SPA/600/4-78/048
  EFft
    ital Protection
Environmental Monitoring and Support . EPA-600 • 4-78-048
Laboratory          August 1978
Research Triangle Park NC 27711
            Research and Development
A Method for
Sampling and
Analysis of
Polychlorinated
Biphenyls (PCB's) in
Ambient Air

-------
                RESEARCH REPORTING SERIES

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

      1.  Environmental  Health Effects Research
      2.  Environmental  Protection Technology
      3.  Ecological Research
      4  Environmental  Monitoring
      5  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7.  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

This  report has been assigned to the ENVIRONMENTAL MONITORING series.
This  series describes research conducted to develop new or improved methods
and  instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations.  It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia  22161.

-------
   A (lETi-iOO FOR THE SAMPLING MD ANALYSIS Of
POLYCHLORINATEO dlPHEMYLS (PCBs) IN AMBIENT AIR
                 Prepared by

              Charles L. Stratton
              Stuart A. Whitlock
                 J .  r4ark Al 1 an

  ENVIRONMENTAL SCIENCE AND ENGINEERING, INC.
      P.O. BOX 13454, UNIVERSITY STATION
          GAINESVILLE, FLORIDA  32604
          EPA Contract No. 68-01-2978
               January 3U, 1978
                     for
           QUALITY ASSURANCE BRANCH
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
      OFFICE OF RESEARCH AND DEVELOPMENT
     U.S. ENVIRONMENTAL PROTECTION AGENCY
       RESEARCH TRIANGLE PARK, NC  27711

                      and

          OFFICE OF TOXIC SUBSTANCES
     U.S. ENVIRONMENTAL PROTECTION AGENCY
            WASHINGTON, D.C.  20460
             EPA PROJECT OFFICERS:
             John H. Hargeson
             EPA, EHSL Research Triangle Park
             George E. Parris
             Vincent J. JeCarlo
             EPA Office of Toxic Substances

-------
This report has been reviewed by the Office of Research and Development,
EPA, and approved for publication.  Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names of commercial products
constitute endorsement of recommendation for use.
                                   11

-------
                            TABLE OF CONTENTS
                                                                  Page
     ACKNOWLEDGEMENT	  iv
     ABSTRACT	   v
1.0  INTRODUCTION	   1
     1.1  COMPLEXITY Of AIRBORNE PCB MEASUREMENT	   1
     1.2  REQUIREMENTS OF A METHOD TO MEASURE AIRBORNE PCB	   3
2.U  RESULTS AND DISCUSSION	,	   5
     2.1  SELECTION OF THE SAMPLING METHOD	   6
     2.2  DESCRIPTION OF THE SAMPLING APPARATUS	   8
     2.3  ANALYTICAL PROTOCOL	  14
          2.3.1  CLEANUP OF MATERIALS AND APPARATUS	  14
          2.3.2  FIELD SAMPLING	  17
          2.3.3  SAMPLE EXTRACTION	  21
          2.3.4  SAMPLE CONCENTRATION	  21
          2.3.5  SAMPLE CLEANUP	  22
          2.3.6  GAS CHROMATOGRAPHIC ANALYSIS	  22
          2.3.7  PERCHLORINATION	  27
          2.3.8  PERCHLORINATION RUGGEDNESS TEST	  38
3.0  RECOVERY STUDIES AND FIELD TESTS	  44
     3.1  RECOVERY STUDIES	  44
     3.2  PRELIMINARY FIELD TESTS	  69
     3.3  FIELD TESTS OF THE FINAL METHOD	  78
          3.3.1  TESTS IN GAINESVILLE, FLORIDA	  78
          3.3.2  TESTS IN NEW BEDFORD, MASSACHUSETTS	  88
4.0  CONCLUSIONS AND RECOMMENDATIONS	 103
5.0  REFERENCES	 105
     APPENDIX A--ANALYTICAL PROCEDURE	 107

-------
ACKNOWLEDGEMENT
The authors wish to gratefully acknowledge the project direction and
valuable technical contributions of the Co-Project Officers, John H.
Margeson of the Environmental  Monitoring and Support Laboratory, Quality
Assurance Branch, Research Triangle Park, and George E. Parris of the
Office of Toxic Substances.  In addition, valuable technical advice was
provided by Robert G. Lewis, Analytical Chemistry Branch, Health Effects
Research Laboratory, Research Triangle Park, and Vincent J. DeCarlo of
the Office of Toxic Substances.  Special thanks is extended to Terry F.
Bidleman who conceived of the use of porous polyurethane foam for
collection of airborne PCBs, and who provided invaluable assistance to
the authors in the early stages of this effort.
                                    iv

-------
ABSTRACT
A method has been developed for the sampling and analysis of polychlo-
rinated biphenyls (PCBs) in air.  An easily constructed, high-volume
sampling system is employed with porous polyurethane foam as the collec-
tion medium.  The sample is collected at the rate of 0.6 to 1.0 m3
per minute.  Laboratory procedures described in this document permit the
quantitative analysis of even the most volatile PCB species in an air
sample.  A perchlorination technique for the quantitative analysis of
PCB has been adapted for use.  The technique is shown to convert even
the most volatile PCB species to decachlorobiphenyl for simple and
direct quantitative analysis.  Data is presented to show conversion
efficiencies of a variety of PCBs to decachlorobiphenyl of 101 _+
6 percent over the range of 0.103 to 10.0 ug.  A ruggedness test was
conducted which indicates the proposed perchlorination technique can
yield reliable inter!aboratory results.  The perchlorination technique
is generally necessary for the analysis of low (i.e., less than 25 ng/m3)
airborne levels of PCB.  The analytical method is effective for the
analysis of airborne PCB levels within at least the range of 1 ng/m3 to
50 ug/m3.  The mean collection efficiency for Aroclor 1016/1242 was
101 jf 10 percent for sampling periods of 20 minutes to 12 hours and air
volumes of 16 to 720 cubic meters.  Field tests of the method under a
variety of ambient conditions are described.  This report was submitted
as part of fulfillment of contract 68-01-2978 by Environmental Science
and Engineering, Inc. under sponsorships of the Environmental Protection
Agency.

-------
                             List of Tables

Table                                                             Pag_e

   1     Perchlorination Recovery vs. Concentration
           Test Resul ts	  38

   2     Eight Combinations of Seven Factors Used
           to Test the Ruggedness of an Analytical  Method	  39

   3     Perchlorination Ruggedness Test Conditions	41

   4     Perchlorination Ruggedness Test Format and Data	  42

   5     Perchlorination Ruggedness Test Results	  43

   6     Molecular Composition of Some Aroclors (After
           Hutzinger, Safe, and Zitko, 1974)	  45

   7     Results of Airborne PCB Collection Efficiency
           Study Using Aroclor 1221	  49

   8     Results of Aroclor 1016 Recovery Studies	  52

   9     PCB Test Mi xture Composi tion	  56

  1U     Results of Initial Recovery Study Using
           PCB Test Mixture	  57

  11     Results of 2-Hour Recovery Studies Using TCB,
           DiCB, and MCB	  58

  12     Results of 12-Hour Recovery Studies Using
           Aroclor 1242, MCB, and DiCB	  60

  13     Results of 6-Hour Recovery Studies Using
           Aroclor 1242	  62

  14     Results of 2-Hour Recovery Studies Using
           Aroclor 1242	  66

  15     Summary of Recovery Tests	  68

  16     PCB in Ambient Urban Atmosphere	  73

  17     Results of High-Volume Ambient Air Sample Taken
           in the Vicinity of an Electrical Substation and
           Transformer Storage Facility	  77
                                   VI

-------
                        List of Tables  (cont'd)

Table                                                            Page

  18     Results of Field Tests Conducted in
           Gainesville,  Florida	 79

  19     Results of Field Tests Conducted in
           New Bedford,  Massachusetts,  in June,  1977	 89

  20     Results of Field Tests Conducted in
           New Bedford,  Massachusetts  in  January,  1978	99
                                 VII

-------
                            List of Figures

Figure                                                            Page

   1     Assembled Sampler and Shelter With Exploded
           View of the Filter Holder	,	  9

   2     Exploded View of Hi-Vol Sampling Apparatus
           Designed For Use With a Rotometer Showing
           Modified Throat to Accomodate Polyurethane
           Foam PI ugs	 10

   3     Analytical Scheme for PCB in Ambient Air	 15

   4     Apparatus for Drying Polyurethane Foam Plugs	 18

   5     Chromatogram of a Typical Unexposed Polyurethane
           Foam Extract at Recorder Attenuation 4 x 1Q~9	 19

   b     Chromatogram Of a Typical Glass Fiber Filter
           (Gelman Type A) Extract at Recorder Attenuation
           4 x 10-9.	 20

   7(a)  Chromatogram of an Urban Air Sample, First
           Polyurethane Foam Plug on OV-17/QF-1 Column	 24

   7(b)  Chromatogram of Second Polyurethane Foam
           Plug on OV-17/QF-1 Column	 25

   7(c)  Chromatogram of Third Polyurethane Foam
           Plug on OV-17/QF-1 Column	 26

   7(d)  Aroclor 1221 on OV-17/QF-1 Column	 27

   8(a)  Chromatogram of an Urban Air Sample, First
           Polyurethane Foam Plug on SE-30/OV-210 Column	 28

   b(b)  Chromatogram of Aroclor 1221 on SE-30/OV-210 Column	29

   9     Chromatogram of a Perchlorinated Air Sample	 32

  10     Perchlorination Reaction Vial	 34

  11     Chromatograms of the Aroclor 1221
           Collection Efficiency Study	 47

  12     Aroclor 1016 Recovery Experiment.	 51

  13     Chromatograms From an Aroclor 1016
           Collection Efficiency Study	 54

  14     GC/EC Chromatogram of Aroclor 1242
           Standard (Attn:  16 x lO"9)	 63
                                  VI 1 1

-------
                     List of Figures (cont'd)

                                                               Page

15(a) GC/EC Chromatogram of First PDF Plug Extract
        (Attn:  16 x 10~9) from 6-Hour Aroclor 1242
        Recovery Study	  64

15    GC/EC Chromatogram of (b) Second and (c) Third
        PUF Plug Extracts (Attn:  16 x 1CT9)  from
        6-Hour Aroclor 1242 Recovery Study	  65

16    Minimum Ambient Airborne PCB Concentrations
        in an Urban Environment	  70

17    Chromatogram of an Urban Ambient Air Sample	  72

18    (a)  Standard Aroclor 1016; (b) Chromatograph
        of Air Sample Taken in the Vicinity of a
        Major Industrial User	  75

19    GC/MS Analysis of an Ambient Air Sample  Taken in
        the Vicinity of a Major Industrial User of PCB	  76

20    (a)  GC/EC Chromatogram of an Unexposed PUF Plug
      (b)  GC/EC Chromatogram of the Glass Fiber Filter
        Used During the Field Test on 7/26/77	  60

21    GC/EC Chromatogram of an Aroclor 1242 Standard	  81

22    GC/EC Chromatogram of the First PUF Plug Extract
        (Attn:  8xlO-9)  From the 2-Hour Ambient Air
        Sample Taken 7/26/77 in Gainesville,  Florida	  82

