Prepublication issue for EPA libraries
and State Solid Waste Management Agencies
DESTROYING CHEMICAL WASTES
IN COMMERCIAL-SCALE INCINERATORS
(Facility Report 6)
This final report (SW-l22c.5) describes work performed
for the Federal solid waste management program
under contract no. 68-01-2966
and is reproduced as received from the contractor
Copies will be available from the
National Technical Information Service
U.S. Department of Commerce
Springfield, Virginia 22161
U.S. ENVIRONMENTAL PROTECTION AGEM™
1977
-------
This report as submitted by the grantee or contractor has been tech-
nically reviewed by the U.S. Environmental Protection Agency (EPA).
Publication does not signify that the contents necessarily reflect
the views and policies of EPA, nor does mention of commercial products
constitute endorsement by the U.S. Government.
An environmental protection publication (SW-122c.5) in the solid
waste management series.
-------
FOREWORD
The tests described in this report are part of a program designed to
evaluate the environmental, technical, and economic feasibility of dis-
posing of industrial wastes via incineration. This objective is being
pursued through a series of test burns conducted at commercial incinerators
and with real-world industrial wastes. Seven incineration facilities and
fourteen different industrial wastes will be tested under this program.
The incineration facilities were selected to represent the various design
categories which appear most promising for industrial waste disposal.
The wastes were selected on the basis of their suitability for disposal
by incineration and their environmental priority.
This report describes the test conducted at Rollins Environmental
Services, Inc. (Deer Park, Texas), which was the sixth facility of the
series. Facility reports similar to this one have been published for the
first five tests which were conducted at the Marquardt liquid injection
facility in Van Nuys, California; the Surface Combustion pyrolysis
facility in Toledo, Ohio; the Systech fluidized bed reactor in Franklin,
Ohio; the Zimpro wet air oxidation unit in Rothschild, Wisconsin: and the
3 M rotary kiln incinerator in Cottage Grove, Minnesota. The facility
reports are primarily of an objective nature presenting the equipment
description, waste analysis, operational procedures, sampling techniques,
analytical methods, emission data and cost information. Facility reports
are published as soon as possible after the testing has been completed
at a facility so that the raw data and basic results will be available
to the public quickly.
In addition to the facility reports, a final report will also be
prepared after all testing has been completed. In contrast to the
facility reports which are primarily objective, the final report will
provide a detailed subjective analysis of each test and the overall
program.
ACKNOWLEDGMENTS
TRW wishes to express its sincere appreciation to Rollins Environ-
mental Services, Inc., particularly Messrs. Donald Matter and Jerry Neel,
for their cooperation in conducting these facility tests. The project is
also deeply indebted to Mr. Eugene Grumpier of the Office of Solid Waste
Management Programs, U.S. Environmental Protection Agency, for his advice
and technical direction.
111
-------
CONTENTS
Page
1. Summary 1
2. Introduction 4
3. Process Description 6
3.1 Facility Process 6
3.1.1 Incineration System 6
3.1.2 Waste Feed Systems 6
3.1.3 Auxiliary Fuel Feed Systems 6
3.2 Instrumentation 8
3.3 Emission Control System 8
4. Test Description 9
4.1 TRW Tests of PCB-containing Capacitors 9
4.1.1 Waste Tested 9
4.1.2 Operational Procedures 11
4.1.3 Sampling Methods 15
4.1.4 Analysis Techniques 24
4.1.5 Problems Encountered 30
4.2 ADL Tests of Nitrochlorobenzene Waste 31
4.2.1 Waste Tested 31
4.2.2 Operational Procedures 32
4.2.3 Sampling Methods 33
4.2.4 Analysis Techniques 36
4.2.5 Problems Encountered 37
5. Test Results 39
5.1 Results of PCB Tests 39
5.1.1 Operating Conditions for Tests 39
5.1.2 Composition of Combustion Zone Gas 41
5.1.3 Final Emissions 44
iv
-------
CONTENTS (CONTINUED)
Page
5.2 Results of NCB Tests 58
5.2.1 Operating Conditions for Tests 58
5.2.2 Destruction Efficiency and Composition of
Combustion Zone Effluent Gas 58
5.2.3 Final Emissions 66
6. Waste Incineration Cost 69
6.1 Capital Investment 69
6.2 Annual Operating Costs 69
7. References 78
Appendixes
Appendix A - Assessment of Environmental Impact of
Destroying Chemical Wastes 79
Appendix B - Sample Volume Data - PCB Tests 82
Appendix C - Analytical Chemistry Details - PCB Tests 86
Appendix D - PCB Analysis by GC/MS 88
Appendix E - Sample Volume Data - NCB Tests 96
Appendix F - Analytical Chemistry Details - NCB Tests 100
Appendix G-l - Rollins Environmental Services, Inc.
Operating Data 123
Appendix G-2 - Incineration System Raw Data 145
Appendix H - Calculation of Waste Destruction Performance 160
v
-------
FIGURES
Page
3-1 Schematic of Rollins Environmental Incinerator 7
4-1 Fiber Drums Being Fed to Rotary Kiln 14
4-2 Sampling Sites at Rollins 17
4-3 Sampling System for On-line Instruments 18
4-4 Instrument Racks 19
4-5 Combustion Zone Sampling Train Schematic 20
4-6 Water-cooled Probe Design 22
4-7 Combustion Zone Sampling Site 23
4-8 Stack Sampling Site 25
4-9 Hot Zone Sampling Train for NCB Tests 35
5-1 Filters From Combustion Zone and Stack Gas Sampling Trains 45
5-2 Photograph of Solid Residue Samples 53
F-l Sorbent Trap Extractor 103
F-2 GC Calibration Curve for NCB Standards 106
VI
-------
TABLES
Page
1-1 Results Summary 2
4-1 Composition of PCB Survey Sample 10
4-2 Composition of PCB Representative Sample 11
4-3 Description of On-line Instruments 21
4-4 Summary of Analytical Methods 29
5-1 Incinerator System Parameters Data Summary 40
5-2 Gas Composition Data Summary 41
5-3 Summary of Survey Analysis on the Combined Probe Wash
and Particulate Filter Extracts 43
5-4 Summary of Survey Analysis on Sorbent Trap Extracts 43
5-5 Survey for Trace Metals in Combustion Zone Samples by ICPOES 46
5-6 Detection Limits for Elements Not Found in the Combustion
Zone Samples by ICPOES 47
5-7 Particulate Loading in the Effluent Gas 47
5-8 Survey for Trace Metals in Stack Samples by ICPOES 48
5-9 Detection Limits for Elements Not Found in the Stack Samples
by ICPOES 49
5-10 Summary of Survey Analysis of Scrubber Water Extracts 50
5-11 Survey for Trace Metals in Scrubber Water Samples by ICPOES 51
5-12 Detection Limits for Elements Not Found in the Scrubber Waters
by ICPOES 52
5-13 Results of Test III Solid Residue Analysis for PCBs
by GC/MS 54
5-14 Summary of Survey Analysis of Solid Residue Extracts 54
5-15 Analysis of Solid Residues by ICPOES 56
5-16 Analysis of Solid Residues by SSMS 57
VI1
-------
TABLES (CONTINUED)
Page
5-17 Operating Conditions for Tests on NCB Wastes 59
5-18 Summary of Quantities of Organic Materials in Hot Zone
Effluent Sample Extracts 60
5-19 Hot Zone Emission Rates and Calculated Destruction Efficiencies 61
5-20 Quantitative Analysis Results Obtained from On-line
Instruments and Gas-Detecting Tubes 63
5-21 Organic Chemical Species Found in Hot Zone Effluent
Sample Extracts 65
5-22 Chloride and Nitrate Analyses of Scrubber Wastes 68
6-1 Capital Investment - 5000 Metric Tons/Year Used PCB Capacitor
Waste Incineration Plant 70
6-2 Capital Investment - Central Facility for Incineration
4540 Metric Tons/4-Month Period NCB Waste 72
6-3 Capital Investment - 4540 Metric Tons/Year NCB Waste
Incineration Plant (On-Site) 73
6-4 Annual Operating Cost - 5GOO Metric Tons/Year Used PCB
Capacitor Waste Incineration Plant 75
6-5 Pro-Rata Annual Operating Cost - Central Facility for
Incineration - 4540 Metric Tons/4-Month Period NCB Waste 76
6-6 Annual Operating Cost - 4540 Metric Tons/Year NCB Waste
Incineration Plant (On-Site) 77
B-l Sampling System Data Summary 83
B-2 Rollins Sample Gas Volumes at Standard Conditions 84
B-3 Collected Water Volume Data 85
D-l AMUs to Be Searched for Raw Counts 94
E-l Stack Sampling Data 97
E-2 Hot Zone Sampling Data 98
E-3 Estimated Total Gas Effluent Flow Ratio 99
vin
-------
TABLES (CONTINUED)
Page
F-l Volumes of Impinger Solutions 109
F-2 Gravimetric Data for Probe Wash, Filter, and
Dry Impinger Samples H°
F-3 Results of Gravimetric Analyses on Concentrated
Organic Extracts '''
G-l Test I - TRW Background I43
6-2 Test II - Hammermi11ed PCB Capacitors 146
6-3 Test III - Whole PCB Capacitors 148
6-4 Run R1(B) Background Test I50
6-5 Run R2 - NCB to Kiln I53
6-6 Run R3 - NCB - Diesel Fuel to Loddby 156
6-7 Run R4 - NCB - Diesel Fuel to Loddby 158
IX
-------
1. SUMMARY
Incineration tests of selected chemical wastes were conducted at a
contract disposal facility operated by Rollins Environmental Services,
Inc., in Deer Park, Texas. This incineration system, consisting of a
rotary kiln and a liquid injection burner feeding a common afterburner, is
representative of commercial waste destruction equipment presently in use.
Total heat release of the system was 28 million kcal/hr (110 million Btu/
hr). These tests were performed to determine the effectiveness of thermal
destruction of two different industrial wastes: 1) discarded electrical
capacitors containing polychlorinated biphenyls (PCBs), and 2) waste from
production of nitrochlorobenzene (NCB).
The PCB-containing capacitors were incinerated in the rotary kiln both
as a hammermilled "fluff" and as whple capacitors. PCB tests were performed
at maximum incinerator temperatures and residence time to operate under
conditions of maximum destructive effectiveness.—NCB waste was destroyed
in the liquid injection burner at two different waste feed conditions to
determine the effects of operating variables. Number 2 oil was utilized as
auxiliary fuel in the destruction of each waste. A background test with
auxiliary fuel only was performed prior to initiation of the PCB tests
and again just before the NCB burns to obtain baseline combustion and emission
data.
During each test, combustion products were continuously monitored by
on-line instrumentation. Samples were taken of the combustion zone effluent
to evaluate the destruction effectiveness. Stack gases and waste products
(scrubber water, ash, etc.) were collected to verify the environmental
safety of the tests. The PCB tests were monitored by a TRW sampling team,
while the NCB test samples were taken by a team from Arthur D. Little, Inc.,
under subcontract with TRW. Laboratory analyses of test samples were
performed by each organization for their respective waste tests.
Results of the waste destruction tests are summarized in Table 1-1.
It should be noted that particulate stack emissions were higher for the whole
capacitor test (53 mg/m3) compared to the hammermilled capacitor test
(35 mg/m3). Of significance also is the fact that while emissions of HC1
from the incinerator alone were high (circa 6 g/m3), the wet scrubber
removed 99.8 percent of the HC1 from the effluent and resulting stack
emissions were approximately 13 mg/m3.
No waste constituents were found in any of the combustion gas samples
above the minimum detection limits indicated in Table 1-1. Incineration
of each waste was accomplished with high effectiveness, with overall waste
destruction efficiencies of over 99.999 percent for every test except
incineration of whole capacitors, where waste residuals in the ash reduced
the overall destruction efficiency to approximately 99.5 percent. No PCB
concentrations above 0.1 mg/kg were found in the ash from destruction of
1
-------
TABLE 1-1. RESULTS SUMMARY
K1ln Flame Temperature (°C)
Afterburner Temperature (°C)
Calculated Residence Time (sec)
Solid Waste Feed (kg/hr)
Liquid Waste Feed (1/hr)
Number 2 Oil Feed (1/hr)
Quality of Stack Emissions:
Partlculate (mg/m3)
Hydrochloric Add (mg/m3)
3
Trace Metals (mg/m )
Quality of Combustion Gas:
Total Organics (mg/m )
Waste Content (mg/m3)
Trace Metals (mg/mj)
Quality of Scrubber Water:
Total Oranglcs (mg/1)
Trace Metals (mg/1)
Chloride Ion (mg/1)
pH
Quality of Ash:
Waste Content (mg/kg ash)
Destruction Efficiency:
Total Organics (percent)
Waste Constituents (percent)
PCB-Containinq
Hammermil led
1252
1331
3.2
210
None
2411
35
Not Analyzed
2.7 Pb
14
Not Detected
(<0.005)
12Pb,1.0Sn
0.19
1.7Pb, l.SZn
Not Analyzed
5
Not Detected
99.98
>99.999
Capacitors
Whole
1339
1332
3.0
360
None
2300
53
Not Analyzed
3.3 Pb
23
Not Detected
(<0.005)
HPb,2.6Sn
0.23
4.0Pb,1.3Zn
Not Analyzed
5
470
99.96
99.5
Nitrochlorobenzene Waste
1st Test
Not Used
1307
2.3
None
404
1616
14
13 (a)
<0.01Ti, Ni.Cr
42
Not Detected
(<0.05)
<0.01Ti,Ni ,Cr
1.0
<0.1Ti,Ni,Cr
1725
3
No Ash
99.84
>99.999
2nd Test
Not Used
1332
2.3
None
350
1410
16
13
<0.01Ti,Ni ,Cr
53
Not Detected
(<0.05)
<0.01Ti,Ni ,Cr
1.4
<0.1Ti, Ni.Cr
1815
3
"
No Ash
99.87
>99.999
(a)
Trace metal contents for NCB effluents were calculated from spark source mass
spectrometric analysis of representative waste sample.
(b)
Overall destruction efficiency reduced due to high waste content in ash.
ciency based on combustion gas sampling was >99.999 percent.
Effi-
-------
hammermilled capacitors. Total organic destruction efficiencies ranged
from 99.84 to 99.98 percent* In general, the capability of the Rollins
incineration system to destroy either solid or liquid wastes was effectively
demonstrated during these tests.
Capital and operating cost estimates were made for incinerator systems
similar to the Rollins unit to destruct each of the two wastes. Estimated
capital investment, not including land cost, for a facility to hammermill
and incinerate 5000 metric tons of waste capacitors per year is $3.65
million, with the total fixed and variable operating costs equivalent to
$741 per metric ton of waste destroyed. Over $300 per ton of operating
cost can be saved if a high heat content waste is utilized instead of
Number 2 fuel oil to provide temperatures required to destroy the PCB.
Using a waste to incinerate the capacitors would be the actual commercial
destruction method; however, in order to more precisely analyze the PCB
combustion products, a clean-burning fuel oil was used as auxiliary fuel
for these tests.
Estimated operating cost for destruction of 4540 metric tons per year
of NCB waste (one half of estimated total volume generated annually) in a
central, contractor-owned facility such as Rollins is $242 per metric ton,
including cost of transporting wastes. For this central facility, presumed
to operate with NCB waste over a four-month period of each year, the esti-
mated capital investment is $3.75 million; therefore, the pro-rata capital
investment for NCB destruction would be $1.25 million (one-third use of
facility). For an on-site facility, scaled to a size adequate for destruction
of 4540 metric tons per year of NCB waste only, the estimated capital in-
vestment is $2.82 million and the estimated operating cost is $283 per
metric ton.
*Destruction efficiency for total organics compares the input rate of
combined waste and auxiliary fuel to the emitted rate of all organic material
found in the combustion samples. Waste destruction efficiency compares only
waste input rate to concentration of organic waste constituents in the com-
bustion gas. All combustion zone samples were taken in the afterburner exit
duct prior to the scrubber system. A sample destruction efficiency cal-
culation is presented in Appendix H.
-------
2. INTRODUCTION
The U.S. Environmental Protection Agency has sponsored a program*
to evaluate the effectiveness of a variety of types of commercial thermal
destruction facilities for destroying chemical wastes. This report de-
scribes test operations and results of incinerating two different wastes
PCB-containing capacitors and nitrochlorobenzene (NCB), at a contract dis-
posal facility operated by Rollins Environmental Services, Inc., in Deer
r cl i K j I GXd S •
The Rollins incineration system, consisting of a rotary kiln and a
liquid injection burner feeding a common afterburner, was selected as
representative of full-scale waste destruction equipment presently in
commercial use. Rotary kiln incineration represents a well-established
technology with widespread applicability to a variety of hazardous wastes
J^LSfr4.1"616"6 is 28 mil11on kcal/hr, with combustion gas temperatures
ot 1300 C in the afterburner. Stack emissions are controlled by a venturi
scrubber system.
The two wastes tested at the Rollins facility, PCB-containing capacitors
and nitrochlorobenzene production waste, were selected on the basis of their
suitability for destruction in a rotary kiln incinerator. A conveyor system
was used to feed either hammermilled or whole capacitors into the kiln
The afterburner provided additional residence time at temperatures sufficient
to insure efficient destruction of the PCB in the capacitors. NCB waste was
selected to be burned in the kiln, but did not atomize properly by itself
in the kiln burner nozzle at Rollins. After being mixed with No. 2 oil the
NCB was fed through the liquid injection burner which fired the afterburner.
Polychlorinated biphenyls (PCBs) have been used extensively in capa-
citor dielectric fluids, with an estimated 11,000 metric tons per year
produced for this use. Although other materials are now being used as
substitutes for PCB, thousands of tons of capacitors, each containing 20
to 40 percent PCB by weight, are taken out of service annually due to
failure or obsolescence. PCB is very resistant to chemical and bioloqical
degradation; therefore, incineration or thermal degradation is the most
feasible means of destruction. Because of the extreme stability of PCB,
sufficient residence'time at high temperature is essential for destruction.
The principal use of nitrochlorobenzene is as an intermediate material
in the manufacture of synthetic dyes, drugs, pesticides, and photochemicals.
tstimated annual production of NCB waste similar to that tested is 9 100
metric tons.
*Contract No. 68-01-2966
-------
The following report sections describe in detail the incinerator
process equipment {Section 3), and the wastes destroyed and test and
sampling procedures followed (Section 4). Test results are presented and
discussed (Section 5), including effectiveness of destruction of the
wastes. An estimate of the capital investment and operating costs of
disposing of wastes using this type of incinerator equipment is also
included in the report (Section 6).
-------
3. PROCESS DESCRIPTION
3.1 FACILITY PROCESS
The Rollins incineration facility process is shown schematically in
Figure 3-1. Basic system components include:
• Rotary kiln incinerator
• Waste liquid burner
• Afterburner
t Waste feed systems
• Auxiliary fuel feed systems
• Instrumentation
• Emission control system
Following is a description of the incineration and feed systems.
Facility instrumentation and emission controls are discussed in subsequent
Sections 3.2 and 3.3, respectively.
3.1.1 Incineration System
The incineration system consists of a rotary kiln and a liquid in-
jection burner, both feeding a common afterburner (Figure 3-1). Total heat
release is 28 million kcal/hr. The kiln is 4.9 meters long and 3.2 meters
in diameter. Flame temperature in the kiln is nominally 1300°C.
The waste liquid burner (Loddby) measures 4.9 meters in length by 1.6
meters diameter. Loddby flame temperature is approximately 1500°C. Re-
sulting afterburner gas temperature is typically 1300°C. Afterburner
dimensions are: 10.6 meters overall length, 4.0 meters high, and 4.3 meters
wide. Overall retention time of the incineration system is from two to three
seconds.
3.1.2 Waste Feed Systems
Solid wastes, usually packed in fiber drums, are fed into the rotary
kiln by a conveyor. Liquids and sludges may also be pumped into the kiln.
Liquid wastes that are to be burned in the Loddby can be fed directly from
tank trucks or from blending tanks.
3.1.3 Auxiliary Fuel Feed Systems
Both the kiln and the Loddby are equipped with natural gas ignitors and
gas burners for initial refractory heat-up, flame stability, and supple-
mental heat, if necessary. Number 2 fuel oil was also used as auxiliary
fuel for these tests. Fuel oil was used to provide heat for the PCB tests,
since the capacitors by themselves did not have sufficient heat capacity.
-------
CONVEYOR
FIBER PACKS
ASH RESIDUE
SAMPLE
EXIT GAS
ON LINE
GAS
MONITORS
SOLID WASTE
FEED CHUTE
EPA METHOD 5
SAMPLING TRAIN
HOT ZONE
SAMPLING
TRAIN
KILN EXIT DUCT
AFTERBURNER
HOT DUCT
MIST ELIMINATOR
ABSORBTION TRAYS
FRESH
WATER
FEED
WASTE FEED
SAMPLE
FEED WASTE LIQUID BURNERS
LODDBY
(WASTE LIQUID BURNER)
INDUCED DRAFT FANS
SCRUBBER SYSTEM
HYDRATED LIME
SLURRY FEED
SAMPLE
PORTS
SCRUBBER
LIQUID SAMPLE
DISCHARGE
SCRUBBER WATER
Figure 3-1. Schematic of Rollins Environmental Services incinerator.
-------
The nitrochlorobenzene waste was mixed with fuel oil in order to be fed
into the Loddby for incineration.
3.2 INSTRUMENTATION
Measurements were made of all process parameters, including temperatures
pressures, and flow rates. On-line gas analysis instruments were also
monitored and recorded in the sampling trailer. The following operating
temperatures (see Figure 3.1) were recorded during each test: 1) kiln
flame, 2) kiln exit duct, 3) Loddby flame, 4) afterburner, and 5) hot duct.
Measured flow rates included: 1) natural gas, 2) auxiliary fuel oil,
and 3} waste feed. In addition, scrubber venturi pressure drop and lime
consumption were also monitored during tests.
3.3 .EMISSION CONTROL SYSTEM
Atmospheric emissions from the combustion of solid and liquid wastes
during the Rollins incineration tests were controlled by a venturi scrubber
as shown in Figure 3-1. Lime is injected to neutralize the scrubber water
Used scrubber water enters settling ponds where it is analyzed and further
treated, if necessary, before discharge.
Exhaust gases also pass through absorption trays and a mist eliminator
before entering the stack. The exhaust stack is 30 meters high with
standard sampling ports and a platform about 16 meters above ground level.
-------
4. TEST DESCRIPTION
This section describes the manner in which the tests were conducted.
It is divided into the following subsections, listed in order of discussion:
• Physical and chemical description of the wastes that were tested.
• Operational procedures used and a test-by-test commentary.
• Sampling methods.
• Analysis techniques.
• Description of the problems encountered related to the facility
and sampling.
These topics will be discussed separately for the TRW and the ADL por-
tions of the tests conducted at Rollins. TRW first tested PCB-containing
capacitor materials; then nitrochlorobenzene (NCB) waste was tested by ADL.
4.1 TRW TESTS OF PCB-CONTAINING CAPACITORS
The PCB-containing capacitors were tested both as a hammermilled "fluff"
and as whole capacitors. Test burns of these materials were conducted
using the rotary kiln incineration system at Rollins, and samples were taken
for analysis of the combustion gases, stack gases, scrubber waters, and
solid residues of the wastes remaining after incineration.
4.1.1 Waste Tested
A survey sample of assorted'whole capacitors was received early last
year, and a representative sample was later taken during the hammermilling
of the capacitors. These waste samples were analyzed for both organic and
inorganic composition, and the results are reported in this section.
The survey sample consisted of several types of small capacitors con-
taining PCB fluid. Because each capacitor was different and because it was
uncertain whether they represented what would actually be burned in the
field tests, it was deemed impractical and unproductive to characterize
the whole capacitors (i.e., percent aluminum shell, percent bakelite, per-
cent paper packing, etc.). As a result, analyses were performed only on
the fluid obtained by puncturing and draining several of the survey sample
capacitors.
The drained PCB fluid was colorless and had a specific gravity of
1.4 at 15.6°C (60 F). The fluid would not ignite in a calorimeter, thus
its heating value is estimated to be less than 3,000 kcal/kg (5,000 Btu/lb).
-------
Quantitative elemental analyses for C, H, N, S, and total halogens (as Cl)
gave the following results:
Element Composition, %
C 54.80
H 2.63
N 0.008
S 0.019
Cl 42.59
Based on the elemental composition, the capacitor fluid approximates
Aroclor 1242. The infrared spectrum of the capacitor fluid waste, mea-
sured as a thin film between NaCl windows, was also consistent with that
of a mixture of polychlorinated biphenyl compounds.
Analysis for trace elements by X-ray fluorescence found Si and P
at >500 ppm and 50-500 ppm, respectively, along with chlorine and bromine.
The survey sample of PCB capacitor fluid was chromatographed on a
fifty meter SE 30 capillary column which was programmed from 50°C to 250°C
at 10°C per minute. The total ion monitor of the mass spectrometer was
used as the detector. Chlorinated biphenyl compounds from the monochloro-
(Ci2HgCl) up through pentachloro biphenyl C^HsCls were seen. A search
specifically for higher molecular weight compounds was conducted, but none
were seen. Most, if not all, of the isomers were detected, but it was
considered unnecessary to identify each individual isomer. The presence
of the nonchlorinated biphenyl species was also noted. The compounds found
in the GC/MS of the survey sample are presented in Table 4-1. The esti-
mated concentrations are based on the mass spec total ion monitor and may
be somewhat different from normalized percent concentrations from other
GC detectors, but the data serve the purpose of showing the distribution
of the PCB compounds in the survey sample.
TABLE 4-1. COMPOSITION OF PCB SURVEY SAMPLE
Compound
Estimated Concentration,
(percent w/w)
Biphenyl
Monochlorobiphenyl (2 isomers)
Dichlorobiphenyl (3 isomers)
•Trichlorobiphenyl (6 isomers)
Tetrachlorobiphenyl (13 isomers)
Pentach1orobipheny1 (8 isomers)
0.2
2
18
41
37
2
10
-------
Only two types of capacitors were selected for the test burns. A
portion of these were hammermilled, and a portion were kept to be burned
whole. Since the hammermi11ing process helps to mix and composite the
material, a sample of the hammermilled "fluff" was taken to be representa-
tive of the feed for both the fluff and the whole capacitor test.
A 20 g portion of the representative (fluff) sample was extracted
twice with pentane in a Soxhlet for 24 hours for a total of 48 hours of
extraction. The extract was then made to a standard volume of 500 ml.
An aliquot of this extract was evaporated to remove all solvent and then
weighed. The weight of organic residue represented 29 percent of the
extracted fluff sample. To determine nonextractable organics (e.g.,
paper and plastics), the solid material remaining in the Soxhlet thimble
after the pentane extraction was ashed at 800°C for one hour. The remain-
ing ash was 47 percent by weight of the solids sample or 32 percent of the
initial 20 g fluff sample.
The pentane extract of the representative sample was analyzed by
GC/MS to quantify the types of PCB compounds present. These results are
presented in Table 4-2. The composition of the fluff sample is slightly
different from that of the survey sample (Table 4-1) but still shows a
predominance of the di-, tri-, and tetra-PCB classes. Detection limits for
the classes of PCBs (7 to 10 chlorine atoms/molecule) not found in the
fluff extract were 0.03 g/kg fluff.
TABLE 4-2. COMPOSITION OF PCB REPRESENTATIVE SAMPLE
Compound
Monochlorobiphenyl
Dichlorobiphenyl
Trichl orobi phenyl
Tetrachl orobi phenyl
Pentachl orobi phenyl
Hexachl orobi phenyl
Estimated Concentration
g PCB/kg Fluff
1.0
94
150
180
55
3.3
Percent of Total PCB
Compounds
0.2
20
31
37
11
0.7
4.1.2 Operational Procedures
Detailed operating procedures, including both a test plan and a safety
plan, were reviewed and approved prior to the arrival of the TRW sampling
team on site. Procedures and operating conditions were also recorded dur-
ing the field tests. Following are brief summaries of both plans, a test-
by-test conmentary on events that occurred in the field, and information
on the disposal of the waste residues.
