United States                 EPA-600/2-8l-055a
               Environmental Protection
               Agency                    April 1981
&EPA        Research  and
               Development
               EVALUATION OF PCB DESTRUCTION

               EFFICIENCY IN AN INDUSTRIAL BOILER
               Prepared for
               Office of Pesticides and Toxic Substances
               Prepared  by

               Industrial Environmental Research
               Laboratory
               Research Triangle Park NC 27711

-------
                 RESEARCH REPORTING SERIES


Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology.  Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

    1. Environmental Health Effects Research

    2. Environmental Protection Technology

    3. Ecological Research

    4. Environmental Monitoring

    5. Socioeconomic Environmental Studies

    6. Scientific and Technical Assessment Reports (STAR)

    7. Interagency Energy-Environment Research and Development

    8. "Special" Reports

    9. Miscellaneous Reports

This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
                       EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.

This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                EPA-600/2-81-055a
                                                April 1981
                 EVALUATION OF PCB
             DESTRUCTION EFFICIENCY IN
               AN INDUSTRIAL BOILER
                         by
       Joanna Hall, Frank Record, Paul Wolf,
            Gary Hunt, Steven Zelenski
                  GCA CORPORATION
              GCA/TECHNOLOGY DIVISION
              Bedford,  Massachusetts
              Contract No. 68-02-3168
               Work Assignment No.  9
          EPA Program Element No. ClYLlB
                EPA Project Officer
                 David C. Sanchez

   Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
        Research Triangle Park, N.C. 27711
        OFFICE OF RESEARCH AND DEVELOPMENT
       U.S.  ENVIRONMENTAL PROTECTION AGENCY
              Washington, D.C.  20460

-------
                                DISCLAIMER
     This Final Report has been reviewed by the Industrial Environmental
Research Laboratory, U.S.  Environmental Protection Agency, and approved for
publication.  Approval does not signify that the contents necessarily re-
flect the views and policies of the U.S. Environmental Protection Agency
nor does mention of trade  names or commercial products constitute endorse-
ments or recommendation for use.

-------
                                 ABSTRACT
     According to EPA's final ruling on the disposal of polychlorinated bi-
phenyls  (PCBs) , waste oils which contain PCBs in the range 50-500 parts per
million  (ppm) can be fired with fuel oil and burned in a high efficiency
industrial boiler.  In May 1980, waste oil containing approximately 500 ppm
of PCBs was co-fired in accordance with applicable state and federal regula-
tions.  This combustion took place in a high efficiency industrial boiler
owned and operated by General Motors Corporation at Bay City, Michigan.  This
report describes the evaluation program which was undertaken to document the
PCB destruction efficiency which occurred during the verification burn.  Also
investigated was the environmental and the workplace impact which occurs
during the handling and combustion of PCB-contaminated waste oils.

     No PCBs were detected in the stack gas within the detection limits of the
sampling and analytical techniques used.  The data collected during this veri-
fication burn indicate that, by following the equipment and performance re-
quirements stated in EPA's PCB regulations (40 CFR 761),  the 99.9 percent
destruction efficiency requirement implied by the regulation can be achieved
by high efficiency (industrial power) boilers.  Furthermore, monitoring of the
downwind ambient air,  the workplace environment, and employee blood levels
has indicated that PCB destruction can be conducted with no measurable effect
on either the workplace environment or the external environment.

     This report was submitted in fulfillment of Contract 68-02-3168,  Work
Assignment No.  9 by CCA/Technology Division under the  sponsorship of the
U.S.  Environmental Protection Agency.  This report covers the period of
January 29, 1980 to October 31,  1980, and work was completed as of October 31
1980.
                                     ill

-------
                                CONTENTS
Abstract	
Figures	vl
Tables	viii
List of Abbreviations	     x

   1.  Background	     1
            GCA Test Program	     2
   2.  Results	     4
            PCB Destruction Efficiency   	     4
            Dibenzof uran/Dioxin Emissions	     4
            In-plant PCB Concentrations  	     4
            PCB Blood Chemistry Levels   	     7
            Boiler PCB Residual Levels   	     7
            PCB Fuel Concentrations	     8
            Ambient PCB Concentrations   	     8
            Data Interpretation	10
   3.  Boiler Operating Conditions   	    11
   4.  Sampling Procedures   	    19
            Stack Emissions - PCB Train	19
            Stack Emissions - Method 5 Train	21
            In-plant PCB Monitoring  	    23
            PCB Blood Chemistry Sampling 	    23
            Boiler PCB Residual Sampling 	    25
            Fuel Sampling for Waste, Spiked and Mixed Oils   ...    25
            Ambient PCB Monitoring   	    25
   5.  Analytical Procedures and Results 	    35
            PCB Analysis	35
            Method 5 Train Analysis  	    36
            In-plant PCB Analysis	44
            PCB Blood Chemistry Analysis .........    44
            Boiler PCB Residual Analysis 	    44
            Fuel Analysis	44
            Ambient PCB Analysis ..........    45
   6.  Air Quality Model Calculations	49
   7.  Quality Assurance/Quality Control 	    62
            Quality Assurance Protocol   	    62
            Investigative Actions	62
            Quality Control Procedures   ........    62

References	69
                                   iv

-------
                           CONTENTS (continued)
Appendices
   A.  Analytical Equipment Operating Parameters	      71
   B.  Work Plan for Evaluation of PCB Destruction Efficiency of
       No. 3 Industrial Boiler at Chevrolet-Bay City	      80
   C.  Procedure for PCB Air Sample Analysis	     114
   D.  Method of Testing for Polychlorinated Biphenyls (PCBs) in Bulk
       Material (Solid) 	     115
   E.  Procedures Used by Chevrolet Central Office and Chevrolet Bay
       City Laboratories to Analyze the PCB Waste Oil	     116
   F.  GM Boiler Operation Measurements 	     117
   G.  Summary of Waste Oil Analysis	     129
   H.  Trace Elements Inc. Waste Oil Analysis Methodology 	     131
   I.  Phoenix Chemical Laboratory, Inc. Reclaim Oil Analysis Results    133
   J.  GM Breathing Zone and Powerhouse Monitoring Results	     136
   K.  GM Boiler Residue Analysis 	     140
   L.  GM Blood Chemistry Results 	     141
   M.  GCA Field Calibrations	     142
   N.  Discussion of Reclaim Oil Data	     144
   0.  Log of Significant Actions	     152
   P.  PCB Waste Oil Supply System Operational Problem	     153
   Q.  Premature Termination of Third PCB Run	     154
   R.  Data Interpretation	     156

-------
                                  FIGURES
Number
                                                                        Page
   1   Flue gas continuous monitoring schematic	   17

   2   PCB sampling train	   20

   3   Organic (Modified Method 5)  sampling train	   22

   4   Plot plan of plant facilities	   26

   5   Map of Bay City near the GM plant	   27

   6   Normalized concentration as  a function of crosswind distance
         and angular bearing for C  stability.  Distance from stack is
         480 meters	   29

   7   Normalized concentration as  a function of distance from stack
         for three stability conditions	   29

   8   Ambient monitor location - day 1	   30

   9   Ambient monitor location - days 2 and 3	   31

  10   Ambient monitoring locations - day 4	   32

  11   Hi-Vol sampling apparatus 	   34

  12   Modified Method 5 train organic analysis flow scheme:   particu-
         late filters	   38

  13   Modified Method 5 train organic analysis flow scheme:   resin.  .   39

  14   GC/ECD chromatogram of Aroclor 1242 standard	   47

  15   GC/ECD chromatogram of day 1—downwind site  1—preburn	   48

  16   Locations of monitors and flow direction vectors during test
         day 1 (no burn)	   52

  17   Location of monitoring and wind direction rose during burn
         No.  1 (test day 2)	  .  .   53
                                     vi

-------
                            FIGURES (continued)

Number                                                                  page

  18   Location of monitors and wind direction rose during burn No. 2
         (test day 3)	.   54

  19   Location of monitors and wind direction rose during burn No. 3
         (test day 4)	55

  20   Predicted facility impact on PCB concentration field during burn
         No.  1 (test day 2)	56

  21   Predicted facility impact on PCB concentration field during burn
         No.  2 (test day 3)	57

  22   Predicted facility impact on PCB concentration field during burn
         No.  3 (test day 4)	58
                                    vn

-------
TABLES
Number
1
2
3
4
5
6
7
8
9
10
11
12

13
14

15

16
17
18
19

Destruction Efficiency of PCB 	
Stack Emission Rate of PCB 	
Stack Emissions Data for Dibenzofurans and Dioxins 	
PCB Concentrations in Fuel 	
Results of Ambient PCB Analysis (yg/m^) 	
Combustion Parameters Monitored Routinely by GM 	
Flue Gas Analyses for 02, CO, C02» THC 	
Summary of Boiler Operating Parameters 	 	 	
Boiler Combustion Efficiency Based on C02 and CO 	
Samples Collected at General Motors Boiler House 	
Key to Stability Categories 	
Factors to Mathematically Convert Decachlorobiphenyl to an
Equivalent Amount of Aroclor 	
Results of Stack Sample Analyses for PCB 	
GC/MS Characteristics of Standards for Dioxin/Dibenzofuran
Analysis 	
Organic Emissions Using Modified EPA Method 5 Sampling
Train (Impinger Extract/Rinse Extract/Train Rinses) . . .
Method 5 Train Chloride Analysis 	
Particulate Weight 	
Model Input Parameters for PCB Burns 	
Location of Monitors During Test Periods 	
Page
5
5
6
8
9
12
14
15
18
24
28

37
37

40

41
42
44
50
51
  viii

-------
                              TABLES (continued)

Number                                                                Page
  20    Monitor Heights During Test Periods 	   51

  21    Comparison of Maximum Expected Facility Impact  and Observed
          Concentration of PCB	60

  22    Distribution of Wind Direction During Various Sampling
          Periods	61

  23    Quality Control Criteria/Blanks 	   63

  24    Quality Control Criteria/Field-Biased Blanks	64

  25    PCB Quality Control - Florisil Spikes 	   64

  26    PCB/Aqueous Media 	   65

  27    PCB Quality Control - Train Recovery	66

  28    PCB Quality Control - Foam Plug Spikes	66

  29    PCB Quality Control - Foam Plug Retention Efficiency	67

  30    Quality Control - PCBs/Fuel Oil	67
                                     ix

-------
                           LIST OF ABBREVIATIONS
aq      — aqueous
Btu     — British thermal unit
CO      — carbon monoxide
C0£     — carbon dioxide
DIG     — distilled in glass
dscf    — dry standard cubic foot
dscm    — dry standard cubic meters
g       — gram(s)
gr      — grain(s)
GC/ECD  — gas chromatograph/electron capture detector
GC/MS   — gas chromatograph/mass spectrometer
gpm     — gallons per minute
HC1     — hydrogen chloride
kg      — kilograms, 1C)3 gram(s)
Ib/hr   — pounds per hour
m/e     — mass to charge ratio
nH      — cubic meter
mg      — milligram(s), 10~3 gram
Mlb/hr  — thousand pounds per hour
mph     — miles per hour
ng      — nanogram(s) , 10~~9 gram
02      — oxygen
PCB     — polychlorinated biphenyl
ppb     — parts per billion
ppm     — parts per million
psig    — pounds per square inch gauge pressure
QA/QC   — quality assurance/quality control
TCO     — total chromatographable organic
THC     — total hydrocarbons
yg      — microgram(s), 10~6 gram
VMSTD   — flue gas volume, corrected to standard conditions,
           through dry gas meter
w.c.    — water column
                                     x

-------
                                 SECTION 1

                               INTRODUCTION
BACKGROUND

     In May 1980, waste oil contaminated with polychlorinated biphenyl (PCS)
was combusted by Chevrolet Motor Division of General Motors at Bay City»
Michigan in a high efficiency industrial boiler.  The "burn" was conducted
in conformance with the conditions established by the U.S. Environmental
Protection Agency for PCB destruction in high efficiency boilers (40 CFR  761)
and applicable state and local regulations.  Federal regulations state that
waste oil containing PCBs in the range of 50 to 500 parts per million (ppm)
can be disposed of in a high efficiency boiler provided that:

     •    The boiler is rated at a minimum of 50 million Btu/hr.

     •    If the boiler uses natural gas or oil as the primary fuel,
          the carbon monoxide concentration in the stack is 50 ppm
          or less, and the excess oxygen is at least three  (3) percent
          when PCBs are being burned;

     »    The waste does not comprise more than 10 percent  (on a
          volume basis) of the total fuel feed rate;

     •    The waste fluid is not fed into the boiler unless the
          boiler is operating at its normal operating temperature
          (this prohibits feeding these fluids during either start
          up or shut down operations);

     •    The owner or operator of the boiler continuously monitors
          and records the carbon monoxide concentration and excess
          oxygen percentage in the stack gas while burning waste
          fluid.

     •    The primary fuel feed rate, waste fluid feed rate, and
          total quantities of both primary fuel and waste fluid fed
          to the boiler are measured and recorded at regular intervals
          of no longer than 15 minutes while burning waste fluid.

     •    The carbon monoxide concentration and the excess oxygen per-
          centage are checked at least once every hour that waste fluid
          is burned.  If either measurement falls outside the criteria
          specified in this rule, the flow of waste fluid to the boiler
          shall be stopped immediately.

-------
     The PCB burn was monitored by GCA Corporation under contract to the U.S.
Environmental Protection Agency, by the Michigan Department of Natural Re-
sources , and by General Motors in order to ensure conduct of the burn in
accordance with governmental guidelines and in order to document the PCB
destruction efficiency of the burn.

     Operating plans (test plans) were developed by GCA Corporation and General
Motors Corporation to ensure definition and coordination of responsibilities
during the burn and to ensure the use of consistent sampling and analysis pro-
cedures during and subsequent to the burn.

     Measurements of PCB concentrations in the waste oil feed and in the com-
bustion exhaust gases were made to establish the PCB destruction efficiency.
Combustion parameters (CO, 02) were monitored to evaluate the boiler perfor-
mance and to establish compliance with federal regulations governing the burn.
Measurements were also made for PCB concentrations in the ambient and in-plant
air and in workers' bloodstreams before, during, and after the burn in order
to evaluate the environmental effect of the burn.

     This report documents the results of the PCB destruction test burn at
Bay City, Michigan, and describes the procedures used to acquire and assess
the supporting data.

GCA TEST PROGRAM

     The GCA test program provided protocol to conduct sampling and analysis
of stack emissions and ambient air quality data.  GCA conducted stack tests
using an EPA Method 5 sampling train (40 CFR 60) modified specifically for
capturing PCBs, as well as a second train modified to capture organic com-
pounds and chlorinated hydrocarbons which could be combustion by-products.
Simultaneously with the stack testing, GCA operated four ambient air monitors
—one upwind and three downwind—to assess the impact of the verification
burn on ambient PCB concentrations.

     Three separate PCB tests were conducted with the boiler operating under
normal load conditions.  The usual boiler fuel, No. 6 fuel oil, was co-fired
with 10 percent waste oil prepared to contain 500 ppm PCB; hence, the PCB
concentration in the fuel being burned was approximately 50 ppm.  In order
to provide background data on emissions from the boiler, 1 day of testing
was also conducted with No. 6 fuel oil only.

     Extensive background information for the test burn is provided in the
appendices of this report.  These appendices include:  operating parameters
on all analytical equipment employed (Appendix A); work plan for the overall
test effort (Appendix B), the air sampling analysis (Appendix C), the evalua-
tion of PCBs in solid materials  (Appendix D), and the analysis of PCB waste
oil (Appendix E); measurement data on the boiler (Appendix F), the PCB waste
oil (Appendices G, H, I), in~plant monitoring (Appendix J), boiler residue
(Appendix K), blood chemistry (Appendix L), and stack test measurements
(Appendix M), discussions of the fuel oil data  (Appendix N), a log of sig-
nificant actions  (Appendix 0), operation problems encountered with the waste
oil supply  (Appendix P), the premature termination of the third PCB run
(Appendix Q), and data interpretation using standard statistical techniques
(Appendix R).
                                     2

-------
     Collectively,  these appendices serve to augment and clarify the various
aspects of the program described in the body of the report.

-------
                                  SECTION 2

                                  RESULTS
PCB DESTRUCTION EFFICIENCY

     Destruction efficiency test results presented in Table 1 indicate that at
least a 99.99 percent PCB destruction efficiency was achieved during the test
burn.  This is significantly greater than the 99.9 percent destruction effi-
ciency expected by EPA's PCB regulation.  The actual destruction may be greater
than that stated because the calculated destruction was limited by the detec-
tion sensitivity of the sampling and analytical techniques used in this project.
The range in PCB inlet concentrations shown in Table 1 results from the uncer-
tainties inherent in the analytical techniques used to measure PCBs.  This un-
certainty is discussed in more detail in the subsection entitled "PCB Fuel
Concentrations."

     The destruction efficiency determination is based on a mass balance of
PCBs into and out of the boiler.  Calculated PCB emission rates are presented
in Table 2.  Because quantities of PCB in the stack gases were below detectable
levels, PCB out in milligrams per minute (mg/min)  had to be expressed as quan-
tities less than (<) an emission rate calculated at the detectable level.

DIBENZOFURAN/DIOXIN EMISSIONS

     Concurrent with sampling for PCBs, a second Modified EPA Method 5 sampling
train was used to collect organic combustion products emitted during the PCB
burn.  Analyses for chlorinated dibenzofurans and dioxins were undertaken.
This train was operated simultaneoulsy with the PCB sampling train during two
of three PCB burns and during the burning of uncontaminated No. 6 fuel oil.
These sampling trains are discussed in detail in Section 4, "Sampling
Procedures."  As with PCB stack samples, all analyses were performed at the
CCA laboratory.

     The results of the analysis of stack samples for dibenzofurans and
dioxins are shown in Table 3.  The stack gas particulate and the XAD-2 resin
(used to trap organic vapor) were independently analyzed for organic combus-
tion products.  No dioxin or dibenzofuran was detected in these samples at
the limits of detectability.

IN-PLANT PCB CONCENTRATIONS

     During the verification burn and the 1-day background burn, 42 tests were
conducted by GM to measure employee workplace exposure to PCB.  The tests were
conducted by the GM Industrial Hygiene Department using modified sampling and

-------
                   TABLE 1.  DESTRUCTION EFFICIENCY OF PCB

Run
number
PCB-2
PCB-3
PCB-4

Test
day
2
3
3
PCB in fuela
concentration
range
(mg/kg)
34-72
34-76
34-76
PCBb
in
(mg/min)
480-1000
480-1100
480-1100
PCB
out
(mg/min)
<5.8 x 10~2
<5.6 x 10~2
<5.6 x ID'2
PCB°
destruction
efficiency
(%)
>99.99d
>99.99d
>99.99d
     Note:  Density of 1:10 Dilution of Waste Oil:  No. 6 Fuel Oil is
            3.43 kg/gal.
     o
      Based on GM and GCA interlaboratory results.
      Fuel combustion rate:  3.43 kg/gal x 4 gal/min fuel
      flow = 14 kg/min.
     c_     _ ,  ^    .   inn [PCB in - PCB out]
      Percent destruction 100  	=——:	
                              L    PCB in      J
      This assumes 100 percent sample collection efficiency.
                    TABLE 2.  STACK EMISSION RATE OF PCB


Run
number
PCB-ld
PCB-2
PCB-3
PCB-4
PCB-5


Test
day
1
2
3
3
4
Gas
volume
sampled
(DSCM)a
4.898
5.380
5.785
5.964
3.207

Aroclor 1242
detected
(mg)b
<0.001
<0.001
<0.001
<0.001
NAe
Stack gas
concentration of
Aroclor 1242
(mg/m3)
<2.0 x 10~4
<1.9 x 10~4
<1.7 x 10~4
<1.7 x 10~4
NAe

Stack
flow
(DSCMM)C
274
307
328
346
264

PCB
out
(mg/min)
<5.6 x 10"2
<5.8 x 10~2
<5.6 x 10~2
<5.6 x 10~2
-
 Dry standard cubic meters.
 Based on GC/ECD results (<1.0 yg Aroclor 1242/extract).
f+
 Dry standard cubic meters/min.

 This was a run to measure background levels of PCBs in the stack gas.   No
 PCB contaminated waste oil was  burned during this run.
Q
 NA - Sample extract was inadvertently spilled during analysis sequence.
      This loss was not integral to calculating destruction efficiency
      with confidence because of redundant sampling.

-------
                TABLE  3.  STACK EMISSIONS DATA FOR DIBENZOFURANS AND DIOXINS


Run
number
M5-1C
M5-2
M5-3


Test
day
I
2
4d
Gas
volume
sampled
(DSCM)a
4.808
5.446
3.290

Weight in
particulate
extract (rag)
<1.0 x 10~3
<5.0 x 10~4
<5.0 x 10~4

Concentration
in particulates
(mg/m3)b
<2.1 x 10~4
<9.2 x 10~5
<1.5 x 10"4

Weight in
resin extract
(rag)
<8.5 x 10~4
<8.5 x 10~4
<8.5 x 10~4
Concentration
in resin
extract
(mg/m3)
<1.8 x 10~4
<1.6 x 10~4
<2.6 x 10~4
Detection
limit
(yg/ml
extract)
0.5
0.5
0.5
ury standard cubic meters.

M5-1 Particulate is 50 percent aliquot, M5-2 and M5-3 are 100 percent aliquot.

This was a run to measure background levels of dibenzofuran and dioxins in the
stack gas.  No PCB-contaminated waste oil was burned during this run.

No dibenzofuran or dioxin data were taken on day 3.   On day 3 two trains were
run simultaneously for PCB sampling to test reproducibility.   Dioxin and
dibenzofuran sampling was resumed on day 4.

-------
analysis techniques recommended for the isomers of PCB as specified in Physical
and Chemical Analytical Method—Number 253 of the National Institute of Occu-
pational Safety and Health (NIOSH) Manual.!  Sixteen of the tests were made in
the breathing zone of the boiler operators.  The PCB detection limit for these
samples was on the order of 50 yg/m^.  The remaining 26 were general conditions
tests made at two selected sites in the powerhouse.  The PCB detection limit
for these samples was approximately 10 yg/nr*.  Samples were analyzed for PCB by
GM's Industrial Hygiene Department.  The sampling procedures followed and
equipment used are more fully described in Appendix B.  Breathing zone and
powerhouse monitoring results are presented in Appendix J.  All tests showed
nondetectable levels of PCB.  This indicates that concentrations were less
than the recommended Occupational Safety and Health Administration (OSHA)
8-hour levels of 1 mg/m^ and 0.5 mg/m^ for Aroclor 1242 and Aroclor 1254,
respectively (29 CFR 1910.1).

PCB BLOOD CHEMISTRY LEVELS

     Testing was performed on GM powerhouse operators to measure blood PCB
levels before and after the verification burn.  Preburn levels ranged from
7 to 36 parts per billion (ppb) with an average of 18.0 parts per billion.
Post-burn levels were 5.4 to 40 ppb with an average of 18.6 ppb.  Within the
limits of accuracy of this testing, the pre- and post-burn samples showed no
significant changes in blood levels of PCB.  These blood chemistry results
are reported in Appendix L.

     A comparison study was performed by GM on 39 office workers who presumably
would have no work association with PCB.  The study is cited in Appendix L.
It showed that:  those who eat in excess of 8 oz of Great Lakes fish per week
have in their blood a PCB range of 4 to 71 parts per billion and an average
of 23.5 parts per billion.  Those who average less than 8 oz of Great Lakes
fish per week have a range of 3.4 to 50.3 parts per billion and an average of
18.03 parts per billion.

     The powerhouse employees therefore showed blood,levels comparable to
those office employees with the least exposure to environmental PCB.

BOILER PCB RESIDUAL LEVELS

     Before and after the PCB verification burn, scrapings from various points
inside the boiler were taken to determine PCB residue.  These residual scrap-
ings were analyzed by the GM Industrial Hygiene Department to determine the
presence of PCBs.  The test  method employed for this analysis is discussed in
Appendix D and the results are reported in Appendix K.

     The analytical technique used for residue analysis was sensitive to PCB
concentrations in the parts  per billion (ppb) range.   Yet, PCB concentrations
could be detected in only one of the  24 preburn residue samples and  only 8
of the 20 post-burn samples.   The residual level of PCBs for all but  one of
the samples was substantially less than 0.1 parts per million (ppm).   The
sole exception was the oil burner tip  residue, where the PCB concentration

-------
was in the 0.8 ppm range.  These results indicate that there was no signifi
cant level of PCBs within the boiler, either before or after the PCB test
burn.

PCB FUEL CONCENTRATIONS

     Fuel samples were analyzed for PCB by GM and GCA laboratories by using
a gas chromatograph with an electron capture detector (GC/ECD) .  Sampling
and analysis procedures followed by GCA are detailed in Appendix N; those
followed by GM are listed in Appendix E.  The PCB concentrations reported
by each laboratory are listed in Table 4.  The variation demonstrated in
Table 4 is not unusual with interlaboratory analysis of a complex mixture
such as PCBs.  An audit was performed at the GM and GCA laboratories to
verify that procedures used on fuel samples were analytically sound.  The
results of the audit verified the comparability of the GCA and GM analyses.

                    TABLE 4.   PCB  CONCENTRATIONS  IN FUEL
                                    GM results   GCA results
                     Sample  ID         mg/kg)      mg/kg)

                  Spiked Waste  Oil      496          750

                  Fuel  Feed
Run PCB- 2
Fuel Feed
Runs PCB- 3 & 4
Fuel Feed
Run PCB-5
34
34
34
72
76
88
      The range  in  concentration,  as  reported  above, was  further  investigated
 by an independent  EPA laboratory  which  analyzed  split  samples  from GCA's  spiked
 waste oil and GM's spiked waste oil.  The  EPA laboratory reported  values  of
 614 and 640  mg/kg  for aliquots of the GM and  GCA samples,  respectively.   These
 values are intermediate  to  the GCA and  GM  values and within  the  range  of  ex-
 pected agreement for  analysis of  a complex mixture containing  PCBs.

 AMBIENT PCB  CONCENTRATIONS

      The ambient monitoring network  was designed to measure  PCB  concentrations
 in the ambient  air on plant property and in the  immediate  populated area  down-
 wind of the  stack  during the test burns.   Meteorological conditions on the
 date of each burn  were used to position the monitors in  areas  on the plant
 property where  maximum downwash from the stack effluent  was  predicted.  In
 this way, PCB background levels,  increases in concentrations of  PCBs,  and
 PCB concentration  in  the nearest  populated area  which  could  have been  exposed
 to PCBs as a result of the  test burn were  measured.

-------
     Ambient sampling occurred on the same days as the stack sampling; i.e.,
1 day was specified as a background burn  (No. 6 fuel oil only) followed by 3
days of burning a combination of No. 6 fuel oil and PCB-contaminated waste
oil.  The monitors were operated in a synchronized manner with the collection
of stack gas samples.

     Four runs, each 3 hours in duration, comprised the 12-hour ambient moni-
toring sample period.  The four runs were:  a preburn, a post-burn, and two
runs during the burning of waste oil.  Testing coordination was designed so
that no waste oil would enter the boiler until the 3-hour preburn run was
complete.  The post-burn run was not started until a time lag, derived from
prevailing wind velocity, sufficient to ensure that effluent from the stack
would have passed by the monitors, had occurred.

     Analysis of ambient samples was performed at the GCA laboratory.  Results
showed that ambient concentrations of PCBs at the downwind test sites were less
than the upwind concentrations by factors which ranged between 200 to 400 per-
cent.  This indicates emissions from the test burn did not increase the ambient
air loading of PCBs in any significant manner.  The ambient tests results are
shown in Table 5.
            TABLE 5.  RESULTS OF AMBIENT PCB ANALYSIS

                                                 Test day
            Site identification        1       2      3
       UPWIND-:;itc lb                0.19   0. ib   0.39      O.i2

       DOWNWIND- (preburn) Sue /    1,4    O.'b   1,041     ''.^i

       DOWNWIND-(burn)-Si lc 2d       0.43   O.CH^  0.019     (' '3

       DOl/NWIND-(post-burn)-Site .'L  O.OVJ  0.24   0.019  No!  s.'.r,.p Led

       DOWNWIND- (preburn)-Si to 3°    0.50   0.09_:  0.0!,      0.024

       DOWNWirO"(bunO-S;io V       0.29   0.08?  '''.055     0.02&

       DOWNWIND-(post-burn)-Site JC  0.015  0.013  3.019  Vot sampled

       POPULATED AREA - Site 4b      0.063  0.067  0.028     0.051

        Aroclor 1242.