23    GC/EC Chromatogram of (a) the Second and (b) the
        Third PUF Plug Extracts (Attn:  8xlO'9) From
        the 2-Hour Ambient Air Sample Taken 7/26/77 in
        Gainesville, Florida	  83

24    (a)  GC/EC Chromatogram of the First PUF  Plug Extract
        (Attn:  4xlO~9)  From tfie 2-riour Ambient Air Sample
        Taken 10/31/77 in Gainesville, Florida; (b) GC/EC
        Chromatogram of the Same Sample (Attn:  16xlO~9)
        After Perch!orination	  85

25    GC/EC Chromatogram (Attn:  16xlO"9) of 6-Hour
        Ambient Air Sample (PUF #1) Taken 11/20/77 in
        Gainesville, Florida	  86

26    GC/EC Chromatograms of (a) Second PUF Plug, (b)
        Third PUF Plug,  and (c) Glass Fiber Filter
        (Attn:  16xlU~9) of 6-Hour Ambient Air Sample
        Taken 11/20/77 in Gainesville, Florida	  87
                                IX

-------
                        List of Figures (cont'd)

Figure                                                            Page

   27    GC/EC Chron,_togram of the Filter Extract
           (Attn:  16xlO"9) From the 60-Minute Ambient Air
           Sample Taken at New Bedford Landfill	  90

   28    GC/EC Chromatogram (Attn:  16xlQ-9) of an
           Aroclor 1242 Standard	  91

   29    GC/EC Chromatogram of the First PUF Plug Extract
           (Attn:  16xlO-9 Diluted to 1/8) of the 60-Minute
           Ambient Air Sample Taken at New Bedford Landfill	  92

   30    GC/EC Chromatogram of the Second PUF Plug Extract
           (Attn:  16xlO-9) of the 60-Minute Ambient Air
           Sample Taken at New Bedford Landfill	  93

   31    GC/EC Chromatogram of the Third PUF Plug Extract
           (Attn:  16x10-9) of the 60-Minute Ambient Air
           Sample Taken at New Bedford Landfill	  94

   32    GC/FID Chromatogram of Biphenyl at 2.8 ng	  96

   33    GC/FID Chromatogram of the First PUF Plug Extract
           From the 60-Minute Ambient Air Sample Taken at
           New Bedford Landfi 11	  97

   34    GC/EC Chromatogram of Air Sample Taken over
           New Bedford, Massachusetts, Landfill in
           January, 1978	 100

   35    GC/EC Chromatogram of Ambient Air Samples Taken
           (a) Upwind and  (b) Downwind of a PCB User Facility
           in New Bedford, Massaschuetts, in January, 1978	 101

-------
l.U  INTRODUCTION
     Polychlorinated biphenyls (PCBs) have been detected by numerous research-
     ers in all  environmental  media.  Significant levels of contamination have
     been detected in waters,  aquatic sediments, soils, and biota.  Data is
     conspicuously limited, however, for the assessment of airborne PCB con-
     tamination.  Since the atmosphere may serve as both a transport medium
     for PCB from a source to  the environment and as a mechanism for direct
     human exposure, it is important to evaluate airborne PCB levels.  No com-
     monly accepted method is  presently available for the sampling of airborne
     PCB.  Investigations were, therefore, undertaken to identify and evaluate
     a usable ambient air sampling method with adequate detection limit, sen-
     sitivity and specificity  to reliably assess airborne pollution by this
     class of compounds.  The  purpose of this report is to describe one such
     method for the measurement of airborne levels of PCB and to present data
     documenting the performance of the method under laboratory and field
     conditions.

1.1  COMPLEXITY OF AIRBORNE PCB MEASUREMENT
     There exist a number of complicating factors which must be taken into
     consideration in the measurement of airborne PCB levels.  Principal among
     these are:

     a.  The term "PCB" applies, not to a single chemical species, but to a
         class of chemical compounds.  These compounds are related by chemical
         structure and by degrees of halogen substitution on the molecule.
         PCB is seldom manufactured or used in the pure isomeric state.  For
         industrial application, PCBs are marketed in the form of mixtures,
         each mixture containing a significant fraction of the 209 possible
         isomeric species.  For example, Aroclor® 1242, a commercial PCB
         mixture, is comprised of 54 identified isomers (Hutzinger, Safe and
         Zitko,  1974).  Such mixtures have been detected in numerous environ-
         mental  media.  The fact that a class of chemical compounds and not a
         single identifiable chemical species is to be detected and measured
         greatly multiplies the complexity of the chemical analysis in an
         already complex environmental medium such as the ambient atmosphere.

-------
b.  Our concern is with extremely low airborne concentrations of PCB.
    Measurements reported to date indicate ambient levels are in the
    ng/m3 or part per trillion (ppt) range.

c.  Not only are airborne PCB levels expected to be low, but they must be
    detected in the presence of a large number of other organic compounds
    which may be present in much higher concentrations than the PCB.
    Organic compounds from both natural and manmade sources will compli-
    cate PCB collection and analysis.

d.  Analytical methods have been developed for the analysis of PCB in
    most environmental media.  These methods generally employ gas-liquid
    chromatography (GC).  They have been developed after years of lab-
    oratory work required to perfect extraction, concentration, cleanup,
    and analytical techniques.  There are numerous variations to each of
    these techniques, but they all entail lengthy, exacting methods of
    sample handling and treatment.  Since the present interest is with
    airborne PCB, it may be expected to be necessary to deal with the
    more volatile species in this class of compounds.  The  more volatile
    species have proven to be more difficult to handle analytically than
    the less volatile, or more highly substituted, PCB species, and
    hence, require even more attention to analytical detail.

e.  The state-of-the-art of quantisation of PCB by gas chromatographic
    analysis is not well-developed.  Currently used methods are sub-
    jective.  This leads to considerable variability among analysts.

There has always been difficulty in accurately quantifying PCB present
in environmental media.  In most cases, the multipeak GC elution pattern
observed after separation of pesticides and other interfering compounds
by extensive extract cleanup procedures is subjectively compared to the
elution pattern of a commercial PCB mixture.  A decision is made as to
what commercial product the sample elution pattern most closely repre-
sents.  The quantity of PCB present is then calculated by comparison of
several matching GC peaks between the  sample and a known concentration of

-------
     the commercial  product.  Inherent in this method of quantisation is the
     assumption that all  PCB isomers are present in the sample in the same
     proportion as in the commercial mixture chosen for quantitation purposes.

     There is no reason to believe the GC elution pattern of ambient airborne
     PCB would closely resemble that of a commercial  product.  The more
     volatile PCB species of a given commercial mixture would likely be
     selectively enriched in the airborne phase as compared to the solvent
     phase.  This phenomenon of selective airborne enrichment of the lower-
     substituted PCB isomer has been demonstrated by Mieure (1975) for vapors
     above a standard Aroclor 1016 mixture.  One would expect even greater
     apparent enrichment of more volatile species as samples are collected
     further from the source of the airborne contamination.

     The most common method of PCB analysis, using the electron capture detec-
     tor in conjunction with gas chromatography, introduces still another
     complicating factor.  The response factor or sensitivity of the various
     PCB isomers in  a given mixture will vary by several orders of magnitude
     depending upon  not only the degree of chlorination, but on the position
     of the chlorine atom on the PCB molecule (Zitko, Hutzinger and Safe,
     1971).  Tne response is not linear with substitution.  The analyst must,
     therefore, be absolutely certain he is comparing, for quantitative
     purposes, GC peaks generated by the identical isomer.

1.2  REQUIREMENTS OF A METHOD TO MEASURE AIRBORNE PCB
     Any method wnich is selected for the measurement of airborne PCB must
     conform to a demanding set of performance criteria.  The most important
     of these criteria are dictated by the anticipated difficulty in analyzing
     airborne PCB.  These criteria are:

     a.  The selected method must incorporate the sampling of large volumes of
         air.  To avoid excessively lengthy sampling periods, it was con-
         sidered important that the sampling method chosen be a high-volume
         technique.   This is necessary to achieve the desired low detection
         limit for the ambient atmosphere within a reasonable sampling period.

-------
b.  The chosen sampling method must demonstrate a high collection
    efficiency for airborne PCB.  A sampling technique which approaches
    quantitative collection of the airborne PCB is essential to achieve
    the desired low detection limit, precision, and accuracy of the
    measurement.

c.  Since one of the complicating factors which must be overcome in the
    measurement of ambient airborne PCB is the presence of other organic
    materials in the atmosphere, many in much higher concentration than
    the PCB, the sampling and analytical techniques must be specific to
    PCB.  The difficulties encountered  in accurately quantifying PCB in
    environmental media have been discussed.  These difficulties are even
    more greatly magnified when contaminants, which are not of immediate
    interest, are present in the final  analytical extract.  When this is
    the case, the multipeak chromatogram is increasingly difficult to
    interpret.

d.  It is desired to select a method that is relatively easy to perform
    and that lends itself to standardization.  This applies to both the
    sampling activity in the field and  to laboratory analytical
    procedures.

e.  The selected method should permit qualitative assessment of the PCB
    present as well as quantitative assessment.  It should be possible to
    ascertain the general distribution  of the total PCB present according
    to species.  For example, if the ambient PCB should resemble a
    specific commercial mixture, this information is desired in order to
    evaluate the potential source of the contamination and to assess the
    potential health and environmental  hazard.

f.  It would be highly preferred that the selected method had been shown
    by previous researchers to perform  well.  This would greatly reduce
    the effort necessary to adapt the method of routine use.

-------
RESULTS AND OISCUSSIUN
A review was conducted of potential sampling and analytical techniques
tor airborne organic pollutants that would apply to PCB.  A sampling
method and analytical methods were selected for evaluation based on the
reported experience of previous researchers.  A high-volume sampling
method using polyurethane foam as  the collection medium was considered to
be tne candidate method that most  closely matched the desired criteria.
This sampling method was combined  with analytical methods reported by
researchers to be effective in determining low concentrations of PCB in
environmental media, to arrive at  a complete step-by-step procedure for
assessing airborne PCB concentrations.

The selected method was evaluated  in the laboratory and in the field.
The evaluation consisted of an analytical recovery study, a sampling
efficiency study, an analytical ruggedness test, a number of preliminary
field tests, and a final field test.  The analytical recovery study was
necessary to establish procedures  and to overcome difficulties encounterd
in the laboratory analysis of the  airborne samples collected on poly-
urethane foam so that quantitative results could be achieved.  Collection
efficiency studies were conducted  to identify the limits of the high-
volume sampling technique selected.  Preliminary field tests were neces-
sary to obtain experimental  samples of the ambient atmosphere during
method development.  These field data were used to evaluate the method
from the standpoint of the practicality of application under field condi-
tions.  A final field test was conducted to demonstrate the application
of the method.

-------
2.1  SELECTION OF THE SAMPLING METHOD
     It was desired to select a sampling technique for airborne PCB that has
     been evaluated by other researchers and that conforms as closely as
     possible to the performance criteria set forth above.  A review of pub-
     lished literature indicated some experience had been reported in the use
     of wet impingers, solid adsorbents, silicone oil, and polyurethane for
     the collection of airborne PCB.