11
-------
4.1.2.1 Test Procedures
The PCB tests conducted at Rollins consisted of one test burn of
waste capacitors that had been hanmermilled and sealed in 130 liter (35 gal
Ion) fiberpacks and one test burn of whole waste capacitors. The basic
procedures for each test are listed below:
• Verify instrumentation and sampling systems ready..
• Connect auxiliary fuel (No. 2 oil) lines to rotary kiln and
Loddby.
• Obtain a sample of the fresh scrubber water.
t Switch the kiln and Loddby to fuel oil, stabilize temperatures,
and purge for 1 to 2 hours.
t Activate on-line combustion gas analyzer system.
• Initiate waste feed to rotary kiln.
• Stabilize flows and temperatures.
• Extended burn duration:
- Process data acquisition.
- Combustion gas composition data acquisition.
— Combustion zone and stack gas sampling.
— Scrubber liquid sampling.
• Collect samples of residue from the rotary kiln.
• Switch the kiln and Loddby burners back to Rollins waste materials.
The two tests consisted of burn periods during which a 3-hour combus-
tion gas sample was to be acquired at steady-state operating conditions.
A 3-hour sampling run while burning with No. 2 oil only was also required
to obtain background data. As explained in the test commentary (Section
4.1.2.3), it was not always possible to obtain a 3-hour sample. Stack
gases were to be sampled by the standard EPA Method 5 procedure.
Because of the limited amount of available waste material, Rollins
was asked to use their operating experience and set their optimum operating
conditions for the tests. Target temperatures were as follows:
Afterburner 1260 to 1320°C
Hot duct 1090°C
Loddby flame 1480°C
Kiln flame 1320°C
12
-------
4.1.2.2 Safety Procedures
Safety requirements established and maintained for handling and incin-
erating these specific wastes included the following:
• Only authorized personnel with prior approval were permitted
in the test area during operations.
• Waste handling was performed only by personnel wearing suitable
protective clothing and trained in handling such materials.
• Visual observation of the test system was maintained at all times
during operation.
• Canister gas masks and emergency oxygen resuscitation units were
available in the immediate area.
t Telephone numbers of emergency agencies were posted near the
test area.
4.1.2.3 Test Commentary
Two test burns of PCB-containing capacitors were conducted at Rollins,
one using whole capacitors and one using a "fluff" obtained by hammermilling.
Since auxiliary fuel was required for the combustion of both forms of
capacitors, a baseline test was also conducted with No. 2 oil.
Test I - No. 2 Oil Background Test
A baseline test was performed with No. 2 oil to acquire background
data on the performance of the rotary kiln system while burning only auxil-
iary fuel. A 2.0-hour combustion zone sample was procured while the system
was operating at steady-state conditions of 1306°C kiln and 1308°C after-
burner average temperatures. No stack sample was obtained due to severe
storm conditions at the time. This problem is discussed further in Sec-
tion 4.1.5. Because of the resulting lack of stack velocity data, residence
time was not calculated for this test. Scrubber water samples were secured.
There was no solid residue from the kiln generated during this test.
Test II - PCB "Fluff" Test
The hammermilled fluff of PCB-containing capacitors had been packaged
for storage in polyethylene-lined fiber drums. These drums were fed whole
via the conveyor belt into the rotary kiln as shown in Figure 4-1. The kiln
was first purged with a 1.5-hour burn of No. 2 oil alone before beginning
testing of the PCB fluff. As the first fiber drum was fed, the on-line
monitors were watched for any sign of combustion problems. The kiln main-
tained essentially steady-state conditions, and no surge of unburned hydro-
carbons was observed. Rollins then proceeded to feed drums at a rate of one
every four or five minutes, and sampling was begun. The fiber drums burned
in this test weighed an average of 25 kg with an approximate tare weight of
9 kg, indicating that roughly 210 kg per hour of PCB fluff was fed into the
rotary kiln. Temperatures averaged 1252°C in the kiln and 13310C in the
afterburner. Residence time was calculated to be 3.2 seconds.
13
-------
DUCT
ROLLINS
CONTROL
ROOM
I AFTERBURNER
FIBER DRUMS
CONTAINING
CAPACITOR
FLUFF
ROTARY
KILN
CONVEYOR
BELT
Figure 4-1. Fiber drums being fed to rotary kiln,
-------
The sampling train at the combustion zone collected sample for
2.25 hours before a power failure at the sampling site forced the train
to be shut down. A full one-hour stack sample was obtained. Scrubber
water samples were acquired and 2 kg of the solid residue from the test
was taken as a sample.
After completion of the sampling efforts, the Rollins facility changed
from No. 2 oil to their normal alky! waste fuel and continued to feed the
few remaining fiber drums of the PCB fluff while watching the on-line gas
monitors. Composition of the gas leaving the afterburner remained the
same while burning with the alkyl fuel, and no increase in unburned hydro-
carbons was observed.
Test III - Whole PCB-Containing Capacitors
After a 1.25-hour purge on No. 2 oil, the first whole capacitors were
fed. The whole capacitors were removed from the steel drums they had been
shipped in and were fed into the rotary kiln individually via the conveyor
belt. The initial feed rate proved too fast as 20 minutes into the capaci-
tor burning, an increase of 20 to 30 ppm of unburned hydrocarbons passing
the afterburner was observed on the on-line hydrocarbon monitor. The feed
rate was subsequently reduced to approximately two drums of capacitors per
hour. A total of six drums was burned at an average weight of 205 kg
each. Tare weight of the drums was approximately 24 kg, yielding an average
feed rate of 360 kg of capacitors per hour.
Steady conditions were then maintained throughout the remainder of
the test except for a 15-minute period during which no capacitors were fed
because of a conveyor belt malfunction. Temperatures averaged 1339°C in the
kiln and 1332°C in the afterburner, with a calculated residence time of
3.0 seconds.
After 2.25 hours of sampling, the combustion zone train was shut down
when Rollins ran out of capacitors to feed. A full one-hour stack sample
was obtained along with scrubber water and solid residue samples.
4.1.2.4 Disposal of Haste Residues
All of the PCB capacitor material consigned by Rollins for these tests
was consumed in the tests. The steel drums that had contained the whole
capacitors were landfilled in an approved Class 1 site. The scrubber
waters were piped to treatment ponds at the Rollins site, and the solid
residues were trucked from the kiln's hopper to an approved Class 1 site
and landfilled.
4.1.3 Sampling Methods
Sampling methods used in the tests at Rollins were chosen to cover
three basic areas:
1) Continuous, on-line monitoring of gas composition to deter-
mine and follow steady state conditions.
15
-------
2) Collection and concentration of hot zone combustion products
to identify and quantify the trace organic and inorganic
species formed.
3) Collection of final emission and waste products to evaluate
the environmental safety of the tests.
Following is a brief summary of the methods for each of these areas.
More detailed discussions can be found in the Rollins Analytical Plan
(Reference 1). The locations of the trailer and sampling trains at the
site are shown in Figure 4-2.
4.1.3.1 On-Line Gas Monitoring
Gases were drawn continuously from the hot zone just downstream of
the afterburner through a ceramic probe, and then through a heated Teflon
sample line to the trailer. The gas then entered the system shown in
Figure 4-3. The gas conditioner supplied a cool, dry, particulate-free
sample to all of the analyzers except the hydrocarbon (HC) monitor which
used an untreated sample. A heated Teflon line carried the HC gas sample
from a tee in the unconditioned sample line to the HC analyzer.
The monitoring instruments used are listed with their operating
ranges in Table 4-3. Data was recorded on Hewlett-Packard 680M strip
chart recorders. Figure 4-4 shows the instrument racks mounted in the
sampling trailer. The analyzers, recorders, and manifold valves were all
located in racks to provide ease of operation and accessibility.
4.1.3.2 Combustion Zone Sample
The sampling train used to collect hot zone gases, vapors, and partic-
ulate matter is shown schematically in Figure 4-5. It consists of a
standard EPA Method 5 train with the following important modifications
presented in order according to flow direction through the train:
• There was a stainless steel jacketed, water-cooled probe (shown
schematically in Figure 4-6) with a quartz liner. The liner
provides an inert surface for the sample gas, and the cooled,
stainless steel jacket cools this gas in order to quench any
further reactions of the sample constituents and to yield a gas
temperature compatible with train materials. Further cooling
of the gas was achieved by aspirating an air/water mixture into
the space between the steel jacket and quartz liner.
t A chromel/alumel thermocouple was potted into the dogleg going
from the quartz probe liner to the filter housing to check the
temperature of the gas stream at that point.
• An ultra high-purity glass fiber filter was used, Gelman Spectro-
quality Type A. The filters, which were muffled to remove organics,
have extremely low background levels of inorganics. They were
16
-------
HOT
DUCT
COMBUSTION
ZONE SAMPLING
SITE
STACK
SAMPLING
SITE
AFTERBURNER
EPA
SAMPLING
TRAILER
Figure 4-2. Sampling sites at Rollins.
17
-------
00
2B
l-3
HCF
l-2
Lyv
PROBE
r
A-.-__n____n_
x_
314 R6
r
(HEAT TRACED)
BLOW
BACK I
LINE
2 WAY BALL VALVE
WHITEY SS-4234
3 WAY BALL VALVE
WHITEY SS-43X54
SS-4SX38
5 WAY BALL VALVE
WHITEY SS-432
FLOWMETER WITH INTEGRAL
CONTROL/SHUTOFF VALVE
HYDROCARBON REMOVAL
FILTER
PRESSURE REGULATOR WITH
GAGE (2 STAGE)
DRYER/FILTER
CONDENSATE TRAP
1/4' TUBE
\fi' TUBE
IN
SAMPLE
IN
PPO
IN
HC
OUT
CONDITIONING
SYSTEM
SAMPLE
OUT
BYPASS
OUT
3B,
~-W
VENT
R.A.
•VINT
Figure 4-3. Sampling system for on-line instruments.
-------
MANIFOLD
ALVES AND
FLOWMETERS
Figure 4-4. Instrument racks.
19
-------
ro
o
4 INCH
FILTER
HOLDER
INCINERATOR
WALL
SOLID
SORBENT
TRAP
GAS
i SAMPLE
VALVE
WATER
COOLED
PROBE
QUARTZ
LINER
HEATED AREA
THERMOMETER
CHECK
VALVE
THERMOMETERS
ORIFICE
ICE
BATH
BVALVES IMPING ERS VACUUM
\ (MAXIMUM SIX) GAUGE
VACUUM LINE
MAIN
VALVE
DRY TEST METER AIR-TIGHT
PUMP
Figure 4-5. Combustion zone sampling train schematic.
-------
tared by desiccating and weighing on consecutive days to a constant
weight (±0.1 mg), then stored and handled throughout the tests and
analyses in glass petri dishes.
A solid sorbent trap, designed to absorb the organic constituents
in the sample gas stream, is located downstream of the heated filter
and upstream of the first impinger. The sorbent trap, with overall
dimensions of 170 x 45 mm, contained ^40 g of XAD-2, an Amberlite
resin of the type commonly used as a chromatographic support.
TABLE 4-3. DESCRIPTION OF ON-LINE INSTRUMENTS
Species Analyzed
Total hydrocarbons (HC)
Carbon monoxide (CO)
Carbon dioxide (C02)
Oxygen (Op)
Sulfur Dioxide (S02)
.Oxides of nitrogen (NO )
/\
Manufacturer
and Model
Beckman
Model 402
Beckman
Model 865
Beckman
Model 864
Taylor
OA 273
Enviro-
metrics
Thermo
Electron
Model 10- A
Range*
0.05 ppm - 10?£ with
eight ranges
2-200 ppm
10-100 ppm
0.05 - 5%
0.02 - 20%
0.05 - 5%
0.25 - 25%
1 - 1 00%
0 - 100 ppm
0.05 - 10,000 ppm
with eight ranges
*A11 of these manufacturers report an accuracy of ±1 percent of full
scale for their instruments.
This hot zone train was operated at a flow rate of approximately
30 liters/min for two hours during each test, thereby sampling an average
of 3 to 4 cubic meters. Gas volumes were measured to 0.03 liter, with a
leak rate of less than 0.6 liter/min. Operating parameters for the train
and sample volume data are tabulated in Appendix B. The location of the
hot zone sampling train at the test site is shown in Figure 4-7.
The following hot zone samples were obtained from the sampling train
of each test:
t Solvent probe wash.
• Aqueous probe wash.
21
-------
• 10-cm diameter particulate filter.
• Solid sorbent trap.
• Combined impinger solutions.
• Acidified split of combined liquid impingers.
• Solvent wash of impinger glassware.
• Spent silica gel.
In addition to the samples from the hot zone train, portions of the
combustion gas were collected in Tedlar bags from the sample bypass port
of the on-line hydrocarbon analyzer.
COOLING
WATER
INLET
PROBE
LINER
AIR/WAT£R COOLANT
FOR SAMPLE GAS
INLET
COOLING
WATER
OUTLET
Figure 4-6. Water-cooled probe design.
22
-------
ro
CO
<• WATER-COOLED
PROBE
HOT
DUCT
(COMBUSTION ZONE
GAS SAMPLING
TRAIN
CERAMIC PROBE AND
HEAT-TRACED LINE
FOR ON-LINE MONITORS
DUCT FROM WET
SCRUBBER TO
STACK
Figure 4-7. Combustion zone sampling site.
-------
4.1.3.3 Sampling Emissions and Waste Products
Samples of the stack effluent, spent scrubber water, and solid com-
bustor residue were taken during and after each test to evaluate the
environmental safety of the final emissions. An EPA Method 5 test was
performed at the stack for particulate mass loading and composition deter-
minations. Location of the sampling train at the test site is shown in
Figure 4-8. Two diameters of the 2-meter diameter stack were traversed.
Sampling was for one hour at approximately 20 liters/min. Gas volumes
were measured to 0.03 liter, with a leak rate of less than 0.6 liter/min.
Operating parameters for the train and sample volume data are tabulated
in Appendix B.
Oxidizing agents (^2 and (NH^SgOs) were added to the impingers to
aid scrubbing of trace metals. The following samples were obtained for
each test from the stack sampling train:
• Acetone probe wash.
0 10-cm diameter particulate filter.
• Impinger solutions.
• Acidified split of impinger solutions.
• Spent silica gel.
Spent scrubber water samples were taken from the end of a drain pipe
that carried the used scrubber water to treatment ponds. The fresh
water being fed into the single pass wet scrubber was sampled from a pipe
transporting the well water into the scrubber system.
The solid residues remaining after combustion of the wastes in the
rotary kiln were accumulated in a hopper below the kiln throughout each
test. Samples were then taken from this collected residue at the end of
each test. The residue (consisting of steel hoops from the fiber drums,
large chunks of metal, fine ash, etc.) was nonhomogeneous in nature, thus
a truly representative sample could not be obtained without the use of
grinding and milling equipment.
4.1.4 Analysis Techniques
Samples taken as described in Section 4.1.3 were analyzed for both
organic and inorganic constituents. When necessary, extractions were
performed first to concentrate the sample in a suitable form for analysis.
Techniques used for these extractions and analyses will be briefly summa-
rized here. For more detailed discussions, see the Rollins Analytical
Plan (Reference 1 ) .
4.1.4.1 Extractions and Sample Preparation
Both solvents and acids were used to extract organics and inorganics,
respectively, from the appropriate samples. These procedures and the basic
sample preparation steps are listed in this section by sample type.
24
-------
ro
C71
MONORAILS
Figure 4-8. Stack sampling site.
-------
• Probe Washes
Combustion Zone
The quartz liner had been rinsed first with acetone to remove
organic matter. A water rinse was added to the procedure to remove
the remaining particulate matter upon which the acetone had little
effect. The acetone probe rinses were filtered. The solids were
weighed and added to the filter from the sampling train in a Soxhlet
apparatus. The acetone filtrate was added to the pentane extraction
solvent. The combined filters were then extracted with this acetone/
pentane mixture for 24 hours.
The parti cell ate matter from the water probe rinses was recovered,
dried and weighed.
Stack
The acetone probe rinses were filtered through 5y Teflon Millipore
filters and the filtered solids weighed. This weight was added to
the weight of the particulate matter on the filter for total mass
loading calculations. The filtered solids were then combined with
the particulate filters for acid extraction.
t Filters
Combustion Zone
The tared sample filters plus one control were desiccated and
weighed on consecutive days to a constant weight +_! mg, combined
with probe rinses, and then extracted in a Soxhlet apparatus for
24 hours with pentane. Solvent extracts were dried by being passed
through a column of pre-extracted sodium sulfate and were then con-
centrated to 10 ml for analysis. At that time the filters were
extracted with constant boiling agua regia for six hours. This acid
extract was made to 100 ml for analysis.
Stack
The tared sample filters were weighed as for the combustion zone
filters, combined with the solids filtered from the probe washes,
and extracted with constant boiling aqua regia for six hours. The
acid extracts were made to 100 ml for analysis.
• Solid Sorbent Traps
Combustion Zone
The XAD-2 resin samples were transferred from the traps to glass
Soxhlet thimbles and extracted in a Soxhlet apparatus with pentane
and methanol for 24 hours with each solvent. These extracts were
dried by being passed through a column of pre-extracted sodium
26
-------
sulfate and were then concentrated to 10 ml for analysis. One
unused trap was extracted for background values and a blank on the
solvent was also run. Another control was prepared by adding
0.1 mg of Aroclor 1242 to a trap and extracting as for the other
samples.
Stack
No solid sorbent traps were used in the stack sampling train.
• Grab Gas
Combustion Zone
No special preparation was required.
Stack
No grab gas samples were taken at the stack.
t Impingers
Combustion Zone and Stack
The volume of liquid in the impingers was measured and the spent
silica gel was weighed in the field after each test burn to
determine the amount of water collected. The aqueous liquid
impingers from the combustion zone were combined and 150 to 300 ml
aliquots acidified in the field to stabilize the metals for analysis.
The stack impinger samples were also acidified. No extractions or
other special preparation steps were performed on any of the
aqueous impinger samples.
The acetone rinses of the combustion zone impinger glassware were
combined and concentrated to a 10 ml volume for analysis.
• Scrubber Waters
1.5 liter aliquots of the scrubber water samples were extracted
for organics, according to the procedure for the separatory funnel
extraction for oil and grease from water recommended by the EPA
Handbook on Methods for Chemical Analyses of Water and Wastes, with
the substitution of pentane for Freon (National Environmental
Research Center, Cincinnati, Ohio 45268, EPA-626-/6-74-003). How-
ever, instead of evaporating the material to the dried residue, the
extracts were concentrated to a 10 milliliter sample by use of a
Kuderna-Danish concentrating evaporator. A control was also pre-
pared by adding 0.2 mg of Aroclor 1242 to 1.5 liter of the fresh
city water from Test II (PCB fluff). These scrubber water extracts
were all dried before their concentration by being passed through
a 200 x 10.5 mm glass column containing a 50 mm bed of sodium sul-
fate which had been pre-extracted with pentane in a Soxhlet for
24 hours.
27
-------
t Solid Combustion Residues
Approximately 100 g portions of the solid residue samples from
the rotary kiln were extracted in a Soxhlet apparatus for 24 hours
with pentane. The solvent extracts were dried and then concentrated
to 10 ml for analysis. A control sample was prepared by adding
0.1 mg of Aroclor 1242 to 100 g of the residue from Test II (PCB
fluff) and extracting with pentane in a Soxhlet.
To prepare the residue samples for inorganic characterization,
100 mg aliquots were dissolved by a combination of digestion in
a Teflon Parr bomb with aqua regia/HF, addition of boric acid,
filtration, and ash and fusion of the remaining solids with sodium
carbonate. The final sample volume was 100 ml. A reagent blank
was also prepared.
4.1.4.2 Analytical Methods
After extraction of the samples for organic material and other prepara-
tion for inorganic material, the concentrated extracts, and aqueous solu-
tions were analyzed by several methods which are summarized in Table 4-4.
A general treatment of the sample preparation and analytical procedures is
discussed below.
Organic Analyses. The concentrated solvent extracts of the filters,
sorbent traps, scrubber waters, and solid residue samples were analyzed by
gravimetry, IR, LRMS, and GC/MS techniques. An aliquot of each extract was
evaporated at ambient conditions to remove the solvent. The residue was
weighed and analyzed by IR and LRMS.
The IR and LRMS analyses yield qualitative information about the
classes or types of compounds (e.g., hydrocarbons, phenols, ROMs, PCBs,
etc.) present as well as an idea of the complexity of the concentrated
sample. Knowledge of the classes of compounds present provides a measure
of the toxicity, if any, of the residue. The detection limits for these
analytical techniques vary somewhat with the type of compound (see
Table 4-4).
The grab gas samples contained in the Tedlar®film bags were analyzed
by a mass spectrometer. A portion of the gas sample was vacuum transferred
into the inlet system of constant volume and measurable pressure.
Quantisation of organic compounds, specifically PCBs, was performed
by combined gas chromatography/mass spectrometry (GC/MS) using a Finnigan
system. Detailed procedures for the analysis of PCBs are included in
Appendix D.
Inorganic Analyses. Inorganic analyses were performed using inductively
coupled plasma optical emission spectrophotometry (ICPOES). Samples from
the acid extractions of the stack particulate filters, the acidified splits
of the impingers and scrubber waters, and the dissolved solid residues were
surveyed for trace metals by ICPOES. The ICPOES analysis determines
28
-------
TABLE 4-4. SUMMARY OF ANALYTICAL METHODS
Method
Instrument Manufacturer
and Model
Detectability for a Compound or
Element Being Sought
Organic Analyses
Gravimetry
Infrared
Spectrophotometry
(IR)
Low Resolution Mass
Spectrometry
(LRMS)
Combined Gas
Chromatography/Mass
Spectrometry (GC/MS)
Inorganic Analyses
Inductively Coupled
Plasma Optical Emission
Spectrophotometry
Spark Source
Mass Spectrophotography
(SSMS)
Mettler, microbalance
Perkin Elmer, 521
Hitachi-Perkin Elmer,
RMU-6 Mass
Spectrometer
Finnigan, 9500 GC and
Finnigan, 3100D Quadrapole
Mass Spectrometer
Applied Research
Laboratories, QA-137
AEI Scientific
Apparatus Ltd., MS 702R
1
^3-5% of the sample
being examined
MO ug
(1% of a 1 mg sample)
^100 ng per yl of sample
^0.5-2000 ppb
^50-100 ppb
-------
32 elements, including most of the toxic elements of interest to the
program, down to ppb levels with an accuracy of 10 to 100 percent.
Spark source mass spectrophotography (SSMS) was also used to analyze
the solid residue samples for a broad survey of trace elements. This
method is also sensitive to ppb levels and has an accuracy range of from
100 to 500 percent.
4.1.5 Problems Encountered
In spite of detailed planning and preparation for these tests, a few
incidents occurred in the field that had not been anticipated. Minor
problems were corrected immediately so the test schedule was not delayed.
However, three incidents took place which either prevented or postponed
testing; these are described in the following paragraphs. No problems were
encountered in the analytical portion of the tests.
4.1.5.1 Combustion Zone Sampling
An existing port in the hot duct was utilized as the sampling port
for the combustion zone, by welding an 8-cm nipple to the port to accommodate
the sampling probe. The immediate problem encountered was that the platform
to be used by the sampling personnel was approximately two meters below the
sampling port. Because of the windy weather conditions and the askew position
of the welded nipple, it was impossible to safely support and operate the
sampling probe and train. This problem was corrected by installing a uni-
strut monorail above the port and by adding an additional section of platform
so that the sampling port was then only one meter above the platform. These
corrective items caused a one-day delay in the testing.
4.1.5.2 Storm Conditions
A storm front was moving in as the purge for the background test on
No. 2 oil was begun. Conditions continued to worsen with high winds and
heavy rain. Members of the sampling crew were just starting up the stack
sampling train when lightning began striking tanks at nearby chemical
plants. For the safety of the sampling personnel, it was then decided to
forego acquiring a stack sample during the background test.
4.1.5.3 Line Plugging
During the purge before the scheduled hammermilled fluff test, the
line feeding the No. 2 oil to the kiln became plugged. The problem was
corrected but not until it was too late in the day to run the test burn,
so the test was postponed until the following day.
30
-------
4.2 ADL TESTS OF NITROCHLOROBENZENE WASTE
4.2.1 Waste Tested
The waste which ADL tested at Rollins Environmental Services, Inc.,
(Rollins) was a stream generated in the production of nitrochlorobenzene
(NCB). The principal use of the desired NCB product is as an intermediate
material for the manufacture of synthetic dyes, drugs, pesticides, and
photochemicals. The estimated total annual production of NCB waste similar
to that selected for testing was 9100 metric tons (20 million Ibs) in 1974.*
A survey sample of the NCB waste was received and analyzed by TRW, Inc.,
several months prior to testing. The survey analysis indicated that the
waste was a black liquid with some suspended crystalline material, a
specific gravity of 1.395 g/ml at 15.6°C, a higher heating value of 5050 Kcal/kg
(9100 Btu/lb), and a low ash content (<0.001%). The elemental analysis of
the sample was:
Weight Percent
Carbon 47.6
Nitrogen 8.7
Hydrogen 3.0
Chlorine 20.6
Sulfur trace
The major component of the waste was nitrochlorobenzene (mixture of isomers,
95% of total waste) with nitrobenzene as the next most abundant species.
Inorganic elements identified at ppm levels were calcium, titanium, potassium,
iron, antimony, silicon and phosphorus.
A second survey sample was obtained just prior to testing because
the waste supplier indicated that the production process had changed. That
sample was physically very different from the first. It contained much more
crystalline material which had agglomerated into a mass of pale brown crystals
which occupied about 50% of the volume of the sample container. The
remainder of the sample was a black liquid. The mass of solid material had
clearly formed in the sample container since it was several times larger than
the mouth of the bottle. There was not sufficient time to do an extensive
survey analysis on this second survey sample.
The waste actually tested was received at Rollins in a heated tank
truck and was expected to resemble the second survey sample. However, the
samples of representative waste feed drawn from the batch actually tested
did not contain any noticeable quantity of crystalline material even after
cooling. It is possible that the production process had been changed again,
or that the second survey sample was not representative of the new waste,
or that the waste within the tank truck was stratified.** (The crystalline
*Destructing Chemical Wastes in Commercial Scale Incinerators, Technical
Summary, Volume I, PB-257 709/6WP
The entire contents of the tank truck were not required for testing because
it was decided to blend the waste with diesel oil for destruction
(see Sec. 4.2.5). The "representative" sample represents the material
withdrawn from the truck for blending with the oil.
31
**
-------
material had been identified as p-chloronitrobenzene in the first survey
analysis, and it is more dense than the ortho-isomer.) Whatever the reason,
the waste material which was actually tested was found to be liquid at room
temperature. Final analysis of the representative waste sample showed that
it quite closely resembled the first survey sample.
The details of the analyses of the representative waste sample were
given in Appendix F, Section 4. The NCB waste itself was found to be about
95% by weight nitrochlorobenzene, primarily the ortho-isomer. Nitrobenzene
and dinitrobenzen^ were also present, as well as some compounds tentatively
identified as styrene-substituted di-tolylethers. The elemental analysis
results, similar to those expected for nitrochlorobenzene, were:
Height Percent
Carbon 46.14
Hydrogen 2.80
Nitrogen 8.92
Chlorine 23.28
Sulfur 0.02
The ash content of the waste was estimated as 0.8%. No trace elements
were found by SSMS at concentrations high enough to cause concern for
emissions of toxic metals at the feed rates used in the tests.
