCES BOARD'S (AR8) EVALUATION
TO DETERMINE EMISSIONS FROM SUNOHIO'S MOBILE PCS TREATMENT PROCESS
Three evaluation tests were conducted to allow determination of emissions from SUNOHIO's
mobile PCS treatment process. The mobile unit was tested while treating contaminated oils at three
locations: Chevron's USA refinery at El Segundo, California; Pacific Gas and Electric (PG&E)
Company's facility at Union City, California; and Maxwell Laboratory's facility in San Diego,
California.
PCBs were not detected in samples of emissions taken at two of the tests. However, relatively
high benzene and aliphatic hydrocarbon concentrations were measured in the units' uncontrolled
exhaust gases. Measured concentrations ranged from 18 to 7,000 ppm for benzene and 700 to
2,700 ppm for aliphatic hydrocarbons. Low process volumetric flow rates resulted in the mass
emission rates for benzene and the aliphatic hydrocarbons to values of approximately 0.1 !b/hr. and
below. Data 'from samples taken by the South Coast Air Quality Management District (SCAQMD)
staff during the El Segundo test indicate that dioxins and furans are not present above the limit of
detectability.
The efficiency of the carbon adsorption control system for preventing emissions to the
atmosphere of benzene and aliphatic hydrocarbons was found to have varied from approximately
99 to 30 percent, indicative that carbon adsorption breakthrough occurred. The 30 percent
efficiency calculation was based upon concentration measurements from the evaluation test
conducted at the PG&E facility.
An alternative control system was tested that utilized an oil fired furnace, a component of the
PCBX process, to incinerate the emissions. Furnace exhaust gas samples indicate a general reduc-
tion in concentration of compounds measured across the furnace. Benzene and toluene were not
detected in the furnace exhaust gas.
INTRODUCTION
On December 2 and 10, 1982 and November 30, 1983, the ARB's Engineering Evaluation
Branch conducted evaluation tests on chemical processing equipment designed to reclaim trans-
former oils contaminated with PCBs. The process tested is known as the "PCBX" process and was
developed by SUNOHIO, a partnership between the Sun Company of Radnor, Pennsylvania and the
-24-
-------
Ohio Transformer Corporation of Louisville, Ohio. The equipment is installed on two mobile
trailers and can be driven to different geographical localities to treat contaminated oils on site. The
first test was conducted on December 2, 1982 at Chevron's USA El Segundo refinery. This was a
joint venture between SCAQMD and the ARB. The second test was conducted on December 10,
1982 at a PG&E facility in Union City. The third test was conducted on November 30, 1983 at the
San Diego facility of Maxwell Laboratory.
The objectives of the evaluation tests were to allow determination of emissions from the unit's
vacuum degasser and determination of the efficiency of two prototype emission control systems.
One control system consists of an oil mist eliminator in combination with an activated charcoal
filter to control emissions from the vacuum degasser's vent pipe. The other control system utilizes
an oil fired furnace that is a component of the PCBX process to incinerate the emissions.
CONCLUSIONS AND RECOMMENDATIONS
• - »
Based on the analytical results and staff experience obtained from the ARB's evaluation test
conducted on the SUNOHIO PCBX process, the following observations are made:
1. The activated carbon adsorption canister used during testing is not big enough to provfde
effective emissions control for an extended period of time, the control system should be (1)
redesigned to have a larger activated carbon adsorption unit, or (2) revise the maintenance
schedule for the present carbon canister to require canister replacement with a frequency
commensurate with a demonstrated breakthrough time.
2. If emissions are to be prevented from the carbon canister, the carbon canister breakthrough
should be monitored with a continuous analyzer.
3. Based upon the results of this test, the oil fired furnace as a control device appears to be
effective.
EXPERIMENTAL PROCEDURES
o
Process Description
SUNOHIO's PCBX process is a chemical treatment process designed to strip chlorine atoms
from the PCB compound to form sodium chloride. The resulting biphenyl is polymerized into
polyphenylene. A process schematic is presented in Figure 3-1.
•25-
-------
FIGURE 3-1
SUNOHIO Process Schematic
38'
UNTREATED
OIL TANK
WHERE:
A - ADJUSTABLE RATE. POSITIVE
DISPLACEMENT PUMPS
B - AN AUTOMATICALLY
CONTROLLED HEATER
C - A METERED FLOW REAGENT
INJECTION DEVICE
D - REAGENT STORAGE
E - A MOTOR-ACTUATED MIXING
CHAMBER
F - A SPECIALLY DESIGNED
REACTION VESSEL
G --A SERIES OF HEAT EXCHANGERS
H - DUAL FILTER BEDS
I - A VACUUM DEGASSER
J - A 1,000 GALLON RETENTION
TANK
-------
PCB-contaminated oils are pumped from either a storage tank or an energized transformer to
an automatically controlled heater at flow rates ranging from 500 to 900 gallons per hour. Initial
f C8 concentrations in the untreated oil are reported to be as high as 10,000 ppm. Process tempera-
tures are maintained below 302" F (150°C) at an average operating pressure of 40 psi. Reagent is
injected into the oil downstream of the heater. A mixing chamber is used to ensure complete
blending between the reagent and oil. The residence time required for reaction to occur is provided
by passing the mix through an extended length of tubing located between the mixing chamber and
heat exchangers. Solids are removed from the treated oil by a centrifuge (not shown in schematic)
and gases by a vacuum degasser. Both the centrifuge and vacuum degasser are process equipment
from which emissions to atmosphere can occur. The contaminated oil is continuously recirculated
and reagent injected until PCB concentrations in the oil are reduced to approximately 2 ppm.
Exhaust gases from the vacuum degasser and centrifuge are vented through two 2-inch
diameter pipes directed out the side of each trailer. The vacuum degasser is in the main trailer and
the centrifuge in the auxiliary trailer. There is a vent pipe for each of the process units.
Description of the Two Proposed Emission Control Systems
Two emission control systems, designed by SUNOHIO,were proposed for testing. One system
consists of condensate knockout drums arranged in series with a carbon cartridge. The other
proposed system utilizes the on-board PCBX process furnace to combust emissions.
The principle components of the knockout drum/carbon cartridge emission control device are
two 55-gallon drums arranged in series with an oil mist •eliminator/activated charcoal filter unit
attached to the outlet of the last drum. Also, a condensate knockout is placed in between the
drums. This system is intended to be attached directly to the vacuum degasser and centrifuge vent
pipes to control emissions before being discharged to the atmosphere. Each process unit will have a
control system dedicated to it.
The proposed combustion control system was connected to the vacuum degasser's vent pipe.
A vacuum pump draws vapors from the degasser and discharges them into a 30-gallon pressure vessel
for storage. The PCBX process furnace fires intermittently. When the burner is on, a furnace
activated solenoid valve opens and vapors are drawn from the pressure vessel to the furnace. When
the burner is idle, the solenoid valve is closed. As a safety measure when the furnace is off, if the
system's pressure exceeds a pre-set value, a pressure relief valve opens and vapors are vented to the
atmosphere through an activated carbon cannister.
-27
-------
Test Methods
1. SUNOHIO Evaluation Test Conducted at Chevron's USA El Segundo Refinery
The ARB staff conducted tests to evaluate uncontrolled emissions from the vacuum degasser
while the SUNOHIO unit was treating PCB-contaminated oil directly out of a transformer.
Samples representing uncontrolled emissions were taken from the vent hose at the inlet to the
emission control system. The types of samples taken and the methods of sampling are sum-
marized in Table 3-1, Summary of Sampling Methods. Organic compounds were collected on
Tenax resin, florisil and in bags and analyzed by GC/MS, GC/FID, and GC/ECD.
The ARB staff did not sample emissions across the activated charcoal filter. Per discussions
held at a pre-test meeting, the participants agreed that the SCAQMD staff would sample at
these points to obtain information they would require for their permitting process.
Low molecular weight hydrocarbons were sampled using two sampling trains: a Tenax tube
was placed after a cold trap in the series arrangement. A total of 7 liters of exhaust gases were
sampled at a rate of 0.2 liters per minute. The adsorbed gases were thermally desorbed and
analyzed by GC/FID and GC/ECD. For the bag sampling, three-bag samples were analyzed
using GC/FID and GC/MS. A 27-minute sample was collected in the first bag at a rate of 0.8
liters per minute. Simultaneous samples for 15 minutes at 0.8 liters per minute were drawn
into the last two bags.
Two samples for the PCB analysis were collected on florisil adsorbent. A total of 8.92 liters
were sampled through one of the florisii tubes at a rate of 0.12 liters per minute and 5.27 liters
through the other at 0.1 liters per minute. A florisil blank was taken before the start of the
. first PCB, florisil test run.
Two ambient samples for background analysis were taken in bags, but without the probe and
impinger/ice bath. One bag sample was given to SCAQMD and the other to the ARB's Haagen-
Smit Laboratory for analysis.
•28-
-------
The results of all laboratory analyses are discussed in Result and Discussion, Summary and
Discussion of Test Results, and presented in Table 3-2.
SCAQMD conducted emission tests concurrently with the ARB's evaluation test. SCAQMD
tested to determine the efficiency of the carbon canister and the emissions to atmosphere from
the control system. The sampling locations were primarily used in testing the effectiveness of
the emission control system's carbon canister for both the vacuum degasser and the centrifuge.
A summary of the SCAQMD test results is presented in Table 3-3.
Flow rates were measured by the SCAQMD test team using a Roots meter attached to the
outlet of the carbon canister. These flow rates were used by the ARB in calculating pollutant
mass rates. _
Additionally, Chevron USA contracted TRW Inc. to perform tests concurrent with the ARB
and SCAQMD for audit purposes.' Results of analyses done by TRW on samples of the vacuum
degasser effluents are summarized in Table 3-4.
2. SUNOHIO Evaluation Test Conducted at a PG&E Facility in Union City
In addition to testing the uncontrolled emissions from the vacuum degasser, tests to allow
determination of emissions to atmosphere after the control system were also conducted. At
the request of the ARB, SUNOHIO did not replace the activated carbon canister used at El
Segundo. The control system's configuration for the Union City test was identical to that used
during testing in El Segundo. However, the mobile unit rather than being used to treat the oil
directly out of a transformer, as was done for Chevron USA at El Segundo, the mobile unit
treated PCB-contaminated oil taken from a storage tank.
The test methods and strategy used at the PG&E facility were the same as those applied at El
Segundo and are summarized in Table 3-1. Results of the test are presented in Table 3-2.
Flow rates were measured by the ARB test personnel using a Roots meter attached to the
outlet of the carbon cannister.
The Bay Area Air Quality Management District (BAAQMD) observed the ARB evaluation test
and took grab samples of the exhaust gases being vented to the atmosphere. The BAAQMD
Summary of Results is presented in Table 3-5.
-29'
-------
3. SUNOH1O PCBX Evaluation Test Conducted at Maxwell Laboratory's San Diego Facility
The previous two tests addressed'the effectiveness of the emission control device consisting of
two 55-gallon drums arranged in series with an oil mist eliminator and an activated charcoal
canister. For the San Diego Test, SUNOHIO modified the configuration by eliminating
one of the drums and the oil mist eliminator. Samples of uncontrolled and controlled
emissions across the system were taken during the test. In addition to the condenser/carbon
control system, a combustion control system was also tested.
The condenser/carbon and combustion systems were plumbed off of the vacuum degasser's
vent pipe with a "tee" valve. To divert the vapors from one control system to the other
merely required the turning of the valve.
The San Diego test strategy remained-the same as that applied to the previous tests. However,
the test methods differed slightly. Carbon, rather than Tenax, was used as the adsorbent
media to collect hydrocarbons. Combustion gases generated by the PCBX oil furnace were
analyzed by continuous monitors for Q^, CC^, CO, THC, NQX, and SC^ in addition to being
samples with carbon tubes and bags. Florisil samples were not taken because PCS was not
detected in the vent gas at the other two tests. A summary of results is presented in Table 3-6;
Flow rates from the condenser/carbon control system were measured with a Roots meter.
Flow rates from the furnace exhaust duct were calculated from fuel oil analytical data.
Results and Discussion
1. Results From the Test Conducted at Chevron's USA El Segundo Refinery
Results of analyses performed on samples taken from the PCBX process during the ARB
evaluation test conducted at Chevron's USA El Segundo refinery are summarized in Table 3-2.