     Harvey (1974) employed a sampling device consisting of a tube packed with
     silicone-oil coated ceramic saddles through which a large volume of air
     can be drawn at the rate of 0.6 m3/min.  Harvey determined the coll lection
     efficiency of this method using trichlorobiphenyl, hexachlorobiphenyl,
     and Aroclor 1254.  He tested the collection efficiency by applying known
     quantities of these PCB species and the commercial mixture to a glass
     fiber filter which preceeded the adsorbent medium in the collection sys-
     tem.  The average collection efficiency was 70 percent.  Concentrations
     in the ambient atmosphere over the western North Atlantic of 0.05 to
   .  5.3 ng/m^ were reported by Harvey using this method.  He expressed
     the opinion that the method can only approximate PCB concentrations in
     ambient air.

     Giam, Chan, and Neff (1975) recommend a method for the detection of PCB
     in air which makes use of a solid adsorbent, Florisil, packed in a short
     glass column.  Air is pumped through this column at the rate of 2 to
     4 1/min.  By placing several of these sample columns in series, it was
     found that PCB spiked onto the first column was retained only on the
     first column for sampling time as long as 60 hours.  Using this arrange-
     ment to sample laboratory atmosphere, Giam, Chan, and Neff reported
     levels of 35 to 90 ng/m^ PCB.  No attempt was made to identify the
     PCB species present.

     Bidleman and Olney (1974a) reported ambient concentrations of PCB over
     the western North Atlantic and Rhode Island of 0.21 to 9.4 ng/m3.

-------
These researcners (Bidleman and Olney, 1974b) devised a high-volume
sampling method employing polyurethane foam  as the collection medium.
Cylindrical polyurethane foam plugs, of 1U cm thickness, are placed  in a
high-volume air sampler behind the standard  filter holder.  Air is drawn
through the sampling apparatus at the rate of 0.5 to 1.0 m^/tnin.  The
authors evaluated the collection efficiency  of this medium by placing the
sampling device near the source of tri-, tetra-, and pentachlorobiphenyl.
The device was equipped with two polyurethane foam plugs.  A glass fiber
filter was placed in the filter holder.  In  this manner, these research-
ers determined less than 1 percent of the PCB isomers tested was trapped
on the glass fiber filter and 96 to 99 percent of the total collected was
trapped on the first polyurethane foam plug.  Bidleman and Olney at  the
same time, and in the same manner, evaluated the collection efficiency of
Greenburg-Smith impinger systems charged with ethylene glycol.  They
found this impinger system trapped 75 to 82 percent of the total tetra-
and pentachlorobiphenyl  while that remaining passed through the sampling
system.  Bidleman and Olney concluded, based on this preliminary evalu-
ation, that the polyurethane foam was an effective collection medium for
airborne PCB.

The high-volume sampling method of Bidleman  and Olney (1974b) was
selected for evaluation and application to assessing ambient airborne
levels of PCB.  Of the methods investigated by previous researchers, this
one appeared to most closely meet the performance requirements set forth
above for the following reasons:

a.  It is a high-volume method that should provide the capacity of
    determining low ambient levels of airborne PCB without excessively
    long sampling periods.

b.  A high collection efficiency has been demonstrated for the tri-,
    tetra-, and pentachlorobiphenyl species.

c.  The polyurethane foam used as a collection medium has been used  for
    the concentration of PCB in natural waters.  It was selected for this

-------
         purpose,  partly  because  it demonstrates a degree of selectivity for
         PCB.   If  this  proved to  be true  for airborne organic pollutants, it
         would be  of considerable advantage, since this technique would
         provide a means  to selectively increase the  ratio of PCB to other
         organic materials which  may interfere analytically in the sample.

     d.   The apparatus  and the manipulation required  in the field are rela-
         tively simple  and straightforward, so that it has potential for
         adaption  to routine monitoring use.

     e.   The method should enable both a  qualitative  assessment of the species
         distribution of  the PCB  present  and a quantitative assessment of the
         amount of PCB  present.

     f.   Bidleman  and Olney (1974b) have  conducted a  preliminary evaluation  of
         the method and expressed confidence in its application to assess
         ambient airborne contamination.

     Several authors (Rhoades, et aJL, 1977; Rice, et al_., 1976; Lewis, et
     a1., 1977) have recently reported on experience  they have gained in the
     laboratory testing and application of polyurethane foam for the sampling
     of  PCB in ambient  air.  Margeson (1977) discusses the state of the art
     for ambient airborne PCB measurement.

2.2  DESCRIPTION OF THE SAMPLING  APPARATUS
     A sampling apparatus similar to that used by Bidleman and Olney (1974b)
     was designed  using readily available equipment.   A number of modifica-
     tions were incorporated into the system to improve performance as a
     result of laboratory and preliminary field evaluation.  The sampling
     apparatus, including all modifications, is illustrated in Figure 1.
     Figure 2 is an exploded view of the  sampler.  This apparatus was used for
     all  evaluations reported in  this document.

     The basic sampling system is a modification of a standard high-volume
     sampling apparatus.   These units are commonly used for airborne

-------
                                              Faceplate
                                                Stainless  Steel  Throat  Extensior
                                                        Polyurethane  Foam
                                                        Plug Location
                                               Throat  Extension

                                                    Wire
                                                    Retainer
                                                     Motor Unit
                                                   Adapter
                                       Exhaust Duct
                                       (3m minimum length)
Figure 1.   Assembled sampler and shelter with exploded view of the filter holder,

-------
                                           Throat Modification
Figure 2.   Exploded view of hi-vol  sampling apparatus  designed for use
           with a rotometer showing modified throat to accomodate
           polyurethane foam pluos.

                                    10

-------
participate sampling.  They consist of  a  stainless  steel  filter  holder
and throat assembly coupled to a high capacity centrifigal blower.   The
standard high-volume sampler used  is manufactured by General Metal  Works
of Cleves, Ohio.  It was modified  by removal  of  the 3 cm  long  by 8  cm
diameter cylindrical throat assembly between  the filter holder and  the
blower housing and replacement by  a throat  assembly of  similar design but
3U cm in length.  The extended throat assembly was  welded  to the filter
holder assembly.  This provided a  space for placement of  the polyurethane
collection medium.  The neoprene seal located on the backside  of the
faceplate was removed and all adhesive  material  was scrupulously removed
with organic solvents.  Teflon tape was then  used to provide a seal
between the faceplate and the filter holder.  These modifications
resulted in a high-volume sampling system which  contained  only stainless
steel and teflon components and surfaces  ahead of the collection medium
in the air stream.  This was considered important to eliminate the
potential for contamination of the sample by  organic materials that may
leach from the seating gasket normally  used.

Initially, a standard high-volume  sampler housing was used to  support the
apparatus.  This proved unsatisfactory  for  two reasons.   It was  large,
cumbersom, and difficult to transport.  The blower motor  operates under
greater load than the conventional high-volume sampling use, and hence,
tended to heat the enclosed sampler housing to an excessive degree.
Elevated temperatures were considered potentially disadvantageous to the
capability of the adsorbent medium to retain  the PCB.  A  simplified, open
construction sampler supporting stand was therefore designed.  The  basic
design of this supporting stand is shown  in Figure  1.  Four legs con-
structed of steel pipe are fitted  into  couplings welded to a rectangular
steel support ring.  This support  ring  is fashioned to  the dimensions of
the filter holder assembly.  The sampling apparatus is  suspended in  this
support assemoly.  An aluminum sheet cover  is positioned  over  the filter
assembly and supported by threaded bolts  welded  to  the  support ring.  The
cover is attached to these bolts with wing  nuts.  This cover protects the
glass fiber filter and the collection medium  from rainfall in  the same
manner a conventional high-volume  sampler housing is used.
                                   11

-------
High-volume samplers are normally operated from a conventional 60-cycle,
120-volt power supply.  With a glass fiber filter in place, a typical
sampler will draw air at approximately 1.7 m^/min.  The air flow rate
is reduced to approximately 1 m^/min with the polyurethane foam
inserted into the throat and the glass fiber filter in place.  Better
flow control is achieved by use of a motor speed controller.  A potentio-
metric controller or variac of suitable capacity should be used.  A flow
rate curve is generated for each individual sampler using the various
size orifice plates supplied by the manufacturer.  After this curve is
plotted, the flow may be determined at any time by measurement of the
exhaust pressure, in inches of water, using a manometer, rotometer, or
calibrated gauge.  The flow calibration is described in the Code of
Federal Regulations (1975).  For the purpose of sampling airborne PCB, a
flow rate of 0.8 m^/min was found to be suitable.

Some difficulty was experienced during initial sampler runs with the
movement of the polyurethane foam down the throat assembly and seating
against the inner housing assembly of the blower.  The polyurethane foam
could  not enter the blower, and hence, this was not an immediate problem.
However, it was considered a simple precaution to place a retainer in the
throat assembly to prohibit the polyurethane foam from moving.  This
simple retainer device is shown in the exploded view in Figure 1.  It
seats  on the inner blower housing and keeps the polyurethane  approxi-
mately 5 cm above the blower intake.  A piece of rigid hardware cloth
would  serve the same purpose.

Lewis  and Zimmerman (1976) have reported the existence of a capacitor
containing polychlorinated naphthalenes (PCN) in the motor assembly of
certain standard high-volume samplers.  When this unit is used for
sampling PCB, using polyurethane foam plugs, PCN is found on  the plugs.
This implies the high-volume samplers used recirculate at least a portion
of the sampled air.  Significant air recirculation will result in low
analytical results of the air  sampled.  In addition, PCN contamination
severely interferes with both  qualitative assessment and quantitative
                                   12

-------
analysis of PCB.  Recirculation must be prevented.  This is accomplished
by placing an exhaust duct three or more meters in length on the high-
volume motor housing, as shown in Figure 1.  Collapsible ducting is ideal
for this purpose.  In this way, the sampler exhaust can be located remote
from the sampler intake to prohibit any opportunity for recirculation.

The high-volume samplers used throughout the laboratory investigations
and field studies reported in this document do not contain the capacitor
discussed by Lewis and Zimmerman (1976).  Only samplers without the
PCN-containing capacitor should be used.

A 20 cm x 25 cm glass fiber filter (Gelman Type A) is placed in the
filter holder assembly for each sample run.  The purpose of this glass
fiber filter is not to separate particulate from gaseous PCB in ambient
air, since this is not considered possible with a compound of insuffi-
ciently low vapor pressure, but to keep airborne particulate from
embedding in the polyurethane foam.  The glass fiber filter is normally
analyzed along with the polyurethane after sample collection.  Experience
has shown, in most cases, that a significant amount of PCB is seldom
retained on the glass fiber filter.  Once this is established, it is not
necessary to include the glass fiber filter in the analysis.

The collection medium used is a commercially available polyurethane foam
material.  For purposes of this study, a 7.6 cm thick sheet of the
material of density O.U25 g/cc was obtained.  Cylindrical  plugs of 9 cm
diameter are cut from the foam sheet.  The plugs are cut oversize so that
they fit snugly into the sampler in such a manner that air cannot bypass
them.  Three polyurethane foam plugs, each 7.6 cm in thickness, were
used, for a total thickness of 23 cm, for most of the studies conducted.
The extensive polyurethane foam cleanup procedures that must be followed
before the material is used will be discussed in a later section of this
report.
                                   13

-------
  2.3  ANALYTICAL PROTOCOL
       The step-by-step procedure recommended for the analysis of ambient air-
       borne PCB is included as Appendix A to this report.   This procedure is
       drawn from the technical literature,  personal  contact with persons
       knowledgeable in the field,  and both  laboratory and  field experience.
       The purpose of this  chapter is to relate experience  gained during
       development of the method.