        Twelve-hour sampling period.
       c
        Three-hour sampling period.

        Six-hour sampling period.

-------
     Section 4, Sampling Procedures, provides a further description of the
protocols and equipment used in conducting the ambient air sampling for PCBs.
Section 6, Air Quality Model Calculations, describes and presents the modeling
results used to locate the ambient air quality samplers.

DATA INTERPRETATION


     A statistical analysis of the measured ambient PCB data was undertaken
to determine the significance, if any, of plant contributions to background
PCB levels.  A paired t-test, using ambient PCB monitoring data collected
at each of the four monitoring stations before, during and after each PCB
burn, was employed for this analysis.  The t-test results concluded that
there had been no apparent increase in PCB levels in downwind or nearby
residential areas due to any of the PCB burns.


     Analysis of the stack PCB emissions was also to be conducted to determine
the statistical validity of the PCB destruction efficiencies achieved.  How-
ever test results for the third burn day were lost, and there were no mea-
surable PCB concentrations from which to determine variability on each of the
remaining 2 days.  Therefore the planned statistical test could not be meaning-
fully applied to the test data.  A more thorough description of the statistical
methods employed in these analyses is presented in Appendix R.
                                     10

-------
                                 SECTION 3

                        BOILER OPERATING CONDITIONS
     The Chevrolet Plant at Bay City, Michigan, includes manufacturing and
administrative facilities as well as a powerhouse and a wastewater treatment
facility.  The wastewater treatment process was the source of the PCB-
contaminated reclaim oil that was used for the verification burn.  Reclaim
oil was separated from the manufacturing facility process water and then
filtered (20 micron) and centrifugated to ensure that water and solids con-
tent were each less than 0.1 percent.  This reclaim waste oil was diluted,
by volume, to a 1:9 ratio with No. 6 oil for firing to a high efficiency
boiler in conformance with federal regulations (40 CFR 761.10).

     The plant complex uses a powerhouse with three high efficiency boilers
to produce steam, compressed air, and surface cooling water for the manu-
facturing plant.  The three boilers in the powerhouse are Wickes Type "K"
65-4K-7 package boilers with a 60,000 Ib/hr steam generation capacity at
8 gal/min No. 6 fuel oil feed rate.  The No. 3 boiler that was used for the
PCB waste oil burn was installed in 1965.  During the verification burn,
continuous monitors were used to measure carbon monoxide (CO), and oxygen (02),
as prescribed by Federal regulations (40 CFR 761.10).  Carbon dioxide (C02)
and total hydrocarbons (THC) were also continuously monitored at the stack
above boiler No. 3.  Provision was made for shutdown of waste fuel feed unless
the following criteria were met:

     •    Boiler flame temperature - normal (^2500°F)

     •    CO concentration - less than 50 parts per million (ppm)

     •    Excess Q£ - greater than 3 percent, and

     •    Opacity - less than 5 percent

An additional permit requirement was for steady-state operation with 4 gallons
per minute fuel feed.  Steady state operation was maintained during the burn
by setting boiler No. 3 at base load and allowing boilers No. 1 and No.  2 to
compensate for fluctuating steam demand.  This is documented in both Boiler
Operator Data (Appendix F) and Flue Gas Analysis (Appendix M).

     Steady operation of the boiler was the responsibility of GM personnel.
Table 6 presents the specific boiler parameters routinely monitored by GM
during the burn.  The flue gas parameters and composition were monitored by
GCA.  Any changes in flue gas readings were noted and relayed to GM personnel
so that appropriate action could be taken.

                                     11

-------
   TABLE 6.  COMBUSTION PARAMETERS MONITORED ROUTINELY
             BY GM
        Parameter
        monitored
 Instrument
    range
    Normal
   parameter
     range
Fuel oil temperature
Fuel oil flow rate
Atomizing steam pressure
Steam flow rate
Smoke density (in-stack)
Windbox pressure
Firebox pressure
    For air preheater
Air temperature in
Air temperature out
Gas temperature in
Gas temperature out
Air pressure in
Air pressure out
50-300 F
0-8 gpm
0-160 psig
0-70 Mlb/hra
1-100%
0-10" w.c.b
Ambient-200 F
1-8 gpm
15-110 psig
10-70 Mlb/hra
5-20%
0.5-7.5T1 w.c.b
           b
0-10" w.c.    0-5.5" w.c.
0-600°F
0-600°F
0-600°F
0-600°F
0-15" w.c.
0-15" w.c.
Ambient
250-350°F
350-580°F
200-300°F
0.5-10.5" w.c.
0.5-8.0" w.c.
 nlb/hr - thousand pounds per hour.
 w.c. - water column.
                             12

-------
     Data obtained by GCA from their continuous monitoring for C>2> C02» CO,
and THC are presented in Table 7.  The remaining parameters, those measured
by GM, are summarized in Table 8.  These latter boiler parameters were re-
corded at 15-minute intervals during each PCB test run.  A full log of these
recorded values is presented in Appendix F.  Boiler operation was very steady
with the exception of Day 4.  Problems encountered on Day 4, though only
apparent in the off-scale oxygen reading in Table 7, can be seen in Appendix F
and Table 8.  While none of the acceptable boiler parameter limits were ex-
ceeded during Day 4, the range of recorded values for several parameters were
from two to five times greater than the corresponding range for the other
test days, as shown in Table 8.  Items such as air heater temperature and
pressure and windbox and furnace pressure varied significantly more on Day 4
than on Days 1, 2 or 3.  Of all the variations, the one most indicative of a
change in boiler operating conditions is that of the back-wall furnace tem-
perature.  The average temperature for this parameter on Day 4 was 10 percent
lower than the other 3 days and the range of recorded values was four times
as great.  This lower back-wall temperature was apparently related to a plugging
of the fuel feed line.  This plugging problem caused premature termination of
the 4th Day's burn and is explained in Appendices 0 and P.

     The stack gas concentrations of 02, C02, CO, and THC for each burn period
were determined by GCA using continuous emission monitors calibrated twice
daily using certified (±1 percent) calibration gases from Scott Environmental
Technology Accublend-^ cylinders.

     The aforementioned gas sample was extracted continuously from a 2-inch
port in the flue gas duct directly above the boiler and drawn through a Thermo-
Electron Model 600 Flue Gas Conditioner to remove moisture (by condensation)
and particulates (by filtration through glass fiber-filter media).  The clean,
dry, flue gas sample was then pumped to each continuous monitor for analysis
as outlined in Figure 1.

     In addition to the boiler operating measurements listed in Table 8, the
following instruments and parameters were added during the verification burn:

     •    An infrared pyrometer for flame temperature.

     9    A custom thermocouple for furnace gas temperature.

     »    An u> t, rasor.ic flowmeter £uc relaim oil flow.
     A •,-Grip; ete discussion rf all ruon.; torii^; equipment used is included in
/• > 0 •:>. n/j J :-. A ,

-------
                            TABLE 7.  FLUE GAS ANALYSES FOR 02, CO, C02, THC
                    Oxygen (%)      Carbon dioxide (%)  Carbon monoxide (ppm)  Total hydrocarbons  (ppm)
 Date sampled    Min   Max     Avg  Min    Max   Avg    Min     Max     Avg     Min      Max      Avg
Day 1 -
Day 2 -
Day 3 -
Day 4 -
5-05-80
5-08-80
5-09-80
5-10-80
5.6
5.3
3.9
5.3
6.7
7.3
• 8.0
>10.23a
6.3
6.1
5.6
6.5
11.2
12.0
8.3
6.7
13.4
21.3
12.6
13.12
12.3
14.1
11.3
11.2
12.3
6.8
8.0
7.2
16.7
9.7
11.0
12.4
14.2
8.7
9.6
8.0
1.3
0.2
0.7
0.5
3.1
0.8
2.4
2.3
2.2
0.4
1.4
0.8
Lowest minimum   3.9                 6.7                 6.8                    0.2
value

Highest maxi-        >10.23               21.3                  16.7                     3.1
mum value

Average 4-day                  6.1               12.4                   10.2                      1.2
test period
value

Standard devi-   6.1 ± 0.46         12.4 ± 1.36         10.2±2.48             1.2 ± 0.8
ation for 4-day
test period
average value


 Reading off scale

Note:  All readings are 15-minute averages unless stated otherwise.

-------
TABLE 8.  SUMMARY OF BOILER OPERATING PARAMETERS
Parameter
Monitored
Fuel Oil
Pressure
(psig)

Fuel Oil
Flow Rate
(gpm)

Steam
Flow Rate
(Mlb/hr)

Air Heater
Temperature
Air In (°F)

Air Heater
Temperature
Air Out (°F)

Air Heater
Temperature
Gas In (°F)

Air Heater
Temperature
Gas Out (°F)

Windbox
Pressure
(in. w.c.)

Furnace
Pressure
(in. w.c.)

Flue Gas
Opacity
'(%)

Statistical
Classification
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Background
May 5, 1980
15/120
39
2
0.87
3.9/4.1
4.0
0
0
10/70
39
3
1.0
ambient
78
9
3.2
250/350
355
8
2.38
350/580
478
12
3.99
200/300 -
N.O.
N.O.
N.O.
0.5/7.5
4.3
0.5
0.21
0.5/5.5
3.2
0.5
0.20
5 or less
2
1.5
0.4
First PCB
May 8, 1980
15/120
37
3
1.1
3.9/4.1
4.0
0.1
0.04
10/70
40
5.5
1.6
ambient
66
5
1.7
250/350
354
9
3.13
350/580
506
20
4.41
200/300
N.O.
N.O.
N.O.
0.5/7.5
5.7
1.0
0.28
0.5/5.5
4.3
1.0
0.28
5 or less
0
0
0
Second PCB
May 9, 1980
15/120
39
7
1.8
3'. 9/4.1
4.0
0.2
0.055
10/70
39
4.5
1.3
ambient
65
4
0.79
250/350
362
10
2.67
350/580
522
20
5.36
200/300
N.O.
N.O.
N.O.
0.5/7.5
6.1
0.5
0.20
0.5/5.5
4.7
0.5
0.24
5 or less
0.4
1.0
0.5
Third PCB
May 10, 1980
15/120
31
10
2.9
3.9/4.1
4.1
0.3
0.96
10/70
32
10
3.3
ambient
61
8
2.7
250/350
329
44
13.0
350/580
453
108
30.0
200/300
N.O.
N.O.
N.O.
0.5/7.5
4.1
3.0
0.97
0.5/5.5
3.0
3.0
0.94
5 or less
2
0
0
(continued)


15




-------
TABLE 8 (continued)
Parameter
Monitored
Fuel Oil
Temperature
(°F)

Atomizing
Steam Press.
(psig)

Air Heater
Pressure In
(in. w.c.)

Air Heater
Pressure Out
(in. w.c.)

Steam Drum
Water Level
(in.)

Flame
Temperature
(°F)
Back Wall
Temperature
(°F)
Oil Flow
Ultrasonic
(gpm)

Wind Speed
(mph)

Wind
Direction
(Comp. Pt.)
Statistical
Classification
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Average
Range
Std. Dev.
Average
Range
Std. Dev.
Acceptable
Average
Range
Std. Dev.
Average
Range
Std. Dev.
Average
Range
Std. Dev.
Background
May 5, 1980
160/220
192
6
1.62
15/110
49
5
1.2
0.5/10.5
6.2
0.5
0.18
0.5/8.0
4.8
0.5
0.18
+5" of ctr.
1.0
0.75
0.13
2630
50
15.4
1690
35
11.5
(ref.)
N.O.
N.O.
N.O.
9
10
2.8
WNW
90U
/•»
40°
First PCB
May 8, 1980
160/220
202
8
2.45
15/110
46
5
1.5
0.5/10.5
8.2
1.0
0.42
0.5/8.0
6.1
1.0
0.28
+5" of ctr.
1.0
0.25
0.051
2600
75
24.4
1660
45
11.5
(ref.)
3.90
2.30
0.852
12
7
2.0
NW
f\
70
o
20
Second PCB
May 9, 1980
160/220
194
15
3.02
15/110
50
7
1.3
0.5/10.5
8.9
0.5
0.21
0.5/8.0
6.6
1.0
0.25
t5" of ctr.
1.0
0
0
2580
50
19.4
1660
120
22.2
(ref.)
2.28
3.22
0.878
13
15
3.3
W
f>
90
o
20
Third PCB
May 10, 1980
160/220
200
12
3.92
15/110
39
14
3.3
0.5/10.5
5.9
4.5
1.3
0.5/8.0
4.3
3.5
1.0
+5" of Ctr.
1.0
0
0
2630
100
29.1
1500
420
83.8
(ref.)
2.25
5.10
1.50
19
16
3.9
SSW
O
45^
~«o
20
          16

-------
   PROBE
                     (PROBE  IS HEAT TRACED
                      FROM  STACK TO CONDITIONER)
           2  PORT
   t

t      t
FLUEGAS
PARTICIPATE
FILTER
                            CONDITIONER W/
                            PARTICIPATE AND
                            MOISTURE  REMOVAL,
                            AND SECONDARY  PUMP
                                                                      EXHAUST
                                                                                           CO
                                                                                         RECORDER
                                                                     C02
                                                                   RECORDER
                                                                                           THC
                                                                                         RECORDER
                                                                      °2
                                                                   RECORDER
                     Figure  1.  Flue gas continuous monitoring  schematic.

-------
     The boiler combustion efficiency, as presented in Table 9, is computed
based on average CC>2 and CO measurements, Table 7.  The calculation of combus-
tion efficiency is based on the following equation taken from Annex I of the
PCB regulations (40 CFR 761).   Table 9 indicates that the combustion efficien-
cy for each test run was nearly identical, and that the combustion efficiency
on Day 4, as measured by CO and C02 concentrations, was unaffected by the
variations in certain boiler parameters.

                                    Cone CO2
                      %C'E- = Cone C02 + Cone CO X 10°

               TABLE 9.  BOILER COMBUSTION EFFICIENCY BASED ON
                         C02 AND CO

                                                     % Combustion
                              C02 (ppm)a  CO (ppm)a   efficiency
Day
Day
Day
Day
1
2
3
4
(5-05-80)
(5-08-80)
(5-09-80)
(5-10-80)
12.
14.
11.
11.
3
1
3
2
X
X
X
X
10
10
10
10
4
4
4
4
14.
8.
9.
8.
2
7
6
0
99.
99.
99.
99.
990
994
992
993
             o
              Based on an average of 15-minute average values.
                                       18

-------
                                  SECTION 4

                             SAMPLING PROCEDURES
STACK EMISSIONS - PCB TRAIN

     The PCB sampling of combustion flue gas included three complete days of
boiler operations at normal load, burning No. 6 fuel oil and approximately
500 ppm PCB contaminated waste oil at a 9 to 1 ratio.

     Waste oil was fed to the boiler only after combustion of No. 6 oil was
steady and a normal burn temperature was established.  It was calculated that
a minimum sampling period of 6 hours would be required to yield approximately
10 yg of PCB for analysis.  This was based on a destruction efficiency of
99.9 percent at the sampling conditions described below.

     The sampling train used for PCBs was a modified Research Appliance
Corporation (RAC) train as described in "Measurement of PCB emissions from
Combustion Sources," (EPA-600/7-79-047).   A schematic of the train is shown
in Figure 2.

     The sampling and velocity traverse was along two diameters of the stack.
A total of 44 points were thus sampled to provide a representative sample of
the flue gas composition.  Total sampling time was 308 minutes; 7 minutes per
point on test days 1 through 3.  The expected flow rate was approximately
0.6 ft^/min.  Sampling on test day 4 was only 210 minutes because an operator
error caused shutdown of the fuel feed during the test, which resulted in a
blockage of the feed line.  This problem is more fully discussed in Appendix Q.

     Sampling was isokinetic (±10 percent) with readings of flue gas param-
eters recorded at every sampling point during the traverse.  In the event
that isokinetic sampling could not be maintained, provisions were made for
the train to be shut down and the problem remedied.  No such problem was ex-
perienced during any of the test runs.  In the event that either steady oper-
ation was not maintained, or monitored gas parameters (CO, C02, 02) were out
of the specified limits (40 CFR 761.10), the testing would have been stopped
until conditions were stabilized.  All monitored gas parameters remained
within specified limits for each test run, as demonstrated in Table 7 and
Appendix F.

     The PCB sampling train uses a series of four impingers and a solid
adsorbent tube as shown in Figure 2.  The first two impingers contain de-
ionized, distilled water.  The third impinger is dry, while the fourth
impinger contains silica gel.   The solid adsorbent tube is a glass column


                                      19

-------
                                                            THERMOMETER
                                                                   CHECK
                                                                   VALVE
*CV£RSE-TYPE
PITOT TUBE
       ORIFICE  t_
      MANOMETER
                                                            I - OEIONIZED,
                                                               DISTILLED  WATER
                                                            2 - DEIONIZED,
                                                               DISTILLED  WATER

                                                            3- DRY
                                                            4- SILICA  GEL
                      Figure 2.   PCB sampling train.
                                      20

-------
 prepacked  in  the  laboratory with  7.5  grams of pre-extracted  30/60 mesh
 Florisil activated  at  130°C for 16 hours.  The Florisil  is secured  in the
 tube with  pre-extracted  glass wool.

     The PCB  sampling  method3 requires  that a field-biased blank train be
 set up  and recovered with the sampling  trains.  One blank train was re-
 covered with  the  sampling trains  for  each day of sampling.

     In an attempt  to  provide a qualitative estimation of recovery  effi-
 ciency, sampling  trains were spiked in  the field with deuterated tetra-
 chlorobiphenyl.   This  was accomplished  by adding a known amount of  the
 deuterated isomer (dissolved in acetone) into the first impinger before
 the final  leak check.  Recovery and analyses of these spiked samples were
 inconclusive.  Problems encountered with these samples are discussed in
 Section 7,  Quality  Assurance/Quality  Control.

     The sampling equipment was prepared for PCB testing at the GCA labora-
 tory in Bedford, Mass.^  The sequence used was the following:  acid soak,
 alcoholic  KOH soak, distilled deionized water rinse, acetone rinse, hexane
 rinse (in  field).   Glass sample bottles with Teflon-lined caps were also
 prepared in a similar manner.

 STACK EMISSIONS - METHOD 5 TRAIN

     The sampling train used to collect other organic compounds from the
 stack gas was a standard EPA Method 5 particulate train, modified to include
 an adsorbent column, as shown in Figure 3.   This train was operated simul-
 taneously with the PCB sampling train to provide real-time comparison samples
 for HC1, dibenzofuran, chlorinated dibenzofurans and chlorinated dioxins
 analyses.   The sampling included a total of 44 points along the two stack
 diameters.   Sampling was a 7 min per point for a total duration of 308
minutes.   Samples of flue gas particulate were collected on a 4-inch glass
 fiber filter.   In addition to total particulate, a temperature controlled
 adsorbent column was included in the sampling train to collect organic emis-
 sions from the flue gas.   This column, shown in Figure 3, contains approxi-
mately 25  grams of XAD-2 resin.   The temperature of the condenser is main-
 tained at  70°F and the condensate  is collected in the first impinger of the
 sampling train.

     This sampling train was also  used to determine HC1 emissions from the
 flue gas.   The second impinger of  the Modified Method 5 train contained
200 ml of 8 percent  Na2C03/H20 as  a trapping solution for HC1 emissions.

     In addition to  the second impinger solution,  an aliquot  (20 percent) of
the first  impinger water (condensate for XAD-2 resin trap)  was taken and pre-
served by addition of Na2C03.   The remaining condensate (80 percent) was pre-
served in the  field  by the addition of distilled in glass (D.I.G.)  methylene
chloride.
                                     21

-------
                                             THERMOMETER
       PROBED
«EVEKSE-TYPE
PlTOT  TUBE
           J
                   THERMOSTATIC
                   WATER BATH ^.
                                                           THERMOMETER
                                                                   CHECK
                                                                   VALVE
      UNGREASEO   \\
STACK  FITTINGS \  II
WALL           /\ U-
       ORIFICE  fc_
      MANOMETER
                                        IMPINGERS
                                     I.  EMPTY
                                     2.  8%
                                     3.  EMPTY
                                     4.  SILICA  GEL
          Figure  3.   Organic  (Modified Method 5)  sampling train.
                                      22

-------
     The sampling equipment was prepared in the same manner as the PCB train
with the addition of a terminal rinse with dichloromethane (in field).  Sample
bottles used for organic samples were cleaned by the EPA Level 1 procedure,
which substitutes a dichloromethane rinse for the final hexane rinse.5
Nalgene sample bottles were used for HC1 samples and were rinsed with 1:1
nitric acid followed by deionized distilled water rinses.

     To provide background data on emissions from the boiler, a preliminary
test was conducted with No. 6 fuel oil only, one day before the initiation of
the 3-day PCB test burn.  This test was for the same duration and at the same
conditions as those to be used for the PCB sampling.  The test included both
a PCB train and a standard Method 5 particulate train.  This test provided
background information on the test conditions and also provided emissions
data for the boiler as operated at normal load and firing No. 6 fuel oil.

     Duplicate PCB trains were simultaneously operated on the second day of
waste oil burning.  The replicate measurements would serve as indication of
the precision of stack collection of PCBs.  The Modified Method 5 particulate
train from that day was therefore replaced by the second PCB train.  A recent
study determined that chlorinated dibenzofurans were not a significant PCB
combustion product of high temperature incineration."  In light of these
findings, it was believed that the elimination of the Modified Method 5 train
from one day's sampling would not eliminate any information that would be
critical to the destruction efficiency calculations.

     A complete listing of samples which were taken in the flue gas stream
is found in Table 10.  Field sample recovery adhered to those methods defined
in the GCA test plan,4 including Quality Assurance/Quality Control (QA/QC)
procedures, as described in that report.

     The PCB train measured the PCB output in the flue gas.  In order to
determine the PCB destruction efficiency during the test, it is necessary
to measure the PCB input from the fuel being combusted.

IN-PLANT PCB MONITORING

     Mine Safety Appliances Model-G personnel air sampling pumps were used
to collect 16 breathing zone and 22 general area samples during the PCB
verification burn.  The sampling flow rate was controlled at either 200 cubic
centimeters per minute (cc/minute) or at 500 cc/minute.  The sample volumes
ranged from approximately 25 to 170 liters of air, varying in proportion to
the sample time as well as the flow rate.  A complete listing of the breathing
zone and powerhouse monitoring results is provided in Appendix J.  A discussion
of these results is provided on page 44 of Section 5.

PCB BLOOD CHEMISTRY SAMPLING

     Blood samples were drawn from the 12 boiler operators and 39 other
employees who had no work experience with PCB.   The samples were split and
sent to the State of Michigan Department of Health Laboratory in Lansing,
Michigan and the Raltech Laboratory in Madison, Wisconsin for analysis.  The
results are given in Appendix L and discussed on page 44 of Section 5.

                                     23

-------
                       TABLE 10.   SAMPLES COLLECTED AT GENERAL MOTORS BOILER HOUSE
    Sample
       Source
                                                      Container
                      Species analyzed
                                Number
                                  of
                                samples
Flue gas particulate

Flue gas particulate

XAD-2 resin

Condensate, preserved

Condensate extracted

Na2C03 solution

Florisil resin

Condensed water
impinged particulates


Train wash


Florisil blank

Train wash blank

Filter blank


XAD-2 blank

CH2C12 blank

Na2C03 blank

Fuel feed
Method 5 filter

Method 5 rinse

Method 5 trap                Amber glass

Aliquot, Method 5 condenser  Nalgene

Method 5 condenser

Method 5 second impinger
Aluminum foil  Polychlorinated dibenzofurans

Amber glass    Total particulate

               Total organic residue

               HC1
PCB train


PCB train impingers

D.I.G. hexane
D.I.G. acetone

Blank PCB train

Blank PCB train

Filter lot sample


XAD-2 lot sample

CH2C12 sample

Solution blank

Trickle value (integrated)
Amber glass    Total organic

Nalgene        HC1
Amber glass


Amber glass

Amber glass

Amber glass

Amber glass

Petri dish


Amber glass

Amber glass

Nalgene

Amber glass
PCBs


PCBs

PCBs

PCB background

PCB background

Background polychlorinated
dibenzofurans

Background organics

Background organics

Background HC1

Polychlorinated dibenzofurans,
PCB
3

3

3

3

3

3

5


5

5

4

4

1


2

2

2

4

-------
BOILER PCS RESIDUAL SAMPLING

     Samples of residue were scraped from eight locations in the boiler and
exhaust system before the verification burn.   Samples from 10 locations were
scraped after the burn.  All the samples were sent to the General Motors
Industrial Hygiene Laboratory for analysis.  The results are given in
Appendix K and discussed on page 45 of Section 5.

FUEL SAMPLING FOR WASTE, SPIKED AND MIXED OILS

     A sample of the waste oil to be fired during the verification test was
taken from the waste oil storage tank after thorough mixing.  The sample was
split and sent to the four laboratories listed in Appendix G.  The waste oil
was analyzed for PCBs, dibenzofurans, dioxins, as well as nitrogen, sulfur,
chlorine, carbon, hydrogen, ash, water, sediment, calorific value, carbon
residue and flash point.  GM retained a fuel sample for their own analytical
determinations.  The spiked oil sample was prepared in the laboratory at
Chevrolet-Bay City from this retained waste oil sample and the spiked fluid.
The spiked oil sample was split and sent to the GCA and Chevrolet laboratories
for analysis.  A requisite QA/QC audit which included a separate analysis, was
performed by the Health Effects Research Laboratory of O.R.D.   The mixed oil
samples were extracted from the fuel oil line to the boiler during each test
run.  These samples were split and sent to the GCA and Chevrolet laboratories
for analysis.  The results of the various oil analysis are given in Appendix G
and discussed in Section 5.

AMBIENT PCB MONITORING

Network Design

     The ambient monitoring network was designed to measure PCB concentra-
tions in the ambient air on plant property and in the immediate populated
area expected to be downwind of stack emissions from the test burns.  Me-
teorological conditions on each test day were used to position the monitors
in areas on the plant property where maximum impact from the stack effluent
was predicted.  In this way, it was hoped to have a basis for determining
background levels, population exposure, and increases in concentrations of
PCBs due to the test burn.  Figure 4 shows the location of the powerhouse
with respect to nearby plant facilities.  Figure 5 shows the locations of
neighboring residential and commercial areas.

     Monitors were positioned each morning based on weather forecasts, onsite
wind conditions (measured at the stack by a continuously monitoring wind
system) and predicted plume dispersions.  Table 11 was used with forecasted
meteorological conditions, to predict atmospheric stability category for the
burn period.
                                     25

-------
 100
  I
  200 ft
  	I
 1
29
SCALE
50
                                            RETENTION  PONDS
                                      PUMP  HOUSE
      POWER  HOUSE
       STACK
                                                 WASTEWATER
                                                   FACILITY
                                                CHEVROLET
                                               MANUFACTURING

                                                     26'
 Figure 4.  Plot plan of plant facilities,
            are indicated in feet.
                                      Structure heights
                             26

-------
Figure 5.  Map of Bay City near the GM plant.
                                             SCALE  1cm = 100m
                      27

-------
     After establishing an expected mean downwind direction, plume width
 information, such as that provided in Figure 6, was used in conjunction
 with Figure  7 to locate the two monitors for maximum ambient plume impact.
 The curves shown in Figures 6 and 7 have been calculated for 3-hour re-
 lease  times.  The goal was to select a compromise distance suitable for all
 expected stabilities during the 12-hour monitoring time.  Consideration was
 also given to any anticipated hotspots from plume downwash, if effluent
 dispersion was predicted over a nearby building.  This selection was con-
 ducted with  the State of Michigan representative at the test burn.
TABLE 11. KEY TO STABILITY CATEGORIES'
Surface wind
speed (at 10 m)
m sec 1
2
2-3
3-5
5-6
6

Incoming
Strong
A
A-B
B
C
C
Day

solar radiation
Moderate
A-B
B
B-C
C-D
D
Slight
B
C
C
D
D
Night
Thinly overcast
or
> 4/8 low cloud

E
D
D
D
< 3/8
cloud

F
E
D
D
     The monitor locations that were used during the test program are shown
in Figures 8, 9 and 10.