Benzene and Cg-C^ hydrocarbons were the predominant components measured in the gases
vented directly out of the vacuum degasser during the treatment of transformer oil. Benzene
concentrations ranged from 400 ppm to 7,000 ppm and hydrocarbon concentrations from 7
ppm to 1,600 ppm. TRW's analytical results for samples taken at the same location as the
ARB, Table 3-4, show benzene concentrations exceeding 5,000 ppm and Cc hydrocarbon
concentrations as being 1,400 ppm. TRW results for benzene are in the range of concentra-
tions determined by the ARB and the hydrocarbon concentrations, while not directly
comparable to the ARB results, are probably not inconsistent. SCAQMD did not take samples
from the vacuum degasser outlet
-30-
-------
PCBs were not detected above the detection limit of the analytical method for any of the ARB
samples taken. TRW also sampled for and could not detect PCBs above the limit of detection
for their analytical method (1 ppm). '
As discussed previously, SCAQMD took samples before and after the activated carbon adsorber
and analyzed those samples for PCBs, furans, and dioxins. A summary of test results is pre-
sented in Table 3-3. Because they sampled at a different location and the emphasis of their
analytical work was different, no direct comparison can be made between the SCAQMD and
ARB test results. Test results indicate that furans and dioxins are not present above the
detectable limits, less than 4—30 parts per trillion. However, in two samples taken from the
centrifuge vent at the inlet and outlet of the carbon cannister, detectable amounts of PCSs
were measured; 1.7 (10*3) ppm at the inlet and 8.7 (1CT6) ppm at the outlet
2. Results From the Test Conducted at PG&E's Union City Facility
Benzene and aliphatic hydrocarbons were the major components measured in the ARB samples
taken from the degasser vent during the treatment of PCB-contaminated oils stored in a tank at
a PG&E facility located in Union City. Referring to Table 3-2, the range of concentrations
determined for benzene and Cg-C12 hydrocarbons was 50 to 950 ppm and 1,900 to 2,700
ppm, respectively.
Benzene and aliphatic hydrocarbons were also the significant components in the treated
vacuum degasser vent gas as sampled at the outlet from the control system's activated charcoal
adsorber. The range of benzene concentrations was 600-700 ppm and the range of Cg-C^
hydrocarbons was 1,400 to 1,800 ppm. BAAQMD also took samples at the charcoal adsorber
outlet and test results, presented in Table 3-5, showed: a comparable benzene concentration
"of 840 ppm (average); a comparable DC-C^2 hydrocarbon concentration of 1,500 ppm; and
total organic and nonmethane organic compound concentrations of 2,600 ppm (average) and
2,400 ppm (average), respectively.
The control system's charcoal canister was the same one used at the El Segundo test. Results
of a comparison between the concentrations determined for the inlet and outlet of the control
system are indicative of carbon adsorption breakthrough. The reduction of both benzene and
aliphatic hydrocarbons was approximately 30 percent across the control system.
PCBs were not detected above the detection limit of the analytical method for any of the ARB
samples taken. BAAQMD did not analyze their samples for PCS.
-31 -
-------
TABLE 3-1
SUMMARY OF SAMPLING METHODS
PARAMETER TO
BE MEASURED
Hydrocarbons
PCS
PCS in oil
to be treated
Ambient Air
Flow Rate
SAMPLING
METHOD
Tenax
.Bag
Florisil
Liquid
Grab
Bag
Roots Meter
ANALYTICAL
METHOD
GC/ECD
~~ GC/FID
GC/FID
GC/MS
GC/ECD
GC/ECD
GC/FID
GC/MS
Not Applicable
•32-
-------
TABLE 3-2
SUMMARY OF ARB TEST RESULTS
FACILITY
Chevron
USA.
El Segundo
Refinery
PG&E,
Union City
SAMPLING
LOCATION
Inlet to
the Control
System
PCBX Oil
Treatment
Line
Ambient
Air
inlet to
the Control
System
Outlet from
the Control
System
Ambient
Air
PCSX Oil
Tr raiment
Line
ANALYTICAL
METHOD
GC/MS;
GC/FIO;
GC/ECO
GC/ECO
GC/MS;
GC/FIO;
GC/EC
GC/MS;
GC/FIO
GC/ECO
GC/MS;
GC/FIO
GC/ECO
GC/MS;
GC/FID
GC/ECD
SAMPLING
METHOD
Tenax
Bag*
Bag*
F!orisi|3
Florist|3
Grab
1
Bag
Bag
Bag
Florist
Bag
Bag
Florili|3
Bag
Grab
COMPOUND
DETECTED
Benzene
Other Aroma tics
C8-C12
hydrocarbons
Trtchloro benzene
Benzene
Other Aromatics
Cg-Ci2
hydrocarbons
Trtehlorohpnzene
Benzene
Other Aromatics
C8-C12
hydrocarbons
Triehlorohertzenff
PCS
PCS
PCS in untreated
oil
PCS in treated oil,
end of ARR't tMfS
Benzene
Other Aromatici
ca-ci2
hydrocarbons
Trichloro benzene
Benzene
Other Aromatici
Alcohols
C6-C,2
hyrlrncurfrnn*
Rpp7«ne
Other Aromatic*
Alcohols
C6-C12
hvrirnrarhnn*
PCS
Benzene
Other Aromatics
Alcohols
C6-C12
hydrocarbons
Benzene
Other Aromatics
Alcohols
C6-C12
hvrinorarhon*
pea
Benzene
Other Aromatics
Alcohols
Ce-C12
hvrJroearhon*
PCS in un-
treated oil
PCS in treated oil
end of ARB's te«5
CONCENTRATION
(ppm)
400
10
1.100
trace, not determined
7,000
90
1300
Trarp nnf ritffMrminpd
350
3
700
rrnrff nor rieTBrmined
N.D.I
N D 2
1,330ppm
1,030 ppm (w/w)
no detectable
compounds
under identical
analytical
conditions
50
4
2
2,700
950
30
10
1500
N.0.2
600
30
60
1,400
700
20
10
1300
N D 2
no compounds were
detected under
identical
analytical
ennriition*
188 ppm (w/w)
4S2pomWw!
1 N.D. • Not detected. Limit of detection with respect to volume sampled; 10 H)/m3.
2 N.D. » Not detected. Limit of detection with respect to volume sampled; 20 /Jg/m3.
3 Detection limit of the analytical method; 0.002 pg/m^.
4 Samples taken simultaneously.
5 The conclusion of the ARB evaluation test did not correspond to thecompletionof the oil treatment process.
-33-
-------
TABLE 3-3
SUMMARY OF SCAQMD TEST RESULTS
(Chevron USA, El Segundo Refinery)
Sampling
Location
Centrifuge
Vent
Vacuum
Oegasser
Vent
Ambient .
Air
Sampling
Method
Ethylene
Glycol in an
impinger tfain
Ethylene
Glycol in an
impinger train
Bag
Analytical
Method
GC/EC;
GC/MS
GC/EC;
GC/MS
GC/EC;
: GC/MS |
Compound
Polvchlorobionenvl
Furans1
Dioxins1
Polychlorobipnenvl
Furans1
Dioxins1
Viny (chloride
Concentrations Across
The Activated Carbon Adsorber
(com)
Inlet Outlet
0.0017 8.66(10~S1
<0.0000357 <8.33(10-6)
<0.0000342 <7.99(10-s)
<3.92(10~6) <3.64(1Q-S)
<3.77(1Q-*} <3.50(10~6)
<3.62(10-6) <3.36(1Q-S)
<0.01
Below limit of detectabilrty
-34-
-------
TABLE 3-4
SUMMARY OF TRW INCORPORATED TEST RESULTS
(Chevron USA, El Segundo Refinery)
SAMPLING
LOCATION
Inlet to
the Control
System
SAMPLING
METHOD
Bag1
Tenax1
Florisil1
NUMBER
OF
SAMPLES
1
2
1
ANALYTICAL
METHODS
GC/MS
GC/MS
GC/MS
COMPOUND
DETECTED
Methane
Ethane
C3 Hydrocarbons
C4 Hydrocarbons
C5 Hydrocarbons
Benzene
Benzene
CONCENTRATION
(PPM)
74
22
52
77
1,400
<5.000a
<3.QO
-------
TABLE 3-5
SUMMARY OF BAAQMD TEST RESULTS
(PG&E, Union City)
SAMPLING
LOCATION
Outlet of
Control
Device .
SAMPLING
METHOD
Evacuated
Cylinder with
Silica Gel
COMPOUNDS OR
CLASS OF
COMPOUNDS
Total Organic
Carbon
Nonmethane
Organic Carbon
Benzene
CONCENTRATIONS (PPM)
Run A
2,768
2,605
938
Run B
. 2,473
2,203
743
Average
2,621
2,454
841
• 36-
-------
3. Results From the Test Conducted at the Maxwell Laboratory in San Diego
A summary of test results for the ARB evaluation test conducted in San Diego is presented in
Table 3-6. Emissions were measured during two distinct phases of the PC8X process: (i)
degassing and (ii) dechlorination. Concentration values in Table VII footnoted with a (1) were
measured during the degassing phase and those footnoted with a (2) were measured during the
dechlorination phase.
Benzene concentrations in the uncontrolled vacuum degasser emissions, measured at the
common inlet to both control systems, ranged-from 18 ppm to 220 ppm. Toluene was also
measured at concentrations ranging from 0.1 ppm to 290 ppm. Concentration values
proceeded by a less than sign «) implies that the corresponding compound was below the
limit of detectability.
At the outlet from the condenser/carbon control systems, ro°,asured benzene concentrations
ranged from 10 ppm to 23 ppm and toluene from 0.1 ppm to 26 ppm. Benzene and toluene
were not measured above the 0.1 ppm limit of detection in the exhaust gases from the com-
bustion control device.
Products of combustion in the furnace exhaust gas were monitored when the furnace was fired
on fuel oil only and when it was fired with a mixture of fuel oil and vacuum degasser vapors.
When combusting fuel oil only, the following compounds were measured: 75 ppm 502, ^2
ppm NOx, 9.3 ppm THC, 133 ppm CO, 10.2 percent C02 and 9.7 percent ©2- When vacuum
degasser vapors were added for combustion in the furnace, measured concentrations were: 78
ppm SO2, 92 ppm NOx, 24 ppm THC, 200 ppm CO, 10.2 percent C02, 9.1 percent 02,
0.1 ppm benzene and 0.1 ppm toluene.
Samples taken before and after the control devices were speciated to determine the types of
compounds present in the emissions. The results are tabulated in Table 3-7. For comparison
purposes, a sample of ambient air was taken. The compounds detected in the ambient sample
were all at the part per billion (ppb) level. The magnitude of the other concentration values
measured in the uncontrolled and controlled emissions were at the part per million level.
Columns 2, 3, and 6 present the results of speciating the samples of the uncontrolled emissions
from the vacuum degasser and column 4, 5, and 7, the controlled emissions.
-37-
-------
TABLE 3-6
SUMMARY OF ARB TEST RESULTS
(Maxwell Laboratory, San Diego)
FACILITY
Maxwell
Laboratory
San Diego
•
SAMPLING
LOCATION
Inlet to
Control
Systems
Outlet from
Condenser/Carbon
Control System
Outlet from
Combustion
System
F-Tank
Lazer Tank
PCS in Treated
oil. end of
ARB Test*
ANALYTICAL
METHOD
GC
GC
GC
GC
GC
U.V.
Chemilumin-
o*pnrp
FID
IMDIR
NOIR
Paramaanetie
U.V.
Chemilumin-
p<;pnrp
FID
NDIR
NOIH
Paramagnetic
GC/EC
GC/EC
GC/EC
SAMPLE
COLLECTION
MEDIA
Carbon
Baa
Carbon
Baa
Carbon
Continuous
Analvzer
a
H
tt
"
"
It
If
It
It
"
tt
Grab
Grab
Grab
COMPOUND
DETECTED
Benzene
Toluene
Benzene
Benzene
Toluene
- rfieozene
Benzene
Toluene
so2
NOX
THC
CO
CO-
°7~
so2
N0x
THC
CO
CO
o,2
PCS
PCS
PCS
CONCENTRATION
(PPM)
1Ra
ai*»
?->rP
<0.1a
90°
•7orf>
5(1*
103
15*
9-7"
<0.1a
<0.1a
?fib
20"
-------
TABLE 3-7
Quantitative Analysis of Compound! Present In the On Strum
•Enuring and Leaving the Activated Carbon Cannlttar and Oil Burner
Compound Name
1. Chloroethylene (Vinyl Chloride)
2. Chloromethane (Methyl Chloride)
3. Bromomethane (Methyl Bromide)
4. Dichlorofluoromethana (Freon 12)
5. Dichloromethane (Methylena Chloride)
6. Trlchlorofluoromethane (Freon 1 1 1
7. 1, 1 -Dichloroethylene (Vinylidena Chloride)
B. trani-1. 2-Dicliloroethylone
9. Trichloromethana (Chloroform)
10. Trichlorotrifluoroethane (Freon 113)
11. 1 ,2-Dichloroethane (Ethylene Chloride)
12. 1.1, 1 — Trlchloroethane (Methyl Chloroform)
13. Tetrechloromethane (Carbon Tetrachloride)
14. Bromodichloromethane
16. 1, 2— Dichloropropane
16. Trichloroathylene (TCE)
1 7. Chlorodlbromomethane
IB. 1,2 Oibromoethane (Ethylene Bromide)
19. Tribromornethana (Bromoform)
20. Tetrachloroethylene (PERC)
21. Benzene
1
Ambient
Sample
PPBV
<1.0E+OO
<2OE400
<5.0E-01
6.7E-01
6.9E-O1
3.9E-01
O.OE-01
O.OE+00
<2.0E-02
9.6E-01
<1. OE-01
6.5E400
4.2E-02
1 E<00
2E-02
<1E+01
<7E*01
9E-01
<1E+O1
<1 E+01
<1E*01
>1E«)0
3
Carbon
Cannister
Inlet
' During
Declilor.
