United States
Environmental Prot>- '
Agency
Industrial Environmenta :    EPA 6(X.
Laboratory             • /9
Research Triangle Park NC 2771 1
Measurement of PCB
Emissions from
Combustion Sources

Interagency
Energy/Environment
R&D Program Report

-------
                  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 INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND  DEVELOPMENT series. Reports in this series result from the
effort funded  under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants  associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of. control  technologies for energy
systems; and integrated assessments of a wide-range of energy-related environ-
mental issues.
                       EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, 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-600/7-79-047

                                           February 1979
Measurement  of PCB  Emissions
     from  Combustion Sources
                          by

              P.L. Levins, C.E. Rechsteiner, and J.L. Stauffer

                     Arthur 0. Little, Inc.
                        Acorn Park
                 Cambridge, Massachusetts 02140
                    Contract No. 68-02-2150
                        T.D. 10102
                   Program Element No. INE624
                 EPA Project Officer: Larry D. Johnson

              Industrial Environmental Research Laboratory
                Office of Energy, Minerals, and Industry
                 Research Triangle Park, NC 27711
                       Prepared for

              U.S. ENVIRONMENTAL PROTECTION AGENCY
                 Office of Research and Development
                    Washington, DC 20460

-------
                      TABLE OF CONTENTS

                                                       Page


      LIST OF FIGURES	    iv

      LIST OF TABLES	     v

  I.   INTRODUCTION 	     1

 II.   BACKGROUND   	     2

III.   DEVELOPMENT OF AN ALTERNATIVE PCB ANALYSIS
        METHOD 	     4

      A.  Discussion of GC/MS Method   	     4

      B.  Selection of MS Data Acquisition Masses  .     5

      C.  GC Retention Time Criteria	     7

      D.  Quantitative Calibration   	    11

 IV.   EVALUATION OF THE PCB ANALYSIS PROCEDURE ...    17

      A.  Calibration Procedure Evaluation   ....    17

  V.   VERIFICATION OF METHOD   	    26

      A.  Evaluation of Test Samples	    26

      B.  Recovery from Flyash	    30

 VI.   CONCLUSION	    38

VII.   REFERENCES   	    39




          Appendix A   	A-l

          Appendix B   	B~*

          Appendix C   	C-1
                             iii

-------
                       LIST OF FIGURES


Figure No.                                                 Page

   1       Hypothetical Mass Spectra of PCB's 	    6

   2       GC/MS Subset Mass Chromatogram of
             Aroclor 1248	10

   3       Analytical Mass Chromatograms for an
             Aroclor 1248 Sample	12

   4       GC/MS Subset Mass Chromatogram of
             Aroclor 1242 Standard	19

   5       GC/MS Subset Mass Chromatogram for
             Aroclor 1254 Standard	   20

   6       GC/MS Subset Mass Chromatogram for
             Aroclor 1260 Standard	21

   7       GC/MS Subset Mass Chromatogram for
             Waste Extract Mixture	28

   8       a) GC/MS Subset Mass Chromatogram,
             b-e) Single Ion Chromatograms for
             Analytical Ions of Aroclor 1260 (326,
             362, 394 and 428) for Waste Extract
             Mixture Dosed with Aroclor 1260.  Solid
             Lines Delineate Relative Retention Time
             Windows Used   	29

   9       a) GC/MS Subset Mass Chromatogram,
             b-g) Single Ion Chromatograms for the
             Analytical Ions of Aroclor 1242 (188,
             224, 258, 292, 326) for Waste Extract
             Mixture Dosed with Aroclor 1242.  Arrows
             Delineate the Relative Retention Time
             Windows Used   ,	32

  10       a) GC/MS Subset Mass Chromatogram,
             b-d) Prominent Ion Chromatograms for
             Analytical Ions 292, 326, 362 of Ferro-
             Alloy Smelter Sample Dosed with Aroclor
             1254.  Arrows Delineate Relative Retention
             Time Windows Used	33

  11       GC/MS Subset Mass Chromatogram for Extracted
             Flyash Sample Dosed with Aroclor 1254 (a)
             and Aroclor 1254 Standard  (b)	36
                                iv

-------
                         LIST OF TABLES


Table No.                                                  Page

   1       Mass Spectrometry Data Acquisition
             Subsets  	    7

   2       Gas Chromatography RRT Windows Relative to
             p,p'-DDE = 100   	    9

   3       Aroclor 1248 Composition Analysis  	   13

   4       Possible Aroclor Calibration Standards ....   14

   5       PCS GC/MS Sensitivity Data:  GC Areas  ....   16

   6       GC/MS Conditions 	   22

   7       Analysis of Aroclor 1242 .	   23

   8       Analysis of Aroclor 1254	   24

   9       Anaylsis of Aroclor 1260	   25

  10       Organic Mixture Composition	   27

  11       Analysis of Mixture of Wastes and Waste
             Extract	   31

  12       PCB Analysis of Ferroalloy Sample	   34

  13       PCB Analysis for Recovery of Aroclor 1254
             in Dosed Flyash	   37

-------
                             I.  INTRODUCTION


     The purpose of the contract award to Arthur D. Little, Inc.,  (EPA
Contract No. 68-02-2150) by the Process Measurements Branch of IERL/RTP
is to provide advanced organic sampling and analysis capabilities and
method development.  The effort covered by the contract has been divided
into several areas, including the development of procedures for more
complete organic analyses which will be required in Level 2 (Technical
Directive 10202).  As part of the effort in this category, EPA requested
an examination of the methods used for polychlorinated biphenyl (PCB)
analysis that would provide reliable data when applied to emissions from
combustion sources.  Verification of a recommended PCB procedure was
part of the request.

     In preparation of this report, three areas of effort have been de-
fined. 1) Pertinent publications have been reviewed to allow drafting of
a tentative PCB analysis procedure.  2)  Laboratory experiments have
been conducted to test the tentative procedure using well characterized
reference materials.  3)  Verification of the PCB analysis procedure has
been made using complex organic samples, representative of the types of
samples expected in environmental assessment studies.  The recommended
procedure for PCB analysis is described in detail in Appendix A.

-------
                            II.  BACKGROUND

     The analysis and measurement of PCB emissions from combustion
sources encounters problems not dealt with in most reported PCB analy-
sis methods.  Many reviews d"1*) have been published which present dif-
ferent analytical methodologies developed for PCB analysis and present
a good understanding of the PCB analysis problem.  Of the methods typi-
cally utilized for PCB analyses, gas chromatography has been widely
implemented as a low cost, yet sensitive analytical technique.

     Gas chromatographic analyses for PCB can be classified as using
either a pattern recognition approach or the measurement of individual
PCB peaks.  The EPA Federal Register Method for PCB analysis in indus-
trial effluents^5) (Vol. 38, No. 78, pt. II) uses the pattern recogni-
tion approach utilizing conventional GC techniques with an electron cap-
ture detector.  In this method the gas chromatogram of the sample is
examined for similarity to one of the known Aroclors, after a series of
clean-up steps, as appropriate for each sample.  If a match is made, the
PCB content is reported as the suspected Aroclor.  The several difficul-
ties associated with the pattern recognition approach have been recog-
nized by researchers conducting this type of analysis.  In particular,
interferences introduced by co-eluting pesticides are especially common
for water and sediment samples.  Elemental sulfur and other common
species such as phthalates are common interferences.
     The quantitative method proposed by Webb and McCall   ' was partially
incorporated into the Federal Register method as being recommended for
use only if the pattern recognition method seems to be nonapplicable.
The method of Webb and McCall measures each individual PCB peak, obvia-
ting the need for prior assessment of which PCB is present to determine
the proper calibration material.  Even if there are interferences in the
chromato grams, use of individual peaks minimizes the errors caused by
non-PCB species.

     Most of the EPA work on PCBs to date, has been on water and sediment
samples.  As these methods have been applied to samples from combustion
sources, a variety of problems have been encountered.  Samples from coal-
fired power plants and incinerators have been difficult to analyze for
PCBs.  One of the problems Is that the PCBs exposed to a combustion pro-
cess have a substantially different pattern than the known Aroclor PCBs.
Most of these samples show a loss of the low end PCBs (mono-, di, and
trichloro) with a residual of the higher molecular weight PCBs (tetra-,
penta-, and hexachloro)   .  In addition, the nature of the GC inter-
ferences is different in these combustion effluent samples than In the
water and sediment samples.

     To simplify the GC/ECD measurement procedure and increase the detec-
tion limit sensitivity for PCB1, Haile(7a) and Armour  C8) have proposed
perchlorination of the PCBs using antimony pentachloride.  This proce-
dure converts all of the individual PCB species to a single species,
decachlorobiphenyl (DCB) .  Measurement of a single species can be a

-------
great improvement over the analysis of a complex mixture.   However,
Armour cautions that the perchlorination method is only for confirmatory
purposes and to enhance the sensitivity of the analysis.  In the Armour
procedure, the conventional pattern recognition approach must be used to
establish the identity of the sample prior to DCB formation.  Since dif-
ferent amounts of DCB are produced from the individual PCBs, adjustment
factors ranging from 0.4 to 0.8 are required to convert DCB values to
quantities of specific Aroclor mixtures, and ultimately the accuracy of
the PCS measurement depends upon correct identification of the PCB
present*

     Application of the perchlorination procedure for source measure-
ment of PCBs has been made to incinerator emissions  (7a~9', with the
procedures described in detail in EPA document EPA-600/4-77-048^7a).
After perchchlorination, DCB was detected indicating the presence of
PCBs.  However careful examination by GC/MS of aliquots of the original
samples, blanks, and incinerator fuels  (10) showed that no PCBs were
present.

     Because of the false positives problem and the difficulties in
accurate quantitative calibration, the perchlorination  (DCB) procedure
is not sufficiently reliable to warrant use in measuring PCB emissions
from combustion sources.  The Federal Register procedure for PCB mea-
surement in industrial effluent is not applicable because of the changes
in relative composition during combustion.  Thus a new procedure was
needed for combustion sources.

-------
             DEVELOPMENT OF AN ALTERNATIVE PCS ANALYSIS METHOD


A.  Discussion of GC/MS Method

     Automated gas chromatography/mass spectrometry  (GC/MS) appeared  to
offer the greatest potential for a sensitive quantitative analysis method
that would provide reliable PCB concentration data from the complete
range of environmental samples, including those from combustion sources.

     Since chlorine in natural abundance exists as approximately a 3/1
ratio of 35C1/37C1, the Isotope clusters produced in the mass spectra
of PCBs provides a unique opportunity to both confirm identity and make
quantitative concentration measurements.  The GC/MS  method developed  by
DudenbostelC12) and used in the EPA Region II laboratories makes use  of
these factors.  In their method, GC/MS data are collected in narrow mass
ranges corresponding to the molecular ion clusters of the mono-, di,  etc.,
chlorobiphenyls.  Selected spectra throughout the chromatogram are ex-
amined to verify that the data collected do correspond to PCBs according
to the isotope patterns expected in each mass range.  The reconstructed
chromatogram is then pattern-matched to one of the Aroclors.  Quantita-
tive measurement is based upon the ratio of the most intense peaks in
the sample and standard.  The work by others, and in this report, is  con-
centrated on the Aroclor PCBs, because they are the  only PCB's that are
found with regularity in environmental studies in the United States.

While the Dudenbostel method^11' has many of the desirable specificity
features inherent in GC/MS,  it still relies upon pattern recognition  and
analysis of PCB as one of the Aroclors.   This approach would not be ac-
ceptable for combustion source samples.

     The paper by Eichelberger, Harris and Budde'12' lays the groundwork
for an alternate approach using GC/MS.  The essence  of their method is
to use PCB subset mass scanning with a particular mass chosen for each
of the monochloro	heptachlorobiphenyl groups.  They selected particu-
lar masses for each chlorobiphenyl so that there would be minimum overlap
between chloro groups.  The approaches to using the method quantita-
tively were only lightly touched upon in that pap-  .  Quantitative
analysis was by pattern recognition and measurements as one of the Aro-
clors.  Interference by pesticides, etc., was minimized by selection  of
the subset masses.  Although not stated explicitly in the paper, this
method clearly leads to the possibility of reporting PCBs in terms of
the amount of each chlorobiphenyl group (mono-,  di-, etc.) present.
This kind of approach would overcome the difficulties found in combustion
source samples where the original Aroclor distribution patterns are
altered.

     The approach developed in this study is based upon the excellent
groundwork developed by Webb and McCall<5),  DudenbostelCIL)
Eichelberger, Harris and Budde<12>.  The essence of the analysis method
is as follows:

-------
     1)  Acquire GC/MS data in PCB subset mass windows large enough
         to encompass all of  the  isotope cluster(s).

     2)  Examine selected mass spectra  to verify  PCBs by  their
         chlorine  isotope abundance  patterns.

     3)  Generate  mass chromatograms from a  single  mass chosen  to
         represent each  chlorobiphenyl.

     4)  Integrate the areas  of each mass chromatogram only  in  the
         relative  retention time  (RRT)  region  corresponding  to  the
         mono-, di, etc., chlorobiphenyls.

     5)  Quantify  either from selected  peaks in Aroclor reference
         standards or with pure chlorobiphenyl isomers.   The details
         of the quantitative  calibration are the  subject  of  follow-on
         studies, discussed in section IV.  A complete description  of
         the  analytical  procedure is found  in  Appendix A.

B.  Selection of MS Data Acquisition Masses

    The previous EPA workX11*12)  has clearly shown  the increased sensi-
tivity to be gained for PCB analysis by using  a selected  subset  masses
for each GC/MS scan.  The question then becomes one of how to select
the correct mass or mass range for each chlorobiphenyl.   We  feel that
one must collect data over the full  isotope  cluster range in order to
be able to verify  the PCB composition by means of the isotope abundance
pattern.  The book by Safe and Hutzinger'13' provides a great deal of
information on the mass spectra of PCBs.  Significant (>  5%) 37C1  iso-
tope peaks are found as  follows for  the various PCB groups.

                   PCB Group            Significant  Ions

                    Cli                    M, +2
                    C12                    M, +2, 44
                    C13                    M, +2, +4
                    Cljt                    M,  42, 44
                    C15                    M,  +2, 44, 46
                    C16                    M,  +2, 44, 46
                    C17                    M,  +2, 44, 46, +8
                    C18                    M, 42, 44, 46, 48
                    C19                    M,  42, 44, 46, 48
                    C110                   M,  42, 44, 46, 48, 410

     The initial data acquisition subset then  should at least include
most of these ions.  The mass spectrum  of a  chlorobiphenyl is dominated
by mass clusters at regions corresponding to the  molecular ion  (M*") and
M*--C1, M*--HC1, M*--C12-   The K^-Cl and M^-HC1 clusters usually have only
about 10% relative intensity, whereas the itf" and  lf*"-Cl^ ions are of about
equal intensity.   The WT*"-C1 ion will be offset one  mass lower than the
corresponding PCB  isomer, while the  M*"-HC1 and M^-ca.2 ions will be off-
set by two mass units from their  corresponding PCB  isomer.  Simplified

-------
180
          188           224
                                            Analytical m/e's
                                          2S8             292
                                                                          326
                                                                                             362
                                                                                                             394

1
1



fl *
1
, 1
1 1
h




,
1
1 1
II




l .
I
1 1
il
ii
-------
 spectra of the PCB groups containing one to seven chlorines are shown
 in Figure 1.  In each case only the ion clusters due to M+, M^-C1 and
 M+-C12 are shown,  M+ and M+-C12 are shown as equal intensity and M+-C1
 is shown at 50% intensity.  From an examination of these data, a  sub-
 set mass range was chosen for each PCB group as shown in Table 1.  The
 ranges were chosen to include the fragment ions from higher PCBs and
 enough of the isotope cluster to allow  unambiguous identification of
 the PCB.

      The analytic method described in this report is applicable to
 PCBs containing from one to ten chlorines.  The samples analyzed during
 the verification phase of this project were spiked with PCBs containing
 no more than 8 chlorines.  In the test samples, PCBs with higher chlo-
 rine content were not found and, therefore, are not discussed in the
 following sections.

      Once the PCB identity has been confirmed, there are a variety of
 ways of treating the data to obtain a quantitative measure of the PCB
 groups.   A single mass has been chosen which represents the most in-
 tense mass in the molecular ion cluster with the minimum fragment ion
 interference.  When the GC retention window criteria are superimposed,
 as discussed in the next section, it is only necessary to eliminate
 (or minimize) the interferences from the one higher chloro group (i.e.,
 tf^-Cl.  The analytical masses felt to represent the best choices are
 shown in Figure 1 and are listed in the last column of Table 1.


 C.  GC Retention Time Criteria

     The M*-Cl2 fragment  ion in the mass spectra of chlorinated biphenyls
 could represent a difficult problem in the quantitative analysis of PCB
 groups.   The results of the careful work by Webb and McCallT6"*1*', how-
 ever, present a means for overcoming this problem.  A careful examina-
 tion of their data reveals that there is no GC overlap between chlori-
 nated biphenyls differing by two chlorines and only slight overlap by
 biphenyls differing by one chlorine.   The single chlorine overlap does
 not present a serious problem for the GC/MS analysis, because the M+-C1
 or -HC1 peaks are generally of only  10% relative intensity,  and the
 analytical masses can be chosen to minimize this problem.

       From  the Aroclor composition and GC  data given by Webb and McCall(6)>
 a tentative set of relative retention time (RRT) windows have been
 chosen for each PCB group as shown in Table 2.

     The PCB analysis procedure would then be  to obtain the GC/MS data
in the subset mass ranges  for  the entire chromatogram.  After PCB identi-
ties are confirmed by isotope  ratio checking, mass chromatograms would
be obtained for each analytical mass.  The area in the proper RRT window
would  then be integrated  for a measure of each PCB chloro group.