Test Scheduling

     Ambient sampling occurred on the same days as the stack sampling; i.e.,
1 day specified as a background burn (No. 6 fuel oil only) followed by
3 days at burning a combination of No.  6 oil and PCB-contaminated reclaim
oil.  The monitors were operated in a synchronized manner with the stack
samples.

     Four runs, each three hours in duration, comprised the 12-hour ambient
monitoring sample period.  These four runs were one preburn, one post-burn,
and two runs during the waste oil burning which were combined for analysis.
Testing coordination was designed so that no reclaim oil would enter the
boiler until the 3-hour preburn run was complete.   The post-burn ambient sam-
pling run was not started until after a time, derived from prevailing wind
velocity, when any PCB contaminated effluent from the stack would have passed
by the monitors.

Sampling Method

     High volume sampling of ambient air was performed with samplers modified
as described in "A Method for Sampling Analysis of Polychlorinated Biphenyls
(PCBs) in Ambient Air" (EPA-600/4-78-048).8  This modification involved an
extension of the throat assembly at the sampler with a piece of cylindrical
aluminum, threaded on one end.  The other end was fitted with an aluminum
flange, which supported the motor housing.  Motor exhaust was directed away


                                     28

-------
O
2
UJ
O


o
<->
O
LJ

ISI
o
z
      5 X 10
         -5
         -5
      4X 10
3X 10
    -5
         -5
      2 XIO
      i x 10
         ~5
              -200  -160  -120 -80  -40   0    40   80   120  160   200  METERS
                           CROSSWIND  DISTANCE


            -25   -20   -15   -10   -50    5   10    15    20   25 DEGREES

                             ANGULAR  BEARING


      Figure 6.   Normalized  concentration  as  a function  of crosswind

                  distance and  angular bearing for C stability.  Dis-

                  tance from  stack is 480 meters.
          5 XIO
    o
    H-

    cc
    l-
    z
    UJ
    o
    z
    o
    o


    o
    UJ
    N
    cc
    o
                       aoo
                          400      6OO      80O


                           DISTANCE FROM SOURCE, m
                                                          1000
                                                                   1200
       Figure  7.   Normalized  concentration as a function of distance

                   from stack  for three stability conditions.
      X = g/m3 pollutant

      N = m/sec wind  speed

      Q = g/sec pollutant emission rate
                                      29

-------

-------
Co

-------
        l-t
      0"

      rt>
     3
     fT



    O
    3
    H-
    rf
    O
    i-t
   H-
   3
  CW
  O
  n
  a
  n
  H-
  O
 3
 CO
Q.
0}

-------
 from the sampler intake by flexible ductwork (Figure 11)  positioned downwind.
 A sheet metal attachment was required and added to adapt  the motor housing
 to the ductwork.  Airborne PCB was collected on a series  of two precleaned
 polyurethane foam plugs housed in the sampler throat.   Flow rate through the
 sampler was between 10 and 30 ft3/min.   These flow rates  were selected since
 experimental data has shown that this range maximized  PCB capture while
 minimizing the possibility of breakthrough of PCB isomers through the poly-
 urethane foam plugs.  8

      The sampling sites consisted of two hi-volume samplers which were
 designed to provide continuous sampling  for the 12-hour period.   Two  timer
 systems were required for this.   One timer automated the  switching sequence
 between samplers at 3-hour intervals.  The second timer provided precise
 measurement of sampling time at  each location.   These  measurements coupled
 with  flowmeter readings at the beginning and end of the sampling period were
 used  to determine the actual sampled air volume.

 Equipment Preparation

      Extensive cleaning procedures were  required to adequately  prepare any
 surfaces that  would come in contact  with materials used for analytical
 samples.   All  metal work;  e.g.,  foam plug extensions,  filter holder and
 throat,  flange to motor support,  followed the following cleaning sequence:
 acid  soak,  water rinse,  distilled deionized water rinse,  acetone rinse,
 hexane  rinse (in field).

      All equipment that  was used  in  contact  with the foam plugs  and filters
 were  hexane washed prior to their use.   This included  tongs,  throat exten-
 sions, Teflon  gaskets and  filter holder.  Quality  control procedures used
 with  the ambient  sampling  equipment  are  further  discussed in  Section 7.

 Filter Preparation

      Filters used  for the  test were Celman  type A-E, glass  fiber filters,
 heated  (200°C) and wrapped  in rinsed, aluminum foil.

 Foam-Plug Preparation

     The polyurethane foam was prepared using the procedures described in
 "A Method for the Analysis of Polychlorinated Biphenyls (PCB) in Air"
 (EPA-600/4-78-048)8 and "High Volume Collection of Atmospheric Polychlorinated
Biphenyls,"9 requiring overnight soxhlet extraction using 9:1 hexane:ethyl
ether.

     Plugs were prepared and labeled in lots of 15.  Two foam plugs from
each lot were chosen at random and re-extracted in hexane  for quality con-
trol purposes.
                                     33

-------
                              Filter holder support
                              Teflon gasket
                              Teflon gasket
                              Aluminum throat extension
                              and foam plugs
                              Aluminum flange & motor
                              support
                              HiVol motor
                              Sheet metal adaptor for
                              exhaust duct
                              Rotometer
                              Power cord
                              Exhaust duct
Figure 11.   Hi-Vol sampling apparatus.
                34

-------
                                 SECTION 5

                     ANALYTICAL PROCEDURES AND RESULTS
PCB ANALYSIS

     Five sampling runs were conducted for PCB emissions.  These consisted
of four samples taken while the PCBs were present in the fuel and  a back-
ground sample.  All sampling runs were conducted in the same manner.  As
described in Section 3, all nonblank trains were spiked to assess PCB sam-
pling efficiency.  Each of these trains generated three types of samples:
(1) a Florisil cartridge, (2) a series of water impingers, and (3) combined
hexane and acetone rinses.  An analysis flow scheme is outlined in the test
plan.4

     Contents of each of the Florisil adsorbent tubes was Soxhlet extracted
with a given volume of hexane (200 ml) for at least 4 hours.  Upon completion
of extraction, the apparatus was allowed to cool and the contents combined with
the appropriate impinger extract and train rinse.

     The impinger waters were returned to the lab as one combined sample.
The aqueous sample in each case was transferred to a liter separatory funnel.
The sample container was rinsed with acetone and hexane which were in turn
added to the separatory funnel.  The sample was extracted with three 100 ml
portions of hexane which were then combined with the Florisil adsorbent tube
extract and the concentrate of the hexane and acetone rinses.  This combined
extract was dried via a sodium sulfate column and rotary evaporated (40°C)
to less than 10 ml.

     This extract was transferred to a 50 ml separatory funnel and partitioned
against concentrated sulfuric acid.  The layers were allowed to separate and
the acid layer discarded.  Further cleanup was not necessary at this point
since all samples extracts were colorless.

     The extract was aliquotted for each of three quantitative procedures plus
a 1 ml reserve for prescreening by gas chromatography with electron capture
detection (GC/ECD) and chromatography/mass spectrometry  (GC/MS).  Results in-
dicated that further cleanup was unwarranted, and the remaining 9 ml was di-
vided into three 3 ml aliquots.  To each portion was added a d^Q-anthracene
internal standard for quantitation by GC/MS selected ion monitoring (SIM).
The GC/MS quantitative data was collected on an HP-5985 GC/MS data system,
using the parameters listed in Appendix A, Table A-l.
                                     35

-------
     Instrument calibration was accomplished using standards obtained from
RFR, Inc., Hope, Rhode Island, which were weighed and diluted into hexane.
The GC/MS characteristics of calibration standards are listed in Appendix A,
Table A-2.

     The extract was analyzed by gas chromatography coupled with an electron
capture detector (GC/ECD).  A Hewlett-Packard 5840A with Model 7671A automatic
liquid sampler was used for this analysis.  The GC/ECD parameters for the
analysis are listed in Appendix A, Table A-3.

     Instrument calibration was accomplished using an Aroclor 1242 solution
distributed by Applied Science Labs, Inc., State College, Pennsylvania.
Calibration standards of chlorinated biphenyl isomers were supplied by RFR,
Inc., Hope, Rhode Island.

     After GC/ECD analysis, one of the 3 ml aliquots was divided into two
portions, and then perchlorinated.  The analytical procedure employed is
summarized here.

     Chloroform was successively added to the hexane extract and azeotrophic
evaporation carried out in a microconcentrator apparatus to a sample volume
of 1.0 ml.  The sample was then transferred to a reaction vial and the volume
reduced to 0.1 ml for perchlorination.  A 0.2 ml portion of SbCls was added
and the capped reaction vial heated at 160 (± 3°C) for 3 hours.

     At the conclusion of this reaction period, the vial was air-cooled and
finally cooled in an ice bath to 0°C.  The residual SbCls was neutralized by
the addition of 1 ml of 6 N HC1.  Subsequent steps include 3 X 1 ml extrac-
tions of hexane, followed by a Na2SO^ drying procedure.   The total hexane
volume is eventually reduced to 1 ml prior to quantitation.  A GC/ECD pro-
cedure was used to quantitate the decachlorobiphenyl (DCB) in the sample by
comparison of the peak area with that of a known concentration of a DCB
standard.  Computed DCB values were then converted to approximate equivalent
PCB values by utilizing the values summarized in Table 12.  A Perkin-Elmer
3920 gas chromatograph with a Ni63 electron capture detector was used to
quantitate DCB in the extract.  Instrumental parameters are listed in
Appendix A, Table A-4.  Detection of DCB in sample extracts would indicate
need for GC/MS confirmation and correction for any biphenyl contained in the
unperchlorinated extract.

     Results of GC/ECD and GC/MS analysis of stack samples presented in
Table  13 show that levels of Aroclor  1242 found in the samples were below
detectable levels in both analyses;  i.e., less than 0.001 mg and 0.002 mg,
respectively.

METHOD 5 TRAIN ANALYSIS

     Operated simultaneously with the PCB train,  the Method 5 train provided
a  total  of two  runs during PCB burning plus  a blank run with No. 6 fuel oil
only.  Each Method 5 train produced  four  types of samples:   (1) particulates
collected on a  filter,  (2) XAD adsorbent  resin,  (3) impinger condensates  and
(4)  rinses.  Their respective analytical  schemes  are shown in Figures  12  and
13.
                                      36

-------
           TABLE  12.  FACTORS TO MATHEMATICALLY CONVERT
                      DECACHLOROBIPHENYL TO AN EQUIV-
                      ALENT AMOUNT OF AROCLORa
Aroclor Average No. Cl^ MWa Xc
1221
1232
1242
1016
1248
1254
1260
1262
DCB
1
2
3
3
4
5
6
7
10
188.5
223
257.5
257.5
292
326.4
361
395.3
499
0.38
0.45
0.52
0.52
0.59
0.65
0.72
0.79
1.00
            Molecular weight of Aroclor based on the
            average whole number of chlorines calculated
            from percent chlorine substitution.

            Average whole number of chlorines calculated
            from percent chlorine substitution for a
            specific Aroclor.
           c
            X = molecular weight Aroclor/molecular weight
            DCB (499).   To convert ppm DCB to ppm for a
            specific Aroclor,  multiply ppm x DCM by X for
            the Aroclor.
        TABLE 13.  RESULTS OF STACK SAMPLE ANALYSES FOR PCB

Run
number
PCB-1
PCB-2
PCB- 3
PCB- 4
PCB- 5

Test
day
1
2
3
3
4
Extract
volume
(ml)
10.0
10.0
10.0
10.0
NAb
Weight (mg)
detected of
Aroclor 1242
(GC/ECD)
<0.001
<0.001
<0.001
<0.001
-
Weight (mg) Weight (mg) detected
detected of of chlorinated
Aroclor 1242
(GC/ECD)3
<0.002
<0.002
<0.002
<0.002
-
biphenyl isomer
(GC/MS)
<0.01
<0.01
<0.01
<0.01
-
Using perchlorination technique.
NA
Sample extract lost during analysis sequence.
                                 37

-------
                             FILTER
                              WEIGH
                               V
                             ALIQUOT	  RESERVE
                               V
                  EXTRACT W/CYCLOHEXANE  (2k  hr)
                   SOLVENT  PARTITION  W/DMF/H  0
                      CONCENTRATE  TO  1.0  ml
                                V
                      SPIKE  d1Q-ANTHRACENE
                                V
                         DUPLICATE  GC/MS
           (DIBENZOFURANS,  CHLORINATED DIBENZOFURANS,
                DIOX1NS,  AND CHLORINATED DIOXINS)
Figure 12.   Modified Method 5 train organic analysis flow scheme;
            particulate filters.
                                38

-------
         PROBE AND
       FILTER RINSES
                                                          RESIN     IMPINGERS
                                              EXTRACT W/METHYLENE  CHLORIDE
                                 COMBINE
                          CONCENTRATE TO 10 ml
                                      DRY
         3 ml
                       3 ml
      TOTAL            TOTAL
CHROMATOGRAPHABLE  CHROMATOGRAPHABLE
    ORGAN ICS           ORGAN ICS
                      GRAVIMETRIC
                       ANALYSES
GRAVIMETRIC
 ANALYSES
1  ml
                                            RESERVE
3 ml
                                                      CONCENTRATE (0.5 ml)
                                                     SPIKE d1Q-ANTHRACENE
       GAS  CHROMATOGRAPHY/
        MASS  SPECTROMETRY
 (DIBENZOFURANS,  CHLORINATED
  DIBENZOFURANS,  DIOXINS,  PAH)
           Figure  13.  Modified Method  5  train organic  analysis  flow
                      scheme:  resin.
                                      39

-------
     The recovered filters were returned to the GCA laboratory for particu-
late analysis as outlined in Figure 12.   Once particulate weights were re-
corded, the filter(s)  were aliquotted prior to extraction.  The combined
filters from each sample run were soxhlet-extracted for a period of 24 hours
in cyclohexane.  The resultant extract then underwent a solvent exchange with
a mixture of dimethylformamide/l^O followed by back extraction into cyclo-
hexane.  The sample was then concentrated on a rotary evaporator to 1.0 ml.
The extract was spiked with deuterated anthracene and scanned by means of
selected ion monitoring GC/MS for chlorinated dibenzofurans and dioxins.

Dibenzofurans/Dioxins

     The XAD resin was recovered from the sampling train in the field and
returned to the laboratory for subsequent analyses.  Figure 13 presents a
flow scheme of the resin organic analysis.  Each resin sample was soxhlet-
extracted for a period of 24 hours with methylene chloride.  The extract was
combined with the probe and filter rinses and impinger extract, concentrated
to a working volume of 10 ml, and dried.  A 3 ml aliquot was rotary evaporated
to 0.5 ml, spiked with deuterated anthracene and analyzed for the same organic
compounds as the filters.  The quantitative procedure proceeded as outlined
earlier, by GC/MS with selected ion monitoring.  The resin extract was analyzed
for dibenzofurans and dioxins not absorbed on the particulates.  Results of
the dioxin/dibenzofuran analyses are listed in Table 3.  Quality control data
on the analyses is presented in Table 14.

             TABLE 14.  GC/MS CHARACTERISTICS OF STANDARDS FOR
                        DIOXIN/DIBENZOFURAN ANALYSIS

                                         Retention^  Parent  Secondary
                                           time       ion      ion
                Compound3                  (min)      m/e      m/e
Dibenzofuran
Dibenzo-p-dioxin
2-chlorodibenzo-p-dioxin
1-chlorodibenzo-p-dioxin
2 , 8-dichlorodibenzof uran
2 , 6-dichlorodibenzo-p-dioxin
1,2,4-trichlorodibenzo-p-dioxin
1,2,3, 4-tetrachlorodibenzo-p-dioxin
DIQ anthracene
14.9
15.5
17.7
17.6
19.0
19.5
21.2
23.0
18.0
168.1
184.1
218.0
218.0
236.1
252.0
288.0
322.0
188.2
139.1
128.2
220.1
220.0
238.0
254.0
286.0
320.0
—
     Detection Limit:  0.5 ng/yl

     bGC/MS  Instrumental Parameters -  Table A-6.
                                      40

-------
     The 7 ml remainder of the XAD extract was aliquotted into a 1 ml reserve
and two 3 ml portions for assessment of total organic residues (see Figure
13).  The latter analyses consisted of total chromatographable organic (TCO)
and gravimetric analyses.

Organics

     Chemical analysis of moderately volatile organic materials (B.P. 100°-
300°C was accomplished by gas chromatography coupled with flame ionization
detection.   A summary of instrument operating conditions is presented in
Table A-5.   This technique is designed to provide both qualitative and quan-
titative information on appropriate sample extracts.  The qualitative analysis
divides each sample into discrete boiling point ranges, which are defined by
a mixture of n-alkanes (Cy-C^y).  This is accomplished through the use of
retention time data recorded for standard mixtures of these materials.

     Retention time window calibration of the procedure was accomplished by
the injection of a mixture of the normal hydrocarbons Cy through Cjy  (B.P.
100°-300°C).  Quantitative calibration was based on the integration of detec-
tor response to known concentrations of normal CQ, Ci2 and C^g hydrocarbons.
Retention times and integrated peak areas were measured using a Hewlett-
Packard Model 3380S Integrator.

     The purpose of the gravimetric analysis is to quantify organic constitu-
ents with boiling points greater than 300°C.  Typical analyses consist of
pipetting a 1.0 ml aliquot of each 10.0 sample extract into a precleaned
and tared aluminum weighing pan.  After a period of sample desiccation, to
remove residual solvent, each sample is weighed to a constant weight of ±0.1
mg.  Gravimetric determinations were accomplished by use of a Mettler ME-22
Electronic Microbalance.

     The total chromatographable or volatile organics (TCO) results reported
in Table 15 have been corrected by subtracting the integrated area response
to field-biased component blank extracts for the EPA Modified Method 5 train
resin and rinses.   The corrected TCO results correspond to the total quantity
of organic material in the 100° to 300°C boiler range not collected on the
EPA Method 5 train particulate filter.

               TABLE 15.   ORGANIC EMISSIONS USING MODIFIED EPA
                          METHOD 5 SAMPLING TRAIN (IMPINGER
                          EXTRACT/RINSE EXTRACT/TRAIN RINSES)
                                               Test day
                                      No PCB   PCB burn  PCB burn
                                      in fuel   day 1     day 3
             Total Chromatographable    33.2     54.2      35.6
               (volatile) Organics,
               (TCO) (yg/m3)
             Nonvolatile Organics      667      678       842
               Grav. (yg/m3)
             Total Organics            700      732       877
               (ug/m3)
                                     41

-------
     The gravimetric results reported in Table 15 have been corrected by sub-
tracting the gravimetric weight of field-biased component blank extracts for
the EPA Modified Method 5 train resin and rinses.  The corrected gravimetric
result quantitates the organic sample components with boiling points higher
than 300°C which were not collected on the particulate filter of the sampling
train.

Chloride

     The chloride analyses of the sodium carbonate impingers were accomplished
with a Dionex Model 14 Ion Chromatograph.  The column system employed:
(1) a pre-column (3 x 150 mm) to remove particulates, strongly retained anions,
and organic species; (2) a separator column (3 x 500 mm) in the HC03- form to
chromatographically separate the anions, and (3) a suppressor column (5 x 250
mm) in the H+ form to remove the background conductivity of the eluent.  A
solution of 0.003 M NaHC03/0.0024 M Na2C03 was issued for the eluent at a 30
percent flow rate (about 140 ml/hr).  The injection loop had a capacity of
100 pi and the sample was introduced from a 5 ml disposable syringe fitted
with a 0.22 pm Millipore filter to remove particulate matter.

     The height of the chloride peak was measured and converted to the appro-
priate concentration by comparison with standard solutions.  The chloride
concentrations in the standards ranged from 0.2 to 20 ppm.  A 3 x 500 mm
separator column was used to separate the chloride peak from the anionic
impurities in the sodium carbonate impinger solution.

     The results from the first impinger analysis have been aliquot corrected.
Correction for blank contributions was unnecessary because both the distilled
water and the Na2C03(aq) preservative contains no measurable chloride (0.2
g/ml).  Results of the analysis are listed in Table 16.

                  TABLE 16.  METHOD 5 TRAIN CHLORIDE ANALYSIS
Run
number
M-5-1
M-5-2
M-5-3
1st
impinger
(mg)
19
231
sample
lost
2nd and 3rd
impingers
(mg)
2
3
7
Chloride
(mg)
21
234

VMSTDa
(m3)
4.808
5.446
3.290
ug/m3
cl
4,000
43,000

            3VMSTD = Volume of gas sample through the dry gas meter
              (corrected to standard conditions).

     The chloride analyses was performed in an attempt to measure the relative
 contribution  of PCBs to stack chloride emissions.  However, the test results
 were inconclusive.  As indicated in Table 16, the background  (no PCB) test run,
 M-5-1, had a  measurable chloride emission of 4000 pg/m3.  The initial PCB run,
                                      42

-------
M-5-2, showed greater than a tenfold increase over this background value.
While all of this increase might appear to be attributable to chlorine from
PCB destruction, a closer examination of the chlorine input values tends to
discount this assumption.  The chlorine content of the residual oil used for
the test was measured to be 0.97 percent, or 9700 ppm, as shown in Appendix
G.  The maximum PCB concentration of the input fuel stream on test day 2
was 72 ppm, as shown in Table 4.  If one assumes that the PCB utilized during
the test was Aroclor 1242, then three equivalent weights of chlorine are
liberated per weight of PCB destroyed, as shown in Table 12.  The net contri-
bution of PCB to overall chlorine emissions would therefore be expected to
be minimal.  The tenfold increase in chloride concentration resulting from
the PCB burn cannot be fully explained.  Interferences with background or-
ganic substances may have led to false chloride peaks on the Ion Chromatograph,
which were interpreted to be chloride.  Unfortunately, the chloride sample
for run number M-5-3 was lost so a confirmation of the tenfold chloride in-
crease could not be substantiated and a statistically significant relation-
ship between PCB destruction and chloride emissions could not be established.

Particulate

     The recovered Method 5 sampling train filters were returned to the GCA
laboratory for analysis as outlined in the test plan.4  Once particulate
weights had been recorded, the filter(s) were aliquotted prior to extraction.
The combined filters from each sample run were soxhlet-extracted for a
period of 24 hours in cyclohexane.  The resultant extract was solvent ex-
changed with a mixture of dimethylformamide/H20 followed by back extraction
into cyclohexane.  Prior to concentration, each extract was passed through
a column of anhydrous sodium sulfate.  Each sample was concentrated on a
rotary evaporator to 1.0 ml.  Extracts were spiked with deuterated anthracene
and scanned by means of selected ion monitoring GC/MS for the compounds listed
in Table 14.  The GC/MS instrument parameters are shown in Appendix A, Table
A-6.

     The pH of the impinger water samples was adjusted first to 2.0, then to
12 for duplicate extractions (3 at each pH) with 20 percent (V:V) portions
of dichloromethane.  The extracts were held for combination with the XAD
extract and train rinses.

Particulate Weight

     Particulate filters were not available for particulate weight determina-
tion in two of three Modified EPA Method 5 runs, Table 17.   This is contrary
to the original test plan for the following reason:  after the respective run,
particulate filters were wrapped in hexane rinsed aluminum foil, and stored
in glass petri dishes.   The filters were found adhered to the aluminum foil,
making a complete weight determination virtually impossible without adversely
affecting the TCO determination.  The assumption was then made that, with
steady-state boiler operation and visibly similar particulate loadings, the
one particulate filter would be representative of the particulate weight in
the two other runs.
                                     43

-------
                       TABLE 17.  PARTICULATE WEIGHT
                                    Particulate concentration
                      Particulate   	
               Run    weight (mg)   (gr/dscf)a        (Ib/hr)

              M-5-1      89.47        0.0081          0.0658

              M-5-2  Not available       ->               •>

              M-5-2  Not available       ->               ->•

              o
               gr/dscf = grams per dry standard cubic foot.


IN-PLANT PCB ANALYSIS

     The analytical techniques used were those prescribed by the National
Institute for Occupational Safety and Health1 (P&CAM No. 253, see Appendix
C).   Approximately 350 milligrams of Florisil were packed into each glass
sampling tube to capture airborne PCBs by means of adsorption.  All tests
showed nondetectable levels of PCB which indicates that concentrations were
less than 3 yg/m3 and 12 yg/m^ for Aroclor 1242 and Aroclor 1254, respec-
tively.  The results are given in Appendix J.

PCB BLOOD CHEMISTRY ANALYSIS

     The blood samples were split and analyzed by the State of Michigan
Department of Health Laboratory in Lansing, Michigan and the Raltech Lab-
oratory in Madison, Wisconsin.   The results are given in Appendix L.  The
results showed that, within the limits of accuracy of the testing, there
were no significant changes in the preburn and post-burn PCB blood samples
of the GM powerhouse operators.  Furthermore, the PCB blood levels of these
operators was comparable to the GM office employees with the least exposure
to environmental PCB.

BOILER PCB RESIDUAL ANALYSIS

     The analytical technique used to detect PCBs in the boiler residual
samples (solid bulk material or scrapings) is outlined in Appendix D.
The results, which are given in Appendix K, indicate that, except for an
expected buildup around the burner tips, (1) the readings were either close
to or below the nondetectable levels, and  (2) no significant buildup was
occurring.

FUEL ANALYSIS

     The waste oil, spiked oil, and mixed oil samples were analyzed by several
laboratories.  The analytical procedures used by Chevrolet Central Office and
Chevrolet Bay City laboratories are referenced in Appendix E.  The procedures
used by Trace Elements Inc. and Phoenix Chemical Laboratories Inc. are refer-
enced in Appendices H and I, respectively.  The results are given in
Appendix G.

                                     44

-------
     GCA's  fuel analysis procedures can be differentiated by  the type of
 analysis performed.  For the PCB  analysis, fuel oil  sample bottles were
 heated  at 40°C in  a water bath  and stirred for 1 minute prior to aliquotting
 to ensure sample homogeneity.

     A  0.5  g  portion of each sample was transferred  to a clean, pretared
 5.0 ml  volumetric  flask.  The flask was reweighed  and the sample diluted  to
 5.0 ml.  The  well-mixed sample  was removed to a clean, 10 ml  vial contain-
 ing 5.0 ml  of concentrated sulfuric acid for cleanup per the  procedure cited
 in the  Federal Register, May 31,  1979  (40 CFR 761).  The oil/acid phases
 were mixed  and the oil fraction quantitatively transferred to a clean vial
 containing  5.0 ml  of 10 percent NaHC03  (aq).  A portion of the oil fraction
 was then transferred to a septum  vial for analysis by GC/ECD  using the
 instrument  parameters detailed  in Table A-3.

     Qualitative identifications  and quantitations were performed according
 to the  procedures  outlined for  ambient analyses.   Results of  these analyses
 are shown in  Table 1.

     For the  dibenzofuran/dioxin  analysis, fuel oil  sample bottles were
 heated  to 40°C in  a water bath  and stirred for 1 minute prior to aliquotting
 to ensure sample homogeneity.

     A  1.0  g  portion of each sample was transferred  to a clean, pretared
 5.0 ml  volumetric  flask.  The flask was reweighed  and the sample diluted  to
 5.0 ml  with dichloromethane.  GC/MS analyses were  performed as indicated  in
 Table 14 and  Table A-6 in Appendix A.

     A  summary of  the waste oil analysis is presented in Appendix G.  A dis-
 cussion of  the waste oil data is  given in Appendix N.  These  data indicate
 that, apart from the variation  in interlaboratory  fuel oil PCB measurements
 disucssion  in Section 7, the residual oil used in  the test series contained
 no unusual  or atypical concentration of any metallic or organic species that
 would tend  to affect the destruction efficiency results.

 AMBIENT PCB ANALYSIS

     Polyurethane  foam plugs and Hi-Vol filters were collected from ambient
monitors and  returned to the laboratory in precleaned amber glass jars with
 Teflon-lined  caps.   The samples were combined as necessary and soxhlet ex-
 tracted overnight with hexane.   After cooling,  the extracts were transferred
 to round bottom flasks and the volume reduced to 2 ml via rotary evaporation.
The extract was then divided into a reserve plus a portion for GC/ED
analysis of PCB.   A 0.5 ml aliquot was subjected to alumina cleanup as re-
commended in  "Sources of Emissions of Polychlorinated Biphenyls in Ambient
Atmosphere and Indoor Air."10  Each extract was further cleaned by extraction
with concentrated  sulfuric acid and analyzed using the instrumental param-
eters listed  in Table A-l.