Mode
PPMV
<1E+00
<2E+00
<2E-01
3E-02
3E+00
6E-01
OE-01
<1E+00
BE-02
4E-OI
OE-01
>1E-«00
OE-O2
OE-02
• <5E-01
4E-01
OE-02
OE-02
OE-O2
>1E40°
5E+01
4
Carbon
Cannister
Outlet
During
Dechlor.
Mode
PPMV
<1E400
<2E+00
<2E-01
3.E-02
1E+00
OE-02
OE-02
<2E-01
"* 2E-O1
.OE-02
OE-02
OE-02
>1EtOO
2E+01
6
Carbon
Canniitar
Outlet
During
Dechlor.
Mode
PPMV
<1E+00
<2E-02
6E-OU
1E+00
BE-01
OE-02
OE-01
4E-03
4E-02
OE-02
2E-O1
OE-03
OE-03
<5E-01
2E-02
OE-03
OE-O3
OE-03
1E-O1
fl
Burner
Inlet
During
Dechlor.
Mode
PPMV
<1E+00
<2E+00
1E-»OO
1E-O2
OE-02
<8E-01
4E-01
OE-02
OE-02
OE-02
>1E+00
7
Burner
Outlet
During
Dechlor.
Mode
PPMV
<1E«K)
<2E-01
<2E-02
2E-01
3E-O1
3E-03
<1E-02
<1E-01
<1E-03
2E-01
<1E-02
4E-02
<1E-03
<1E-03
<2E-O2
3E-02
<1E-03
<1E-03
<1E-03
6E-02
Note: 1. Less than symbol (<• preceding a concentration value - below the limit of detection.
2. Greater than symbol C>) preceding a concentration value *• minimum value.
-------
4. Discussion of Results
During the El Segundo test, sample was drawn from the inlet to the vacuum degasser's control
system and introduced simultaneously into two sample bags. However, laboratory results
indicate that one of the bags may have leaked, introducing unwanted dilution air into the
sample via the bag material itself or between connectors. The samples in question are identified
in Table 3-2 by a "d" footnote: "BAGd." Analytical results are tabulated in Table 3-8.
Comparing these results, a 95 percent difference in benzene concentrations and a 96 percent
difference in the "other aromatics" concentrations were observed between the two samples. A
similar comparison between aliphatic hydrocarbons indicates a 56 percent difference. This
anomaly points strongly towards the supposition that dilution air leaked into one of the
sampling bags. Therefore, the sample bag having lower concentration values will be reported
but not factored into the ARB's evaluation process.
Benzene, "other aromatics," and aliphatic hydrocarbon concentrations measured in the
t
vacuum degasser vent gas versus process lapse time are plotted for each test. This measurement
site represents the uncontrolled emissions from the vacuum degasser. Figure 3-2 is a plot of
the El Segundo test data, Figure 3-3 the Union City test data, and Figure 3-4 the San Diego
test data. In all figures, benzene concentrations appear to increase as the treatment of PCS
contaminated oil proceeds to completion. This dependence appears strongest for the treat-
ment process conducted at El Segundo than at Union City and significantly less so at the San
Diego test. Although lower in magnitude, the same trend was observed for the concentrations
of "other aromatics" in the vacuum degasser's vent gas. The dependence on processing time
was not evident for the aliphatic hydrocarbons. A gradual growth of aliphatic hydrocarbon
concentrations with time was observed at the El Segundo test while a decay was evident for
the Union City test. Concentration data for "other aromatics" in the vacuum degasser's vent
gas. The dependence on processing time was not evident for the aliphatic hydrocarbons. A
gradual growth of aliphatic hydrocarbon concentrations with time was observed at the El
Segundo test while a decay was evident for the Union City test. Concentration data for "other
aromatics" and aliphatic hydrocarbon were not available from the San Diego test.
The cause behind the observed growth of benzene and "other aromatics" concentrations is
merely speculative at this time, per telephone conversations with SUNOHIO personnel. Sub-
sequent measurements and observations made while treating oils in other states lead them to
believe that this phenomena may be dependent upon the type of oil being treated rather than
inherent with the PCBX process itself J4' Resolving the cause of the concentration growth
would be of interest in future evaluation studies.
.40-
-------
TABLE 3-8
Concentration of Benzene, "Other Aromatics," and Aliphatic?
Measured at the Inlet to the Vacuum Degasser's Control System
With Respect to Process Lapse Time
FACILITY
Chevron USA,
El Segundo
Refinery
PG&E
Union City
Maxwell
Labs,
San Diego
v.
SAMPLING
LOCATION
Inlet to the
Vacuum Degasser
Control System
In let to the
Vacuum Degasser
Control System
•
Inlet To
Control Systems
COMPOUND
Benzene
OTher
Aromatics
Aliphatics
Benzene
Other
Aromatics
Aliphatics
Benzene
MEAS.
CONC.
(ppm)
400
7.0008
350* -b
10
90"
30
1,100
1,600"
7009 .b
SO
950
4
30
2,700
1.900
18
31
50
220
TIME
Start of
Treatment
Process
1058
"
it
it
tt
ii
-•*•
••
if
0945
"
»t
n
"
"
1625
"
»i
n
Time
Sample
Taken
1144
1655
1655
1144
1655
1655
1144
1655
1655
0945
1256
0945
1256
0945
1256
1640
2150
2226
0012
Process
Lapse
Time (min.)
46
357
357
46
357
357
46
357
357
Start Up
191
Start Up
191
Start Up
191
15
325
363
467
3 Simultaneous bag sample.
b Remit not used in the ARB evaluation process because of a suspected leak into this sample bag.
-41 -
-------
Low vacuum degasser flow rates offset the effect of high concentrations on the pollution mass
emission rate. Flow rates measured from the vacuum degasser were very low, especially those
measured at the San Diego test, resulting in emission mass rates on the order of 0.1 pounds per
hour and less. The uncontrolled and controlled mass rates based on the ARB data are
presented in Table 3-9. Benzene, toluene, and the aliphatic hydrocarbons were the only
components in the vacuum degasser emissions that were present in amounts significant to a
mass emission calculation.
The furnace's exhaust gas flow rate of 267.9 cubic feet per minute was estimated from
*
calculations based on fuel analysis and fuel flow .rates. However, benzene and toluene-were
not detected in the furnace exhaust gas, resulting in zero emissions from the furnace for these
two compounds.
At the ARB's request, SUNOHIO left.the activated carbon canister in place for the first two
tests: the SCAQMD staff tested for emissions at the inlet and outlet of the carbon canister at
the first test and the ARB staff tested at the inlet and outlet of the enitre control device during
the second test. The PC8X system treated a total of 48,750 gallons of oil at the Chevron USA
refinery before being moved to Union City. Depending upon the process flow rate, this would
correspond to a continuous treatment time of 54 to 98 hours, or approximately 2 to 4 days.
Based on the SCAQMD's El Segundo test results, adsorber effeciency with respect to PCS
was 99 percent during the first 3 hours of oil treatment. Assuming this level of control to be
applicable for other aromatics and aliphatics as well, carbon canister breakthrough had not
occurred. However, analysis of the ARB's subsequent PG&E data indicates that breakthrough
occurred prior to this test. Therefore, sorbent saturation was reached sometime between the
conclusion of the SCAQMD's El Segundo test and the start of the ARB's Union City test, a
span of 51 to 94 hours of uncertainty. San Diego test data is also indicative that the lifetime
of the carbon cartridge is limited to a few hours of PCBX processing time. Upon review of
preliminary ARB test data and subsequent tests of their own, SUNOHIO modified its operat-
ing schedule to replace the carbon canister after 48 hours of oil treatment processing time.
However, more tests are recommended to determine the actual lifetime of the activated carbon
for this particular application.
-42-
-------
TABLE 3-9
MASS RATE OF BENZENE. TOLUENE. & ALIPHATIC HYDROCARBONS
FROM THE VACUUM OEGASSER
(AR8 Data Only)
FACILITY
Chevron
USA
PG&E
Maxwell
Labi
SAMPLE
LOCATION
Inlet to
Control
System
Inlet to
Control
System
Outlet from
Control
System
Inlet to
Condenser/
Carbon
Control
System
Outlet from
Condenser/
Carbon
Control
System
Outlet from
Combustion
System
FLOW RATE
(cfm)
.645*
1.55°
1.550
•
0.03*
0.03°
2675°
COMPOUND
Benzene1
Aliphatic
Benzene?
Aliphatic
Hydrocarbons?
Benzene?
Aliphatic
Hydrocarbons?
Benzene
Toluene
Benzene
Toluene
Benzene
Toluene
CONCENTRATION
400
7,000
700
1.600
50
950
2.700
1.600
600
700
1,400
* 1.800
18
31
220
50
<0.1
90
290
10
15
23
20
<3.1
(M.W.) Q (1.58 x 10~7 - Ibs^hr.)
' (dry) scfm
t> (wet) cfm
c Calculated volumetric flow rate. Calculations in APPENDIX V of this report.
•43-
-------
FIGURE 3-2
.Chevron U.S.A., El Segundo Refinery
Concentrations of Benzene, Other Aromatic* & Aliphatic
Measured at the Inlet to the Control Systems
vs.
Process Lapse Time
Q.
a
§ 80
o
r»
o
c
o
o
w
•O
I
.2
7000 ._
6000
5000 ..
4000 ._
| 60 .9- 3000
O <
o
c
I 40
I
o
o
as 2000 . _
o
o
20
8 1000 - -
A
X
Benzene cone.
Other Aromatics cone.
_ Aliphatic H.C. cone.
-i-
1 f—
75 150 225 300
Process Lapse Time (minutes)
375
-44-
-------
FIGURE 3-3
PG&E, Union City
Concentrations of Banzan*. Otfiw Aromatic* & Aliphatic
Maacurad at th« lnl«t to tha Control Systarna
vs.
Procaa Lapsa Time
7000 4.
Benzene cone.
Other Aromatics cone.
Aliphatic H.C. cone.
150 225 300
Process Lapse Time (minutes)
375
-45'
-------
FIGURE 3-4
Maxwell Laboratory, San Diego
Benzene Concentration Measured at the Inlet to the
Control System vs. Process Lapse Time
Q.
_a
a
c
a
N
C
4>
a
c
o
250 ..
200 . .
150 ..
r 100
3
J
50 ..
75 150 225 300
Process Lapse Time (minutes)
375
400
• 46-
-------
BIBLIOGRAPHY
1. "The Reclamation of PCB-Contaminated Transformer Oils by the SUNOHIO PCBX Process,
Interim Report;" Ken Chen, P.E.; The Tennessee Valley Authority, Office of Power, Division
of Energy Demonstration and Technology; March 1982. «•'
2. Personal Communication with Mr. Hubert Wilson, South Coast Air Quality Management
District
.')
3. "Measurement of PCS Emissions from Combustion-Sources;" Levins, P.L., et al, Arthur D.
Little, Inc.; prepared for Industrial Environmental Research Lab, Research Triangle Park, N.C.,
EPA 600/7-79-047; February 1979.
4. Personal communication with Mr. James Kozak, SUNOHIO.
-47
-------
SECTION 4
COMMERCIAL DEMONSTRATION OF WET AIR OXIDATION
OF HAZARDOUS WASTES
SUMMARY
Wet Air Oxidation by Zimpro, Inc., is a process which has been used to oxidize dissolved or
suspended organic substances at elevated temperature and pressures. The process is thermally
self-sustaining with relatively low organic feed concentrations and is, therefore, most useful for
wastes which are too dilute to incinerate economically yet too toxic to treat biologically.
The purpose of this project was to demonstrate wet air oxidation of toxic and hazardous
wastes at a full-size installation which was located at Casmalia Resources, a commercial waste
treater in California. In the operation of the full-scale Zimpro Wet Air Oxidation unit, wastes
selected from classified groups of organic wastes were detoxified. These classified groups were:
phenolic wastes, organic sulfur wastes, general organic wastes, cyanide wastes, pesticide wastes, and
solvent still bottoms wastes. This section contains detailed evaluation of these six classified wastes'
treatment, the effectiveness of the wet air oxidation unit, and sample analysis of feed and effluent.
Oxidation of a petroleum refining spent caustic waste resulted in 99.77 percent total phenol
reductions, 94.0 percent organic sulfur reduction, and 89.3 percent chemical oxygen demand
[COD] reduction. Oxidation of spent caustic wastewater from a petroleum refinery resulted in
sulfide reduction of >99.7 percent, phenol reduction of 98.8 percent, and COD reduction of 81.3
percent. Very effective treatment of the general organic waste was obtained by wet air oxidation.