     An example of this procedure can be seen from some preliminary work
conducted in the ADL laboratories.  Figure 2 shows the reconstructed
chromatogram obtained from a GC/MS run using the subset mass ranges

-------
                             TABLE 1



              Mass Spectrometry Data Acquisition Subsets


                                               Total       Analytical*
PCB Cluster        H*        m/e Range          m/e            m/e


                                                 5             188

                                                 7             224

                                                 7             258

                                                 7             292

                                      **         6             318

                                      **         7             326

                                                 9             362

                                                 9             394

                                                 9             428

                                                 9             464

                                                11             498
Cli
C12
C13
Clj,
p,p'-DDE
C15
C16
C17
C18
Clg
Clio
188
222
256
290
316
324
358
392
426
460
494
186
220
254
288
316
322
356
392
426
460
494
- 190
- 226
- 260
- 294
- 321'
- 328'
- 364
- 400
- 434
- 468
- 504
    40 m sec integration time per m/e.
  **
    p,p*-DDE (l,l-di[p-chlorophenyl]  dichloroethylene) is used as
    internal standard and its mass range may be combined with that
    for pentachlorobiphenyl to meet the constraints of the data
    system.

-------
                          TABLE 2


  Gas Chromatography RRTWindows Relative to p,p'-DDE* = 100


                     Relative Retention
PCH Cluster           Time (RRT) Window           Analytical m/e


     Clj                 0(5) - 20                      188


     C12                   15   35                      224
C13
cu
cis
C16
C17
C18
C19
ciio
25
40
70
125
160
275
400
650
- 55
- 100
- 150
- 250
- 350
- 600
- 1000
- 1200
258
292
326
362
394
428
464
498
 p,p'-DDE (l,l-di-[p-chlorophenyl[dichloroethylene) is used as
 internal standard.

-------
RROCLOR  1248.  60  NG/UL, OV-1,  185  ISO,  SMS
100
              70
                     100       150
  FIGURE 2:  GC/MS Subset Mass Chromatogram of Aroclor 1248
        RRTs for specific peaks labeled above peak
                     10

-------
 chromatogram obtained from a GC/MS using the subset mpss ranges
 given in Table 1.   The RRT's were assigned bv comparison of this chroma-
 togram to that published by Webb and McCall^ '.  The chromatogram rep-
 resents a 60 ng injection analyzed on a 2 m OV-1 column operated iso-
 thermally at 185°C.  The mass chromatograms constructed from each of
 the analytical masses are shown in Figure 3.  Each chromatogram is self-
 normalized to 100.

      The appropriate RRT windows for each chlorobiphenyl group are shown
 on their respective mass chromatograms.  One can see that using this
 procedure eliminates the overlap problem from fragment ions appearing
 in a particular mass chromatogram, due to another chlorobiphenyl.  For
 instance, the M-Cla peaks due to trichloro and tetrachlorobiphenyl,
 respectively, are seen in the 188 and 224 mass chromatograms, but the
 RRT window criteria eliminate these interferences.

      These windows may have to be adjusted alightly for some samples
 in order to properly integrate the GC areas.  The PCS isotope pattern
 criteria should be used in making these adjustments.


     The  chromatographic peak area  found  for each  PCB  group  of  this
Aroclor 1248 sample within the designated relative  retention  time window
are listed in Table 3.  Comparison of the area  % data  found  for this
test sample with  the weight %  data of Webb and  McCall^6)  for  a different
Aroclor 1248 sample shows excellent agreement,  using the  assumption that
each PCB cluster has the same weight/response sensitivity.  The close
agreement shown In Table 3 may be a fortuitous  concidence in  view of the
fact that two different samples of Aroclor 1242 are  being compared.  How-
ever, the closeness of the data suggests  that relatively  simple calibra-
tion procedures may yield accurate quantitative determinations.
D.  Quantitative Calibration

    Several approaches are possible for quantitative calibration of the
different PCB groups.  For the simplest case, if one assumes that each
of the different PCB clusters has the same weight/response sensitivity,
then a single PCB species would allow calibration of all of the PCB
isomers found in a sample.  A slightly more complex method would in-
volve preparation of a calibration mixture containing a single isomer
for each PCB group.  This method assumes that all of the isomers within
a single PCB group have identical weight/response sensitivities.
A third, readily available method would be to use previously calibrated
Aroclor reference standards, such as those evaluated by Webb and McCall
for the GC/ECD analysis method, under conditions Identical to those
used with samples.

     Since the PCBs are always found as complex isomer mixtures and the
various Aroclors have similar complexity, the third method offers self-
correcting advantages not available with the other two methods.  Selec-
tion of either one or two isomer peaks for each chlorobiphenyl group
which are well defined by GC, have no other chlorobiphenyl overlap and
                                   11

-------
            100
100
IS)
          t

          I
             Cl,
                             MflSS 188
             r i Ji i i i i' I I [ ' I M I I ' I ' | ' I ' I i I M I | I I

                     50      100      150
                 MRSS 224
                                             CL2
                                                RRT

                                                36
          I i| I | I | I I I [ I I I I ' I I I ' | ' I

          50     100      ISO
          I
          1
                      50     100
                         M/e
                                                                                          MflSS 258
                                        50  ,  100      150
                                           Mr*
                         FIGURE 3  ANALYTICAL MASS CHROMATOGRAMS FOR AN AROCLOR 1248 SAMPLE

-------
                               TABLE 3
                  Aroclor 1248 Composition Analysis
PCB Group
Clj
C12
C13
CL,
C15
Cls
m/e
188
224
258
292
326
362
RRT Window
1-20
15-35
25-55
40-100
70-150
125-180a
Area
—
1452
17423
36896
11629
0
%D
(0)
2
26
55
17
0
Known %
0
1.2
24.7
57.8
19.8
0.4
a.
b.
   Limit of chromatogram

   Percentage each group was off the sum of the individual group
   areas assuming that each PCB group has the same sensitivity.

c" Webb and McCall (reference 6, Table 4).
                                    13

-------
                             TABLE 4


               Possible Aroclor Calibration Standards

                                                       % RRT peak
PCS Cluster        Aroclor         RHT Peak8         is of Aroclor

    Cl!              CBP*              11                  100

    C12             1242             20, 21                 11.3


    C13             1242             37, 40                 22.6


    Cl^             1254               47                    6.2


    C15             1254               84                   17.3


    C16             1254              174                    8.4


    C17             1260              280                   11.0

    C18               c

    C19               d


    C110             DCB**          1,100                   100
  Relative to p,p'-DDE = 100

b Webb and McCall (Ref. 6) (Appendix B)
c
  Octachlorobiphenyl  may be calibrated from Aroclor 1260 although
  it is present at about 1% by weight.  As an alternate method,
  calibration against a pure Octachlorobiphenyl is recommended.

  Nonachlorobiphenyl is not found in common Aroclors; calibration
  against pure nonachlorobiphenyl is recommended.  If nonachloro-
  biphenyl is unavailable, assume equal sensitivity between octa-
  chloro and nonachlorobiphenyl.
**
  Decachlorobiphenyl
                                 14

-------
whose abundance  is  reasonably high, or by using the entire envelope of
each PCB group, allows  calibration  under conditions of sample complexity
similar to that  expected  in actual analyses, as well as allowing routine
monitoring of  all experimental equipment.

     Table 4 lists  a set  of peaks  abstracted from the work of Webb and
          which  could  be  used for  PCB calibration.  Other sets of peaks
could be chosen  for calibration purposes, or even the entire RRT window
for each PCB group, as noted previously. Use of the entire RRT window
affords optimum  self-correction by averaging the individual chlorobi-
phenyl isomer  sensitivities, which may be of significance if the indi-
vidual isomer  sensitivities show large variations.  The other previously
proposed methods make  the tacit assumption  that the individual isomer
sensitivities  do not vary too greatly.  Use of a calibration based upon
the complete RRT window for any chlorobiphenyl group gives directly an
ensemble averaged calibration factor which  should accurately describe
the average GC/MS system  response  to each chlorobiphenyl group.

     Preliminary study of some individual chlorinated biphenyl isomers
have been conducted to evaluate the validity in GC/MS response between
isomers.   A selection  of  pure compounds was obtained from RFR Corpora-
tion (1 Main Street, Hope, Rhode Island 02831).  Certain of these were
selected on a  random basis within  PCB groups for initial study.  Solu-
tions were prepared in benzene containing 5r-15 ug/mL.  A 2-yL aliquot
of each solution was analyzed by GC/MS using a Finnigan 4000 with their
6100 data system.  An  OV-1 glass column was used and operated isother-
mally at either  185°C  or  200°C depending on the isomer groups.   A single
injection of each solution was made, and the data for each compound were
tabulated for  comparison.  A set of subset masses slightly different
from that recommended  in  the proposed method was used with an integra-
tion time of 20  msec/mass.   The PCBs have a significant mass defect
(exact masses  less  than nominal mass) and the spectrometer was not tuned
to compensate  for this fact.  Thus, the data obtained are in some cases
from the sides of mass peaks and are not the optimum achievable.

     The GC area response obtained for each of the isomers studied is
shown in Table 5.  For this small  set of data, standard deviations be-
tween 12% and  37% were found.  Also, the average area/ng responses for
the different  chloro groups vary by up to a factor of 2.  As an extreme
case, consider the  comparison of the tetrachlorobiphenyls.   The
area/ng sensitivities  for these species vary by a factor of 3.   There-
fore any method  which  is  based upon a single calibrating isomer is poten-
tially subject to sizeable error.  Of the alternatives presented for
calibration, a calibration based upon the complete envelope of peaks
for each chloro-containing group seems to be the most attractive.   The
procedure is essentially  the same  as that recommended by Webb and McCall
for the GC/ECD analysis of PCBs.   Since the complexity of the calibration
samples approximates that found in environmental samples,  the error due
to sensitivity differences between isomers Is minimized.

-------
                                            Table 5
                            PCB GC/MS Sensitivity Data;  GC Areas
RPC-27
RPC-30
RPC-33
RPC-38
RPC-41
RPC-47
RPC-51
Conq> ound
2 , 3 , 5-trichlorobiphenyl
2 ,2', 5-trichlorobiphenyl
2 ,4 , 5-trichlorobiphenyl
2,3,4, 5-tetrachlorobiphenyl
i i
2,2,4, 5-tetrachlorobiphenyl
2 , 3*, 4 i 5-tetrachlorobiphenyl
2 , 2', 3 , ' , s'-pentachlorobipheny 1
2, 2*, 3', , 6-pentachlorobiphenyl
2 , 2*, 4 , 4', 5 , S'-hexachlorobiphenyl
2 , 2*, 3 , 5 , 5', 6-hexachlorobiphenyl
Quantity (ng)
27.7
29.5
28.0
28.4

24.0
8.32
9.40
12.06
19.26
18.20
GC Area
206991
261941
191635
268218

284449
158440
135933
208174
167214
112066
 avg. 7730
   ±  1040 (13%)

      9400
     11900
     19000
avg. 13400
  ±   5000 (37%)

     14500
     17300
avg. 15900
  ±   2000 (12%)

      8700
      6200
 avg. 7400
   ±  1800 (24%)

-------
               IV.  EVALUATION OF THE PCB ANALYSIS PROCEDURE


     In the previous section, calibration of the different chlorobi-
phenyl groups  using the entire RRT window for  that group  from a well
defined Aroclor reference material was  recommended as the method of
choice.  That  procedure was evaluated in terms of its ability to give
consistent results over a period of time.

A.  Calibration Procedure Evaluation

     To evaluate the calibration procedure, samples of  three Aroclor
reference materials, Aroclor 1242, Aroclor  1254, and Aroclor 1260, were
obtained from  the batches analyzed by Webb^6^.  The solutions were ob-
tained at a concentration of 1 mg Aroclor per  1 mL of solution in iso-
octane.  Aliquots of each of the three  Aroclor reference materials were
diluted to 20  yg/mL and p,p'-DDE was added  to  each to a level of 1 ug/mL.
Each Aroclor was then analyzed a minimum of six times over a period of
one week by the recommended GC/MS procedure using the conditions listed
in Table 6.  After the completion of a  run, the ion chromatograms for
each of the analytical ions and the internal standard were obtained.

     Overlapping the ion chromatograms  of PCBs differing by two chlorines
(e.g., M and M+-2C1) serves to clearly delineate the retention time win-
dows.  For cases where the PCB group containing one more chlorine (MfCl)
is at a much higher concentration than  M, the  mass spectra of the peaks
in the ion chromatogram is used to pinpoint the location of RRT window
for M.  The total ion chromatograms for the different Aroclors, Figures
4-6, show a pattern for these three Aroclors similar to those found by
GC/ECD methods.  The results for the different Aroclors are shown in
Table 7 for Aroclor 1242, Table 8 for Aroclor  1254, and Table 9 for
Aroclor 1260.

     The  relative  standard deviations  for the analysis  of each  of  the
Aroclor reference materials  show excellent  agreement.   A majority  of
the  deviations are  less  than 3%  and all are below 8%.

     One  may assess  the validity of the equal weight/response  sensiti-
vity method of calibration discussed in the previous  section from
Tables  7,  8,  and 9.

      For  some of the PCB chloro  group,  typically the major component(s),
the  agreement between the known  and found results was  very good.   How-
ever,  the  chloro groups  present  at less than 20% by weight can show
considerable error.   Thus,  using one group of chloro-containing PCBs,
such as the tetrachloro  biphenyls,  for calibration of  the different
chlorobiphenyls may  give acceptable results.

      The  method detailed in Appendix A, which uses calibration of  the
specific  chloro groups against a well characterized reference  standard
will not  be subject  to the errors observed when the equal sensitivity
assumption is used.   With this recommended analysis method more


                                 17

-------
accurate concentration data can be obtained for a small increase in
analysis time.  For the remaining sections, the recommended PCB pro-
cedure was used exclusively.
                               18

-------

I r | i i i i r I > i i i i i i  i i I i i r i i i i i i n i i i i i | i i i i M i i i I i  i i  i I r 1 i \ i  i i i i
     I > i i i i i i i i I
              too       iso      2oa       250       300
FIGURE 4  GC/MS SUBSET MASS CHROMATOGRAM OF AROCLOR 1242 STANDARD

-------
to
o
                                  IOQ       ISO       2QQ       250      3QQ
                   FIGURE 5  GC/MS SUBSET MASS CHROMATOGRAM FOR AROCLOR 1254 STANDARD

-------
            100
                     50
200
250
3DD
350
400
450
N>
            100
            500     550      6DO  " " 650	?b'o	750     800



                        FIGURE 6    GC/MS SUBSET MASS CHROMATOGRAM FOR AROCLOR 1260 STANDARD

-------
                           TABLE 6


                       GC/MS Conditions


 I.  Gas Chromatographic Conditions

     a)  Finnigan Model 9610 GC

     b)  6-ft glass column packed with OV-1 coated
         Supelcoport 100/120.

     c)  Multilinear temperature program

         1)  isothermal program at 185°C for 30 mln.

         2)  linear program from 185°-300° at 258C/min.

         3)  isothermal program at 300°C for 25 min.

     d)  2-3 yL injections



II.  Mass Spectrometric Condition

     a)  Finnigan Model 4000 mass spectrometer

     b)  mass range - 186-190, 220-226, 254-260, 288-294,
           316-328, 356-364, 392-400, 426-434

     c)  integration time    - " 50 msec/amu

     d)  electron multiplier -  -1800V

     e)  electron energy     -  50 eV

     f)  filament emission   -  30 ma

     g)  scan rate           -  3 sec/spectrum
                                22

-------
                              TABLE 7
                     Analysis of Aroclor 1242
                                             % By
PCB
Cluster
Cli
C12
C13
Cli+
C15
Cl6
m/e
188
224
258
292
326
362
Known1
Results
1.1
16.95
39.19
31.83
9.64
.49
Weight2
Found
4.74
9.16
45.67
32.17
7.62
.63
Rel. Std.
Deviation
.0217
.0330
.0101
.0319
.0016
.0120
1  Webb & McCall, Ref.  6

   Weight percent found by normalizing raw area ratio (PCB cluster
   area/internal standard area)  by the isotopic abundance of the
   analytical ion and the molecular weight for the chloro group.
                                 23

-------
                         TABLE 8






                   Analysis of Aroclor 1254
PCB
Cluster
C12
C13
Clk
m/e
224
258
292
Known1
Results
	 3
_3
13.80
% By2
Weight
Found
.17
1.03
21.68
Rel.Std.
Deviation
.003
.0056
.0106
 C15          326         61.92          56.88           .0204





 C16          362         23.28          19.08           .0019





 C17          394          1.00           1.17           .0071
1 See Table 7




2 See Table 8




  Webb reported components only >1%
                               24

-------
                          TABLE 9
                    Analysis of Aroclor  1260
PCB
Cluster
C12
Cl3
Cli,
C15
C16
C17
C18
m/e
224
258
292
326
362
394
428
Known1
Results
	 3
	 3
__3
11.52
46.14
34.84
6.10
% By
Weight2
Found
.20
.34
1.52
17.85
47.39
28.24
4.46
Rel. Std.
Deviation
.0009
.0029
.0032
.0776
.0152
• 0619
.0102
1 see Table 7



2 see Table 7



3 Webb reported only components >1%
                              25

-------
                     V.  VERIFICATION OF METHOD

A.  Evaluation of Test Samples

    In order to demonstrate the application of the PCB procedure to
actual samples, two test mixtures were obtained, dosed with different
Aroclor materials and analyzed by the proposed procedure.  One of the
test mixtures was a complex combination of several wastes and waste
extracts (see  Table 10 for the mixture composition)  while the other
mixture was the extract of a sample collected from a ferroalloy smelter.
Both samples had a wide variety of organic compounds ranging from low
mass to above mass 400.