Analysis of a sample of the diesel oil - NCB waste blend actually fed
to the incinerator via the Loddby burner (see Section 4.2.5) confirmed that
this mixture was 20% by volume or 28% by weight of nitrochlorobenzene. The
blend.was found to contain excess chlorine (10.0% Cl by weight .total)
compared to the levels calculated from its NCB content (6.92% by weight Cl
calculated). No explanation for the apparent excess was determined, but
the higher value was confirmed in three separate analyses and must be
presumed to be real. Small amounts of PCBs (j< 0.1%) were found in the
blended waste - oil feed; these were found to be present in the #2 diesel
oil used at Rollins and not in the NCB waste itself. The ultimate source
of the PCBs was not traced (it may have been contamination of an oil
truck, storage tank or the waste blend tank). That there is only a remote
possibility of adverse environmental impact from incineration of these
quantities of PCBs can be seen by consulting other sections of this
report.
4.2.2 Operational Procedures
The preliminary detailed operating procedures including test plan and
safety plan were reviewed prior to the arrival of the ADL sampling team
on-site. Due to problems encountered in burning the NCB waste alone in
the kiln, however, (these problems are described in more detail in
Section 4.2.5.1) the NCB was dissolved in diesel oil and this solution
was fired through the Loddby burner. For the test where the NCB waste
alone was burned in the kiln (ADL Run R2), the kiln and afterburner were
32
-------
preheated and purged by burning diesel fuel (in the kiln and Loddby burner) for
one hour prior to the feeding of the NCB waste. For the tests where NCB
waste was mixed with diesel fuel and fed to the Loddby burner, there was no diesel
fuel purge of the kiln and Loddby burner prior to firing the waste, since there
was no means readily available for feeding straight diesel oil to the Loddby
burner. (The only available feed tank for the Loddby burner was filled with the
diesel fuel/NCB mixture.)
The operating procedure for the two test burns of the solution of
NCB waste in diesel oil (ADL Test Runs R3 and R4) was as follows:
• The feed to the incineration system was switched over from
commercial industrial waste being incinerated by Rollins to
diesel fuel to the kiln and NCB waste/diesel oil blend to
the Loddby burner (afterburner).
• The feed rates of oil and waste blend to the kiln and Loddby
burner were adjusted to give the desired temperatures in
the hot duct and afterburner.
• The on-line instruments were activated and test equipment
and sampling trains were checked and prepared for use.
• After the NCB waste/diesel oil blend had been fed to the
incinerator for at least one hour, the probes were inserted
into the hot duct and stack and sampling begun.
• After the conclusion of sampling at the hot zone and stack
for the day, the incineration system was switched over to
other commercial wastes being incinerated by Rollins.
4.2.3 Sampling Methods
Sampling methods used in the NCB tests at Rollins are described
briefly below.
Five distinct samples were taken during each waste test:
• Composite sample of waste feed material.
• Sample of combustion zone effluent fed to on-line
instruments for continuous monitoring of test.
t Grab sample of combustion zone effluent to evaluate
process effectiveness.
• Grab sample of stack gases to verify that test program was
environmentally acceptable.
• Sample of fresh and spent scrubber water.
33
-------
4.2.3.1 Waste Feed Sample
A composite sample of the blended waste-oil feed was obtained by
collecting a portion of the material in the waste feed tank during each
test. The two feed samples were blended to yield one representative sample.
4.2.3.2 On-line Gas Monitoring
A portion of the combustion effluent was sampled through a ceramic
(12.5 mm) probe.and passed through a heated Teflon line to a gas conditioning
system. The gas conditioner was designed to deliver a cool, dry, particulate-
free sample to-the CO, C02, and NOX analyzers. A fraction of the sample was
also supplied, untreated, to the hydrocarbon analyzer.
The instruments used and their ranges were:
Hydrocarbons Beckman
Model 402 0.05 ppm - 10%
Carbon Monoxide Beckman
Model 865 2-220 ppm
Carbon Dioxide Beckman
Model 864 0.05 - 20%
Oxygen Taylor
OA 273 0.05 - 100%
Nitrogen Oxides Thermo Electron
Model 10A 0.05 ppm -1%
In addition, portions of this stream were sampled with gas-detector
tubes.
4.2.3.3 Combustion Zone Sample
The train used for collecting this three-hour sample is shown
schematically in Figure 4-9. The principal components in this comprehensive
sampling train were:
• A 1.2 cm (0.5") quartz-lined sampling probe.
• A knockout trap consisting of an oversized impinger to
condense water.
• A quartz fiber filter.
• A sorbent trap filled with XAD-2^-' resin to collect
organics of moderate volatility.
• Impingers containing aqueous sodium acetate to collect
acidic gases. Sodium acetate, rather than sodium hydroxide,
was used in the impingers to trap HC1. This was done to
minimize the scrubbing of C0? that would tend to exhaust
the capacity of the impinger solution and could also cause
plugging problems by precipitation of sodium carbonate.
•"Trademark of Rohm and Haas Company.
34
-------
co
tn
To Control
Module
Silica Gel
Figure 4-9. Hot zone sampling train for NCB tests.
-------
In addition, a portion of the combustion zone effluent was collected
in gas sampling bulbs from the bypass line of the hydrocarbon analyzer.
4.2.3.4 Stack Gas Sample
The stack gas effluent was sampled isokinetically, according to the
EPA Method 5 procedure, along two perpendicular traverses. The train was
a typical EPA Method 5 type, the RAC Staksamplr.* The impingers contained
aqueous sodium acetate to trap hydrochloric acid.
4.2.3.5 Scrubber Hater
Samples of spent scrubber water were composited, using a proportional
sampling pump, over a three-hour period coincident with the combustion
zone gas grab sample. Samples of fresh scrubber water were taken in
1 liter bottles from a convenient tap.
4.2.4 Analysis Techniques
4.2.4.1 Extractions and Sample Preparation
General descriptions of the techniques used are in the Phase I Final
Report of this contract. A detailed description of the specific solvents,
etc., used for the Rollins NCB test samples is given in Appendix F.
4.2.4.2 Analytical Methods
The techniques which were chosen during Phase I of this contract for
evaluation of the effectiveness of thermal destruction of industrial
wastes were:
Low Resolution Mass Spectrometry (LRMS)
Infrared Spectrometry (IR)
Gas Chromatography
Elemental Analysis
Inorganic Analyses were done by:
X-ray Fluorescence (XRF)
Spark Source Mass Spectrometry (SSMS)
Atomic Absorption Spectroscopy (AAS)
Specific Ion Electrode Methods (SIE)
These techniques were applied to the Rollins NCB samples where appropriate.
Details of analytical procedures are given in Appendix F.
trademark of Research Appliance Corporation.
36
-------
4.2.5 Problems Encountered
4.2.5.1 Process
In the first test burn of NCB waste (Test No. R2), the NCB waste
alone was fed under pressure from a heated tank truck directly to the
kiln and diesel fuel was fed to the Loddby burner to maintain temperature
in the afterburner chamber. The NCB waste by itself would not atomize
properly (in the air atomization nozzle in the kiln) and burned with
considerable generation of carbon soot. The carbon soot was subsequently
incinerated in the afterburner, but the NCB feed rate was limited to about
190 liters (50 gallons) per hour. Because of the poor feed rate and
combustion of NCB in the kiln, it was decided to dissolve the NCB in diesel
fuel and fire this solution through the Loddby burner which has more
capacity than the kiln.
Approximately 20% by volume of NCB waste was blended with No. 2 diesel
fuel in an agitated tank that feeds the Loddby burner. A laboratory test
at Rollins indicated that up to 50% (by volume) of the NCB waste could be
dissolved in the diesel oil. However, since the feed tank for the Loddby burner
was about 80% full with diesel oil, enough NCB waste was added to fill the
feed tank and give a 20% (vol.) solution of NCB in diesel oil.
4.2.5.2 Sampling
The water pressure at the elevation of the hot zone probe was
insufficient to supply adequate cooling water to the jacket of the hot
zone probe. A booster pump was used to increase the water pressure, but
the small diameter of the water lines to the jacket of the probe created
so much back pressure at the water flow rate necessary to cool the probe
(with 2-1/2 ft. of probe in the hot duct), that the pump motor would
overheat and shut down. It was, therefore, necessary to back the probe
out of the duct (so only about 1 foot remained in the duct) allowing
the amount of cooling water to be reduced, thus preventing overheating
of the pump.
4.2.5.3 On-line Instruments
The oxygen analyzer failed when it was turned on the day before the
R3 test. Trouble-shooting, after telephone consultation with the manufacturer's
service representative, identified a defective analyzer cell as the most
probable source of the symptoms observed. This could not be field repaired,
and a replacement could not be obtained in time for completion of testing.
For the R3 and R4 tests, the oxygen content in the combustion zone gas was
estimated from the measured carbon dioxide concentration with the assumption
that the sum of oxygen and carbon dioxide concentrations would be the same
as for the background test.
The chemiluminescent NO/NOX analyzer appeared to be ineffective in the
NO- to NO conversion mode. The instrument consistently gave the same
concentration reading for the hot zone combustion gas whether in the NO mode
37
-------
or the NOX mode. Nitrogen dioxide was an expected product of combustion
from the nitrochlorobenzene waste and, in fact, tests with gas detecting
tubes which were on hand in the sampling trailer indicated more than
300 ppm of N02 in the combustion zone effluent. Telephone consultation
with the manufacturer's service representative failed to produce an
explanation for this phenomenon or to identify a remedy.
38
-------
5. TEST RESULTS
As described in the previous section, the test burns at Rollins
consisted of tests conducted by TRW on PCB-containing capacitors and tests
conducted by ADL on nitrochlorobenzene manufacturing waste. The results
obtained from these experiments are presented here in the following terms:
t Operating conditions and other data obtained in the field.
• Data from the analysis of the combustion gases for organic and
inorganic species, particularly as relates to destruction of the
waste This will include a discussion of destruction efficiency
for the ADL tests which had no solid residues left in the
combustion zone.
• Data from analyses of the final emissions (i.e., stack gases,
scrubber waters, and any solid residues from the kiln).
5.1 RESULTS OF PCB TESTS
5.1.1 Operating Conditions for Tests
The data presented in this section were collected from the operation
of:
0 Rollins rotary kiln incinerator facility, and
• TRW on-line gas composition monitors.
All recorded data for the incinerator system operating conditions
were provided by Rollins and are summarized in Table 5-1. TemperatureS
in the rotary kiln system stayed fairly constant, with temperatures in
the afterburner, for example, varying over a range of not more than
±40°C during any of the three tests.
Readings from all on-line gas monitors except the SOX analyzer were
continuously recorded on strip charts. The resulting scans and readings
were averaged over the 2-3 hour long test runs. Concentration values
obtained for the dry combustion gas are shown in Table 5-2 HC1 concen-
trations were determined by titration of the hot zone sampling train
impingers.
Percent excess air was calculated according to the equation in EPA
Method 3 (Reference 3) and was found to be 90, 84, and 91 percent for the
resoective tests. Percent moisture was calculated according to EPA
Method 4 and was found to be 7.1, 6.5, and 8.5 percent in the combustion
gas for the respective tests.
39
-------
TABLE 5-1. INCINERATOR SYSTEM PARAMETERS DATA SUMMARY
_
Flow Rates
No. 2 011 to the K1ln
(Uters/min)
No. 2 011 to the Lodd-
by (Uters/min)
Natural Gas (m3/m1n)
Waste (kg/hr)
Fresh Scrubber Water
Feed (I1ters/m1n)
L1me Slurry Feed
(I1ters/m1n)c
Temperatures
Kiln Flame (°C)
K1ln Duct (°C)
Loddby Flame (°C)
Afterburner (°C)
Hot Duct (°C)
Venturl Pressure Drop
(cm H20)
Calculated Residence
Time (sec)
Test I
No. 2 011
Background
34. 8a
0.989
3200
6.4
1306
373
1485
1308
1091
104
-
Test II
Hammermilled
Capacitors
4.48
35.7
1.51
21 Ob
3200
6.4
1252
488
1499
1331
1089
102
3.2
Test III
Whole
Capacitors
9.34
29 0
™ -f • V
1 46
1 • ~ W
360
3200
8.4
1339
493
1509
1332
1096
102
3.0
a)
b)
Only the total flow could be measured for this test. Subsequent
changes separated the two feed systems, thus feed rates could
also be measured separately.
This is the feed rate of the fluff material alone. The feed rate
including the weight of the fiber drums was 330 kg/hr.
c) ^30% solution by weight.
40
-------
TABLE 5-2. GAS COMPOSITION DATA SUMMARY
Test
No.
I
II
III
°2
(percent)
10.1
9.8
10.1
C02
(percent)
9.1
8.8
9.3
N2a
(percent)
80.8
81.3
80.5
CO
(ppm)
15
5-10
5-10
NOX
(ppm)
55
60
60
S02
(ppm)
35
40
35
HC1
(ppm)
-
470
670
HCb
(ppm)
<20
<20
<30
By difference
As methane
Attempts were also made to use Gastec® tubes to detect hydrocarbon
species at the stack. However, because of the high moisture content of
the stack gases and low ambient temperatures, the detection tubes became
saturated with condensed water, so accurate readings could not be
obtained. The condensation problem has been discussed in previous
reports (Reference 2).
5.1.2 Composition of Combustion Zone Gas
Using the sampling train described in Section 4.1.3.2, samples of
the combustion products were taken from the hot duct leading from the
afterburner to the venturi scrubber. These samples were then separated
into their organic and inorganic constituents and analyzed by appropriate
techniques. Analysis of the combustion products was aimed mainly at
identifying and quantifying any unburned waste material or hazardous
partial combustion products. The production of potentially toxic levels
of trace metals from burning these wastes was also examined. Where
quantified species are calculated back to mg/m3 in the sample gas
stream, the gas volume data used is summarized in Appendix B.
5.1.2.1 Organic Composition
Quantitation for Specific Compounds
The sample extracts were analyzed for specific toxic compounds found
in the survey waste sample, i.e., PCB isomers, using a combined gas
chromatography/mass spectrometer (GC/MS) method. The concentration level
of interest for this program is defined as 0.1 mg/m3 of sample gas, the
threshold level of nearly all of the most hazardous species described in
OSHA and other health and safety documents. However, currently there is
much interest being given to PCB compounds, and it is known that they are
considered hazardous at lower levels. Therefore, efforts were made to
lower the detection levels for these specific compounds.
41
-------
Two samples for each test were analyzed to determine PCBs in the
combustion zone gas: 1) extracts of the participate filter and probe
wash and 2) extracts of the solid sorbent traps. No PCBs of any class
were found in any of these samples. The minimum detectable quantities
for the analysis were ~10 micrograms in the total 10 ml concentrated
sample solution for each of the PCB classes. These minimum detectable
quantities, when related to the average combustion zone sample gas
volume of 3.9 m3, represent about 5 micrograms per cubic meter (or
0.005 mg/mj) for the filter and sorbent trap samples combined.
A recovery study was conducted by adding 0.100 mg of Aroclor 1242 to
"» unused XAD-2 sorbent trap as a doped control. The doped trap was then
extracted and the extract treated along with the test sample extracts.
Analysis of the doped control extract by GC/MS found 0.2 mg at total PCBs,
indicating a recovery of 200 percent. This difference between added PCBs
and recovered PCBs was within the expected accuracy range for PCBs at
these low levels.
Qualitative Surveys
Combustion zone samples were also surveyed by gravimetric, infrared
spectrometry (IR), and low resolution mass spectrometry (LRMS) techniques
The qualitative results are discussed in the following order:
• Combined particulate filter and probe wash extracts
• Sorbent trap extracts
• Acetone impinger rinses
• Grab gas samples
Combined Probe Wash and Particulate Filter Extracts. The types of
organic material discovered in this survey are typical of those compounds
found in most of the other organic residues obtained in the analyses from
these tests. Hydrocarbon oil, dioctyl phthlate, and silicone were the
predominant species present in these samples. In addition, traces of
palmitic acid, stearic acid, substituted aromatics, and an unknown
substance with a molecular weight of 382 were identified. Table 5-3
shows the amounts of the various materials that were present. The values
have been corrected for those obtained from blanks and controls which
were considerably lower.
Sorbent Traps. The amounts of material extracted from the sorbent
traps and found as a residue after mild evaporation are presented in
Table 5-4. The quantities have been corrected for the unused, control
sorbent trap extract. The types of matter found in the, trap extracts by
IR and LRMS analyses were essentially the same for all samples. The
major compound found in all three test samples was silicone, with minor
amounts of hydrocarbon oil and substituted aromatics. In addition,
traces of a low molecular weight amine and a glycol ether were present
in the extract samples from Tests I and II, respectively.
42
-------
TABLE 5-3. SUMMARY OF SURVEY ANALYSIS ON THE COMBINED
PROBE WASH AND PARTICULATE FILTER EXTRACTS
Test No.
I
II
III
Amount of
Material Found
as Residue3
(mg)
15.4
38.2
76.0
Volume of
Sampled Gasb
(m3)
3.65
3.88
3.88
Concentration
in Sample
Gas
(mg/m3)
4.21
9.84
19.6
a Corrected for blank on extraction glassware and solvent.
Includes water vapor.
TABLE 5-4. SUMMARY OF SURVEY ANALYSIS ON SORBENT TRAP EXTRACTS
Test No.
I
II
III
Doped
Control0
Material
Extracted by
Pentane
(mg)a
34.1
16.9
12.8
1.23
Volume of.
Sample Gas
(m3)
3.65
3.88
3.88
-
Concentration
of Extractables
in Sample Gas
(mg/m3)
9.3
4.4
3.3
-
a Corrected for blank control sorbent trap extract weight.
Includes water vapor.
c An unused trap with 0.100 mg of Aroclor 1242 added.
The primary purpose of preparing the doped control sample was to
provide recovery data for the quantitative organic analysis. Thus, the
residue obtained from this sample was not qualitatively analyzed beyond
the gravimetric determination.
43
-------
Acetone Impinqer Rinses. As an added check for PCBs on the
combustion zone sampling train, the impinger glassware was rinsed with
acetone at the end of each test. For maximum sensitivity the rinses
were combined, concentrated, and qualitatively examined for PCBs. The
IR and LRMS analyses of this sample found only silicone and dioctyl
phthlate. There was no indication of PCBs.
Grab Gas Samples. The contents of the Tedlar® gas sampling bags
were analyzed by introducing a portion of the gas into the mass spectro-
meter and measuring its pressure at constant volume. Low molecular weight
hydrocarbons were detected at ppm levels, but no PCBs were disclosed.
5.1.2.2 Inorganic Characterization
Inorganic elemental concentrations were determined by analysis of the
particulate filters and aqueous impinger samples. Figure 5-1 shows a
photograph of the particulate filters obtained from sampling the combustion
gases from Tests I through III (from left to right in the top row). Trace
metals on the particulate filters were put into solution by acid digestion
of the filters. The filter digests and the impinger solutions were then
surveyed by ICPOES to determine the major elements. The results of this
analysis are shown in Table 5-5. The major elements found in these samples
were boron, cadmium, copper, iron, potassium, phosphorus, lead, and tin.
Major levels of aluminum may also have been present; however the high
background of this element in the glass fiber filters makes that impossible
to determine. Other components which were analyzed by the ICPOES study
but were undetected in combustion zone samples are shown in Table 5-6.
5.1.3 Final Emissions
Emissions from the Rollins rotary kiln process were sampled and
analyzed to evaluate the environmental safety of the waste burns. All of
the final process effluents were sampled; these included stack gas,
scrubber water, and solid residue.
5.1.3.1 Stack Gas
Stack effluents were sampled during the tests with a standard EPA
Method 5 train. The samples obtained were analyzed to determine particu-
late loading in the effluent gas and elemental composition of the
particulate matter. Figure 5-1 shows the particulate filters obtained
from sampling the stack gas from Tests II and III (from left to right in
the bottom row). No stack sample was procured from Test I for the
reasons discussed in Section 4.1.5.
Particulate loading was determined by adding the weight gain on the
filters to the weight of residue in the probe washes. This total was
then divided by the dry sample gas volume. The loading values acquired
are listed in Table 5-7.
44
-------
Figure 5-1. Filters from combustion zone and stack
gas sampling trains.
-------
TABLE 5-5. SURVEY FOR TRACE METALS IN COMBUSTION ZONE SAMPLES BY ICPOES
Element
Al
Ag
B
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
P
Pb
Si
Sn
Sr
Ti
V
Zn
Concentration in the Sample Gas (ng./m )
Test I
Filter
<0.91a
0.001
<0.46
<0.19
ND(<0. 00003)
<0.88
0.001
0.001
0.01
0.036
0.31
ND(<0.06)
<0.26
fO.003
0.003
<3.6
0.010
0.093
0.36
<0.38
ND(<0.002)
<0.008
0.016
0.003
<0.12
Inpingers
0.014
<0.003
0.11
ND(<0.036)
<0.0002
<0.036
ND(<0.001)
ND(<0.002)
0.005
0.002
<0.048
ND(<0.22)
ND(<0. 00003)
0.0005
0.005
_b
ND(-O.OOl)
ND(-0.054)
ND(<0.010)
2.6
ND(<0.005)
ND(<0.0001)
0.006
0.0003
0.002
Total
<0.9
< 0.004
<0.6
;0.2
<0.0002
<0.9
<0.002
<0.003
0.02
0.04
<0.4
ND(<0.3)
<0.3
;0.004
0.008
<4
0.01
<0.1
0.4
3
ND(<0.007)
<0.008
0.02
0.003
<0.1
Test II
Filter
<54
0.006
<0.66
<0.075
<0.0003
<1.2
0.18
0.002
0.025
1.2
10
1.1
<0.29
<0.084
0.013
<8.0
0.033
0.98
12
<0.24
0.89
<0.006
0.33
0.11
<2.3
Impingers
ND(<0.003)
ND(< 0.00003)
36
ND(<0.030)
ND(<0. 00001)
<0.007
N0(<0.001)
ND(<0.002)
ND(<0.0006)
ND(<0. 00005)
<0.012
ND(-0.18)
ND(<0. 00003)
ND(<0. 00009)
0.17
_b
ND (--0.0009)
ND(<0.046)
ND(<0.008)
0.67
0.083
ND(<0. 00008)
<0.001
ND(<0. 00006)
0.002
Total
_.50
0.006
• 4
_-0.1
-_ 0.000 3
<1
0.2
<0.004
0.03
1
10
1
<0.3
<0.08
0.2
<8
0.03
1
10
<0.9
1
<0.006
0.3
0.1
<2
Test 111
Fi 1 ter
<7.3
0.018
<0.12
-0.065
ND (-0.00003)
<0.'54
0.70
0.002
0.018
8.3
12
0.85
<0.16
<0.043
0.014
<3.7
0.019
1.7
11
<0.38
2.6
<0.003
0.022
0.011
<0.37
Impingers
ND(<0.004)
<0.001
0.14
ND(<0.036)
^0.0001
<0.009
ND(<0.001)
ND(<0.002)
0.002
0.001
<0.005
ND(<0.22)
ND(<0. 00003)
ND(<0. 00009)
0.059
_b
ND(<0.001)
0.11
ND(<0.010)
0.60
ND(<0.005)
N0(<0.0001)
<0.002
ND(<0. 00008)
0.002
Total
<7
0.02
<0.3
<0.1
<0.0001
<0.5
0.7
<0.004
0.02
8
10
1
<0.2
<0.04
0.07
<4
0.02
2
10
<1
3
<0.003
0.02
0.01
<0.4
-Pa
CTl
"<", a less than or equal to sign indicates those elements which were detected but not significantly above background levels.
Not appropriate because NaOH was added to these impingers.
-------
TABLE 5-6. DETECTION LIMITS FOR ELEMENTS NOT FOUND IN THE
COMBUSTION ZONE SAMPLES BY ICPOES
Element
Au
As
Eu
Se
Te
U
W
o
Average Detection Limit (mg/m )
Filter
0.0002
0.001
0.0004
0.002
0.002
0.002
0.003
Impingers
0.0005
0.004
0.002
0.006
0.007
0.008
0.009
Total
0.0007
0.005
0.002
0.008
0.009
0.01
0.01
TABLE 5-7. PARTICULATE LOADING IN THE EFFLUENT GAS
Test
II-"Fluff"
Ill-Whole Capacitors
Weight
on
Filter
(nig)
44.5
33.6
Weight
in Probe
Wash
(mg)
9.0
21.5
Total
Weight
(mg)
53.5
55.1
Sample Gas
Volume, Dry
(m3)
1.55
1.04
Participate
Loading
mg/m
35
53
Grains/scf
0.015
0.022
After weighing, the filters were acid digested and along with the
impinger solutions were surveyed for trace metals by inductively coupled
argon plasma optical emission spectroscopy (ICPOES). The results of this
survey are shown in Table 5-8. The predominant elements found in the
combustion gas were also the major inorganic constituents of the stack gas.
Copper, iron, lead, phosphorus, tin, and cadmium were the major elements
in the stack, out of which only lead and cadmium are considered to be
significantly toxic. Of the 32 elements that were determined by the
47
-------
TABLE 5-8. SURVEY FOR TRACE METALS IN STACK SAMPLES BY ICPOES
Element
Al
Ag
B
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Mg
Mn
Mo
Na
Ni
P
Pb
Si
Sn
Sr
Ti
V
Zn
Concentration in the Stack Gas (mg/m )
Test II
Filter
<0.16a
0.003
<0.081
<0.021
;0. 00003
;0.66
0.14
ND(<0.001)
0.012
0.21
1.6
;0.26
<0.006
0.019
<5.9
0.019
0.17
2.7
<0.41
0.12
_0.003
0.007
0.006
_2.1
1st Jmpinger
0.018
0.003
0.58
ND(<0.013)
0.00008
ND(<0.001)
ND(<0.003)
ND(<0.004)
ND(<0.002)
ND(<0.0001)
<0.024
ND(<0. 00008)
ND(<0.0003)
0.016
ND(<0.16)
ND(<0.003)
ND(<0.13)
ND(<0.024)
0.22
ND(-0.013)
ND{- 0.0002)
0.002
: o.ooos
0.005
2nd Impinger
0.009
_b
0.11
0.15-
ND(<0. 00002)
0.023
ND(<0.002)
0.021
ND(<0.001)
0.002
<0.014
ND(<0. 00005)
ND(<0.0002)
ND(<0.002)
0.52
ND(<0.002)
ND(<0.087)
ND(<0.016)
0.70
ND(<0.009)
ND(<0.0002)
0.002
ND(-O.OOOl)
ND(-0.0004)
Total
<0.2
0.006
<0.8
<0.2
fO.OOOl
<0.7
0.1
<0.03
0.01
0.2
2
<0.3
<0.006
0.04
<7
0.02
;0.4
3
<1
0.1
fO.003
0.01
-0.007
_2
Test III
Filter
<0.44
0.010
<0.30
<0.26
<0. 00004
<0.88
0.44
ND(<0.002)
0.034
7.5
5.5
<0.29
<0.02
0.16
<6.3
0.059
1.9
3.3
:i-7
0.72
;0.011
0.27
0.009
_0.95
1st Impinger
ND(<0.013)
0.011
0.52
HD(<0.020)
0.0002
ND(<0.002)
ND('0.004)
ND(<0.006)
ND(<0.002)
0.014
<0.012
ND(<0.0001)
ND('G.0004)
0.079
0.79
ND(<0.004)
ND('0.20)
ND(<0.035)
0.27
ND(<0.020)
ND( 0.0004)
0.002
_0.0012
0.008
2nd Impinger
0.011
_b
0.043
ND(<0.008)
NO (<0. 00002)
0.055
0.003
0.020
ND(<0.001)
0.012
fO.016
ND(<0. 00005)
ND(<0.0002)
ND(<0.002)
0.92
ND{<0.002)
ND(<0.077)
ND(<0.014)
O.GO
ND(-O.OOS)
ND(--O.OOOl)
0.003
ND(-O.OOOl)
0.003
Total
<0.5
0.02
<0.9
<0.3
0.0002
<0.9
0,4
<0.03
0.04
8
6
<0.3
;0.02
0.2
<8
0.06
2
3
<3
0.7
:o.oi
0.3
_0.01
J
00
";", a less than or equal to sign indicates those elements which were detected but not significantly above background levels.