Oxidation resulted in 96.7 percent COD reduction. However, an increase in soluble chloride con-
centration in the oxidized effluent was observed following wet air oxidation. In case of cyanide
experimentation, 99.7 percent reduction of cyanide and 88.8 percent reduction in COD were
observed upon completion of wet air oxidation. High levels of destruction—98.0 to greater than
99.8 percent-were observed for the full-scale wet air oxidation of four pesticides. COD and BOD
removals of 72.3 and 73.6 percent were observed in a full-scale wet air oxidation demonstration of a
solvent still bottoms type wastewater.
-43-
-------
INTRODUCTION
Zimpro, Inc., has been granted an EPA-funded contract through the State of California,
Department of Health Services, for the purpose of demonstrating wet air oxidation of toxic and
hazardous wastes at a full-scale installation located at Casmalia Resources, Inc., a secure chemical
waste landfill in Santa Barbara County, California. This report includes operating and performance
data obtained during demonstration of wet air oxidation of six classified groups of wastes.
DESCRIPTION OF WET AIR OXIDATION PROCESS
The Zimpro Wet Air Oxidation unit for this demonstration processed aqueous wastes at a
designed reactor temperature of 550° F, a designed reactor pressure of 2,000 psig a liquid waste
flow rate of 10 GPM, and a compressed air rate of approximately 190 SCFM. In the wet oxidation
process, liquid waste, exiting from a high pressure pump, is combined with compressed air and
directed through the cold, heat-up side of the heat exchanger. The incoming waste-air mixture exits
from the heat-up side of the heat exchanger and enters the reactor where exothermic reactions
increase the temperature of the mixture to a desired value. The waste-air mixture exits the reactor
and enters the hot, cool-down side of the heat exchanger and, after passage through the system
pressure control valves, is directed to the separator. In the separator, the spent process vapors
(noncondensible gases) are separated from the oxidized liquid phase and are directed into a two-
stage water scrubber-carbon bed adsorber, vapor treatment system.
In the wet air oxidation process, organic substances can be completely oxidized to yield highly
•
oxygenated products and water. For example, organic carbon-hydrogen compounds can be
oxidized to carbon dioxide and water, while reduced organic sulfur compounds (sulfides,
mercaptans, etc.) and inorganic sulfides are easily oxidized to inorganic sulfate, usually present in
the oxidized liquor as sulfuric acid. Inorganic cyanides and organic cyanides (nitrites) are easily
oxidized to carbon dioxide, ammonia, or molecular nitrogen. It should be noted that oxides of
nitrogen such as NO or N02 are not formed in wet air oxidation.
When incomplete oxidation of organic substances occurs, the easily oxidized reduced sulfur
and cyanides are usually still oxidized to sulfate and carbon dioxide-ammonia provided a sufficient
degree of oxidation is accomplished. However, incomplete oxidation of other organic compound
results in the formation of low molecular weight compounds such as acetaldehyde, acetone, and
acetic acid. These low molecular weight compounds are volatile and are distributed between the
process off-gas phase and the oxidized liquid phase. The concentration of these low molecular
• 49-
-------
weight compounds (measured as total hydrocarbons (THC) expressed as methane) in the process of
off-gas is dependent on their concentration in the oxidized liquid phase, which is determined by the
degree of oxidation accomplished, the waste being oxidized, and the influent organic concentration
of the waste.
A. PHENOLIC AND ORGANIC SULFUR WASTE CLASSES
Summary and Conclusion
Wet air oxidation of phenolic and organic sulfur classes of waste has been demonstrated
at the Casmalia Resources, Inc., full-scale wet air oxidation installation. Oxidation of a
petroleum refining spent caustic waste at a process temperature of 515°F (268" C) and a
nominal residence time of 113 minutes resulted in 99.77 percent total phenols reduction, 94.0
percent organic sulfur reduction, and 89.3 percent chemical oxygen demand (COD) reduction.
Gas chromatographic/mass spectroscopic (GC-MS) analysis identified acetic acid, benzoic acid,
and several sulfide derivatives as the major components present in the effluent oxidized waste.
Analysis of treated process off-gases indicated a total hydrocarbon (THC) concentration of
only 84.5 ppm (expressed as methane).
Therefore, performance data obtained during full-scale oxidation of the spent caustic
waste demonstrates wet air oxidation to be effective for treatment of wastewaters in the
— phenolic and organic sulfur classes of waste.
Demonstration Program
1. Operating Conditions
The wet air oxidation demonstration of petroleum refining spent caustic waste was
performed during an eight-hour period of operation at steady state conditions on May 19,
1983. During the demonstration period, the wet air oxidation unit was operated at an
average reactor temperature of 515° F (268°C), a compressed air rate of 190 SCFM, and a
reactor pressure of 1,610 psig. Waste was processed at a liquid flow rate of 5.3 GPM
corresponding to a nominal residence time of 113 minutes. Residual oxygen concentra-
tions in the process off-gas averaged 3.7 percent during the demonstration period.
•50-
-------
2. Sampling Procedures
Sampling procedures used during the wet air oxidation demonstration period were
performed according to the sampling program outlined in the Proposal for U. S. EPA
Commercial Scale Demonstration Wet Oxidation of Hazardous Waste.1 Liquid
composites of the influent raw waste and the effluent oxidized waste were obtained by
collecting one-liter samples of each stream on an hourly interval throughout the
demonstration period. Samples were composited in thoroughly cleaned, five-gallon
stainless steel containers.
Upon completion of the demonstration period, split samples of the composite streams
were transferred to one-liter glass jars. Composite feed and effluent samples were then
returned to Zimpro, Inc., for analyses. A one-liter composite sample of each stream was
also retained on-site and made available to the EPA for analysis.
Duplicate off-gas grab samples were obtained from the wet air oxidation unit during the
demonstration period. Off-gases were sampled after treatment but prior to discharge to
•
the atmosphere. One off-gas sample was analyzed on-site by Zimpro personnel for
oxygen, nitrogen, carbon dioxide, carbon monoxide, total hydrocarbons, and methane.
The remaining off-gas sample was retained on-site and made available to the EPA for
analysis.
3. Analytical Procedures
Procedures used by Zimpro, Inc., for analysis of the composite influent raw waste,
composite effluent oxidized waste, and grab off-gas sample are contained in the Quality
Assurance Project Plan^ for the commercial demonstration of wet air oxidation of
hazardous wastes.
Gas chromatographic/mass spectroscopic (GC-MS) analysis of composite raw and
oxidized waste samples were performed using a Finnigan 4021 GC-MS. Aliquots of each
sample were placed in 40 ml vials and sealed with a Teflon-lined septum such that no air
bubbles were present. The 40 ml vial samples were then screened for volatile compounds
-51 -
-------
by direct injection using a 6' x 2 mm 1% SP-1000 on CP-C packed column (1 min. @
50°C, 10°C/min. to 220°C) to chromatograph the volatile components.
Additional volumes of the raw and oxidized waste samples were extracted with
«
metnylene chloride according to EPA Method 625° for the acid and base-neutral
fractions. The raw waste acid fraction extract was concentrated to a final volume of 50
ml while the base-neutral fraction extract was concentrated to 10 ml. Both oxidized
waste extracts were concentrated to a final volume of 2.0 ml. The acid extracts were
then chromatographed on a 6' x 2 mm 1% SP-1240 DA on Supelcoport packed column (2
minutes at 70° C, 8°C/min. to 200° C). Base-neutral extracts were chromatographed on a
6* x 2 mm 3% SP-2250 DB on Supelcoport (4 minutes at 45°, 8°/min. to 270°C).
Results and Discussion
1. Waste Selection
The waste selected for the wet air oxidation demonstration run was obtained from a
petroleum refinery located in California. Previous laboratory screening tests consisting of
treatability and materials of construction testing indicated the waste to be acceptable for
treatment by wet oxidation.
The waste selected is a spent caustic waste generated by various petroleum refining
processes and contains high concentrations of phenolic and organic sulfur compounds.
Since both phenols and organic sulfur compounds are present in the waste, this
demonstration run was used to determine treatment by wet air oxidation of both the
phenolic and organic sulfur classes of wastes.
2. Performance Results
Analyses of the composite feed and effluent samples collected during the demonstration
period are reported in Table I. The composite feed had a chemical oxygen demand
(COD) of 108.1 g/l and a pH of 13.0. Total phenols and organic sulfur concentrations of
the composite feed were 15,510 and 3,010 mg/l, respectively.
Wet air oxidation resulted in very effective treatment of the spent caustic waste. COD of
the composite feed was reduced to 11.6 g/l in the composite effluent representing a COD
•52-
-------
reduction of 89.3 percent. A total phenols reduction of 99.77 percent was achieved with
only 36 mg/l residual total phenols observed in the oxidized effluent. Oxidation resulted
in good conversion of organic sulfur to sulfate sulfur with sulfur data indicating an
organic sulfur reduction of 94.0 percent.
The pH of the spent caustic waste decreased from 13.0 to 8.3 upon oxidation. Similar
pH decreases are typically observed during wet oxidation of wastes containing reduced
sulfur compounds. Oxidation of these materials results in the formation of inorganic
sulfate, normally present as sulfuric acid, resulting in a decreased pH in the.oxidized
effluent.
Soluble chloride analyses indicated an apparent decrease from 1,510 to 550 mg/l
following oxidation. Similar results were observed during laboratory screening of the
spent caustic waste and have been observed during past industrial waste oxidation studies.
This apparent decrease in soluble chloride is likely due to the presence of materials in the
raw waste which exert a positive interference during the chloride analysis. Oxidation of
these materials eliminates the interference and should result in an accurate chloride
analysis for the oxidized waste.
3. GC-MS Analysis
GC-MS analysis was performed on the composite feed and effluent samples obtained
during the petroleum refinery spent caustic waste demonstration run. Compounds
identified during analysis of the volatile, acid, and base-neutral fractions of the feed and
effluent samples are reported in Tables 4-2A to 4-7A.
Major volatile components identified in the composite feed sample include phenol,
methylphenol isomers, and a few sulfides. Volatile components in the effluent sample
were composed primarily of acetic acid. Phenol, methylphenol isomers, and
dimethylphenol isomers were the major components identified in the feed acid fraction
extract. Benzoic acid and methylbenzoic acid isomers were the major components of the
effluent acid fraction extract. Although wet chemical analysis indicated 36 mg/l total
phenols in the effluent composite sample, no phenols were identified in the effluent acid
fraction extract. Several peaks in the effluent acid extract were not identified and may
•53-
-------
be phenolic compounds which are not included in the GC-MS library. Many of the feed
components of the base-neutral fraction extract could not be identified by the GC-MS
library. The effluent base-neutral fraction extract contained several sulfide derivatives,
4. Process Off-Gas Analysis
A grab sample of the wet air oxidation unit process off-gas was collected during the
demonstration period. The sample was collected at a point after treatment but prior to
discharge to the atmosphere. Results of chromatographic analysis of the off-gas sample
are reported in Table 4-8A. Analysis of the off-gas indicated carbon dioxide, oxygen, and
nitrogen concentrations of 11.3, 4.3, and 83.2 percent, respectively. Carbon monoxide
was not detected irrthe off-gas sample.
A total hydrocarbon (THC) concentration of 85.4 ppm (expressed as methane) and a
methane concentration of 12.1 ppm were determined for the process off-gas sample.
Results of off-gas analysis by an on-line continuous THC analyzer (Beckman, Model 400
THC Analyzer) indicated a THC concentration of 80 ppm at the approximate time the
grab off-gas sample was collected.
-54-
-------
REFERENCES
1. Proposal for U. S. EPA Commercial Scale Demonstration Wet Oxidation of Hazardous Waste,
Zimpro, Inc., June 1982.
2. Quality Assurance Project Plan, Commercial Demonstration Wet Air Oxidation of Hazardous
Waste, Casmalia Resources, California, Zimpro, Inc., February 16, 1983.
3. "EPA Method 625," Federal Register, December 3, 1979.
4. Dietrich, J. J., "Casmalia Resources Wet Air Oxidation Unit Screening; Petroleum Refinery
Spent Caustic Wastewater," January 18,1983.
-55-
-------
TABLE 4-1A
WET AIR OXIDATION DEMONSTRATION
PETROLEUM REFINERY SPENT CAUSTIC WASTEWATER
Wet Air Oxidation Conditions:
Oxidation Temperature
Nominal Residence Time
Waste Flow Rate
Compressed Air Flow Rate
Reactor Pressure
Residual Oxygen Concentration
515°F (268°C)
113min.