       Figure 3 is a flow diagram of the analytical  scheme.  The procedure
       consists of preparation of materials, field sampling, extraction, con-
       centration, cleanup, and gas chromatographic analysis.  To avoid the
       difficulty and uncertainty in accurately quantifying the PCB present in a
       sample, it is often  necessary to employ the perchlorination technique
       developed by Armour  (1973).   Considerable effort was applied to adopting
       this technique for routine use in quantitation of ambient airborne PCB.
       Some modifications of the perchlorination technique  were necessary to
       achieve quantitative derivization of ambient airborne PCB.

2.3.1  Cleanup of Materials and Apparatus
       It has been emphasized many times by numerous researchers that PCB are
       ubiquitous in the environment.  They are present in  the laboratory
       atmosphere, on glassware, and occasionally as contaminants in chemical
       reagents.  Strict attention must be given to every detail of sample
       handling in order to avoid PCB contamination.  It was found necessary to
       employ rather extensive cleanup measures to avoid PCB contamination from
       materials and apparatus.  All materials that come into physical contact
       with the samples are sequentially cleaned by detergent wash, tap water
       rinse, deionized water rinse, acetone rinse, petroleum ether rinse, and
       heating to remove any residual organics.  This applies to the sampling
       system, to glassware, aluminum foil, and any items used to handle the
       polyurethane foam.

       The polyurethane foam that is used as an adsorbent medium requires exten-
       sive cleanup by solvent extraction in a soxhlet extraction apparatus.
       Extraction is performed with nanograde hexane for a period of 12 hours.
                                          14

-------
                          CLEANUP OF POLYURETHANE
                     FOAM  PLUGS, GLASS  FIBER  FILTERS
                        AND  SAMPLING APPARATUS
                            Field Sampling
                         Soxhlet Extraction of
                        Polyurethane Foam Plugs
                        and Glass Fiber Filters
                             Silica Gel
                              Separation
                                                              Pesticides
                                                               Fraction
  GC/EC Analysis
  of PCB Mixture
         1
GC/FID Analysis
 of Biphenyl
Perchlorination
  PCB Quantitation
by Pattern Matching
                                   ±
                           GC/EC Analysis of
                           Decachlorobiphenyl
            Figure 3.   Analytical  scheme for PCB in ambient air.
                                      15

-------
Following this, they are dried with organic-free nitrogen in a heated
vacuum desiccator and sealed in cleaned glass containers until use.  This
is similar to the polyurethane foam cleanup procedure that is recommended
by Bidleman and Olney (1974b).  If it becomes necessary to prepare a
larger number of polyurethane foam plugs, it has been found convenient to
use a glass apparatus similar to that shown in Figure 4 for drying rather
than the vacuum desiccator, which has a limited volume.  It is important
to take every precaution to remove residual organics from the nitrogen
used for drying the plugs and for other steps in the analytical proce-
dure.  An organic trap of the type shown in Figure 4 has been found to be
effective if the components are replaced at routine intervals when they
become saturated with organic gases.  The polyurethane foam used in this
trap has been extracted with solvents as described above prior to use.

A typical polyurethane foan extract is depicted in Figure 5.  This is an
electron capture chromatogram of an extract that has been prepared as
described above.  It was selected randomly from plugs stored before use
in glass containers.  The unexposed plug can be carried through the
analytical procedure in a manner identical to a sample in order to demon-
strate contamination-free laboratory procedures.

It is essential that the extensive pre-cleaning procedures for the poly-
urethane foam plugs as described in Appendix A be followed.  All of the
foam materials tested contained PCB and other electron capture sensitive
contaminants that must be removed before use of the material as a collec-
tion medium.  Shorter than 12-hour extraction periods did not adequately
reduce contaminant levels.  A 5-hour extraction of plugs of the proper
dimensions cut from the stock foam material described in Appendix A
yielded a mean blank value of 113 ng as DCB with a range of 42 to 200 ng
and a standard deviation of 58 ng for five samples.  Twelve-hour extrac-
tion of the same five plugs yielded blank values as follows:
                       Mean DCB residual     54 ng
                       Range of values    30-78 ng
                       Standard deviation     9 ng
                                   16

-------
       Both the mean residual DCB and the range of values from one plug to
       another is reduced to manageable levels by 12-hour extraction.  The
       residual DCB level (blank value) for each lot of polyurethane foam plugs
       that are prepared for sample collection must be determined before they
       are used in order to validate the adequacy of the pre-cleaning procedure.

       Polyurethane foam plugs may be reused after sample extraction for
       collection of further samples by simply drying and returning to glass
       containers for storage.  It is not necessary to repeat the 12-hour
       soxhlet extraction if the plug is not exposed to contamination.

       The glass fiber filter (Gelman Type A) is pre-conditioned by heating for
       several hours to remove any organics present.  These filters should not
       be stored in cardboard boxes or manila envelopes, since these materials
       and other paper products sometimes have very high levels of PCB con-
       tamination.  A typical chromatogram of a filter blank is reproduced in
       Figure 6.

2.3.2  Field Sampling
       Field sampling procedures are sufficiently straightforward that they can
       be performed without undue burden.  It is necessary, however, that the
       individual responsible for sampling be conscientiously aware of the
       potential for PCB contamination and take routine precautions to avoid
       it.

       Recirculation of exhaust from the sampling apparatus to the intake has
       been discussed.  This is a problem that must be avoided by placement of
       the end of the exhaust duct so that recirculation is unlikely to occur.
       An exhaust duct 3 m in length has been found to be effective; however, if
       there is any question concerning the possibility of recirculation in a
       given situation, a longer exhaust duct is recommended.  Collapsible
       ducting of the type used to vent household clothes dryers is ideal  for
       this purpose.

       The placement of pre-cleaned polyurethane foam plugs in the sampling
       apparatus is done at the sampling location.  Only pre-cleaned metal tongs
                                          17

-------
                                                                        . Hood
                       Nitrogen
                Polyurethane
                    Foan
                   Activated
                   Charcoal
                     Polyurethane
                    — Foam
                    Drying Agent
                    —(Dri-Rite)
                      Glass Wool
                                         Variable
                                         Voltage
                                         Control
                                          Heating
                                           Tape
                                     Glass Tubing
                                     120cm x 10 cm
                                    Polyurethane
                                    Foam Plug
Figure 4.   Apparatus for drying polyurethane foam plugs,
                               18

-------
                                                                          o


                                                                          X
                                                                          c:
                                                                          o
                                                                          ro
                                                                          3
                                                                          C
                                                                          O)
                                                                          •M
                                                                          +J
                                                                          03
                                                                          O)
                                                                         •o
                                                                          S-
                                                                          o
                                                                          (J
                                                                          O)
                                                                          S-

                                                                         -M
                                                                          "3
                                                                          O
                                                                          CO
                                                                          S-
                                                                          4->
                                                                          X
                                                                          O)
                                                                          O)
                                                                          c
                                                                          ro
                                                                          CD
                                                                          o
                                                                          CL
                                                                          ro
                                                                          O
                                                                          O.
                                                                          E
                                                                          ro
                                                                          i_
                                                                          CT-
                                                                          o
                                                                         •(->
                                                                          ro

                                                                          O
                                                                          S-
                                                                         jc:
                                                                         o
                                                                          01
                                                                          s_
19

-------
                                                                        CTi
                                                                          I
                                                                          o
                                                                           X

                                                                          •vt-
                                                                           c
                                                                           o
                                                                           rO
                                                                           3
                                                                           C
                                                                           
                                                                           fC

                                                                           s_
                                                                           QJ
                                                                          TD
                                                                           S-
                                                                           O
                                                                           o
                                                                           OJ
                                                                           s-
                                                                           (J
                                                                           «
                                                                           s-
                                                                          -p
                                                                           X
                                                                           O)
                                                                           O)
                                                                           Q.
                                                                           O)
                                                                           OJ
                                                                           s_
                                                                           QJ
                                                                          JD
                                                                           to
                                                                           to
                                                                           Ol
                                                                           S-
                                                                           cn
                                                                           o
                                                                           O
                                                                          CJ
                                                                           cu
20

-------
       are used to handle the plugs.  When plugs are placed in the apparatus,
       the individual  should assure they fit snugly to avoid any possibility of
       leaks around them.  There should be a minimum of exposure of the cleaned
       plugs to the laboratory or ambient atmosphere.  Immediately after
       sampling is complete, the polyurethane foam plugs are removed from the
       sampling apparatus and returned to the appropriately-labeled storage
       container for transport back to the laboratory.  The glass fiber filter
       is handled in the same manner.  When storage is necessary, it is prefer-
       able to keep the exposed plugs in the dark and under refrigeration in
       order to avoid any decomposition that may possibly take place.

2.3.3  Sample Extraction
       Exposed polyurethane foam plugs are transferred to a soxhlet extraction
       apparatus and extracted for 3 hours with nanograde hexane.  The glass
       fiber filter is extracted in the same manner as the plug.  All plugs may
       be extracted together; they may be extracted separately and combined; or
       the glass fiber filter and the plugs may be analyzed separately.  The
       polyurethane foam plugs can be reused if they are immediately dried in
       the proper manner and returned to storage in a sealed glass container.
       Even high concentrations of PCB (e.g., 128 ug as DCB) spiked onto a plug
       were quantitatively removed during a 3-hour extraction period.

       Contamination of the sample was sometimes noted when the paper thimbles
       normally used in a soxhlet extraction apparatus were employed.  It was
       found that the extractors performed satisfactorily without these thimbles
       and they were subsequently eliminated.  Another point of caution is worth
       mentioning.  An interference problem was encountered in the use of
       boiling chips ("Boileezers" marketed by Fisher Scientific Co.) in the
       soxhlet extraction step.  Broken glass which has been heated to remove
       organics serves as a most satisfactory boiling aid without causing
       interference problems.

2.3.4  Sample Concentration
       Larger volumes of sample extract are concentrated satisfactorily using a
       Kuderna-Danish evaporator.  When it is necessary to reduce extract volume
                                          21

-------
       to less  than 5  ml,  a  modified micro-Snyder apparatus  is  used.   Extreme
       caution  must be exercised to avoid evaporation  of any sample  to dryness.
       In the early phases of evaluation  of the laboratory procedures, excellent
       recovery efficiencies (average 104 percent)  were achieved with Aroclor
       1260.  Since airborne PCB is likely composed of more  volatile  species
       than those present in Aroclor 1260, analytical  recovery  studies were con-
       ducted using Aroclor  1016.   This mixture contains principally  di-,  tri-,
       and tetrachlorobiphenyls.  Recovery efficiencies were found to be erratic
       and poor.  Step-by-step evaluation of the analytical  technique isolated
       the problem to  two steps in the procedure where the sample extract was
       evaporated to "near dryness."  During these steps, it was found that a
       significant amount of the dichlorobiphenyl present in Aroclor  1016 was
       lost by  volatilization resulting in modification of the  chromatographic
       pattern.  This, of course,  would introduce serious error in the quantita-
       tion of  airborne PCB.  The analytical procedure was,  therefore, modified
       to eliminate procedural steps resulting in the  evaporation of sample
       extracts to dryness or "near dryness."  Acceptable analytical  recovery
       efficiencies (average of 97 percent) were subsequently achieved with
       Aroclor  1016 using the modified procedures.