     Qualitative identification of PCB was performed using the Aroclor pattern-
matching technique.  Instrument calibration was accomplished by injection of
0.2,  2.0 and  10.0 ppm (g/ml)  dilutions of  an Aroclor 1242 standard prepared

                                      45

-------
in hexane.  These dilutions were prepared from a commercially available stock
solution distribution by Applied Science Labs, Inc., State College, Pa.
Calibration was checked every sixth sample to ensure a uniform detector
response.  Figure 14 is a typical Aroclor 1242 standard chromatogram.  The
retention times of the five peaks used for area summation calibration were
2.92, 3.95, 5.20, 6.73, and 9.80 minutes.

     Calibration curves were prepared daily from a linear regression analysis
of the above three Aroclor dilutions.  Each of these data points was obtained
by the summation of five appropriate peak areas.

     Calibration curves were rejected if the correlation coefficient of the
linear regression analyses was less than 0.996.  All samples were quantitated
by entering the area summation into the appropriate calibration curve.  Sample
extracts were diluted, if necessary, to be within the concentrations bracketed
by the calibration standards.

     Results of GC/ECD analyses of ambient PCB samples are listed in Table 5.
A typical sample chromatogram of the extract derived from 1 day's ambient
sample is shown in Figure 15.
                                     46

-------
                                                                          El
                                                                          N
->TIME, minutes
   Figure  14.   GC/ECD  Chromatogram of Aroclor 1242 Standard .

-------
                               197
               <\J

               SI
                           in

                           n
C»
                                                                                                          05
                                                                                                          OJ
                          TIME, minutes

                          Figure  15.   GC/ECD chromatogram of day  1—downwind site 1—preburn.

-------
                                  SECTION 6

                       AIR QUALITY MODEL CALCULATIONS


      The  rural  version of the  RAM Gaussian plume model' developed by  the EPA
 Meteorology  Laboratory was used  to calculate  the average  PCB  concentration
 fields  downwind from  the  facility during the  three  test periods.  This ver-
 sion of the  model  is  based on  algorithms similar to those used  in CRSTER.
 It  is designed  for application to one  or several days  of  meteorology  rather
 than a  full  year.   Plume  rise  is calculated by  the  Briggs equations.1 The
 requisite model input parameters were  developed from stack test data  and hourly
 onsite  meteorological measurements.

      In carrying out  the  model calculations,  receptors were located to ensure
 area coverage sufficient  to define the approximate  location and magnitude of
 the highest  concentration.   Multiple model runs were carried  out.  One run was
 carried out  for each  trial to  define the concentration field  that would be
 expected  at  a height  of 2 m in the absence of building structures.  Additional
 runs were made  to  determine concentrations expected at monitors located at
 other heights.

      Table 18 lists the meteorological and source parameters  used in  the model
 calculations.   Table  19 shows  the monitor locations with  reference to the
 boiler  stack during each  test  day,  and Table  20 gives  the heights of  the
 monitors  above  ground.

      Figures 16 through 19 are combined plots showing  the locations of the
 monitors  and the prevailing wind direction on each  of  the four  test days.
 In  these  figures wind direction  percentages are indicated by  vectors  originat-
 ing at  the boiler  and flowing  with the wind.  In Figure 16, the vectors repre-
 sent winds during  the entire test day  prior to  any  PCB burn.  On days with a
 PCB burn  (Figures  17,  18,  and  19),  however, the vectors represent winds during
 the period of the burn  only.  As  can be  seen  from Figures  17  through  19, place-
 ment  of the  ambient monitors to measure  expected maximum PCB  concentrations
 during burn days was generally good.  Burn number  3 had the best overall
monitor placement,  with monitor 2 located directly  downwind of the principal
wind direction,  as  shown in Figure 19.   All downwind monitors  were within  30
compass degrees  of  the prevalent  wind direction  during  burn number 1   (Figure
17).  Only during burn number 2 (Figure 18) were the three downwind monitors
as much as 30 to 60 compass degrees from the true wind  direction for  the burn
period.
     Figures 20, 21, and 22 show  the 2m concentration isopleths predicted  by
 the model for each of the PCB burns.  Maximum impact of the boiler during
 these test burns at ground level   (2 meters) is of  the  order of 7-10 x 10~3


                                     49

-------
TABLE 18.  MODEL INPUT PARAMETERS FOR PCS BURNS
Test
day Date Period
2 5/08/80 1430-1530
1530-1630
1630-1730
1730-1830
1830-1930
1930-2030
3 5/09/80 1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
4 5/10/80 1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
Exit
velocity
(m/sec)
8.5
8.5
8.5
8.5
8.5
8.5
9.3
9.3
9.3
9.3
9.3
9.3
7.1
7.1
7.1
7.1
7.1
Stack height (above
Stack
temp.
(°0
129
129
129
129
129
129
132
132
132
132
132
132
116
116
116
116
116
ground)
Stack diameter
Assumed source strength
Wind
direction
(deg)
305
305
285
290
300
305
260
240
240
240
255
260
210
205
190
185
170
18.3 m
1.07 m
1.0 x 10-6
Wind
speed
(m/sec)
5.6
6.0
6.0
5.6
5.4
4.2
5.6
6.0
5.6
5.4
4.2
3.8
9.8
8.3
5.6
5.8
6.0


g/sec
Stability
class
C
C
D
D
D
D
C
C
C
D
D
D
D
D
D
D
D



                      50

-------
    TABLE 19.  LOCATION OF MONITORS DURING TEST PERIODS
                        Site identification
TC'Ht	
day         1              2
       x(m)   y(m)   x(m)   y(m)   x(m)   y(m)   x(m)   y(m)

 1     -91    -96     76     84    133    126    553     503

 2     -82     52    238    -24    254    -20    445    -198

 3     -82     52    238    -24    254    -20    445    -198

 4     -91    -96     49    258    201    805    402     872

Note:  Origin of coordinate system is at boiler stack.
             TABLE  20.   MONITOR HEIGHTS DURING
                        TEST PERIODS
Test
day
1
2
3
4

Site
1
2.3
2.3
2.3
2.3
Height (m)
identification
2 3
1.0 1.0
8.9 8.9
8.9 8.9
1.0 1.0


4
2.0
2.0
2.0
2.0
                              51

-------
        330
300-
270«
240°
        210"

      Figure  16.
                   180
ISO'
Locations of monitors  and  flow direction vectors
 during test day 1  (no burn).
                                   52

-------
       330
                                                                 30'
300"
270
 240*
                Location of monitoring and wind direction  rose  during
                burn No. 1 (test day 2).
                                   53

-------
300*
270"-
240°
                 Location of monitors  and wind direction rose during
                 burn No. 2  (test  day  3).
                                     54

-------
        330
                                             30'
300°
270"
240"
         210°

        Figure  19.
                                     180
                                                                  150'
Location of monitors  and wind direction rose
during burn No,  3  (test  day 4).
                                      55

-------
                            1—200
Ui
                N
         -200
                             --200
SCALE IN METERS
M = MONITOR
                                             200
                             --400
                                                                                         4XIO'3
                             "—-600
            Figure  20.   Predicted facility impact on PCB concentration field during burn No.  1  (test
                         day 2).   (Distance scale is in meters,  concentration isopleths are in ng/m3,  M
                         designates a monitor).

-------
            600i—
            400
             200
       M- I
     -100
            -aoo1—
                        SCALE  IN METERS
                        M = MONITOR
                                                     M-4
Figure 21.
Predicted facility impact on  PCB concentration field during burn No. 2  (test
day 3).  (Distance scale is in meters,  concentration isopleths are in ng/m^,  M
designates a monitor).

-------
                           1—2000
      -500
                                             -3
                     M- I
                                           500
                                   SCALE IN METERS
                                   M = MONITOR
                          I—-500
                                                              1000
                                                             _J
Figure 22.
Predicted facility impact on PCS concentration  field  during
burn No. 3 (test day 4).  (Distance scale  is  in meters,  con-
centration isopleths are in ng/m^, M designates a monitor).
                                    58

-------
ng/m .  These estimates are based on a source strength of 1 x 10 ^ g/sec PCB,
which is the indicated upper limit of possible source strengths during the
test burns.  (Stack sampling indicated source strengths less than 9.7 x 10~^
g/sec).  It can be seen from these figures that the monitors were satisfac-
torily positioned except for Test Day 3 (Burn No. 2) when little, if any, of
the boiler plume impacted on any of the monitors.  This unsatisfactory monitor
placement for Test Day 3 can also be seen in the wind rose plot for this
day, Figure 18.

     Table 21 compares the maximum expected boiler impact during the various
trials with the observed ambient concentrations.  The results are striking.
In every case the observed concentrations greatly exceed the values predicted
by the model, differing, on the average, by about 5 orders of magnitude.
Thus, even allowing for modeling error, it is quite clear that the impact of
the boiler upon the existing ambient PCB levels is insignificant.  This is
in agreement with the statistical analysis presented in Table R-l,  Appendix R,
where upwind and "no-burn" concentrations are compared with downwind "burn"
concentrations.

     It is also of interest to compare the stack concentration of PCB during
the test burns with ambient concentrations.  Assuming a PCB emission rate of
1 x 10~6 g/sec yields stack concentrations ranging from about 120 to 160 ng/m3
during the various burns.   This is less than the average of the 20 ambient
samples (203 ng/m3) collected at upwind locations and during no burn periods.
How much lower the actual stack concentration was than the assumed 1 x 10~6
g/sec is not known because of limitations in the sampling procedure.

     Table 22 gives the wind direction distributions for the various sampling
periods.  Examination of these distributions in conjunction with the observed
concentrations (Table 5) shows the following feature of note.  The highest
concentrations of all (1400 and 500 ng/m3) were observed at Sites 2 and 3 on
Test Day No. 1 prior to conducting any PCB incineration.  During this period
winds were from the southwest (230-250 degrees).   When the wind shifted to
340 degrees the concentrations dropped dramatically to 50 and 15 ng/m3,
respectively at the two sites.   This behavior is suggestive of a PCB source(s)
to the southwest of the monitors.
                                    59

-------
                    TABLE 21.  COMPARISON OF MAXIMUM EXPECTED FACILITY  IMPACT
                               AND OBSERVED CONCENTRATION OF PCB
Test
day
1
2
3
4




Upwind — 1
Max . exp .
_ 1
3
3
1
Obs.
.9
.5
.9
.2
X
X
X
X
102
102
10 2
102
Site
2
Max. exp. Obs.
4.3 x
0.19 x 10~3 0.84 x
0.17 x 10~3 0.45 x
1.80 x 10- 3 1.3 x
identification


102
102
10 2
102

Max.

0.11
0.29
2.56
3
exp. Obs.
2.9 x 102
x 10-3 0.82 x 102
x 10-3 0.55 x 102
x 10~ 3 0.26 x 102
Population — 4
Max . exp .
0
4.21 x 10-3 0
0 0
3.39 x 10-3 0
Obs.
.63 x
.67 x
.28 x
.51 x

102
102
102
102
Notes:  Concentrations are in ng/m3.

        Concentrations are for entire sampling day at Sites 1 and 4; for Sites 2 and 3,
        concentrations are for period of burn only.

-------
           TABLE 22.   DISTRIBUTION OF WIND  DIRECTION DURING VARIOUS  SAMPLING PERIODS
Wind3
direction "Pre-
(deg) burn"
160
170
180
190
200
210
220
230 1
240 4
250 1
260
270
280
290
300
310
320
330
340
Test day 1 Test day 2 Test day 3 Test day 4
"Post- "Pre- "Post- "Pre- "Post- "Pre- "Post-
"Burn" burn" Total burn" "Burn" burn" Total burn" "Burn" burn" Total burn" "Burn" burn" Total
2 2
2 2
1 1
11 3 3
1 1 2
53 8
1 1
1
4 617
12 22
11 3317
55 314
1 2 3
1 3 3
1517
1 123 5
1 12 13
22 11
167
 Degrees indicate direction  from which the wind is blowing based on 30 minute averages.
Note:  Observations are 30-minute averages.

-------
                                 SECTION 7

                     QUALITY ASSURANCE/QUALITY CONTROL


     As designated in the program test plan, a detailed quality control pro-
gram was conducted throughout the course of the sampling and analysis effort.
The primary elements of the program are highlighted below.

QUALITY ASSURANCE PROTOCOL

     Quality assurance audits were conducted during the course of the program.
These included reviews of field and laboratory test data, notebooks, and
standard operating parameters, in addition to periodic reviews of program
quality control data.  The results of these audits, fortified further the
quality of sampling and analysis data.  The tests of these audits are on file
with the contractor for this report.

INVESTIGATIVE ACTIONS

     Investigative actions initiated  during the program were prompted by
interlaboratory data discrepancies in the reclaim waste oil analysis.  As
shown in Table 4, the GM and GCA analyses for fuel oil PCB concentrations
varied considerably.  To ascertain the cause of this discrepancy, EPA
assigned Research Triangle Institute  (RTI) to conduct a program audit of
GCA and GM data, Quality Assurance/Quality Control (QA/QC)  procedures and
recordkeeping and procedural documentation related to the sampling of the
test burn oil.  These findings are reported in a companion report.   RTI
found that the differences among the  GM and GCA laboratories appear to
represent the analytical variability  that is inherent in the analysis for
complex mixtures containing PCBs.  A  complete documentation of the fuel oil
sampling and analysis procedures followed by GCA is provided in Appendix N.

QUALITY CONTROL PROCEDURES

     The quality control program consisted of the following:  (a) detailed
quality control acceptance criteria for all reagents, solvents, and adsorbent
media utilized during the course of the program,  (b) quality control of
analytical procedures, including collection and analysis of appropriate
method blank, and  (c) validation of analytical quality.

Reagent Quality Control Criteria

     Acceptance criteria were implemented for all reagents, solvents, and adsor-
bent media utilized in the program.  The Laboratory Analysis Department of GCA
routinely monitors all solvents  for gravimetric residues and chromatographic

                                     62

-------
interferences (GC/FID).  Each of  these measurements were  collected for the
solvents  utilized on this program including methylene chloride,  hexane and
acetone.  Where appropriate, electron capture acceptance  criteria were imple-
mented  to quantitate potential  interferences in all GC/ECD measurements;
e.g., PCBs.

     Preparatory cleanup procedures were used on all adsorbent media including
solvent extraction and conditioning of all Florisil, polyurethane foam plugs
and XAD-2.   Acceptance criteria established for each of these media were met
prior to  distribution for field use.

Quality Control Criteria for Blanks

     The  quality control program  required pretest analysis of all solvents,
reagents  and sample media for possible interference via the proposed analyt-
ical method.   The criteria for  quality assessment are summarized in Table 23.
                 TABLE 23.  QUALITY  CONTROL CRITERIA/BLANKS
            Matrix
  Analysis
  procedure
                                        Criterion for evaluation
                         Number of
                          samples
      Hexane  (DIG)

      Acetone (DIG)

      Florisil (30/60 mesh)

      Polyurethane foam
      plugs (2)  + HiVol
      filter  (1)

      Dichloromethane
        (DIG)


      XAD-2 resin
      Cyclohexane
        (DIG)

      8% Na2C03(aq)
GC/ECD

GC/ECD

GC/ECD


GC/ECD


GC/MS
TCO
Gravimetry

TCO
Gravimetry

GC/MS

Ion
Chromatography
<1.0 ug/1 Aroclor 1242          2
<1.0 ug/1 Aroclor 1242          2

-0.5 pg Aroclor 1242/7.5g       2


O.05 ug Aroclor 1242          11


<10 ug/1 Dibenzofuran/Dioxin     2
<0.1 mg/1                     2
<0.5 mg/1                     2

<10 rag/kg                     2
<20 mg/kg                     2

<10 ug/1 Dibenzofuran/Dioxin     2

<0.2 mg/1 Cl~                 2
     Another  requirement of the quality  control program was the  analysis of
a series of field-based blanks.  A list  of  the samples taken and the  analyt-
ical results  is  presented in Table 24.

Quality Control  of  Analytical Procedures

     In accordance  with the program text requirements, a variety of quality
control measures were implemented during the  analysis period.  These  included
use of a series  of  spiked samples provided  by the GCA Organic Laboratory Quality
Control Coordinator.   Each of the representative matrixes were spiked with
a known quantity of Aroclor 1242 and submitted into the analytical scheme
as a "blind"  spiked sample.  The spiking solution was prepared by the QC
Coordinator from an RTP/EPA standard reference material.  (Lot No. 8937,
100 percent Purity  Received 5/27/80 from Quality Assurance Section, Analyt-
ical Chemistry Branch, U.S. EPA, HERL, ETD  (MD-69), Research Triangle Park,
North Carolina 27711).
                                      63

-------
   TABLE  24.   QUALITY CONTROL  CRITERIA/FIELD-BIASED BLANKS
                                                               Number
                             Analysis                             of
         Matrix              procedure      Analytical results    samples
PCB traina - florisil
          - impinger
            waters
          - hexane/acetone
            rinses
GC/ECD         <1.0  yg Aroclor 1242
Method  5 train3 - filter     GC/MS           <0.5  yg Dibenzofuran-
                                                 Dioxin           2
          -XAD-2 resin      GC/MS           <0.85 yg Dibenzofuran-
                                                 Dioxin
          -XAD-2 resin      TCO/Gravimetry  <10 mg/kg/<20 mg/kg      2
          -distilled, de-   Ion
           ionized  water    Chromatography  <0.2  mg/1 Cl~            2
         - 8% Na2C03(aq)    Ion
                           Chromatography  <0.2  mg/1 Cl~            2

Polyurethane foam plugs  (2)
  + HiVol filter (1)         GC/ECD          <0.3  yg Aroclor 1242      2
 One complete train was assembled  for each day of  testing.

 Train components only.
       TABLE  25.  PCB QUALITY  CONTROL  - FLORISIL  SPIKES


A
B
C
X
Aroclor
Actual
13
13
13
13.0(13.4)
1242 (vig)
Observed
15
12
None
13.0(13.5)

% Recovery
112
89
-
100
        aNone  - Sample extract soiled  during analysis
                 sequence  and could not be  recovered.
                                   64

-------
Florisil Spikes--
     Each of the three aliquots of Florisil adsorbent were spiked prior to
extraction with a specified quantity of Aroclor 1242.  These were prepared as
an indication of the measurements accuracy of PCBs adsorbed on Florisil.  The
results of these measurements are summarized in Table 25.  The average results
of these determinations indicate acceptable recoveries for Aroclor 1242 ad-
sorbed on Florisil adsorbent.

Impinger Waters—
     Quality control monitoring of PCBs collected in impinger waters consisted
of the use of "blind" spiked water samples submitted during the appropriate
analysis sequence.  Each of three water samples were prepared by the Organic
Laboratory Quality Control Coordinator and submitted for analysis.  The results
are recorded in Table 26.

                       TABLE 26.  PCB/AQUEOUS MEDIA
                        Aroclor 1242 (yg)

A
B
C
Actual
(yg)
6.7
6.7
6.7
Observed
(yg)
7.5
7.2
8.9
% Recovery
112
107
133
                        6.7         7.9           117
     The results above, while somewhat high, are consistent with acceptance
criteria maintained for the analysis of PCBs in aqueous media.

PCB-Train Spikes—
     As designated in the program test plan, sampling trains were spiked in
the field with a solution of deuterated tetrachlorobiphenyl in acetone to
provide procedural recovery data.  The deuterated isomer was supplied  (97
percent d) by KOR Isotopes, Cambridge, Mass.  The spike was applied prior to
the start of each sample run to the first impinger.  Stack samples were quan-
titated for d6-tetrachlorobiphenyl isomer during the GC/MS analysis sequence
for PCBs.  The results of these analyses are summarized in Table 27.

     As can be seen in Table 27, the recovery efficiency of deuterated tetra-
chlorobiphenyl was low.  The cause of this low recovery was investigated by
GCA personnel.   Potential sources of this problem were the spiking procedure
used in the field and/or the laboratory procedure used to analyze the sample.
Discussions with field and laboratory personnel indicated that normal pro-
cedures were followed.  However, due to the relative newness of this sampling
and analysis procedures and the infrequency of its use, an inadvertent error
may have occurred.  This could not be substantiated.
                                     65

-------
     An attempt was made by the laboratory personnel to isolate any analysis
error.  These investigations included determination of the extraction effi-
ciency of the d6~isomer from an aqueous sample under a number of pH conditions,
These laboratory checks failed to pinpoint the source of the low recovery
efficiency.

              TABLE 27.  PCB QUALITY CONTROL - TRAIN RECOVERY
Run No.
PCB-1
_2
-3
_4
-5
Test
day
1
?.
3
3
4
Air volume
sampled
(dscm)'
4.898
5.330
5.785
5.964
3.207
dQ-'to-Liac'i'lorobip-'ic-ny I
actual
(nr.)
u
31
11
11
11
Observed
(i-s)
2.0
1.0
4.7
3.1
N/A
X
% Recovery
18
9.1
43
28
-
= 25
   L)ry standard cubic meters.

  N/A - Sample lost during analytic'1! sequence
 Polyurethane Foam Spikes—
      Each of three sets  of polyurethane foam plugs were prepared by the Organic
 Laboratory Quality Control Coordinator.   These were submitted during the ambi-
 ent portion of the analysis program as "blind" spiked samples.   The results
 of these are summarized  in Table  28.

                      TABLE 28.  PCB QUALITY CONTROL -
                                FOAM PLUG SPIKES


A
B
C
X
Aroclor
Actual
6.7
6.7
6.7
6.7
1242 (ug)
Observed
6.0
6.2
6.2
6.1

% Recovery
90
93
93
91
      The average results of these measurements  are consistent with recoveries
 of 96 percent reported by other investigation.°
                                      66

-------
Polyurethane Foam Plugs - Retention Efficiency Study—
     Additional quality control measures included a collection efficiency
test conducted at the GCA/Technology Division Laboratory in Bedford,
Massachusetts.  This test consisted of spiking a set of polyurethane foam
plugs and Hi-Vol filter with a known amount of Aroclor 1242.  This spike was
applied directly to the filter in an assembled Hi-Vol apparatus.  The spiking
solvent was allowed to evaporate before startup of the apparatus.  The
sampler was operated at a flow rate of 26 ft^/min for 3 hours.  A blank
sampler was operated simultaneously to provide background correction.

     The samples were collected and analyzed for PCB according to previously
described methods for ambient samples.  The results are summarized in
Table 29.

                TABLE 29.  PCB QUALITY CONTROL - FOAM PLUG
                           RETENTION EFFICIENCY

Spike
Blank
Volume of
sampled air
(dscm)
131
112
Aroclor 1242 (yg)
Actual Observed
5.0 5.0
<0.5
% Recovery
100%
          Dry standard cubic meters.

     The results indicate sufficient retention of Aroclor 1242 on the foam
plug matrix using a 26 ft^/min flow rate and a 3-hour sampling time.

Fuel Oil Analysis—
     In accordance with the program quality control protocol, a series of
spiked fuel oil samples were also analyzed.  Each of three 0.5 grams aliquots
of No. 6 fuel oil were spiked with 25 yg of Aroclor 1242 (Applied Science) and
analyzed per the procedure outlined earlier in the analytical method section
of this report.  The results of these analyses are summarized in Table 30.

                TABLE 30.  QUALITY CONTROL - PCBs/FUEL OIL

                          Aroclor 1242 (yg)
         Sample     Actual (yg)     Observed (yg)     % Recovery
A
B
C
X
S
25
25
25
25
-
24
24
23
24
0.58
96
96
92
96
_
             X
                                      67

-------
     The above data indicates acceptable results  for  the  determinations  of
PCBs in a fuel oil matrix.

Data Quality

     The quality of program data was  maintained jointly by  the  GCA/Technology
Division Quality Control Committee in conjunction with the  program staff in-
cluding the services of a Program Statistician, Mr. Ralph D'Agostino.

     The details of the data validation sequence  included the following:
daily verification of all program measurements by appropriate staff members;
e.g., laboratory notebook entries, periodic  audits of program data by  GCA QA
Manager, periodic submission of quality control samples with subsequent  data
reports by the organic laboratory QC  coordinator, and statistical validation
of program test results prior to final report submission  by the program
statistician.
                                     68

-------
                                 REFERENCES
 1.    NIOSH Manual  of  Analytical Methods,  2nd  Edition, Volume  1.  U.S. Depart-
      ment  of  Health,  Education, and Welfare,  Public Health  Service, Center
      for Disease Control,  National Institute  for Occupational Safety and
      Health,  Cincinnati, Ohio, April  1977.

 2.    Collins, P. F.,  and G.  F. Hunt.   Evaluation of PCB Destruction Effi-
      ciency in an  Industrial Boiler:   Audit Report.  EPA-600/2-81-055b,
      Office of Research and  Development,  U.S.  Environmental Protection Agency,
      Washington, D.C.   April 1981.

 3.    Levins,  P.  L., C.  E.  Rechsteiner,  and J.  L. Stauffer.  Measurement of
      PCB Emissions  from Combustion Sources.   EPA-600/7-79-047.  Office of
      Research and  Development, U.S. Environmental Protection  Agency,
      Washington, D.C.   February 1979.

 4.    Zelenski, S.,  J. Hall,  and S. Haupt.  Applying for a Permit to Destroy
      PCB Waste Oil.   Volume  I.  Summary.  EPA-600/2-81-033a.  U.S. Environ-
      mental Protection  Agency, Washington, D.C.  March 1981.

 5.    IERL-RTP Procedures Manual:  Level I Environmental Assessment  (Second
      Edition).  EPA-600/7-78-201.  Office of  Research and Development, U.S.
      Environmental  Protection Agency,  Washington, D.C.  October 1978.

 6.    Flynn, N. W.   Analysis  of Polychlorinated Dibenzofurans  and the Iden-
      tification of  Selected  Products  from Thermal Destruction of Poly-
      chlorinated Biphenyls.   Proceedings  of Symposium on Process Measure-
      ments for Environmental Assessment.  Atlanta.  1980.

 7.    Turner,  D.  B.  Workbook of Atmosphere Dispersion Estimates.  U.S.
      Department of  Health, Education,  and Welfare.  Cincinnati, Ohio.  1970.

 8.    Stratton, C.  L., S. A.  Whitlock,  and J.  M. Allan.  A Method for the
      Analysis of Polychlorinated Biphenyl  (PCB) in Air.  EPA-600/4-78-048,
      U.S.  Environmental Protection Agency, Research Triangle  Park, N.C.
      August 1978,

 9.    Bidleman, T.  F., and  C. E. Olney.  High-Volume Collection of Atmospheric
      Polychloriated Biphenyls.  Bulletin  of Environmental Contamination and
      Toxicology 11:5  (1974).

10.    Macleod, K.   Sources  of Emissions of Polychlorinated Biphenyls in the
      Ambient  Atmosphere and  Indoor Air.  March 1979.


                                      69

-------
11.    Chemistry Laboratory Manual  for  Bottom Sediments  and Elutriate Testing.
      EPA-905/4-79-014,  U.S.  EPA,  Surveillance and Analysis Division, Region V,
      Chicago.   March 1979.