    The wastes and waste extracts comprising the first test sample are
listed in Table 10. These represent a complex system with numerous
potential interferences.    The perchloroethylene waste did, in fact,
have some PCB content which added to the complexity of that sample.
Aliquots of the waste  solution were dosed with either Aroclor 1242 or
Aroclor 1260 to a level of 10 yg per mL, and the internal standard
p,p'-DDE added to a level of 1 yg per mL prior to analysis.  The amount
of dissolved material in this solution was about 0.28 g per mL, greatly
in excess of the amount of PCBs present.  Figure 7 shows the reconstruc-
ted gas chromatogram (RGC) of the waste sample obtained from the inten-
sities of the ions in the mass ranges listed in Table 1 for each of the
PCB groups.

    This sample contained no added Aroclor.  With no preliminary sepa-
ration of the sample as is the case with the first sample, the normal
GC/ECD pattern recognition approach would have been unable to even indi-
cate the presence of PCBs in the sample.  The presence of PCBs would be
even further obscured if ions had been acquired over the entire mass
range of interest, the normal GC/MS data acquisition mode.

    Figure 8a shows the RGC for this mixture dosed with Aroclor 1260.
Since this sample consisted of primarily low molecular weight species,
the higher mass PCBs show relatively little inter*" ^rence from back-
ground species.  The RRT windows used for each s^-.ies are shown in
Figures 8b-d, which is an RGC constructed from only the ions of analyti-
cal interest.  Table lla lists the recovery data for the Aroclor 1260
dosed sample.  For the major Aroclor containing components (Clg and Cly
biphenyls) the recoveries are within +5% and the relative standard
deviations are less than 3%.  The other large component (Cls) has a
higher error (+10%) than for the two major components but it still
represents a reasonably accurate measurement.  The minor Gig biphenyl
group shows large error (22%) and standard deviations (35%). but the
absolute level of this component was quite low (about 600 pg injected
or 1.4 x 10 12 mole), approaching the instrumental detection limit.
Measurement at such levels will show large errors for replicate runs
due to instrumental fluctuations.
                                 26

-------
                              TABLE 10

                     Organic Mixture Composition
Styrene waste
mixture of aromatic hydrocarbons
API waste extract
mixture of aliphatic unsaturated
hydrocarbons and aromatic hydro-
carbons
Lucidol waste
mixture of a-methylstyrene, cumene,
cumyl alcohol and acetophenone
Perchloroethylene waste
hexachlorobutadiene, hexachloro-
benzene and a mixture of other
chlorinated hydrocarbons
Simulated coke waste
 extract
mixture of phenol, cresol, amines
and benzoic acid
p-Toluene sulfonic acid
                                  27

-------
Is,
00
                                                                                300
' ' ' ' I  '
   350
                         FIGURE 7  GC/MS SUBSET MASS CHROMATOGRAM FOR WASTE EXTRACT MIXTURE

-------
KJ
VC
                                                                         r-f-r- t- r-r-p*
                                                                          100   I'll
                     50   100

                                                                                       ?so   300   350   «0fl
                                                                                                ',0   SUO   SSI
                     FIGURE 8    a) GC/MS SUBSET MASS CHROMATOGRAM, b-fl) SINGLE ION CHROMATOGRAMS FOR
                                 ANALYTICAL IONS OF AROCLOR 1260 (326, 362,394 AND 428) FOR WASTE EXTRACT
                                 MIXTURE DOSED WITH AROCLOR 1260.  SOLID LINES DELINEATE RELATIVE
                                 RETENTION TIME WINDOWS USED

-------
     When the above sample was dosed with Aroclor 1242 instead of Aro-
clor 1260, the interference due to the lower weight species became more
important.  Figure 9 shows the RGC for the dosed sample using the set
of selective scanned masses (Figure 9a) and the chromatograms obtained
for each of the analytically important masses (Figures 9 b-g).  The in-
terference at mass 188 (Clj biphenyl) swamps out the signal for that
chlorog group making it impossible to quantitate the level of monochloro-
biphenyl present.  The other chloro groups from the added Aroclor 1242
show a decreasing amount of interference as the mass of the chloro group
increases.  The results are tabulated in Table lib.

    The second test mixture collected from a ferroalloy smelter repre-
sented a substantially different test for the PCB analysis procedure
than the previous mixture.  This mixture contained aromatic hydrocarbons
and various oxygenated compounds ranging in mass to above 500, with
major components found in the 220-300 mass range.  This range covers the
Cl^-Cli, PCBs, and the high levels of high mass species present a chal-
lenge for the determination of 01$ and higher PCBs not encountered with
the previous sample.

     Examination of the ferroalloy extract prior to any Aroclor dosing
showed no detectable quantity of any PCBs.  An aliquot of the extract
was then dosed with Aroclor 1254 to a level of 10 ug/mL, and repeti-
tively analyzed for PCBs.  Figures 10 a-d show the various chromatograms
obtained for the ferroalloy sample extract dosed with Aroclor 1254.  The
large variety of high mass species caused a nearly uniform total signal
during the elution of the PCBs, in contrast to the decreasing total
signal during the course of the PCB elution seen in the previous samples.
The recoveries for the various PCB groups are listed in Table 12.

     The low recovery observed for the Clg biphenyl is in part due to
the high background observed for that ion throughout the chromatographic
run.  Fluctuations in the background level near the RRT window for the
hexachloro group will cause error in the estimation of the background
level, with overestimation of the background level leading to low re-
coveries.  However the recoveries found are acceptable for these low
level analyses, particularly since no pretreatmei   >r separation proce-
dures were used.

B.  Recovery from Flyash

    As a final test of the recommended PCB procedure, a sample of flyash
was thoroughly extracted with methylene chloride, dosed with Aroclor
1254 and re-extracted with methylene chloride to ascertain the recovery
of PCB from a flyash sample.  To 2 g of the pre-extracted flyash, 10 pg
of Aroclor 1254 in 50 mL of methylene chloride was added and the mixture
evaporated to dryness.  The sample was then re-extracted with ten 10 mL
portions of methylene chloride.  The ten portions were combined and
evaporated to 1 mL.  The concentrated extract was dosed with 1 yg of
the internal standard material, p,p'-DDE, and analyzed via the recom-
mended procedure.
                                  30

-------
                              TABLE 11
           Analysis of Mixture of Wastes and Waste Extract
a)  Mixture dosed with Aroclor 1260:
                    Mixture +  Aroclor  Aroclor
  PCB     Mixture    Aroclor    1260
Cluster  Background   1260      Found
            (ug)      (yg)      (yg)
  Cl,.       .234      1.497     1.263
            .351      4.919     4.568
           1.059      4.739     3.680
            .066       .813      .747
  cle
  Cl.
     J
  Cl,
Aroclor
1260
Added
(vg)
1.152
4.614
3.484
.610
Recovered
(%)
109.6
99.0
105.6
122.5
Relative
Standard
Deviation

.037
.015
.029
.353
b)  Mixture dosed with Aroclor 1242
                    Mixture +  Aroclor  Aroclor
                                                           Relative
PCB
Cluster

Cli
C12
C13
Cl,,
C15
cic
Mixture
Background
tygj
—
—
—
.013
.234
.351
Aroclor
1242
(yg)
—
—
3.83
2.68
,905
.372
1242
Found
(yg)
—
5.85
3.83
2.67
.671
.021
1242
Added
(yg)
.11
1.70
3.92
3.18
.96
.05
Recovered
(%)
—
34.4*
97.7
84.0
69.9
42.0
Standard
Deviation

—
.090
.099
.050
.051
.255
*From reference 15, test sample was a mixture of each of the waste
 ans waste extracts listed.
                                31

-------
                    H   !•   IH  m   SCO   3

                                   H/C It*
K
                        i'li	"fti	'MB	in   »
                                   ivtnt
                    ta   log   fa  ni   aa   m  xo
                                                                   n   m   IM   tn   m   JOB
                                                                                         30B   3EB   408   4H
                                            se   IBB   tso
                                                            Z6D   JOB   3bO   4Bd   4Sfl
                         FIGURE 9   a) GC/MS SUBSET MASS CHROMATOGRAM, b-g) SINGLE ION CHROMATOGRAMS
                                   FOR THE ANALYTICAL IONS OF AROCLOR 1242 (188.224,258.292,326) FOR
                                   WASTE EXTRACT MIXTURE DOSED WITH AROCLOR 1242. ARROWS DELINEATE
                                   THE RELATIVE RETENTION TIME WINDOWS USED

-------
               SO    tOO    160   2W    260   300   360    ««
         too
               60    tOO    160   200    260    300   360    «»
               60    100    160   200    260
FIGURE 10    a) GC/MS SUBSET MASS CHROMATOGRAM. b-d) PROMINENT ION
            CHROMATOGRAMS FOR ANALYTICAL IONS 292.326,362 OF
            FERRO-ALLOY SMELTER SAMPLE DOSED WITH AROCLOR 1254.
            ARROWS DELINEATE RELATIVE RETENTION TIME WINDOWS USED
                              33

-------
                              TABLE 12
                 PCS Analysis of Ferroalloy Sample1
PCB
Cluster2
Cl,,
C15
C16
C17
PCB Found
1.35
4.98
1.61
— an
Aroclor 1254
Added
1.38
6.19
2.33
.10
Recovered
97.8
80.5
69.1
••MB
1 Values reported are for 1 mL of sample volume.

2 €13 cluster was observed In reference material  but Interferences
  in the RRT window from other species prevent measurement.
                                 34

-------
     The RGC obtained from the extracted flyash sample was virtually
identical to the RGC obtained for the Aroclor 1254 reference material.
Figure 11 allows comparison of the Aroclor 1254 reference sample with
the re-extracted flyash sample.  The recoveries were quite good, at
least 80% for the major PCB components (see Table 13).  The Cla bi-
phenyl was seen in both the reference material and the dosed flyash
sample and the relative responses agreed within 5% for the reference
standard and the extracted flyash.
                                  35

-------
                        150     200      260     300      35C
  I I I I I I I | I I I I ' I ' I I [ I I I I ' M I ' [ I I ' I I I ' M | ' I ' I ' I r | I [ I | I I i | I M | ' I I I I I I I ' [ ' I
        50      100      ISO      200       250      300      350

FIGURE 11    GC/MS SUBSET MASS CHROMATOGRAM FOR EXTRACTED FLY ASH
           SAMPLE DOSED WITH AROCLOR 1254 (a) AND AROCLOR 1254
           STANDARD (b)

-------
                             TABLE 13



       PCS Analysis for  Recovery of Aroclqr 1254  ia  Dosed Flyash1



        PCB                            Aroclor 1254
      Cluster2        PCB  Found            Added         Recovered
                          (Vg)               (ug)             (%)


         CL*              1.17              1.38           84.8


         C15              5.11              6.19           82.6


         Cle              2.47              2.33          106.0


         C17                —               .16            —



  Notes:  See notes for this table on following page.
1 Values reported are for 1 ml of sample volume.

2 Cla cluster observed in both reference material and sample.
  Agreement between the 013 cluster to internal standard ratio
  for the reference and the standard is within 5%.
                                37

-------
                            VI.  CONCLUSION
     The analysis procedure for the measurement of PCB emissions from
combustion sources, Appendix A, has been tested for several representa-
tive samples.  The recoveries found were quite good without any form of
extraction or sample cleanup.  For complex samples, use of a prelimi-
nary separation scheme should improve the quantitative measurement of
PCBs by eliminating low mass interferences which can be quite sizeable.
Standard separation procedures such as those described in the Federal
Register Method for PCBs may be used for pretreatment of samples prior
to analysis of PCBs.

-------
                           VII.   REFERENCES


 1.  J. Mieure, et al., Characterization of Polychlorinated Biphenyls,
     in National Conference on Polychlorinated Biphenyls (November 19-21,
     1975, Chicago, Illinois), published by Research Triangle Institute,
     March 1976, NTIS No. PB-253248, page 84.

 2.  S. Safe, Overview of Identification and Spectroscopic Properties,
     ibid, page 94.

 3.  Panel on Hazardous Trace Substances:  Polychlorinated Biphenyls,
     Environ, Rsch., .5, 247-362 (1972).

     a.  Properties, Production and Uses, ibid, p 258.

     b.  Analytical Methods, ibid, p 338.

 4.  0. Hutzinger, _et jl., "The Chemistry of PCB's", CRC Press,
     Cleveland, Ohio (1974).

 5.  Method for Polychlorinated Biphenyls (PCB's) in Industrial Effluents,
     Fed. Reg. J8, No. 75, Pt. II (1973).

 6.  R. Webb and A. McCall, Quantitative Standards for Electron Capture
     Gas Chromatography, J. Chromat. Aci., 11, 366 (1973).

 7.  a.  C. Haile and E. Baladi, Methods for Determining  the Total Poly-
         chlorinated Biphenyl Emissions  from Incineration and Capacitor-
         and Transformer-filling Plants, Final Report on  EPA Contract
         No. 68-02-1780, Nov. 1977.

     b.  Personal communication, J. Clausen, TRW Systems Group,
         January, 1977.

 8.  J. Armour, Quantitative Perchlorination of Polychlorinated Biphenyls
     as a Method for Confirmatory Residue Measurement and Identification,
     JAOAC, 56, 987 (1973).

 9.  W. Mitchell, QAB/EMSL/EPA/RTP, personal communication
     February, 1977.

10.  C. Haile, PCB Interlaboratory Verification Analysis, Midwest
     Research Institute, Final Report on EPA Contract No. 68-02-1399,
     December 27, 1976.

11.  B. Dudenbostel, Tentative Method of Test for Polychlorinated
     Biphenyls in Water, EPA, Region II, Edison, N.J., internal memo,
     January 22, 1976.

 12.  J. Eichelberger, et al., Analysis of the Polychlorinated Biphenyls
     Problem, Anal. Chem., 46, 227  (1974).
                                   39

-------
References (continued)

13.  S. Safe and 0. Hutzinger, Mass Spectrometry of Pesticides and
     Pollutants, CRC Press, Cleveland, Ohio, 1973.

14.  R. Webb and A. McCall, Identities of Polychlorinated Biphenyl
     Isomers and Aroclors, JAOAC, 55, 746 (1972).
                                  40

-------
                          APPENDIX A

                        Proposed Method
                         (Abbreviated)
             Measurement of Polychlorinated Biphenyls
                   (PCBs) in Combustion Sources
1.  Abstractof the Method

    The method is designed primarily to address the problem of
    measurement of PCB emissions from combustion sources, but
    should be applicable to PCB measurements from any source.

    The method uses an automated gas chromatograph/mass spectrom-
    meter.  Data are acquired in a select subset of masses and
    integrated according to gas chromatographic retention time
    criteria.  Data are reported as quantity of monochloro-,
    dichloro-	decachlorobiphenyl.

2.  Interferences

    Interferences in the PCB analysis are minimized with this
    procedure.  Isotope abundance patterns are used to verify
    the composition as a PCS.  Selected mass chromatograms and
    retention time windows provide a high degree of specificity
    in the analysis of a specie as a PCB.

3.  Sample Extraction

    Sample extractions should be done using distilled-in-glass
    pentane or methylene chloride (Burdick and Jackson).  Samples
    should be concentrated to 1.0 yl using a Kuderna-Danish evapo-
    rator.  If necessary to achieve sensitivity samples may be
    further concentrated to 0.1 ml using a gentle stream of
    nitrogen.

4.  Sample Cleanup

    It may be possible to analyze the extracted samples directly
    without further cleanup.  The analysis itself should be the
    criteria for•determining the need for further cleanup as de-
    scribed for the Standard EPA method for PCBs in industrial ef-
    fluents. (5)   If cleanup is required, use the florisil/silica
    gel procedures described in the EPA method.
                                A-l

-------
5»  Analysis

    A.  GC Conditions

        Use a 2 m x 2 mm I.D., glass column containing any of several
    phases.  OV-1, OV-101,  OV-17,  Dexsil 300 and Dexsil 400 at 3% on
    80/100 Chromosorb have  all been used successfully for the PCB
    analysis.   Temperature  programming from about 150°-280°C has been
    used,  but the chromatograms are more reproducible when run in the
    isothermal mode.   A temperature of 185 °C on OV-1 is good for the
    Aroclors through  1248.   A temperature of 200°C is used for the
    higher Aroclors.   A 2-5 yl sample size injection is made dependent
    on the concentration in the sample.   Use of less than 2 yl will
    lead to poor reproducibility.   The GC gas stream is diverted for
    the first 30 sec  allowing the  solvent to elute and be vented and
    then the diverter is closed and data acquisition initiated.

    B.  MS Conditions

        Extract Conditions  will depend on spectrometer type and condi-
    tion.   Care should be taken to calibrate the mass scale to accom-
    modate the significant  mass defect of the PCBs.  It is recommended
    that an Aroclor mixture be used in place of PC-43 (or other PFKs)
    to construct an alternate mass calibration scale for the PCB analy-
    sis.   Set the mass ranges for  data acquisition as follows:

          PCB Group             Range            Analytical m/e
                               186 - 190               188

             C12               200 - 226               224

             C13               254 - 260               258

             Cli*               288 - 294               292

             C15               322 - 328               326

             C16               356 - 364               362

             C17               392 - 400               394

             C18               426 - 434               428

             C19               460 - 468               464

             Cllo              494 - 504               498


    Integration times will vary with instruments.  A setting of 50-
    64 msec/amu is recommended.
                               A-2

-------
 6.  Qualitative  Identification of  PCBs

    A  total  chromatogram is  constructed  from the  sum  of all  the mass
    used  in  data acquisition.   Individual mass  spectra are obtained
    at GC peak maxima.   These  spectra are examined  to determine
    whether  the  proper  isotope abundance patterns are present  for
    the given  chlorobiphenyl group.