2.3.5  Sample Cleanup
       The silicic acid cleanup procedure developed by Snyder and Reinert (1971)
       is employed to  remove chlorinated  pesticides and other interfering sub-
       stances  from the sample.  This technique is not effective in eliminating
       DDE or toxaphene.

       The sulfuric acid cleanup procedure of Murphy (1972)  was found to remove
       a significant number  of contaminants from certain of  the sample extracts;
       however, it was not found to be necessary to use a sulfuric acid cleanup
       step routinely.

2.3.6  GAS CHROMATOGRAPHIC ANALYSIS
       Analysis was conducted on a Varian 2760 gas chromatograph using a tritide
       foil electron capture detector.  A 1.5%/1.95% OV-17/QF-1 packed glass
       column was found to be suitable.  Confirmation was performed on a 4X/6X
                                          22

-------
       SE-30/OV-210 column.   Operating conditions for both these  columns  are
       described in the appendix.   Several  analyses were also confirmed by  gas
       chromatography/mass  spectrometry (GC/MS)  using a AEI 30 (Associated
       Electronic Industries)  instrument.   The OV-17/QF-1 column  was  used with
       this instrument under the same operating  conditions as described in  the
       appendix, with the exception that helium  was used as the carrier gas in
       place of nitrogen.

       The importance of confirmation of peaks cannot be overstressed.   This
       can be demonstrated  by the  following two  figures.  The sample  is one of
       the initial  field samples of urban  atmosphere.  The chromatogram is
       complicated  by numerous peaks which could not be identified as PCB.
       Figures 7(a), (b), and (c)  are chromatograms of the first, second, and
       third polyurethane foam plugs, respectively, as they are located in  the
       sampling apparatus.   Figure 7(d) is a chromatogram of Aroclor  1221 on  the
       same column.  Those  peaks indicated by a  'V' match corresponding peaks in
       Aroclor 1221 and/or  Aroclor 1016.  The first major apparent PCB peak,
       eluting at 1.33 minutes, appears to be present in roughly  equivalent
       amounts on all  three  plugs  indicating a poor collection efficiency.   This
       did not seem consistent with other  data,  so the sample was run on  the
       confirmatory GC column.  Figure 8(a) is a chromatogram of  the  first  plug
       on the confirmatory  column.  Figure 8(b)  is a chromatogram of  Aroclor
       1221 on the  same column.  2-Monochlorobiphenyl (2-MCB), 4-monochlorobi-
       phenyl (4-MCB), 2,2'-dichlorobiphenyl  (2,2'-DCB) and 2,4-dichlorobiphenyl
       (2,4-DCB) have been  identified by comparison with these pure PCB isomers.

       The peak which appeared to  match 2-MCB on the OV-17/QF-1 column, was not
       confirmed on the SE-30/OV-210 column.   Its position falls  between  the
       peaks eluting at 1.18 minutes and 1.49 minutes.  2-MCB was not present at
       all.  This was later confirmed by GC/MS analysis.

2.3.7  Perch!orination
       Quantitation of PCB in environmental samples is sometimes  an extremely
       difficult and, unfortunately, often an inexact procedure.   There exists
       little difficulty in those  cases where the GC elution pattern  closely
                                          23

-------
                                                                 o
                                                                 o
                                                                 I
                                                                Lu
                                                                Cf

                                                                r^
                                                                r—
                                                                 I
                                                                >
                                                                O

                                                                 c
                                                                 o

                                                                 o>
                                                                 to
                                                                 o
                                                                 cu
                                                                 c
                                                                 -p
                                                                 QJ
                                                                 S-
                                                                 o
                                                                 a.
                                                                 OJ

                                                                "o.

                                                                 rO
                                                                 to

                                                                 s_
                                                                •I—
                                                                 to
                                                                 (O
                                                                 .a
                                                                 s-
                                                                 ra
                                                                 14-
                                                                 o
                                                                 s-
                                                                 C71
                                                                 o
                                                                 o
                                                                 S-
                                                                 CJ
                                                                 (U
                                                                 s-
24

-------
                                                                  c
                                                                  E
                                                                  o
                                                                  o
                                                                  I
                                                                 Ll_
                                                                 O
                                                                  i
                                                                 >
                                                                 o
                                                                  c.
                                                                  o
                                                                  E
                                                                  to
                                                                  o
                                                                  O)
                                                                  c
                                                                  OJ
                                                                  S-
                                                                  o
                                                                  Q.
                                                                  O
                                                                  O
                                                                  S-
                                                                  cn
                                                                  O
                                                                 -I-J
                                                                  O
                                                                  S-
                                                                 _a

                                                                 PC*
                                                                  O)
                                                                  S-
                                                                  CD
                                                          O
25

-------
                                                                  c
                                                                  E
                                                                  3
                                                                  O
                                                                  u
                                                                  c
                                                                  o
                                                                   CD
                                                                   E
                                                                   (O
                                                                   o
                                                                   O)
                                                                   c
                                                                   fO
                                                                   OJ
                                                                   S-
                                                                   3
                                                                   o
                                                                   ex
                                                                   ro
                                                                   s-
                                                                   cn
                                                                   O
                                                                  +->
                                                                   (Q

                                                                   O
                                                                   s_
                                                                  J=
                                                                  CJ
                                                                   O)
                                                                   $-
                                                                   en
26

-------
                                                                  c:
                                                                  o
                                                                  o
                                                                  I
                                                                 u_
                                                                 o-
                                                                  i
                                                                 >
                                                                 o
                                                                 C\J
                                                                 00
                                                                  s_
                                                                  o

                                                                  o
                                                                  o
                                                                  s_
                                                                 "O

                                                                 r^
                                                                  OJ
                                                                  s_
                                                                  en
27

-------
                                                            LO
                                                           ,LO
                                                            O
                                                                     o
                                                                     (J
                                                                    CM
                                                                     I
                                                                    o
                                                                    o
                                                                     c
                                                                     o
                                                                     en
                                                                     E
                                                                     «3
                                                                     O
                                                                     O)
                                                                     E
                                                                     to
                                                                     o>
                                                                     o
                                                                     Q.

                                                                    4-1
                                                                     I/)
                                                                     S-
                                                                     
                                                                     ro

                                                                     O
                                                                     ro

                                                                    co'
                                                                     QJ
                                                                     S-
                                                                     en

                                                                    LU
28

-------
       ca
       o
 i
CM
              DO
              C_>
              Q
               I

              <3-
               f\

              CM
            CD
            O


             I


            CM
             r\

            CM
  0
                                            10
                                                                 MIN
Figure 8(b).   Chromatogram of Aroclor 1221 on SE-30/OV-210 column.


                                     29

-------
resembles that of an Aroclor, but this is often not the case.  With
ambient airborne PCB in particular, there is little reason to expect a
close correspondence with any single Aroclor unless the sample is taken
very near the source of the pollution.  The more volatile PCB isomers
will likely be enriched in the atmosphere and airborne transport will not
likely be uniform for isomers of different volatility.  Each of these
factors will alter the nature of the collected sample and distort the GC
elution pattern, making quantitation a difficult task.  When peak-to-peak
area ratios in the sample are different from those in the standard used
for purposes of quantitation, results may be reported which are
significantly in error.

Another approach to the task of quantitation of PCB in ambient air is by
comparison of individual PCB isomers found to be present in the sample
with known concentrations of that isomer.  Quantitation in this manner is
possible for the monochlorobiphenyl and dichlorobiphenyl species, since
there exists a manageable number of isomeric variations of these species
in commercial PCB mixtures.  Three monochlorobiphenyl isomers and four
dichlorobiphenyl isomers have been detected in Aroclors (Hutzinger, Safe,
and Zitko, 1974).  The task is rapidly complicated, however, when tri-
chlorobiphenyl isomers are present.  There have been 17 trichlorobiphenyl
isomers detected in commercial PCB mixtures.  Perhaps by use of capillary
column GC, quantitation in this manner would be feasible for more highly
substituted isomers.  However, it would certainly entail a long, arduous
process when greater than trichloro isomers are present.

The most logical approach to accurate quantitation of airborne PCB is by
use of the perchlorination technique recommended by Armour (1973).  This
technique permits derivitization of all PCB present to decachlorobiphenyl
(DCB).  When this is done, the task of quantitation is greatly simplified
and the detection limit is improved, since DCB is a single isomeric
species.  Confirmation of the PCB is also accomplished by this technique.
Armour (1973) achieved essentially quantitative conversion of Aroclors
1016, 1242, 1248, 1254, 1260, and 1262 to DCB using this technique at
microgram levels.   If all PCB present in an ambient air sample could be
                                   30

-------
chemically derivatized to DCB in this manner, quantisation would be
accurate, precise, and reproducible from one analyst to the next with
little variability due to subjective GC pattern interpretation.  The
nature of the PCB present can be qualitatively assessed before
perchlorination.

Investigations were conducted into the use of the perchlorination tech-
nique for the analysis of ambient air samples for PCB collected on
polyurethane foam.  The method of Armour (1973) was initially used.  The
utility of this procedure is demonstrated by Figures 7(a) and 9.
Figure 9 is a chromatogram of the same air sample shown in Figure 7(a)
after perchlorination.  Quantitation of PCB is obviously greatly simpli-
fied.  The PCB is clearly separated in elution time from other materials
which are not removed by the extensive cleanup procedures applied to the
sample.  The analytical detection limit is also greatly enhanced by per-
chlorination since the electron capture detector is much more responsive
to DCB than lower-substituted isomers.  The first polyurethane foam plug,
shown in Figures 7(a) and 9, contained 1,790 ng of PCB as DCB; the
second, 920 ng.  The concentration on the third plug was equivalent to
blank values observed for unexposed plugs.

Two problems in application of the Armour method had to be overcome,
however, to achieve satisfactory results for airborne PCB.  First, the
method did not yield quantitative results when attempts were made to
perchlorinate PCB species of less than trichloro substitution.  Secondly,
even for Aroclors, quantitative recovery was not achieved for less than
1 ug quantities of PCB.  Both of these problems, it was discovered, were
the result of the volatile loss of PCB during one step in the procedure
where the extract is evaporated to near dryness.  Furthermore, the recom-
mended reaction temperature appeared to be sufficiently high to result in
decomposition of the lower-substituted PCB species.  After rather exten-
sive experimentation and several modifications to the original Armour
method, quantitative recovery was achieved for even the most volatile PCB
species in sub-microgram quantities.
                                   31

-------
        (a)   Decachlorobiphenyl  Standard
                         (b)  Air Sample
                     i  i  i  i  i  i   i  i • i1 -i  i  i
                                                          Minutes
Figure 9.   Chromatogram of a Perchlorinated Air Sample.
                                  32

-------
The sample is extracted from the polyurethane foam with hexane, concen-
trated, and cleaned up using silica gel.  Since the perchlorination
reaction cannot be carried out in hexane, all residual hexane must be
removed from the extract prior to perchlorination.  Even small amounts
of residual hexane will result in the formation of a black, solid resi-
due upon the addition of the antimony pentachloride (SbC^).  This
severely reduces PCB recovery.  Attempts to  remove the hexane by direct
evaporation, even when very carefully conducted, invariably resulted in
the volatile loss of dichlorobiphenyl and monochlorobiphenyl from test
sofutions.  The hexane is, therefore, removed by azeotropic evaporation
from a hexane/chloroform mixture.  To the 1.0 ml extract in a Kuderna-
Danish (K-D) apparatus, is added 10 ml of chloroform.  This is concen-
trated by slow boiling to 4 ml.  Repeat this azeotropic evaporation three
additional times by adding 10 ml increments  of fresh chloroform in order
to remove all traces of hexane.