12.   Kehlf,  F. James, and Robert K.  Sokal.   Statistical Tables W.  H.  Freeman
      and Co.   San Francisco, Calif.   1969.   pg.  161, Critical Values  of
      Students to Distributor.
                                       70

-------
                                  APPENDIX A

                    ANALYTICAL EQUIPMENT OPERATING PARAMETERS


          TABLE A-l.   GC/MS  INSTRUMENTAL PARAMETERS FOR PCB ANALYSIS
Parameter
          Gas Chromatographic Conditions
Column:      6 ft X 2 mm (I.D.) Pyrex packed with 1% SP2250 on 80/100 mesh
             Supelcoport

Injection
Temperature: 275°C
Temperature
Program:
•  isothermal for 4.0 min at 160 C

o  linear program from 160°-270° at 10°C/min

•  isothermal for 15 min

           Mass Spectrometer Conditions
Emission:    300 ya

Integration
Time:

Election
Energy:
250 msec/amu


70 E.V.
Operating
Condition:    Selected Ion Monitor
                                      71

-------
     TABLE A-2.   GC/MS CHARACTERISTICS OF STANDARDS FOR PCB ANALYSIS
o
Compound
Biphenyl
2-Monochlorobiphenyl
3, 3'Dichlorobiphenyl
2,4,5 Trichlorobiphenyl
2, 3, '5 Trichlorobiphenyl
2, 3 ',4 ',5 Tetrachlorobiphenyl
2,2,'4,4',6,6~ Hexachlorobiphenyl
2, 2 ',4, 5 ',5 Pentachlorobiphenyl
2, 2', 3, 4, 5, 5', 6-Heptachlorobiphenyl
Decachlorobiphenyl
dl 0 -Anthracene
dg -Tetrachlorobiphenyl
Retention
timeb
(min)
3.0
5.2
10.7
12.0
12.4
16.4
16.5
17.2
22.3
28.7
11.7
19.0
Parent
ion
(m/e)
154.1
188.1
222.1
256.0
256.0
292.0
359.9
325.9
395.8
497.7
188.2
298.1
Secondary
ion
(m/e)
153.1
152.1
224.1
258.0
258.0
290.0
361.9
323.9
393.9
499.7

296.1
- Detection Limit - 1.0 ng/pl
- GC/MS Instrumental Parameters - Table A-l.
                                    72

-------
               TABLE A-3.  GC/ECD PARAMETERS FOR PCB ANALYSIS
Instrument:
Column:
Temperatures:
Flow Rate:
Detection Limit:
Hewlett-Packard Model 5840A with a Ni63 electron capture
detector and a Model 7671A automatic liquid sampler
6 ft X 2mm (I.D.) 1.5% OV-17/1.95%
QF-1 on 100/120 mesh Chromosorb WHP
Column    175°C
Injector  270°C
Detector  300°C
50 cc/min argon methane
0.1 ng/yl Aroclor 1242
               TABLE A-4.   GC/ECD PARAMETERS FOR DCB ANALYSIS
  Instrument:

  Integrator:
  Column:
  Temperatures;
 Flow Rate:
 Retention Time:
 Detection Limit:
  Perkin-Elmer Model  3920 with  a  Ni63  electron
  capture  detector.
  Hewlett-Packard Model  3380S
  6  ft x 2 mm(I.D.)   3%  SE  30 on  100/120 mesh Gas Chrom Q
  Column     -   240°C
  Injector  -   270°C
  Detector  -   300°C
  50cc/min argon methane
  3.5 min
  O.lng per injection
                                      73

-------
       TABLE A-5.   TOTAL CHROMATOGRAPHABLE ORGANICS (TCO)  - OPERATING
                   PARAMETERS
        Parameter
                   Description
Instrument:

Recorder/Integrator:

Chart Speed:

Column:


Carrier Gas:

Oven Temperature:


Injector Temperature:

Detector Temperature:

Injection Volume:
Tracer 560 gas chromatograph with dual FID

Hewlett-Packard 3380S

1 cm/min

10 percent SP-2100 on Supelcoport 100/120
6 ft x 2 mm I.D. glass column

20 ml/min prepurified N2

50°C for 5 min, programmed 50°-250°C
at 10°/min, hold at 250°C for 2 min

275°C
300°C
2.0 yl
                                     74

-------
       TABLE A-6.   GC/MS  INSTRUMENTAL  PARAMETERS  FOR DIOXIN/DIBENZOFURAN
                   ANALYSIS
Column:

Injection:

Temperature
Program:
Emission:

Integration
Time:

Election
Energy:

Operating
Condition:
       GAS CHROMATOGRAPHIC CONDITIONS

SP2100 (50m) Wall Coated Open Tubular (WCOT): fused silica

Splitless (held for 48 sec)


•  isothermal for 2.0 min at 35°C

•  linear program from 35° - 135° at 25°C/min

•  linear program from 135° - 270° at 10°C/min

•  isothermal for 10 min

        MASS SPECTROMETER CONDITIONS

300 pa


250 msec/amu


70 E.V.


Selected Ion Monitor
                                     75

-------
                 TABLE A-7.  STACK PARAMETERS MEASURED BY GCA
Parameter measured
                                      Method
    in flue gas
CO in flue gas
02 in flue gas
THC
                     Continuous gas  analyzer infrared ind.  703-353
                     Continuous gas  analyzer Beckman 865  (IR)
                     Continuous gas  analyzer Beckman 741
                     Continuous gas  analyzer Beckman 108A
Flue gas velocity    Calibrated pitot tube/manometer
Flue gas temperature           Thermocouple
Flue gas pressure              Pitot tube/manometer
Grain loading (particulate)     Results of Method 5 tests
Percent moisture in flue gas   Moisture gain in Method 5  tests
                TABLE A-8.   STACK PARAMETERS MEASURED BY GM
   Air  Temp  In & Out:
   Gas  Temp  In & Out:
   Fuel Oil  Temperature:
   Windbox Pressure:
   Firebox Pressure:
   Air  Pressure In & Out:
   Atomizing Steam Pressure:
   Steam Flow Rate:
   Fuel Oil  Flow Rate:
   Smoke Density:

   Flame Temp:
   Fire Box  Temp:
                              Bailey A26 temperature transmitter
                              Bailey E101 temperature recorder
                              H.O. Trerice dial indicating
                              Bailey DA226W multipointer pressure  gauge
                              Bailey DA226W multipointer pressure  gauge
                              Bailey DA228W multipointer pressure  gauge
                              Ashcroft dial indicating
                              Bailey El01; a flow recorder & transmitter
                              Bailey AO & E101 flow recorder & transmitter
                              Bailey B101; UB5000A; UL5000A smoke  density
                                recorder, light detector & light source
                              Modline Q series infrared pyrometer
                              Type K thermocouple
                                       76

-------
     Table A-9 lists analyzer specifications for all flue gas monitoring
equipment.

     Oxygen concentrations were determined using a Beckman Model 741 Para-
magnetic 02 Analyzer with a measuring range of 0 to 10 percent 02 full scale.
The Analyzer was calibrated at 0 percent 02 with ultrapure nitrogen and with
8.03 percent 02 certified calibration gas before and after each test.

     Carbon dioxide concentrations were determined using an Infrared Industries
Model 702 NDIR Carbon Dioxide Analyzer with a measuring range of 0-30 percent
C02 full scale.  The Analyzer was calibrated at 0 percent C02 with ultrapure
nitrogen and with 7.99 percent C02 certified calibration gas before and after
each test period.

     Carbon monoxide concentrations were determined using Beckman Model 65 NDIR
CO Analyzer with a measuring range of 0-50 ppm CO full scale and was calibrated
at 0 ppm CO with ultrapure nitrogen and with 39.9 ppm certified calibration gas
before and after each test period.

     Total hydrocarbon concentrations were determined using a Beckman Model
108-A Flame lonization Hydrocarbon Analyzer with a measuring range of 0-50 ppm
full scale.  It was calibrated at 0 ppm with ultrapure nitrogen and with 40.1
ppm methane certified calibration gas before and after each test period.

     Four  strip chart recorders were used to record all monitoring data.  The
data were  later corrected for calibration drift, if any, and reduced to 15-
minute averages.  Maximum and minimum values for each test burn period were
also determined.  These are based on a statistical treatment of actual measure-
ments, not the 15-minute averages.

     Other physical parameters were also measured in this testing program.
These parameters included the flue gas velocity, static pressure and tempera-
ture.  These measurements were made before each run to determine proper nozzle
size and sampling rate.  Additional physical parameters were measured during
the course of  each test.  A complete list of all parameters monitored by GCA
at the stack is included in Appendix A, Table A-7; and those monitored  by GM
is included in Appendix A, Table A-8.

                      COMBUSTION TEMPERATURE MONITORING

Instrumentation was installed on No. 3 boiler, for the verification-burn,
to measure and record flame temperature and furnace gas temperature.  Flame
temperature was measured with an optical pyrometer mounted on the burner
front of the boiler.  Furance gas temperature was measured with a grounded
junction thermocouple inserted through the rear wall of the boiler.

FLAME TEMPERATURE

Flame temperature was measured with a Modline "Q" Series Infrared Pyrometer,
Type Q45F05, Serial 100796.  This instrument has a range of  1500°F to 4500°F
and incorporates analog readout.  The unit was factory calibrated.  Emissivity


                                     77

-------
                       TABLE A-9.  ANALYZER SPECIFICATIONS OF CONTINUOUS MONITORS
oo

Operating
sensitivity
ranges


Operating
temperature
ranges

Linearity

Accuracy


Stability
(precision)






Infrared
industries
Model 702/703
C02 analyzer
0-10% C02, FS

0-30% C02, FS


32°F - 120°F



± 1% of
full scale
± 1% of
full scale
a
± 1% of
full scale
in 24 hours

a



Beckman
Model 864
C02 analyzer
0-10% C02, FS

0-50% C02, FS

0-100% C02, FS
32°F - 120°F



± 1% FS

± 1% of
full scale
a
± 1% of
full scale
in 24 hours

a



Beckman
Model 10 8- A
hydrocarbon
analyzer
Adjustable
from 1 ppm
CH4 full scale
to 2% CHit
Full scale
32°F - 110°F





± 1% of
full scale

± 1% of
full scale
in 24 hours





Beckman Model 865
CO analyzer
0-50 ppm CO, FS

0-100 ppm CO, FS

0-1000 ppm CO, FS
30°F - 120°F



± 1% FS

± 1% of
full scale
a
± 1% of
full scale
in 24 hours

a



Beckman
Model 741/742
02 analyzer
0-1% 02, FS
0-5% 02, FS
1-10% 02, FS
0-25% 02, FS

32°F - 110°F
for sensor
-20°F - 122°F
for analyzer
± 1% of
full scale
± 1% of
full scale

± 0.25% of full
scale at constant
temperature ±1%
of full scale

for temperature
deviations of
-20°F to 122°F
       Performance specifications based on ambient temperature shifts
       at a maximum rate of 20 Fahrenheit degrees per hour.
of less than 20 Fahrenheit degrees

-------
was set to "1.0" for the verification-burn.  The pyrometer will be mounted
to the burner front of the boiler in such a manner that line of sight access
to the burner flame is realized.

Flame temperature was continuously recorded on an Esterline Angus Miniservo
Strip Chart Recorder, Model MS401BB, Serial S-22243-1A.  This recorder has a
range of 0-100 mv. (0-100 percent).  The unit was calibrated on March 26,
1980 by General Electric.

The optical pyrometer was installed to monitor the flame which exhibits the
lower observed temperature (of the two burner flames).   A normal operating
temperature range was established for this burner using the optical pyrometer.
Flame temperature was manually read and recorded once each 15 minutes during
the verification-burn.  Input of PCB contaminated oil to the boiler was to
be stopped immediately if the observed flame temperature dropped 100°F or
more below the established normal operating temperature.

FURNACE GAS TEMPERATURE

Furnace (fire box) gas temperature was measured with a grounded junction,
Inconel sheath, Type "K" Megopack thermocouple (two foot length).  The thermo-
couple was inserted through an observation port in the rear wall of the
boiler.  The thermocouple junction was approximately 1 foot from the inside
face of the furnace rear wall.

Furnace gas temperature was continuously recorded on a Honeywell strip chart
recorder, Model 143C10PS-24W7-2, Serial' 552.  The recorder has a range of
0-2400°F.

Furnace gas temperature was manually read and recorded once each 15 minutes
during the verification-burn..
                                     79

-------
                      APPENDIX B

WORK PLAN FOR EVALUATION OF PCB DESTRUCTION EFFICIENCY
   OF NO. 3 INDUSTRIAL BOILER AT CHEVROLET-BAY CITY
                      April,  1980
                      Prepared by
                   A.P.  Garwick,  Jr.
                 Chevrolet - Bay City
               Chevrolet Motor Division
              General Motors Corporation

                           80

-------
                         Contents
                                                         PAGE

INTRODUCTION 	   83

   Objective	   83
   Description of the Verification Burn	   83
   Description of Facility	   84
   Work Plan Organization	   87

OPERATING PROCEDURES	   90

   Organizational Structure	   90

     Plant Engineer	   90
     Chief Powerhouse Engineer  	   90
     Senior Environmental Engineer  	   92
     Sampling Team Leader	   92
     Monitoring Team Leader	   92
     Powerhouse Boiler Operator	   93
     Sampling Team	   93
     Industrial Hygienist	   93
     Governmental Agency Representative	   94

   Scheduling	   94

     Day One	   94
     Day Two	   95
     Day Three	   96
     Day Four	   96
     Day Five	   96
     Day Six	   97

   Instrumentation  Provided  By  Chevrolet 	   97

     Existing Boiler  Controls	   97
     Flame Temperature	   98
     Furnace Gas Temperature 	   99
     Reclaim Oil Flow Rate	100
   Security
                                                         100
     Facilities For Visitors  	  101
     Instrument Protection  	  101

   Emergency Procedures	1°2

EMPLOYE HEALTH EFFECTS  PROTOCOL	104

WORKPLACE MONITORING  	  108
                              81

-------
                                                       PAGE

   Employe Breathing Zone Sampling	108
   Boiler Residue Sampling  	  109

PCB RELATED ACTIVITIES  	   Ill

   Spill Prevention and Control	   Ill
   Waste Handling and Disposal	•	   112
   Employe Training 	   112
                            82

-------
                       INTRODUCTION
Objective
EPA has promulgated a final rule to implement provisions of the
Toxic Substances Control Act prohibiting the manufacture, pro-
cessing,  distribution in commerce,  and use of PCB's.  Subpart B
of 40 CFR Part 761 relates  to disposal of PCB's.   The final rule
permits the disposal of liquid wastes with 50 to  500  ppm PCB
in a case-by-case approval  basis in high efficiency boilers.

In order to gather data relative to PCB destruction efficiency
of industrial boilers,  the  United States Environmental Protec-
tion Agency and General Motors Corporation have developed a
sampling and analysis protocol for a verification burn to
evaluate the PCB destruction efficiency of the No. 3  industrial
boiler at Chevrolet - Bay City in Bay City, Michigan.

Description Of The Verification Burn
The verification burn will  require isokinetic sampling for
PCB's in the combustion flue gas during three complete tests
of the boiler at normal load operation and burning No. 6 fuel
oil containing approximately 10 percent reclaim oil and 50 ppm
PCB (500 ppm PCB in the reclaim oil).  The No. 6  fuel oil/
reclaim oil mixture will be fed to the boiler after normal burn
temperature has been established.  The sampling will  require up
to 6 hours for each run, based on an assumed destruction effi-
ciency of 99.9 percent.
                             83

-------
In order to provide background data on emissions from the boiler,



a preliminary test will be conducted with No. 6 fuel oil only,



one day prior to the initiation of the three day PCB verifica-



tion burn.   This test will be for the same duration and at the



same conditions as those to be used for the PCB sampling.  This



test will provide background information on the test conditions



and also provide emissions data for the boiler as operated at



normal load and firing No. 6 fuel oil.








Description Of The Facility



Chevrolet - Bay City is a manufacturer of component parts for



General Motors Corporation passenger cars and trucks.  The



manufacturing facility covers in excess of 1,000,000 square



feet.  Approximately 4,000 employes work at this location.  In



addition to the manufacturing and administrative facilities,



the plant complex includes an industrial waste water treatment



facility and a powerhouse  (See Plot Plan, Figure 1).








The industrial waste water treatment facility is the source of



the reclaim oil for the verification burn.   The reclaim oil has



been separated from the process waters collected within the



manufacturing facility.  The PCB's present in the reclaim oil



are residuals which have persisted in purged hydraulic systems



within the  manufacturing facility.  Prior to being burned in



the No. 3 boiler, the reclaim oil will be processed through an



on-site oil reclamation process to remove dirt and water.  The



reclamation process includes 20 micron diameter filtration





                             84

-------
00
Ui
              E f-4 u i-f> 7-f-e
             . SCffLP- { C7UflB.DHOUSf-
      Z Z. tyUfl^jy HPi'f P-
                                       Figure 1.   Chevrolet-Bay  City Plot Plan.

-------
followed by a centrifuge to remove water and more solids.



The reclaim oil for the verification burn will contain less



than 0.17o solids and less than 0.17o water.







The powerhouse at Chevrolet - Bay City provides steam, com-



pressed air, and surface cooling water for the plant complex.



Steam is generated in one or more of three oil-fired boilers.



The steam requirement for the plant complex is dependent on



heating load requirements.   One or two boilers are required



as the demand changes.   A boiler normally remains in service



for two months before being secured for inspection, cleaning,



and repair.   Each boiler is further inspected on an annual



basis by the Michigan Department of Labor or its designated



representative for license renewal.







The No. 3 boiler is a Wickes Type "A" 65-4K-7 package boiler



installed in 1965, and having a stearn generating capacity of



60,000 pounds per hour at 8 gallons per minute of No. 6 fuel



oil.   The No. 3 boiler is essentially identical to the other



two boilers  at this location except for the additional equip-



ment added for the verification burn.







The boilers  are normally fired on No. 6 fuel oil from two



200,000 gallon storage tanks.   Each boiler may also be fired



on No.  2 fuel oil from a 15,000 gallon back up system.  The



boilers are  also equipped to burn natural gas, although no





                           86

-------
contractual supply is currently available.








For the verification burn, No. 3 boiler will be fired on a



mixture of approximately 90 percent No. 6 fuel oil and 10



percent reclaim oil from a 10,000 gallon reclaim oil tank



at the powerhouse.  The No. 6, No. 2, and reclaim oil systems



are individually piped to the burner front of No. 3 boiler in



a manner such that the fuel supply can be changed readily.



(See Figure 2).








Work Plan Organization



The remainder of this work plan will describe the operating



procedures, employe assignments, health protection, and PCB



handling and containment procedures for the verification burn,



This work plan complements GCA's "TEST PLAN FOR EVALUATION OF



PCB DESTRUCTION  EFFICIENCY IN INDUSTRIAL BOILERS" of March



1980.








The Operating Procedures section describes the organizational



structure,  day-by-day scheduling,  instrumentation to be



provided by Chevrolet,  security provisions,  and emergency pro-



visions for the  verification burn.








The Employe Health Effects Protocol section describes the



medical examination procedures for determining general well-



being of the powerhouse operators  prior to the verification



burn and to identify any possible  health concerns associated





                             87

-------
00
oo
                                                                                          200,000  Gallori
                                                                                          No.  6  Fuel  Oil
                                                                                           Storage Tank
                                                                                             (Heated)
No.  6 Fuel Oil From Tank Truck

              Reclaim Oil From Waste Treatment
                               10,000 Gallon
                                Reclaim Oil
                               Storage Tank
                                 (Heated)
                                                     200,000 Gallon
                                                     No.  6 Fuel Oil
                                                      Storage Tank
                                                        (Heated)
                                                          Burner 1
                                                             M
                                                       No.  6 Fuel Oil
                                                       Transfer Pump
                                                                         No.  3 Boiler
                                                                      Chev. Tag S  24567
                                                                         60,000 LB/HR
                         Circulating
                            Heater
              ,-(}-J--Flow Control Valve

                    -Flow Meter
                                                                                            No.  6 Fuel Oil
                                                                                                Heater
                  Circulating Pump
                  No.  6 Fuel Oil/
                    Reclaim Oil	:
                  Transfer Pump
                                       No.  6  Fuel Oil/
                                      Reclaim Oil Heater
                  No.  2 Fuel Oil Transfer  Pump
                                    Reclaim Oil  System  Contained Within Dashed Lines

                                            Figure 2.  Fuel Oil Piping.

-------
with previous burning of PCB contaminated oil.







The Workplace Monitoring section describes the  employe breath-



ing zone testing program and boiler residue analysis program



which will be conducted in conjunction with the verification



burn.







The PCB Related Activities section describes the handling,



clean up,  and disposal provisions for the verification burn.
                              89

-------
                  OPERATING PROCEDURES








Organizational Structure




The verification burn will be conducted by General Motors



Corporation, in close cooperation with the United States En-



vironmental Protection Agency (EPA),  the Michigan Department



of Natural Resources (MDNR),  and the  GCA Corporation, EPA's



contractor.  General Motors will provide and operate the



boiler, monitor and record boiler operating parameters, monitor



the workplace,  prevent and control spills, handle and dispose



of wastes, and provide security.  GCA will provide and operate



the flue gas and ambient air sampling equipment,  and perform



the analytical work necessary to evaluate these samples.  The



organizational structure for the project is shown in Figure 3,



Specific project assignments are as follows:








Plant Engineer (Chevrolet - Bay City  (CBC)).  The Plant Engi-



neer has complete responsibility for  and control  of the veri-



fication burn.   He will direct all aspects of the burn through



the Chief Powerhouse Engineer and the Senior Environmental



Engineer.  The Plant Engineer will be the verification burn



spokesman for Chevrolet - Bay City.








Chief Powerhouse Engineer (CBC).  The Chief Powerhouse Engineer



has responsibility for the safe operation of the  boiler and



associated equipment.   He will direct the Powerhouse Operators,



who will adjust and maintain the boiler and monitor boiler





                             90

-------
                                  ENGINEER
                                                         ENVIRONMENTAL
POWERHOUSE
                                                            ENGINEER
 ENGINEER
                                   AGENCY
                                MONITORING
                                  SAMPLING
                                  ANALYSIS
SECURITY
               BOILER
              OPERATION
                             BOILER
                           MONITORING
                                  WORKPLACE
                                 MONITORING
          Figure 3.  Organizational Structure.

-------
operating parameters.   The Chief Powerhouse Engineer will



initiate, interrupt, and stop the verification burns, as



required.  All operational questions and comments will be



directed to him.








Senior Environmental Engineer (CBC). The Senior Environmental



Engineer has responsibility for all  sampling,  monitoring,



and coordinating  activities,  other than safety and security.



He will direct the General Motors monitoring activities and



serve as the plant representative for the GCA sampling and



analysis team. He will also  be the  plant representative for



the EPA and MDNR.








Sampling Team Leader (GCA).   The Sampling Team Leader has



responsibility for all flue gas and  ambient air sampling and



analysis.  He will be  responsible to inform the Chief Power-



house Engineer when the verification burn may begin,  should



be interrupted, and may be stopped,  with respect to the sampling



program.  He will notify the  Chief Powerhouse Engineer immediate-



ly if any monitored parameter exceeds its allowable range.








Monitoring Team Leader (CBC).   The Monitoring Team Leader has



responsibility for all boiler parameter monitoring and record-



ing.  He will notify the Chief Powerhouse Engineer immediately



if any monitored  parameter exceeds its allowable range.  The



Monitoring Team Leader will be assisted by a Powerhouse Boiler



Operator who will observe all readings.
                              92

-------
Powerhouse Boiler Operators.  The Powerhouse Boiler Operators



will perform their normal job assignments under the direction



of the Chief Powerhouse Engineer.  The Operator who is firing



will adjust and maintain the boilers and monitor boiler oper-



ating parameters.  He will start and stop flow of the reclaim



oil to No. 3 boiler upon the command of the Chief Powerhouse



Engineer.  He will assist the Monitoring Team Leader by ob-



serving all readings that are logged, and will notify the



Chief Powerhouse Engineer immediately if any monitored parameter



exceeds its allowable range, or if any other operational problems



arise.  This Operator will be relieved and supplemented as



required by other Powerhouse Boiler Operators.








Sampling Team (GCA).   The GCA Sampling Team will operate the



flue gas and ambient air sampling equipment and analyze samples,



under the direction of the Sampling Team Leader, as described



in the GCA "TEST PLAN FOR EVALUATION OF PCB DESTRUCTION EFFICIENCY



IN INDUSTRIAL BOILERS" dated March 1980.   The Sampling Team



will notify the Sampling Team Leader immediately if any monitored



parameter exceeds its allowable range.   The Sampling Team will



notify the Chief Powerhouse Engineer directly in the absence of



the Sampling Team Leader.








Industrial Kygienist (GM).   The Industrial Hygienist has respon-



sibility for employe breathing zone sampling within the power-



house.  He will take breathing zone air samples before,  during



and after the burning of the PCB contaminated oil.   He will be
                            93

-------
responsible for the analysis of the samples.  The Industrial



Kygienist will also observe the workplace during the verification



burn and report any concerns to the Chief Powerhouse Engineer.



The Industrial Hygienist will coordinate his activities through



the Senior Environmental Engineer.








Governmental Agency Representatives (EPA, MDNR).  Representatives



of the United States Environmental Protection Agency and Michigan



Department of Natural Resources will observe the verification



burn and monitor procedures and activity as directed by their



respective agency management.   Agency Representatives will coordi-



nate their activity through the Senior Environmental Engineer.








Scheduling



The verification burn will include equipment set-up and prepara-



tion,   a five-hour background burn, three five-hour burns with



PCB contaminated oil,  clean-up,  securing of the boiler, and



removal of the test equipment.   The entire verification burn



will require approximately one week as detailed below,  and



in the GCA "TEST PLAN FOR EVALUATION OF PCB DESTRUCTION



EFFICIENCY IN INDUSTRIAL BOILERS".






Day One Day One will begin with the arrival of the GCA Sampling



Team and equipment.  Chevrolet Plant Engineering  Personnel will



orient and assist the GCA Sampling Team by providing power and



utility connections, laboratory and parking provisions, and



security for equipment.   Day One will precede the calendar week
                            94

-------
of the background and PCB contaminated oil burns.








Day Two  The calendar week of the burns will begin with the



background burn with No. 6 fuel oil, only, on Day Two.  At



approximately 7 A.M., the GCA Sampling Team will position and



start the ambient air monitors and ready the flue gas monitoring



equipment.








At approximately 9:30 A.M.,  the No. 3 boiler, which will already



be at temperature and operating on No.  6 fuel oil, will be set



to 4 gpin under manual control.  The GCA Sampling Team Leader will



notify the Chief Powerhouse  Engineer when the flue gas monitoring



equipment is ready.   The Chief Powerhouse Engineer will verify



that the boiler is stable and within allowable operating parameter



limits.   At approximately 10:00 A.M., the Chief Powerhouse Engineer



will instruct the Sampling Team Leader and the Monitoring Team



Leader to begin the  test.








The test will continue for five hours,  but possibly longer,  until



approximately 4:00 P.M.   The Sampling Crew Leader will notify



the Chief Powerhouse Engineer when the  flue gas sampling is  com-



plete.   Monitoring and recording will cease,  except for the  ambient



air monitors and normal  monitoring for  the boiler.  The boiler



firing rate will be  adjusted consistant with the steam demand



from the plant complex.   The special equipment will be secured



for the  day.  At the completion of the  ambient air monitoring



for the  day, the GCA Test Crew will secure those samplers.
                             95

-------
Day Three.   Day Three will proceed as did Day Two, except that



the PCB contaminated oil will be substituted for No. 6 fuel oil.



The substitution will be accomplished after the Sampling Crew



Leader has  indicated that the test may begin.  The boiler will



already have been stabilized on No. 6 fuel oil at 4 gpm.   Upon



the instruction of the Chief Powerhouse Engineer, the Powerhouse



Boiler Operator will start the reclaim oil supply pump and verify



system pressure.  He will then open the manual stop valve in the



reclaim supply line, and close the manual stop valve in the No. 6



fuel oil supply line.  The reclaim oil flow rate and temperature



will be adjusted as required.  When the boiler is stable, the



Chief Powerhouse Engineer will instruct the Sampling Team Leader



and the Monitoring Team Leader to begin the test.








Upon completion of the test,  at approximately 4:00 P.M.,  the



boiler will be returned to operation on No. 6 fuel oil by revers-



ing the valve and pump sequence above.  The boiler firing rate



will be adjusted as required, and the special equipment secured



as in Day Two.








Day Four.   Day Four will be a repeat of Day Three.








Day Five.  Day Five will be a repeat of Day Three, until the com-



pletion of  the sampling program.   After the flow of reclaim oil



to the boiler has been stopped and the boiler has been stabilized



on No.  6 fuel oil, steam load will be shifted from No. 3  boiler



and it will be secured and allowed to cool for inspection.   The



special equipment will be secured as in Day Two.




                               96

-------
Day Six.  The special equipment will be removed from the boiler



during Day Six.   Chevrolet personnel will open the boiler and



sample residue from the interior of the firebox and from the



hot side of the  air preheater wheel.  Protective clothing will



be worn by the personnel collecting the samples.  The boiler will



be resealed until the results of the residue analysis are avail-




able.








Instrumentation  Provided By Chevrolet



Existing controls,  typical of this type of boiler, will be used



to monitor and record the operation of No. 3 boiler during the



verification burn.   Additional instrumentation will be installed



by GCA and Chevrolet for the verification burn only.  The equip-



ment provided by GCA is detailed in their "TEST PLAN FOR VERIFI-



CATION OF PCB DESTRUCTION EFFICIENCY IN INDUSTRIAL BOILERS."  The



equipment provided by Chevrolet will measure and record flame



and furnace gas  temperature and oil flow rate as described below.