Not appropriate because AgN03 was added to this impinger.
-------
ICPOES analysis, eight were not detected in the stack samples. These eight
components are listed in Table 5-9, with a calculation of the average
detectable limit for e*_.'. in the flue gas.
5.1.3.2 Scrubber Water
The Rollins scrubbing system consisted of an essentially single
pass, wet scrubber. Samples were obtained for each of the three tests
of both the well water feed (FSW) and the spent water (SSW) exiting the
system. Aliquots of these samples were both solvent extracted for
analysis of organics and acidified to stabilize inorganics for analysis.
Comparison of the analytical results for the FSW and SSW samples yields an
estimate of how much, if any, hazardous species were added to the scrubber
water during each test.
Organic Composition
As with other samples for analysis of organics, the scrubber water
extracts were quantitatively analyzed for specific compounds and qualita-
tively surveyed for overall composition.
Quantitative Results. The solvent extracts of the scrubber water
samples were concentrated and examined for PCBs by GC/MS. One city water
input and all three effluent water samples were analyzed. No PCBs of any
class were found in any of the samples at a minimum detection limit of
0.007 mg/liter.
TABLE 5-9. DETECTION LIMITS FOR ELEMENTS NOT FOUND IN THE
STACK SAMPLES BY ICPOES
Element
Au
As
Eu
K
Se
Te
U
W
o
Average Detection Limit (mg/m )
Filter
0.0004
0.003
0.001
0.2
0.005
0.005
0.006
0.007
Impingers
First
0.002
0.01
0.005
0.7
0.02
0.02
0.03
0.03
Second
0.0008
0.007
0.002
0.3
0.01
0.01
0.01
0.01
Total
0.003
0.02
0.008
1
0.04
0.04
0.05
0.05
49
-------
A recovery study was also conducted wherein 0.200 mg of Aroclor 1242
was added to 1.5 liters of the input well water sample from Test II.
The doped sample was then extracted and the extract treated along with the
test sample extracts. GC/MS analysis of the doped sample found 0.13 mg of
total PCBs, yielding a recovery of 65 percent. This indicates that the
combined effect of recovery efficiency and analytical limits was adequate
to detect PCBs down to a level of 0.01 mg/liter.
Qualitative Survey Data. Aliquots of the solvent extracts were
evaporated at ambient conditions. The residue obtained from this step
was then weighed and analyzed by IR and LRMS techniques. The concentra-
tion of extractable species in the scrubber waters, as calculated from
the extract residue weights, is presented in Table 5-10. The data from
the IR analysis indicate that silicone was the major component of the
residues. The residues from these samples were generally quite low.
The doped control sample was prepared primarily for recovery data
for quantitative analyses. For this reason the residue from this sample
was not qualitatively analyzed beyond obtaining the gravimetric value.
Inorganic Characterization
The acidified aliquots of the scrubber water samples were surveyed by
ICPOES to determine the major elements present. The results of this analysis
are shown in Table 5-11. Substantial increases were observed for aluminum,
copper, lead, and zinc compared to the concentrations in both the input city
water and the effluent from the background test (Test I). Elements not
detected by the ICPOES investigation in any of the scrubber water samples
are listed in Table 5-12 along with their detection limits.
TABLE 5-10. SUMMARY OF SURVEY ANALYSIS OF
SCRUBBER WATER EXTRACTS
Test and
Sample
Identification
I - FSW
SSW
II - SSW
III - SSW
Doped Control9
Volume of Water
Extracted
(liters)
1.5
1.5
1.5
1.5
1.5
Weight of Residue
in Extract
(mg)
0.24
0.29
0.28
0.35
1.07
Concentration
in Scrubber
Sample
(mg/liter)
0.16
0.19
0.19
0.23
-
Water with a known amount of Aroclor 1242 added.
50
-------
TABLE 5-11. SURVEY FOR TRACE METALS IN SCRUBBER WATER SAMPLES BY ICPOES
Element
Al
B
Ba
Be
Ca
Cu
Fe
K
Mg
Mn
Na
N1
P
Pb
Si
Sr
T1
V
Zn
Concentration in Scrubber Water (mg/liter)
FSW
Test II
ND(<0.04)
0.15
ND(<0.05)
ND(<0.001)
23
0.007
0.065
30
5.0
0.055
99
0.05
1.1
ND(<0.09)
11
0.20
0.010
0.007
0.05
Test III
0.09
0.16
ND(<0.05)
ND(<0.001)
24
0.010
0.10
30
5.1
0.055
100
0.05
ND(<0.5)
ND(<0.09)
12
0.20
0.009
0.005
0.03
ssw
Test I
1.7
0.20
0.23
0.013
860
0.089
2.4
130
12
0.16
120
0.12
2.1
ND(<0.09)
17
0.72
0.063
0.042
0.16
Test II
3.0
1.6
0.14
0.003
310
0.15
0.58
60
6.4
0.048
110
0.04
.1.4
1.7
12
0.40
0.020
0.028
1.5
Test III
4.7
0.24
ND(<0.05)
0.003
600
7.3
1.7
100
8.5
0.10
120
0.05
3.0
4.0
15
0.58
0.031
0.025
1.3
-------
TABLE 5-12.
DETECTION LIMITS FOR ELEMENTS NOT FOUND IN THE
SCRUBBER WATERS BY ICPOES
Element
Ag
As
Au
Cd
Co
Cr
Eu
Detection Limit (ppb)
0.3
40
5
11
16
6
15
Element
Mo
Se
Sn
Te
U
W
Detection Limit (ppb)
11
60
50
65
80
90
5.1.3.3 Solid Residue
The objective of analyzing the solid residue samples taken at the
end of the tests was to determine whether any residual hazardous materials
were present which would affect disposal methods or which would in them-
selves constitute an environmental hazard. Figure 5-2 presents a photograph
of the two solid residue samples. They were quite similar in physical appear-
ance except that the Test II sample was slightly reddish in color.
Organic Composition
Portions of the residue samples were extracted in a Soxhlet apparatus
with pentane. The resulting extracts were then analyzed by the quantitative
and qualitative methods described previously.
Quantitative Data. The results of the GC/MS analysis for PCB compounds
in the solid" residue extracts are presented in Table 5-13. PCBs were found
in the Test III residue (whole capacitor feed) at a level of 470 mg/kg total
PCBs, but were undetected in the Test II residue (hammermilled fluff feed).
The lower limit of detection for PCBs for this study corresponded to
0.1 mg/kg or 0.1 ppm by weight. The distribution of PCBs in the Test III
sample over the five classes found is significantly different from that
of the waste feed material (see Table 4-2). The differing profiles
indicate a much higher destruction efficiency for the lower chlorinated
PCBs and a decreasing efficiency as the number of chlorine atoms per PCB
molecule increases.
A doped control was also prepared by adding 0.100 mg of Aroclor 1242
to M).l kg of the residue from Test II. This doped sample was then
extracted and the extract treated in the same manner as the test samples.
Analysis of the doped sample extract did not detect the presence of any
PCBs, indicating a recovery efficiency of less than ten percent. Thus
the extracted PCB quantity for the Test III sample could actually be as
much as 4,700 ppm (M).5 percent) or higher. The detectable limit in the
Test II sample is probably closer to 1 to 10 ppm.
52
-------
OJ
-
.
Figure 5-2. Photograph of solid residue samples
-------
TABLE 5-13.
RESULTS OF TEST III SOLID RESIDUE
ANALYSIS FOR PCBs BY GC/MS
Compound
Estimated Concentration
mg PCB/kg Residue
Percent of Total PCB
Compounds
Monochlorobi phenyl
Dichlorobiphenyl
Trichlorobiphenyl
Tetrachlorobiphenyl
Pentachlorobiphenyl
Hexachlorobiphenyl
NDC
8
109
300
52
4
2
23
63
11
1
ND = less than 0.1 mg/kg
Qualitative Survey Data. Evaporated aliquots of the solvent extracts
were weighed, and the gravimetric results are shown in Table 5-14. Analysis
by IR of the test sample residues from Test II found only silicone and
ketone compounds. Major amounts of PCBs were found in the Test III
samples. LRMS analysis of these samples confirmed the presence of PCBs in
the Test III residue, as well as indicating their existence in the Test II
residue. Since the LRMS method is more sensitive than the GC/MS analysis,
this result denotes that PCBs are present in the Test II residue at levels
below the GC/MS detection limits, or <0.1 ppm.
The doped control sample was prepared to determine the recovery of
PCBs from the solid residue material for the quantitative analyses. Thus
this sample was not qualitatively ana'lyzed beyond the gravimetric
determination.
TABLE 5-14. SUMMARY OF SURVEY ANALYSIS
OF SOLID RESIDUE EXTRACTS
Test No.
II-"Fluff"
Ill-Whole Capacitors
Doped Control3
Amount
of Sol ids
Extracted
(kg)
0.102
0.111
0.110
Amount of
Organic Residue
in Extracts
(mg)
1.86
44.7
1.23
Concentration of
Extractables
(mg/kg)
18
404
Solid residue from Test II with O.lCmgof Aroclor 1242 added.
54
-------
Inorganic Characterization
Approximately 100 mg aliquots of the solid residues from Tests II and
IHwere completely dissolved for inorganic analysis. Major elements were
determined within an accuracy factor of 100 percent by ICPOES, and a complete
trace element survey was performed by SSMS within an accuracy factor of 500
percent. The results of the ICPOES and SSMS analyses are presented in
Tables 5-14 and 5-15, respectively. Considering the relative accuracies of
ICPOES and SSMS, the results of the two analyses agree generally well. The
value for silicon determined by SSMS was lower than that determined by ICPOES.
Silicon was probably lost as a volatile fluorosilicate during the preparation
of the samples for the SSMS analysis.
The major elements composing the residues were found to be aluminum,
iron, silicon, copper, tin, phosphorus, lead, zinc, and possibly potassium,
all of which are the same constituents found to have predominated in
samples from other areas of the kiln system.
55
-------
TABLE 5-15. ANALYSIS OF SOLID RESIDUES BY ICPOES
Element
Al
Ag
As
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
K
Concentration (mg/g)
Test II
360
0.4
ND(<0.1)
0.2
ND(<0.001)
3
0.04
<0.1
0.7
50
300
<180
Test III
180
0.5
0.9
1
0.002
6
0.07
<0.1
0.7
43
400
<160
Element
Mg
Mn
Mo
Ni
P
Pb
Si
Sn
Sr
Ti
V
Zn
Concentration (mg/g)
Test II
0.7
0.9
<0.6a
0.3
17
7
85
13
0.01
0.9
0.1
2
Test III
1
1
<0.5
0.4
13
4
100
7
0.02
1
0.1
5
a) The less than or equal to values indicate that the sample results
were equal to or lower than the reagent blank.
56
-------
TABLE 5-16. ANALYSIS OF SOLID RESIDUES BY SSMS
Element
Pb
Pt
W
Pr
Ce
La
Ba
I
Sb
Sn
Cd
Ag
Mo
Zr
Y
Sr
Br
Se
As
Ge
Z
Concentration (mg/g)
Test II
4
0.3
<0.3a
0.02
0.06
0.06
0.3
0.02
<0.01
6
0.1
10.08
0.07
0.1
0.02
10.07
£0.09
0.08
I3
0.008
<2
Test III
3
0.5
ll
0.02
0.05
0.05
3
0.01
10.07
3
0.09
<1
0.1
0.2
0.009
10.06
<0.4
0.1
I5
0.03
I8
Element
Cu
Ni
Co
Fe
Mn
Cr
V
Ti
Sc
Ca
K
Cl
S
P
Si
Al
Mg
F
Li
Ga
Concentration (mg/g)
Test II
MCb
<0.6
<0.02
MC
2
ll
0.1
<1
£0.004
MC
MC
MC
<1
MC
<1
MC
MC
MC
10.02
0.2
Test III
MC
<1
10.09
MC
1
10.09
0.09
ll
10.007
MC
MC
MC
10.9
MC
10.9
MC
MC
MC
10.004
0.1
a) "i", a less than or equal to sign indicates those elements
which were detected but not significantly above background
levels.
b) Major Constituent -present above quantifiable levels.
57
-------
5.2 RESULTS OF NCB TESTS
5.2.1 Operating Conditions for Tests
Table 5-17 presents the operating conditions for the fuel oil background
burn, R1(B), and the two test burns of NCB waste in diesel oil. More details
of the operating conditions can be found in Appendix G. The major difference
between Tests R3 and R4 was the temperature in the hot duct.
5.2.2 Destruction Efficiency a^nd Composition of Combustion
Zone Effluent Gas
Data on the composition of the combustion zone effluent were obtained
from a variety of samples and types of analyses. The analyses, which included
both quantitative and qualitative characterizations of the effluent and the
analytical results, are described in detail in Appendix F. In this section
of the report the analytical results are presented in a reduced form which
facilitates interpretation of the data.
5.2.2.1 Quantitative Characterization
The principal criteria for assessing the effectiveness of the incineration
process for treatment of the waste were the calculated destruction efficiencies
(DE's). The calculations of DE were based on the quantities of total organics
and of nitrochlorobenzene found in the organic solvent extracts of the
various hot zone sampling train components. Those data for the background
test and two NCB tests are presented in Table 5-18. In estimating the total
quantity of organic material collected, the estimates obtained by gas
chromatographic analysis (of the unconcentrated extracts) and by gravimetric
analysis (of an aliquot of concentrated extract evaporated to dryness) have
been summed. This was done because other work in ADL laboratories has shown
that many species with GC retention times similar to those found in these
samples are lost by evaporation when an organic extract is dried to constant
weight. As a result of this conservative method of estimation, the total
organic emissions may be slightly overestimated.
In Table 5-19 the organic emission rates are expressed in terms of
mg/cu m of hot zone effluent. These values are compared with the maximum
organic loading that could have occurred if no destruction was accomplished.
The latter'quantity is simply the feed rate (waste and/or auxiliary fuel),
in mg/min, divided by the total hot zone gaseous effluent flow, in standard
cu m/min.
The calculated values of DE total and DE waste both suggest that
destruction of the NCB waste was virtually complete in both tests. There
was no decrease in the overall destruction efficiency (DE total) when the
NCB waste was fed. The calculated destruction efficiency based only on
NCB waste (DE waste) was within experimental error of 100%.
Other quantitative characterizations of the combustion zone effluent
gas were provided by the on-line instruments, gas-detector tubes,
58
-------
TABLE 5-17. OPERATING CONDITIONS FOR TESTS
ON NCB WASTES
Hot Duct Temp. (°C)
Hot Duct Gas Flow /^
s.cu.m/min (dry basis)
Afterburner Temperature (°C)
Fuel Oil /NCB Solution Rate
to Loddby Burner (liters/hr)
Fuel Oil Rate to Kiln
(liters/hr)
Loddby Burner Flame Temp. (°C)
Lime Consumption [32% Ca(OH)2by wt]
(liters/hr)
Natural Gas (1000 cu.m./hr.)
Retention Time (seconds)
Fuel Oil
Background
RUB)
1074
890
1288
1780^
400
1495
380
174
2.6
NCB Waste
R3
1075
1000
1307
2021
249
1548
400
144
2.3
NCB Waste
R4
995
1000
1332
1764
394
1510
620
124
2.3
(1) At 21°C
(2) Oil only
59
-------
TABLE 5-18. SUMMARY OF QUANTITIES OF ORGANIC MATERIALS IN HOT ZONE
EFFLUENT SAMPLE EXTRACTS
Run
R1(B)
R3
R4
Sample '
PWF/CH2C12
I1/CH2C12
ST/Pentane
ST/Methanol
PWF/CH2C12
I1/CH2C12
ST/Pentane
ST/Methanol
PWF/CH2C12
I1/CH2C12
ST/Pentane
ST/Methanol
Estimated Total Organic Content, mq
Total By Total By Sum of GC
19.4 48.4 67.8
<1.0 6.8 7.8
14.4 23.8 38.2
1.8 97.4 99.2
36.6 176.4 213-0
17.8 76.2 94.0
<1.0 1.6 2.6
3.8 3.4 7.2
1.2 150.0 151.2
23.8 231.2 255.0
13.0 53.2 66.2
<1.0 47.2 48.2
23.5 9.0 32.5
0.4 53.8 54.2
37.9 163.2 201.0
Estimated NCB
Content mg
(by GC)
<0.2
<0.02
0.06
<0.01
<0.3
<0.2
<0.2
0.02
<0.01
<0.25
<0.2
<0.02
<0.01
<0.01
<0.24
f See Appendix F for sample codes.
* Based on residue remaining after evaporation of extract to dryness.
-------
TABLE 5-19
HOT ZONE EMISSION RATES AND CALCULATED DESTRUCTION EFFICIENCIES
(see text for explanation)
.RUB}.
R3
R4
Total Organics in HZ Effluent, mg/mw
(PWF/CH2C12 plus I1/CH2C12 Plus
ST/Pentane plus ST/Methanol)
Estimated NCB in HZ Effluent, mg/nT
(PWF/CH2C12 plus Il/CH2Cl2 plus
ST/Pentane plus ST/Methanol)
Calculated Maximum Total Organics in Effluent,
if DE = 0, mg/m3
Calculated Maximum NCB in Effluent,
if DE = 0, mg/m3
DE
DE
Total
NCB Waste
42.2
<0.06
26,500
99.84
52.6
<0.05
33,400
8,350
99.84
99.99,,
42.1
<0.05
31,500
7,290
99.87
99.99
-------
and analysis of the hot zone train impinger solutions.* Data obtained from
the on-line instruments and gas-detector tubes are presented in Table 5-20.
The "not detectable" notation for volatile hydrocarbons means that
no difference in signal intensity could be observed between the sample
gas and cylinder "zero" air.
Because the gas-detector tubes indicated very low levels of chlorine
phosgene and sulfur dioxide, no more precise quantitative methods were
applied for these species.
It is unfortunate that the chemiluminescent analyzer appeared to be unable
to detect nitrogen dioxide during these tests (see Section 4.2.5.3 ). Calculations
based on the nitrogen content of the NCB waste and total hot zone effluent
flow rates suggest that about 2500 ppm of nitrogen dioxide is the theoretical
maximum yield. The data obtained during field testing do not allow
determination of the actual nitrogen dioxide concentration in the combustion
zone effluent. The apparently low level of hydrochloric acid in the gas-
detector tube assay is probably due to scrubbing of the acid by droplets
of water that condensed in the sampling line after it left the hydrocarbon
analyzer.
The hydrochloric acid content of the combustion zone effluent was also
estimated by analysis of the dry impinger and sodium acetate impinger
components of each hot zone sample. The estimates thus obtained for chloride
content of the gaseous effluent were**:
R1(B) 0.51 g/m3
R3 5.26 g/m3
R4 6.96 g/m3
The calculated maximum theoretical loadings of hydrochloric acid, based on
the feed rate and elemental analyses were:
Rl (B) = 0.23 g/m3
R3 3.02 g/m3
R4 2.67 g/m3
Except for the background tests, the actual estimated chloride loadings were
much higher than the calculated maxima. It does not seem probable that the
discrepancy is due to systematic errors in the analysis. The most that can
* The gas bulb samples were not useful. See Appendix F.
** Value stated is a sum of data from dry impinger plus sodium acetate
impingers for each run; Cl" in g/m3 = (Cl~ in ppm x total volume of
solution in 1) T (total volume sampled in m3 x 1000).
62
-------
TABLE 5-20. QUANTITATIVE ANALYSIS RESULTS OBTAINED FROM
ON-LINE INSTRUMENTS AND GAS-DETECTING TUBES
On-Line Instruments
Volatile Hydrocarbons
Carbon Monoxide
Carbon Dioxide
Oxygen**
Nitrous Oxide (NO)
Gas-Detector Tubes
Chlorine
Hydrochloric Acid
Phosgene
Sulfur Dioxide
Nitrogen Dioxide '"
R1(B)
R3
R4
Concentration (volume basis)*
N.D. '
<5 ppm
7 ppm
7.3%
9.4%
50 ppm
N.D.f
<1 ppm
N.D.*
<2 ppm
j.
N.oJ
<1 ppm
<10 ppm
-
N.D.*
<5 ppm
7 ppm
7.0%
9.7%
170 ppm
N.D.'
<1 ppm
N.D.*
<2 ppm
N.D.I
<1 ppm
12 ppm
>300 ppm
N.D. '
<5 ppm
8 ppm
6.7%
10.0%
165 ppm
N.D.'
<1 ppm
N.D.*
<2 ppm
N.D.I
<1 ppm
15 ppm
>300 ppm
** Estimated from C02 concentration for R.3 and R4
* Wet gas basis
** Estimated fror
f Not detectable
U High range tubes were not available on-site when the NOX analyzer
IT fa-ilpH tn ript.prt. Nfl.
63
-------
be said, based on the available data, is that observed hydrochloric acid
hot zone emissions agree with expected emissions to within a factor of two
or three.
5.2.2.2 Qualitative Characterization
Qualitative characterization of the organic material found in the
solvent extracts of the various hot zone sampling train components is
somewhat less straightforward than the quantitative treatment above.
It should be noted at the outset, however, that none of the organic
species identified in the various effluent samples appear to be
especially hazardous. The only chlorinated hydrocarbon, for example,
found in any of the samples was a species with a molecular weight of 200
and an empirical formula of CioHiaC^Cl, in the R3 sorbent trap pentane
extract; the total quantity found corresponded to about 0.3 mg/m3.
The organic chemical species which were identified in samples but
not present in the blanks are listed in Table 5-21. Estimated upper limits
on concentration are given for species which may be present at > 1 mg/m3
The concentrations shown are almost certainly overestimates in all cases.
They were calculated by assuming that the relative abundances indicated
by LRMS were applicable to the entire sample and by assuming that the
total sample weight was the sum of the GC and gravimetric estimates.
The compounds found can be grouped into several categories:
• The first four species are normally expected products
of combustion of aromatic fuels (#2 diesel oil or NCB
waste) and thus not of particular concern.
• The presence of dibutyl tin dichloride probably indicates
that the purge time allowed before each test was insufficient
to eliminate all residues from the wastes burned during the
previous shift. The organo tin species is not a component
of either the diesel oil or the NCB waste. Tin, in any
form, is not present in the feed in sufficient quantity
to account for this species in the effluent. It is
interesting to note that the dibutyl tin dichloride was
found primarily in the pentane extract from the sorbent
trap.
• The "nitrogen-containing compounds" which are indicated
as present in substantial quantities in the R3 and R4 hot
zone samples are high molecular weight species of unknown
structure. Their presence in the methanol extracts of the
two NCB test sorbent traps was inferred from the LRMS
spectra. Fairly low molecular weight nitrogenous fragments
appeared in the spectra at probe temperatures (>300°C)
much higher than would be consistent with their molecular
weights.
It is possible that these compounds are contaminants, but
since they do not appear in the blank or in the background
test sample, it seems most reasonable to presume that they
64
-------
TABLE 5-21. ORGANIC CHEMICAL SPECIES FOUND IN
HOT ZONE EFFLUENT SAMPLE EXTRACTS*
3
Approximate Concentration, mg/m
Species R1(B) R3 R4
Benzoic Acid
Alkyl Benzenes
Toluic Acid
Ethyl Benzoate
Dibutyl Tin Dichloride
Nitrogen-Containing Compounds
Si li cones
MW 326 (Hydroxy Octoxy Benzophenone?)
MW 346 (Oxygenated)
Nonyl Phenol
Phenol
C10H1302C1, Mw 200
Dihydrofuran
Furfural
m/e 530
1.8 1.3
n.f. <1.0
n.f. <0.1
n.f. <0.1
1 .1 <1 .0
n.f. 10.0
3.0 <0.1
n.f. n.f.
n.f. n.f.
n.f. n.f.
n.f. <0.1
n.f. <0.1
<1 .0 n.f.
<1.0 n.f.
n.f. n.f.
n.f.f
n.f.
n.f.
n.f.
3.1
6.0
3.9
10.0
2.0
<1.0
n.f.
n.f.
n.f.
n.f.
<1.0
* Sum of quantities extracted from probe wash, filter, dry impinger
and sorbent trap for each run. Quantitative estimates are based
on LRMS abundance data (Appendix F) and total organic emissions
, (Table 5-2).
| Not found.
65
-------
arise from some degradation of the NCB waste. These species
are not chlorinated and are not highly aromatic. The possible
hazard associated with these species cannot be assessed, but
the types of compounds implied by the available analytical
data are not among those presently known to be especially
hazardous.
t The silicones and the MW 326 species, tentatively
identified as hydroxy octoxy benzophenone, are probably
contaminants. The latter species is found in some of
the blanks (the mixed solvent used as a dry impinger
blank and the ST/methanol blank) as well as in some
sample extracts. The MW 346 compound found in the R4 probe
wash and filter extract is probably also a contaminant.
• The various oxygenated species found at concentrations of
<1 mg/m3 in all three tests are probably combustion
products. None appears to be present at high enough
concentration to cause serious concern.
The conclusion to be drawn from these results is that the organic
species identified in the hot zone effluent during the NCB tests do not
include any compounds that are known to be hazardous at the concentrations
found in these tests. The only species identified in the hot zone effluent
which would clearly require some form of emission control are hydrochloric
acid and nitrogen oxides.
Although the nitrate concentration in the sodium acetate impingers
was measured, this medium was not selected for this purpose and is not
well suited for trapping of nitrogen oxides. The data obtained indicated
little difference between the background and the waste tests, which implies
that trapping efficiency was indeed low. The calculated maximum yield* of
nitrogen oxides, as N02, is 2.46 g/m3 for R3 and 2.14 g/m3 for R4. In
the absence of reliable data, it must be presumed that actual levels
approached these maxima.
5.2.3 Final Emissions
5.2.3.1 Stack Gas
The total particulate loading as measured by EPA Method 5 was:
R1(B) 38.1 mg/m3
R3 13.9 mg/m3
R4 16.4 mg/m3
The estimated emissions of hydrochloric acid, based on analysis of
the sampling train impingers, were:
* Assuming all nitrogen in the waste was converted to nitrogen dioxide.
66
-------
3
R1(B) 2.5 mg/m as Cl
R3 12.5 mg/m
3
R4 13.0 mg/m
These data, when compared to the chloride levels estimated for the
hot zone effluent, imply that the scrubber efficiency is excellent.
Calculated chloride removal efficiencies are 99.5% for R1(B) and 99.8%
for R3 and R4.
No other species were sought in the stack train samples.
5.2.3.2 Scrubber Water
There was no evidence from gas chromatography, IR spectra, or LRMS
analyses of the methylene chloride extracts of the scrubber water samples
that any significant quantity of organic material was present. The total
quantity of material (GC estimate plus gravimetric determination) was
2 0 mg/liter for the fresh scrubber water and <1.4 mg/liter for the back-
ground test spent scrubber water. For the two waste tests, the spent
scrubber water levels were estimated as <1 mg/liter (R3) and <1.4 mg/liter
(R4). No particular organic species were identified for these samples
because the total quantities were so low.
The scrubber water samples were also analyzed for chloride and
nitrate content. The data are presented in Table 5-22. The chloride
data are in good agreement with the estimates obtained from analyses of
the hot zone sampling train impingers. This confirms the implication
of the stack impinger data that the scrubber efficiency for removal of
HC1 is high (see data above). The nitrate data indicate that the
scrubber did not remove substantial quantities of nitrogen oxides from
the combustion zone effluent.