5.3 GPM
190SCFM
1,610 PSIG
3.7%
SAMPLE DESCRIPTION
COD, g/l
COD Reduction, %
Total Phenols, mg/1
Total Phenols Reduction, %
Total Sulfur, mg/1
Sulfate Sulfur, mg/1
Organic Sulfur, mg/1 *
Organic Sulfur Reduction, %
Sulfide Sulfur, mg/1
PH
Total Solids, g/1
Total Ash, g/l
Volatile Sol ids, g/l
DOC, mg/1
Soluble Chloride, mg/1
Soluble Fluoride, mg/1
COMPOSITE
INFLUENT
RAW
WASTE
108.1
15,510
3,580
570
3,010
1.0
13.0
88.6
57.1
31.5
1,510
4.4
COMPOSITE
EFFLUENT
OXIDIZED
WASTE
11.6
89.3
36
99.77
3,090
2,910
180
94.0
1.0
8.3
59.7
50.2
9.5
3,680
550
1.3
Organic Sulfur - Total Sulfur Minus Sulfate Sulfur.
-------
TABLE 4-2A
CASMALIA SPENT CAUSTIC FEED
VOLATILE REACTION
SCAN NUMBER COMPOUND
55 Methanethiol
147 Ethanediol
349 Dimethyldisulfide
455 ?
575 Phenol
733 Column Bleed
823, 889, 928 Methyl phenol isomers
TABLE 4-3A
CASMALIA SPENT CAUSTIC OXIDIZED
VOLATILE FRACTION
SCAN NUMBER COMPOUND
106 Ethanol
137 Methylene Chloride
285 Acetic acid
523 Prapanedioic acid
733 Column Bleed
• 57-
-------
TABLE 4-4A
CASMALIA SPENT CAUSTIC FEED
ACID FRACTION
SCAN NUMBER COMPOUND
56 Benzenethiol
101 2-Methyl-benzenethiol
131,140,148,171,210 ?
235 Phenol
245, 274 Methylphenol isomers
300, 309, 323 Dimethylphenol isomers
336 Propylphenol
351 4-Ethyl-2-methylphenol
369,389 Dimethylbenzaldehyde
374 Trimethylphenol
TABLE 4-5A
CASMALIA SPENT CAUSTIC
OXIDIZED ACID FRACTION
SCAN NUMBER COMPOUND
36 Solvent
69 Propanoic acid
138 Hydroxybenzaldehyde
162 Butanoicacid
194 3,4-Dimethyl-2,5-furandione
244 Sulfonylbis-methane
392, 410 Methylbenzoic acid ixomers
492 ?
506 Hydroxybenzoic acid
516 Hydroxybenzaldehyde
526, 552, 630, 729 ?
366 Benzoic acid
-58
-------
TABLE 4-6A
CASMALIA SPENT CAUSTIC
FEED BASE/NEUTRAL FRACTION
SCAN NUMBER COMPOUND
32 Solvent
60 ?
223 Diethyldisulfide
267 Dimethyltrisulfide
320 Methylpyridine or benzenamine
353 " Methylbenzenamine
360,392 Dimethylphenol
383 4-Methyl-2-pyridinamine
414 Dimethylphenol and ?
456, 495, 507, 526 ?
694,720 ?
TABLE 4-7A
CASMALIA SPENT CAUSTIC OXIDIZED
BASE/NEUTRAL FRACTION
SCAN NUMBER COMPOUND
33 Solvent
220 Pyridine
297 Benzofuran
328 Benzaldehyde
356 3-Phenyl-2-propenal or 7-Methylbenzo-furan
383 Methylbenzaldehyde
407 Sulfonylbismethane ?
462 Methylbenzofb] thiophene
473 Quinoline or isoquinoline
492 Aliphatic hydrocarbon
507 2(1H)-Quinolinone + T(2H)-Phthalazincne ?
537 Methylsulfonylbenzene
573 [(Methylsulfonyl)methyl] benzene ?
585 4-Methylbenzenesulfonamide
698 9H-Xanthen-9-one
723 4-Methoxybenzo[c] cinnoline
753 1,2-Dihydro-2-phenyl-3H-indazol-3-one
-59-
-------
TABLE 4-8A
WET AIR OXIDATION DEMONSTRATION
PETROLEUM REFINERY SPENT CAUSTIC WASTEWATER
PROCESS OFF-GAS ANALYSIS
Carbon Dioxide
Oxygen
Nitrogen
Carbon Monoxide
Methane
Total Hydrocarbons
11.3%
4.3%
83.2%
Not Detected
12.1 ppm
85.4 ppm as methane
• 60-
-------
B. PETROLEUM REFINERY SPENT CAUSTIC WASTEWATER
Summary
Wet air oxidation at 280°C for 60 minutes has been performed on a sulfide and phenof
containing spent caustic wastewater from a petroleum refinery located in California.
Oxidation resulted in sulfide reduction of >99.7 percent, phenols reduction of 98.8 percent,
and COD reduction of 81X3 percent. Materials of construction studies indicated compatibility
with titanium.
Thus, treatability and materials of construction testing indicate the petroleum refinery
spent caustic wastewater to be an acceptable candidate for treatment in the Casmalia wet air
oxidation system.
Discussion
Treatability
Laboratory autoclave wet air oxidation has been performed on a petroleum refinery spent
caustic wastewater at the Casmalia wet oxidation unit nominal operating conditions of 280°C
for 60 minutes. The waste as received had a high COD (151.0 g/l) and was diluted to 40.0 g/l
COD prior to oxidation. Treatability by wet oxidation was determined by sulfide, phenols,
and COD reductions.
Oxidation resulted in >99.7 percent sulfide sulfur reduction, 98.8 percent total phenols
reduction, and 81.3 percent COD reduction. Sulfide sulfur and total phenols concentrations
were reduced to >1.0 mg/l and 66 mg/l, respectively, upon oxidation. Further reduction in
residual total phenols concentration would likely be achieved by postoxidation treatment with
ozone or hydrogen peroxide.
The raw spent caustic wastewater had a BOD5/COD ratio of 0.49 compared to 0.64 for
the oxidized product indicating a slight increase in biodegradability following oxidation. The
oxidized product would likely be easily biodegradable since highly biodegradable materials
generally have BODg/COD ratios in the range of 0.5 to 0.6.
-61 -
-------
The wastewater pH decreased from 12.6 in the raw waste to 8.7 upon oxidation, likely
due to the conversion of reduced sulfur compounds to sulfuric acid.
An apparent decrease in soluble chloride concentrations from. 2,190 mg/l in the diluted
feed to 470 mg/l in the oxidized product was observed following wet oxidation. Similar
results have been observed with previous industrial waste oxidation studies and are likely due
to a positive chloride interference in the raw wastewater caused by organic or reduced
inorganic materials. Oxidation of these materials should result in a more accurate chloride
analysis for the oxidized product.
Oxidation results and analytical data are reported in Table 4-1B.
Materials of Construction Compatibility
Autoclave corrosion tests were performed at 140°C for 70 hours and 280°C for 90 hours
to determine the effects of partial and complete oxidation on titanium. No localized corrosion
was observed at either condition. A general corrosion rate of 2.5 mils per year was determined
at 280° C. Carbon steel tanks with Ceilcote lining 103 are recommended for on-site storage of
the spent caustic wastewater.
Hazardous Waste Compatibility
A hazardous waste compatibility chart is attached to this report. Materials capable of
hazardous reactivity with the major components of the spent caustic wastewater-sulfides,
phenols, and caustic—are indicated.
Experimental
Oxidation was performed using a titanium laboratory autoclave. The autoclave was
charged with 100 ml diluted feed and air containing oxygen in excess of the sample oxygen
demand, placed in a heater/shaker mechanism, heated to 280°C, and held at temperature for
60 minutes. The autoclave was cooled with tap water immediately after oxidation.
For corrosion testing, as-welded titanium samples were exposed to the diluted waste. A
150 ml sample was charged to a titanium autoclave containing the titanium corrosion samples.
• 62-
-------
The test was then run at 140°C for 70 hours. After completion of this test, a new titanium
coupon was exposed to a new waste sample at 280°C for 90 hours. After test completion,
general corrosion rates were determined and samples were evaluated for localized corrosion.
Oxidation off-gases were analyzed by gas chromatography. Other analyses reported in
Table 4-18 were performed according to Standard Methods.1
REFERENCES
1. Standard Methods for the examination of Water and Wastewater, APHA, AWWA, WPCF,
15th Ed., 1980.
, >.
2. Zimpro Research Notebook No. 2082, p. 8.
•63-
-------
Table 4-1B
PETROLEUM REFINERY SPENT CAUSTIC WASTEWATER
WASTE CODE NO. 5003
SAMPLE DESCRIPTION
Oxidation Temperature, °C
Time at Temperature, min.
COD, g/l
COD Reduction, %
02 Used, g/l
AOD, g/l**
Total Phenols, mg/l
Total Phenols Reduction, %
Total Sulfur, mg/l
Sulfate Sulfur, mg/l
Sulfide Sulfur, mg/l
Sulfide Sulfur Reduction, %
BOD5, mg/l
BODg/COD, g/g/
PH
Soluble Chloride, mg/l
Total Solids, g/l
Total Ash, g/l
Volatile Solids, g/l
Off-gas CH^ ppm
Off-gas THC, ppm
Sample No.
WASTE AS
RECEIVED
—
—
151.0
—
—
—
20,200
—
12,260
1,890
1,440
—
74,600
0.49
12.6
8,280
232.3
201.8
30.5
—
—
5003-3-1-1
2082-8-1
DILUTED*
FEED
—
—
40.0
—
—
—
5,350
—
3,250
500
382
—
19,770
0.49
—
2,190
61.6
53.5
8.1
—
—
OXIDATION
PRODUCT
280
60
7.5
81.3
31.7
39.2
66
98.8
3,430
2,610
<1.0
>99.7
4,830
0.64
8.7
470
60.7
53.7
7.0
5.0
201
5003-3-1-2
2082-8-2
Calculated values based on 265 to 1,000 ml dilution of raw wastewater.
Autoclave Oxygen Demand - 0-, Used +• Residual CCO.
•64-
-------
C. GENERAL ORGANIC WASTE CLASS
INTRODUCTION
An eight-hour wet air oxidation demonstration of a general organic wastewater was
performed at the Casmalia Resources, Inc., full-scale wet oxidation installation on July 28,
1983.
This report contains operating conditions and performance results obtained during the
•
wet air oxidation demonstration. GC-MS analyses of the influent raw waste and effluent
oxidized "waste samples have not been completed, and therefore, GC-MS results are not
included in this report. A detailed interim report discussing the complete general organic
wastewater category wet air oxidation demonstration will be prepared by Zimpro, Inc., when
GC-MS analyses of the composite waste "Samples are completed.
Results and Discussion
The wastewater processed during the wet air oxidation demonstration was generated by a
polyester resin manufacturing process and was assigned to the general organic wastewater
category. The waste was expected to contain high concentrations of propylene glycol and
various mixed organic ethers and esters. Previous laboratory screening tests performed by
- Zimpro, Inc., indicated the waste to be treatable by wet air oxidation and also compatible with
respect to materials of construction.
The general organic wastewater was processed continually during the eight-hour wet air
oxidation demonstration. During the demonstration period, the wet air oxidation unit was
operated at an average reactor temperature of 531° F (277°C), a compressed air flow rate of
190 SCFM and a reactor pressure of 1,515 psig. Waste was processed at an average liquid flow
rate of 5.0 GPM resulting in a nominal residence time of 120 minutes. Residual oxygen
concentrations in the process off-gas averaged 4.1 percent during the demonstration period.
Analyses of the composite feed and effluent samples obtained during the wet air
oxidation demonstration are shown in Table 4-1 C. The data indicates the waste to be relatively
high strength with a chemical oxygen demand (COD) of 76.0 g/l and a dissolved organic
carbon (DOC) concentration of 20,830 mg/l. The raw waste had a pH of 1.9.
-65-
-------
Very effective treatment of the general organic waste was obtained by wet air oxidation.
Oxidation resulted in 96.7 percent COD reduction with the waste COD reduced to 2.5 g/l. A
DOC reduction of 96.7 percent was obtained with the oxidized waste having a DOC concentra-
tion of 685 mg/l.
An increase in soluble chloride concentrations from 212 mg/l for the raw waste to 880
mg/l in,*he oxidized effluent was observed following wet air oxidation. Similar results were
observed du -ing the initial laboratory screening of the waste and may be due to the presence of
chlorinated organic materials in the raw waste which are broken down during oxidation
releasing free chloride.
Results of analysis of a process off-gas grab sample collected during the wet air oxidation
demonstration are shown in Table 4-2C. Analysis of the off-gas sample indicated carbon
dioxide, oxygen, nitrogen, and canbon monoxide concentrations of 12.9, 5.9, 81.2, and 0.3
percent, respectively. Total hydrocarbon (THC) and methane concentrations of 29.1 ppm
(expressed as methane) and 10.0 ppm, respectively, were determined for the process off-gas
sample.