Parameters monitored by Chevrolet will be manually read and



recorded once each 15 minutes during the verification burn.







Existing Boiler  Controls.  The existing boiler controls which



will be used to  monitor the operation of the boiler during the



verification burn are listed below:



Parameter Monitored       Instrument Range    Normal Parameter Range



Fuel oil temperature      50-300°F.           Ambient-200°F.



Fuel oil flow rate         0-8   gpm          1-8 gpm
                               97

-------
Parameter Monitored
                          Instrument Range
Air pressure in
Air pressure out
                          0-10" w.c.
                          0-10" w.c.
Atomizing steam pressure  0-160 psig
Steam flow rate           0-70  MLB/Hr.
Smoke density (in stack)  1-100%
Windbox pressure
Firebox pressure
  For Air Preheater:
Air temperature in
Air temperature out
Gas temperature in
Gas temperature out
0-600°F.
0-600°F.
0-600°F.
0-600°F.
0-15" w.c.
0-15" w.c.
 Normal Parameter Range
 15-110 psig
 10-70  MLB/Hr.
  5-20%
0.5-7.5"w.c.
  0-5.5"w.c.

 Ambient
 250-350°F.
 350-580°F.
 200-300°F.
 0.5-10.5" w.c.
 0.5- 8.0" w.c.
The existing boiler controls for No.  3 boiler were checked and
calibrated by the Baily Meter Company on April 15, 1980.  The
opacity meter for measuring smoke density in the exhaust stack
of No. 3 boiler will be cleaned and calibrated each night of
the verification burn,  after the day's test has been completed.

Flame Temperature  will be measured with a Modline "Q" Series
Infared Pyrometer,  Type Q45F05,  Serial 100796.  This instrument
has a range of 1,500°F. to 4,500 F and incorporates analog read-
out.  The unit has  been factory calibrated.  Emissivity will be
set to "1.0" for the verification burn.   The pyrometer will be
mounted to the burner front of the boiler in such a manner that
line of sight access to the burner flame is realized.
                             98

-------
Flame temperature will be continuously recorded on an Esterline


Angus Miniserve Strip Chart Recorder, Model MS401BB, Serial


S-22243-1A.  This recorder has a range of 0-100 mv. (0-100%).


The unit was calibrated on March 26, 1980 by General Electric.





The optical pyrometer will be installed to monitor the flame


which exhibits the lower observed temperature (of the two burner


flames).   A normal operating temperature range will be established


for this burner using the optical pyrometer.  Flame temperature


will be manually read and recorded once each 15 minutes during


the verification burn.  Input of PCB contaminated oil to the


boiler will be stopped immediately if the observed flame temper-

               o
ature drops 100 F. or more below the established normal operating


temperature.





Furnace Gas Temperature.Furnace (firebox) gas temperature will be


measured with a grounded junction, Inconel sheath, Type "K" Megopack


thermocouple (two foot length).   The thermocouple will be inserted


through an observation port in the rear wall of the boiler.  The


thermocouple junction will be approximately one foot from the in-


side face of the furnace rear wall.





Furnace gas temperature will be continuously recorded on  a Honey-


well strip chart recorder, Model 143C10PS-24W7-2,  Serial 552. The


recorder  has  a range of 0-2,400°F.
                              99

-------
Furnace gas temperature will be manually read and recorded once



each 15 minutes during the verification burn.








Reclaim Oil Flow Rate.  Oil flow to the boiler will be measured



and controlled by the existing boiler controls.  The flow rate to



the boiler during the verification burn will be maintained at



4 gpm.  For the verification burn,  oil flow to the boiler will



also be measured and recorded by a Clampitron portable ultra-



sonic clamp-on flow meter, Series 240, Model 241MP-12-A-T, Serial



932CA.  This instrument has a velocity range of 0 to 30 feet



per second.  The unit has been factory calibrated, and requires



zero flow adjustment only prior to the verification burn.  The



unit provides instantaneous digital readout in gallons per minute



and accumulates total gallons registered.   This instrument will



not be used to control oil flow to No. 3 boiler, but will be used



only to collect additional data for the verification burn.








Security



The PCS contaminated reclaim oil is stored in tanks within diked



areas on Chevrolet property.  These areas are restricted to



authorized personnel and are included in scheduled security patrols.



Access to Chevrolet property is through secured entrances.  Additior



al security requirements for the verification burn concern facil-



ities for visitors and instrument protection.
                               100

-------
Facilities For Visitors  Visitors may gain access to the Chevrolet



plant through the lobby of the administration building between



the hours of 6:30 A.M. and 5:00 P.M., and through the west employe



entrance at other times.  Visitors should identify themselves



and their business to the employe on duty and sign in.  Each



visitor will be issued safety glasses and an identification badge



before entering the plant complex.  Chevrolet personnel will escort



visitors to and from the Powerhouse.  Visitors requiring early



entry or late exit should make prior arrangements through the



Plant Engineering Department.  Messages for visitors should also



be directed through the Plant Engineering Department.








Instrument Protection  Instruments for the verification burn will



be located within the Powerhouse,  on the roof of the Powerhouse,



at three remote locations on Chevrolet property, and at one location



in a residential area of Bay City.








The Powerhouse is a restricted area which is kept locked at all



times.  It is manned 24 hours per day by Powerhouse Boiler Op-



erators, and is included in scheduled security patrol checks.



Only authorized personnel will be admitted to the Powerhouse during



the duration of the verification burn.   The instruments within



and on the roof of the Powerhouse will be included in the security



protection of the Powerhouse.
                               101

-------
The ambient air monitors which will be located at the remote



locations on Chevrolet property will be within the secured area



of the plant fence.  Access to this area will be through secured



entrances.  These monitors will be left in the field each night



of the test and will only be relocated as required by changes in



weather conditions.  Dependent on placement on the property, some



or all of the three monitors may be within observation limits of



Chevrolet's closed-circuit television surveillance.   The monitors



will also be included in the scheduled security patrol checks.








The single ambient air monitor which will be located off Chevrolet



property will be truck mounted.   This monitor will be located each



morning of the test and will be returned to the plant each night,



after completion of the sampling program.  The off-site monitor



will be attended by Chevrolet personnel at all times that it is off



Chevrolet property.








Emergency Procedures








The Chief Powerhouse Engineer will have complete responsibility



for the safe operation of the boiler and associated  equipment



during the verification burn.  He will be assisted in carrying



out his responsibility by the Powerhouse Boiler Operators,  the



GCA Sampling Team,  the Chevrolet Monitoring Team,  and the



personnel and resources of Chevrolet - Bay City.   The Chief



Powerhouse Engineer will initiate,  interrupt,  and stop the veri-



fication burn,  as required by the sampling program and safety



consideration.   Safety will be of the greatest concern at all



                             102

-------
times.  Established Chevrolet procedures for fire prevention,



spill prevention and control, personal injury,  and natural dis-



aster will remain in effect for the verification burn.
                               103

-------
               EMPLOYE HEALTH EFFECTS PROTOCOL








Chevrolet - Bay City burned about 100,000 gallons of waste oil



containing low concentrations of PCB's in our boilers between



1974 and 1977, with estimated destruction efficiencies of 99.970.



Because of concern by plant Powerhouse Operators that they were



possibly exposed to excess levels of PCB's during past years,



comprehensive medical examinations and blood tests were conducted



on these employes.  Their concerns resolved themselves into two



questions which could be paraphrased as:








      1.  "What damage,  if any have I sustained?"



      2.  "What residual PCB is there in  my body, and could



           it push me over a critical threshold if I absorb



           more?"








      With the endorsement of the Divisional and Corporation



      Medical Directors  and with extensive consultation with



      experts in many specialized fields  the plant set out to



      answer these two questions as directly and completely



      as possible.








      With each employe, the plant medical department went into



      a searching medical history, detailed physical examination,



      special testing with electrocardiograms and lung capacity



      tests.  However, it ultimately came down to blood testing



      for the answers.  The ansx^er to "What damage, if any have
                               104

-------
I sustained?" was none.








All of the operators showed normal results in the 28



analytic tests of the chemistry of the blood, blood cell



examinations, general cellular well-being, and the Medical



Director's special concern, liver function.  The question:



"What residual PCB is there in my body, and could it push



me over a critical threshold if I absorb more?", requires



some background information.  PCB testing is done only



in a limited number of laboratories in the United States.



One of these is that of the State of Michigan Department



of Health which has had extensive experience testing for



both PCB and PBB.  The Health Department was not only co-



operative in its execution of test procedures but offered



advice and the benefit of their accumulated experience.



Another experienced laboratory is the Raltech Laboratory



of Madison,  Wisconsin which did the PBB testing for Dr.



Selikoff when he surveyed the farm families involved in



the Michigan PBB contamination incident.








Blood samples were split and sent, half of each, to these



two laboratories both to check accuracy and assure back-up



in case specimens met disaster in the mails.  No such dis-



aster occurred, and after consulting with the directors of



each laboratory, good agreement was found in the test results



The widespread low level contamination of soil,  water,  fish,



and wildfowl requires sampling of an average population for
                        105

-------
comparison (There is no normal amount of PCB in human



blood).   The State of Michigan had previously conducted



testing in Michigan and was able to provide data from



an average population group with no industrial exposure.



Their reference group was divided in two parts on the



basis of average weekly consumption of lake fish, (In ex-



cess of 8 oz. or less than 8 oz.)-  Michigan residents,



whose daily intake of lake fish exceeds 8 oz. per week,



ranged from 25 to 366 parts per billion.  The range for



Michigan residents who consume less than 8 oz. per week



is 7 to 42 parts per billion.   Powerhouse employes ranged



from 7 to 36 parts per billion.







The average level for plant boiler operators was 18 parts



per billion.   For non-fish consuming Michigan residents



it is 20 parts per billion.  For those exceeding 8 oz.



of Great Lakes area fish per week it is 73 parts per



billion.







Each Powerhouse employe was given a copy of the results



of all of his testing accompanied by reference normals



where applicable.  This tabulation was then reviewed in



a personal summary with the Medical Director.







At this time, Chevrolet was able to provide the answer to



question number two above:
                        106

-------
The residual PCB in the blood streams of plant operators



is equal to or lower than that of comparison groups through-



out the state.  Furthermore, the proposed test-burn is



designed so that under "worst-case" conditions no signifi-



cant additional burden of PCB would be introduced into the



bodily systems of anyone.
                         107

-------
              WORKPLACE MONITORING








In order to determine the extent to which the Powerhouse



Boiler Operators are exposed to PCB by burning the con-



taminated oil, the General Motors Industrial Hygiene



Department will conduct employe breathing zone tests



during the verification burn.   Samples of the residue from



within the boiler will also be taken and analyzed to de-



termine the extent to which the Powerhouse Boiler Op-



erators could be exposed to PCB during cleaning of the



boiler.








Employe Breathing Zone Tests



Continuous samples will be taken from the breathing zone



of each Powerhouse Boiler Operator during the entire time



that PCB contaminated oil is being burned, and during



the corresponding period of the background burn.  In



addition to the samplers on each Operator, four stationary



samplers will also be located in the Powerhouse.  Two of



the stationary samplers will be located on the first floor



of the Powerhouse in the immediate area of the burner front



of No. 3 Boiler.  The other two stationary units will be



located on the third floor of the Powerhouse, in the im-



mediate area of the exhaust duct expansion joint.








All the samplers will utilize solid absorbent tubes con-
                         108

-------
 taining approximately 350 rag. each of Florisil.  The Operator



 breathing zone samples will be drawn with SKC Model 222-3S



 Personal Sampling Pumps, operating at approximately 150 cc/min.



 Two 3 hour samples will be taken for each operator during each



 day (the samplers will be transferred to the Second Shift Op-



 erators at shift change).   The PCB detection limit for the



 personal samples will be on the order of 50 ug/m3 or less.








 The two stationary locations will each include one SKC sampling



 pump operating as described above, and one MSA Model "G" Pump



 operating at approximately 350 cc/min. for approximately six



 hours.  The six hour tests with the MSA pumps will provide a



 PCB detection limit on the order of 10 ug/^ or less.   The



 sampling and analytical techniques which will be used for the



 employe breathing zone tests follow the recommendations of the



 NIOSH Manual of Analytical Methods, Method P & CAM-253 for



 PCB's in Air.








 Boiler Residue Sampling



 Samples of the residue from the interior of the boiler firebox



 and from the hot side of the air preheater wheel will  be taken



 after the boiler has been  secured,  cooled,  and opened.   The



 samples will be analyzed by Chevrolet Laboratory in Detroit to



 determine the extent to which Powerhouse Boiler Operators could



be exposed to PCB during cleaning of the boiler.
                             109

-------
Disposable protective clothing and full-face,  positive pressure,



air supply masks will be worn by the personnel collecting the



samples within the boiler.   After the samples  have been collected,



the boiler will be resealed until the results  of the analysis



are reviewed.   Based on the residue analysis results,  appropriate



protective clothing will be provided for  the Powerhouse Boiler



Operators who  will clean the boiler.
                            110

-------
                 PCB RELATED ACTIVITIES
Special handling and cleanup procedures will be used during



the verification burn, due to the presence of the PCB con-



tamination in the reclaim oil.  The reclaim oil will have been



spiked to near,  but less than 500 ppm PCB and then will have



been mixed with No. 6 fuel oil so that the reclaim oil accounts



for 10 percent or less of the total mixture by volume.  The



concentration of the contaminated oil to the boiler will be



approximately 50 ppm PCB.  The oil will be considered to be at



least 50 ppm PCB for safety and handling procedures.








Spill Prevention and Control



The prevention of spills of oil,  fuel, and other hazardous



materials is far superior to clean-up.  Operation methods and



procedures have  been established to prevent the occurrence of



polluting spills.  However,  accidental discharges could occur



due to operator  error or equipment failure while transferring,



storing,  distributing, or transporting these materials.   These



accidental discharges could enter the waterways directly as



surface run-off  or indirectly through either sanitary or



storm sewer systems.








Chevrolet - Bay  City has established procedures to prevent



polluting discharges from storage, use,  distribution,  and



bulk receiving areas and to  provide procedures  for prompt
                            111

-------
clean-up and containment of accidental discharges.   These



procedures are contained within the Chevrolet - Bay City



Spill Prevention Control and Countermcasuro Plan (SPCC)/



Pollution Incident Prevention Plan (PIPP),  which was last



reviewed and revised on November 10,  1979.   Copies  of the



SPCC/PIPP are on file in the Powerhouse,  Plant Engineering



Department, Maintenance Department, Metallurgical Laboratory,



and at Plant Security Headquarters.








Waste Handling and Disposal



PCB contaminated materials will be collected and deposited



in scalable 55 gallon containers which will be located in



the Powerhouse Near No. 3 boiler.  Contaminated materials



will include oil drippings, absorbents, rags, gloves, dis-



posable clothing, boiler residue, etc.  The containers will



be identified as containing PCB, and will be serialized and



logged.  Full containers will be sealed,checked, and trans-



ported directly to the secured PCB storage area to await



proper disposal.  All PCB materials will be disposed of in



accordance with the applicable provisions of the law.








Employe Training



The Powerhouse Boiler Operators at Chevrolet - Bay City are



all licensed by the City of Saginaw, Michigan.  Each is ex-



perienced with the operation of the Powerhouse, and  is thorough-



ly familiar with  the operation of No.  3 boiler and its assoc-




iated  equipment.





                             112

-------
Prior to the week of the verification burn, training burns will



be conducted using the reclaim oil system and No. 6 fuel oil



only.  These practice sessions will allow the Operators to



familiarize themselves with the operation of the reclaim oil



system, and with the procedure for switching No. 3 boiler from



No. 6 fuel oil to reclaim oil and back to No.  6 fuel oil.







Also prior to the week of the verification burn, this work



plan will be reviewed with each Operator who will be directly



involved with the verification burn.   Each Operator xvill also



receive his own personal copy of the  Work Plan for ready



reference.
                           113

-------
                            APPENDIX C

               PROCEDURE FOR PCS AIR SAMPLE ANALYSIS


1.   Transfer the Florisil into a 10 ml volume flask.

2.   Pipette 4 ml of n-hexane (Pesticide Grade) into the
     volume glask.   Wait for at least 10 minutes.

3.   Run the final solution on Gas Chromatography with
     Electron Capture Detector.

4.   Compare the fingerprints with known Aroclor to
     determine what kind of PCB is in the sample.

5.   Determine the concentration of the PCB in samples
     by comparing the average peak height of the sample
     and the liquid standards.
                                114

-------
                            APPENDIX I>

      METHOD OF TESTING FOR POLYCHLORINATED BIPHENYLS (PCBs)
                     IN BULK MATERIAL (SOLID)
1.   The bulk sample is ground with a porcelain mortar and
     pestle to uniform particle size.

2.   The sample is then put in a vacuum desiccator over-
     night.  Prior to extraction, the sample is kept in a
     regular desiccator.

3.   The weighed sample (5 - 10 g) is added to the thimble
     and a wad of glass wool is placed on the top.  The
     extraction is then carried out in a soxhlet extractor
     using 150 ml n-Hexane for 4 hours.

4.   The extract is then transferred to a beaker and
     concentrated to 1 - 2 ml in a heated vacuum
     (temperature set at 5) desiccator.

5.   Rinse the beaker with n-Hexane and transfer the liquid
     through a Florisil tube (a disposable pipette packed with
     2-1/2 inches of Florisil) to a 10 ml volume flask.  The
     final solution is 10 ml.

6.   Analyze the final solution by gas chromatography
     with electron capture detector to determine the
     concentration of PCBs.  Inject 5 yl with a 1 yl
     back-up.

GC Condition:

     Column:  SE30 glass column 1/4 in. x 6 ft.  Temp.:  200°

     Detector temperature:  Electron capture detector 300°

     Injector temperature:  280°

     Standard current:  0.5


7/7/80
                                115

-------
                                 APPENDIX E

               PROCEDURES USED BY CHEVROLET CENTRAL OFFICE AND
        CHEVROLET BAY CITY LABORATORIES TO ANALYZE THE PCB WASTE OIL
Chevrolet Central Office Laboratory

Aroclor 1242 and 1254 were analyzed with gas chromatography using the pro-
cedure outlined on page 7 of the April 1980 EPA Newsletter (Volume 3,
Number 2).

Metal analysis for arsenic, cyanide, mercury, nickel, selenium, and zinc
used the methods from the EPA document titled "Test Methods for Evaluating
Solid Waste," SW-846, May 1980.

Phenols were analyzed using the August 1977 EPA document titled "Interim
Methods for the Sampling and Analysis of Priority Pollutants in Sediments
and Fish Tissue."

Chevrolet Bay City Laboratory

Fire point was done by ASTM Method D-92.  Sediment was done by ASTM Method
D-96.  Water was done by ASTM Method D-95.  Viscosity was done by ASTM
Method D-88.
6ARF/9292
10/01/80
                                      116

-------
                    APPENDIX  F




        GM BOILER OPERATION  MEASUREMENTS
Boiler Operating Parameters -* Background Run On May 5. I960

Time
Acceptable
Range
1510
1525
1540
1555
1610
1625
1640
1655
1710
1725
1740
1755
1810
1825
1840
1855
1910
1925
1940
1955
2010
2025
Average
Range
Std. Dev.
fuel Oil
Pressure
(PSIC)
157
120
38
38
38
38
38
38
38
39
40
39
39
40
40
40
40
40
40
40
40
39
39
39
39
2
0.87
Fuel Oil
Flow Rate
(CPH)
3.9
4.1
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
' 4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
. 4.0
4.0
0
0
Steam
Flow Rate
(MLB/HR)
10/
70
40
40
40
41
41
40
40
39
38.5
39
38.5
38.5
38.5
38
38
38
38
38
38
38
38
38
39
3
1.0
Air Heater
Temperature
Air In (°F)

Ambient
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
79
76
74
72
71
71
78
9
3.2
Air Heater
Temperature
Air Out (°F)
250/
350'
355
355
355
352
352
351
352
352
353
353
353
358
358
358
359
358
356
356
355
355
356
355
355
8
2.38
Air Heater
Temperature
GBB In ( F)
350/
580
475
475
475
474
472
471
472
473
476
477
478
480
460
480
480
481
481
482
483
483
483
482
478
12
3.99
Air Heater
Temperature
Gas Out ( F)
200/
300
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
                     Page 1 of 3
                         117

-------
                         Boiler Operating Parameters - Background  Run  On May  5,  1980
CO


Time
Acceptable
Range
1510
1525
1540
1555
1610
1625
1640
1655
1710
1725
1740
1755
1810
1825
1840
1855
1910
1925
1940
1955
2010
2025
Average
Range
Std. Dev.
Windbox
Pressure
(in. w.c.)
0.5/
7.5
4.5
4.5
4.25
4.0
4.5
4.0
4.0
4.0
4.5
4.0
4.5
4.5
4.25
4.0
4.5
4.25
4.5
4.25
4.25
4.5
4.25
4.5
4.3
0.5
0.21
Furnace
Pressure
(in. w.c.)
0.5/
5.5
3.0
3.0
3.25
3.0
3.5
3.0
3.0
3.0
3.5
3.0
3.25
3.5
3.0
3.0
3.25
3.0
3.25
3.25
3.25
3.5
3.25
3.5
3.2
0.5
0.20
Flue Gas
Opacity
1*1
5 or
less
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2
2
2.5
2
2.5
2
2
3
3.5
3
2
1.5
0.4
Fuel Oil
Temperature
J&L
160/
220
191
190
191
194
196
195
190
192
194
191
192
192
190
191
191
191
190
191
192
192
192
192
192
6
1.62
Atomized
Steam
Press. (PSIG)
15/
110
48
48
48
46
48
48
48
50
50
50
50
50
50
50
50
50
50
50
51
50
50
50
49
5
1.2
Air Heater
Pressure In
(in. w.c.)
0.5/
10.5
6.0
6.0
6.0
6.25
6.25
6.0
6.25
6.0
6.0
6.0
6.0
6.25
6.25
6.25
6.25
6.25
6.25
6.25
6.5
6.5
6.5
6.5
6.2
0.5
0.18
Air Heater
Pressure Out
(in. w.c.)
0.5/
8.0
5.0
5.0
5.0
4.5
4.75
5.0
4.5
5.0
5.0
4.5
5.0
4.75
5.0
4.75
4.75
5.0
4.75
4.75
5.0
4.75
5.0
4.75
4.8
0.5
0.18
                                                Page 2 of 3

-------
                   Boiler Operating Parameters - Background Run On May 5, 1980
Time

Acceptable
Range

1510
1525
1540
1555
1610
1625
1640
1655
1710
1725
1740
1755
1810
1825
1840
1855
1910
1925
1940
1955
2010
2025

Average
Range
Std. Dev.
Steam Drum
Water Level
(Inches)
5" of
center
+0.5
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.25
1.25
1.0
1.0
1.0
1.0
1.0
1.0
1.0
+1.0
0.75
0.13
Flame
Temperature
( F)


2625
2650
2625
2600
2625
2625
2625
2650
2650
2625
2650
2650
2650
2650
2625
2625
2640
2625
2625
2640
2600
2625
2630
50
15.4
Back Wall
Temperature
(°F)


1680
1680
1670
1670
1675
. 1675
1680
1675
1675
1690
1680
1695
1690
1685
1695
1695
1705
1695
1705
1700
1700
1700
1690
35
11.5
Oil Flow-
Ultrasonic
(GPM)

(Ref.)
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
Wind
Speed
(MPH)


20
•15
25
25
20
26
20
18
20
19
19
20
19
19
18 •
15
23
20
20
15
10
12
19
16
3.9
Wind
Direction
(Comp. Pt.)


W
W
W
W
W
WNW
W
W
W
W
W
W
W
W
WNW
N
NNW
N
N
N
N
N
WNW
90°
40°
                                           Page 3 of 3

-------
                      Boiler  Operating Parameters - First  PCS  Run  On May  8,  1980
Time

Acceptable
Range

1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
1745
1800
1815
1830
1845
1900
1915
1930
1945

Average
Range
Std. Dev.
Fuel Oil
Pressure
(PSIG)
15/
120
38
38
38
38
38
38
38
38
36
36
36
35
36
36
36
36
36
36
38
38
38
38
36
36
37
3
1.1
Fuel Oil
Flow Rate
(GPH)
3.9
A.I
4.0
4.0
4.0
4.0
4.0
4.0
3.9
4.0
3.9
4.0
4.0
-3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
" 3.9
4.0
4.0
4.0
0.1
0.04
Steam
Flow Rate
(HLB/HR)
io/
70
39
39.5
38
37.5
39.5
41
41.5
43
40.5
40
40
43
40
41.5
43
41
39
38
39
39.5
41
40
38
38
40
5.5
1.6
Air Heater
Temperature
Air In (°F)

Ambient
68
70
69
69
69
68
66
67
67
67
68
67
65
65
65
65
65
65
65
65
65
65
65
65
66
5
1.7
Air Heater
Temperature
Air Out(°F)
250/
350
352
356
358
358
358
355
352
352
350
350
352
352
355
354
354
350
353
355
359
359
358
350
352
350
354
9
3.13
Air Heater
Temperature
Gas In (°F)
350/
580
508
506
508
509
505
505
502
500
495
500
506
506
510
508
505
502
510
510
515
512
510
505
505
502
506
20
4.41
Air Heater
Temperature
Gas Out (°F)

  200/
  300

  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.
  N.O.

  N.O.
  N.O.
  N.O.
                                            Page 1 of 3

-------
                     Boiler Operating Parameters - First PCB Run On May 8,  1980
Time

Acceptable
Range

1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
1745
1800
1815
1830
1845
1900
1915
1930
1945

Average
Range
Std. Dev.
WIndbox
Pressure
(in. w.c.)
0.5/
7.5
5.5
5.5
5.5
5.5
5.0
6.0
5.5
6.0
5.5
6.0
5.75
5.5
6.0
5.5
5.5
6.0
6.0
6.0
6.0
5.5
5.5
6.0
5.5
5.5
5.7
1.0
0.28
Furnace
Pressure
(in. w.c.)
0.5/
5.5
4.0
4.25
4.25
4.25
3.75
4.5
4.0
4.5
4.0
4.5
4.5
4.0
' 4.75
4.25
4.0
4.5
4.5
4.5
4.5
4.5
4.0
4.75
4.0
4.0
4.3
1.0
0.28
Flue Gas
Opacity
(%)
5 or
less
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fuel Oil
Temperature
(°F)
160/
220
203
206
206
206
204
202
200
200
198
200
201
202
202
200
199
198
202
203
204
202
201
201
199
198
202
8
2.45
Atomized
Steam
Press.(PSIG)

    15/
   110

    48
    48
    48
    48
    48
    46
    45
    43
    44
    45
    45
    45
    46
    45
    45
    45
    46
    48
    47
    48
    46
    45
    45
    45

    46
     5
     1.5
Air Heater
Pressure In
(in. w.c.)

   0.5/
  10.5

   7.5
   7.5
   7.5
   7.5
   7.5
   8.0
   8.0
   8.0
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5
   8.5

   8.2
   1.0
   0.42
Air Heater
Pressure Out
(in. w.c.)

   0.5/
   8.0

   6.0
   6.0
   5.5
   6.0
   6.0
   6.5
   6.0
   6.0
   6.0
   6.5
   6.0
   6.0
   6.0
   6.5
   6.0
   6.0
   6.5
   6.5
   6.0
   6.5
   6.5
   6.5
   6.0
   6.0

   6.1
   1.0
   0.28
                                            Page 2 of 3

-------
                   Boiler Operating Parameters - First PCB Run On May 8, 1980
Time

Acceptable
Range

1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
1745
1800
1815
1830
1845
1900
1915
1930
1945

Average
Range
Std. Dev.
Steam Drum
Water Level
(Inches)
5" of
center
+0.75
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
+1.0
0.25
0.051
Flame
Temperature
(°F)


2625
2625
2625
2650
2650
2625
2625
2600
2600
2600
2575
2575
2625
2575
2575
2575
2575
2575
2600
2600
2600
2600
2600
2575
2600
75
24.4
Back Wall
Temperature
(°F)


1690
1675
1680
1675
1670
1650
1670
1655
1660
1655
1660
1655
1650
1660
1660
1660
1645
1650
1655
1660
1645
1645
1660
1660
1660
45
11.5
Oil Flow-
Ultrasonic
(GPM)

(Ref.)
5.10
4.20
3.90
3.50
2.80
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
3.90
2.30
0.852
Wind
Speed
(MPH)


10
10
13
13
16
13
20
16
13
13
10
13
15
10
12
10 .
15
12
19
15
10
5
10
15
13
15
3.3
Wind
Direction
(Comp. Pt.)