67
-------
TABLE 5-22. CHLORIDE AND NITRATE ANALYSES OF SCRUBBER WASTES
CT
mg/liter
mg/liter
Fresh Scrubber Water
RO - SI
Spent Scrubber Water
R1(B) - SO
R3 - SO
R4 - SO
24
208
1750*
1815*
0.3
47.8
57.2
53.5
*These values correspond to estimated concentrations of 5.94
g/m3 (R3) and 6.16 g/m3 (R4) of chloride in the incinerator
effluent gas. Analysis of the hot zone impinger solutions implied
5.26 g/m3 and 6.96 g/m3, respectively, which is in very good
agreement with the scrubber water assays.
68
-------
6. WASTE INCINERATION COST
Individual economic analyses were performed to determine the commer-
cial (i.e., contract off-site) and industrial costs of incinerating the
used PCB-containing capacitor and NCB process wastes tested at the Rollins
Environmental Service Incorporated incinerator near Houston, Texas. The
economic analyses were divided into capital investment and annual operating
costs. For the commercial disposal facilities (one to be used exclusively
for used capacitor wastes and one to be used four months per year for NCB
process wastes), equipment prices, fuel consumption, and manpower require-
ments estimates for the conveyor fed venturi scrubber-equipped incinerator
system were based on data obtained from Rollins Environmental Services
Incorporated. For the on-site industrial facility for disposal of NCB
process wastes, equipment prices, fuel consumption, and manpower require-
ments estimates were based on engineering scaling of data obtained for the
Chemolite plant incinerator system from the 3M Company. The costs of other
portions of the disposal facilities and associated labor were estimated
using the method of Happel ("Chemical Process Economics," second edition,
John Happel and Donald G. Jordan, 1974), data from Guthrie ("Capital Cost
Estimating," Chemical Engineering, March 24, 1969), and standard engineering
reference methods. Equipment costs were adjusted to March 1976 prices using
the Engineering News Record Index. Land prices are not included in any of
the disposal plant cost estimates. Transportation costs were included for
the economic analyses of commercial facilities for disposal of used capaci-
tor and NCB process wastes, which were premised upon central facilities at
Houston, Texas, incinerating waste materials transported 850 kilometers
(500 miles). Transportation costs were not included for the economic
analyses of the on-site industrial facility for disposal of NCB process
wastes, since the incinerator was assumed to be located at the chemical
plant generating the waste to be disposed.
6.1 CAPITAL INVESTMENT
The capital investment for the facility to incinerate 5,000 metric
tons per year of used PCB-containing capacitors, as shown in Table 6-1, is
based upon a design concept which employs an enclosed one-metric-ton-per-
hour hammermill to mill the used capacitors, an enclosed conveyor belt feed,
a 3.2 meter diameter solid waste kiln, and a 1.6 meter diameter Loddby
liquid waste burner exhausting into a 10.6 meter long afterburner. The
system is equipped with a venturi scrubber, Flexitray absorber, demister,
and stack system. The facility costs include exhaust of the enclosed hammer-
mill and conveyor to the incinerator, induced draft fans (600 KW total), fuel
oil storage tank farm (one-month capacity), lime slurry tank, fuel oil feed
pumps, 30.5 meter stack, and stabilization lagoon system for disposal of
neutralized lime slurry/scrubber wastewater discharge. It was assumed that
fuel oil and natural gas consumption rates per kilogram of milled capacitor
69
-------
TABLE 6-1. CAPITAL INVESTMENT
5000 METRIC TONS/YEAR* USED PCB CAPACITOR WASTE INCINERATION PLANT
Equipment
1-Venturi Scrubber, absorber, pumps, and stack
2- Induced draft fans, including motors
2-Feed pumps
1-Tank farm, including lime slurry tank
l-Hammermill , enclosed, fan and ducting
Instruments (10% of equipment)
(Key Accounts)
Insulation (10% of key accounts)
Piping (45% of key accounts)
Foundations ( 4% of key accounts)
Buildings ( 4% of key accounts)
•Structures ( 4% of key accounts)
Fire Protection (0.75% of key accounts)
Electrical (4.5% of key accounts)
Painting and Clean-up (0.75% of key accounts)
Subtotal
Installed Costs of Special Equipment
1 -Incinerator, conveyor fed
1 -Stabilization lagoon system
Lquipment and Labor
Overheads (30% of Equipment & Labor)
Total Erected Cost
Engineering Fee (10% of Erected Cost)
Contingency (10% of Erected Cost)
Total Capital Investment
Estimated
Size Equipment*
5,300 1pm $ 88,000
300 KW each 69,000
1 ,300 Iph each 4,600
1,900 cu.m. 84,400
0.9 metric tons/hr. 9,900
$255,900
25,600
$281,500
28,200
126,700
11,300
11,300
11,300
2,100
12,700
2,100
$487,200
24 million Kcal/hr.
30,000 cu.m.
Costs**
Labor
$ 25,600
3,800
$ 29,400
42,200
126,700
16,900
7,900
2,300
13,700
19,000
13,700
$271 ,800
$ 759,000
1,500,000
80,000
$2,339,000
701,700
$3,040,700
304,100
304,100
$3,648,900
--J
o
16 metric tons/day
Mar. 1976 basis
F.O.B, cost
-------
waste were 3.6 liters and 136.7 cubic meters, respectively, based on
extrapolation of test results to maximum waste feed rate by Rollins
Environmental Services Incorporated.
The size of the facility was based upon three-shift, seven-day per
week, 52-week per year operation to dispose of 5,000 metric tons of used
PCB-containing capacitors. Annual production of medium and small sized
PCB-containing capacitors had been approximately 45,400 metric tons per
year through 1976, based on use of 11,340 metric tons of PCB per year for
this purpose, at an average PCB weight content of 25 percent. The used
capacitor incineration plant has a nominal thermal capacity of 24 million
Kcal per hour; the actual thermal load is estimated at 94 percent of nominal
capacity, based on 2411 liters per hour of No. 2 fuel oil and 90.6 cubic
meters per hour of natural gas. Heating value of the milled capacitor
waste is less than 3,000 Kcal per kilogram. The total capital investment
for the used capacitor waste facility is estimated at $3,648,900.
The total capital investment shown in Table 6-2 for the contract
disposal facility required to incinerate 4,540 metric tons of nitrochloro-
benzene process wastes in four months is based upon the foregoing Rollins
Environmental Service incinerator used to burn a blend of thermally
equivalent quantities of No. 2 fuel oil and NCB wastes at the rate of
18 100,000 Kcal per hour. Additional waste storage tankage, to store
1140 cubic meters of NCB process wastes, has been included in the capital
outlay estimate. The total capital outlay of $3,749,800 shown for the
central facility has been divided pro-rata to an allocation of $1,249,900
for contract NCB waste incineration, since the facility spends only one-
third of its time in this service.
Since the estimated total annual quantities of approximately 9100
metric tons of NCB process wastes for the United States appear to be large
enough to support an on-site incinerator installation in addition to the
contract facility, capital cost estimates for an on-site plant have been
developed. It was assumed, for incineration at the generating site, that
less auxiliary fuel would be used (30 percent of the total heat input), and
that a smaller, continuous service incinerator would suffice. Consequently,
the capital estimate of $2,816,800 shown in Table 6-3 has been based on
engineering scale-down of the 22.7 million Kcal per hour incinerator used
by the 3M Company at their Chemolite plant to the 4.3 million Kcal per hour
intinerator system required for continuous on-site NCB waste disposal.
6.2 ANNUAL OPERATING COSTS
The annual operating costs for the commercial and industrial facilities
consist of labor, fuel, other utility, hydrated lime and freight costs (where
applicable), plus cost of capital, equipment depreciation, maintenance,
taxes and insurance. The labor costs for the contract disposal facilities
located near Houston, Texas, have been calculated on the number of personnel
assigned to operate the system at the rates given by Rollins Environmental
Services Incorporated. The labor costs for the on-site NCB incineration
facility have been calculated on the basis of data from the 3M Company. Costs
for supervision, supplies, and payroll-related expense have been included for
all facilities at rates prevalent in the chemical industry.
71
-------
rv>
TABLE 6-2.
CENTRAL FACILITY FOR INCINERATION
CAPITAL INVESTMENT
- 4540 METRIC TONS/4-MONTH PERIOD NCB WASTE
Equipment
1-Venturi scrubber, absorber, pumps and stack
2-Induced draft fans, including motors
2- Feed pumps
1-Tank farm, including lime slurry tank
Instruments (10% of equipment)
(Key Accounts)
Insulation (10% of key accounts)
Piping (45% of key accounts)
Foundations ( 4% of key accounts)
Buildings ( 4% of key accounts)
Structures ( 4% of key accounts)
Fire Protection (0.75% of key accounts)
Electrical (4.5% of key accounts)
Painting and Clean-up (0.75% of key accounts)
Subtotal
Installed Costs of Special Equipment
1 -Incinerator, conveyor fed
1 -Stabilization lagoon system
Equipment and Labor
Overheads (30% of Equipment & Labor)
Total Erected Cost
Engineering Fee (10% of Erected Cost)
Contingency (10% of Erected Cost)
Total Capital Investment
Pro-rata Total Capital Investment for
NCB Waste Incineration
Estimated
Size Equipment**
5,300 1pm $ 88,000
300 KW each 69,000
1,300 Iph each 4,600
3,040 cu.m. 116,200
$277,800
27,800
$305,600
30,600
137,500
12,200
12,200
12,200
2,300
13,800
2,300
$528,700
24 million Kcal/hr.
30,000 cu.m.
Costs*-
Labor
$ 27,800
4,200
$ 32,000
45,800
137,500
18,300
8,600
2,400
14.900
20,600
14,900
$295,000
$ 823,700
1,500,000
80,000
$2,403,700
721,100
$3,124,800
312,500
312,500
$3,749,800
$1,249,900
-------
OJ
TABLE 6-3. CAPITAL INVESTMENT
4540 METRIC TONS/YEAR NCB WASTE INCINERATION PLANT (ON-SITE)
Equipment
1 -Scrubber, absorber, pumps and stack
3-Induced draft fans
1-Well water supply system
1 -Waste water treatment system
1-Incinerator system, including feed system
Instruments (10% of equipment)
(Key Accounts)
Insulation (10% of key accounts)
Piping (45% of key accounts)
Foundations ( 4% of key accounts)
Buildings ( 4% of key accounts)
Structures ( 4% of key accounts)
Fire Protection (0.75% of key accounts)
Electrical (4.5% of key accounts)
Painting and Clean-up (0.75% of key accounts)
Subtotal
Installed Costs of Special Equipment
Refractories
Equipment and Labor
Overheads (30% of Equipment & Labor)
Total Erected Cost
Engineering Fee (10% of Erected Cost)
Contingency (10% of Erected Cost)
Total Capital Investment
Estimated
Size Equipment
1,100 1pm $ 98,200
75 KW, total 45,500
1,100 1pm 21,200
1,100 1pm 72,700
4.3 million Kcal/hr. 318,200
$555,800
55,600
£611,400
61 ,100
275,100
24,500
24,500
24,500
4,600
27,500
4,600
$1,057,800
Costs
Labor
$ 55,600
8,300
$ 63,900
91,700
275,100
36,700
17,100
4,900
29,800
41 ,300
29,800
$590,300 $1,648,100
157,600
$ 1 ,oUb ,/UU
541,700
$2,347,400
234,700
234,700
$2,816,800
-------
The utility costs include those for electricity. Water costs, based
on actual consumption data, have been included only for the on-site NCB
incinerator, since the Houston, Texas, located contract disposal facility
has its own well-water system and its water costs were covered under
electric power costs. The amount of No. 2 fuel oil consumed was based on
extrapolation of actual test data for the two wastes at the Rollins Environ-
mental Services, Deer Park, Texas, site.
The annual operating costs for the contract incineration of 5,000
metric tons per year of used PCB-containing capacitors are summarized in
Table 6-4. The estimated annual operating expense for the plant, based on
21-shift per week operation, is $3,704,600 or $740.92 per metric ton
Almost $307 of this very high disposal cost per metric ton is cost of the
fuel oil and natural gas required to yield the temperatures needed for
total incineration of the PCB, since the waste itself has no appreciable
heating value. Disposal costs would be significantly reduced if a high
heating value liquid waste could be utilized in place of fuel oil to
support the incineration of the PCB.
The annual operating costs for the contract disposal and on-site
plants for incineration of NCB wastes are shown in Tables 6-5 and 6-6
respectively. For contract disposal, using the pro-rata capital invest-
ment value of $1,249,900 given in Table 6-2, and four months of 21-shift
per week operation on NCB process wastes, the estimated total operating
cost is $1,099,200, equivalent to $242.11 per metric ton. The total annual
corresponding operating cost for the disposal of the same tonnage (4540
metric tons) of NCB waste on-site, given in Table 6-5 is $1,283 300
equivalent to $282.67 per metric ton. This figure is also based on'2]-
shift per week operation.
The cost of capital shown is based on the assumption that private
debt financing is used for each facility.
74
-------
TABLE 6-4. ANNUAL OPERATING COST
5000 METRIC TONS/YEAR* USED PCB CAPACITOR WASTE INCINERATION PLANT
Item Cost - $/Year
Depreciation (15% of plant investment) 5 ^47,300
Cost of Capital (10% of plant investment) oSJ'Snn
Maintenance (9% of plant investment) c?!'5nS
Utilities 1,675,400
Electric power [1260 KW (24)(310)] $0.015 = $ 140,600
Water [3218 1pm (1440) (310) G> $0.008/1000 liter = --- **
Fuel Oil, No. 2, 114,400 bbl @ $13.00/bbl = 1,487,200
Natural Gas [90.6 cu.m./hr (24)(310)] G> $70.63/1000 cu.m. = 47,600
Solid Waste Disposal -,no~c™**
Freight (500 mi. @ $1.71/cwt) 5,000 metric tons @ $37.70/metric ton 188,500
Chemicals
Hydrated Lime, 3,750 metric tons @ $35.00/metric ton 131,300
Labor onn 395>800
Chief Operator - 1 x 24 x 260 x $8.05 = 50,200
Operators - 3 x 24 x 365 x $6.95 = 182,600
Supervision (15% of Operating Labor) = 34,900
Supplies (20% of Operating Labor) = 46,600
Payroll Related Expense (35% of Operating Labor) = 81,500
Taxes and Insurance (2% of plant investment) 73'000
Total 53,704,600
Cost per metric ton of used PCB capacitor waste 0 5,000 metric tons/year S 740.92
* 16 metric tons/day '
** From wells-cost included in Total Capital Investment and Electric Power Costs
*** Offset by aluminum recovery
-------
TABLE 6-5. PRO-RATA ANNUAL OPERATING COST
CENTRAL FACILITY FOR INCINERATION-4540 METRIC TONS/4-MONTH PERIOD NCB WASTE
Item Cost - $/4 Month Period
Depreciation (15% of pro-rata plant investment) $ 187 500
Cost of Capital (10% of pro-rata plant investment) 125*000
Maintenance (9% of pro-rata plant investment) 112*500
Utilities 282*100
Electric power [1260 KW (24)(104)] $0.015 = $ 47,200
Water [3218 1pm (1440)(104)] @ $0.008/1000 1 — *
Fuel Oil, No. 2, 16,490 bbl G> $13.00/bbl = 214,400
Natural Gas [(114.6 cu.m./hr)(24)(104) @ $70.63/1000 cu.m. = 20,200
Freight (500 mi. @ $1.71/cwt) 4540 metric tons @ $37.70/metric ton 171 200
Chemicals '
Hydrated lime, 1,817 metric tons @ $35.00/metric ton 63 600
Labor 132*300
Chief Operator - 1 x 24 x 87 x $8.05 = 16,800 '
Operators - 3 x 24 x 122 x $6.95 = 611000
Supervision (15% of Operating Labor) = 11,700
Supplies (20% of Operating Labor) = 15,600
Payroll Related Expense (35% of Operating Labor) = 27,200
Taxes and Insurance (2% of pro-rata plant investment) 25,000
Total $1,099,200
Cost per metric ton of NCB waste $ 242 11
-------
TABLE 6-6. ANNUAL OPERATING COST
4540 METRIC TONS/YEAR NCB WASTE INCINERATION PLANT (ON-SITE)
Item
Cost - $/Year
Depreciation (15% of plant investment)
Cost of Capital (10% of plant investment)
Maintenance (5% of plant investment)
Utilities
Electric power [120 KW (24)(330)] $0.015
Water [1100 1pm (1440)(330)] @ $0.009/1000 1
Fuel Oil, No. 2, 7040 bbl G> $13.00/bbl
Chemicals
Hydrated lime, 1,817 metric tons
Labor
Operators - 2 x 24 x 365 x $7.00
Supervision (15% of Operating Labor)
Supplies (20% of Operating Labor)
Payroll Related Expense (35% of Operating Labor)
Taxes and Insurance (2% of plant investment)
= $ 14,300
4,200
91,500
= 122,600
= 18,400
= 24,500
= 42,900
Total
$ 422,500
281,700
140,800
110,000
63,600
208,400
56,300
$1,283,300
Cost per metric ton of NBC waste @ 4540 metric tons/year
282.67
-------
7. REFERENCES
TRW Document #27033-6005-RU-00, "Analytical Plan for Facility No. 8
Tests to be Conducted at Rollins Environmental Services, Inc.,"
November 1976.
"Destroying Chemical Wastes in Commercial Scale Incinerators Facility
Report No. 3, Systems Technology," published under EPA Contract
No. 68-01-2966, November 1976.
Environmental Protection Agency, Standards of Performance for New
Stationary Sources, Federal Register, Vol. 41, No. Ill, June 8, 1976.
78
-------
APPENDIX A
ASSESSMENT OF ENVIRONMENTAL IMPACT OF DESTROYING CHEMICAL WASTES
AT
ROLLINS ENVIRONMENTAL SERVICES, INC.
DEER PARK, TEXAS
The Rollins Deer Park facility is located about 20 miles east of
Houston in a highly industrialized area near the Houston ship channel.
This rotary kiln incinerator will be evaluated for its capability of
thermal destruction of the following industrial chemical wastes:
1) Nitrochlorobenzene - 7,000 gallons
2) PCB wastes in capacitors - 4,000 Ibs whole capacitors
10,000 Ibs hammermilled capacitors
State and federal permits have been issued to Rollins authorizing both
air and water emissions, including:
Texas Air Control Board, Permit No. R-679
Texas Water Quality Board, Permit No. 01429
National Pollutant Discharge Elimination System, Permit No. TX0005941
The Rollins incineration system consists of a rotary kiln and a liquid
injection burner feeding into a common afterburner. Total heat release is
110 million Btu/hr. Solid wastes are conveyor fed into the rotary kiln and
reacted at temperatures from 1200 to 1500°F. Combustion gases are passed
through an afterburner, fired by a liquid injection burner, to attain
temperatures of 2300 to 2400°F. Estimated total residence time is 2.6 sec-
onds. Ash from the kiln is disposed of on-site in an approved landfill.
A venturi scrubber is used to remove particulate from the exhaust gases,
Lime is injected to neutralize the scrubber water in a single-pass system.
The used scrubber water then enters settling ponds and is further treated,
if required, before eventually entering the Houston ship channel. Exhaust
gases also pass through absorption trays and a mist eliminator before
entering the stack. The exhaust stack is 100 feet high with a sampling
platform and ports about 55 ft above ground level. Standard sample ports
90° apart are provided. The exhaust plume was white with slight fallout on
the day of the site visit.
Operating at this site since 1971, Rollins incinerates approximately
250,000 Ibs of waste per day, consisting of bulk liquids, solids, and
slurries. The facility operates 24 hours a day, 7 days a week, and has
42 employees. About 10-15 trucks enter the site for unloading each day.
Routine maintenance is performed four times a year. The facility is rela-
tively clean and odor-free for a commercial destruction facility.
79
-------
Numerous refineries and chemical plants surround the Rollins facility.
There are grass, brush, and small trees in the area. Local wildlife include
quail, dove, rabbits, and fox. There are some cattle grazing in the area.
The nearest residences are 3 to 4 miles from the site. Prevailing winds
are 10-12 miles per hour from the southeast.
A maximum of 600,000 gallons of waste are stored on-site for periods
averaging two weeks. A spill prevention plan is in effect. Wastes are
analyzed prior to incineration by an on-site laboratory. Operation of the
incinerator is not considered to be noisy.
The potential detrimental environmental effects are expected
to result from: 1) storage and handling of wastes prior to destruction,
2) emissions occurring during tests, and 3) disposal of liquid and solid
residue remaining after combustion. The most significant hazard would
result from contact with waste liquid and/or fumes during a spill.
The nitrochlorobenzene waste is a brownish liquid containing 90%
isomers of nitrochlorobenzene. The p-nitrochlorobenzene, the desired
product at the waste generation source and probably also the major con-
stituent of the nitrochlorobenzene waste, is a significant toxic hazard.
The 1973 Toxic Substances List provides the following data on
p-nitrochlorobenzene and m-nitrochlorobenzene:
p-ni trochlorobenzene
Oral Toxicity: LD5Q (rat) - 420 mg/kg
Dermal Toxicity: OSHA (skin) - 1 mg/M3
m-nitrochlorobenzene
Oral Toxicity: LD5Q (rat) - 2460 mg/kg
In addition, the mixture has a flash point of 264°F. The waste m-
and o-nitrochlorobenzenes are volatile solids which give off flammable
vapors when heated and may form explosive mixtures with air. The waste
is also incompatible with strong oxidizing agents and should not be stored
near strong oxidizing agents. Since the nitrochlorobenzenes are also toxic
by inhalation, inhalation of the vapors should be avoided at all times. In
the case of skin contact, the skin should be washed off immediately with
water and soap. The effects of overexposure to the nitrochlorobenzene
waste include blue tint to fingernails, lips and ears indicative of
cyanosis, and headache, drowsiness and nausea followed by unconsciousness.
The capacitor wastes contain membranes impregnated with PCB. The
general environmental contamination of PCB, resulting in adverse effects on
certain forms of animal life, is well known. The studies $onducted so far
have indicated that some persons carry a body burden of PCB in their fat
tissue. In addition, PCBs have been shown to accumulate in fish and
80
-------
aquatic invertebrates to levels of 75,000 times that present in water.
Thus PCB is definitely bioaccumulative. The 1973 Toxic Substances List
provides the following ^ta on the type of PCB used in capacitors:
Aroclor 1242 [Chlorobiphenyl (42% C1)]
o
Inhalation Toxicity: Toxic Concentration (human) - 10 mg/M°
Dermal Toxicity: Lethal Dose (rabbit) - 794 mg/kg
OSHA (skin) - 1 mg/M3
Oral Toxicity: LD5Q (rat) - 8650 mg/kg
Storage and Handling
Liquid wastes will be received by tank truck and transferred to
storage/run tanks. Solid wastes will be received by truck and stored in
the original containers until incinerated. Leaks and any spills will be
washed down with water or absorbed. All rinse or wash down liquids, or
absorbent, will be incinerated, or neutralized prior to disposal.
Incineration Tests
Operating temperature and residence time of the incinerator and after-
burner should provide essentially complete combustion of the wastes,
resulting in harmless exhaust emissions. On-line monitoring of gases from
the combustion zone will be utilized as an indication of combustion
efficiency. In addition, stack emissions (downstream of scrubber) will be
checked for hazardous gaseous species using Gastec® analyzer tubes for
specific gases and vapors.
Disposal of Residues
Residue material from the incineration process will consist of
scrubber water and ash. Liquid residue from the scrubber will be analyzed
by Rollins personnel before discharge to the channel. Solid residues
(ash) will also be tested prior to on-site landfill. Wastes remaining in
the storage/run tanks and residual solids will be incinerated at the con-
clusion of the test program.
81
-------
APPENDIX B
SAMPLE VOLUME DATA - PCB TESTS
82
-------
TABLE B-l. SAMPLING SYSTEM DATA SUMMARY
Run/Train
Run I
Stack
Combustion Zone
Run II
Stack
Combustion Zone
Run III
Stack
Combustion Zone
Sampling
Time
e(min)
-
120
60
138
60
150
Gas Volume
vm (ft3)
-
116.4
54.5
132.3
36.5
129.6
Liquid
Volume
Vw(nl)
-
185-
512
187
335
246
Site
Temp.
TS(°F)
-
1091
130
1089
135
1096
Dry Gas
Meter Temp.
y°F>
-
86
70
91
66
91
Nozzle
Diameter
Dn (in.)
-
*
3/8
*
3/8
*
Gas
Velocity
Vs (ft/sec)
-
-
30.14
88
32.25
90
Moisture
Content (%)
-
7.13
30.6
6.45
33.1
8.48
Pressure
Drop
AH
(in. H20)
-
3.5
3.5
3.5
1.5
3.2
Percent
Isokinetic
Average
-
-
106
-
69
-
00
co
*No nozzle was used for the combustion zone sampling train.
-------
TABLE B-2. ROLLINS SAMPLE GAS VOLUMES AT STANDARD CONDITIONS
Test
No.
I
II
III
Stack
Dry
n3
-
54.82
36.80
m3
-
1.55
1.04
Wet
ft3
-
78.99
54.97
m3
-
2.24
1.56
Hot Zone
Dry
ft3
113.70
128.06
125.29
m3
3.22
3.63
3.55
Wet
ft3
122.43
136.88
136.90
m3
3.65
3.88
3.88
84
-------
TABLE B-3. COLLECTED WATER VOLUME DATA
Test
I Hot Zone
Stack
II Hot Zone
Stack
III Hot Zone
Stack
Water Volumes in Impingers
1st Imp. (ml)
Initial
100
—
100
100
100
100
Final
270
—
190
405
245
410
2nd Imp. (ml)
Initial
100
—
100
100
100
100
Final
80
~
145
270
140
160
3rd Imp. (ml)
Initial
_
™
-
_
Final
5
~
10
15
10
5
4th Imp. (g)
Initial
_
*
_
—
_
Final
30
™
42
22
51
10
Total Collected
Liquid
Sample (ml)
185
"
187
512
246
385
CD
-------
APPENDIX C
ANALYTICAL CHEMISTRY DETAILS - PCB TESTS
Many of the analyses of the samples from the PCB tests were performed
by subcontracted laboratories. All sample preparation as described in
Section 4.1.4.1 was carried out in the TRW chemistry laboratories. In
addition, the gravimetric and IR organic survey techniques were performed
at TRW. The following laboratories were used for the analyses indicated:
• Ultrachem - quantitative analysis for PCBs by GC/MS.
• West Coast Technical Service, Inc. - qualitative organic
surveys by LRMS.
• Barringer Research Ltd. - quantitative inorganic surveys
by ICPOES.
• Commercial Testing and Engineering - semi-quantitative
inorganic surveys by SSMS.
Ultrachem, located in Walnut Creek, California, was utilized for its
experience and demonstrated ability in PCB analyses. They used a Finnigan
Model 9500 gas chromatograph connected, via a single stage glass jet
separator, to a Finnigan Model 3100D quadrupole mass spectrometer controlled
by a System Industries System/250 computer system.
Five yl of each sample was injected onto a 5 ft x 1/4" OD OV-17 on
100/120 mesh Chromosorb W AW-DMCS column, temperature programmed from 125QC
to 250°C at 8°C/min. The vacuum divertor valve (solvent dump valve) was on
for 1-1/2 minutes after injection; an additional 30 seconds was allowed for
the column flow to equilibrate before temperature programming and mass spec
scanning were started. The following respective sets of ions were used for
the detection of the monochloro through octachlorobiphenyls - 188-198, 222-
232, 256-266, 290-300, 324-334, 358-368, 392-436. Integration time was 17
milliseconds per atomic mass unit; the seven sets of ions were scanned
every 1.9 seconds in a cyclic, continuous recording. The procedure yielded
a detection limit of 3 to 5 ng for a 5 yl injection. Every sample derived
from the PCB tests at Rollins was analyzed by Ultrachem, including all
blanks, controls, and standards.