-66-
-------
Table 4-1C
WET AIR OXIDATION DEMONSTRATION
GENERAL ORGANIC WASTE WATER CLASS
Wet Air Oxidation Conditions:
Oxidation Temperature 531°F (277°C)
Nominal Residence Time 120 min.
Waste Flow Rate 5.0 GPM
Compressed Air Flow Rate 190 SCFM
Reactor Pressure 1,515 PSIG
Residual Oxygen Concentration* 4.1%
Composite Influent Composite Influent
Sample Description . Raw Waste Oxidized Waste
COD, g/l 76.0 2.5
COD Reduction, % - 96.7
DOC, mg/l 20,830 685
DOC Reduction, % - 96.7
pH '1.9 ,3.3
Total Solids, g/l 10.0 "3.1
Total Ash, g/l 0.7 1.1
Volatile Solids, g/l 9.3 ' 2.0
~BOD5, mg/l 29,680 880
BODg/COD, g/g 0.39 0.39
Soluble Chloride, mg/l 212 866
Soluble Fluoride, mg/l 0.7 <0.5
• 67
-------
Table 4-2C
WET AIR OXIDATION DEMONSTRATION
GENERAL ORGANIC WASTEWATER CLASS
Process Off-Gas Analysis
Carbon Dioxide 12.9%
Oxygen 5.9%
Nitrogen 81.2%
Carbon Monoxide 0.3%
Methane 10.0 ppm
Total Hydrocarbons . 29.1 ppm as methane
•68-
-------
CYANIDE WASTE CLASS
INTRODUCTION
;
Treatment of cyanide wastewaters by wet air oxidation has been demonstrated at the
Casmalia Resources, Inc., full-scale wet air oxidation installation. The wet air oxidation
demonstration was performed on July 29 and August 18, 1983 during a combined, six-hour
period of steady state operation.
•
Operating conditions and performance results obtained during treatment of the
cyanide wastewater are discussed in this report. GC-MS analyses of composite raw and
oxidized waste samples are currently being performed and results are therefore not included
in this report. A detailed interim report discussing the complete wet air oxidation demonstra-
tion of the cyanide wastewater wiH be prepared by Zimpro, Inc., when GC-MS results are
available. ,
Results and Discussion
The wastewater processed during the wet air oxidation demonstration was a mixture of
cyanide wastes generated by various metal plating processes. Laboratory screening tests
performed by Zimpro, Inc., indicated the individual wastes contained in the wastewater
mixture to be treatable by wet air oxidation and compatible with respect to materials of
construction.
The cyanide waste treatment demonstration consisted of a four-hour period of steady
state operation on July 29, 1983 and a two-hour period of steady state operation on
August 18, 1983. The short process runs obtained during treatment of the cyanide waste was
due to the presence of high concentrations of metal-cyanide complexes in the waste. As the
cyanide waste was processed through the wet air oxidation system, the soluble metal-cyanide
complexes were destroyed, resulting in precipitation of the metals. The precipitated metals
caused scale formation in the oxidation system which required processing of the waste to be
discontinued in order to remove the scale from the system. Therefore, the cyanide waste wet
air oxidation demonstration was limited to two separate periods of operation for a combined
total of six hours treatment at steady state operating conditions.
During the combined demonstration period, the wet air oxidation unit was operated at
an average reactor temperature of 495°F (257°C), a compressed air flow rate of 190 SCFM,
-69-
-------
and a reactor pressure of 1,220 psig. Waste was processed at an average liquid flow rate of
7.5 GPM, resulting in a nominal residence time of 80 minutes. Residual oxygen concentra-
tions in the process off-gas averaged 7.1 percent during the demonstration period.
Analyses of the influent raw waste and effluent oxidized waste samples composited
during the demonstration are reported in Table 4-1 D. The analyses indicate the composite
raw waste to be a typical high strength cyanide waste with a cyanide concentration of 25,390
mg/l, chemical oxygen demand (COD) of 37.4 g/l, and pH of 12.6.
Wet air oxidation resulted in very effective treatment of the cyanide waste. The
cyanide concentration of the raw waste was reduced to 82 mg/l, representing a cyanide
reduction of 99.7 percent. A COD reduction of 88.8 percent and a dissolved organic carbon
(DOC) reduction of 88.4 percent were obtained by wet air oxidation. COD and DOC
concentrations in the composite oxidized waste were 4.2 f/! .and 1,710 mg/l, respectively.
The scale formation which occurred in the oxidation unit during the cyanide
demonstration period is reflected by the total ash data. The composite raw waste had a total
ash concentration of 112.9 g/l compared to only 77.4 g/l for the composite oxidized-waste.
Since the ash is expected to pass through the oxidation unit as inert material, the data
indicates as much as 35 g/l of inert solids were deposited in the oxidation system during
treatment of the cyanide waste.
Results of analysis of a process off -gas grab sample collected during the demonstration
period are shown in Table 4-2D. Analysis of the off-gas sample indicated 1.5 percent carbon
dioxide, 8.5 percent oxygen, and 82.8 percent nitrogen. Carbon monoxide was not detected
in the off-gas sample. A total hydrocarbon (THC) concentration of 61.1 ppm (expressed as
methane) and a methane concentration of 9.0 ppm was determined for the process off-gas
sample.
GC-MS analyses are currently being performed on the composite raw waste and
oxidized waste samples. Results of these analyses will be published by Zimpro, Inc., in a
detailed interim report which will be prepared following completion of GC-MS testing.
- 70-
-------
Table 4-1D
WET AIR OXIDATION DEMONSTRATION
CYANIDE WASTEWATER CLASS
Wet Air Oxidation Conditions:
Oxidation Temperature 495° F (257°C)
Nominal Residence Time 80 min.
Waste Flow Rate 7.5 GPM
Compressed Air Flow Rate 190 SCFM
Reactor Pressure 1,220 PSIG
Residual Oxygen Concentration 7.1%
Composite Influent Composite Influent
Sample Description Raw Waste Oxidized Waste
COD, g/l 37 4 4 2
COD Reduction, % - _ 33 3
Cyanide, mg/l ' 25,390 82
Cyanide Reduction, % _ ' 99.7
DOC, mg/l 14,710 1,710
DOC Reduction, % _ 33,4
PH 12.6 Q.o
Total Solids, g/l 135.3 91.2
Total Ash, g/l ' 112.9 77.4
Volatile Sol ids, g/l 22.4 13.8
BOD5, mg/l _ 603
Soluble Chloride, mg/l _ 773
Soluble Fluoride, mg/l 30 29
-71 -
-------
Table 4-2D
WET AIR OXIDATION DEMONSTRATION
CYANIDE WASTEWATER CLASS
Process Off-Gas Analysis -
Carbon dioxide 1.5%
Oxygen 8.5%
Nitrogen 82.8%
Carbon Monoxide Not Detected
•Methane 9.0 ppm
Total Hydrocarbons 61.1 ppm as methane
•72-
-------
E. PESTICIDE WASTE CLASS
SUMMARY
High levels of destruction-98.0 to greater than 99.8 percent-were observed for the
full-scale wet air oxidation of four pesticides in the Casmalia Resources wet air oxidation unit.
Dinoseb, methoxychlor, carbaryl, and malathion were added to an acidic distillate waste which
had previously been processed in the Casmalia wet air oxidation unit.
An eight-hour wet air oxidation demonstration of a pesticide containing wastewater was
performed at the Casmalia Resources, Inc., full-scale wet air oxidation installation on
March 28, 1984. This report contains operating conditions and performance results obtained
during the wet air oxidation demonstration.
Results and Discussion ' >
Wet air oxidation of four pesticides—dinoseb, methoxychlor, carbaryl, and malathion—
was evaluated in a full-scale demonstration at Casmalia Resources, Casmalia, California, on
March 28, 1984. Since wastewaters containing relatively high concentrations of a variety of
pesticides were not easily available, the above compounds were spiked into an acidic distillate
wastewater which had previously been processed in the Casmalia wet air oxidation unit.
Prior to the full-scale pesticides wet air oxidation demonstration, bench scale autoclave
oxidations of a variety of pesticides had been evaluated. Results of these screening tests, also
performed in the acidic distillate wastewater matrix, are reported in Table 4-1E. Greater than
99 percent destruction was observed for seven pesticides, including the four subsequently
demonstrated in the Casmalia full-scale unit.
The pesticide wastewater was processed continuously during the eight-hour wet air
oxidation demonstration. During the demonstration period, the wet air oxidation unit was
operated at an average reactor temperature of 537° F (280° C), a compressed air flow rate of
190 SCFM, and a reactor pressure of 1,600 psig. Waste was processed at an average liquid flow
rate of 3.3 GPM, resulting in a nominal residence time of 182 minutes. Residual oxygen
•73
-------
concentrations in the process off-gas averaged 2.3 percent during the demonstration period.
Feed and effluent composite samples were prepared by combining hourly grab samples from
each sampling point.
Analyses of the composite feed and effluent samples obtained during the wet air
oxidation demonstration are shown in Table 4-2E. Removals of the four pesticides ranged
from 98.0 to greater than 99.8 percent. Analyses of pesticides in the feed and effluent
composites were by gas and liquid chromatography. Due to an interference in the gas
chromatographic analysis of effluent malathion, this compound was quantified using GC-MS
techniques.
COD, BODg, and DOC removals were quite similar to results obtained during oxidation
of the acidic distillate waste alone.' COD, BODg, and DOC removals of 95.3, 93.8, and 96.1
percent were observed.
r
Results of gas chromatographic analyses of a process off-gas grab sample collected during
the wet air oxidation demonstration are shown in Table 3. Carbon dioxide, oxygen, nitrogen,
and carbon monoxide concentrations of 14.2, 3.5, 79.0, and 0.7 percent, respectively, were
observed. Total hydrocarbon (THC) and methane concentrations of 153 ppm (expressed as
methane) and 61.9 ppm, respectively, were determined for the process off-gas sample.
1 Dietrich. M. J., "Commercial Demonstration of Wet -Air Oxidation of Hazardous Wastes: General Organic Waste Class/
October 17,1983. Zimpro, Inc., Intsrim Report.
-74-
-------
Table 4-1E
•WET AIR OXIDATION DESTRUCTION OF PESTICIDES
ADDED TO ACIDIC DISTILLATE WASTE
Pesticide
D8CP
Malathion
Carbaryl (sevin)
Carbofuran (Furadan)
Dinoseb
Methoxychlor
Linuron (Lorox)
*«Feed,mg/l
42.5
39.5
53.5
46.5
65.0
77.3
87.0
Oxidation
Product, mg/l
0.28
<0.3
<0.1
<0.1
<0.1
<0.1
• <0.1
% Destruction
99.3
<99.2
<99.8
<99.8
<99.8
<99.9
<99.9
Oxidation at 280 760 minutes.
Compounds added to acidic distillate watte which had been diluted 1 to 4.
-75.
-------
Table 4-2E
WET AIR OXIDATION DEMONSTRATION
PESTICIDE WASTE CLASS
Wet Air Oxidation Conditions:
Oxidation Temperature 537°F
Nominal Residence Time 182 min.
Waste Flow Rate 3.3 GPM
Compressed Air Flow Rate 190 SCFM
Reactor Pressure 1,600 PSIG
Residual Oxygen Concentration . 2.3%
Composite Influent Composite Influent
Sample Description Raw Waste Oxidized Waste
Dinoseb, mg/l 37.1 0.19
Dinoseb Reduction, % " - 99.5
Methoxychlor, mg/l 8.84 > - • <0.018
Methoxychlor Reduction, % . - >99.8
Carbaryl, mg/l 30.0 0.59
Carbaryl Reduction, % - 98.0
Malathion, mg/l 93.1 0.13
Malathion Reduction, % — 99.9
COD, g/l 110 5.15
COD Reduction, % - 95.3
BOD5, mg/l 41,000 2,530
BOD5 Reduction, % - 93.8
BODs/COD 0.37 0.49
DOC, mg/l 26,600 1,040
DOC Reduction, % - 96.1
pH 2.0 4.4
Total Solids, mg/l 8,090 740
Total Ash, mg/l 1,110 460
Total Volatile Solids, mg/l 7,980 280
Soluble Chloride, mg/l 384 186
Soluble Fluoride, mg/l <0.5 <0.5
-76-
-------
Table 4-3E
WET AIR OXIDATION DEMONSTRATION
PESTICIDE WASTE CLASS
Process Off-Gas Analysis
Carbon Dioxide 14.2%
Oxygen 3.5%
Nitrogen 79.0%
Carbon Monoxide 0.7%
Methane 61.9 ppm
Total Hydrocarbons . 153 ppm as methane
• 77-
-------
F. SOLVENT STILL BOTTOMS WASTE CLASS
SUMMARY
COD, BOD, and DOC removals of 72.3, 73.6, and 47.4 percent were observed in a full-
scale wet air oxidation demonstration of a solvent still bottoms type wastewater. Relatively
high off-gas total hydrocarbon (THC) levels and rapid exhaustion of the gas train carbon
adsorption bed indicated significant stripping of volatile organics during the oxidation run.