 7.  Quantitative Measurement of PCB  Groups

    When  the species have been confirmed as PCB's,  individual mass
     chromatograms are obtained for the analytical masses corres-
    ponding to the PCB groups, 188,  224, 	 498.  An Aroclor sample
     such  as Aroclor 1232 and Aroclor 1254 is used to establish a
    relative retention time (RRT) scale using the data given by
    Webb  and McCall^6^.  The area for each PCB group is integrated
     over  the RRT regions indicated below:

           PCB Group        Analytical m/e         RRT Region

               Cli                188              0(5) -  20

               C12                224                 15 -  35

               C13                258                 25 -  55
               Cli,                292                40 - 100

               C15                326                 70 - 150

               C16                363               125 - 250

               C17                394               160 - 350
               C18                428               275 - 600

               C19                464               400 -1000

               C110               498               650 -1200

    The RRT windows may need to be adjusted slightly  for proper
    measurement of total areas.  Use of these windows minimized
    interferences from other PCB&.

    Complete details of the quantitative calibration have not been
    worked out at this time.  It is tentatively recommended that
    calibration be based upon specific GC peaks in Aroclor reference.

8.  Sensitivity

    The sensitivity of this method has not  been established,  but is
    expected to be at least 0.1 ng/injected sample.
                              A-3

-------
                                               APPENDIX B
 Quantitative   PCB   Standards   for

 Electron   Capture   Gas  Chromatography
 by Ronald G. Webb and Ann C. McCall, Southeast Environmental Research Laboratory, National
 Environmental Research Center—Corvallis Environmental Protection Agency, Athens, Georgia 30601
                      Abstract

    The weight of PCB represented by each electron capture
 gas chromatographic (EC-GC) peak in solutions of Aroclors
 1221-1260 has been determined.  The  Aroclor samples from
 which these solutions were prepared are proposed as quan-
 titative PCB standards. Their compositions were determined
 by elemental analysis, GC with a Coulson conductivity de-
 tector, and  combined GC MS. Retention times  relative to
 p.p -DDE  are recommended to designate individual GC peaks
 of PCB's. A table is given for each Aroclor showing the weight
 percent of each EC-GC peak in the mixture. A procedure us-
 ing Aroclors 1242,1254, and 1260 is recommended for analyz-
 ing environmental  samples containing more than one Aroclor
 mixture.  Stock  solutions  of the  Aroclors in isooctane are
 stable except when directly exposed to sunlight  Ampoules
 of the Aroclor solutions are offered.
                    Introduction

    Quantitation of polychlorinated biphenyls  (PCB's)
 from electron capture (EC) chromatograms is compli-
 cated because the EC detector responds differently to
 each PCB isomer '1,2). Quantitation by direct com-
 parison of an unknown EC chromatogram with those of
 Aroclor standards is difficult because individual peaks
 in environmental samples are sometimes obscured by
 pesticide residues, are  completely missing,  or  have
 considerably different relative intensities.
    To avoid these difficulties, Berg and co-workers (3)
 have proposed that the PCB's in a sample be converted
 to decachlorobiphenyl,  and the EC signal from  this
 derivative be compared with that from conversion of
 a known amount of Aroclor 1254 to decachlorobiphenyl.
 This method appears to be ideally suited to a  monitor-
 ing program designed for rapid and sensitive measure-
 ment of the total quantity of PCB's without regard to
 composition. In many cases, this derivative approach
 is  undesirable because an extra analytical  step is re-
 quired and because any  evidence  of  metabolism or
 degradation of the sample is destroyed. The disadvan-
 tages of both the direct comparison method  and the
 derivative method can be largely  overcome by using
 Aroclor standards in which the quantitative composi-
 tion of each EC-GC peak is  known. We have prepared
 these standards and recommend procedures for  their
 use.
                   Experimental
    Aroclor 1221   C, 72.94; H, 4.45; Cl, 22.74

           1232   C, 64.44; H, 3.46; Cl, 31.96

           1242   C, 54.64; H, 2.70; Cl, 42.85

           1248   C, 49.50; H, 2.17; Cl, 48.54

           1254   C. 44.10; H, 1.61; Cl, 54.33

           1260   C. 38.18: H, 0.94; Cl, 60.97

    The Food and Drug Administration provided the
primary standard p.p'-DDE, which  was used as a re-
tention time standard and calibration standard for the
conductivity detector.
    A Microtek 220 gas chromatograph was equipped
with a Coulson conductivity detector. The column was
a 6 ft x Vi in., o.d., glass U-shaped tube packed with 3%
SE-30 on  80/100 Gas  Chrom Q. The carrier gas was
helium at a flow rate of 60 ml/min. All Aroclors were
chromatographed iaothermally; Aroclors 1221 and 1232
at 175*C;  1242, 1248 and 1254 at 185*C; and 1260 at
190*C. The chromatograms were quantitated by mea-
suring peak areas with either a planimeter  or disc in-
tegrator.
    A Microtek 220 gas chromatograph with Ni-63 EC
detector was operated  at  15-30  V (DO and 275'C.
The column was a 6 ft x Vi in., o.d.,  glass U-shaped
tube  packed with 3% SE-30 on 80/100  mesh Gas
Chrom Q. The carrier  gas was nitrogen at 90 ml/min.
Aroclors 1221 through 1254 were chromatographed iso-
thermally  at 200'C and Aroclor  1260 at 215"C.
    An F&M 700  gas chromatograph with tritium EC
detector at 205'C was  operated at a pulse  interval  of
15  microseconds.  The coiled glass column was  8 ft  z
Vi  in., o.d., packed with 3% SE-30  on  80/100 Gas
Chrom Q.  The carrier gas was 95% argon and 5%
methane  at 80-100 ml/min.  AH  samples were chro-
matographed isothennally at 195° C.
    Mass spectra  (70 eV) were obtained on a Finnigan
1015-C quadrupole mass spectrometer interfaced with
a Gohlke  separator to a  modiifed  Varian  1400 GC.
GC  conditions were set  to produce  chromatograms
equivalent to those from EC-GC. The spectrometer was
controlled  by a DEC  PDP-8 computer,  and spectra
were collected on magnetic tape and printed or plotted
under computer control.
'Mention of product* or companies doet not imply endorte-
 ment by tk« Environmental Protection Agency.
    The  Monsanto  Company"  provided  the Aroclor
 samples,  which were not marked with lot numbers.
 Elemental analysis by Galbraith Laboratories, Knox-
 ville, Tennessee, showed the following percent composi-
 tions (average of triplicate analyses):
Reorinted with permission of the copyright"   holder
 1. Gregory, N. L., J. Chem. Soc. (B). 1968. 295 (1968).
 2. Zitco. V., HuUinger. O.. and Safe, S.. Bull. Environ
   Contain. Toxicol. 6,160 (1971).
 3. Berg, O. W., Dioaady, P. L., and Rees, G. A. V. Bull
   Environ. Contain. Toxicol. 7,338 (1972).

-------
               Results and Discussion
    The weight of PCB present in each GC peak of a
 given Aroclor can  be calculated from two pieces of
 information:
    1) the empirical formula of the compound
       represented by the peak, and
    2) the absolute  amount of chlorine represented by
       the peak
 Combined GC/MS determines  the  first, and a  GC
 equipped with an electrolytic conductivity detector can
 determine the second.
    GC/MS examination of Aroclors  showed that sev-
 eral peaks were mixtures of PCB's with different num-
 bers of chlorines. To estimate the composition of a GC
 peak  containing  PCB's  with  different numbers  of
 chlorines,  the  following observation  was used:  equal
 weights of two PCB's that differ only by one chlorine
 give the same sum, within 25%, when the intensities
 of all the signals from the molecular  ion, or parent,
 cluster, the parent-minus-one-chlorine cluster and par-
 ent-minus-two-chlorine cluster of each PCB are added.
 This  rule was derived from a limited study of quadra-
 pole mass spectra of a series of  synthetic PCB's  (4)
 and may not hold for  other types of spectrometers.
    Figure 1 is the mass spectrum from an Aroclor 1242
 GC peak that is a mixture of one or more trichloro-
 biphenyls and one or more tetrachlorobiphenyls. The
 molecular  ion pattern at m/e 290-298 is  typical of
 four-chlorine  molecules.  These  molecules lose one
 chlorine, producing  a  three-chlorine fragment  pattern
 (parent-minus-one-chlorine) at m/e 255, 257, 259, and
 261. However, between these signals, there is also a
 strong three-chlorine pattern at m/e 256, 258, 260, and
 262; this  is the parent ion duster of the trichlorobi-
 phenyl(s). The signals at m/e 220-224 are seen after
 loss of two chlorines from the tetrachlorobiphenyl (the
 trichlorobiphenyl (s) does not show a significant signal
 for loss of one chlorine), and  the signals at m/e 186
and 188 are seen after loss of two chlorines from tri-
 chlorobiphenyl.
   In Figure  1,  the tetrachlorobiphenyl  intensities
 (peak heights)  totaled 612 mm and the trichlorobi-
phenyl 304 mm. Tetrachlorobiphenyl is thus about two-
            Jf
            I V
        tFtrnui MIVDI
       170  10 ISO 200 210 22D ZB XO ZS 268 BO HO 250 JOO
Figure 1.  A limited portion of the mass spectrum  of a mix-
ture of tetrachlorobiphenyl(s) and trichlorobiphenyl(s)  from
RRT peak 54 of Aroclor 1242  (See Figure 5). The abcissa is
           marked in atomic mass units (m/e).
                     1242
                               "J?4
 Figure 2. Gas chromatogram of Aroclor 1242 on SE-30 with
 an electrolytic conductivity detector. The peak identification
 numbers correspond to the retention  time  relative  to p,p'-
 DDE=100. From injection, at the  arrow,  to peak 146  was
                     about 20 min.
thirds of the mixture.  These data were used to cal-
culate the average molecular weight of the material in
the GC peak.
   The Coulson conductivity detector responds linearly
to chlorine. Linear response to PCB's was shown with
individual PCB isomers  (4)  containing one  to six
chlorines, and the detector response was checked for
reproducibility several times each day with p,p'-DDE.
A typical Coulson chromatogram for 1242 is shown in
Figure 2. Peak resolution was not completely optimized
here or in the GC-MS  work so that these separations
would be typical of those in the  pesticide literature
(5-10). Other studies  (4, 11, 12) have show that the
Aroclors are much more complicated than shown here,
but this resolution is adequate for routine analysis.
   The area of each Aroclor peak was determined and
the weights (nanograms) of chlorine and PCB present
were  calculated using  the response of p,p'-DDE as
follows:
     ng DDE injected
   DDE peak area (cm=)
   ngCl
 4xat.wt. Cl    ngCI
mol. wt. DDE ~  cm=
    cm1
         X PCB peak area (cm2) =ngCl
                                                              ngCl  X;
                    Gram molecular weight
            ' No. of chlorines in molecule x 35.46 g
             ngPCB
   Tables I-VI present these results as the percent of
 4.  Webb, R. G., and McCall. Ann C.,  J. Assoc.  Offic.
    AnaL Chemists 55, 746 (1972).
 5.  Zitco,  V.,  Bull. Environ. Con tarn.  Toxicol. 6. 46-1
    (1971).
 6.  Biros, F. J.. Walker,  A. C., and Medberry, A., Bull
    Environ. Contain. Toxicol. 5,317 (1970).
 7.  Grant, D. L., Phillips. W. E. J., and Villeneuve. A..
    Bull. Environ. Contain. Toxicol. 5,317 (1970).
 8.  Holmes, D. C., Simmons, J. H-, and Tatton. J. O'G..
    Nature 216, 227 (1967).
 9.  Bagley, G. E., Reichel, W. L., and Cromartie, E..  J.
    Assoc. Offic. Anal. Chemists 53, 251  (1970).
10.  Rote, J. W.. Murphy. P. G..  Bull. Environ. Contain.
    To*icol. 6. 377 (1971).
11.  Stalling, D. L., and  Huckins, J. N..  J. Assoc.  Offic
    Anal. Chemists 54.801 (1971).
12.  Sissons, D., and Welti, D., J. Chroraatog 60. 15 (1971).
JOURNAL OF CHROMATOGRAPHIC SCIENCE  •  VOL. 11
  B-2
          JULY 1973  •  3S7

-------
        Table I.  Composition of Aroclor 1221
RRT'
11
14
16
19
21
28

32

37
40
Total
Mean
Weight
Percent
31.8
19.3
10.1
2.8
20.8
5.4

1.4

1.7

93.3
Relative
Std. Dev.''
15.8
9.1
9.7
9.7
9.3
13.9

30.1

48.8


No. of
Chlorines1'
1
1
2
2
2
2~i
3 '
2H
3 !
3
3






85%
15%
10%
90%



   'Retention time relative to p,p'-DDE=100. Measured
from  first appearance of solvent. Overlapping peaks
that are quantitated as one peak are bracketed.
   Standard  deviation of seventeen  results as a per-
centage of. the mean of the results.
   From GC/MS data. Peaks  containing  mixtures of
isomers of different chlorine numbers are bracketed.
Table


RRT-
11
14
16
20
21
28

32
37
40
47
54

58
70

78
Total
1 1. Composition of Aroclor
Mean
Weight
Percent
16.2
9.9
7.1
17.8

9.6

3.9
6.8
6.4
4.2
3.4

2.6
4.6

1.7
94.2

Relative
Std. Dev."
3.4
2.5
6.8
2.4

3.4

4.7
2.5
2.7
4.1
3.4

3.7
3.1

7.5

1232

No. of
Chlorines0
1
1
2
2
2
2—40%
3J60%
3
3
3
4
3-^33%
4_J67%
4
4- 90%
5 110%
4

   Retention time relative to p,p'-DDE=100. Measured
from  first appearance of solvent. Overlapping peaks
that are quanti' .ted as one peak are bracketed.
   Standard d-   ation of four results as a means of the
results.
   From GC/MS data. Peaks  containing mixtures of
isumers of different chlorine numbers are bracketed.
    total Aroclor  weight  represented by  each GC peak.
    The peaks are identified by their retention times rela-
    tive to p,p'-DDE. We recommend that this be adopted
    as a standard method for  designating individual PCB
    GC peaks. The separate  percentages given for  over-
    lapping peaks were obtained by dividing the area with
    a perpendicular to the baseline from the minimum point
    between the two peaks. The accuracy of the Coulson
    determinations was checked by comparing  each  Aro~
    clor's calculated percent  chlorine with  its elemental
    analysis. The amount found by Coulson GC was 98-
    102%  of  the  elemental analyses  except  for Aroclor
    1221.
       Seventeen  analyses with 1221 were performed and
    the average of the data was used to prepare  Table I.
    Willis  and Addison (13) have recently reported  semi-
    quantitative values for the composition of Aroclor 1221.
    Their analysis was based on EC-GC and flame ioniza-
    tion GC.  They  found  12.7%  biphenyl  present  and
    about the same amounts of other materials as in Table
    I. Willis  and Addison accounted  for 92.3% of the
    weight of materials in  1221. If their 12.7%  biphenyl
Table !
RRT"
11
16
21
28

32
37
40
47
54

58
70

78
84
98
104
125

146

Total
III. Composition of Aroclor
Mean
Weight Relative
Percent Std. Dev."
1.1
2.9
11.3
11.0

6.1
11.5
11.1
8.8
6.8

5.6
10.3

3.6
2.7
1.5
2.3
1.6

1.0

98.5
35.7
4.2
3.0
5.0

4.7
5.7
6.2
4.3
2.9

3.3
2.8

4.2
9.7
9.4
16.4
20.4

19.9


1242
No. of
Chlorines1"
1
2
2
2~ 25%
3J75%
3
3
3
4
3-j 33%
4 J 67%
4
4~, 90%
5J10%
4
5
5
5
5~~j 85%
6J15%
5~1 75%
6J25%

      "Retention time relative to p,p'-DDE=100. Measured
    from first appearance of solvent.
      •"Standard deviation  of six results as a percentage
    of the mean of the results.
      •'From GC/MS data. Peaks containing mixtures of
    isomers of different chlorine numbers are bracketed.
    13.  Willis, D. E.. and  Addison, R. F., J.
        Board Can. 29,592 (1972).
Fisheries  Res.
368  •  JULY 1973
B-3
                                                          inilRNAI  OF r.HROMATnRRttPHir

-------
 is added  to the PCB values  in  Table I,  the weight
 IK'rcvnt Aroclor is  106.0 A mixture composed of the
 rdVs that Willis and Addison quantitated would con-
 tain 19.6% by weight chlorine; the Coulson determina-
 tions pave 22.9%; elemental analysis gave 22.7%.
    When  the  data in Tables I-VI are compared to
 Aroclor chromatograms from an EC detector operated
 in the DC mode (Figures 3-8), peak size obviously  is
 not a valid indication of concentration. For example,
 in Aroclor 1242 (Figure 5a) peaks 21, 28, 37, and 40
 each represent about 11%  of the mixture  (See  Table
 III), but their areas differ  by as much as 65%. There
 are also major differences in peak ratios when the
 Aroclors are measured with a detector operated in the
 pulsed mode as shown in Figure 5b.