NW
NW
NW
NNW
NNW
NW
NW
NNW
NW
NW
NW
WNW
WNW
WNW
NW
W
WNW
NW
WNW
NW
NW
NW
NW
NW
NW
70
20
                                          Page 3 of 3

-------
Boiler Operating Parameters - Second  PCS Run on May  9.  1980


Time
Acceptable
Range
1300
1315
1330
1345
1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
1745
1800
1815
Average
Range
Std. Dev.
Fuel Oil
Pressure
(PSIG)
15/
120
38
38
38
38
38
38
38
38
38
38
36
43
42
42
42
40
38
40
38
38
38
38
39
~7
1.8
Fuel Oil
Flow Rate
(GPM)
3.9
4.1
4.0
3.9
4.0
4.0
4.0
4.05
4.0
4.0
4.0
4.0
3.9
4.05
' 4.0
4.05
4.1
4.1
4.05
4.0
4.1
4.0
4.05
. 4.1
4.0
0.2
0.055
Steam
Flow Rate
(MLB/HR)
10/
70
39
39
39
39
38.5
38
37
39
40
40
36.5
37
35.5
37
38
39.5
39
39
40
40
40
40
39
4.5
1.3
Air Heater
Temperature
Air In (°F)

Ambient
68
65
65
65
65
64
64
64
65
65
65
65
65
65
65
65
65
65
66
65
65
65
65
4
0.79
Air Heater
Temperature
Air Out (°F)
250/
350
362
361
361
361
361
361
361
361
360
360
360
365
366
370
365
365
360
360
360
360
360
360
362
10
2.67
Air Heater
Temperature
Gas In (°F)
350/
580
525
512
523
525
522
522
523
523
520
510
515
530
530
530
530
525
521
520
520
520
520
520
522
20
5.36
Air Heater
Temperature
Gas Out (°F)
200/
300
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
                        Page 1 of 3

-------
Boiler Operating Parameters  -  Second  PCS Run On May 9,  1980


Time
Acceptable
Range
1300
1315
1330
1345
1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
1745
1800
1815
Average
Range
Std. Dev.
Windbox
Pressure
(in. w.c.)
0.5/
7.5
6.0
6.0
6.0
6.25
6.5
6.0
6.0
6.0
6.0
6.0
6.0
6.5
6.0
6.0
6.0
6.0
6.5
6.0
6.0
6.0
6.0
6.5
6.1
0.5
0.20
Furnace
Pressure
(in. w.c.)
0.5/
5.5
4.5
4.5
4.5
5.0
5.0
4.5
4.5
4.5
4.5
4.5
4.5
5.0
4.5
4.5
4.5
5.0
5.0
4.5
4.5
5.0
4.5
5.0
4.7
0.5
0.24
Flue Gas
Opacity
!%i
5 or
less
1
0
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0.4
1
0.5
Fuel Oil
Temperature
(°F)
160/
220
205
190
192
195
195
196
196
196
194
193
195
194
193
194
193
192
192
192
191
192
192
192
194
15
3.02
Atomized
Steam
Press. (PSIG)
15/
110
50
50
50
50
50
50
50
50
50
50
45
50
52
52
50
50
50
50
50
50
50
50
50
7
1.3
Air Heater
Pressure In
(in. w.c.)
0.5/
10.5
8.5
9.0
8.5
8.5
9.0
8.5
9.0
9.0
8.5
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
8.9
0.5
0.21
Air Heater
Pressure Out
(in. w.c.)
0.5/
8.0
6.5
6.5
7.0
7.0
6.5
6.5
6.5
6.5
6.5
6.5
6.5
7.0
7.0
6.5
6.5
6.5
6.5
6.5
6.5
6.0
6.5
7.0
6.6
1.0
0.25
                        Page  2  of  3

-------
                  Boiler Operating  Parameters -  Second PCS Run On May 9, 1980
Time

Acceptable
Range

1300
1315
1330
1345
1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
1745
1800
1815

Average
Range
Std. Dev.
Steam Drum
Water Level
(Inches)
5" of
center
+1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
+1.0
0
0
Flame
Temperature
(°F)


2600
2575
2600
2575
2600
2600
2575
2550
2550
2550
2550
2575
2575
2575
2600
2600
2575
2575
2550
2600
2575
2600
2580
50
19.4
Back Wall
Temperature
(°F)


1670
1675
1690
1670
1675
1660
1650
1650
1660
1660
1570
1670
1650
1660
1660
1660
1650
1650
1660
1650
1660
1670
1660
120
22.2
Oil Flow-
Ultrasonic
(GPM)

(Ref.)
4.40
3.70
3.60
3.30
3.00
2.80
2.60
2.50
2.40
2.20
1.80
2.20
1.90
1.75
1.88
1.50
1.44
1.65
1.51
1.42
1.36
1.18
2.28
3.22
0.878
Wind
Speed
(MPH)


10
12
12
10
12
15
12
15
13
12
10
10
12
15
10
15 '
12
10
8
12
12
10
12
7
2.0
Wind
Direction
(Comp. Pt.)


W
NW
W
U
W
W
WSW
sw
sw
W
W
sw
W
W
W
W
W
W
U
W
W
W
W
90°
20°
                                         Page  3 of  3

-------
                        Boiler Operating Parameters - Third PCS Run on May 10, 1980
r-o


Time
Acceptable
Range
1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
Average
Range
Std. Dev.
Fuel Oil
Pressure
(PSIG)
15/
120
40
36
32
30
30
30
30
30
30
30
30
30
30
30
30
31
10
2.9
Fuel Oil
Flow Rate
(GPM)
3.9
4.1
4.0
4.3
4.3
4.2
4.2
4.1
4.1
4.1
4.1
4.0
4.1
• 4.2
4.0
4.1
4.1
4.1
0.3
0.96
Steam
Flow Rate
(MLB/HR)
io/
70
38
34
35
36
36
34
32
31
31
30
30
29
28
28
28
32
10
3.3
Air Heater
Temperature
Air In (°F)

Ambient
57
60
60
60
60
60
60
60
60
60
65
65
65
65
65
61
8
2.7
Air Heater
Temperature
Air Out (°F)
250/
350
362
350
340
330
320
320
320
321
325
335
325
322
320
320
318
329
44
13.0
Air Heater
Temperature
Gas In (°F)
350/
580
530
510
470
452
445
440
445
450
445
450
440
435
430
425
422
453
108
30.0
Air Heater
Temperature
Gas Out (°F)
200/
300
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
N.O.
                                               Page 1 of 3

-------
   Boiler Operating Parameters - Third PCS Run On May 10,  1980


Time
Acceptable
Range
1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730
Average
Range
Std. Dev.
Windbox
Pressure
(in. w.c.)
0.5/
7.5
6.0
6.0
4.5
4.0
4.0
4.0
4.0
5.0
4.0
4.0
4.5
3.0
3.0
3.0
3.0
4.1
3.0
0.97
Furnace
Pressure
(in. w.c.)
0.5/
5.5
5.0
5.0
3.0
3.0
2.8
3.0
3.0
3.0
3.0
3.0
2.5
2.0
' 2.0
2.0
2.0
3.0
3.0
0.94
Fine Gas
Opacity
SSL
5 or
less
2
2
2
2
2
2
2
2
2
2
2
2
2.
2
2
2
0
0
Fuel Oil
Temperature
(°F)
160/
220
200
200
197
195
193
194
197
199
200
201
202
204
205
205
204
200
12
3.92
Atomized
Steam
Press. (PSIG)
15/
110
50
42
38
38
36
37
38
38
40
38
38
38
38
38
38
39
14
3.3
Air Heater
Pressure In
(in. w.c.)
0.5/
10.5
9.0
8.5
6.5
6.0
6.0
6.0
6.0
6.0
5.5
5.5
5.0
4.5
5.0
4.5
4.5
5.9
4.5
1.3
Air Heater
Pressure Out
(in. w.c.)
0.5/
8.0
6.0
6.5
5.0
4.5
5.0
4.0
. 4.5
4.5
4.5
4.0
4.0
3.5
3.0
3.0
3.0
4.3
3.5
1.0
f
                           Page 2 of 3

-------
                             Boiler Operating Parameters - Third PCS Run On May 10, 1980
00
Time

Acceptable
Range

1400
1415
1430
1445
1500
1515
1530
1545
1600
1615
1630
1645
1700
1715
1730

Average
Range
Std. Dev.
Steaa Drum
Water Level
(Inches)
5" of
center
-i-1.0
l.C
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
i .0
1.0
1.0
1.0
1.0
+1.0
0
0
Flame
Temperature

2600
2550
2600
2625
2625
2650
2650
2640
2650
2650
2650
2650
2650
2650
2650
2630
100
29.1
Back Wall
Temperature

1670
1500
1525
1525
1525
1510
1520
1520
1525
1450
1530
1510
1490
1500
1250
1500
420
83.8
Oil Flow-
Ultrasonic
(GPM)
(Ref.)
5.50
4.00
3.70
3.30
3.20
2.80
2.30
2.30
1.90
1.10
1.10
0.80
0.80
0.50
0.40
2.25
5.10
1.50
Wind
Speed
(MPH)

11
10
15
12
12
10
10
10
6
10
6
7
6
10
5
9
10
2.8
Wind
Direction
(Cotnp. Pt.)


SW
SW
sw
sw
sw
sw
sw
sw
sw
sw
sw
s
s
s
s
ssw
45°
20°
                                                    Page 3 of 3

-------
                                   APPENDIX G

                         SUMMARY  OF WASTE OIL ANALYSIS
Analysis Values requested by both EPA and DNR, unless otherwise noted.
N.D. means not detected.   N.D./E.S. means not detected by emission
spectrographic analysis.

PHYSICAL AND THERMAL PROPERTIES (by Phoenix Chemical)
Ash
Calorific Value

Carbon Residue
Flash Point
Fire Point (EPA only)
Sediment and Water
Viscosity (Saybolt@100°F) (EPA only)
0.45%
18,600 BTU per pound gross
17,495 BTU per pound net
0.93%
34 5° F
400°F (by Chevrolet)
0.1%
175 seconds (by Chevrolet)
NON METALLICS AND INORGANICS (by Phoenix Chemical)
Carbon
Chlorine
Cyanide (DNR only)
Hydrogen
Hypochlorite (DNR only)

Nitrogen (DNR only)
Sulfur
82.62%
 0.97%
N.D.@0.5 ppm ( by Chevrolet )
12.11%
(Rapid oxidation of this chemical precluder
 presence in this oil.)
 0.28%
 0.98%
METALLICS REQUESTED (by Phoenix Chemical)
                                        N.D./E.S.
                                        N.D./E.S.
                                        N.D./E.S.
                                        8.8 ppm
                                        N.D./E.S.
                                            ppm
                                            ppm
                                            /E.S.
N.D
N.D./E.S
N.D./E.S
Arsenic
Barium (EPA only)
Cadmium (EPA only)
Chromium
Cobalt (DNR only)
Copper (DNR only)                      12
Lead                                   23
Mercury
Nickel (DNR only)
Selenium
Zinc

OTHER METALLICS (by Phoenix Chemical)

Aluminum                                120 ppm
Iron                                    43.2 ppm
Magnesium                               31 ppm
Silicon            .        '             0.08%
Sodium                                  254 ppm
Titanium                                23.5 ppm
          (N.D.@0.5ppm by Chevrolet)
          (N.D.@0.04ppm by Chevrolet)
          (7 ppm by Chevrolet)
          (N.D.@0.5ppm by Chevrolet)
Trace (40ppm by Chevrolet)
                                       129

-------
Siimm ay y _of _ Waste Oil Analy s_is (cont' d.)
PagTl
ORGANICS (by Trace Elenents, except as noted)
Aniline (DNR only)
P Chloro M Cresol
Dichlorophenol
Chlorophenols (DNR only)
Chlorinated Dibenzofurans
Dibenzofurans
Dioxln
Phenols
Polychlorinaced Biphenyls
Tri N Cresol Phosphate
Tri Phenol Phosphate
Tri Cresol Phosphate
Triaryl Phosphate Ester (DN'R only)
N.D. @ 100 ppm
100 ppm
 45 ppm
145 ppm  (total)
N.D. @ 20 ppm (by CCA)
N.D. @ 20 ppra (by GCA)
N.D. @ 20 ppm (by CCA)
478 ppm (by Chevrolet)
(See Note Below)
70,920 pptn
 9,450 ppm
   450 ppm
80,820 ppm (8.1%)   (total)
NOTE ON ANALYSIS FOR POLYCIILORKJATED BIPHENYLS

The waste  (reclaim) oil to be used contains 132.5 ppm PCB.  For  the  verification-
burn the waste oil will be spiked to approximately 500 ppm PCB.   In  order to
detect the presence of any Chlorinated Dibenzofurans, Dibenzofurans, or Dioxins
present in the spiking fluid, the waste oil sample was also spiked with the
same fluid.  The PCB content of the spiked sample was 598 ppm.   The  amount of
spike  fluid added  to the waste oil for the verification-burn will be adjusted
accordingly so that the concentration is near, but less than, 500 ppm.   The
actual PCB content of the spiked waste oil for the verification-burn will be
reported when available.

SUMMARY OF WASTE OIL ANALYSIS
 Laboratories  Providing  Analy

 Phoenix Chemical Laboratory
 3953 West Shakespeare Avenue
 Chicago, Illinois 60647

 CCA/Technology Division
 GCA Corporation
 Burlington Road
 Bedford, Massachusetts  01730
         Trace  Elements Laboratories
         460  South Northwest Highway
         Park Ridge, Illinois 60068

         Chevrolet Laboratories
         Chevrolet Motor Division
         General Motors Corporation
         30007 Van Dyke
         Warren, Michigan  48090
                                        130

-------
                                 APPENDIX H

                             TRACE ELEMENTS INC.
                       WASTE OIL ANALYSIS METHODOLOGY

                      (Retyped because copy illegible)


                                                 July 29, 1980


Mr. Al Garwick
Chevrolet Motor Division
100 Fitzgerald Street
Bay City, Michigan  48706

Dear Mr. Garwick:

The following is the information you requested per report #6225 of April 1980:

              Sample Preparation:

    Five milliliters of sample were diluted to 1 liter in water.  One
    such sample was made acidic with 1+1 HC1 and extracted three times
    with CH2C12.  The first extraction was done with 100 mol CH2C12 and
    the second and third with 50 ml CH2C12•  Due to the nature of the
    extract (suspected to be highly concentrated), a Kurderna-Danish
    concentration was not performed on the extract.

    A second extraction was performed with 85%/15% hexane/CH2C!2 on a
    neutral sample.  The extracts were again not concentrated by Kurderna-
    Danish.

Both samples were run GC-MS.

              Instrumentation:

    GC-MS:    Hewlett Packard 5990A
    Columns:   OV17  10m Capillary
                Tenax GC  6' packed

    Conditions:   Injector 250°C
                  Injection amount 2 pi
                  Program:  40° /4 min, 8°/min •+ 250°

             Results:

    Acid Extraction:

             Dichlorophenol

             p chloro m cresol
                                      131

-------
Mr. Al Garwick                                           July 29, 1980
Chevrolet Motor Division
Bay City, Michigan  48706                                page  -2-
    Neutral Extraction

             PCB

             tri m cresol phosphate

             tri phenyl phosphate

             tri cresyl phosphate


             Quality Control

    A phenol standard, PCB standard,  and phosphate pesticide standard
    were all run prior to the samples to determine instrument
    sensitivity and proper temperature program for good retention
    times and resolution.  Standards  are also run daily for instrument
    tune-up.


I hope this information is what is needed.

Thank you for doing business with TEI.   We  hope to continue working with
you in the future.


                                        Sincerely,

                                        TEI ANALYTICAL
                                        Susan Krejci
SK/
as
                                     132

-------
           APPENDIX I

PHOENIX CHEMICAL LABORATORY, INC.
  RECLAIM OIL ANALYSIS RESULTS
               133

-------
                 Pi       •   Tl     •    ]   i    l               f
                 I  h o e n i x vj n e m i c a l   La! • oral o r v, inc.
                           r'.'Ei AND uTkiCAX r T:-CHSO:.O' -,iYrc.
                                3' ••_..- :,}-: .1- • ••;:•• r i.  AV • : . ,
                                   CHICAGO  ILL 6064''
                                   /•iJ'just TJ, 10V)
                 C;,:,vrol :t V.t;>r  division
                 "/.raral  ' otors Corporation
                 'Jay City Plant
SAMPLE OF         ™ cnyfn?ih.S48706                 i ABORMORV NO  0 3 23  18
MAKKhl)                :d Oil
                 3-26-SO

       Litrogen,  S  ^Ti'. D322S                                         0.23
       Sulfur,  X   /^5T"; 0129                                            0.0
       Ciilorlne,  ?  AST!' Do^O                                          0.97
       Caruon,  «     )                                                   C2.T2
                                                                       IT. 11
       /.sn,  «  AST. I  J-'C2                                                L'..'i5
       Hater by Distillation, Z  ?ST\1 D95                              Tr,;cc
       Sediment, mg/100 ml  ASTM F312                                 CZvGOv
       Heat  of Cunb'jstion, 3TU/lb. AST;-', D240
              Gross                                                    1C, 500
              Net                                                     17,405
       Conraci^on Carbon Residua, «  ASTil C139                              n.'J3
       Flash Point (COG), F  AS7K irxj                                    345
       Sodiun, ppn1.    )                                                  2C4
       AluHiirium, pp,:,  )                                                  12^
       Iron, ppiii      )                                                   '"'3.2
       Magnesium, pp.i )   Atomic Absorption                               31.0
       Lead, pps.i      }                                                   23. 1,
       Titciiiiu-n, pp."  )                                                   ^:3.:»
       Cooper,  ppn    )                                                   12.C
       Lnrc.:!iu"i, ppn  )                                                    "••"
       billed,  '/.  (?oc  :ncj  sh-'2t)                                          -"!'
                                           134

-------
 I koemx Unemical  Laboraiorv, I
          FUEL AND LUBRICANT TECHNOLOGISTS
               3r>'_>3 SHAKE SPFARh AVFXUE
                 CHICAGO. ILL. 606-7
                  August 15, 19JO

Chevrolet  ''.otcr  n vis ion
                                                               nc.
SAMPLE OF
                                     I AUOKATOia NO  0 3 ?.B  18
MAKkLD
      Silica  is determined by  ashing the sample,  extracting with hydrochl.^-
      and treating the insolubles with sulfurlc acid and hydrofluoric acid.
      The volatiles liberated  by hydrofluoric acid  are calculated as silica.
                       Emission  Spectrographic Analysis  (Jarrell-Ash instru.-nent)

              Major:       Sodium, phosphorus, Silicon

              Minor:       Alunvinun,  Iron, Lead,  Chromium, Copper,
                           iiagnesium, Titanium

              Trace:       Coron,  Cclciuni, Potassium, ilanganeso,
                           ilolybdenuii, Nitrogen,  Tin, Vanadium, Zinc
                                         .
                                        135

-------
                            APPENDIX J

                   GM BREATHING ZONE AND POWERHOUSE
                        MONITORING RESULTS
                    Chevrolet Motor Division
                       Bay City,  Michigan

                         May 5-10,  1980
Results of Tests for Polychlorinated Biphenyls (PCS's)
Test
 No.
Location and Description
Sampling
  Time
(minutes)
 PCB's
(mg/m3)*
        MIOSHA Permissible Exposure Limit

        Tests 1-12 were made during the
        background burn on May 5, 1980.

  1     Breathing zone  (BZ) of D. Mika
        (374-42-7560).  Operator,
        boiler No. 3, afternoon shift.

  2     Repetition of Test 1.

  3     BZ of J. Wiznerowicz  (379-34-
        4679).  Operator, boiler
        No. 3, afternoon shift.

  4     Repetition of Test 3.

  5     General condition  (GC) test made
        approximately 5 feet west of
        boiler No. 3 exhaust stack,
        second floor.

  6     Repetition of Test 5.
                                      163

                                      156



                                      164

                                      154
                                      160

                                      153
              ND**

              ND



              ND

              ND
              ND

              ND
 *mg/m3  =  milligrams  per cubic meter of air.
**ND = none  detected
                              136

-------
Chevrolet Motor Division
Bay City, Michigan
Page Two
Results of Tests for Polychlorinated Biphenyls (PCB's)
Test
No.
7
8
Location and Description
GC test made atop boiler No. 3
control panel, first floor.
Repetition of Test 7.
Sampling
Time
(minutes)
156
153
PCB's
(mg/m3)
ND
ND
   9     GC test made simultaneously with
        Test 5 near boiler No. 3 exhaust
        stack.

  10     GC test made simultaneously with
        Test 6 near boiler No. 3 exhaust
        stack.

  11     GC test made simultaneously with
        Test 7 atop boiler No. 3 control
        panel.

  12     GC test made simultaneously
        with Test 8, atop boiler No.  3
        control panel.
156
153
156
 153
ND
ND
ND
 ND
        Tests 13-22 were made during
        the first day of the PCS burn
        on May  8, 1980.

  13     BZ of J. Fryzel  (369-18-1522).
        Operator, boiler No. 3, after-
        noon shift.

  14     Repetition of Test  13.

  15     BZ of D. Mika, operator
        boiler  No. 3, afternoon shift.

  16     Repetition of Test  15.

  17     GC test made atop boiler No.  3
        control panel.
174
175
179
175
ND
ND
ND
ND
 190
 ND
                              137

-------
Chevrolet Motor Division
Bay City, Michigan
Page Three
Results of Tests for Polychlorinated Biphenyls  (PCB's)
Test
No.
18
19
20

21
22

23
24
25
26
27
28
29
30

31
Location and Description
Repetition of Test 17.
GC test made simultaneously
with Tests 7 and 18.
GC test made near boiler
No. 3 exhaust stack.
Repetition of Test 20.
GC test made simultaneously
with Tests 20 and 21.
Tests 23-32 were made during
the second day of the PCB burn
on May 9, 1980.
BZ of E. Markiecki (375-14-
7126). Operator, boiler
No. 3, afternoon shift.
Repetition of Test 23.
BZ of J. Fryzel. Operator,
boiler No. 3, afternoon shift.
Repetition of Test 25.
GC test made atop boiler No. 3
control panel.
Repetition of Test 27.
GC test made simultaneously
with Tests 27 and 28.
GC test made near boiler No. 3
exhaust stack.
Repetition of Test 30.
Sampling
Time
(minutes)
162
340

186
153
339

153
183
152
172
150
180
332

152
180
PCB's
(mg/ni3)
ND
ND

ND
ND
ND

ND
ND
ND
ND
ND
ND
ND

ND
ND
                              138

-------
Chevrolet Motor Division
Bay City, Michigan
Page Four
Results of Tests for Polychlorinated Biphenyls (PCS's)
Test
 No.
Location and Description
Sampling
  Time
(minutes)
 PCB's
(mg/m3)
 32     GC test made simultaneously
        with Tests 30 and 31.
                                     332
                                                          ND
        Tests 33-42 were made on the
        third day of the PCS burn on
        May 10, 1980.

 33     BZ of J. Fryzel.  Operator,
        boiler No. 3,  afternoon shift.

 34     Repetition of Test 33.

 35     BZ of D. Mika.  Operator,
        boiler No. 3,  afternoon shift.

 36     Repetition of Test 35.

 37     GC test made atop boiler No. 3
        control panel.

 38     Repetition of Test 37.

 39     GC test made simultaneously
        with Tests 37 and 38.

 40     GC test made near boiler No. 3
        exhaust stack.

 41     Repetition of Test 40.

 42     GC test made simultaneously
        with Tests 40 and 41.
185
105
184
102
183
104
ND
ND
ND
ND
ND
ND
                                     286


                                     181

                                     101


                                     282
              ND


              ND

              ND


              ND
WBK:SZ
                              139

-------
                                            APPENDIX K


                                    GM  BOILER RESIDUE ANALYSIS



                                 BOILER PCB RESIDUAL ANALYSIS
Scraping
 Number
           Location
          1 Aualys 1-*
           Stack »1
           Preheater Air
            Outlet
           Preheater Air
            Inlet
           Furnace  North
            Wa 11 it 1
           Furnace  South
            Wall «1
           Furnace  Floor-Center
           Furnace  Floor-Rear
           Furnace  Rear Wall
                                  Date
                     10/15/79
                               Aroclor 1242
                              Concentration
                                   (PPb)
                                     ND

                                     ND

                                     ND

                                     ND

                                     ND
                                     ND
                                     ND
                                     ND
           Aroclor 1242
            Sensitivity
                (PPb)

                 47

                126

                448

                 55

                 61
                 26
                 20
                 48
           Aroclor 1254
          Concentration
               (PPb)


                 ND

                 ND

                 ND

                 ND

                 ND
                 ND
                 ND
                 ND
           Aroclor 1254
            Sensitivity
               (PPb)

                 51

                137

                486

                 59

                 66
                 28
                 21
                 52
11

12

1 J

14
Preheater Hot
Inlet
Furnace North
UaJl #2
Furnace South
Wall 12
Stack f-2
                        Gas
                                4/18/80
                                    ND

                                    ND
                                               NU
                                               ND
                                 15

                                 ND
                                                                    ND
                                                                    ND
   9
  10
burner Tips
I'reliejter Air Inlet
Furnace Floor-Center
Furnace Floor-Rear
Furnace Rear Wall
Preheater Hot Gas
 Inlet
Stack
Preheater Hot Gas
 Outlet
Furnace North Wall
Furnace South Wall

ND - Mon-Detectable
                                5/15/80
771
 ND
 ND
 ND
 ND

 ND
 ND

 ND
 ND
 ND
187
 24
 25
  9
 20

 12
 11

 10
 10
 18
413
 ND
 54
 12
 49

 20
 17

 ND
 ND
 27
203
 26
 27
  9
 22

 13
 12

 11
 11
 20
                                                140

-------
                                   APPENDIX L

                           GM BLOOD CHEMISTRY RESULTS
                        (Retyped because copy illegible)


To       Powerhouse Operators           Location  Bay City

From     Dr. Thomas A. Hockman          Location  Bay City

Subject                                 Date      July 11, 1980


I have received and reviewed the PCB blood level reports on the samples drawn
after the test burn.

We can say the following:

     1.   The range of results before the burn was 7 to 36 parts per billion.

     2.   The range after the burn is 5.4 to 40 parts per billion.

     CONCLUSION:  This is About as Close to Identical as Analytical Data Can Get.

     3.   The average (or mean) before the burn was 18.0 parts per billion.

     4.   The average after burn is 18.6.

     CONCLUSION:  No Change

     5.   In comparison, 39 people who work in the office and presumably have
          no work association with PCB show that:

          A)   Those who eat in excess of 8 oz of Great Lakes fish per week
               have a range of 4 to 71 parts per billion and an average of
               23.5 parts per billion.
          B)   Those who average less than 8 oz of Great Lakes fish per week
               have a range of 3.4 to 50.3 parts per billion and an average
               of 18.03 parts per billion.

     CONCLUSION:  The Powerhouse Employes Show Blood Levels Comparable to Those
                  Office Employees With the Least Exposure to Environmental PCB.

                  And Those Individuals With the Highest Blood Levels in any of
                  the Groups Analyzed Remain Low Relative to Danger Levels and
                  to State-wide Levels.

Any participants in the above sampling program who wish to review their
individual reports are welcome to consult with me.  Call Ext. 392 for an
appointment.
                                        Thomas A.  Hockman
                                        Medical Director
TAH:T
                                      141

-------
                                  APPENDIX M
                            GCA FIELD CALIBRATIONS
            Continuous Monitor  Data -  corrected  15 minute  averages.

