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 1300C.

      The waste liquid burner (Loddby) measures  4.9 meters in  length  by 1.6
 meters diameter.  Loddby  flame  temperature  is  approximately 1500C.  Re-
 sulting  afterburner gas temperature  is  typically 1300C.  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  15.6C  (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 50C to 250C
 at 10C 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 800C 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  1320C
                 Hot duct                   1090C
                 Loddby flame               1480C
                 Kiln flame                 1320C
                                    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  1306C kiln and 1308C  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 1252C 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 1339C in the
 kiln  and  1332C 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/WATR 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 Tedlarfilm 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 15.6C, 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
 40C  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 21C

(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 (>300C)
          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 1500F.  Combustion gases are passed
through an afterburner, fired by a liquid injection burner, to attain
temperatures of 2300 to 2400F.  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 264F.   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.
yF>

-
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  250C at 8C/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 ,  Nanogradeor  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 100C to 275C at 12C/min.
             Injector - 200 to 220C
             Transfer lines - 150 to 175C
             Separator - 200 to 220C
          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 - -40C
             Manifold - ~40C
             Electron multiplier - ~40C
          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 295C
          injector = glass lined, 275C
          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 50C for 3 mln.  Linear program
                    at 20C/min to 275C.   Hold for 5-10 min at
                    275C.

     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  350C 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 850C.

 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  2C) 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
(^300C)   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 2000F 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                    1996F
               Kiln duct                    703F
               Afterburner                 2387F
               Loddby flame                2705F
               Kiln flame                  2382F
               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  703F
                                     A^te.ibaine.1 2387 F
Feed Wa4.e Liquid
    Bu.ine.1*
                                                                 1996F
                                                                    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
1963F
 688F
2351F
2723F
2354F
  36 inches of water
 101 GPH
4923 CFH
 2.6 seconds
                                 130

-------
                                     ROLLINS  ENVIRONMENTAL SERVICES
                                    INCINERATOR SVSTEM CONFRGURATIOW
                                              HOUSTON  PLANT
                                                                   1963F
                                                                   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
1993F
 910F
2428F
2731F
2'286F
  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 910F
CO
en
                                                    2428F
                                                                     1993F
                                                                       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
2005F
 919F
2429F
2748F
2442F
  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                    1966F
               Kiln duct                    790F
               Afterburner                 2385F
               Loddby flame                2818F
               Kiln flame                  1714F
               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
                                                                1966F
o.
-------
                                      ROLLINS ENVIRONMENTAL SERVICES
                                     INCINERATOR 5V5TEH CONFRGURATIOM
                                               HOUSTON PLANT
                                                                  1823F
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                    1823F
               Kiln duct                    747F
               Afterburner                 2430 F
               Loddby flame                2750F
               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\
             293K at standard  conditions    1.824m  hot duct area
                                         A                       A
              327K 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

-------