 A Technique to Quanlitate PCB's in
 Environmental Samples

   The chromatograms of  PCB's from environmental
samples usually show some evidence of degradation or
 metabolism. A sample may contain a single partially
degraded Aroclor or a combination of Aroclors. Such
samples can be quantitated  by using the standard Aro-
clors, the data in Tables I-VI, and  some simple compu-
tation rules. The key principle is that the total amount
Table


RRT-
21
28
32
37
40

47
34

58
70

78
84
98
101

112
125

146

Total
IV. Composition of Aroclor
Mean
Weight
Percent
1.2
5.2
3.2
8.3
8.3

15.6
9.7

9.3
19.0

6.6
4.9
3.2
3.3

1.2
2.6

1.5

103.1

Relative
Std. Dev."
23.9
3.3
3.8
3.6
3.9

1.1
6.0

5.8
1.4

2.7
2.G
3.2
3.G

0.6
5.9

10.0


1248

No. of
Chlorines1
2
3
3
3
3-] 85%
4_J15%
4
3— 1 10%
4J90%
4
4-] 80%
5J20%
4
5
5
r ' 10%
5 j 90%
5
5~" 90%
6J10%
5") 85%
« J 15%

   •Retention time relative to p,p'-DDE=100. Measured
from first appearance of solvent.
  'Standard deviation of six results as a percentage of
the mean of the results.
  • From GC MS data. Peaks containing mixtures of
isomers of different chlorine numbers are bracketed.
of PCB present is  the sum of the amounts  from all
the individual peaks.
   To  quantitate  PCB's,  chromatograph   known
amounts of the standards.  Measure the area  for each
peak. Using the tables, determine  the response factor
                    1221
                 2»
                                                       Figure 3.  EC chromatogram of Aroclor 1221 chromatographed
                                                       on SE-30 with a Ni-63 detector operated in the DC mode. The
                                                       peak identification numbers  correspond to the retention
                                                                    time relative to p,p'-DDE=100.
Table


RRT"
47
54
58
70

84
98
104
125

146

160
174
COS
232
Total
V. Composition of Aroclor
Mean
Weight
Percent
6.2
2.9
1.4
13.2

17.3
7.5
13.6
15.0

10.4

1.3
8.4
1.8
1.0
100.0

Relative
Std. Dev."
3.7
2.6
2.8
2.7

1.9
5.3
3.8
2.4

2.7

8.4
5.3
18.G
26.1

1254

No. of
Chlorines1
4
4
4
4—, 25%
5 '75%
5
5
5
5~ 70%
6 '30%
5— 30%
6 :70%
6
G
G
7

  "Retention time relative to p,p'-DDE=100. Measured
from first appearance of solvent.
  "•Standard deviation of six results as a percentage of
the mean of the results.
  Trom GC,MS data. Peaks containing mixtures of
isomers are bracketed.
JOURNAL OF CHROMATO&RAPHIC SCIENCE  • VOL. 11
                                                      E-4
                                   JULY 1973
369

-------
                    1232
                                                                                  1242
          11
          I
              47 *•
                                                                        j
                                                                     ''  ' '
 Figure 4.  EC chromatogram of Aroclor 1232 chromatographed
 on SE-30 with a Ni-63 detector operated in the DC mode. The
 peak identification numbers correspond to the retention time
                 relative to p.p'-DDE=100.
        Rgure 5a.  EC chromatogram of 03 ng Aroclor 1242 chromato-
        graphed  on SE-30 with a Ni-63 detector operated in the DC
        mode. The peak identification numbers correspond to  the
                  retention time relative to p,p'-DDE=100.
Table


RRT>
70
84
98
104

117
125

146
160

174
203

232
_244

280
332
372
448
528
Total
VI. Composition of Aroclor
Mean
Weight
Percent
2.7
4.7
3.8


3.3
12.3

14.1
4.9

12.4
9.3


9.8

11.0
4.2
4.0
.6
1.5
98.6

Relative
Std. Dev.»
6.3
1.6
3.5


6.7
3.3

3.6
2.2

2.7
4.0


3.4

2.4
5.0
8.6
25.3
10.2

1260

No. of
Chlorines*
5
5
1"
5 J60%
6J40%
6
5~|15%
6J85%
6
6~|50%
7J50%
6
en 10%
7J90%
ne
6 |10%
7J90%
7
7
8
8
8

  "Retention time relative to p,p'-DDE=100. Measured
from first appearance of solvent. Overlapping- peaks
that are quantitated as one peak are bracketed.
  ''Standard deviation of six results as a mean of the
results.
  'From GC MS data. Peaks containing mixtures of
isomers of different chlorine numbers are bracketed.
  •'Composition determined at the  center of peak 104.
  • Composition determined at the  center of peak 232.
                                                                                 1242
                                                                    n
                                                                  1  W"-.
                                                                             •   «   m
                                                           Figure 5b.  Pulsed mode EC chromatogram of Aroclor 1242
                                                           chromatographed on  SE-30 with a  tritium foil detector. The
                                                           peak identification numbers correspond to the  retention time
                                                                           relative to p,p'-DDE=100.
                                                                                            1248
       Figure 6.  EC chromatogram of Aroclor 1248 chromatographed
       on SE-30  with a Ni-63  detector operated in  the  DC mode.
       The peak  identification  numbers correspond to the retention
                    time relative to p,p'-DDE=100.
•nn
       Illl V 1977
B-5
                                                             JOURNAL OF CHROMATOGRAPHIC SCIENCE  •  VOL 11

-------
                                1254
                              t*o       m
 Figure 7.  EC chromatogram of Aroclor 1254 chromatographed
on SE-30 with a Ni-63 detector operated in the DC mode. The
peak identification numbers correspond to the retention time
                relative to p.p'-DDE=100.
                                1260
               r *• «'„
                                »  10

                                iV
                                ll
 Figure 9a.  DC mode EC chromatogram of a fat sample frc^i
 a turkey that had been fed fishmeal contaminated with A-o-
 clor 1242. The peak identification numbers correspond to the
          retention time relative to p,p'-DDE=100.
 Figure 9b.  Standard Aroclor 1242 run under the  same ccn-
                  ditions as Figure 9a.
Figure 8.  EC chromatogram of Aroclor 1260 chromatographed
on SE-30 with a Ni-63 detector operated in the DC mode. The
peak identification numbers correspond to the retention time
                relative to p,p'-DDE=100.
(ng  PCB'cm-1  for each  peak. Chromatograph the
sample and measure the area of each peak. Multiply
the area  of each peak  by the response factor for that
peak. Add the nanograms of PCB found in each peak
to obtain the total nanograms of PCB present.

Environmental Samples Containing Only One Aroclor

   An example of a sample containing a single Aroclor
that is partially metabolized or degraded  is seen in
Figure 9a. This is a chromatogram of fat extract from
a turkey that had been fed fishmeal contaminated with
Aroclor 1242. Figure 9b is standard Aroclor 1242 run
at the same conditions.
   Several peaks  present  in the standard  are  com-
pletely missing in the sample,  e.g., those at relative
retention times (RRT) 28 and  54. Some quantitation
methods  do not make  an adjustment when a peak is
missing from  the  sample.  For  example, one method
used  to quantitate  this sample  compared the sum of
.ill the major sample peak heights with the  sum of all
the peak  heights in the standard.  This approach  as-
sumes that all the  peaks present in the standard are
also present  in  the sample  and that all PCB  peaks
have  the same electron capture response. Neither  as-
sumption   is  valid.  Quantitation  by  calculating the
amount of PCB present in each individual peak is the
solution of this missing peak problem. Calculation by
the sum-of-heights  method  indicates 0.28 ng of PCB
is present; the individual peak  method gives 0.18 ng.
   The standard Aroclor (Figure 9b) shows the peak
at RRT  84  separated only as a  barely discernible
shoulder on peak 78; in the sample their proportions
 are different and  they elute as two  separate  peak?.
 Since individual values are  given for these peaks in
 Table III,  their individual  values can be calculated.
 In Figure 9, peaks 37 and  40 elute as a sincle peak
 for both the sample and the standard.  They are qu.in-
 titated by  combining  the values in  Table  III.  The
 sample peak at RRT 203 was not quantitated because
 no accurate comparison from the standard  was avail-
 able. Here, fortunately, the peak is  only a small  por-
 tion of the total and can be  ignored without seriously
 biasing the results.
    Peaks arising from pesticides can  be mistaken for
 PCB's; DDE and DDT were present in this sampl".
 The PCB peaks eluting in these RRT  regions are only
 about 4% of Aroclor 1242 and can be omitted from the
 total without causing major error. If  there is any qup*-
 tion about the presence of DDT, chlorinated naphtha-
 lenes, chlorinated  terphenyls,  or other  interferences,
 then the total PCB residues computed by this method
 can be confirmed by the derivative  technique • 3,  14
 Methods are also available to separate DDT-type pesti-
 cides  from PCB's if these results are necessary  3.
 15-17).

 Environmental  Samples Containing
 More Than One Aroclor

   Most  PCB contaminated samples of fi?h.  watc-r.
 and  sediment  contain  residues  of  several  Aroclor-.
 Usually the sample  chromatogram  can simply  }»• di-
 vided into three separate areas and peaks in each  arc-n
 quantitated  by using the appropriate Aroclor.  Pf.-.k-
 with  RRT 11-70 are  compared individually  to  cor-
 14.  Hutzinger. O.. Safe. S.. and Zitco. V..  Intern  J. F:i
     viron. Anal. Chem. 2.95 (1972».
 15.  Armour. Judith, and Burke. J.. J. Assoc Oiric. An,.
     Chemists 53. 7C1 (1970).
 1C.  Burke. J. A., J. Assoc. Offic  Anal. Clu-mi-i- 33 ^
     (1972).
 17.  Leoni, V.. J. Chromatog. 62. 03  (1971).
JOURNAL OF CHROMATOGRAPH 1C SCIENCE  •  VOL. 11
B-6
JULY 1973
                                                  371

-------
 responding peaks in Aroclor  1242, peaks 84-174 with
 Aroclor 1254, and peaks with larger RRT with Aroclor
 1260. These three Aroclors have been chosen for rou-
 tine use as standards because  they are the PCB's most
 often found in environmental samples, they were sold
 in largest quantities (particularly 1242 and 1254), and
 their chromatograms include all the EC peaks normally
 found in other Aroclor mixtures. In addition,  this sys-
 tem of dividing the chromatgojams generally matches
 the changes in chlorine numbers; i.e., the peaks used
 from the  1242  standard are the mono-through-tetra
 chlorobiphenyls,  the peaks calculated with  1254  are
 penta-  and hexachlorobiphenyls, and the peaks from
 the 1260 standard are used to measure the hepta- and
 octachlorobiphenyls.
    Figure 10 is the chromatogram of  a composite liver
 extract of three  bluegills from a PCB-contaminated
 river. This sample obviously contains residues of sev-
 eral Aroclors and was quantitated as described above.
 However, some sample chromatograms require a more
 rigorous division. A schematic of this division pro-
 cedure is shown in Figure 11.
   The logic of this division  is based on several key
 facts. Since Aroclor 1254 shows no appreciable peaks
 (see Figure 7) before RRT 47, any peaks with a lower
RRT indicate the  presence of some Aroclor  of  less
chlorine content. The most likely  compound, from ox-
perience  and commercial usage, is 1242.  Ideally, all
peaks through  RRT  78  would be calculated  against
1242  (though many of them  are  present in TJ54 as
well),  because peak  78  is unique to 1242 and  not
ordinarily discernible in 1254 (compare Figures 5  and
6). However, when chromatographic resolution is not
optimum, mixtures of the two Aroclors may show peak
78 as only a shoulder on peak 84. Experiments with
known mixtures,  described  below, show  that  in  this
case better results are obtained by making  the division
at RRT 70 and treating peak 78 + 84  as though it were
all peak 84 from 1254.
   A  peak with RRT 117  is present in Aroclor 1260.
but absent from 1254. The  presence or absence of  this
peak in the sample chromatogram  determines the stan-
dard used  to quantitate the remaining peaks. Since
some columns do not resolve peak 117, the investiga-
tor will not always  know that 1260 is present. Experi-
ments with mixtures show  that calculations  based on
1254 are  adequate in  this case.
   All peaks of RRT larger than 174 are calculated
with Aroclor 1260 because the relative standard devia-
tions for  these peaks are much lower in 1260  (Table
VI) than in 1254 (Table V). These rules  were tested
with known-weight-ratio  Aroclor  mixtures that were
chromatographed through a low resolution column with
the GC detector operated in the pulsed mode. Table
VII gives the amount of FCB measured as a percentage
Figure 10.  Division for quantitation of a DC mode EC chro-
matogram of PCB's from bluegill liver extract. The  column
substrate was SE-30. The  large peak at RRT 58 is an artifact
CM till tor
.«• tm f«r
CM tit* 1
i. WT m ^
             \      //
!• i
;**»
*!*•-*»*: tin
                                                         Table VII.  Percent Recoveries of Aroclor Mixtures
                                                                Using Chromatogram Division Rules
Mixture
—1242
! 1254
~1242
1254
—1242
1254
~1242
I 1254
f-1242
1260
~1254
1260
—1254
1260
~1254
1260
"1242
! 1254
' 1260
Weight
Ratio
1
1
3
1
4
1
5
1
2
1
1.5
1
1
1.5
1
4
4
1
2
% Re-
covery
1242"
98

96

96

99

103'
102"-'









%Re-
covery
1242"
103

100

100

102

107»
104-1-'









%Re-
covery
1254*










109

103

105

99»
105"

% Re-
covery
1254'1










101

94

96

99='
103"

       Figure 11.  Chromatogram Division Flowchart.
                                                       "Peaks through RRT 70 calculated as 1242 (Table III).
                                                       "Peaks through RRT 84 calculated as 1242 (Table III'».
                                                       cPeaks through RRT 104 calculated as 1254 (Table V >.
                                                       •"Peaks through RRT 174 calculated as 1254 (Table V>.
                                                       'All other peaks calculated as 1260 (Table VI).
372  •  JULY  1973
                                                B-7
                                                          JOItRNAI OF rHRnMiTnri&flDuir

-------
of th.it  injccli'd.  Thcsr rcsulls were c.-ilculatcd using
pi';ik  heights from the chromatograms.  Tables III, V.
and VI. and  the rules given above.
   These computation rules and  Tables  I-VI should
also  apply to chromatogram run on DC-200, OV-17
and  OV-101  columns.  However, QF-1 and  OV-225
elute  PCB's in a different order and the tables do not
apply.

Standard Solution Stability

   To test the stability to  UV light of Aroclor solu-
tions  in concentrations typically  used in GC analyses
118), several sealed glass ampoules of Aroclors 1242
and 1254 (10 ng/^1 isooctane) were prepared. Some
samples  were stored in the dark, some were continu-
ously  exposed two feet from a fluorescent light fixture
fitted  with a decorative clear plastic shield, and sev-
eral were stored in a window that  received  several
hours of sunshine each day. No  measurable  changes
in peak ratios were observed by EC-GC in five obser-
vations over two months' time. However, after identi-
cal samples were exposed to direct  sunlight for nine
dnys, some peak  ratios changed significantly. For  ex-
ample, in 1254 the apparent weight of material in RRT
peak  125 decreased  by 50% and RRT peak 104  in-
creased by 15%. Therefore, direct exposure to sunlight
should be avoided.
   Limited supplies of Aroclors 1242, 1254,  and 1260
as dilute isooctane  solutions in glass ampoules  are
available from  the authors as reference Aroclor kits.
Aroclors 1221,  1232, and  1248 are  also available if
there is a special  need.
                 Acknowledgment

   The authors wish to thank Alfred Thruston for the
use of the turkey fat chromatograms and William Loy
for the  fish liver chromatogram.                    •

           Manuscript received October 2,1972
18.  Safe. S.. and Hutzinger. O.. Nature 232. 641 (1971).
                                                    B-8

-------
Appendix C. Determination of Total PCB Emissions from Stationary Sources
                               (Draft Method)*
                PART A.  INDUSTRIAL, SEWAGE SLUDGE.  AND
                     MUNICIPAL REFUSE INCINERATORS
1.  Principle and Applicability

     1.1  Principle.  Gaseous and particulate PCBs are withdrawn isokinet-
ically from the source using a sampling train.  The PCBs are collected in
the Florisil adsorbent tube and in the impingers in front of the adsorbent.
The total PCBs In the train are determined by perchlorination to decachloro-
biphenyl (DCB) and gas chromatographic determination of the DCB.

     1.2  Applicability.  This method is applicable for the determination
of PCB emissions (both vaporous and particulate) from industrial, sewage
sludge, and municipal refuse incinerators.

2.  Range and Sensitivity

          The range of the analytical method may be expanded considerably
through concentration and/or dilution.  The total method sensitivity is also
highly dependent on the volume of gases sampled.  However, the sensitivity of
the total method as described here is about 10 ng DCB for each analytical
replicate.

3.  Interferences

          Excessive quantities of acid-resistant organics may cause signifi-
cant interferences obscuring the analysis  of DCB in the perchlorinated ex-
tracts.  Blphenyl, although unlikely  to be present in samples from combus-
tion sources, can form DCB in the perchlorination processes.

          Throughout all stages of sample handling and analysis, care should
be  taken to avoid contact of samples and extracts with synthetic organic
materials other than TFE® (polytetrafluoroethylene).  Adhesives must not be
used to hold TFE® liners on lids, and lubricating and sealing greases must
not be used on any sample exposed portions of the sampling train.

4.  Precision and Accuracy

          From sampling with identical and paired sampling trains,  the pre-
cision of the method has been determined to be 10 to 157. of the PCB concentra-
tion measured.  Recovery efficiencies on source samples spiked with PCB com-
pounds ranged from 85 to 957..
(*) Method found in reference 7a
                                     c-i

-------
5.  Apparatus

     5.1  Sampling Train.  See Figure A-l; a series of four impingers with a
solid adsorbent trap between the third and fourth impingers.  The train may
be constructed by adaptation from a Method 5 train.  Descriptions of the
train components are contained in the following subsections.