West Coast Technical Service (WCTS) in Cerritos, California, is
frequently used for LRMS and other analytical work. In this particular
case, TRW's mass spectrometer was down because of facility modifications
underway to create a mass spectrometry center to house the Hitachi Perkin-
Elmer RMU-6 instrument and TRW's new Finnigan GC/MS system. Organic survey
samples were thus sent to WCTS, who also has a Hitachi Perkin-Elmer RMU-6
MS instrument.
Barringer Research in Rexdale, Ontario, Canada, was utilized to perform
inorganic analyses because their ICPOES technique provides data on a
relatively large number of elements (32) with good accuracy and at very low
86
-------
cost. Analyses performed by Barringer on samples from waste test burns at
Systems Technology and St. Lawrence Cement (facility nos. 3 and 7) yielded
data that was on the average 14% lower than AAS data with a range from -56%
to +68% using 30 data points. Because of the good accuracy obtained on
these previous samples, which were in the same or similar solution matrices,
and the generally low concentrations found, no AAS work was performed on the
Rollins samples.
Commercial Testing and Engineering (CT&E) located in Golden, Colorado,
was selected to run complete inorganic surveys by SSMS on samples of dissolved
solid residues from the Rollins kiln. SSMS is a very specialized and expen-
sive analysis technique that is available from only a few commercial labora-
tories in this country. CT&E is one of these labs and also has experience
with analyzing environmental samples.
87
-------
APPENDIX D
PCB ANALYSIS BY GC/MS
The analysis of the Rollins sample concentrates for PCBs was carried out
using the method in this appendix as a guide. The instrument and data
system used in the analysis was not the same as the system for which the
method is written so slight changes in the method were necessary. In
particular, an EPA Aroclor standard of known PCB composition was used
for determination of response factors.
D.I SCOPE AND APPLICATION
Scope
This method is a direct analysis for polychlorinated biphenyl
compounds (PCB). A search for these compounds is carried out primarily
because of 1) their carcinogenic properties and 2) their tendency to be
bio-accumulative and to resist degradation. If these materials are found
to exist in environmental samples at significant levels (to be defined
later), it is almost certain that a Level 2 analysis will follow.
Sensitivity
The actual instrument sensitivity is expected to vary with factors
including operating parameters and the efficiency of the GC/MS interface/
sample enrichment device. In practice 5-10 nanograms (ng) will yield an
MS signal with a usable signal noise ratio of >10:1. A dynamic range of
approximately 1000 is typical. If the detectors are saturated a sample
dilution is required.
Detection Limit
There will typically have to be at least 50 yg of a PCB compound
extracted from a sample and concentrated to 10 ml final sample solution.
This presumes a 1 yl sample injection volume and the typical instrument
sensitivity specified above. Larger sample injection volumes are possible.
Interferences
The possibility of non-PCB compounds interfering with the analysis by
causing a false positive reading is about 3 percent. This is based on the
number of non-PCB compounds which have a spectrum containing one or more
of the specific PCB subset ions in the 11-100 percent relative abundance
range. No interferences have been identified which would mask the presence
of PCB. False positives are easily detected by the additional printed out
data during the quantification step.
D.2 SUMMARY OF METHOD
This is a combined gas chromatography/mass spectrometry method
(GC/MS). Microliter quantities of concentrated sample extracts are
injected into the chromatpgraph. The concentrated extracts result from
the extraction of the various samples obtained from the sampling activity.
88
-------
The extraction and concentration procedures are specified in Sections 7.3
and 7.4 of this document.
Micro!Her sized samples are injected onto a gas chromatographic
column and are separated by the differences in the retention character-
istics between the sample components and the column material. As the
mixture components elute from the column they are transported via an
instrument interface to a mass spectrometer preprogrammed to act as a
detector specifically for PCBs. The signal from the mass spectrometer
is converted into a mass fragmentogram, a specialized gas chromatogram.
The presence of peaks at levels significantly greater than those for the
reagent blank samples indicate PCB presence at high confidence levels.
These levels are quantitatively determined using external standardization.
D.3 DEFINITIONS
• Subset masses - a group of ions whose masses are characteristic
to a particular class of compounds.
• Mass fragmentogram - a gas chromatogram, using the MS as a detector
for specifically selected fragment ions.
t TIM - Total Ion Monitoring
• SIM - Selective Ion Monitoring
• PCB - Class - Group of PCB having the same chlorine content.
• 1C1-PCB, 2C1-PCB, etc. - This refers to the number of chlorine
atoms on the PCB molecule. The maximum number is ten.
D.4 SAMPLE HANDLING AND PRESERVATION
These samples are organic solutions resulting from extraction and
concentration. They are contained in stoppered or septum sealed ampules
or flasks typically containing less than 5 ml. These samples should never
be exposed to direct sunlight. These samples shall be stored in a refrig-
erator when not being used. Solvent losses through evaporation shall be
minimized by maintaining good seals on the containers. Abnormal or unusual
losses in solvent volume shall be reported on the data report sheet. Con-
tact of the sample with hands and other sources of outside contamination
shall be avoided. PCBs are toxic materials and due care shall be used in
handling all samples and standards. Gloves and the fume hoods are appro-
priate considerations.
D.5 APPARATUS
This section specifies the major pieces of required apparatus. A
normal compliment of glassware and various laboratory implements is assumed.
89
-------
Automated Combined GC/MS, Finnigan Model 4023
Several quadrapole or magnetic sector instruments, computer driven,
are available to perform this analysis using the selective ion monitoring
or SIM technique. Basic required capabilities include:
• Resolution sufficient to obtain unit mass resolution in the
40-400 range. This is typically 1000.
• Capability for glass 2-4mm ID packed columns and a sample
enrichment device to achieve the instrument sensitivities
described above.
• Electron multiplier detection system.
• Selective ion monitoring capability for at least 8 selected
ions.
Interactive Data System
Capable of gathering and storing mass abundance data from the MS,
generating fragmentograms, and performing quantitation with peak area or
peak height ratios, normalizing and correcting for background.
GC column, a 6-foot glass column containing 3 percent OV-17 on 100-200 mesh
Gas-Chrom Q or equivalent.
GC injection syringes, 10 yl, Hamilton 701-N or equivalent.
Analytical balance, capable of weighing 0.05 mg.
'Volumetric Flask, 2 ml, hexagonal base, Corning 5630 or equivalent.
D.6 REAGENTS
Supplies of the following solvents, Pesticide Grade, Distilled-in-
Glass® , Nanograde®or equivalent should be kept at hand in several
liter quantities.' These solvents will be used to prepare analytical
standards, make dilutions, do solvent replacement or other similar
activities as required.
• Pentane
• Methylene chloride
• Methanol
• Acetone
• Petroleum ether
90
-------
t Hexane
• Di-ethyl ether
PCB isomers, 99+ percent purity. Available from Analabs, Inc., New Haven,
Connecticut. The stock of available standard PCB isomers should include
but not be limited to the following. There is no evidence of significant
variation in instrument response with position isomers of PCBs with each
chlorine level. Those on the list were selected on the basis of cost and
availability. These compounds are used to determine response factors
(instrument sensitivity) in the calibration steps.
3 - chlorobiphenyl
2, 3 - dichlorobiphenyl
2, 5, 4' - trichlorobiphenyl
2, 4, 2', 4' - tetrachlorobiphenyl
2, 4, 5, 2', 5' - pentachlorobiphenyl
2, 4, 6, 2', 4', 6' - hexachlorobiphenyl
2, 3, 4, 5, 6, 2', 5' - heptachlorobiphenyl
D.7 PROCEDURE
D.7.1 Preparation of the Unknown Samples
The samples in solution shall be prepared in such a way that the
solutions will be concentrated as much as is practically possible:
Take a 2 ml aliquot of the 10 ml sample resulting from the concentration
of the sample extracts. Place it in a clean 2 ml volumetric flask or
vial. This sample will be used for the yl injections. If more concen-
tration is required, put the 2 ml in a receiver ampule and direct a
stream of clean filtered laboratory air over the solution until the
sample is at the desired volume.
D.7.2 Preparation of PCB Standard Samples
Concentrated stock solutions of the desired individual PCB isomers
shall be made. Five milligrams in 50 milliliters of iso-octane yields
a suitable concentration of 100 nanograms (ng) per microliter (yJl).
Portions of these stock solutions can be serially diluted or combined
with each other to produce working standard mixtures. The working
standard solutions will be made up in 100 ml volumetric flasks. The
attached table (courtesy of the EPA) provides a guide of volumes of stock
solution required to yield 100 ml of the desired working standard solution,
91
-------
D.7.3 Preparation of the GC/MS Instrument
The chromatograph shall be prepared according to the manufacturer's
operating manual. Key GC operating parameters to be used are as follows:
• Temperatures
Oven - Programmed from 100°C to 275°C at 12°C/min.
Injector - 200° to 220°C
Transfer lines - 150° to 175°C
Separator - 200° to 220°C
• Carrier gas - Helium at 30 ml/min.
• Solvent divert time programmed to divert solvent and other
species until a preprogrammed start time.
t Other parameters may be added (to be determined).
The mass spectrometer (MS) shall be prepared according to the manu-
facturer's operating manual. Briefly summarized, this procedure includes:
• Monitoring all applicable pressures and confirming that
proper vacuum levels are being achieved. Proper levels are
specified in the manual.
• Checking all applicable temperatures and making adjustments
as necessary:
Ionizer - -40°C
Manifold - ~40°C
Electron multiplier - ~40°C
• Checking performance of ionizer
Emission current - 0.2 to 0.5 milliamps
t Optimizing resolution/sensitivity
Electron multiplier voltage - 1500 to 2000 volts
• Confirming that all electronics associated with data gathering
are functioning properly.
Date and record parameters in instrument logbook.
92
-------
Computer Data Gathering/Data Storage
The data acquisition and storage operations shall be preprogrammed
to select only ions of specific mass to charge ratios that are character-
istic of PCBs. Program the following data gathering parameters into the
instrument controller.
a) Selected mass set voltages 188-198, 222-232, 256-266,
290-300, 324-334, 358-368, 392-436
b) Integration time at each mass set voltage - 17 msec
c) Samples per AMU - 1
d) Counts threshold - 200 to 300
e) Run time - 99 min (operator can intervene)
Date and record in instrument logbook.
D.7.4 Starting the Analysis
When all preparations have been completed, inject the sample into
the gas chromatograph which is set in the solvent divert mode. At a
given time after injection, the solvent divert valve directs the sample
components eluting from the GC into the MS. Concurrently, a signal is
sent to start the MS scanning sequence and data acquisition of the mass
spec/computer to acquire and store data in accordance with the commands
inputed by the operator. The instrument will then gather and store data
over the time interval during which the PCBs are expected to elute. The
analysis will continue for a pre-set duration or until the operator inter-
venes. The system shall be programmed to stop the analysis after 99 min-
utes have elapsed from time of injection.
D.7.5 Data Processing and PCB Quantification Calibration
A known mixture of the pure isomer standards (D.6) is injected into
the chromatograph and analyzed using the same technique and operating
parameters as used for the samples and which are described above. The
instrument can detect as low as 3-5 ng but for calibration purposes,
accurately injected amounts in the 100 ng range yield more accurate PCB
response factors.
The reconstructed chromatogram based on the normalized raw data is
obtained via computer interaction. The scan number or spectrum number
corresponding to the top of each peak is noted. The spectrum number
representing the top of a peak is inputed to the computer, and the command
is given to obtain the raw portioned data counts for the selected AMUs
corresponding to the molecular ion cluster for that PCB. These counts
represent the number of ions actually detected for each AMU. Knowledge
of the standard composition and retention times enables the operator to
identify the level of chlorination represented in each peak, i.e., 1 Cl,
2 Cl, etc. The selected AMUs to be searched are relatively specific to
93
-------
each chlorine level, and these are tabulated below. For example, if the
peak in the chromatogram of the standard is a 3 chlorine PCB, the 256-266
AMU range is selected but only the 256, 258, 260 AMU detected ions are
summed.
The sum of the ions for the selected AMUs representing each PCB
chlorine level is divided by the amount (ng) of each PCB standard injected
into the instrument. A response factor of ion counts per nanogram is
obtained for each of the PCB standards with a different chlorine level.
This standard should be analyzed at least three times, average response
factors determined and the standard deviation calculated. Thereafter, the
standard is injected at the start of each working day to confirm acceptable
instrument performance.
D.8 QUANTIFICATION OF PCBs IN SAMPLES
At the completion of the analysis of a sample (mass spectral data in
storage) perform the following:
1) Obtain the reconstructed chromatogram using all of the subset
data specified in D.7.3. This signal normalized output pro-
vides retention time data from which the composition of the
PCB (chlorine levels) can be estimated. This provides the
data analyst a feel for the type of PCB mixture with which he
is working (e.g., Aroclor 1242, 1260, etc.) but is not needed
for final identification and quantification.
TABLE D-l.
AMUs to Be Searched for Raw Counts
Chlorine Atoms per
PCB Molecule
Selected AMUs
1C!
2
3
4
5
6
7
8
9
190
222, 224, 226
256, 258, 260
290, 292, 294, 296
324, 326, 328, 330
358, 360, 362
392, 394, 396
426, 428, 430
460, 462, 464
94
-------
2) Obtain separate computer prepared chromatograms for each PCB
class, i.e., 1C1-PCB, 2C1-PCB, etc., up to the maximum number
of chlorines indicated by the chromatogram obtained in Step 1
of this section. This is obtained by inputing only the AMUs
specific to each PCB class from Table D-l and obtaining a
chromatogram from those AMUs only.
3) Starting with the chromatogram for detected PCBs with the
smallest number of chlorines, locate by spectrum number the
top of each peak. (For purposes of illustration, assume that
1 chlorine and 2 chlorine PCBs were not detected and hence
one starts with 3 chlorine PCBs). For each spectrum number
corresponding to the top of a peak, a numerical printout is
obtained presenting the absolute value, i.e., number of ion
counts for AMUs 256 through 266 (from Section D.7.3). The
number of ion counts only for AMUs 256, 258, and 260 are
summed for each peak. These ion counts are divided by the
calibration response factor for 3C1-PCB to obtain the weight
of 3C1-PCB having caused that peak. This weight is the amount
of 3C1-PCB injected into the instrument which caused that peak.
wt in PCR (no) = - £ counts Equation 1
wt .jU-rtB ing; Response counts/ng
The weights of 3C1-PCB causing each of the peaks in the 3C1-PCB
chromatogram are summed and multiplied by the appropriate
aliquot or dilution factors to obtain the total weight of
3C1-PCB isomers in the starting 10 ml sample volume:
wt 3Cl-PCB(ng) = E3Cl-PCB(ng) x y x ^ x Bml
Equation 2
where:
wt 3Cl-PCB(ng) = amount of 3C1-PCB in the entire sample
Z3Cl-PCB(ng) = the sum of 3C1-PCB in each of the peaks
in the 3C1-PCB chromatogram.
Ay! = size of the injection (microl iters)
Bml = volume of total sample (mill il iters)
The weights of 1C1-PCB, 2C1-PCB, 4C1-PCB, 5C1-PCB, etc., up
through 10C1-PCB, if detected, are calculated in similar fashion
95
-------
APPENDIX E
SAMPLE VOLUME DATA - NCB TESTS
96
-------
TABLE E-l
STACK SAMPLING DATA
Ol
03
O
12/14
12/15
12/17
12/18
O)
-£D
c E
3 3
Q£ Z
R1(B)-S
R2-S
R3-S
R4-S
Q_
E
i—
o
•(-> O
OO 0
70
59
58
62
i.
in
CO
a. I
-*£ <4-
0 O
to
** §
oo E
763
766
767
764
(U
X
o &«
12.8
12.8
13.1
12.4
a>
X
0
a
c
0
S-
6.2
6.1
5.7
6.3
s-
et
CO
CO
O)
O
X
LlJ *S
150
150
160
140
(U
CO +J
•i- C
o o
29.5
29.9
29.9
30.1
u >,
10 4-»
oo o
o
QJ i—
cn <1J
IO > O
S~ (U
ai co co
•* 0 ^
7.983
8.779
8.560
8.709
u
(U
"E"
C7> X>
Q) C +J
(O ' — O
S_ Q. O
0) E i—
^^^
8.230
8.791
8.465
8.621
u
4J
0) (U
C CD
•r- fO
j^: s-
O O)
CO >
1-1 et
103.1
100.1
98.9
99.0
a> to
3 -M O CO
n— *O VO *r~
o r^ co
"^ QJ •+• CQE
"(B 'o.O -M3
-t-> E 0 <->
o
-------
TABLE E-2
HOT ZONE SAMPLING DATA
0
12/14
12/15
12/17
12/18
x-*
3 C i—
+J QJ O
tn -r^ ^>
0 0
S C_) &«
7.9
Lost
7.1
6.4
O) O
E UD
3 4J 1^
1/1 i — (O
•«- O 13
<0
-------
TABLE E-3
ESTIMATED TOTAL GAS EFFLUENT FLOW RATIO
Run No.
Rl
R3
R4
HZ
SCF/Min
34066
37581
37592
SCM/Min
964
1064
1064
S
SCF/Min
44821
50090
50267
SCM/Min
1269
1418
1423
99
-------
APPENDIX F
ANALYTICAL CHEMISTRY DETAILS - NCB TESTS
F.I. Sample Codes, Preparation, and Analysis Procedures
F.2. Recovery Efficiencies for NCB
F.3. Gravimetric and Volumetric Data
F.4. Analytical Results for Representative Waste Feed
F.5. Analytical Results for Probe Washes and Filters
F.6. Analytical Results for Dry Impingers
F.7 Analytical Results for Sorbent Traps
F.8. Analytical Results for Impingers
F.9. Analytical Results for Scrubber Water
F.10. Analytical Results for Gas Bulb Samples
100
-------
F.I SAMPLE CODES, PREPARATION AND ANALYSIS PROCEDURES
F.I.I. Sample Codes
Samples are identified by multi-syllable codes in which:
t The first syllable identifies the test run:
R1(B) = background test on 12-14-76
R2 = unsuccessful NCB test on 12-15-76
R3 = NCB test on 12-17-76
R4 = NCB test on 12-18-76
RO = general sample such as blended representative
waste feed
• The second syllable identifies the sampling location:
-HZ- = hot zone (combustion zone)
-S- = stack
• The third syllable identifies the type of sample or sample
train component:
-PW = probe wash
-F = filter
-II = dry impinger (See page )
ST = sorbent trap
I = impinger
SI = fresh scrubber water
SO = spent scrubber water
GB = gas bulb
• A final syllable indicates the solvent used, if the sample
is an organic extract.
F.I.2. Sample Preparation Procedures
F.I.2.1. Probe Washes
The probe washes were transferred to tared glass evaporating dishes and
and the solvent evaporated on a hot plate. The dishes and intents were
then dried to constant weight in a desiccator over Oriented.
The residue in the evaporating dish, for the hot zone (HZ) probe
washes only, was taken up in methylene chloride. Several small portions
of solvent were used, with agitation in an ultrasonic bath to facilitate
uptake of soluble portions of the residue. Each resulting methylene chloride
suspension was combined with the corresponding hot zone filter for Soxnlet
extraction as described below.
* Trademark of W.A. Hammond Drierite Company, Xenia, Ohio.
101
-------
F.I.2.2. Filters
TU JT^e/i!fers were dried to constant weight in a desiccator over Drierite®
The dried filters were photographed. unerne.
/o, uThe filters for the hot zone only were folded and placed in preextracted
(24 hours with methylene chloride) cellulose Soxhlet thimbles. The
methylene chloride suspension of the corresponding probe wash sample (above)
was poured through the thimble; the solvent which drained through was added
to the boiling flask of each Soxhlet apparatus. The boiling flask was charged
with 200 ml of methylene chloride. The extraction was allowed to proceed for
24 hours.
An empty, preextracted thimble was extracted as a blank.
F.I.2.3. Dry Impinqer
The "Dry Impinger" was an empty, standard-sized impinger, used between
the filter and the sorbent trap in the hot zone sampling train at Rollins.
The sample obtained from this train component consisted of a measured volume
of condensate, combined with water, acetone, and pentane rinses of the impin-
ger and connecting glassware.
In the laboratory, the total sample volume was measured. An aliquot
of the sample was set aside for chloride and nitrate analysis. The remainder
of the sample was transferred to a tared glass evaporating dish. The liquid
was evaporated on a hot plate. The dish and contents were dried to constant
weight in a desiccator over Drierite®
The residue in the evaporating dish was taken up in methylene chloride.
Several small portions of solvent were used, with agitation in an ultrasonic
bath to facilitate uptake of. soluble portions of the residue. The methylene
chloride solutions were transferred to 10 ml volumetric flasks and made to
volume with fresh solvent.
A blank consisting of 1:1:1 water:acetone:pentane was carried through
the evaporation-solution procedure.
F.I.2.4. Sorbent Traps
The sorbent traps were placed in the specially designed extraction
apparatus shown in Figure F-l. Each trap was extracted for 24 hours with
pentane and then for 24 hours with methanol. The extracts were not combined
for analysis. An unused sorbent trap was extracted; this served as a blank.
F.I.2.5. Impingers
Total volumes of the combined impinger solutions* and distilled water
* Not including the "dry" impinger of the hot zone sampling train. See F.I.2.3,
102
-------
Flexible Teflon Coupling
250 Ml Flask
Figure F-l. Sorbent trap extractor.
103
-------
rinses of glassware from each sampling train were measured.
A portion of each solution was acidified to pH <2 (pH paper) with
concentrated nitric acid. This acidified portion was refrigerated and
stored in a Nalgene®* container.
A separate portion of each of the combined solutions was taken for
chloride and nitrate analyses.
F.I.2.6. Scrubber Water
The fresh scrubber water samples collected on 4 different test days
were combined to give a single sample - RO-SI. A portion of this solution
was acidified to pH <2 (pH paper) with concentrated nitric acid and re-
frigerated. A separate portion of the combined sample was taken for chloride
and nitrate analyses. The pH of the combined sample was found to be 6.
The three spent scrubber water samples - R1(B)-SO, R3-SO, and R4-SO -
were not combined. An acidified (pH <2) portion of each was stored in the
refrigerator. A separate portion was taken for chloride and nitrate analy-
ses. The pH's were measured and found to be 4 for R1(B)-SO; 3 for R3-SO-
and 3 for R4-SO. '
For determination of organics, 2 500-ml portions of each of the four
scrubber water samples above and of a distilled water blank were extracted
with 3 x 20 ml each of methylene chloride. The organic extracts were dried
by passing them through anhydrous sodium sulfate.
F.I.3. Analysis Procedures
F.I.3.1. Gas Chromatographic Analysis of Unconcentrated Organic Extracts
The organic solvent extracts prepared, as described above, from the
probe wash plus filter, dry impinger, sorbent trap and scrubber water from
each test's sampling effort were analyzed by gas chromatography prior to
any concentration step. The total volume of each extract was measured, and
5 p£ portions were taken for injection.
The gas Chromatographic conditions were as follows:
instrument = Varian 2700
detector = FID at 295°C
injector = glass lined, 275°C
column = 6' x 1/8" stainless steel packed1 with 10% OV-101 on
100/120 mesh Supelcoport®**
* Trademark of Nalge Company, Rochester, New York.
** Trademark of Supelco, Inc., Bellefonte, Pa.
104
-------
program = Isothermal at 50°C for 3 mln. Linear program
at 20°C/min to 275°C. Hold for 5-10 min at
275°C.
Quantitative calibration of the detector response was achieved by use
of standard solutions of o- and p-nitrochlorobenzene. These two isomers
were not resolved by the chromatographic conditions used. The retention
time, relative to the leading edge of the solvent front, was found to be
constant at 520 ± 12 sec. over the course of the analyses, The quantitative
calibration changed somewhat from one week to the next. Typical calibration
curves are shown in Figure F-2. Injections of calibration standards were
interspersed with injections of samples each day.
F.I.3.2. Concentration of Organic Extracts
After gas chromatographic analysis, each of the organic extracts was
concentrated to a volume of <10 ml.* Concentration was accomplished by
allowing solvent to evaporate from an open container under a gentle stream
of nitrogen. The concentrated extracts were transferred to 10 ml volumet-
ric flasks and the volume restored to 10 ml using fresh solvent.
F.I.3.3. Gravimetric Quantification of Organic Extracts
A 5 ml aliquot (one half of total sample) was withdrawn by volumetric
pipette from each 10 ml extract and transferred to a tared aluminum weighing
dish. The contents of the dishes were allowed to evaporate in a hood at
ambient temperature until constant weight (± 1 mg) was obtained (repetitive
weighings at least 6 hours apart).
F.I.3.4. Infrared Spectroscopy
A portion of each concentrated organic extract was taken for IR
analysis. Spectra were obtained using KBr micropellets, or macropellets for
the more concentrated solutions, using a Perkin-Elmer Model 521 grating
spectrophotometer.
F.I.3.5. Low Resolution Mass Spectrometry (LRMS)
A portion of each concentrated organic extract was taken for LRMS
analysis. Analyses were done using both the batch inlet and the direct
insertion probe of a DuPont (CEC) 21-110B high resolution mass spectro-
meter.
Both qualitative and quantitative results of the LRMS analyses are
reported. The quantitative data given are expressed as estimated percent
* The I-l/CH2Clp extracts were initially 10 ml in total, volume and were
not concentrated further.
105
-------
1000
Curve Used for Samples / ,
. 100h Run In Week of 1/16/77' /
1
E
.o
'+3
TO
D
OJ
C
O)
E
u
10
O)
0)
I
A
X/
Curve Used for Samples Run
on 1/14/77
•*= S ' A Data of 1/14/77
• Data of 1/16/77
A Data of 1/17/77
O Data of 1/18/77
10 100 10oo
NCB Concentration, ppm
Figure F-2. GC calibration curve for NCB standards.
106
-------
abundance of the indicated species, relative to the total quantity of sample
which vaporized in the mass spectrometer inlet. They are not normalized to
correct for nonvolatile portions of the sample. The precision of the esti-
mated values varies with concentration; data for major components are probably
reliable to within 50%. For components at very low concentration, the error
may be as large as a factor of five.
Mass spectrometry on gas bulb samples was done by Gollub Analytical
Services Corporation, Berkeley Heights, New Jersey.
F.I.3.6. Elemental Analyses
These analyses were performed only on the waste feed and not on the
effluent samples. Elemental analysis for probable major components (C, H,
N, Cl, S) was done by Galbraith Laboratories, Knoxville, Tennessee.
Elemental analysis for trace components, especially metals, was done
using spark source mass spectroscopy (SSMS) by Commercial Testing and
Engineering Laboratories, Golden, Colorado. The sample was thermally low
temperature ashed at 350°C for 1 hour in a laboratory furnace in a quartz
crucible prior to analysis.
The ash content of the waste was determined at ADL by ashing in a
muffle furnace at 850°C.
F.I.3.7. Anion Analyses of Aqueous Samples
The stack and hot zone impinger solutions and the scrubber water
samples were analyzed for chloride and nitrate.
For chloride analyses, a suitable aliquot was taken,.oxidized with
hydrogen peroxide, and treated with barium nitrate to precipitate sulfur
species as barium sulfate. The resulting solution was then titrated using
standardized mercuric nitrate titrant with S-diphenyl carbazone as the
endpoint indicator. Standard 0.1 N hydrochloric acid was used as a reference
standard.
For nitrate analyses, chloride was first removed by precipitation with
.silver sulfate. The samples were diluted to the proper range (0.1 to 2.0 ppm
N03~) and analyzed by the phenol disulfonic acid colorimetric method. The
analysis was done on the Coleman Model 55 spectrophotometer at a wavelength
of 410 ran. Reagent grade anhydrous sodium nitrate was used as a reference
standard.