An eight-hour wet air oxidation demonstration of a solvent still bottoms type wastewater
was performed at the Casmalia Resources, Inc., full-scale wet oxidation installation on
March 29, 1984. The demonstration was performed as outlined in agreement OPR-54 between
Zimpro, Inc., and the State of California Department of Health Services (DHS). This report
contains operating conditions and performance results obtained during the wet air oxidation
demonstration.
-78-
-------
Results and Discussion ~~
Wet air oxidation of a solvent still bottoms type waste was evaluated in a full-scale
demonstration at Casmalia Resources, Casmalia, California, on March 29, 1984. The waste-
water was processed continuously during the eight-hour wet air oxidation demonstration.
During the demonstration period, the wet air oxidation unit was operated at an average
reactor temperature of 514° F (268°C), a compressed air flow rate of 190 SCFM, and a reactor
pressure of 1,550 psig. Waste was processed at an average liquid flow rate of 5.1 GPM, resulting
in a nominal residence time of 118 minutes. Residual oxygen concentrations in the process
off-gas averaged 3.8 percent during the demonstration period. Feed and effluent composite
samples were prepared by combining hourly grab samples from each sampling point.
Analyses of the composite feed and effluent samples are reported in Table 4-1 F. COD,
BOD, and DOC reductions of 72.3, 73.6, and 47.4 percent were "Observed.
During the demonstration run, a tendency to strip volatile organics into the gas phase was
apparent. The vapor phase activated carbon adsorption bed was rapidly exhausted and
relatively high off-gas THC levels were observed. This stripping may have led to the low DOC
reduction compared to the COD and BOD reductions. Much of the volatile organic material
stripped from the reactor by the gas stream would have recondensed upon cooling in the
process heat exchanger. Many of these volatile compounds may have been only partially
oxidized because of their short residence time at oxidation temperatures. Partial oxidation
would lead to a decrease in oxygen demand for these organics, but not as great a decrease in
organic carbon content. Organics which were partially oxidized would also be more likely to
recondense because of higher water solubility generally seen for oxygenated organic species.
Soluble chloride analyses for feed and effluent indicated 7,860 and 505 mg/l, respectively.
These data indicate material in the feed causing a positive interference in the chloride analysis.
Oxidation of this material resulted in an apparent decrease in soluble chloride. This behavior is
frequently seen in the wet air oxidation of industrial wastes.
-79
-------
Soluble fluoride increased from 7.5 to 42.7 mg/l upon oxidation, likely due to the
destruction of fluorinated organic compounds. This fluoride content was not observed in
previous bench scale screening of this wastewater. Fluoride levels much higher than this may
not be acceptable in the Casmalia wet air oxidation unit because of corrosive effects on
titanium system components.
Results of gas chromatographic analyses of a process off-gas grab sample collected during
the wet air oxidation demonstration are shown in Table 4-2F. Carbon dioxide, oxygen,
nitrogen, and carbon monoxide concentrations of 11.8, 3.8, 81.2, and nil percent, respectively,
were observed. THC and methane concentrations of 217 ppm {expressed as methane) and 80
ppm, respectively, were determined for the process off-gas sample. However, on-line off-gas
THC measurements indicated increasing THC concentrations throughout the demonstration
run as the vapor phase activated carbon adsorption bed became exhausted.
-80-
-------
Table 4-1F
WET AIR OXIDATION DEMONSTRATION
SOLVENT STILL BOTTOMS WASTE CLASS
Wet Air Oxidation Conditions:
Oxidation Temperature 515° F
Nominal Residence Time 118 min.
Waste Flow Rate 5.1 GPM
Compressed Air Flow Rate 190 SCFM
Reactor Pressure t,550PS|G
Residual Oxygen Concentration 3.8%
Composit Influent Composit Influent
Sample Description Raw Waste Oxidized Waste
COD,g/l ' 166 45.9
COD Reduction, % - 72.3
BOD5, mg/l 83,000 21,900
BOD5 Reduction, % - 73.6
BODs/COD 0.50 0.48
DOC, mg/l 15,200 7,990
DOC Reduction, % - 47.4
pH 12.5 8.2
Total Solids, mg/l 38,600 30,000
Total Ash, mg/l 25,500 19,800
Total Volatile Solids, mg/l 13,100 10,200
Soluble Chloride, mg/l 7,860 505
Soluble Fluoride, mg/l 7.5 42.7
-81
-------
Table 4-2F
WET AIR OXIDATION DEMONSTRATION
SOLVENT STILL BOTTOMS WASTE CLASS
Process Off-Gas Analysis
Carbon dioxide 11.8%
Oxygen 3.8%
Nitrogen 81.2%
Carbon Monoxide Nil
Methane 80 ppm
Total Hydrocarbons -" 217 ppm as methane
•82-
-------
.SECTION 5
AIR RESOURCES BOARD'S EVALUATION TEST
CONDUCTED ON A WET AIR OXIDATION PROCESS
TO TREAT HAZARDOUS WASTES
INTRODUCTION
The California Air Resources Board (ARB) conducted six evaluation tests on a wet air oxida-
tion unit manufactured by Zimpro, Inc. As it is mentioned in Section 4, the unit designed to treat
toxic wastes, is installed and operated at a Class I waste disposal facility managed by Casmalia
Resources and located in Casmalia, California. The test results of the four category wastes,
namely—phenols, su If ides, acid organics, and cyanides—are discussed in this report.
The ARB evaluation tests were initiated in response to a request by Santa Barbara County Air
Pollution Control District for emission data with which to properly establish and assess the wet air
oxidation unit's waste processing capabilities and the effectiveness of the air pollution control
devices.
-.83
-------
Process Description
,1
1. Wet Air Oxidation System
The wet air oxidation process is a method of oxidizing organic matter suspended or dissolved
in water. The principal components of the treatment process and the air pollution control
system are shown in Figure 5-1.
Six types of wastes are scheduled to be treated at the facility:
a. Cyanide wastes.
b. Solvent still bottom wastes.
c. Phenolic wastes.
r
d. Sulfide wastes.
e. Acid-organic wastes.
f. Pesticide wastes.
Each waste type is stored in fixed-roofed tanks located adjacent to the wet air oxidation
building. The maximum delivery rate of waste to the treatment building is ten gallons per
minute (gpm). Typical process flow rates are between five gpm and eight gpm.
Before being fed into the reaction vessel, the pressure in the line carrying raw, untreated liquor
is boosted to approximately 1,800 pounds per square inch gauge (psig) by high pressure
pumps. The system is rated at 2,000 psig. After pressurization, the untreated liquor is
preheated by a heat exchanger that utilizes the high temperature exit liquid from the reaction
vessel. If required, the final process temperature of approximately 500-540° F is attained by
a second heat exchanger that uses hot oil on the shell side. Under normal operating condi-
tions, the second heat exchanger is not utilized because sufficient heat energy is extracted
from the reaction vessel's effluent to obtain the target process temperatures in the first heat
exchanger. However, the second heat exchanger is usually needed for start-up operation.
Two compressors introduce ambient air into the raw liquor process line to provide the
oxygen necessary to carry out the reaction. This amount is near stoichiometric and is
calculated from the chemical oxygen demand (COD) of the waste stream to be treated
• 34-
-------
00
en
DISCHARGE N
NONCONDENSABLES
TANK
SCRUBBER/
SEPARATOR
SCRUBBER
HIGH PRESSURE
PUMP
tP
AMBIENT
AIR
->
COMPRESSOR
PRESSURE
CONTROL
M^A,
HEAT EXCHANGERS
' AMBIEflt— S
AIR '
COMPRESSOR
<>
-ft-
HEAT EXCHANGER
REACTION
CARBON
ADSB.
I
TO
ATMOSPHERE
FIGURE 5-1
Schematic of the Wet-Air
Oxidation Process and the
Air Pollution Control System
-------
The reaction vessel has enough capacity to provide a hold-up time of approximately 60 minutes
at a maximum system flow rate of ten gpm. At the process temperature and pressure ranges of
500—540°F and 1,300—1,800 psig, the oxidation rate of the aqueous waste-air mixture is
almost instantaneous. If the concentration of oxidizable material, or COD, is high enough, a
self sustaining reaction can be maintained in the reaction vessel. After treatment, the temper-
ature of the aqueous waste-air stream is reduced in a heat exchanger and the pressure returned
to atmospheric through a pressure control valve.
The treated waste stream is then separated into a condensible and noncondensible fraction in
the scrubber/separator. The condensible fraction is pumped out of the condensate well to a
water discharge pond. The noncondensible fraction ia diverted through two water scrubbers
and an activated carbon adsorber before being vented to atmosphere.
2» Zimpro/Casmalia Screening Program .to Determine Candidate Waste for Wet Air Oxidation
Treatment
To determine if a particular waste is suitable for treatment by the waste air oxidation process,
it is subjected to an initial screening process: two one-liter samples of the candidate waste are
obtained from the generator and shipped to Zimpro's Wisconsin laboratory. Here, acceptance
or rejection of the waste's suitability for commercial oxidation is determined by oxidation in a
laboratory autoclave. Figure 5-2 outlines Zimpro's general criteria for scrutinizing each candi-
date waste. After completion of Zimpro's laboratory analysis, Casmalia Resources personnel
are informed about the acceptance or rejection of the candidate waste.
If accepted, Zimpro's analytical results are sent to Casmalia Resources and kept in the generator's
file at the disposal site. Subsequent waste shipments received from the generator are sampled
directly from the hauler's truck upon arrival at Casmalia's disposal facility and analyzed at an
on-site laboratory. Casmalia's on-site analysis is compared to Zimpro's laboratory analysis as a
compatibility check. If it passes the on-site analytical check, the waste is transferred to a
designated fixed roof storage tank adjacent to the treatment building.
As the waste is processed and treated by the wet air oxidation unit, samples of the process
stream are taken from specific points in the system and analyzed. Figure 5-3 schematically
follows the waste from the gate through its treatment in the wet air oxidation process and
illustrates where process samples are taken and the type of analysis performed.
-86-
-------
FIGURE 5-2
ZIMPRO's Wisconsin Laboratory Method for Screening
Potential Wastes for Wet-Air Oxidation
(CANDIDATE WASTE)
SHIP TWO LITER
SAMPLE TO
ZIMPRO'S
LABORATORIES
IS
WASTE
SUITABLE FOR
LABORATORY
AUTOCLAVE
XIOATIO
7
WASTE ANALYSIS
COD, TOTAL SOLIDS. ASH, pH,
SOLUBLE FLUORIDE, SPECIFIC
COMPONENET (CYANIDE, PHENOLS,
SULFIDE, CHLORINATED ALIPHATIC
COMPOUNDS, NONHALOGENATED
PESTICIDES).
SAMPLE RETAINED BY Z1MPRO
YES
LABORATORY
. AUTOCLAVE
OXIDATION
280°C, 1 HOUR
)
r
|
/CALCULATE /
AUTOCLAVE /
Mte
OFFGAS ANALYSIS
Oy N2. CO, C02,
TOTAL HYDROCARBON, CH4
OXYGEN DEMAND
WASTE
SUITABLE
FOR PROCESSING
IN FIELD
UNIT
7
OXIDIZED WASTE ANALYSIS
COO. DOC, TOTAL SOUDS, ASH.
pH, SOLUBLE FLUORIDE, SPECIFIC
COMPONENT (CYANIDE, PHENOLS,
SULF1OE, CHLORINATED ALIPHATIC
COMPOUNDS, NONHALOGENATED
PESTICIDES).
f*
I TO SITE FOR
V DEMONSTRATION
^ RUN
•87
-------
FIGURE 5-3
ZIMPRO/CASMAL1A Resources Process Sampling Locations
and Sample Analysis
SCREENED WASTE
/ DELIVER WASTE / -^
/ TO SITE l~
TRANSFER WASTE
TO DESIGNATED
STORAGE TANK
^
/WAST
WET
OX ID>=
. 1
r
ETO / ^
kTION /
r
WET AIR OXIDATION
PROCESS
WASTE ANALYSIS UPON RECEIPT
COO, pH, SPECIFIC COMPONENT
STORAGE TANK DISPLACED
HEADSPACE GAS ANALYSIS GRAB
SAMPLE
THC. CH4
SAMPLE RETAINED ON SITE
ulDUID WASTE DAILY COMPOSITE
ANALYSIS
COO, TOTAL SOLIDS, ASH, pH.