          5.1.1 Probe nozzle—Stainless steel (316) with sharp, tapered
leading edge.  The angle of taper shall be £ 30 degrees and the taper shall
be on the outside to preserve a constant internal diameter.  The probe noz-
zle shall be of the button-hook or elbow design, unless otherwise specified
by the Administrator.  The wall thickness of the nozzle shall be less than
or equal to that of 20 gauge tubing, i.e., 0.165 cm (0.065 in.) and the dis-
tance from the tip of the nozzle to the first bend or point of disturbance
shall be at least two times the outside nozzle diameter.  The nozzle shall
be constructed from seamless stainless steel tubing.  Other configurations
and construction material may be used with approval from the Administrator.

          5.1.2  Probe liner--Borosilicate or quartz glass equipped with a
connecting fitting that is capable of forming a leak-free,  vacuum tight con-
nection without sealing greases; such as Kontes Glass Company "0" ring spher-
ical ground ball joints (model K-671300) or University Research Glassware SVL
teflon screw fittings.

          A stainless steel (316) or water-cooled probe may be used for sam-
pling high temperature gases with approval from the Administrator.  A probe
heating system may be used to prevent moisture condensation in the probe.

          5.1.3  Pitot tube--Type S, or equivalent, attached to probe to
allow constant monitoring of the stack gas velocity.  The face openings of
the pitot tube and the probe nozzle shall be adjacent and parallel to each
other but not necessarily on the same plane, during sampling.  The free
space between the nozzle and pitot tube shall be at least 1.9 cm (0.75 in.).
The free space shall be set based on a 1.3 cm (0.5 in.) ID nozzle, which is
the largest size nozzle used.

          The pitot tube must also meet the criteria specified in Method 2
and be calibrated according to the procedure in the calibration section of
that method.

          5.1.4  Differential pressure gauge--Inclined manometer capable of
measuring velocity head to within 10% of the minimum measured value.  Below
a differential pressure of 1.3 mm (0.05 In.) water gauge, micromanometers
with sensitivities of 0.013 mm (0.0005 in.) should be used.  However,
                                     C-2

-------
                                                     Thermometer"

                                           Florisil Tube
        Probe (f^
Reverse-Type
Pi tot Tube
      Manometer
                                         Control Box
                                                                        	I
                    Figure A-l.   PCB Sampling Train for Incinerators
                                            C-3

-------
micromanometers are not easily adaptable to field conditions and are not
easy to use with pulsating flow.  Thus, other methods or devices acceptable
to the Administrator may be used when conditions warrant.

          5.1.5  Impingers--Four impingers with connecting fittings able to
form leak-free, vacuum tight seals without sealant greases when connected to-
gether as shown in Figure A-l.  The first and second impingers are of the
Greenburg-Smith design.  The final two impingers are of the Greenburg-Smith
design modified by replacing the tip with a 1.3 cm (1/2 in.) 10 glass tube
extending to 1.3 cm (1/2 in.) from the bottom of the flask.

          5.1.6  Solid adsorbent tube—Glass with connecting fittings able to
form leak-free, vacuum tight seals without sealant greases (Figure A-2).  Ex-
clusive of connectors, the tube has a 2.2 cm inner diameter, is at least 10 cm
long, and has four deep indentations on the inlet end to aid in retaining the
adsorbent.   Ground glass caps (or equivalent) must be provided to seal the
adsorbent-filled tube both prior to and following sampling.

          5.1.7  Metering system--Vacuum gauge, leak-free pump, thermometers
capable of measuring temperature to within 3°C (~ 5°F), dry gas meter with
2% accuracy at the required sampling rate, and related equipment, or equiv-
alent, as required to maintain an isokinetic sampling rate and to determine
sample volume.  When the metering system is used in conjunction with a pitot
tube, the system shall enable checks of isokinetic rates.

          5.1.8  Barometer--Mercury, aneroid, or other barometers capable
of measuring atmospheric pressure to within 2.5 mm Hg (0.1 in. Hg).  In many
cases, the barometric reading may be obtained from a nearby weather bureau
station, in which case the station value shall be requested and an adjust-
ment for elevation differences shall be applied at a rate of -2.5 mm Hg
(0.1 in. Hg) per 30 m (100 ft) elevation increase.

     5.2  Sample Recovery                                   ,

          5.2.1  Ground glass caps--To cap off adsorbent tube and the other
sample exposed portions of the train.

          5.2.2  Teflon FBI® wash bottle--Two, 500 ml, Nalgene No. 0023A59
or equivalent.

          5.2.3  Sample storage containers—Glass bottles, 1 liter, with
TFE®-lined screw caps.

          5.2.4  Balance—Triple beam, Ohaus Model 7505 or equivalent.

          5.2.5  Aluminum foil--Heavy duty.
                                      C-4

-------
                             J 28/12
    10cm
                           -2.5cm O.D.




                           -2.2cm I.D.
                             j28/12
Figure A-2.  Florisil Adsorbent Tube
                    C-5

-------
          5.2.6  Metal can--To recover used silica gel.

     5.3  Analysis

          5.3.1  Glass Soxhlet extractors--40 mi ID complete with 45/50 §
condenser, 24/40 <£ 250 ml round bottom flask, heating mantle for 250 ml
flask, and power transformer.

          5.3.2  Teflon FEP wash bottle—Two, 500 ml, Nalgene No. 0023A59
or equivalent.

          5.3.3  Separatory funnel--!,000 ml with TFE® stopcock.

          5.3.4  Kuderna-Danish concentrators--500 ml.

          5.3.5  Steam bath.

          5.3.6  Separatory funnel—50 ml with TFE® stopcock.

          5.3.7  Volumetric flask—25.0 ml, glass.

          5.3.8  Volumetric flask—5.0 ml, glass.

          5.3.9  Culture tubes—13 x 100 mm, glass with TFE®-lined screw caps,

          5.3.10  Pipette--5.0 ml glass.

          5.3.11  Aluminum block—Drilled to support culture tubes while
heating.                                                         ^

          5.3.12  Hot plate—Capable of heating to 200°C.

          5.3.13  Teflon®-glass syringe--! ml, Hamilton 1001 TLL or
equivalent with Teflon® needle.

          5.3.14  Syrlnge--10 ul, Hamilton 701N or equivalent.

          5.3.15  Gas chromatograph—Fitted with electron capture detector
capable of operation at 300°C and with 2 mm ID x 1.8 mm glass column packed
with 3% OV-210 on 100/120 mesh inert support (e.g., Supelcoport®).

          5.3.16  Electric muffle .furnace--Capable of heating to 650°C.

          5.3.17  Electric oven—Capable of heating to 150°C.

          5.3.18  Disposable glass pipettes with bulbs--To aid transfer of
the extracts.
                                    C-6

-------
          5.3.19  Porcelain casserole—Capable of withstanding temperatures
as high as 650°C.

6.  Reagents

     6.1  Sampling

          6.1.1  Florisil—Floridin Co., 30/60 mesh, Grade A.  The Florisll
is  cleaned  by 8 hr Soxhlet extraction with hexane and then by drying for
8 hr in an oven at 110°C and is activated by heating to 650°C for 2 hr (not
to exceed 3 hr) in a muffle furnace.  After allowing to cool to near 110°C
transfer the clean, active Florisil to a clean, hexane-uashed glass jar and
seal with a TFE®-lined lid.  The Florisil should be stored at 110°C until
taken to the field for use.  Florisil that has been stored more than 1 month
must be reactivated before use.

          6.1.2  Glass wool—Cleaned by thorough rinsing with hexane, dried
In a 110CC oven, and stored in a hexane-washed glass jar with TFE®-lined
screw cap.

          6.1.3  Water—Delonized, then glass-distilled, and stored in hexane-
rlnsed glass containers with TFE®-lined screw caps.

          6.1.4  Silica gel--Indicating type, 6-16 mesh.  If previously used,
dry at 175°C for 2 hr.  New silica gel may be used as received.

          6.1.5  Crushed ice.

     6.2  Sample Recovery

          6.2.1  Acetone—Pesticide quality, Burdick and Jackson "Distilled
in Glass" or equivalent, stored in original containers and used as received.

          6.2.2  Hexane--Pesticide quality, Burdick and Jackson "Distilled
in Glass" or equivalent, stored in original containers and used as received.

     6.3  Analysis

          6.3.1  Hexane--Pestlcide quality, Burdick and Jackson "Distilled
in Glass" or equivalent, stored in original containers and used as received.

          6.3.2  Acetone--Pesticide quality, Burdick and Jackson "Distilled
in Glass" or equivalent, stored in original containers and used as received.

          6.3.3  Water—Deionized and then glass-distilled, stored in hexane-
rlnsed glass containers with TFE®-lined screw caps.
                                   C-7

-------
          6.3.4  Sodium sulfate (Ns^SO^)- -Anhydrous, granular.   Clean by
overnight Soxhlet extraction with hexane, drying in a 110° C oven,  and then
heating to 650°C for 2 hr.  Store in 110°C oven or in glass Jar closed with
    - lined screw cap.
          6.3.5  Sulfuric acid (HjSO^)-- Concentrated,  ACS reagent grade or
equivalent.

          6.3.6  Antimony pentachloride (SbCl~) — Baker Analyzed Reagent or
equivalent.

          6.3.7  Hydrochloric acid (HC1) solution- -ACS reagent grade or
equivalent, 50% in water.

          6.3.8  Glass wool — Cleaned by thorough rinsing with hexane, dried
in a 110° C oven, and stored in a hexane-rinsed glass jar with TFE®- lined cap.

          6.3.9  Decachlorobiphenyl — RFP Corp.,  No.  RPC-60,  or equivalent.

          6.3.10  Compressed nitrogen- -Prepurif led.

          6.3.11  Carborundum boiling stones — Hengar Co. No. 133-B or equiv-
alent, rinsed with hexane.

7.  Procedure

     Caution:  Section 7.1.1 should be done in the laboratory.

     ^•1  Sampling.  The sampling shall be conducted by competent personnel
experienced with this test procedure and cognizant of  the constraints of the
analytical techniques for PCBs, particularly contamination  problems.

          7.1.1  Pretest preparation.  All train components  shall be main-
tained and calibrated according to the procedure described in APTD-0576,
unless otherwise specified herein.

               7.1.1.1  Cleaning glassware.  All glass parts of the train
upstream of and Including the adsorbent tube, should be cleaned as described
in Section 3A of the 1974 issue of "Manual of Analytical Methods for Analysis
of Pesticide Residues in Human and Environmental Samples."  Special care
should be devoted to the removal of residual silicone grease sealants on
ground glass connections of used glassware.  These grease residues should be
removed by soaking several hours in a chromic acid cleaning  solution prior
to routine cleaning as described above.
                                    C-8

-------
               7.1.1.2  Solid adsorbent tube.  Weigh 7.5 g of Florisil,  ac-
tivated within the last 30 days and still warm from storage in a 110°C oven,
into the adsorbent tube (pre-rinsed with hexane) with a glass wool plug in
the downstream end.  Place a second glass wool plug in the tube to hold the
sorbent in the tube.  Cap both ends of the tube with ground glass caps.   These
caps should not be removed until the tube is fitted to the train immediately
prior to sampling.

          7.1.2  Preliminary determinations.  Select the sampling site and
the minimum number of sampling points according to Method 1 or as specified
by the Administrator.  Determine the stack pressure, temperature, and the
range of velocity heads using Method 2 and moisture content using Approxi-
mation Method 4 or its alternatives for the purpose of making isokinetic
sampling rate calculations.  Estimates may be used.  However, final results
will be based on actual measurements made during the test.

          Determine  the molecular weight of the stack gases using Method 3.

          Select a nozzle size based on the maximum velocity head so that
isokinetic sampling  can be maintained at a rate less than 0.75 cfm.  It is
not necessary to change the nozzle size in order to maintain isokinetic
sampling rates.  During the run, do not change the nozzle size.

          Select a suitable probe length such that all traverse points can
be sampled.  Consider sampling from opposite sides for large stacks to re-
duce the length of probes.

          Select a sampling time appropriate for total method sensitivity
and the PCB concentration anticipated.  Sampling times should generally fall
within a range of 2  to 4 hr.

          It is recommended that a buzzer-timer be incorporated in the con-
trol box (see Figure 1) to alarm the operator to move the probe to the next
sampling point.

          In some circumstances, e.g., short batch processes, it may be
necessary to sample  through two or more batches to obtain sufficient sample
volume.  In these cases, sampling should cease during loading/unloading of
the furnace.

          7.1.3  Preparation of collection train.   During preparation and
assembly of the sampling trcin, keep all train openings where contamination
can enter covered until Just prior to assembly or until sampling is about to
begin.  Immediately  prior to assembly, rinse all parts  of the train upstream
of the adsorbent tube with hexane.
                                      C-9

-------
Mark the probe with heat resistant tape or by some other method at points
indicating the proper distance into the stack or duct for each sampling
point.

          Place 200 ml of water in each of the first two ispingers, and
leave the third impinger empty.  CAUTION:  do not use sealant greases in
assembling the train.  If the preliminary moisture determination shows that
the stack gases are saturated or supersaturated, one or two additional empty
impingers should be added to the train between the third iepinger and the
Florisil tube.  See Section 10.1.  Place approximately 200 to 300 g or more,
if necessary, of silica gel in the last impinger.  Weigh each impinger (stem
included) and record the weights on the impingers and on the data sheet.

          Unless otherwise specified by the Administrator, attach a tempera-
ture probe to the metal sheath of the sampling probe so that the sensor is
at least 2.5 cm behind the nozzle and pitot tube and does not touch any
metal.

          Assemble the train as shown in Figure A-l.  Through all parts of
this method use of sealant greases such as stopcock grease to seal ground
glass joints must be avoided.

          Place crushed ice around the impingers.                        """

          7.1.4  Leak check procedure--After the sampling train has been as-
sembled, turn on and set (if applicable) the probe heating system(s) to reach
a temperature sufficient to avoid condensation in the probe.  Allow time for
the temperature to stabilize.  Leak check the train at the sampling site by
plugging the nozzle and pulling a 380 mm Hg (15 in. Hg) vacuum.  A leakage
rate in excess of 47. of the average sampling rate of 0.0057 Q3/min (0.02 cfm)
whichever is less, is unacceptable.

          The following leak check instruction for the sacpling train de-
scribed in AFTD-0576 and APTD-0581 may be helpful.  Start the pump with by-
pass valve fully open and coarse adjust valve completely closed.  Partially
open the coarse adjust valve and slowly close the bypass valve until 380 mm
Hg (15 in. Hg) vacuum is reached.  Do not reverse direction of bypass valve.
This will cause water to back up into the probe.  If 380 cc Hg (15 in. Hg)
is exceeded, either leak check at this higher vacuum or end the leak check
as described below and start over.

          When the leak check is completed, first slowly renove the plug
from the inlet to the probe and immediately turn off the vacuum pump.  This
prevents the water in the impingers from being forced backward into the
probe.
                                     C-10

-------
          Leak checks shall be conducted as described above prior to each
test run and at the completion of each test run.  If leaks are found to be
In excess of the acceptable rate, the test will be considered invalid.  To
reduce lost time due to leakage occurrences, it is recommended that leak
checks be conducted between port changes.

          7.1.5  Train operation—During the sampling run, an isokinetic
sampling rate within 10%, or as specified by the Administrator, of true iso-
kinetic shall be maintained.  During the run, do not change the nozzle or
any other part of the train in front of and including the Florisil tube.

          For each run, record the data required on the data sheets.  An
example is shown in Figure A-3.  Be sure to record the initial dry gas meter
reading.  Record the dry gas meter readings at the beginning and end of each
sampling time increment, when changes in flow rates are made, and when sam-
pling is halted.  Take other data point readings at least once at each sam-
ple point during each time increment and additional readings when significant
changes (20% variation in velocity head readings) necessitate additional ad-
justments in flow rate.  Be sure to level and zero the manometer.

          Clean the portholes prior to the test run to minimize chance of
sampling deposited material.  To begin sampling, remove the nozzle cap,
verify (if applicable) that the probe heater is working and up to tempera-
ture, and that the pitot tube and probe are properly positioned.  Position
the nozzle at the first traverse point with the tip pointing directly into
the gas stream.  Immediately start the pump and adjust the flow to isokinetic
conditions.  Nomographs are available for sampling trains using type S pitot
tubes with 0.85 + 0.02 coefficients (C_), and when sampling in air or a stack
gas with equivalent density (molecular weight, M^, equal to 29+4), which
aid in the rapid adjustment of the isokinetic sampling rate without excessive
computations.  APTD-0576 details the procedure for using these nomographs.
If C  and Md are outside the above stated ranges, do not use the nomograph
unless appropriate steps are taken to compensate for the deviations.

          When the stack is under significant negative pressure (height of
impinger stem), take care to close the coarse adjust valve before inserting
the probe into the stack to avoid water backing into the probe.  If neces-
sary, the pump may be turned on with the coarse adjust valve closed.

          When the probe is in position, block off the openings around the
probe and porthole to prevent unrepresentative dilution of the gas stream.