After removal of hexane, the chloroform extract is quantitatively trans-
ferred to a reaction vial, prepared as shown in Figure 10, using three
chloroform rinses.  The reaction vial is a 16 x 100 mm glass screw top  .
culture tube with teflon-lined screw cap which has been modified by
drawing the bottom into a cone shape.  Reaction vials prepared in this
manner were found to be easier to work with  and much more reasonable in
cost than the vacuum hydrolysis tubes recommended by Armour.  The tapered
bottom of the vial allows the analyst to reduce solvent volume to 0.1 ml
without going to dryness.  Add two boiling chips and immerse the reaction
vial upright in a 70°C water bath to a depth of 6 + 2 cm.  Increase the
water bath temperature slowly until the solvent begins to boil at
72-76°C.  Concentrate slowly to a volume of  approximately 0.1 ml.  Under
no circumstances should the water bath temperature be permitted to exceed
76°C nor should the solvent be evaporated to dryness.  If either of these
occurs, PCB will be lost by volatilization and consequent recoveries will
be low.  The final volume (0.1 ml) may be determined with sufficient
accuracy by comparison of the solvent remaining with another reaction
vial containing 0.1 ml of chloroform.  Cap the reaction vial immediately
after concentration and allow to cool.
                                   33

-------
                     Teflon-lined  cap
                mr
Figure 10.   Perchlorination reaction vial
                         34

-------
Reagent grade SbCl^ from several commercial sources has been found
to contain DCB and bromononachlorobiphenyl  (BNCB), as reported by
Trotter and Young (1975).  The DCB contaminant results in a positive
blank value, and the BNCB interferes with the perch!orination of the
lower-substituted PCB species.  The SbCl5 selected for use must first
be analyzed for these contaminants.  SbCl^ manufactured by Cerac Pure
(Stock Number A1147, Lot Number 15-26A) was found to be suitable.  DCB
contamination of this reagent is approximately 10 ng/ml.  BNCB was not
detected.

The reaction vial containing the sample is  placed into a preheated
(16U + 3°C) aluminum block heater for a period of 15 hours.  It was found
that the use of higher reaction temperatures (185°C) as recommended by
Armour resulted in poor recovery of di- and mono-substituted PCB species.
Apparently these species were not stable at the higher temperatures.
Reaction temperatures less than 150°C also  resulted in poor recovery due
to incomplete conversion of lower-substituted PCB species to DCB.
Shorter reaction periods may be adequate to affect quantitative deriva-
tization; however, the 15-hour period is convenient for overnight,
unattended use.

After the reaction period, the reaction vial is removed from the heater
and allowed to cool  to room temperature.  Then it is cooled in an ice
bath before the addition of 1 ml of 6 ^ HC1 to inactivate the excess
perchlorination reagent.  Pressure from the reaction vial is cautiously
vented in a fume hood, directed away from the analyst, to avoid any
possibility of injury.

To extract the DCB from the reaction mixture, add 1 ml of hexane to the
reaction vial, shake vigorously for 3D seconds, and carefully draw off
the hexane layer with a disposable pi pet.   Place this hexane extract on
the top of a 6 mm x 12 cm disposable pi pet  packed with 2 g of anhydrous
         This column is prewashed with hexane.  Repeat hexane extraction
                                   35

-------
of the reactants in this manner five times.  Tests have shown this repe-
tition assures quantitative recovery of DCB.  After passing all  six
increments of the hexane extract through the ^$04 column, pass two
additional 1-ml portions of fresh hexane through the column.  Collect
all fractions in a 10-ml graduated K-D apparatus.  Connect a modified
micro-Snyder column to the K-D apparatus, add a single boiling chip,
and evaporate in a water bath (70°C) to less than 0.5 ml.  Add fresh
hexane to oring the volume back to exactly 1.0 ml on the graduated K-D
receiver.  Cap the receiver tightly or transfer the contents to a sealed
container to avoid evaporation of solvent prior to gas chromatographic
analysis.  Store in a cold, dark location to avoid any possibility of
photodegradation.

It is recognized that biphenyl, if present in a sample extract, would
perchlorinate to OCB.  Therefore, attempts were made to remove biphenyl
from the extract before the addition of the perchlorinating reagent.  Two
approaches were taken without success.  First, it was attempted to
selectively remove biphenyl from the reaction mixture by either hydroly-
sis or nitrification.  ^SO^ and H2S04/FS03H treatment of a mixture of
biphenyl and PCB resulted  in no effect.  The use of concentrated FS03H
and 502/^504 each resulted in decomposition of both biphenyl and PCB.
HN03/H2$04 caused nitration of PCB and biphenyl.  The use of an activated
alumina chromatographic column to selectively isolate biphenyl from a PCB
mixture was also evaluated.  This method showed some promise; however, a
considerable amount of research would have been required to find the most
suitable conditions (e.g., column dimensions, eluting solvents, activity
grade, etc.) for effective isolation of biphenyl.  Therefore, it was
decided not to pursue this technique further, but to develop GC condi-
tions for the analysis of  biphenyl.

The GC conditions using a  flame ionization detector that are described in
Appendix A (Section 12.u)  were found to provide an adequate detection
limit for biphenyl in a sample extract.  Hence, if the analyst suspects
                                   36

-------
       biphenyl  may be present, the GC/FID analysis is conducted to determine
       the amount present in the sample extract.  The efficiency of conversion
       of biphenyl  to DCB using the method described in Appendix A is 43.3 +
       8.8 percent.  This conversion efficiency is based on ten replicate analy-
       ses at 3.4 ug and U.7 ug biphenyl.  The analyst may correct the results
       obtained upon perch!orination by subtracting that amount resulting from
       the conversion of biphenyl  present in the extract to DCB.  Biphenyl was
       detected during preliminary field tests near a transformer storage
       facility.  Field tests at other locations did not yield detectable
       levels of biphenyl.

       After the modifications  described above were incorporated into the per-
       ch! ori nation method  to eliminate the problems encountered with volatile
       loss of PCB and thermal  decomposition of the lower-substituted PCB
       species,  a series of method recovery studies were conducted.  The modi-
       fied perchlorination method, as described in detail in Appendix A, was
       tested using Aroclor 1016,  Aroclor 1254, and a test mixture of the most
       volatile and least thermally stable PCB species.  The test mixture con-
       sisted of 25 percent monochlorobiphenyl  (2-MCB), 55 percent dichlorobi-
       phenyl (2,2'-DiCB,  2,4-DiCB, 2,4'-DiCB), and 20 percent trichlorobiphenyl
       (2,2',5-TCB).  The  results  of these recovery studies are summarized in
       Table 1.   Quantitative recovery was achieved for even the most volatile
       species over the range of at least 100 ng to 10 ug concentration.  The
       overall mean recovery for all tests was 100+5 percent.  The difficulties
       with poor recovery  of the lower-substituted species at sub-microgram
       concentrations have,  therefore, been overcome by the modifications to the
       original  procedure.

2.3.8  Perchlorination Ruggedness  Test
       A ruggedness test (Youden and Steiner, 1975) was performed on the final
       modified perchlorination method to assess the likelihood of achieving
       comparable inter!aboratory  results for PCB at low concentrations.  The
                                          37

-------
                                 TABLE 1
        PERCHLORINATION RECOVERY VS. CONCENTRATION TEST RESULTS


PCB
Aroclor 1016
Aroclor 1016
Aroclor 1254
Test Mixture*
Aroclor 1016
Test Mixture*
Test Mixture*
PCB Concen-
tration
(ug)
10.0
5.0
3.0
1.18
0.750
0.412
0.103

Number of
Replicates
3
3
1
3
4
2
3

Percent
Recovery
103
105
113
98
93
101
101

Standard
Deviation
3.7
2.1
—
6.5
5.6
1.1
4.7
   Consisting of 25% monochlorobiphenyl, 55% dichlorobiphenyl, and 20%
   trichlorobiphenyl.
statistical theory underlying the ruggedness testing methodology is dis-
cussed by Youden and Steiner (1975).  Seven parameters are chosen for
testing and their nominal values (the value specified in the procedure)
are denoted by A, B, C, D, E, F, and G.

A challenging value for each of those parameters is then selected and
denoted by a, b, c, d, e, f, and g.  The challenging parameters should
represent differences that might be introduced as a result of different
analysts at different laboratories following the same procedure.  A
series of eight experiments are then conducted using the combinations of
conditions specified in Table 2.  Each of the eight experiments, which
may be performed in replicate, produces a result.  The experimental
results are denoted as S, T, U, V, W, X, Y, and Z.  The results can be
grouped as follows, for example, to determine the difference attributable
to variable A:
                  A   - S + T + U + V   U + X + Y + Z
                  A—a -       
-------
                                TABLE 2

          EIGHT COMBINATIONS OF SEVEN FACTORS USED TO TEST THE
                   RUGGEDNESS OF AN ANALYTICAL METHOD
Factor
Value
A or a
B or b
C or c
D or d
E or e
F or f
G or g
Observed
Result
Determination Number
1
A
B
C
D
E
F
G

S
2
A
B
c
D
e
f
g

T
3
A
b
C
d
E
f
g

U
4
A
b
c
d
e
F
G

V
5
a
B
C
d
E
F
g

W
6
a
B
c
d
e
f
G

X
7
a
b
C
D
E
f
G

Y
8
a
b
c
D
e
F
g

z
After Youden and Steiner  (1975),
                                     39

-------
The ruggedness test of the perchlon'nation method incorporated the seven
factors listed in Table 3.  These factors varied from nominal (procedure
specified) conditions to the degree that might be expected of a competent
analyst following the written procedure.  The ruggedness test format and
data are shown in Table 4.  Each determination number consisted of
triplicate analysis of a PCB test mixture containing 25 percent monochlo-
robiphenyl, 55 percent dichlorobiphenyl (three isomers) and 20 percent
trichlorobiphenyl at a total concentration of 0.824 ug PCB.  The test
involved 24 separate analyses and, hence, has statistical reliability.