SOLUBLE CHLORIDE, SOLUBLE
FLUORIDE, SPECIFIC COMPONENT |
(CYANIDE, PHENOLS, SULFIDE,
CHLORINATED ALIPHATIC COM-
POUND, NON-HALOGENATED
PESTICIDES)
SAMPLE RETAINED ON SITE
J
SEPARATION OF PROCESS
OFFGAS AND OXIDIZED
WASTE
OXIDIZED WASTE TO
EVAPORATION PONDS FOR
FINAL DISPOSAL
PROCESS OFFGAS
TREATMENT
' EXHAUST OFFGAS )
ATMOSPHER^/
PROCESS OFFGAS ANALYSIS
GRAB SAMPLE
02, N2, CO, CO2, THC, CH4
SAMPLE RETAINED ON SITE
CONTINUOUS
ANALYZER
THC
HYDROCARBON
OXIDIZED WASTE DAILY COM-
POSITE ANALYSIS
COD, DOC. TOTAL. SOLIDS ASH
pH, SOLUBLE CHLORIDE j
SOLUBLE FLUORIDE, SPECIFICS
COMPONENT (CYANIDE. PHENOLS I
SULFIDS. CHLORINATED!
ALIPHATIC COMPOUND, NON-1
HALOGENATED PESTICIDES).
SAMPLE RETAINED ON SITE
I
• 88
-------
Sampling Methods and Locations
Each waste type treated generates uniquely different liquid and gas phase byproducts. Sampl-
ing and monitoring locations are identified in Figure 5-4.
Untreated and treated liquid waste samples were taken in one-pint, glass mason jars and then
placed in a cool, dark storage area. At each of these sampling points, numerically identified as
1 and 2 in Figure 5-4, four samples were taken; three grabbed at different times during the test
and one composite sample at the conclusion of the test.
•
The composite samples were generated by combining hourly grab samples into a common
container and then taking a sample from this container at the end of the test period. Each
composite represented a sample of the process stream entering and leaving the reaction vessel
averaged over time. . .
Inlet and outlet scrubbing water samples, points 6a, 6b, 7a, and 7b, were taken in one-pint,
mason jars and stored with the other liquid samples.
Gaseous samples were taken from points 3, 4, and 5, shown in Figure 5-4. This provided the
information to determine control equipment efficiency. The sampling train used for cyanide
and sulfide/phenol wastes is shown in Figure 5-5. It consists of four modified Greenburg-
Smith impingers arranged in series. The first and last impingers are empty and the second and
third impingers are each filled with 100 milliliters of 0.1 N sodium hydroxide solution. The
potential gas phase organics generated after treatment of the other waste types were captured
in tenax and charcoal tubes. The sampling arrangement is illustrated in Figure 5-6.
Process volumetric flow rates to atmosphere were determined using ARB Methods 1 and 2.
Summary and Discussion of Results
1. Wet Air Oxidation Process
A summary of test results is presented in Table 5-1. Wet air oxidation appears to be an effec-
tive method for reducing the concentration of liquid phase cyanide and phenol compounds.
Cyanide concentrations measured at the inlet and outlet of the weta.fr oxidation process were
•89-
-------
46,300 parts per million by weight (ppm W) and 6.94 ppm W, respectively. Inlet and outlet
phenol concentrations were 18,700 micrograms per milliliter (ng/m\) and 2.35M9/ml. As
shown in Table 5-2, the reduction in concentration for both compounds treated by the wet air
oxidation process was over 99 percent.
However, wet air oxidation of acid-organics was somewhat less consistent. Two sets of inlet-
outlet samples were taken across the process with one indicating a 64 percent reduction of the
initiaJ concentration and the other, a 97 percent reduction. Each value was obtained by
ratioing the GC/FID generated total peak areas for an inlet-outlet sample pair taken across the
wet air oxidation process. When the output sample's GC/FID trace was compared to that for
the inlet sample, some peaks had disappeared, some were noticeably reduced, and in some
instances, new peaks appeared. Because the availability of standards to identify each peak was
limited, the ratio of total peak areas was used to give a relative indication of the wet air oxida-
tion process.
Sulfide reduction across the wet air oxidation process was not measured because of sample
related interferences to the laboratory's analytical method. The nature of the interference was
unknown and could not be biased out of the analytical procedure.
2. Condensible, Noncondensible Separator
With the exception of acid-organics, the amount of noncondensible cyanide, phenols, and
sulfide measured at the separator was at the microgram level. Referring to Table 5-1, the
quantity of CN, phenol, and sulfide captured was 3.23Mg, and 713 M9, respectively. These
values also represent the prescrubbed gas concentrations at the inlet to the scrubber. The total
volume drawn for each sample was 45 cubic feet. A "less than" symbol «) proceeding a
value implies that it is below the detection limit of the analytical method with respect to the
total volume sampled.
Noncondensible acid-organic samples taken at the separator were used to identify and semi-
quantitate the major organic components present in the gas stream prior to entering the
control equipment The results of this work are presented in Table 5-4 and show that for
the treatment of this particular waste, halogenated (bromo-, chloro-) albenes and benzene
appear to be the major compounds in the gaseous effluents from the separator.
-90-
-------
3. Scrubber
The scrubber was effective in removing sulfide from the gas stream but did not afford any
greater control advantage for cyanide and phenols than was achieved by the wet air oxidation
process. The results of scrubber inlet and outlet sample analysis are shown in Table 5-3.
Gas phase phenols were not detected at the scrubber inlet or outlet.
The calculated cyanide concentration at the inlet was very low, 0.0025 pg/l or 0.0034 ppm.
The measured scrubber efficiency for controlling noncondensible sulfide was 93.4 percent.
The scrubber reduced the inlet concentration of 0.555 pg/l, or ppm, to 0.037 M9/I, or ppm.
The outlet cyanide concentration of. 0.0026 pg/l, <0.0025 ppm is comparable to the inlet
concentration and indicates that'the scrubber has no apparent effect on cyanide when this
compound is introduced into the scrubber at such low levels.
It appears that a major portion of the residual phenols and cyanides remaining after the wet air
oxidation process is retained in the separator's liquid phase and pumped to the facility's water
discharge pond. Any liquid phase reactions that may be occurring are unknown. There is a
minimal contribution of cyanide and phenols to the gas phase for control by the scrubber.
However, whether the scrubber would be an effective control device at higher inlet concentra-
tions of cyanide and phenols, as might occur during an upset condition, is yet to be determined.
4. Carbon Bed
. The carbon bed was most effective in controlling the discharge to atmosphere of gas phase
brominated compounds. The inlet concentration of 450 ppm was reduced to 0.18 ppm at the
outlet of the carbon bed, representing a removal efficiency of over 99 percent. Note that these
concentration values are order of magnitude estimates based on (a) the qualitative speciation
data that identified brominated hydrocarbons as the major noncondensible hydrocarbons
present and (b) quantitative analysis that normalized all significant GC peaks to 1,2-dibro-
methane, which was the major identifiable compound for which a standard was available. The
normalization technique was performed because standards were not available for the other
major peaks. A comparison of these values will give a relative efficiency performance of the
carbon adsorber.
The percentage control for cyanide and sulfide was minimal: 17.8 percent and 26.8 percent,
respectively. Phenols were not detected in the gas phase.
-91
-------
©
CO
ro
Liquid i imp let
flow rite
temperature
preiiurt
Slor.9«
Gateoui semplei
temperatures
preisurei
FIGURE 64
Sampling Location!
&
Sa:np|tt Pdiamelere
-------
FIGURE 5-5
Sampling Train for Gas Phase Cyanide. Sulfide. & Phenols •
Silica Gel-
Sample line;
I
I Manometer
I
I \ J (— t rump
Control Box
•93-
-------
FIGURE 5-6
Sampling Train for Gas Phase Organics
Sample Line
Tenax or
Carbon Sampling
Tube
•94
-------
Table 5-1
SUMMARY OF ARB TEST RESULTS
Compound
CN
Phenols
Sulfide
(As Sulfate)
Acid-Organics
Sampling
Location
1
2
3
4
5
6
7
1
2
3
4
5
6
7
3
4
5
1
2
3
4
5
Phase of
Treatment Process
Untreated Liquid Waste
Treated Liquid Wate
Pre-Scrubbed Gas
Scrubbed Gas
Filtered Gas
Scrubber Water
Scrubber Water
Untreated Liquid Waste
Treated Liquid Waste
Pre-Scrubbed Gas
Scrubbed Gas
Filtered Gas
Scrubber Water
Scrubber Water
Pre-Scrubbed Gas
Scrubbed Gas
Filtered Gas
Untreated Liquid Waste
Treated Liquid Waste
Pre-Scrubbed Gas
Scrubbed Gas
Filtered Gas
Concentration
46,300 ppmW(pH- 13.1)
6.94 ppmW (pH - 13.1)
3.37 A
-------
Table 5-2
PERCENT REDUCTION IN CONCENTRATION
ACROSS THE REACTION VESSEL
Compound
CN
•
Phenols
Acid-Organics2
Concentration
Inlet
46,300 ppm V
1 8,700 M9/ml
—
-
Outlet
6.94 ppm V
2.35 ng/ml
—
—
Percent
Reduction1
99.99
99.99
64
97
1 Percent Reduction -
(Inlet) - (Outlet)
(Inlet)
2 Percent reduction obtained by rationing the GC/FID generated total peak areas for an inlet-outlet sample pair taken across the
reaction vessel.
-96-
-------
Table 5-3
PERCENT REDUCTION IN CONCENTRATION
ACROSS THE SCRUBBER AND CARBON BED
Compound
CN
Penols
Sulfide
Acid-Organ ics
Control
Device
Scrubber
Carbon Bed
Scrubber
Carbon Bed
Scrubber
Carbon Bed
Carbon Bed
Concentration
Inlet
3.23 M9
3.37 M9
<30M9
<30Mg
713 M9
47.33 M9
~450 ppm
Outlet
3.37 M9
2.77 M9
<30M9
<20t c
47.33 M9
34.67 M9
— 0.18 ppm
Percent
Reduction
—
17.80
—
—
93.36
26.75
99.96
•97
-------
Table 5-4
NONCONDENSIBLE HYDROCARBONS DETECTED
IN THE GASEOUS EFFLUENTS FROM
THE SEPARATOR
Compound
a Bromochloroethene
CIS &> Trans— Dibromoethenes
Two Dibromochloroethenes
Tribromoethene
Tetrabromoethane
Bromobenzene
Bromochlorobenzenes
Acetone
Two Bromodichloroethenes
a Dibromodichloroethene
Chlorobenzene
A Dichlorobenzene
Dibromobenzenes
a Bromodichlorobenzene
a Dibromochlorobenzene
Tribromochlorobenzene
Dichloromethane
a Dichloroethene
C5C11 Alkanes
Three Aldehydes (C5, C6, and C7)
Benzene
Tolu nee
Ethylbenzene and Xylenes
Styrene
C3 and C4 Alkylbenzenes
C4 Alkenylbenzenes
Nap.thalene
Carbon Disulfide
Isopropanol
Cyanogene Bromide (possibly)
Cyanobutadiene (possibly)
Dimethyl Sulfonyl (possibly)
Sample
M
M
M
M
M
M
M
m
m
m
m
m
m
tr
tr
m
—
tr
m
—
m
tr
—
-
-
—
-
—
—
tr
tr
—
Blank
—
tr
tr
m
-
m
m
M
—
tr
tr
tr
tr
—
—
—
tr
—
m
tr
tr
m
tr
tr
m
tr
m
tr
tr
—
_
tr
•98-
-------
REFERENCES
1. Thermal Destruction of Chlorinated Hydrocarbons with a High-Temperature Fluid-Wai
Reactor by Arthur W. Hornig and E. Matovich, Thagard Research Cooperation, March 30,
1983.
2. Air Resources Board Evaluation Test to Determine Emissions From SUNOHIO's Mobile PCB
Treatment Process by Peter Ouchida, et al., State of California, Air Resources Board, June
1984.
3. Commercial Demonstration of Wet Air Oxidation of Hazardous Wastes, Phenolic and Organic
Sulfur Waste Classes, Interim Report, Casmalia Resources,*Casmalia by Zimpro Environmental
Control Systems, June 27, 1983.
4. Commercial Demonstration of Wet Air Oxidation of Hazardous Wastes, General Organic Waste
Class, Casmalia Resources, Casamalia, CA by M. J. Dietrich, Zimpro Environmental Control
System, October 17, 1983.
5. Commercial Demonstration of Wet Air Oxidation of Hazardous Wastes, Cyanide Waste Class,
Casmalia Resources, Casmalia, CA by M. J. Dietrich, Zimpro Environmental Control System,
October 21, 1983.
6. Commerical Demonstration of Wet Air Oxidation of Hazardous Wastes, Pesticides Waste Class,
Casmalia Resources, Casmalia, CA by P. J. Canney, Zimpro Environmental Control Systems,
JuneS, 1984.
7. Commercial Demonstration of Wet Air Oxidation of Hazardous Wastes, Solvent Still Bottoms
Waste Class, Casmalia Resources, Casmalia, CA by P. J. Canney, Zimpro Environmental
Control Systems, June 15, 1984.
8. Evaluation Test Conducted on a Wet Air Oxidation Unit Used to Treat Waste A + A Class 1
Disposal Facility, Preliminary Draft, by Air Resources Board, Stationary Source Division
November 1984.
-99-
------- |