     Strip chart measurements were reduced and digitized to 15 minute averages
and corrected for calibration drift using the following procedure:

Theory:
     Instruments calibrations are applied to digitized strip chart data in the
following manner.
Given:
     On date (Dl) in hour (HI)  the recorder open measured ZER01 (ZCR1) chart
units when the input of zero ppm (ZEU1)  was applied to the instrument.  The
recorder pen also measured SPAN1 (SCR1)  chart units when the input of SPAN1
(SEUl)  ppm was applied to the instrument.  On a later date (D2) in hour (H2)
the recorder pen measured ZERO (ZCR2) chart units when an input of zero ppm
(ZEU2)  was applied to the instrument.  The recorder pen also measured SPAN2
(SCR2)  chart units when an input of SPAN2 (SEU2) ppm was applied to the instru-
ment.  On an arbitrary date (D) at some time (T) the recorder pen measured
CV chart units.
Find:
     The actual value on date D at time T in ppm.

Assumption  1;  The base line (zero) of the recorder drifts linearly between
measurements of the zero.
The zero (Z) at time T on date D is therefore
          Z = ZCR1 + (T-H1) *  (ZCR2-ZCR1)/(H2-H1)                     (1)
                                      142

-------
Assumption 2:   The chart scale (ppm/chart unit)  of the recorder drifts
linearly between measurements of the span.
The scale factor (SF1)  on DATE1 at HOUR1 is
          SF1  = (SEU1-ZEU1)/(SCR1-ZCR1)
The scale factor on DATE2 and HOUR2 is
          SF2  = (SEU2-ZEU2)/(SCR2/ZCR2)
The chart scale factor on date D at time T is therefore
          SF = SF1 + (T-H1) *  (SF2-SF1)/(H2-H1)                       (2)

Applying equations (1) and (2) to the chart value CV on date D at time T we
find
          Actual Value = ZEU + (CV-Z) * SF (ppm)
Stack Data  Reduction
      Source sampling data  reduction sheets for  the  PCB  and Modified  Method  5
 (organic) sampling trains  have been retained  by the contractor for this  pro-
ject, under Contract 68-02-3168,  Work Assignment No.  9.   Copies  of these
sheets  may  be  obtained  by  contacting GCA/Technology Division.
                                     143

-------
                                APPENDIX N

                      DISCUSSION OF RECLAIM OIL DATA
                                                GGA CORPORATION
                                                Technology Division
GCA
August 8, 1980
Mr. Peter Collins
RTI
P.O. Box 12194
Research Triangle Park, NC  27709

Dear Mr. Collins:

Enclosed are the GCA/Technology procedures for PCB in Fuel Analysis
used for the GM PCB verification burn.  The procedures document sampling,
chain-of-custody, analysis, and quality control.

Please call me at (617)-275-5444 if you have any further questions.

Sincerely,
                 /k
Johanna M. Hall
Laboratory Analysis Department
JMH:vmk
Enclosures

cc: K.T. McGregor  (w/o enclosures)
    G.T. Hunt  (w/enclosures)
    R.M. Ellersick  (w/enclosures)
    D. Sanchez  (w/enclosures)
                                     144

-------
                       ANALYSIS OF FUEL OIL FOR PCS






Sampling Procedures




     Fuel and waste oils were sampled at the GM-Bay City plant as part of the




PCS verification burn program.  The oil sampling was performed by GM personnel




and the samples shipped to GCA for analysis.  Each sample was provided in a




glass container which had been rigorously cleaned using the following sequence:




acid soak, alcoholic KOH soak, distilled deionized water rinse, acetone rinse,




and hexane rinse.  As a further precaution container caps were lined with




Teflon.




Sample Traceability




     Samples were received at the GCA facility by the Sample Bank Manager




where sample conditions and identification codes were recorded prior to its




distribution for analysis.  All transfers of the samples followed prescribed




chain-of-custody procedures.  All subsequent handling of each sample was




recorded in the sample bank master log accompanied by appropriate signatures




and dates of transfer.  At the conclusion of the prescribed analysis sequence




the remainder of each sample was returned to the sample bank.  (See attach-




ments for excerpts from GCA/Technology Quality Control Manual).




Analysis Procedure




     Fuel oil sample bottles were heated to 40°C in a water bath and stirred




for 1 minute prior to aliquotting to ensure sample homogeneity.




     A 0.5 g portion of each sample was transferred to a clean, pre-tared




5.0 ml volumetric flask.  The flask was reweighed and the sample diluted to




5.0 ml.  The well-mixed sample was removed to a clean, 10 ml vial containing




5.0 ml of concentrated sulfuric acid for cleanup per the procedure cited in




the Federal Register, May 31, 1979 (40 CFR 761-31538).  The oil/acid phases




                                    145

-------
were mixed and the oil fraction quantitatively transferred to a clean vial




containing 5.0 ml of 10 percent NaHC03 (aq).   A portion of the oil fraction




was then transferred to a septum vial for gas chromatographic analysis.




     Instrument operating parameters were as  follows:




     Instrument:  Hewlett-Packard Model 5840  Gas Chromatograph, equipped




                  with an Electron Capture Detector (BCD) and Model 7671A




                  Automatic Liquid Sampler.




     Column:      6 ft x 2 mm pyrex, packed with 1.5 percent OV-17/1.95




                  percent QF-1 on 100/120 mesh Chromsorb WHP.




     Temperatures (°C):  Column = 175




                         Inlet  = 270




                         Detector = 300




     Gas Flow:    50 cc/min 5 percent methane in argon




     Area Rejection:  1000 area counts




     Slope Sensitivity:  0.18




     Instrument calibration was accomplished  by injection of 1.0, 2.0 and




10.0 ppm (pg/ml) dilutions of an Aroclor 1242 standard prepared in hexane.




These dilutions were prepared from a commercially available stock solution




distributed by Applied Science Labs, Inc., State College, Pa.  Calibration




was checked every sixth sample to ensure a uniform detector response.  A




typical Aroclor 1242 standard chromatogram is attached noting retention time




of the five peaks where areas were summed in  obtaining each of the three




calibration points  (2.90, 3.96, 5.21, 6.74, and 9.85 rain.).




     Calibration curves were prepared daily from a linear regression analyses




of the above three Aroclor dilutions.  Each of these data points was obtained




from a sum of the five appropriate areas.





                                     146

-------
     Calibration curves were rejected if the correlation coefficient of the
linear regression analyses was less than 0.996.   All samples were quantitated
by entering the area summation into the appropriate calibration curve.   Sample
extracts were diluted, if necessary, to be within the concentrations bracketed
by the calibration standards.
Results
     Quantitative Aroclor 1242 measurements on each of the verification burn
oil samples are shown in the following Table.  Each was calculated from the
Applied Science curve per the calibration procedure described earlier:
                        Sample ID                      yg/g
                    Day 1 fuel feed                    <1.2
                    Day 2 fuel feed                    72
                    Day 3 fuel feed                    76
                    Day 4 fuel feed                    88
                    Reclaim waste oil                 750
Quality Control
     Three 0.5 gm aliquots of No. 6 fuel oil were spiked by the addition of
25 yl of a 1.0 yg/yl solution of Aroclor 1242 in iso-octane (Applied Science).
As a means of method verification, samples were processed as described above.
The results of these analyses are summarized in the following Table:
                       QUALITY CONTROL - PCBs/FUEL OIL
Sample
A
B
C
X
S
yg Added
25
25
25
25
	
yg Observed
24
24
23
24
0.58
% Recovery
96
96
92
96
_
                                     147

-------
     Aroclor 1242 reference materials (Lots No.  8758 and No.  8937) were ob-




tained from the Reference Standards Repository,  Environmental Toxicology




Division, EPA Health Effects Research Laboratory, Research Triangle Park,




North Carolina.




     Dilution of the above materials were prepared by the laboratory quality




control coordinator and submitted as "blind" checks on the preexisting cali-




bration curves constructed from each of two commercially available Aroclor




1242 stock solutions.




     The initial calibration curve had been prepared from dilution of an




Aroclor 1242 stock solution provided by Supelco, Inc. of Belleforte, Pa.




The observed concentration for each of the "blind" checks as  calculated from




this calibration mixture are shown in the following Table:




                              RTP AROCLOR 1242
Lot No.
8758
8937
(using Supelco calibration)
Nominal
concentration
(yg/ml)
34
6.7

Concentrat ion
observed
(yg/ml)
22
4.3

     Further inquiry included conversations with appropriate personnel at




Supelco.  To their knowledge there was no justifiable evidence that their




stock solution was responsible for the observed concentration discrepancy.




     However, when the same two "blind" checks were quantitated from the




Applied Science calibration mixtures the results were as follows:
                                     148

-------
                                          Nominal        Concentration
                                       concentration        observed
                 Lot No.                  (pg/ml)           (yg/ml)

       8758                                 34                34

       8937                                  6.7               6.7

       (using Applied Science calibration)


     As a further means of verification, quality control check samples pro-

cured from EPA/EMSL (Cincinnati, Ohio) were submitted for analysis.  The

results for these Aroclor 1242 QC mixtures as derived from the same Applied

Science calibration mixture were the following:
                                                   Mean
                                              concentration
               Nominal                           observed
            concentration                   (duplicate analyses)
               (yg/ml)                            (yg/ml


                 6.3                                8.4

                 6.3                                8.6

                 1.9                                2.3
                                     149

-------
                                                               Section 1.2.1
                                                               Page 1 of 1
                                                               Revision 0
1.2.8  SAMPLE ANALYSIS
     The GCA/Technology Division performs a wide variety of programs requiring
reliable chemical analyses.  The Technology Division Analytical Laboratory
performs the bulk of these analyses but field measurements and certain other
air pollution analyses are handled by the Environmental Measurements Depart-
ment.  Both groups strive to include all elements of quality assurance in
their programs.
1.2.8.1  Analytical Laboratory Quality Control
     The Analytical QC Manual documents procedures in use which are implemented
by the entire laboratory staff with direction from the Analytical QC Committee
and the QA Manager.  The Analytical QC Committee consists of the Sample Bank
Manager, an analyst responsible for implementing inorganic analytical QC and
one responsible for organic QC.  These committee members are appointed by the
Laboratory Manager and report to him and to the division QA Manaber.  The
Committee works closely with the QA Mana'ger and the chairman serves on the
division QA/QC Committee.
1.2.8.2  Field Measurements Quality Control
     Environmental Measurements Department staff members with expertise in
ambient, fugitive and source emission measurements serve on the department
QA/QC Committee and ensure that all staff members are following appropriate
quality control procedures.  The Sampling and Field Measurements QC Manual
includes checklists for field instrumentation.
1.2.8,3  Specialized Analyses/QA Project Plan
     If a project requires analytical work that will be done by other than GCA
laboratory staff, the QA Project Plan will outline the quality control pro-
cedures for analysis.  An example is the Asbestos Measurement program which
requires skilled geologists or mineralogists to perform the analyses.  Environ-
mental Engineering Department staff with the necessary expertise identify and
characterize the asbestos and use the appropriate QC procedures.
                                    150

-------
                                                              Section 1.2.7
                                                              Page 2 of 2
                                                              Revision 0
1.2.7.3  Sample Bank/Chain of Custody
     The chain of custody for field measurements is maintained by the ambient
site log or the source measurement project log book.  Sample recovery and
integrity forms are used to identify the person responsible for the sample at
all times from collection to final measurement or analysis.
     All samples submitted to GCA/Technology Division laboratories are brought
to the Sample Bank Manager who establishes a chain of custody by assigning a
GCA Control Number to each sample on receipt which identifies it through all
further handling.  The sample is recorded in the Master Sample Log under this
number and the QA Project Plan is consulted for preservation and storage re-
quirements.  Technology Division maintains large, locked, refrigerated and
nonrefrigerated storage areas with provision for hazardous material storage.
After necessary preservation or subdivision, the Sample Bank Manager stores
each sample in the appropriate area, filed under its GCA Control Number.
Samples which will not be analyzed by GCA's Laboratory Analysis Department
are handled in essentially the same way with the Project Manager responsible
for notifying the Sample Bank Manager of receipt of samples.
     The chain of custody for information gathering projects is outlined in
the QA Project Plan.  Accurate recordkeeping is essential and should include
a log of the numbered survey forms which can follow each form through com-
puterized data handling and storage.
                                      151

-------
CHEVROLET
             •<**
Bay City Plant          APPENDIX 0

          LOG OF SIGNIFICANT ACTIONS
                          LOG OF SIGNIFICANT ACTIONS
        Time
       Action
        Background Run  on May 5,  1980:
        1510
        2025
       Initiated burn and monitoring.
       Terminated burn and monitoring.
        First PCS Run  on  May  8,  1980;
        1350
        1945
       Initiated burn and monitoring.
       Terminated burn and monitoring.
        Second PCS Run  on  May 9,  1980:
        1258
        1535

        1542

        1815
       Initiated burn and monitoring.
       Observed low flame and backwall  temperatures,
       Switched to No. 6 fuel oil  supply system.
       Observed acceptable temperatures.
       Switched to reclaim oil supply  system.
       Terminated burn and monitoring.
        Third PCB Run on May  10,  1980:
        1400
        1738
        1739

        1750

        1840

        1845
       Initiated burn and monitoring.
       Reclaim oil supply system pump  inadvertantly
       stopped.   Oil pressure and  flow dropped.
       Pump immediately restarted.  Oil flow would
       not increase above 1.6 gpm.
       Switched to No. 6 fuel oil  supply system.
       Attempted to regain 4 gpm oil flow.
       Switched to No. 2 fuel oil  supply system.
       Attempted to regain 4 gpm oil flow.
       Unable to regain 4 gpm oil  flow.
       Terminated burn and monitoring.
       Secured boiler.
                                        152
               Chevrolet Motor Division Gcnoial Motors Corporation 109 Fitzgerald Streot, Bay City, Michi'jnn 43707

-------
                                 APPENDIX P

               PCB WASTE OIL SUPPLY SYSTEM OPERATIONAL PROBLEM
  The supply system used for the PCB waste oil during the verification burn
  was assembled primarily from components that were available on-site.  Due
  to its limited application and expected service life,  this system was not
  as sophisticated as  the  permanent oil supply system for No.  6 fuel oil.
  The PCB waste  oil  supply system did not  have  the filtration equipment,
  continuous bypassing piping, nor degree of delivery pressure control that
  the permanent oil supply system has.

  Because of these limitations, maintaining a constant flow from the waste
  oil to  the boiler was a  problem during all three PCB  runs.   Flow rate
  irregularities,  in  the   form   of   abrupt  changes  (spikes)  occurred
  frequently  and  unexpectedly  throughout  the   runs.   However,  through
  continuous monitoring and manual adjustment, we were able to  control the
  amplitude and duration of the spikes and maintain an adequate flow rate in
  such a manner that the other key  parameters  remained at acceptable levels.

  Two revisions should be made to  the waste oil supply system to improve its
  performance.   The first revision should correct the  oil-related problems
  experienced during the verification burn due to the absence of filtration
  of the No. 6 fuel oil.  Rather than delivering the No.  6 fuel oil directly
  to the  waste oil  supply  tank  from  a  tanker truck,  the No.  6  fuel oil
  should be pumped from an existing No. 6 fuel oil bulk storage tank through
  the existing No. 6 fuel oil filtration equipment to  the  waste oil supply
  tank.   The  second  revision should  correct equipment-related  pressure
  control problems of  the  system.  A  waste oil  supply  system  return line
  should be provided from the boiler to the waste  oil supply tank to provide
  an alternate  (loop)  path  for excess oil flow to eliminate  "dead ending"
  pressure fluctuations.
6ARF/9244
10/10/80
                                    153

-------
                               APPENDIX Q

                 PREMATURE  TERMINATION OF  THIRD  PCB  RUN

The third PCB run was terminated prematurely due to  a malfunction within
the waste oil  supply system.   While  attempting  to compensate for a fuel
flow spike, the  boiler  operator inadvertently turned off the  waste oil
supply pump.  Both the  waste  oil  supply  pressure and the waste oil flow
rate dropped significantly.  However, the  pump  was  restarted quickly so
that the burner flames were not lost.  Upon restarting of the supply pump
and re-establishment of supply oil pressure, neither the oil pressure at
the burners  nor the  oil  flow  rate  could be increased to  their  prior
levels.

Because the required  conditions  of the verification burn were not being
met, the boiler was switched to No. 6 fuel oil in accordance with estab-
lished safety and permit procedures.  Oil  pressure and flow  rate were not
improved by  the switch.  The  observed  gauge readings  and  the boiler's
reaction to the control adjustments that were  made indicated the presence
of  a  partial  blockage  upstream of the burners in  the piping which is
common to the  three  oil supply systems  for this boiler.  The boiler was
then switched to the No. 2  fuel oil system in an  attempt  to dissolve or to
dislodge the suspected  blockage with the  lighter fuel oil.   This action
also had no effect on either pressure to the burners  or flow  rate.  Having
exhausted all  available alternatives, the burn was  terminated,  and the
boiler was  shut down.

The available evidence indicates that a partial  blockage occurred in the
boiler oil flowmeter.  This device operates on a piston and  port princi-
ple, and a high viscosity substance could restrict piston movement
and/or reduce the effective port area of  the meter.   This type  of failure
would  account   for  the reduced  oil  pressure and flow  rate which were
observed.  Blockages  at other  possible locations in  the  system would pro-
duce different  observed  conditions.

The boiler operating parameters monitored during the third PCB run indi-
cate that a progressive restricting  action within  the  port area of the
meter was indeed occurring  during  that entire run. Each  of the parameters
which were  dependent on oil  flow rate  had steadily decreased from the
initiation of the run until  the malfunction occurred.  Actual oil flow was
apparently dropping due to the smaller  and smaller effective port areas,
even  though  the flow  meter was  showing  a constant  flow rate due to a
"sticking" condition.

In  support  of  this reduced oil flow rate  during the third PCB run, key
boiler parameters  were  analyzed and  compared with the other three runs.
Air heater  temperature-air in, flue gas  opacity, fuel  oil  temperature,
steam drum water level,  and  flame temperature were found  to be independent
of  fuel  oil   flow  rate  within  the  range  of concern.   Air  heater
temperature-air   out,   air   heater   temperature-gas   in,   air  heater
temperature-gas out, windbox pressure,  furnace pressure, air heater pres-
sure in, and air heater pressure out were found to be related to fuel oil
flow rate within this range, but provided inconsistent results. Fuel oil
pressure,  steam   flow  rate,   atomizing   steam   pressure,   and backwall
temperature were found to be proportional  to fuel oil flow rate  within the
                                 154

-------
range of  concern,  and  were used to determine  that,  in fact, an average
fuel oil flow rate of 3.2 instead of 4.0 gallons per minute was occurring
during the third PCB run, as summarzied below.

                           Average at     Average for     Adjusted gpm
   Parameter Examined        4 gpm       Third PCB Run    Third PCB Run

Fuel Oil Pressure           38 psig        31 psig           3.3 gpm
Steam Flow Rate             39 Mlb/hr      32 Mlb/hr         3.2 gpm
Atomizing Steam Pressure    48 psig        39 psig           3.3 gpm
Backwall Temperature        1,670°F        1,500°F           3.1 gpm

                       Average Fuel Oil Flow Rate            3.2 gpm

After the  termination  of the third PCB run, the meter was disassembled.
After cleaning and reassembly of the meter, the problem was eliminated and
it has  not reappeared  during  several subsequent firings  of the boiler
using No.  6 fuel oil and the primary oil supply  system.
                                  155
6ARF/9243
10/01/80

-------
                                 APPENDIX R

                             DATA INTERPRETATION
DESTRUCTION EFFICIENCY

     A primary objective of this program was to document the PCB destruction
efficiency for a high efficiency oil-fired industrial boiler.  To make this
determination statistically valid, 3 days of test data were required, as well
as a measurement of variability between test days.   Unfortunately, test re-
sults for the third burn day are lost.   Further, there were no measurable
PCB concentrations from which to determine variability on each of the remain-
ing 2 days.  Therefore, the planned statistical tests could not be meaning-
fully applied to the test data.

     The probability of the achievement of greater than 99.9 percent destruc-
tion cannot be assigned a statistical value.  However, the measurement of 2
days of greater than 99.994 percent destruction efficiency strongly indicates
the high probability of that achievement.

     A more detailed description of the statistical methods used is available
in the PCB burn test plan.4

Statistical Test

     On each day there were three samples available at each onsite monitor
(one obtained during boiler test, one obtained before, and one after).  An
appropriate statistical test for testing the null hypothesis of the equality
of PCB levels at the three monitors (i.e., no plant contribution) versus the
alternative hypothesis of differences between the background station and the
other two stations  (i.e., plant adds to PCB levels) is a paired t test.  Let
Ai, BI, and C^ represent the PCB levels at Stations A, B, and C for times 1,
2, and 3, then the t statistic of this t test is:
                                       SD
 where
                                      3
                                      E
                                      156

-------
and
                                  1=1
                            "D          2
with

                                  B. + C.
for i = 1, 2, and 3.  The power of the test (i.e., probability of rejecting
incorrect null hypothesis) depends upon the level of significance and the true
value of k.  If we view the t test as a one-sided test with level of signifi-
cance of 0.05, then some estimated maximum power values for various k values
are:

                                        Estimated
                          k           maximum power
                         1                0.20

                         1.5              0.40
                         2                0.45

                         3                0.50

                         4                0.55
                         6                0.60

     Using the measured ambient PCB concentrations listed in Table 5, the t
statistic results are summarized in Table R-l.

     For this evaluation HQ is accepted if t-test (see Table R-l) is less than
t0 os^2) = 4.303 and HI is accepted if t-test is greater than t0
-------
                         TABLE R-l.   t-TEST FOR UPWIND VERSUS DOWNWIND SITES
Ul
CO
Time
period
Preburn
Burn
Post-burn
Critical value
of student's t
distribution
t Test
Day 1 Day 2 Day 3 Day 4
B.+C. B.+C. B.+C. B.+C.
1 ! A l:LA l:LA 1 *• A
2 Ai 2 Ai 2 Ai 2 Ai
950 190 127 350 30 390 43.5 120
360 190 83 350 50 390 78 120
32.5 190 127.5 350 19 390 - 120
t_ n5(2)=4.303 tn n,(2)=4.303 tn n,(2)=4.303 tn (1)=12.706
\) *\J U • \J J U • U D v»UJ
0.959 -16.10 -39.34 -3.43
            Where A is upwind - Site 1;  B is downwind - Site 2; C is downwind - Site 3.

-------
     For the t-test on population exposure F = the population station and
A = the upwind station, F and B represent the PCB levels at Stations F and
A for days 1, 2, 3, and 4.

     The t statistic for this t-test is:
                                        4

                                     4
                              S. 2
                               D         2

                                  D. = F. - A±


     Using the measured ambient PCB concentrations listed in Table 5, the t
statistic results are shown in Table R-2.

                   TABLE R-2.  t-TEST FOR UPWIND VERSUS
                               POPULATION SITES

Day 1
Day 2
Day 3
Day 4

F
63
67
28
51
t = 310
A
190
350
390
120

Where the critical value   of student's t distribution is to.05(3) = 3.182.

     From the t-test results, HQ is accepted which shows that there is no sig-
nificant difference in PCB concentration between the upwind and population
sites.  This conclusion is evident from the values being the opposite of the
original hypothesis.  Therefore, you can not presume any positive change in
PCB concentration in ambient levels as you have mathematically disproved this
as being possible.
                                     159

-------
                                 GLOSSARY
Aroclor 1242:   Commercial mixture of PCBs manufactured in the United States
     by Monsanto.  The digits "12" refer to biphenyl as the parent compound;
     the digits "42" indicate that the chlorine content of Aroclor 1242 is
     approximately 42 percent by weight.  Aroclor 1242 was the PCB mixture
     burned by General Motors and was the reference standard for PCB
     measurements.

M-5-1:  Three emission test runs conducted to measure hydrogen chloride,
     dibenzofuran, chlorinated dibenzofurans, chlorinated dioxins and or-
     ganics released during the PCB burn.  M-5-1 was conducted as a back-
     ground run with no PCB contaminated oil fed to the boiler.  M-5-2 and
     M-5-3 were conducted during the PCB run.  Particulate and organic vapors
     from all three runs was analyzed for the compounds cited.

Modified Method 5 sampling train:  The sampling train used for test runs
     M-5-1 - M-5-3.  This train is similar to a standard EPA Method 5 par-
     ticulate train except for the addition of a temperature controlled
     adsorbent column of XAD-2 resin between the first and second impingers.
     This column was used to collect organic emissions.  The contents of
     the 4 glass impingers following the particulate filter are also dif-
     ferent from the Method 5 train.  (Refer to Section 4).

PCB-1 - PCB-5:  Five emission test runs conducted to measure stack gas PCB
     emissions.  PCB-1 was a background run with no PCB contaminated oil fed
     to the boiler.  PCB-5 was intended as a valid test run, but the sample
     was inadvertently spilled and the emission data was not used.  PCB-2,3
     and 4 were the three valid test runs used to measure PCB destruction
     efficiency.

PCB sampling train:  The sampling train used for test runs PCB-1 - PCB-5.
     This train differs from a standard Method 5 train in that (a) there is
     no particulate filter and (b) a Florisil cartridge is inserted between
     the third and fourth impingers to collect PCB emissions.  (Refer to
     Section 4).

Spiked waste oil:  The PCB contaminated oil used for the GM test burn.  As
     received from GM's wastewater treatment plant, the waste oil contained
     132.5 ppm of PCBs.  This oil was "spiked" for the destruction efficiency
     tests by adding Aroclor 1242.  Spiking increased the PCB concentration
     to 500 ppm and thereby improved the accuracy of the PCB destruction
     efficiency determination.  (See Section 3).
                                    160

-------
                                TECHNICAL REPORT DATA
                         (Please read Inunctions on the reverse before completing)
 . REPORT NO.
 EPA-600/2-81-055a
                           2.
                                                      3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation of PCB Destruction Efficiency in an
 Industrial Boiler
                                 S. REPORT DATE
                                 April 1981
                                 «. PERFORMING ORGANIZATION CODE
7. AOTHOR(S)
         J. Hall, F. Record, P. Wolf, G. Hunt, and
         S. Zelenski
                                                      8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
GCA/Technology Division
213 Burlington Road
Bedford, Massachusetts  01730
                                  10. PROGRAM ELEMENT NO.
                                  C1YL1B
                                  11. CONTRACT/GRANT NO.

                                  68-02-3168, Task 9
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC  27711
                                  13. TYPE OF REPORT AND PERIOD COVERED
                                  Task Final; 2-11/80	
                                  14. SPONSORING AGENCY CODE
                                   EPA/600/13
is. SUPPLEMENTARY NOTES IERL_RTP project officer is David C. Sanchez, Mail Drop 63,
919/541-2547. Report "b" of this series is an audit report.
is. ABSTRACT rp^ rep0rt describes the evaluation program undertaken to determine the
poly chlorinated biphenyl (PCB) destruction efficiency during a May 1980 verification
co-firing of waste oil containing approximately 500 ppm of PCBs,  in accordance
with applicable state and federal regulations,  in a high-efficiency industrial boiler
owned and operated by  General Motors Corporation at Bay City, MI.  Also investi-
gated was the environmental and workplace impact which occurs during the handling
and combustion of PCB-contaminated waste oils. (According to EPA's final ruling on
the disposal of PCBs, waste oils containing PCBs in the 50-500 ppm range can be
fired with fuel oil and burned in a high-efficiency industrial boiler.)  No PCBs were
detected in the stack gas within the detection limits of the sampling and analytical
techniques used. Data collected during the verification burn indicate that, by follow-
ing the equipment and performance requirements given in EPA's final PCB rule
(44, Federal Register, 31545, May 31, 1979),  at least  99.9% destruction efficiency  of
PCBs  can be achieved by high efficiency boilers. Furthermore,  monitoring of the
downwind ambient air,  the workplace environment, and employee blood levels has
indicated that PCB destruction can be conducted with no measurable  adverse  effect
on either the workplace environment or the external environment.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                     b.IDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Croup
 Pollution
 Chlorine Aromatic
  Compounds
 Biphenyl
 Materials  Handling
 Combustion
Boilers
Oils
Fuel Oil
Wastes
Flue Gases
Pollution Control
Stationary Sources
Polychlorinated Biphe-
  nyls  (PCBs)
Waste Oil
13 B
07C
13 H
21B
13A
11H
2 ID
14H
13. DISTRIBUTION STATEMENT

 Release to Public
                      19. SECURITY CLASS (This Report)
                      Unclassified
                         21. NO. OF PAGES
                            171
                      20. SECURITY CLASS (Thispage)
                      Unclassified
                                              22. PRICE
EPA Form 2220-1 (t-73)
                                        161

-------