          Traverse the stack cross section, as required by Method 1 or as
specified by the Administrator.  To minimize chance of extracting deposited
material, be careful not to bump the probe nozzle into the stack walls when
sampling near the walls or when removing or inserting the probe through the
portholes.
                                    C-ll

-------
                                                                             FIELD DATA
                             PLANT.
                             DATE_
                             SAMPLING LOCATION.
                             SAMPLE TYPE	
                             RUN NUMBER	
                             OPERATOR 	
                             AMBIENT TEMPERATURE .
                             BAROMETRIC PRESSURE .
                             STATIC PRESSURE. IPSI_
                             FILTER NUMBER Is)	
                                                                     SCHEMATIC OF TRAVERSE POINT LAYOUT
                                                PROBE LENGTH AND TYPE.
                                                NOZZLE 1.0	
                                                ASSUMED MOISTURE.'.	
                                                SAMPLE BOX NUMBER	
                                                METER BOX NUMBER	
                                                METER AH,,	
                                                C FACTOR	
                                                PROBE HEATER SETTING _
                                                HEATER BOX SETTING	
                                                REFERENCE ip	
                                                              READ AND RECORD ALL DATA EVERY.
                                           MINUTES
TRAVERSE
POINT
NUMBER























^S. CLOCK TIME
«jp*NXp3ctt
TIME.min \
' ~~~- 	 ~ 	 	























GAS METER READING
IVmi. II3
























VELOCITY
HEAD
(APjl, m H20























ORIFICE PRESSURE
DIFFERENTIAL
IAHI. in. HjOl
DESIRED























ACTUAL























STACK
TEMPERATURE
•T...T























DRY GAS METER
TEMPERATURE
INLET

-------
          During the test run, make periodic adjustments to keep the probe
temperature at the proper value.  Add more ice and, if necessary, salt to
the ice bath, to maintain a temperature of less than 20°C (68"F) at the
impinger/silica gel outlet, to avoid excessive moisture losses.  Also, peri-
odically check the level and zero of the manometer.

          If the pressure drop across the train becomes high enough to make
isokinetic sampling difficult to maintain, the test run should be terminated.
Under no circumstances should the train be disassembled during a test run to
determine and correct causes of excessive pressure drops.

          At the end of the sample run, turn off the pump, remove the probe
and nozzle from the stack, and record the final dry gas meter reading.  Per-
form a leak check.*  Calculate percent isokinetic (see calculation section)
to determine whether another test run should be made.  If there is difficulty
in maintaining isokinetic rates due to source conditions, consult with the
Administrator for possible variance on the isokinetic rates.

           7.1.6  Blank  train—For each series of test runs,  set up a blank
train  in a manner identical to  that described above, but with  the nozzle
capped with aluminum  foil and the exit end of the  last  impinger capped with
a ground glass cap.  Allow the  train to remain assembled  for a period equiv-
alent  to one  test run.  Recover the blank sample as described  in Section  7.2.

     7.2   Sample recovery.  Proper cleanup procedure begins  as soon as the
probe  is removed from the stack at the end of the  sampling  period.

          When the  probe  can  be safely handled, wipe off  all external par-
ticulate matter near  the  tip  of the probe nozzle.  Remove  the  probe from  the
train  and  close off both  ends with aluminum  foil.  Cap  off  the inlet  to the
train  with a  ground glass cap.

           Transfer  the  probe  and  impinger assembly to the  cleanup area.   This
area should be clean  and  protected from the  wind so that  the chances  of con-
taminating or losing  the  sample will be minimized.

           Inspect the train prior to and during disassembly  and note  any  ab-
normal conditions.  Treat the samples as follows:

           7.2.1  Adsorbent tube—Remove the  Florisil tube from the train  and
cap it off with ground  glass  caps.
*  With  acceptability  of  the  test  run  to be based  on  the same criterion as
      In  7.1.4.

                                      C-13

-------
          7.2.2  Sample container No. I—Remove the first three impingers.
Wipe off the outside of each impinger to remove excessive water and other
debris, weigh (stem included), and record the weight on data sheet.  Pour
the contents directly into container No. 1 and seal.

          7.2.3  Sample container No. 2--Rinse each of the first three im-
pingers sequentially first with 30 ml acetone and then with 30 ml hexane,
and put the rinses into container No. 2.  Quantitatively recover material
deposited in the probe using 100 ml acetone and then 100 ml hexane and add
these rinses to container No. 2 and seal.

          7.2.4  Silica gel container--Remove the last impinger, wipe the
outside to remove excessive water and other debris, weigh (stem included),
and record weight on data sheet.  Transfer the contents to the used silica
gel can.

     7.3  Analysts.  The analysis of the PCB samples should be conducted by
chemical personnel experienced in determinations of trace organics utilizing
sophisticated, instrumental techniques.   All extract transfers should be
made quantitatively by rinsing the apparatus at least three times with hex-
ane and adding the rinses to the receiving container.  A boiling stone should
be used in all evaporative steps to control "bumping."

          7.3.1  Extraction

               7.3.1.1  Adsorbent tube.   Expel the entire contents of the
adsorbent tube directly onto a glass wool plug in the sample holder of a
Soxhlet extractor.  Although no extraction thimble is required, a glass
thimble with a coarse-fritted bottom may be used.

               Rinse the tube with 5 ml acetone and then with 15 ml hexane
and put these rinses into the extractor.  Assemble the extraction apparatus
and extract the adsorbent with  170 ml hexane for at least 4 hr.  The ex-
tractor should cy^cle 10 to 14 times per hour.  After allowing the extrac-
tion apparatus to cool to ambient temperature, transfer the extract into a
Kuderna-Danish evaporator.

               Evaporate the extract to about 5 ml on a steam bath and
allow the evaporator to cool to ambient temperature before disassembly.
Transfer the extract to a 50-ml separatory funnel and set the funnel aside.

               7.3.1.2  Sample  container No. 1.  Transfer the aqueous sam-
ple to a 1,000-ml separatory funnel.  Rinse  the container with 20 ml acetone
and then with two 20-ml portions of hexane, adding the rinses to the sep-
aratory funnel.
                                     C-14

-------
               Extract the sample with three 100 ml portions of hexane,
transferring the sequential extracts to a Kuderna-Danish evaporator.

               Evaporate the extract to about 5 ml and allow the evaporator
to cool to ambient temperature before disassembly.  Filter the extract through
a micro column of anhydrous sodium sulfate into the 50 ml separatoty  funnel
containing the corresponding Florisil extract.  The micro column is prepared
by placing a small plug of glass wool in the bottom of the large portion of
a disposable pipette and then adding anhydrous sodium sulfate until the tube
is about half full.

               7.3.1.3  Sample container No. 2.  Transfer the organic solu-
tion into a 1,000 ml separatory funnel.  Rinse the container with two 20 ml
portions of hexane and add the rinses to the separatory funnel.  Wash the
sample with three 100 ml portions of water.  Discard the aqueous layer and
transfer the organic layer to a Kuderna-Danish evaporator.

               Evaporate the extract to about 5 ml and allow the evaporator
to cool to ambient temperature before disassembly.  Filter the extract through
a micro column of anhydrous sodium sulfate into the SO ml separatory funnel
containing the corresponding Florisil and impinger extracts.

          7.3.2  Extract cleanup--Clean the combined extracts (in SO  ml
separatory funnel) by shaking with 5 ml concentrated sulfuric acid.  Allow
the acid layer to separate and drain it off.

          Transfer the hexane layer to a Kuderna-Danish evaporator and evap-
orate to about 5 ml.  Allow the evaporator to cool to ambient temperature
before disassembly.

          The extract should be essentially colorless.  If it still shows
significant color, additional cleanup may be required before assaying for
PCBs.  In this event, further clean the extract by liquid chromatography on
Florisil according to procedures described in Section 5A of the 1974  issue
of "Manual of Analytical Methods for Analysis of Pesticide Residues in Human
and Environmental Samples"  Reduce the Florisil eluant to about 10 ml by
Kuderna-Danish evaporation techniques described above.

          Transfer the cleaned extract to a 25 ml volumetric flask and di-
lute to volume with hexane.  Pipette three 5.0 ml aliquots into culture
tubes for perchlorination.  Retain the remaining 10 ml for later verifica-
tion, If required (see Section 10.2).

          7*3*3  Extract perchlorination—Evaporate the aliquots in the cul-
ture tubes just to dryness with a gentle stream of dry nitrogen. If the ali-
quots will not evaporate to dryness, refer to Section 10.3 concerning special
cases. Add 0.2 ml antimony pentachloride with a 1 ml glass-TFE® syringe and

                                    C-15

-------
seal the tube with a TFEf®-lined screw cap. Heat the reaction mixture to 160°C
for 2 hr by placing the tube in a hole in an aluminum block on a hot plate.

          Allow the tube to cool to ambient room temperature before adding
about 2 ml of 50% HC1 in water to destroy residual antimony pentachloride.
This is a convenient "stopping point" in the perchlorination procedure.

          Extract the reaction mixture by adding about 1 ml hexane to the
tube, shake, and allow layers to separate.  Remove the upper hexane layer
with a disposable pipette and filter through a micro column of anhydrous
sodium sulfate directly into a 5 ml volumetric flask.   Repeat the extraction
three times for a total of four extractions.  Dilute the extract to volume
with hexane.

          7.3.4  PCS determination—Assay the perchlorinated extracts for
decachlorobiphenyl (DCB) by gas chromatographic comparison with DCB stan-
dard solutions and correct this result for the DCB concentration determined
for the blank train.  (Column temperature and carrier gas flow parameters
of 240°C and 30 ml/min, are typically appropriate.  The concentrations of the
standard solutions should allow fairly close comparison with DCB in the sam-
ple extracts.  Standards near 25 to 50 picograms/microliter may be appropriate.)

8.  Calibration

          Maintain a laboratory log of all calibrations.

     8.1  Sampling Train

          8.1.1  Probe nozzle--Using a micrometer, measure the inside dia-
meter of the nozzle to the nearest 0.025 mm (0.001 in.).  Make three separate
measurements using different diameters each time and obtain the average of
the measurements.  The difference between the high and low numbers shall not
exceed 0.1 mm (0.004 in.).

          When nozzles become nicked, dented, or corroded, they shall be re-
shaped, sharpened, and recalibrated before use.

          Each nozzle shall be permanently and uniquely identified.

          8.1.2  Pitot tube--The pitot tube shall be calibrated according
to  the procedure outlined in Method 2.

          8.1.3  Dry gas meter and orifice meter—Both meters shall be  cali-
brated according to the procedure outlined  in APTD-0576.  When diaphragm
                                      C-16

-------
pumps with bypass valves are used, check for proper metering system design
by calibrating the dry gas meter at an additional flow rate of 0.0057 m^/min
(0.2 cfm) with the bypass valve fully opened and then with it fully closed.
If there is more than + 27, difference in flow rates when compared to the fully
closed position of the bypass valve, the system is not designed properly and
must be corrected.

          8.1.4  Probe heater calibration—The probe heating system shall be
calibrated according to the procedure contained in APTD-0576.  Probes con-
structed according to APTD-0581 need not be calibrated if the calibration
curves in APTD-0576 are used.

          8.1.5  Temperature gauges--Calibrate dial and liquid filled bulb
thermometers against mercury-in-glass thermometers.  Thermocouples should
be calibrated in constant temperature baths.

     8.2  Analytical Apparatus

          8.2.1  Gas chromatograph--Prepare a working curve from at least
five standard injections of different volumes of the DCB standard.

9.  Calculations

          Carry out calculations, retaining at least one extra decimal fig-
ure beyond that of the acquired data.  Round off figures after final calcu-
lations .

     9.1  Nomenclature

     G_ = Corrected weight of DCB in nth perchlorinated aliquot (n = 1, 2, 3), pg.

     G  = Total weight of PCBs (as DCB) in sample, ug.
      s

     C  = Concentration of PCBs in stack gas, ug/nr*, corrected to standard
            conditions of 20°C, 760 mm Hg (68°F, 29.92 in. Hg) on dry basis.

     A  = Cross-sectional area of nozzle, nr (ft2).
      n

    B   = Water vapor in the gas stream, proportion by volume.
     W8
      I = Percent of isokinetic sampling.

     It. = Molecular weight of water, 18 g/g-mole (18 Ib/lb-mole).

   Pbar = Barometric pressure at the sampling site, mm Hg (in. Hg).


                                    C-17

-------
     P_ = Absolute stack gas pressure) ran Hg (in. Hg).
      s

   P  . = Standard absolute pressure, 760 mm Hg (29.92 in Hg).

      R = Ideal gas constant, 0.06236 nm Hg-m3/°K-g-mole (21.83 in.
            Hg-ft3/°R-lb-mole).

     Tm = Absolute average dry gas  meter temperature  °K  (°R).

     T8 = Absolute average stack gas  temperature  °K  (°R).

   Tstd = standard absolute temperature,  293°K  (528°R).

    Vlc - Total volume of liquid collected in impingers  and  silica  gel, ml.
            volume of water collected equals  the  weight  increase  in grams
            times 1 ml/gram

     V  = Volume of gas sample as measured by dry gas meter, dcm (dcf).

Vm(std) = Volume of gas sample measured by the dry gas meter corrected to
            standard conditions, dscm (dscf).

V /stcj) = Volume of water vapor  in the gas sample corrected to standard
            conditions, son  (scf).

     Vt = Total volume of sample, ml.

     V  = Stack gas velocity, calculated by EPA Method 2, m/sec (ft/sec).
      S

     AH = Average pressure differential across the orifice meter, mm
            (in. H20).

     pw" = Density of water,  1 g/ml (0.00220 Ib/ml).

      6 = Total Sampling time, min.

    13.6 = Specific gravity of mercury.

     60 = Sec/min.

     100 - Conversion  to percent.
                                     C-18

-------
     9.2  Average drygas meter temperature and average orifice pressure
drop. See data sheet (Figure A-3).

     9.3  Pry gas volume.  Correct the sample volume measured  by  the dry
gas meter to standard conditions [20°C, 760 mm Hg (68°F,  29.92 in. Hg)] by
using Equation A-l).
          V«i(std) " v,
                        Lstd
           ra
pbar
h AH
13.6
?8td
+ AH
= K V , bar 13.6
T
m
where     K = 0.3655 °K/nm Hg for metric units

            = 17.65 °R/in. Hg for English units

     9»4  Volume of water vapor
Vw(std) B vic  rr
                             RT8td
                                                   Equation A-l
                                                    Equation A-2
where     K = 0.00134 m /ml for metric units

            = 0.0472 ft3/ml for English units

     9.5  Moisture content
          BW8  -
                     Vstd)
                  Vm(std) + Vw(8td)
                                         Equation  A-3
     If the liquid droplets are present  in the  gas  stream assume the stream
to be saturated and use a psychrometric  chart to obtain an approximation
of the moisture percentage.
                                   C-19

-------
     9.6  Concentration

          9.6.1  Calculate the total PCB residue (as DCB)  in the  sample  from
the weights of DCB in the perchlorinated aliquots according to Equation  A-4.

          G  = 5(Gi + G2 + G3)                               Equation A-4
          9.6.2  Concentration of PCBs (as  DCB)  in stack gas.  Determine  the
concentration of PCBs in the stack gas according to Equation A-5.
                   P
          C  = K    8                                        Equation A-5
           8     v
                  m(std)

where      K - 35.31 ft3/m3
     9.7  Isokinetic variation

          9.7.1  Calculations from raw data.


                    100  T8tKVlc+  (Vm/Tm)  (Pbflr) + AH/13.6)]
                                   60 8 vs  P8  An

                                                            Equation A-6


     where     K = 0.00346 mm Hg-m3/ml-°K for metric units

                 = 0.00267 in. Hg-ft3/ml-°R for English units

          9.7*2  Calculations from intermediate values.


               r -     Ts  Vstd)  *std  10°
                             6  An  Ps  60
                         Ts  Vm(std)	
                     *s  vs  An     (!~BWS)                  Equation A-7
     where     K = 4.323 for metric units

                 = 0.0944 for English units


                                     C-20

-------
     9.8  Acceptable results. The following range sets the limit on accept-
able isokinetic sampling results:

          If 907. < I < 110%, the results are acceptable. If the results are
low in comparison to the standards and I is beyond the acceptable range, the
Administrator may option to accept the results.

10.  Special Cases

     10.1  Sampling moisture saturated or supersaturated stack gases. One
or two additional modified Greenburg-Smith impingers may be added to the
train between the third impinger and the Florisil tube to accommodate addi-
tional water collection when sampling high moisture gases. Throughout the
preparation, operation, and sample recovery from the train, these additional
impingers should be treated exactly like the third impinger.

     10.2  PCS verification. It is recommended that an unperchlorinated
aliquot from at least one sample be subjected to GC/MS examination to verify
that PCB isomers are present.

          To accomplish this, the unperchlorinated portion of each extract
is first screened by GO with the same chromatographic system used for DCB
determination except for a cooler column temperature, typically 165 to 200°C.
The elution patterns are compared with those of commercial PCB mixtures (in
hexane solution) to determine the most similar mixture.

          After determining what PCB isomers are possible present, the sam-
ple is examined by GC/MS using multiple ion selection techniques for ions
characteristic of the molecular clusters of the PCBs possibly present.

     10.3  Evaporation of extracts for perchlorination. For cases where the
extract will not evaporate to dryness or excessive PCB loss by volatiliza-
tion is suspected, the hexane may be removed by azeotrophic evaporation from
the hexane/chloroform mixture.

          Add 3 ml of chloroform to the aliquot in the culture tube. Add
a boiling chip and concentrate by slow boiling in a water bath to 1 ml.
Repeat the chloroform addition and evaporation three times in order to remove
all residual hexane. Then further concentrate (slowly) to a volume of ap-
proximately 0.1 ml. Under no circumstances should the water bath tempera-
ture be permitted to exceed 76 C or the solvent be evaporated to dryness.
The final volume (0.1 ml) may be determined with sufficient accuracy by
comparison of .solvent level with another reaction vial containing 0.1 ml
of chloroform. When a volume of 0.1 ml is achieved, cap the reaction vial
immediately and allow to cool. Proceed with the perchlorination as described
in Section 7.3.3.
                                 C-21

-------
11.  References

          Martin, Robert M., "Construction Details  of Isokinetic Source
Sampling Equipment," Environmental Protection Agency, Air Pollution Control
Office Publication No. APTD-0581.