F.2. RECOVERY EFFICIENCIES FOR NCB
To determine whether NCB could, in fact, be recovered if present in
effluent samples, small quantities of NCB waste (taken from the survey
sample) were spiked onto a sorbent trap and into some distilled water
107
-------
(simulated scrubber water). These samples were extracted as described
above for the effluent samples and the extracts analyzed for NCB by quanti-
tative gas chromatography.
For the sorbent trap, 90% recovery of NCB (8.97 mg out of a 10 mg spike)
was found in the pentane extract. The subsequent methanol extract was found
to contain no NCB (<0.2 mg or <2% of the spike).
For the aqueous solution, recoveries in the methylene chloride extracts
were found to be: 86% at 1000 ppm NCB; 73% at 500 ppm NCB; and 66% at
100 ppm NCB.
F.3. GRAVIMETRIC AND VOLUMETRIC DATA
Table F-l presents the measured volumes for the various impinger solu-
tions. These data are used in calculating the percent moisture in the
sampled streams. The total final volume of solution is also used to calcu-
late a mass emission concentration from the measured chloride concentration.
Table F-2 presents gravimetric data for the three types of sample (PW,
F, and HZ-I1) that were dried to constant weight before extraction. For each
test and sampling location, the PW and F numbers in this table are the values
which are summed to estimate the total particulate loading.
Table F-3 presents the results of gravimetric analyses on the various
concentrated extracts. Also included are data for the various blanks which were
run in parallel with the samples.
F.4. ANALYTICAL RESULTS FOR REPRESENTATIVE WASTE FEED
It is necessary to distinguish between two different waste feed samples.
The first, RO-REP(l), is the nitrochlorobenzene waste which was shipped by
the waste producer. The second, RO-REP(2), is the blend of 20% (v) NCB:80%
(v) #2 diesel oil which was actually-fed to the incinerator in tests R3 and
R4. A sample of unadulterated #2 diesel oil was also available; this is
RO-REP(3).
All of the data imply that the "pure" NCB waste, as received in a heated
tank truck, consisted mainly of nitrochlorobenzene. Analysis by GC, GC/MS
and LRMS indicate that about 95% by weight of RO-REP(l) is nitrochlorobenzene.
The IR spectrum is consistent with this, suggesting a mixture of nitrochloro-
benzene isomers. Further confirmation is provided by the elemental analysis:
C H N Cl S
Found for RO-REP(l) 46.14 2.80 8.92 23.28 0.02
Calc'd for C6H4N02C1 45.86 2.56 8.92 22.3
108
-------
TABLE F-l
VOLUMES.OF IMPINGER SOLUTIONS
Run1
Rl(B)
R3
R4
Sample
HZ-I1
HZ-I
S-I
HZ-I1
HZ-I
S-I
HZ-I1 .
HZ-I1
S-I
Volume
Charged
Before Test
ml
0
2000
250
0
2000
250
0
2000
250
Volume
Recovered
After Test
ml
194
2075
550
164
2065
652
144
2055
663
Final Total
Volume
Including
Rinses
ml
460
2650
645
370
2540
700
500
2740
700
109
-------
TABLE F-2
GRAVIMETRIC DATA FOR PROBE WASH, FILTER, AND DRY IMPINGER SAMPLES
Residue After Drying
to
Run Sample Constant Weight, mg.
R1(B) HZ-PW 83.9
HZ-F 7.5
HZ-I1 164.6
S-PW 47.3
S-F 5.9
R3 HZ-PW 334.0
HZ-F 130.5
HZ-I1 400.9
S-PW 5.5
S-F 20.5
R4 HZ-PW 240.0
HZ-F 132.8
HZ-I1 354.9
S-PW 11.6
S-F 19.3
110
-------
TABLE F-3
RESULTS OF GRAVIMETRIC ANALYSES ON CONCENTRATED ORGANIC EXTRACTS
Run
Sample
Total Weight*
R1(B)
R3
R4
RO
HZ-PWF/CH2C12
HZ-I1/CH2C12
HZ-ST/Pentane
HZ-ST/Methanol
SO/CH2C12
HZ-PWF/CH2C12
Hz-n/CH2ci2
HZ-ST/Pentane
HZ-ST/Methanol
SO/CH2C12
HZ-PWF/CH2C12
HZ-I1/CH2C12
HZ-ST/Pentane
HZ-ST/Methanol
SO/CH2C12
PWF Thimble Blank/CH2Cl2
11 Mixed Solvent Blank/CH2Cl2
ST Blank/Pentane
ST Blank/Methanol
SI/CH9C1,
mg
48.4
6.8
23.8
97.4
0.4
76.2
1.6
3.4
150.0
0.0
53.2
47.2
9.0
53.8
0.4
1.2
5.8
1.2
2.0
2.0
* Based on value obtained by evaporating one half of total extract
to dryness.
Ill
-------
The density of the waste was measured and found to be 1.329 g/ml. Compari-
son with literature values* (1.368 for ortho, 1.534 for meta and 1.520 for
para-nitrochlorobenzene at 20 ± 2°C) suggests that the ortho isomer is
predominant.
Other species which were identified in the RO-REP(l) sample of undiluted
NCB waste were dinitrochlorobenzene (about 3%) and nitrobenzene (about 1%).
The LRMS also indicated a homologous series of compounds (about 3% total)
which appear to be styrene-substituted di-tolyl ethers. The lowest member
of this series has MW 198 and a molecular formula, confirmed by HRMS peak
matching, of Ci4Hi40. Others are at 302, 406, and 510 MW. Very small
amounts of hexachlorobutadiene and hexachlorobenzene were also detected.
The waste had a low ash content; the loss on ignition was 99.2%
The only elements which were found by SSMS to be present in the waste
at concentrations (yg/ml) greater than 1 ppm were: calcium, chlorine, iron
and magnesium (all at >1000), sodium (9), silicon (6), sulfur (6), phos-
phorus (5), potassium (4), copper (4), aluminum (3), titanium (2), nickel
(1) and chromium (1). None of these are present at high enough concentration
to cause concern for emissions of toxic metals at the actual feed rates
used in the Rollins tests.**
The #2 diesel oil (RO-REP(3)) was not subjected to detailed analysis.
A gas chromatographic analysis of a diluted sample showed a typical diesel
oil profile. The density was measured and found to be 0.862 g/ml. An elem-
ental analysis revealed the following composition: C, 84.31%; H, 12.57%;
N, 0.2%; ci, 0.71%; and S, 0.14%.
In most significant aspects, the blended waste feed sample, RO-REP(2),
conformed to the properties expected for a 20% (v:v) mixture of the NCB
waste, RO-REP(l), and #2 diesel oil, RO-REP(3). For example, the density
of a 20% by volume blend was calculated to be 0.955 g/ml; the value found
was 0.9549. Quantitative analysis by gas chromatography indicated that the
blend was 19% by volume NCB.
Because of the substantially different densities of the two components,
a 20% by volume NCB mixture would be 27.8% by weight of NCB. The LRMS analy-
sis of the RO-REP(2) sample led to an estimate of 28% NCB which again con-
firms the presumed composition of the blend. When the elemental analysis
results were examined, however, an inconsistency appeared in the data.
* CRC Handbook of Chemistry and Physics, 43rd. ed., Chemical Rubber Publishing
Co., Cleveland, Ohio, 1961, pp. 832-3.
** Given the actual feed rates and total gaseous emission rates, 1 uq/ml of
any element in the NCB waste corresponds to a stack emission rate on the
order of 0.004 mg/m^.
112
-------
Samples of REP(l), REP(2) and REP(3) were resubmitted for analysis and the
original results were confirmed. As indicated below, the blended waste sample
has almost 50% more c). i.Mne content than would be expected for a 27.8% by
weight (20% by volume) mixture of NCB with #2 diesel oil.
C H N Cl S
Calc'd. for 27.8% mix 73.70 9.85 ±2.49 6.98 0.11
Found for RO-REP(2) 73.84 10.08 2.48 9.89 0.10
73.95 10.14 2.68 10.05 0.11
Calc'd. for 41.4% mix 68.34 8.53 £3.70 10.05 0.09
Data for the other elements analyzed in the mixture are consistent with the
27.8% by weight composition of the blend. In order to account for the 10%
chlorine found in the REP(2) sample,, it would be necessary to postulate a
mixture which was more than 40% by weight of NCB. The latter mixture would
not have the carbon, hydrogen or nitrogen content consistent with that found
in the blend.
The reason for the discrepancy between observed and expected chlorine
content has not been identified. It does not arise from errors in the
Galbraith analyses or in post-sampling contamination of the ADL samples,
since Rollins' analysis at the time of testing showed 10.2% chlorine. It
is improbable that the entire batch of blended waste could have been
contaminated in a way which would explain the observations. In the
absence of further information, the mean empirical value of 10.0% chlorine
in the blend has been assumed to be the most reliable estimate of the true
composition.
The LRMS analysis of the blended waste feed revealed, in addition to
the 28% nitrochlorobenzene, 20% alkylnaphthalenes and 49% other hydrocarbons,
mainly aliphatic, up to C3o> as major components. These are expected com-
ponents of the diesel oil. The ether-styrene series of compounds was
present at the expected 1% level. In addition, about 0.1% by weight of
PCB's was detected in the mixture. The PCB's were sought and
found at about the same level in the RO-REP(3), #2 diesel oil sample.
F.5 ANALYTICAL RESULTS FOR PROBE WASHES AND FILTERS
F.5.1 Quantitative Analysis by Gas Chromatography
Waste Components: No peaks were observed in the retention time range
520 j^ 25 sec. relative to the solvent front (for NCB standards, R.T. = 520
+_ 12 sec.). The quantity of NCB in these samples was therefore less than
the limit of detection, or.<0.2 mg in each sample.
113
-------
Other Material: The detector response, after elution of solvent, was
integrated for peaks with retention times (RT) up to 900 sec. Major peaks, equi-
valent quantities as NCB, and total chromatographic material as NCB were:
R1(B) R3 R4
RT
750
781
791
815
858
890
TOTAL
mg, as NCB
2.5
1.4
1.3
5.4
1.3
2.4
19.4
RT
755
784
820
863
895
TOTAL
mg, as NCB
2.6
1.6
5.1
0.6
3.1
17.8
Rt
773
810
818
843
TOTAL
mg, as NCB
2.4
4.1
1.2
4.0
13.0
F.5.2 Qualitative and Quantitative Analysis by IR and LRMS
The IR of the Soxhlet thimble blank showed only weak aliphatic CH2 and
CH3 bands. The LRMS of this sample showed:
Fatty acids ^26%
Homologous CnH2n series ^21%
Phthalates ^20%
Other hydrocarbons ^13%
MW 198, Ci4Hi40 (ditolyl ether)
Triglycerides
The IR of the Rl(B)-HZ-PWF/CH2Cl2 sample revealed strong bands charac-
teristic of diesel fuel, moderate carbonyl bands at 1712 and 1740 cm"',
and weak alkene bands. The LRMS of this sample showed:
Homologous CnH2n series, n even ^38%
Homologous CnH2n series, n odd ^35%
Cl6> C18 Fatt.X acids * 25%
Phthalates ^ 2%
For the R3 test, the IR of the PWF/CH2Cl2 extract was similar to that
of the corresponding background test, except that the carbonyl band was
weaker and a weak band attributable to silicone grease was observed. Sili-
cones (^4%) were confirmed by LRMS. The LRMS also showed:
114
-------
Homologous CnH2n series, n even
Paraffin, C2Q-C40
C]g Fatty acids <\,5%
The R4 PWF sample was like the R3 in the IR spectrum, but showed evi-
dence of possible nitrogen groups (-N02, -NH2, or nitrates) and aromatic
ring stretch. The LRMS spectrum was dominated by one major component which
appeared to be hydroxy octoxy benzophenone (MW 326):
MW 326,
MW 346, C24H420 (346.3253)
Cl6~Cl8 ^tty acids
Nonyl Phenol
m/e 530.47897 (?)
Triglycerides -\,1%
F.6. ANALYTICAL RESULTS FOR DRY IMPINGERS
F.6.1. Quantitative Analysis by Gas Chromatography
Waste components; in the retention time range 520 ± 25 sec. relative to
the solvent front (for NCB standards, R.T. = 520 ± 12 sec), one peak was
observed in the R1(B) sample chromatogram and one in the R4 sample. The
areas, however, corresponded to concentrations of ^ 2.2 ppm of NCB or <0.02
mg in the total extract. No peaks in the 520 ± 25 sec. range were found for
R3-HZ-Il/CH2Cl2.
Other Material : Total detector response for all three chromatograms was
equivalent to <1.0 mg, quantified as NCB. The largest peak in the chromato-
gram in each of the Il/CHgC samples had a retention time of 390 ± 10 sec.
F.6.2. Quantitative and Qualitative Analysis by IR and LRMS
The IR spectrum of the mixed-solvent blank suggested diesel fuel and
aromatic compounds, possibly phenol derivatives. The LRMS indicated 56% of
this sample was a species tentatively identified as hydroxy octoxy benzo-
phenone (MW 326). Other LRMS indications were:
MW 326
Triglycerides
Other hydrocarbons ^12%
Nonyl phenol ^9%
m/e 530 ^5%
Cl6> C18 Fatty acids ^3%
Phthalates *1%
For the background test, R1(B), the IR of the dry impinger sample showed
sulfate and bisulfate (strong) and weak hydrocarbon and OH, NH bands. The
LRMS of this sample showed almost nothing except silicones and sulfate decom-
position products.
115
-------
For the R3-HZ-I1/CH2C12 sample, the IR showed only weak bands for
aliphatlcs, -OH or -NH, and silicones. The LRMS revealed nothing, too
little sample volatilized.
The corresponding R4 sample looked more like the mixed-solvent blank
by both techniques. The IR's were very similar. The LRMS showed:
Triglycerides ^44%
Silicones ^26%
Hydroxy Octoxy Benzophenone? (MW 326)
CIG.CIS Fatty acids
Other Hydrocarbons
F.6.3. Analyses for Chloride and Nitrate
The results of chloride and nitrate analyses for the dry impinger
samples were:
Sample C1" (ppm) N03~ (ppm)
R1(B) - HZ-I1 4850. 90.
R3-HZ-I1 64200. 2010.
R4-HZ-I1 56800. 4780.
F.7. ANALYTICAL RESULTS FOR SORBENT TRAPS
F.7.1. Quantitative Analysis by Gas Chromatography
Waste components: In the pentane extracts of the sorbent traps, a
few peaks were observed in the retention time range 520 H^ 25 sec. relative
to the solvent front (for NCB standards, R.T. = 520 ± 12 sec). Using the
NCB quantitative calibration curve, the calculated quantities of material
in this retention time range were:
R1(B) 0.06 mg
R3 0.02 mg
R4 <0.01 mg
In the methanol extracts of the sorbent traps there were no peaks in
the 520 ± 25 sec. retention time range.
Other material : The detector response, after elution of solvent, was
integrated for peaks with retention times up to 900 sec. Predominant peaks,
116
-------
equivalent quantities as NCB, and total chromatographable material as NCB
were, for the pentane extracts:
Rl(B)-HZ-ST/Pentane
R3-HZ-ST/Pentane
R4-HZ-ST/Pentane
RT, sec
390
495
580
650
710
780
TOTAL
mg as NCB RT, sec mg as NCB
3.8 370 0.4
2.6 460 0.2
1.3
0.7
0.3
0.2 825 2.4
14.4 TOTAL 3.8
RT, sec
380
411
492
579
654
765
778
830
886
TOTAL
mg as NCB
0.4
6.9
2.8
0.8
0.4
0.5
0.5
3.6
5.2
23.5
For the methanol extracts, no peaks were observed in the sample
chromatograms that were not present in the RO-ST/methanol blank. The total
areas of peaks eluting after the solvent but in less than 900 sec. were
(as NCB):
Rl(B)-HZ-ST/methanol
R3-HZ-ST/methanol
R4-HZ-ST/methanol
1.8 mg
1.2 mg
0.4 mg
F.7.2.
Quantitative and Qualitative Analysis by IR and LRMS
Pentane Extracts
The IR spectrum of the ST/Pentane blank showed moderate intensity
bands characteristic of hydrocarbons, and weak bands characteristic of a
carbonyl (1730 cm~l), silicon grease, and -OH or -NH.' The LRMS on this
blank showed:
117
-------
Paraffins ^77%
Napthalene °AQ%
Alkyl Napthalenes * 4%
Anthracene % 1%
MW 340 (C22H2803) * 5%
Triglycerides ^ 2%
m/e 530 * IX
For the R1(B) background test, the IR spectrum of the ST/Pentane
extract was similar to that of the blank: silicone grease (moderate);
hydrocarbon (moderate): -OH or -NH (weak) and carbonyl at 1740 cm'1 (weak)
The LRMS on this sample suggested the following:
Si li cones
Benzoic Acid ^23%
Aliphatic Hydrocarbons,
m/e 400-470 M5%
C13' C15' C17 Fatty Acids ^10%
Di butyl tin Chloride *11%
The presence of the organic tin compound was unexpected. Its identity
was confirmed by HRMS peak matching as well as by the distinctive tin isotope
pattern. It is expected that this compound may be due to carry-over or
incomplete purging of the Rollins facility, since "metal alkyls" were
among the wastes burned at the facility during the week of ADL's testing.
The IR spectrum of the R3-HZ-ST/Pentane extract was similar to that
of the background sample: silicone grease (strong), hydrocarbon (moderate),
-OH or -NH [moderate), carbonyl at 1730 cm"! (moderate) and alkene on
C-C1 (weak). The LRMS showed mostly benzoic acid:
Benzoic Acid
Alkyl Benzenes ^ 9%
Toluic Acid * 2%
Ethyl Benzoate ^ 2%
C10H1302C1 - 1%
Phenol ~ 1%
Silicone * 1%
For the R4 test, the IR spectrum was much like that of the R3 sample,
except that the hydrocarbon bands were stronger and the carbonyl band
appeared at 1700 cnH . The LRMS indicated that, as in the background test,
dibutyl tin chloride was a significant component:
118
-------
Di butyl Tin Chloride
Aliphatic Hydrocarbons ^24%
Si li cone
C12* C14' Fatty Acids * 6%
Alkyl Napthalenes * 2%
Anthracene * 2%
s
Again, it is suspected that the organic tin compound is probably a carry
over from a prior Rollins waste burn.
Methanol Extracts
The RO-ST/methanol blank had an IR spectrum with moderate intensity
bands for aliphatic CHg, CH3, -OH, or -NH, SO/ and COf , and weak bands
for amides. The LRMS indicated:
Benzoic Acid ^63%
MW 326 (hydroxy octoxy
_ . . ... benzophenone) „ ..,
Toluic Acid r * 4%
C16' C18 Fatty Acids * 4
Ethyl Benzoate * 4%
m/e 530 * 1%
For the background test, the ST/methanol extract IR spectrum showed
moderate intensity hydrocarbon and -OH or -NH bands and weak bands indicating
an ether or alcohol; There was also strong evidence of an amide linkage.
The LRMS of this sample showed very substantial amounts of HC1 (^78% of the
total LRMS intensity). Other LRMS indications were:
Sulfate Decomposition Products ^ 8%
Dihydrofuran, MW 70 -x. 4%
Furfural or Furan Aldehyde -\, 4%
m/e 84 (possible C5HQ0) ^ 3%
Phthalates ^2%
Di butyl Tin Chloride * 1%
For the R3- and R4- HZ- ST/methanol, the IR spectra were the same
as that of the corresponding R1(B) sample. The LRMS results for the R3
and R4 methanol extracts were:
119
-------
R3 R4
HC1 ^64% ^36%
Various N- Containing Decomposition
Products (C10H14N202, C4H8N02, C^N) ^6% ^9%
Unidentified (numerous small peaks) M6% M8%
Sulfate Decomposition Products ^3% ^2%
Phthalates * 1%
Dibutyl Tin Chloride * 1%
Benzoic Acid -\, 2%
Methyl Benzoate -x, 4%
The species referred to as "decomposition products" in the LRMS data
summary are so called because they appear at much higher probe temperatures
(^300°C) than is consistent with their low molecular weight. These species
are apparently formed by pyrolysis of the sample in the heated probe.
F.8 ANALYTICAL RESULTS FOR IMPINGERS
The impinger solutions originally charged at the start of sample
collection were 0.1N sodium acetate for all stack samples and the R1(B)
hot zone sample and 5N sodium acetate for the R3 and R4 hot zone samples.
The results of the chloride and nitrate analyses were:
Sample C1" NO
R1(B) - S - I 5.5 0.7
R1(B) -HZ - I 120.0 41.8
R3 - S - I 11.5 5.2
R3 - HZ - I 685.0 45.0
R4 - S - I 35.0 4.0
R4 - HZ - I 1775.0 82.0
F. 9 ANALYTICAL RESULTS FOR SCRUBBER WATER
F.9.1 Quantitative Analysis of Organic Extracts by Gas Chromatography
No peaks were found in any of the samples in the retention time
range (520 +_ 25 sec) which included the retention time of NCB. There
were no significant peaks in the R3-or R4- SO/CH2C12 extracts that were
not present in the SI or blank extracts. The only significant peak in all
of the chromatograms had a retention time of 700 sec relative to the solvent
front and an intensity corresponding to <1 mg as NCB.
120
-------
F.9.2 Qualitative and Quantitative Analysis by IR and LRMS
The IR spectrum of the fresh scrubber water extract (RO-SI/CH2C12)
showed bands characteristic of sulfate (very strong), CH2, CHs (weak),
-OH or -NH (weak) and carbonyl at 1730 cnH (moderate). The LRMS of this
extract showed:
Nonylphenol Dimer
Unidentified (numerous peaks)
Nonylphenol
Phthalate
Piperidine * 8%
Fatty Acids ^ 1%
The IR spectra of the R1(B)-, R3-, and R4- SO/CH2C12 extracts were
very similar to those of the RO-SI extract, except that some silicone
bands were observed. The LRMS spectra, again, were similar to those of
the Si-extract. The LRMS spectra did indicate the presence of the
ditolylether species, previously identified in the representative waste
feed. Some mono- and di- brominated analogs appeared to be present in
the R1(B)- SO/CH2C12 extract. Nitrogen-containing species amounted to
about 15% of the total LRMS intensity for the spent scrubber water
effluents.
F.9.3 Analyses for Chloride and Nitrate
Samples Cl". ppm NO^- . ppm
RO - SI 23.5 0.3
R1(B) - SO 208 47.8
R3 - SO 1750 57.2
R4 - SO 1815 53.5
F.10 ANALYTICAL RESULTS FOR GAS BULB SAMPLES
In the GC/MS analysis of the gas bulb samples, collected from the
by-pass line of the hydrocarbon analyses, the analyst was asked to search
specifically for the possible presence of chlorinated hydrocarbons. None
were detected. However, the composition found for these samples by
GC/MS suggest that the gas bulbs had leaked to the extent that the analytical
results were not meaningful. This may have been due to unequal coefficients
of thermal expansion for Teflon (stopcocks) and glass (bulb) if dramatic
temperature changes occurred during sample shipment.
121
-------
Constituents R1(B)- HZ- GB R3- HZ- GB R4- HZ- GB
Concentration, % (v/v)
Nitrogen 78+ 79+ 73+
Oxygen 20.8 16.5 20.7
Argon 0.93 0.93 0.94
Carbon Dioxide 0.042 3.1 0.12
Hydrogen <0.005 <0.005 <0.005
Methane <0.005 <0.005 <0.005
Chlorinated Hydrocarbons <0.005 <0.005 <0.005
122
-------
APPENDIX G-l
ROLLINS ENVIRONMENTAL SERVICES, INC,
OPERATING DATA
FOR THE
THERMAL DESTRUCTION
OF
CHEMICAL WASTE
FOR
TRW DEFENSE AND SPACE SYSTEMS GROUP,
TRW, INC.
123
-------
CONTENTS
Introduction
Background Test - TRW
Background Test - ADL
Hammermilled Capacitors - TRW
Whole Capacitors - TRW
NCB Test No. 1 - ADL
NCB Test No. 2 - ADL
Cost Estimates
124
-------
INTRODUCTION
Rollins Environmental Services, Inc. ("RES") operates at its Houston,
Texas, facility an industrial waste incinerator capable of thermally
destroying liquid, solid and gas wastes. This unit operates at com-
bustion temperatures in excess of 2000°F with a retention time be-
tween 2 to 3 seconds. All combustion gases are quenched and scrubbed
by a venturi, flexitray scrubber.
This unit, operating 24 hours per day - 7 days per week, is unique in
that it has two separate burners feeding a common afterburner. The
main burner (loddby) is for waste liquids only and is a vortex type
horizontal burner. The second burner is a rotary kiln capable of
burning waste liquid, sludges, and solids via fiber packs. Both
burners have natural gas igniters and gas burners for the purpose of
initial refractory heat-up, flame stability and supplemental heat,
if necessary.
The combustion gases from the above mentioned burners combine in the
afterburner with a horizontal cyclone effect. This prolongs the
residence time and completes the combustion process prior to the gases
leaving the afterburner. From here the gases pass through a duct and
enter a wet venturi scrubber, flexitray and demister. The venturi
pressure drop is capable of being varied, however, it is normally
operated sufficiently so as to scrub all but submicron particles.
Two 400 horsepower induced draft fans in series provide the energy
for the scrubber and maintain the entire incinerator at a negative
pressure so as to contain all toxic gases. On leaving the fans the
gases pass up a one-hundred foot stack to the environment.
125
-------
^ot-id (''aa-te Feed Chute
Overall length 35'
ROLLINS ENl/IKO.VMEMTAL SERVICES
INCINERATOR 5V5TEM COWFRGURATION
HOUSTON PLANT
Total:850 gpm
270 gpm to venturi
tangential nozzle^**
T/C F*eA8° gpm t0 flexitray
Hot Vuct tta.te.1
Feed
x^^v
GO.A
Feed Wa-ate Liquid Loddby length 16'
Loddby diameter 5'3"
Sciubbe.fi
l:>ate.fi
12/29/76
-------
Destruction testing on two types of waste products was accomplished,
the first being polychlorinated biphenyls (PCB) and the second nitro-
chlorobenzene tars (NCB). Background data was gathered while burning
No. 2 oil and then the wastes were burned on subsequent burns using
No. 2 oil as the supplementary fuel. The actual conditions of each
test will be discussed separately.
127
-------
BACKGROUND TEST - TRW
12/06/76
Before all tests a purge burn using No. 2 oil was done for a period
of time not less than one hour. This provided for at least six turn-
overs of scrubber water and sufficient purging of the combustion
chambers.
The TRW background test started at 10:45 A.M. with the purge commencing
at 8:45 A.M. Both the loddby and kiln waste liquid burners operated
on No. 2 fuel oil with a combined feed rate of 552 gallons per hour.
Heat release averaged 74.6 million BTU per hour with the average tem-
peratures of the respective points indicated below:
Hot duct 1996°F
Kiln duct 703°F
Afterburner 2387°F
Loddby flame 2705°F
Kiln flame 2382°F
Venturi delta P 41 inches of water
Lime consumption 101 GPH
Natural gas consumption 2095 CFH
The test concluded at 1400 hours due to severe weather conditions.
128
-------
ROLLINS ENVIRONMENTAL SERVICES
INCINERATOR 5VSTEM CONFIGURATION
HOUSTON PLANT
Solid Va&te. Feed Chafe.
T/C Kiln Exit Gat 703°F
A^te.ibaine.1 2387 F
Feed Wa4.£e Liquid
Bu.ine.1*
1996°F
T/C
Hot Vuct Wa.te.ft.
Feed
Hydiate.d Lime. Slutiy Vi&cha.ige.