The ruggedness test results are tabulated in Table 5. All factors showed
a difference of less than that of the blank.  Hence, the challenging
conditions that were applied in no way affected the outcome of the
determination.  Furthermore, the difference observed for the blank (no
variation) determination was +3.25 percent, which is satisfactorily low.
An estimate may be made of the analytical standard deviation as follows:
            standard deviation = /2/7 Di 2

where Dj is the difference (Table 5) measured for each factor.  When
the data from Table 5 are entered into this formula, an estimated analy-
tical standard deviation of 3.0 percent results.  This is an acceptable
level of variance for gas chromatographic analytical procedures.
                                   40

-------
                                TABLE 3




               PERCHLORINATION RUGGEDNESS TEST CONDITIONS
Factor
Extract transfer
rinse volume
Reaction
temperature
SbCl5 reagent
volume
Final extract
vol time
Blank
Solvent evaporation
temperature
Water bath
immersion depth
Procedure
Step No.*
8.2.3
11.4.2
11.4.1
11.3.6
—
11.3.5
11.3.4
Nominal
A = 4 ml
B = 160°C
C = 0.2 ml
D = 0.1 ml
E = e
F = 72°C
G = 4 cm
Challenging
a = 1 ml
b = 165°C
c = 0.1 ml
d = 0.3 ml
e = E
f = 76°C
g = 9 cm
*As described in Appendix A.
                                     41

-------












^£
1—
*t
o

z
*i

(—
e^
2:
cc
o
u_

h-
oo
UJ
oo
LU 00
_J Ul
03 Z
< Q
1— UJ
UJ

DC

z
o
1— 1
L_
I
3Z
\ — I
ca
o
1
a:
o
UJ
Q.






























S.
jQ
E
3
Z

C
O
'£>
*o
c
•r"
E
S_
cu
•p
O)
Q













00



>^


*-O





LO





*^






oo




CM





9—t




















CU
3
^v
»

^.
o
•p
o
eO
U,
o LO «— i >— t i CM cn
• to • • i r>»
.-t .-I O O

O LO CM T— 1 1 to *d"
• to • • i r**-
i— ( i— i O O
O O •— 1 CO 1 to 
^- r-l O O


O O <— « i— 1 1 to CT)
• to • • i t--
•d- r-l O O

r- r- i—
E 0 1 1
o 0 E
O O CM <-l 1 0 U
• to • • 1 CM
^ «— 1 O O r~ •*
CU
S-

>E 0 E E 4JQ. Q.
0 (0 E 0 CU E
CUO LT> >— 1 CO -r-CUO QO
«/>• to • • 4-1 — to
C *~1 CU «-H CU C3 CU CO fO r^^ £T CT)
•i- s- E E > c o
a; n 3 n 3 n 3 n o it -r- it
•P i— r— O -i- tO
^K *o rt5 ^^ ^5 ^^ ^5 ~o c ^^ ^^~ ti- en
CL) S«- ^* ^^ fO ^^
4— Q> t- £
in O. -P 4J O E
Cr-E Cr-Oi— CL ^

t- 1 — O C7) j^_ ^ f ^ c~ ^
1— O O ^ o '^ C£. O UJ O CU -P r^* CQ ^*
U -r- C
foil -p it LO n i— n jx: H cu n s- n
s- u i—  cu
•p< iaca oo CQ coui r— u. -PCU
X Q) J3 -i- r— O 






"—^




t—






oo

















^>
^_
3
to
cu
f^f

T3
cu
>
^
cu
to
^^
o
i-l LO
c %
O-i CO
Oi
CTi to
0 9
r-l O
LO ^
cr> «^-



to

^f 1
CT^


CT^ CM

oo «*
en



CO CM

LO CO


r-* ,-<
m •
O*> "^
CT>



J*-N.. Qfj
* •
CO CO

^^
^

o^
^-*-' C
0
>> -I-
1? +J
cu to
^> *f~"
o >
o cu
CU Q
a:
-o
4-> S-
C ro
cu -a
o c
U rO
CU P
a. oo

^3 "C3
cu cu
> >
S. S-
cu cu
I/) tO
-Q jQ
0 0
42

-------
                                 TABLE 5
                PERCHLORINATIOM RUGGEDNESS TEST RESULTS
                                                            Difference
Rank                  Factor
1
2
3
4
5
6
7
Blank
Reaction Temperature
SbCl5 Volume
Final Extract Volume
Water Bath Immersion Depth
Extract Transfer Method
Solvent Evaporation Temperature
+3.25
+2.95
-2.80
1.40
-1.30
0.75
0.15
The results of ruggedness testing of the modified perchlorination method
indicate the method is "rugged" and therefore immune to modest depar-
tures from habitual routine on the part of the analyst.  Furthermore,
all method variables are in control.  Variation within the specified
limits of the procedure should not affect the outcome of the analysis.
Competent trace analytical  chemists following the above procedure in
different laboratories should therefore achieve comparable analytical
results.
                                   43

-------
3.0  RECOVERY STUDIES AND FIELD TESTS
     After development of suitable laboratory analytical  techniques and dem-
     onstration that these techniques permitted quantitative recovery of PCB
     ranging from mixtures of low substitution (e.g., Aroclor 1221) to mix-
     tures of a high degree of substitution (e.g., Aroclor 1260), and after
     improvement of the perchlorination method, efforts were turned to an
     evaluation of the selected sampling method for airborne PCB.  This evalu-
     ation was conducted in such a manner that it was, in fact, an evaluation
     of the entire method, including both sample collection and laboratory
     analysis.  This chapter describes the results of collection efficiency
     and recovery studies of the method, preliminary field tests, and field
     tests of the final sampling and analytical method.  The collection
     efficiency and recovery studies were conducted to identify the capabili-
     ties and limitations of the high-volume sampling method selected.  The
     preliminary field tests described below were conducted over the course of
     development of the method.  They provided valuable information pertinent
     to development of the method for routine use.  Finally, after method
     development was complete and the method tested in the laboratory, it was
     subjected to several thoroughgoing field tests.

3.1  RECOVERY STUDIES
     PCB recovery studies were first conducted with Aroclor 1221 and, more
     extensively, with Aroclor 1016.  These two commercial mixtures were
     selected for the initial studies for two reasons.  First, they are
     composed principally of the lower-substituted PCB species which are
     generally the more volatile ones.  Secondly, Aroclor 1221 and Aroclor
     1016 are commonly used commercial products.

     It would be anticipated, and has been shown by Mieure (1975), that the
     more volatile PCB species are likely to be found in the atmosphere in
     greater amounts than the less volatile species.  The molecular composi-
     tion of several Aroclors is tabulated in Table 6.  Aroclor 1221 is com-
     posed predominantly of biphenyl, monochlorobiphenyl, and dichlorobiphenyl.
                                        44

-------






















i^O

UJ
_J
03
1
r~~




































^r
§
o
1—
Q
Z

"
u.
c^
oo

*»
fv*
UJ
z
M
1—
^2
~T"

UJ
1—
Li-
's:
"""'
00
o
1
—I
0
0

^c

UJ
o
OO

u_
0
•z.
o
t— 4
1 —
oo
o
Q.
s:
o
o

g£
_i
>
o
UJ
O
22










s*
3
3"
'—
S-
o

u
o
S-
<
c
'""
(U
o
c
^
c
01
a.
•f~
-Q
O
S_
O
r—
0
^-( t— 1 LT>
• • • » — i < — i oo f^o ^0 o
o o o oj 
•o
O)
c
o
c
M
^§
I— I
o

o
V

cu
•p
(O
o
•o
c~
.^

Q
z

-X
45

-------
Aroclor 1016 is largely trichlorobiphenyl  with significant amounts of
dichlorobiphenyl and tetrachlorobiphenyl.   Approximately 48 percent of
the total domestic production of PCB in 1975 consisted of Aroclor 1016.
Another 25 percent consisted of the closely related product Aroclor 1242
(Monsanto Co., 1975).  Aroclor 1221 production quantities were much
lower, comprising less than 1 percent of total annual  production in 1975;
however, Aroclor 1221 was produced in considerably larger quantities
during the years 1957 to 1974.

The sampling system used for the collection efficiency studies is identi-
cal to that described earlier and illustrated in Figure 1.  A Gelman
20 x 25 cm glass filter (Type A) which has been heated to remove any PCB
contaminant is placed in the stainless steel filter holder.  Three pre-
cleaned polyurethane foam (PUF) plugs are fitted into the throat of the
sampler in the manner described in the method (Appendix A).  A known
quantity of Aroclor, dissolved in isooctane or hexane, is quantitatively
transferred to the center of the filter using a volumetric pi pet.  The
sampler is immediately turned on at an adjusted flow rate of 0.8 m3/min
and the sampling period is commenced.  Samplers were run 20 minutes for
each of these initial recovery studies.  Approximately 16 cubic meters of
air were drawn through the glass fiber filter and the three PUF plugs.
The tests were conducted in the laboratory using large enough PCB spike
concentrations and short enough sampling periods that background levels
could initially be ignored in computation of spike recoveries.

Figure 11 illustrates the first, second, and third PUF plug extract chro-
matograms and a chromatogram of the glass fiber filter extract after a
recovery experiment using Aroclor 1221.  Note the recorder attenuation
factors indicated for these chromatograms.  The first PUF plug chromato-
gram is shown at a recorder attenuation of 16 x 1Q~9, while the others
are shown at attenuation 4 x 10~9.  Those peaks designated by a "•"
have been identified as peaks matching those present in Aroclor 1221;
those designated by an "x" are not present in Aroclor 1221.  The labeled
peaks have been specifically identified using pure isomers of

-------
CTl
 I
 o
                                                                                QJ  >,
                                                                               -Q -O
                                                                                01
                                                                                ro -C
  c:
  O)
  "O
    c
M- 
                                                                               O -t-1  (O
                                                                               S_ i —
                                                                               -C T- -M
                                                                               C_> 4-  ro
                                                                              CD
                                                                               I
                                                                               o
                                                                                                                                                  T3
                                                                                                                                                   3
                                                                    l_)

                                                                    OJ

                                                                    U



                                                                    O)

                                                                    E
                                                                    O
                                                                                                                                                   U
                                                                                                                                                   O)
                                                                                                                                                   O
                                                                                                                                                   o
                                                                                                                                                  CO
                                                                                                                                                  CNJ
                                                                    o
                                                                    o
                                                                    i.
                                                                   •=c
                                                                    
-------
2-monochlorobiphenyl (2-MCB), 4-monochlorobiphenyl (4-MCB), 2,2'-
dichlorobiphenyl (2,2'-DCB), and 2,4-dichlorobiphenyl (2,4-DCB).  It was
later determined that many of those peaks on the chromatograms which are
designated by an "x" originated with the sampler exhaust.  Recirculation
from sampler exhaust to intake was apparently occurring.  This problem
was eliminated on subsequent work.

As demonstrated in Figure ll(a), the Aroclor 1221 can be identified on
the first PDF plug, which comprises the first 7.6 cm of the total length
of 23 cm of PUF collection media.  Figure ll(d), the chromatogram of the
glass fiber filter, demonstrates that the Aroclor 1221 spiked onto the
filter at the initiation of the study was quantitatively volatilized.
Very little of the PCB was detected on the second and third PUF plugs, as
shown in Figures 1Kb) and (c).