          1973 Annual Book of ASTM Standards. Part  23, Designation:  D 1179-72,

          Thompson, J. F., Ed.,  "Analysis  of Pesticide Residues  in  Human and
Environmental Samples," Environmental Protection Agency,  Research Triangle
Park, N.C., 1974.
                                     C-22

-------
          PART B.  CAPACITOR- AND TRANSFORMER-FILLING PLANTS
1.  Principle and Applicability

     1-1  Principle.  Gaseous and particulate PCBs are withdrawn isokinet-
Ically from the source.  The PCBs are collected on Florisil and determined
by gas chromatography against an Aroclor® standard.

     1.2  Applicability.  This method is applicable for the determination
of PCB emissions from the room air, room air exhaust and process point ex-
hausts at capacitor- and transformer-filling plants.

2.  Rangeand Sensitivity

          The range of the analytical method may be expanded considerably
through concentration and/or dilution of the extract.  The total method
sensitivity is also  highly dependent on the volume of gases sampled.  How-
ever, sensitivity of the total method is near 1 ug per test or near 10 ng
per test where the perchlorination assay method is used.

3.  Interferences

          Throughout all stages of sample handling and analysis, care should
be taken to avoid contact of samples and extracts with synthetic organic ma-
terials other than TF^® (polytetrafluoroethylene).  Lubricating and sealing
greases should not be used on the sample exposed  portions of the sampling
train.

4.  Precision and Accuracy

          Sampling with identical and paired sampling trains, the precision
of the method should be 10 to 15% of the PCB concentration measured.  Re-
covery efficiencies on source samples spiked with PCB compounds ranged from
85 to 95% of the spike.

5.  Apparatus

     5.1  Sampling Train.  The sampling train, see Figure B-l, consists of a
glass-lined probe, an adsorbent tube containing Florisil, and the appropriate
valving and flow meter controls for isokinetic sampling as described in Part A
of the procedure.  The sampling apparatus in Figure B-l is the same as that in
Figure A-l and Section 5.1 of Part A, except that the Smith-Greenburg impingers
and heated probe are not used.  If condensation of significant quantities of
moisture prior to the solid adsorbent is expected, Part A of the method should
be used.  Since probes and adsorbent tubes are not cleaned up in the field, a
sufficient number must be provided for sampling and allowance for breakage.

                                    C-23

-------
                       Probe (to sample from duct) «*—I
                               Glass-lined Probe
                                                               Florisil

                                                               Glass Wool
                     Integrated  1
                     Flow Meter I
                                                                 Check Valve
Manometer
                                                                      Vacuum
                                                                      Line
       Figure  B-l.   PCB Sampling Train for Capacitor- and
                    Transformer-Filling Plants
                                   C-24

-------
     5.2  Sample Recovery.  Heavy duty aluminum foil must be provided to
cap off the probe prior to shipment.

     5*3  Analysis.  The equipment required for the analysis is identical to
that specified in Part A except that the equipment necessary for perchlorina-
tion of the PCBs collected to the decachlorobiphenyl form is not required.
(Perchlorination of the sample here is optional and should be employed only
if the GC fingerprint technique of this procedure is not applicable.)

6.  Reagents

     6.1  Sampling

          6.1.1  Florisil—Floridin Company, 30/60 mesh, Grade A.  The Flori-
sil is cleaned by overnight Soxhlet extraction with hexane and then drying
overnight at 110°C and is activated by heating to 650°C for 2 hr (not to ex-
ceed 3 hr) in a muffle furnace.  After allowing to cool to near 110°C, trans-
fer the clean, active Florisil to a clean, hexane-washed glass jar and seal
with a TFlf®-lined lid.  The Florisil should be stored at 110°C until taken
to the field for use.  Florisil that has been stored more than 1 month must
be reactivated.

          6.1.2  Glass wool—Cleaned by thorough rinsing with hexane, dried
in a 110°C oven, and stored in a hexane-washed glass jar with TF^®-lined
screw cap.

     6.2  Analysis

          6.2.1  Hexane--Pesticide quality, Burdick and Jackson "Distilled
in Glass" or equivalent, stored in original containers and used as received.

          6.2.2  Acetone—Pesticide quality, Burdick and Jackson "Distilled
in Glass" or equivalent, stored in original containers and used as received.

          6.2.3  Sodium sulfate (Na2804)—Anhydrous, granular.  Clean by
overnight Soxhlet extraction with hexane, drying in a 110°C oven, and then
beating to 650°C for 2 hr.  Store in 110°C oven or in glass jar closed with
TF^®-lined screw cap.

          6.2.4  Sulfuric acid (H2SO^>--Concentrated, ACS reagent grade or
equivalent.

          6.2.5  Glass wool—Cleaned by thorough rinsing with hexane, dried
in a 110°C oven, and stored in a hexane-rinsed glass jar with TFI^-lined cap.
                                      C-25

-------
          6.2.6  Carborundum boiling stones--Hengar Company No. 133-B or
equivalent, rinsed with hexane.

          6.2.7  Standard Aroclor PCB mixtures--Aroclors® 1016, 1221, 1232,
1242, 1248, 1254, 1260, and 1262 may be obtained from the Pesticide Reposi-
tory, EPA/HERL/ETD, Research Triangle Park, North Carolina.

7.  Procedure

     7-1  Sampling.  The sampling shall be conducted by competent personnel
knowledgeable with this test procedure and cognizant of the constraints of
the analytical techniques for PCBs, particularly contamination problems.

          The sampling procedure for capacitor and transformer plants is
identical to that described in Part A with the following exceptions:  (a)
"impingers and a beatable probe are not required prior to the adsorbent
tube; and (b) the PCB concentrations may be considerably higher for ca-
pacitor and transformer plants, compared to most incinerators, thus the
sampling time can be less than the 2 hr specified in Part A.

          The selection of sampling time and rate should be based on the
approximate levels of PCB residues expected in the sample.  The sampling
rate should not exceed 14 liters/min and may typically fall in the range
of 5 to 10 liters/min.  Sampling times should be more than 20 min but
should not exceed 4 hr.

          Because the processes for filling the capacitors and transformers
can vary significantly between plants, isokinetic sampling is required in
the procedure.  However, if it can be shown to the satisfaction of the
Administrator that isokinetic sampling is not necessary, then sampling at
a proportional rate is an acceptable alternative.  Proportional or constant
flow rate sampling may also be necessary in cases where the standard pitot/
nozzle, assembly physically blocks a significant portion of the stack or
where the flow rate is too low (less than 10 ft/min) for the pitot tube.

     7.2  Sample Recovery

          7.2.1  Adsorbent tube--Remove the Florisil tube from the collec-
tion system and cap it off with ground glass caps for shipment to the ana-
lytical laboratory.

          7.2.2  Probe  (where applicable)—Remove the probe from the col-
lection system and cap  it off with aluminum foil.
                                      C-26

-------
     7*3  Analysis.  The analysis of the FOB samples should be conducted by
chemical personnel experienced in determinations of trace organics utilizing
sophisticated instrumental techniques.  All extract transfers should be made
quantitatively by rinsing the apparatus at least three times with hexane and
adding the rinses to the receiving container.  A boiling stone should be used
in all evaporative steps to control "bumping."

          7.3.1  Extraction

               7.3.1.1  Adsorbent tube.  Expel the entire contents of the
adsorbent tube directly onto a glass wool plug in the sample holder of a
Soxhlet extractor.  Although no extraction thimble is required, a glass
thimble with a coarse-fritted bottom may be used.

               Rinse the tube with about 5 ml acetone and then about 15 ml
hexane into the extractor.  Assemble the extraction apparatus and extract
the adsorbent with 170 ml hexane for at least 4 hr.  The extractor should
cycle 10 to 14 times per hour.  After allowing the extraction apparatus to
cool to ambient temperature, transfer the extract into a Kuderna-Danish
evaporator.

               Evaporate the extract on a steam bath to about 5 ml and al-
low the evaporator to cool to ambient temperature before disassembly.  Trans-
fer the extract to a 50 ml separatory funnel and set the funnel aside.

               7.3.1.2  Probe (where applicable).  Rinse the probe with hex-
ane into a Kuderna-Danish evaporator.  Evaporate the extract to about 5 ml
and allow the evaporator to cool to ambient temperature before disassembly.
Add the concentrated extract to the 50-ml separatory funnel containing the
corresponding Florisil extract.

          7.3.2  Extract cleanup—Clean the combined extracts (in 50-ml
separatory funnel) by shaking with 5 ml concentrated sulfuric acid.  Allow
the acid layer to separate and drain it off.

          Transfer the hexane layer to a Kuderna-Danish evaporator and evap-
orate to about 5 ml.  Allow the evaporator to cool to ambient temperature
before disassembly.

          The extract should be essentially colorless.  If it still shows
significant color, additional cleanup may be required before assaying for
PCBs.  In this event, further clean the extract by liquid chromatography on
Florisil according to procedures described in Section 5A of the 1974 issue
of "Manual of Analytical Methods for Analysis of Pesticide Residues in Human
and Environmental Samples."  Reduce the Florisil eluant to about 10 ml by
Kuderna-Danish evaporation techniques described above.

                                    C-27

-------
          Transfer the cleaned extract to a 25-ml volumetric flask and dilute
to volume with hexane for gas chromatographic analysis.

          7.3.3  PCB determination—Assay the cleaned extracts by gas chromato-
graphic comparison with standard solutions of a similar commercial PCB mixture
(A column temperature between 165 and 200°C at a flow rate of 30 ml/min may be
appropriate. Aroclot®standard solutions at concentrations near 10 ng/ul should
be appropriate for calibration of the gas chromatograph.)  If PCB mixtures were
being used at the sampling site, a standard solution of that mixture, e.g.,
Aroclor® 1016, will likely be appropriate.  Quantitation should be based on the
summed areas of at least five major peaks coincident in the chromatograms of
the sample extracts and standards.  The range and sensitivity of the method
may be extended somewhat by diluting concentrated extracts with hexane or
concentrating dilute extracts by evaporation under a gentle stream of dry
nitrogen. If the sample chromatograms do not closely resemble a particular
PCB standard, e.g., in the case of emissions from more than one Aroclor®
product, refer to Section 10.1 concerning Special Cases. Correct the PCB
assays for PCBs determined in the blank train.

8.  Calibration

          Maintain a laboratory log of all calibrations.

     8.1  Sampling Train

          8.1.1  Probe nozzle--Using a micrometer, measure the inside diameter
of the nozzle to the nearest 0.025 mm (0.001 in.).  Make three separate mea-
surements using different diameters each time and obtain the average of the
measurements.  The difference between the high and low numbers shall not ex-
ceed 0.1 mm (0.004 in.).

          When nozzles become nicked, dented, or corroded, they shall be re-
shaped, sharpened, and recalibrated before use.

          Each nozzle shall be permanently and uniquely identified.

          8.1.2  Pitot tube—The pitot tube shall be calibrated according
to the procedure outlined in Method 2.

          8.1.3  Dry gas meter and orifice meter—Both meters shall be cali-
brated according to the procedure outlined in APTD-0576.  When diaphragm
pumps with bypass valves are used, check for proper metering system design
by calibrating the dry gas meter at an additional flow rate of 0.0057 m3/min
(0.2 cfm) with the bypass valve fully opened and then with it fully closed.
If there is more than + 2% difference in flow rates when compared to the
fully closed position of the bypass valve, the system is not designed properly
'ind must be corrected.
                                     C-28

-------
          8.1.4  Temperature gauges—Calibrate dial and liquid filled bulb
thermometers against mercury-in-glass thermometers.  Thermocouples need not
be calibrated.  For other devices, check with the Administrator.

     8.2  Analytical Apparatus

          8.2.1  Gas chromatograph--Prepare a working curve from at least
five standard injections of different volumes of the Aroclor® standard in
hexane solution.

9.  Calculations

          Carry out calculations, retaining at least one extra decimal fig-
ure beyond that of the acquired data.  Round off figures after final calcu-
lations.

     9.1  Nomenclature

     Gs " Total weight of Aroclor® in sample, ug.

     Cg = Concentration of Aroclor® in stack gas, ug/m3, corrected to
            standard conditions of 20°C, 760 mm Hg (68°F, 29.92 in. Hg).

     An - Cross-sectional area of nozzle, m* (£t2).

     I  = Percent of isokinetic sampling.

   Pfcar = Barometric pressure at the sampling site, mm Hg (in. Hg).

     Pg = Absolute stack gas pressure, mm Hg (in. Hg).

   P8td •= Standard absolute pressure, 760 mm Hg (29.92 in Hg).

     R  = Ideal gas constant, 0.06236 mm Hg-m3/°K-g-mole (21.83 in.
            Hg-ft3/°R-lb-mole).

     Tm = Absolute average dry gas meter temperature °K (°R).

     T8 » Absolute average stack gas temperature °K (°R).

   Tstd e Standard absolute temperature, 293°K (528eR).

     VQ * Volume of gas sample as measured by dry gas meter, dcm (dcf).
                                     C-29

-------
        * Volume of gas sample measured by the dry gas meter corrected to
            standard conditions, dscm (dscf ) .

     V8 = Stack gas velocity, calculated by Method 2, Equation 2 to 7, m/sec
            (ft/sec).

     AH = Average pressure differential across. the orifice meter, mm H2<)
            (in. H2<>).

      9 «= Total sampling time, min.

   13.6 • Specific gravity of mercury.

     60 m Sec /min.

    100 - Conversion to percent.

     9.2  Average dry gas meter temperature and average orifice pressure
drop.
     9.3  Dry Gas Volume.  Correct the sample volume measured by the dry
gas meter to standard conditions [20°C, 760 mm Hg (68°F, 29.92 in. Hg)]
by using Equation B-l.
        vn(8td)
                     Tstd
                      im
        AH
?bar + 13.6
    *std
                                                      AH
KVm Pbar + 13.6
                                                  T,
                                                   m
                                                             Equation B-l
where   K = 0.3855 °K/mm Hg for metric units
          = 17.65 °R/in. Hg for English units

     9.4  Concentration

          9.4.1  Concentration of Aroclor® in stack gas.  Determine the
concentration of Aroclor® in the stack gas according to Equation B-2.
                             C8 = ___2 -                    Equation B-2
                                  vm(std)
                                    C-30

-------
10.  Special Cases

     10.1  Quantitation of PCS Residues Not Similar to a Commercial Mixture.
In cases where the composition of the PCB residue does not closely resemble
an available commercial PCB mixture, i.e., from comparison of EC-GC chromato-
grams, direct quantitation against available standard mixtures may be diffi-
cult and inaccurate.  These extracts should be split, perchlorinated, and
total PCBs quantitated by procedures described in Part A, Sections 7.3.2,
7.3.3, and 7.3.4, and the total PCB residue of the sample calculated from
Equation A-4.

     10.2  PCB. Verification.  It is recommended that an unperchlorinated
aliquot from at least one sample be subjected to GC/MS examination to ver-
ify that PCB isomers are present.

          After determining what PCB isomers are possibly present by the
quantitation procedures in Section 7.3.3, the sample is examined by GC/MS
using multiple ion selection techniques for ions characteristic of the
molecular clusters of the PCBs possibly present.

11.  Reference

          Martin, Robert M., "Construction Details of Isokinetic Source
Sampling Equipment," Environmental Protection Agency, Air Pollution Control
Office of Publication No. APTD-0581.

          1973AnnualBook of ASTM Standards, Part 23, Designation:  D 1179-72

          Thompson, J. F., Ed., "Analysis of Pesticide Residues in Human and
Environmental Samples," Environmental Protection Agency, Research Triangle
Park, N.C., 1974.
                                     C-31

-------
                               TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
 1. REPORT NO.
 EPA-600/7-79-047
                          2.
                                                     3, RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
 Measurement of PCB Emissions from Combustion
   Sources
            6. REPORT DATE
            February 1979
            6. PERFORMING ORGANIZATION CODE
7. AUTHOH(S)
 P.L.Levins, C.E.Rechsteiner, and J.L.Stauffer
                                                     8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts  02140
                                                     10. PROGRAM ELEMENT NO.
            INE624
            11. CONTRACT/GRANT NO.

            68-02-2150, T.D. 10102
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: 12/76 - 12/78
            14. SPONSORING AGENCY CODE
             EPA/600/13
15. SUPPLEMENTARY NOTES
2557.
                                              ^    3  JohllSOn , MD-62, 919/541-
16. ABSTRACT
The report describes a gas chromatographic/mass spectrometric (GC/MS) procedure
that overcomes problems encountered when using GC procedures (previously used
to determine polychlorinated biphenyls (PCBs)  in solids and water) on emissions
from combustion sources. The GC/MS procedure, which relies on selected mass
scanning in restricted regions of the chromatograms, was developed because in
the combustion process the distribution pattern of the individual PCBs  changes,
rendering invalid the pattern matching approach used with the gas chromatographic/
electron capture detection (GC/ECD) method.
 7.
                            KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                         b.lDENTIFIERS/OPEN ENDED TERMS
                          COS AT I Field/Group
 Pollution
 Chlorine Aromatic Compounds
 Biphenyl              Mass Spectroscopy
 Combustion
 Measurement
 Gas Chromatography
Pollution Control
Stationary Sources
Polychlorinated Bi-
 phenyls
13B
07C

21B
14B
07D
18. DISTRIBUTION STATEMENT

 Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
    88
20. SECURITY CLASS (This pagel
Unclassified
                        22. PRICE
EPA Form 2220-1 (9-73)
                                      C-32

-------