Feed
101 gph
72/29/76
-------
BACKGROUND TEST - ADL
12/14/76
The oil purge started at 0900 hours and the background test secured
at 1630 hours. As before, both the loddby and kiln were fired on
diesel fuel, however, the combined feed rate averaged 576 gph with
an average heat release of 77.8 million BTU's per hour. The opera-
ting data was as follows:
Hot duct
Kiln duct
Afterburner
Loddby flame
Kiln flame
Venturi delta P
Lime consumption
Natural gas consumption
Retention time
1963°F
688°F
2351°F
2723°F
2354°F
36 inches of water
101 GPH
4923 CFH
2.6 seconds
130
-------
ROLLINS ENVIRONMENTAL SERVICES
INCINERATOR SVSTEM CONFRGURATIOW
HOUSTON PLANT
1963°F
T/C
Ho-t OucX Wa.teft
Feed
C'aa-te Feed Cftu*e
Feed
Bun.ne.iA
101 gph
12/29/76
-------
The test destruction of PCB waste was accomplished burning capacitors
in two different states. The first test was conducted with capacitors
that were hammermilled and sealed in 35 gallon fiberpacks, and the
second test burned whole capacitors.
HAMMERMILLED CAPACITORS - TRW (12/08/76)
In this burn No. 2 oil was fired in the loddby and kiln while the
whole fiberpacks were fed into the kiln at a rate varying from one
every 4 to 5 minutes. Their weights and feed time are set forth in
Figure IV. The total weight of capacitors and fiberpacks was 2558
pounds. As in the background tests, a purge burn was accomplished
and started 0830 hours with the test burn commencing at 1000 hours.
The test was secured at 1330 hours with a total of 46 fiberpacks
burned. The fuel oil burned in the kiln averaged 71 gph or 9.7 million
BTU per hour, while the loddby burn rate was 566 gph or 76.5 million
BTU per hour. The operating data was as follows:
Hot duct
Kiln duct
Afterburner
Loddby flame
Kiln flame
Venturi delta P
Lime consumption
Natural gas consumption
Retention time
1993°F
910°F
2428°F
2731°F
2'286°F
40 inches of water
101 GPH
3200 CFH
3.2 seconds
132
-------
FIGURE IV
HAMMERMILLED CAPACITORS
DRUM NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
TIME DROP
0955
1005
1010
1014
1018
1022
1026
1030
1034
1038
1042
1046
1050
1054
1100
1204
1108
1112
1116
1120
1124
1128
1132
1136
1140
1144
1148
1152
1156
1200
WEIGHT
50 Ibs.
52 Ibs.
51 Ibs.
45 Ibs.
68 Ibs.
45 Ibs.
48 Ibs.
81 Ibs.
67 Ibs.
67 Ibs.
50 Ibs.
46 Ibs.
51 Ibs.
75 Ibs.
52 Ibs.
50 Ibs.
56 Ibs.
49 Ibs.
56 Ibs.
75 Ibs.
65 Ibs.
47 Ibs.
55 Ibs.
69 Ibs.
44 Ibs.
46 Ibs.
47 Ibs.
46 Ibs.
43 Ibs.
43 Ibs.
133
-------
FIGURE IV
HAMMERMILLED CAPACITORS: CONTINUED
DRUM NO. TIME DROP WEIGHT
31 1204 47 Ibs.
32 1208 56 Ibs.
33 1212 63 Ibs.
34 1216 61 Ibs.
35 1220 . 67 Ibs.
36 1225 56 Ibs.
37 1240 53 Ibs.
3« 1245 45 Ibs.
39 1250 80 Ibs.
4° 1255 50 Ibs.
41 1300 43 Ibs.
42 1305 53 Ibs.
43 1310 40 Ibs.
44 1315 78 Ibs.
45 1320 70 Ibs.
46 1325 57 Ibs.
134
-------
ROLLINS EMI/IR0MMENTAL SERVICES
IMCINEKAT0R 5VSTEM C0NFRGURATI0H
HOUSTON PLANT
lotid Watte. Feed Chu-tc
T/C Kiln Ex-U GaA 910°F
CO
en
2428°F
1993°F
T/C rm&k
Hot Vuct Watt*.
Feed
T
L>
:e(
101 gph
Hydiattd Lime. Stuiiy
Feed
12/29/76
-------
WHOLE CAPACITORS - TRW (12/09/76)
As was in the case of the hammermilled capacitors, fuel oil was
burned in the kiln and loddby while whole capacitors were fed evenly
as possible over a 5 hour-15 minute test period. Purge started at
1045 hours with capacitors starting at 1200 hours and ending at
1515 hours. Six drums of capacitors were fed with the average weight
being approximately 450 pounds per drum. Feed rates to the loddby
averaged 460 pgh or 62.2 million BTU per hour while the kiln averaged
148 gph or 20 million BTU per hours. Operating data was as follows:
Hot duct
Kiln duct
Afterburner
Loddby flame
Kiln flame
Venturi delta P
Lime consumption
Natural gas consumption
Retention time
2005°F
919°F
2429°F
2748°F
2442°F
40 inches of water
133 GPH
3100 CFH
3.0 seconds
136
-------
ROLLINS EMVIR0VMEMTAL SERVICES
IVCINERATCR 5/STEM C0NFRGURATI0N
HOUSTON PLANT
00
d ^ ~\
LTV
X "\
Hydiate.d Lime. Stuiiy
Feed Sciufafaet
Feed
Bui.rte/i.4
12/29/76
-------
The destruction of NCB tars was tested twice under similar conditions
except that the hot duct temperature was lowered for the second test.
The NCB tars were diluted with No. 2 oil to a blend of 20% NCB and
80% fuel oil by volume. Analysis of this blend was as follows:
BTU per pound 17584
Pounds per gallon 7.88
BTU per gallon 138.562
Scrub pound per pound .156
Chlorine 10.1% by wt.
Ash 0.8% by wt.
NCB TEST NO. 1 - API (12/17/76)
This test started its purge at 0900 hours with the sampling commencing
at 1045 hours and ending at 1400 hours. The kiln was fed No. 2 oil
at a feed rate of 65.8 gph or 8.9 million BTU per hour while the
loddby was fed the blend at 534 gph or 74 million BTU per hour. The
pounds per hour of NCB burned in the loddby averaged 1099. The opera-
ting data was as follows:
Hot duct 1966°F
Kiln duct 790°F
Afterburner 2385°F
Loddby flame 2818°F
Kiln flame 1714°F
Venturi delta P 43 inches of water
Lime consumption 106 GPH
Natural gas consumption 4065 CFH
Retention time 2.33 seconds
NCB TEST NO. 2 - ADL (12/18/76)
The purge started at 0930 hours and testing began at 1030 hours and
finished at 1330 hours. The same blend was fed in the loddby and
No. 2 oil again was burned in the kiln. The loddby feed rate was
138
-------
ROLLINS ENVIRONMENTAL SERl/ICES
INCINERATOR 5V5TEM C'JNFRGURATION
HOUSTON PLANT
1966°F
o.
-------
ROLLINS ENVIRONMENTAL SERVICES
INCINERATOR 5V5TEH CONFRGURATIOM
HOUSTON PLANT
1823°F
I
olid (>>a&te. Feed Chute.
Feed
Mate.*.
Feed s^ *"\
KV
AbAOibi
T
Li
ee
165 gph
Hydfiatid Lime.
Feed
Induced
Via &t
Vi& chafige.
Sciubfae*
12/29/76
-------
466 gph or 64.6 million BTU per hour, while the kiln was 104 gph or
14 million BTU per hour. The above loddby feed rate gave an average
input of 959 pounds per hour of NCB to the loddby. The operating
date was as follows:
Hot duct 1823°F
Kiln duct 747°F
Afterburner 2430 F
Loddby flame 2750°F
Kiln flame Not available
Venturi delta P 43 inches of water
Lime consumption 165 GPH
Natural gas consumption 3500 CFH
Retention time 2.31 seconds
141
-------
COST ESTIMATES
Capital Investment Needed
for Turnkey installation
including land, tanks, scrubber ponds,
utiltities and incineration system
$5,000,000
Operating personnel requirements
1 Supervisor per shift 7 days per week
3 operators per shift 7 days per week
1 lab technician 2 shifts per day 7 days
per week
4 mechanics one shift per day
5 days per week
Levels of skill include: pipefitting
machinist
welder
millwright
Approximate estimate of yearly maintenance expense
less labor
$
196,000
Approximate estimate of yearly maintenance labor other
than in house- i.e. refractory, fiberglass, and electrical
$ 85,000
142
-------
APPENDIX G-2
INCINERATION SYSTEM RAW DATA
143
-------
TABLE G-l
TEST I - TRW BACKGROUND
Date 12/6/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp ( F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(rtnips)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
0845
1050
750
2100
2540
2370
62
430 ,
7 460
42"
193C86
0900
2050
700
2420
2700
2280
62
420 ,
' 460
41"
0915
2000
800
2370
2600
2350
62
430 .
' 460
42"
0930
2000
700
2440
2760
2370
62
430 .
' 460
41"
0945
2005
700
2500
2710
2400
62
430
' 460
41"
1000
2000
700
2510
2710
2410
62
430 ,
' 460
41"
1015
1950
750
2410
2700
2420
62
430 .
' 460
41"
1030
2000
700
2320
2710
23CC
62
410 .
' 450
41"
1045
1975
700
2330
2700
2380
62
420
1 460
41"
193886
1100
2000
700
2320
2710
2400
62
420 .
7 460
41"
1115
2000
750
2700
62
420
1 460
41"
1130
2000
750
2320
2710
1
2410
62
430
' 460
41"
-------
TABLE G-l
TEST I - TRW BACKGROUND
(continued)
Date 12/6/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
1145
2000
750
2430
2720
2310
62
410 .
1 450
41"
1200
1975
700
2410
2700
2330
62
420
1 460
41"
1215
2000
700
2400
2700
2400
62
420 .
' 450
41"
193894
1230
2000
700
2420
2710
2390
62
420
' 460
41"
1245
2050
750
2430
2720
2380
62
410 ,
' 460
42"
1300
2025
750
2420
2720
2400
62
420
' 460
41"
1315
2000
600
2400
2700
2380
62
410 ,
' 450
41"
1330
1975
600
2410
2710
2400
62
420 .
' 460
41"
1345
2000
700
2410
2700
2400
62
410 .
' 460
41"
1400
1950
700
2400
2680
2370
62
420 .
' 460
41"
193807
1415
1430
-Pi
cn
-------
TABLE G-2
TEST II - HAMMERMILLED PCB CAPACITORS
Date 12/8/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp ( F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
0830
1900
1000
2450
2720
Alkylson
62
420 ,
1 450
38"
194057
0845
2000
1000
2450
2720
Alkylson
62
410
' 440
38"
0900
2000
950
2500
2700
Securing
Alkyls
62
410 ,
' 450
38"
0915
2100
950
2475
2710
Starting
Diesel
Purge
62
410 .
1 440
38"
194059
0930
1900
850
2450
2725
Couldn ' t
get good
reading
62
410
' 450
38"
0945
2050
900
2460
2800
2480
62
430
' 460
38"
1000
2000
850
2480
2810
2490
62
430 .
' 460
40"
194061
1015
2000
800
2390
2680
2400
62
420 .
' 450
40"
1030
1975
800
2310
2700
2240
62
430 .
' 460
40"
1045
2005
850
2330
2740
2500
62
430
' 460
40"
1100
2000
900
2450
2740
2370
62
420 .
' 450
40"
1115
2000
900
2470
2810
2800
62
420 ,
'450
40"
en
-------
TABLE G-2
TEST II - HAMMERMILLED PCB CAPACITORS
(continued)
Date 12/8/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Readiny
1130
2000
900
2420
2740
2240
62
430
' 460
40"
1145
1975
850
2400
2710
2100
62
430
' 460
40"
1200
2000
900
2410
2710
2110
62
420 .
' 450
40"
1215
2000
850
2400
2720
2100
62
410 .
' 450
40"
1230
2000
950
2420
2730
2120
62
420 .
' 450
40"
1245
2000
975
2450
2740
2120
62
420
' 450
40"
1300
2000
1000
2420
2710
2150
62
430
' 460
40"
1315
2000
900
2410
2730
2170
62
430
' 450
40"
1330
2000
950
2450
2720.
2200
62
430
1 460
40"
194073
1345
1950
1000
1985
2720
-
62
430
' 460
flO"
1
1400
1415
1
-p.
•-J
-------
TABLE G-3
TEST III - WHOLE PCB CAPACITORS
Date 12/9/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
1045
1850
950
2330
2720
2420
62
430 .
' 460
40"
194142
1100
1850
1000
2380
2730
2470
62
420 ,
' 450
40"
1115
1900
1000
2370
2700
2470
62
420 .
' 450
40"
1130
2075
1000
2670
2790
2470
62
410 ,
' 450
40"
1145
1950
1000
2480
2720
2450
62
410 .
' 450
40"
1200
1900
1000
2390
2740
2390
62
410
' 450
40"
1215
1975
1000
2420
2700
2470
62
410 .
' 450
40"
1230
1900
950
2330
2710
2420
62
410 .
' 450
40"
1245
2050
975
2530
2790
2460
62
410
' 450
40"
1300
2050
975
2490
2770
2460
62
410
' 450
40"
1315
2050
950
2420
2750
2430
62
420 ,
1 450
40"
1330
2050
950
2430
2740
2470
62
420 .
' 450
40"
oo
-------
TABLE G-3
TEST III - WHOLE PCB CAPACITORS
(continued)
Date 12/9/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp ( F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
1345
2050
900
2420
2750
2460
62
420 .
' 460
40"
1400
2000
900
2440
2760
2480
62
420 ,
1 460
40"
1415
2000
900
2440
2770
2470
62
420 .
' 460
40"
1430
2000
850
2400
2770
2430.
'2450
62
420 ,
' 460
40"
1445
1950
825
2400
2750
2420
62
415 ,
' 440
40"
1500
2050
850
2450
2730
2410
62
415
' 440
40"
1515
2050
850
2450
2750
2400
62
420 ,
' 460
40"
194156
1530
1545
1600
1615
1630
•o
-------
TABLE G-4
RUN R1(B) BACKGROUND TEST
Date 12/14/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
0900
2050
650
2400
2780
2400
62
415
' 415
35"
194600
0915
2000
650
2350
2750
2425
62
415
' 415
35"
START
0930
2000
650
2400
2750
2450
62
415
' 415
35"
0945
1950
600
2425
2710
2345
62
415
' 415
35"
1000
1950
575
2400
2720
2280
62
415
'415
35"
1015
1950
575
2400
2720
2350
63
418
' 415
36"
1030
1925
600
2360
2720
2350
63
418
' 415
35"
1045
1975
625
2360
2730
2370
63
417
' 415
35"
1100
2000"
650
2360
2720
2350
63
417
'415
35"
1115
1950
650
2350
2730
2350
63
117 ,
' 415
35"
1130
1950
650
2350
2730
2340
63
415
'415
36"
1145
1950
650
2360
2710
2320
63
415
'415
36"
en
o
-------
TABLE G-4
RUN R1(B) BACKGROUND TEST
(continued)
Date 12/14/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
1200
1950
650
2330
2730
2350
63
430 .
' 430
36"
1215
1950
650
2360
2740
2350
63
430 .
' 430
36"
1230
1950
650
2360
2740
2350
63
430 .
' 430
36"
1245
1925
650
2320
2700
2300
63
430 ,
'430
36"
1300
1900
625
2290
2600
2310
63
430 ,
' 430
36"
1315
1925
650
2330
2720
2340
63
430 .
' 430
36"
1330
1950
650
2350
2720
2350
63
430 ,
' 430
36"
1345
1950
750'
2340
2720
2350
63
430 .
' 430
36"
1400
1950
750
2310
2750
2310
62
420 .
' 430
36"
1415
1950
750
2330
2750
2320
62
420 .
1 430
36"
1430
1.975
750
2350
2730
2340
62
415
' 415
36"
1445
2000
800
2330
2730
2420
62
415 .
1 415
36"
-------
TABLE G-4
RUN R1(B) BACKGROUND TEST
(continued)
Oate 12/14/76 T1me
Hot Duct Temp (°F)
Kiln Temp (PF)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Gas Meter Reading
1500
2000
800
2330
2710
2400
62
430 ,
' 420
36"
1515
2000
800
2330
2710
2400
62
430 ,
' 420
36"
1530
2000
800
2330
2730
2390
62
430 ,
1 420
36"
1545
1975
800
2350
2730
2360
62
430 .
'430
36"
1600
1975
800
2350
2720
2350
62
430 .
' 430
36"
1615
1950
750
2350
2720
2330
62
430 .
' 430
36"
END TEST
1630
1950
750
2350
2720
2330
62
430 .
' 430
36"
FINISH
194632
1645
1700
•
1715
1730
1745
en
ro
-------
TABLE G-5
RUN R2 - NCB TO KILN
Date 12/15/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Oelta-P Venturi
(In. H20)
Loddby Feed Tank
Level (Inches) *
Gas Meter Reading
0900
1200
600
1950
2400
2200
60
440 ,
' 41C
39"
65"
194674
0915
1300
650
1975
2450
2310
60
440 .
' 410
39"
68"
START
0930
«
1500
700
2150
2525
2430
62
440 ,
' 410
40"
70"
0945
1800
750
2350
2600
2470
62
460 ,
' 430
&2"
73"
1000
1950
750
2400
2650
2500
62
450 .
' 420
42"
75"
1015
1875
750
2360
2650
2570
62
440 ,
' 410
42"
77"
1030
2100
750
2460
2650
3300
62
440 .
' 410
a2"
79"
1045
2000
750
2430
?640
2360
62
450 .
' 410
42"
81"
1100
1800
750
2690
2670
2150
62
460 .
' 420
42"
83"
1115
1850
700
2580
2690
2270
62
440 .
' 410
41"
85"
1130
2000
775
2510
2740
2290
62
460 ,
' 410
41"
87"
1145
1850
. 775
2510
2650
2370
62
460 .
' 410
41"
88"
in
GO
* 1 inch - 69 gal Ions
-------
TABLE G-5
RUN R2 - NCB TO KILN
(continued)
Date 12/15/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Loddby Feed Tank
Level (Inches)*
i
1200
1850
700
2340
2770
2290
62
460 .
' 420
43"
90"
1215
1900
700
2520
2890
2230
62
460 ,
' 420
42"
91"
1230
1950
700
2400
2800
2250
62
470 ,
' 430
40"
93"
«
1245
1950
675
2250
2290
2250
62
470 .
' 430
39"
95"
1300
1200
575
2600
2810
1550
60
470 ,
' 430
37"
96"
1315
1900
600
2470
2600
1800
61
450 .
' 410
42"
97"
1330
2100
650
2750
2900
2000
61
470 .
' 420
42"
99"
1345
1900
600
2470
2660
2000
61
470 ,
' 420
42"
101"
1400
1950
600
2600
2750
2250
61
450 .
' 410
42"
103"
1415
1950
600
2480
2760
2250
61
450 .
' 410
42"
104"
1430
1950
550
2370
2700
2250
61
440 .
' 410
42"
106"
1445
1900
550
2350
2710
2250
61
440 .
' 410
42"
108"
-------
TABLE G-5
RUN R2 - NCB TO KILN
(.continued)
Date 12/15/76 T'me
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Loddby Feed Tank
Level (Inches)
Gas Meter Reading
1500
1900
550
2330
2720
2220
61
450 .
' 420
41"
110"
1515
1900
550
2330
2710
2200
61
450 .
' 420
41"
111'1
1530
1900
550
2320
2710
2170
61
460 ,
' 420
41"
112"
END TEST
1540
1900
550
2300
2700
2150
61
450 .
' 420
41"
113"
FINISH
194713
!
i
en
tn
-------
TABLE 6-6
RUN R3 - NCB - DIESEL FUEL TO LODDBY
Date 12/17/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp ('FJ
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Oelta-P Venturi i
(In. H20) !
Loddby Feed Tank Blend Level
(IN.)80r: =2
22 NCB *
Lime Feed Tank Level
(Inches)**
Gas Meter Reading
0845
36"
40"
START
194893
0900
1900
800
2400
2460
301 - 410
302 - 450
42"
37"
41"
0915
2150
800
2600
2750
301 - 410
302 - 480
42"
39"
41"
0930
1950
750
2350
2700
301 - 410
302 - 450
44"
42"
41"
0945
1900
700
2450
2740
301 -410
302 - 460
42"
44"
41"
1000
1900
750
2400
2780
301 -410
302 - 460
43"
45"
41"
1015
1950
800
2450
2840
301 -410
302 - 460
43"
46"
42"
1030
1950
800
2450
2800
301 - 430
302 - 460
43"
48"
42"
1045
1975
800
2410
2830
2370
301 - 420
302 - 460
43"
50"
42"
1100
1975
800
2400
2830
2350
301 -420
302 - 460
43"
52"
43"
1115
1975
775
2430
2900
2380
301 - 430
302 - 460
54"
43"
1130
2000
790
2440
2920
2400
301 -420
302 - 460
56"
44"
* 1 inch - 69 gallons
** 32% Ca(OH)2 by wt., 1 inch - 110 gallons
-------
TABLE G-6
RUN R3 - NCB - DIESEL FUEL TO LODDBY
(continued)
Date 12/17/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Del ta-P Venturi
(In. H20)
Loaaoy Feeo TanK Bleno
Level (In.) 80% #2
2", NC3-* '
Lime Feed Tank Level
(Inches)**
Gas Meter Reading
1145
1950
790
2380
2790
2300
301 430
302 460
43"
58"
44"
1200
1950
790
2440
2800
2300
301 430
302 460
43"
59"
44"
1215
1950
790
2410
2810
2330
301 430
302 460
43"
60"
44"
1230
1950
790
2440
2820
2300
301 430
302 460
43"
63"
44"
1245
1950
790
2420
2800
2300
301 430
302 460
43"
64"
44"
1300
1950
790
2430
2840
2300
301 410
302 460
43"
67"
45"
1315
1950
790
2360
2780
2300
301 410
302 460
44 "
70"
45"
1330
1950
790
2240
2770
2300
301 410
302 460
44"
72"
46"
1345
2000
790
2290
2770
2310
301 410
302 460
44 "
73"
46"
1400
2000
790
2300
2800
2310
301 410
302 460
44"
74"
46"
1415
76"
48"
STOP
194914
1430
* 1 inch - 69 gallons
** 32% Ca(OH)2 by wt., 1 inch - 110 gallons
-------
TABLE G-7
RUN R4 - NCB - DIESEL FUEL TO LODDBY
Date 12/18/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Loddby Feed Tank Blend
Level (In. ) 80V, #2
2% NCB * :
Lime Feed Tank Level !
(Inches)**
Gas Meter Reading
0930
1725
750
2130
2670
120 .
' 450
40"
75"
42"
START
194962
0945
1900
800
2310
2700
440 .
' 470
43"
77"
43"
1000
1750
700
2210
2660
430 ,
' 460
42"
78"
43"
1015
1900
800
2310
2700
430 .
1 460
42"
80"
43"
1030
1825
790
2350
2620
430 .
' 460
43"
81"
44"
1045
1800
750
2350
2700
430 .
1 460
42"
83"
44"
1100
1800
700
2350
2720
430 ,
' 460
42"
85"
44"
1115
1825
700
2400
2710
430 .
' 460
43"
87"
44"
1130
1800
710
2420
2740
430 .
' 460
43"
89"
45"
1145
1850
?80
2450
2760
430 ,
' 460
43"
90"
45"
1200
1850
800
2450
2780
430 ,
'460
43"
92"
45"
1215
1875
800
2430
2780
430 .
' 460
43"
94"
46"
tn
Co
* 1 inch - 69 gallons
** 32% Ca(OH)2 by wt., 1 inch - 110 gallons
-------
TABLE G-7
RUN R4 - NCB - DIESEL FUEL TO LODDBY
(continued)
Date 12/18/76 Time
Hot Duct Temp (°F)
Kiln Temp (°F)
Afterburner Temp (°F)
Optical
Loddby Flame Temp (°F)
Optical
Kiln Flame Temp (°F)
Optical
Forced Draft Fan
(Amps)
Induced Draft Fan
(Amps)
Delta-P Venturi
(In. H20)
Loddby Feed Tank Blend
Level (In.) 80", #2
2% NCB *
Lime Feed Tank Level
(Inches)**
Gas Meter Reading
1230
1800
800
2500
2800
430 ,
' 460
43"
95"
46"
1245
1790
790
2500
2800
430 .
' 460
43"
97"
47"
1300
1890
700
2450
2800
430 .
' 460
43"
99"
47"
1315
1850
700
2470
2860
430 .
' 460
43"
101"
48"
1330
1750
700
2470
2800
410 .
' 450
42"
102"
48"
STOP
194976
Ul
VO
TEST FINISHED 0 1325 Hours
* 1 inch - 69 gallons
** 32* Ca(OH)2 by wt.. 1 inch - 110 gallons
-------
APPENDIX H
CALCULATION OF WASTE DESTRUCTION PERFORMANCE
Waste destruction performance data were calculated for all waste
burns. Input into these calculations was taken from several sections of
this report indicated in the following examples.
The waste destruction efficiency (DEwaste) calculation is based upon
comparing a waste input rate to a waste emitted rate.
OF
= waste input - waste emitted
waste waste input
Y 1nn
A IUU
Equation (H-l), restated in another form, is
waste
where:
I
waste
VFR
y '
E waste
input rate of organic portion of aqueous waste feed,
grams per minute.
volumetric flow rate of combustion gases from the
afterburner, dry standard cubic meters per minute.
concentration °f organic waste constituents in
combustion gas as determined by GC or GC/ms, grams
per dry standard cubic meters.
Similarly the destruction efficiency for total organics
(DEtotal organics) compares the input rate of combined waste and auxiliary
fuel to the emitted rate of all organic material found in the combustion
zone samples.
* «. , -
total organic
" [VFR
gas total organics
]
I
fuel
„ lnn
x IUO
160
-------
where:
Lfuel
"total organics
input rate of organic portion of waste plus
auxiliary fuel oil, grams per minute.
sum of the concentrations of all organics found
in the combustion zone samples, grams per dry
standard cubic meter.
The calculations for the hammermi11ed PCB capacitors test are
presented below as examples. Initially Iwaste 1S calculated, then VFRgas,
and finally Ew£,ste (which in this case is the analytical detection
limit) is added to Equation H-2 to calculate DEwaste-
T = 210 kg fluff 0.29 kg PCB 1000 g hr
waste hr kg fluff kg 60 min
= 1,015
PCB
mm
VFR
gas
9.19 actual meters
sec
x (1 - 0.306 mole fraction water) x
3.083m cross-sectional stack area x
r\
293°K at standard conditions 1.824m hot duct area
A A
327°K at actual conditions 3.083m2 stack-area
767mm Hg at actual conditions 60 sec
760mm Hg at standard conditions min
630 dry standard cubic meters per minute
DE
waste
i1-015 ^ar) - <630 dsc""m1n> (°-°05jT* iwhs)
1,015
3 PCB
min
x 100
The destruction efficiency for total organics, DEtotal organic> for
the PCB fluff test is calculated in a similar manner using Equation H-3
and inputing the combined waste and auxiliary fuel feed and the concen-
tration of total organics found. The one assumption used in this
particular calculation was that the fiber drums were -50 percent com-
bustible material by weight.
161
-------
I
_ 40.18 liters No. 2 oil 840 g\
total organics \ min 1iter /
/210 kg fluff 0.29 kg PCB v 1000 g hr \
\ Rr kg fluff x kg x 60 minj
+ /120 kg drums 0.5 kg combustibles 1000 g hr
I hr :kg drumskg x 60 min
= 35.800 9 total organics
(35,800 ^ta^anics) . f(
630
total organics ~ 35,SOO g total_organ1cs
m- — m9
mm
x 100
= 99.97%
yo!467e
SW-122c.5
162
• U.S. GOVERWIENT PRINTING OFFICE : 1977 0-720-117/2012
------- |