The quantitative results of the Aroclor 1221 collection efficiency study
are tabulated in Table 7.  Since 51 percent and 32 percent of commercial
Aroclor 1221 is monochlorobiphenyl and dichlorobiphenyl, respectively, as
shown in Table 6, then the Aroclor 1221 spiked onto  the glass fiber
filter contained 15,7UO ng of the monochlorobiphenyl species and 9,840 ng
of the dichlorobiphenyl species.  Of the remainder,  11 percent of the
total composition is biphenyl and 6 percent is PCB of greater than
dichloro-substitution.  A total of 98 percent of the Aroclor 1221 spike
was volatilized from the glass  fiber filter.  Over 97 percent of the
monochlorobiphenyl was captured on the first two PUF plugs.  All of the
dichlorobiphenyl was captured on the first two PUF plugs.  If biphenyl is
excluded from consideration, since the electron capture detector is not
sensitive to this species, a total of 92 percent of  the PCB was captured
on the first PUF plug and 6 percent was captured on  the second PUF plug,
for a total recovery of 98 percent.

This collection efficiency study with Aroclor 1221 demonstrated that the
sampling system is capable of quantitatively trapping even the most
                                   48

-------
                                TABLE 7

          RESULTS OF AIRBORNE PCB COLLECTION EFFICIENCY STUDY
                           USING AROCLOR 1221
  3U,750 ng of Aroclor 1221 spiked on high-volume filter.  Sampler run
  20 minutes, total of 16 m3 air drawn.
             Monochlorobiphenyl
               ng (% of total)
Dichlorobiphenyl      Total  PCB*
ng (% of total)    ng (% of total)
Spike
1st plug
2nd plug
3rd plug
Filter
15,700
13,600 (87%)
1,100 (7*)
500 (<3%)
550 (3%)
9,840
9,840 (100%)
500 (5%)
<100 (<1%)
<100 (0%)
27,400
25,300 (92%)
1,600 (6%)
<600 (<2%)
<650 (2%)
*  Excluding biphenyl.
                                    49

-------
volatile PCB species present in commercial products.  It also demon-
strated the method, including both the sample collection and analysis,
has the capability of accurate measurement of airborne PCB.

Six recovery studies were then conducted using Aroclor 1016.  These were
conducted in the same manner as the previously described recovery study.
Spike concentrations covered the range of 5.6 to 120 ug of Aroclor 1016.
Figure 12 is a series of typical chromatograms resulting from these
studies.  The one depicted is the 20 ug spike.  The peaks indicated by a
"•" match corresponding peaks in Aroclor 1016.  The peaks designated by
an "x" are the result of contaminants present in the laboratory atmos-
phere.  They do not interfere with the Aroclor 1016 chromatogram.  As can
be seen by this figure, the PCB is quantitatively captured on the first
PUF plug; that is, on the first 7.6 cm of PUF adsorbing media.

The quantitative results of all six studies are tabulated in Table 8.
In general, over 95 percent of the Aroclor 1016 was volatilized from the
glass fiber filter during each test.  The collection efficiency shown in
Table 8 is calculated as follows:
     Collection efficiency (%) =   I$_p  J  X 100;

where PI ,2, 3 ""s tne quantity of PCB found on the first, second, and
third PUF plugs, respectively; S is the amount of PCB spiked onto  the
filter; and F is the PCB residue remaining on the filter at the conclu-
sion of the experiment.  The calculated collection efficiency, therefore,
represents the recovery efficiency of the complete method, incorporating
both the collection efficiency of the sampling device plus the analytical
recovery of the laboratory procedure.  The reported variance, likewise,
encompasses both sampling and laboratory analysis.

A mean collection efficiency of 102 percent was achieved for Aroclor 1016
during this short-term collection study.  The range was 89 to 117  percent
and the  standard deviation was 11 percent.  This level of precision is
                                    50

-------
   XI*
                (a) First  polyurethene
                    foam plug
                                                      (d)  Third polyurethene
                                                          foam plug
               (b)  Second polyurethene
                   foam plug
              fc)  Standard Aroclor 1016
                                                       (e)  Filter
Figure 12.   Aroclor 1016 recovery experiment.
                                     51

-------

























CO

UJ
1
CO

1— '













































oo
UJ
o
rs
t—
00

>—
fY'
UJ
>*
o
0
UJ
as
^o
r— 1
O
,— t

o:
o
_j
o
o
u.
0

oo
h-

— *
oo
UJ
a:






















c
O >>
•i- O
•P C
U Ol
CO •.—
r— O
"o C
O 4-
UJ



01
3

CL O)
C
-o ^-

oo




en
3
f— ~ ^"•*
Q- cn
c
-o >—
c:
CM



cn
3

SI 'cn
c

c/)
i— 1


i.
O>
r-
t/> u. cn
ro C
^~ ^ -^-*
CJ3 O)
i^t
•i-
U.





CO

•.- cn
CL C
oo — •





CTi t — - CM OO
OO «— I Ol O
r-H T~l








O O O O
LO =* 00
\& »— (
*i
CM







O O O O
^O CM OD i^>
r--- r- 1 o
^
^-






o o o o
O O >-Q r~»
 CM O O

LO •-! O 30
—i CM CM





0 0

f-l 1— 1








0 0

I**-- C^
r>
co







0 0
oo o
CM 00
r< *N
CM OO
OO





0 0
CM  t^>





0 0
CTt CT*
CO CM
*
OO






O O
O O
o o
<* n
^O t\J
j^ «— 1
r-H




t
1
•








^•5
^J-











^S
T— 4
i — 1








$?5
LO
00






•—
-3
o
1—

M— "O
O O)
•P 0
C OJ
0) .—
0 t—
i- 0
cu o
Q.
QD
C 0
'O O.
cu
s
^
t— 1
,—-4
+
CM
O
i— 1






































O
c
O)
o
i£
4-
UJ

C
o
IP
o
0)
r-~
i—
0
o

c
fO
11
2:

-------
acceptable tor lengthy organic analysis techniques and should be repre-
sentative of results achievable during routine applications of the
method.

It is important to point out in these tests that an average of 85 percent
of the PCS captured was retained on the first PUF plug and 96 percent was
retained on the first two PUF plugs.  Hence, three PUF plugs of the
dimensions described above have proved sufficient to retain over 100 ug
of airborne PCB.  Furthermore, an average of only 4 percent of the PCB
captured was located on the third PUF plug.  Independent analysis of the
third PUF plug extract would, therefore, permit the analyst to judge the
collection efficiency of the system for the particular application
desired without the necessity of conducting recovery studies.

Figure 13 shows the chromatograms of the highest spike recovery (112 ug)
study.  Again, V indicates Aroclor 1016 peaks and "x" indicates uniden-
tified laboratory atmosphere contaminants.  In this 20-minute test, a
clearly detectable amount of Aroclor 1016 was present on all three PUF
plugs.  The entire chromatographic pattern is observed in the extract
from each PUF plug, and the peak-to-peak ratios appear to be largely
unchanged from one plug to another.  From this it may be concluded, at
least for very short sampling periods (e.g., 20 minutes), that the chro-
matographic pattern is a reflection of true ambient conditions.  Over the
short sampling period there seems to be little modification of the chro-
matographic pattern resulting from selected retention of individual PCB
species.  It will be shown later that this is not the case for sampling
periods in excess of 6 hours.

These initial tests confirmed that polyurethane foam can be a highly
efficient collection medium under the conditions recommended in the pro-
posed high-volume sampling method.  When two PUF plugs are used (15.2 cm
of adsorbent media) over 100 ug of PCB is retained with a high degree
of efficiency.  The inclusion of a third section of PUF provides a
                                   53

-------
o

X
C
O
ro
3
C
QJ
 
 3
i —
 a.
                                                                                     3

                                                                                     'o.
                                                                                     fO
                                                                                     o o
QJ  X
c
fO CO
                                                                                      cu  o
                                                                                      5-  -r-
o  c
Q. OJ
                                                                                     CD

                                                                                     -O
                                                                                          ns

                                                                                         -»->
       en
         o
  O1
    I
    O
                                                                                                                                                         o
                                                                                                                                                         c:
                                                                                                                                                         ai
                                                                                                                                                        •r*
                                                                                                                                                         O
                                                                                                                                                         a>
                                                                                                                                                         c
                                                                                                                                                         o
                                                                                                                                                         O
                                                                                                                                                         OJ
                                                                                                                                                         o
                                                                                                                                                         o
«j      X


O)     r—

3  ro  C
>>     O
i—  OVr-
O  3 4->
<"">  r^  rO
    Q. 3

VI  E  CD
S-  rO •*->
•r-  O 4->
U_  «4-  fO

ro
                                                                                      4-3  tD
                                                                                       03  i—

                                                                                       CT) C
                                                                                       3  O
                                                                                      1	  "I—
                                                                                       CL4->
                                                                                          IT3
                                                                                      •a  3
                                                                                       c  c
                                                                                       o  
                                                                                       a)  4->
                                                                                      OO  n3
                                                                    S-
                                                                    O

                                                                    O
                                                                    O
                                                                    S-
                                                                    c
                                                                    fC
                                                                    o
                                                                    s-
                                                                                                                                                         ro
                                                                                                                                                         S_
                                                                                                                                                         CD
                                                                                                                                                         O

                                                                                                                                                         ro

                                                                                                                                                         O


                                                                                                                                                         CJ>


                                                                                                                                                         ro


                                                                                                                                                         O)

                                                                                                                                                         3
                                                                                                                                                         CT,
                                                                                54

-------
safety factor and a mechanise  of  estimating  the  collection  efiiciency by
independent analysis of  the  third  PDF  plug  in  the  samples.   All  PCB
species from mono-substituted  to  tetra-substituted are  efficiently
retained on the adsorbing  median.   It  is  reasonable to  expect  penta-
substituted through deca-substituted PCB  would,  likewise,  be effectively-
captured from the atmosphere.

Further testing of collection  efficiency  and  recovery at  lower PCR con-
centrations and for longer sampling periods  had  to await  improvement of
the perchlorination method for  PCB  quant Station  as described above,  since
no method was readily available for producing  large volumes  of PCE-free
air to conduct the tests.  It  was  apparent that  unless  the  perchlorina-
tion method could be adapted for  use,  background levels of  PCE and other
substances collected in  the  sample would  obscure the  test  results.  It
would be extremely difficult to discern a low-level  Aroclor  pattern  and
nearly impossible to accurately measure the  amount present  by  pattern
matching techniques.  After  improvement of the perchlorination method and
demonstration that it worked effectively  for the full  range  of PCB
species, collection efficiency  and  recovery  studies were conducted using
a synthetic PCB test mixture,  three single volatile PCE isomers,  and
Aroclor 1242.

The composition of the PCB test mixture is listed  in  Table  9.   The test
mixture was spiked onto  the  ol?s°  fiber filter immediately  before  the
test began in the same mariner  as  /Voder  1221 described above.   The  test
was run cut-of-doors in  Gainesville, Floric-?..  Alon^  v/ith this  and each
succeedirg test, a backgrounc  sample was  taken at  the  sar:e  location
during the same period of  tiro  as  the  test spike sample.  Background PCB
concentrations determined  by analysis  of  the background sample glass
fiber filter anc each of the three P'JF plugs were  subtracted from  the
respective ar,a1ytical results  cf  the test spike  sample  to determine  spike
recovery.  The l.a^k'jrcunr!  :..5n."1cs  are  described  in a  l?ter  section.

-------
                                TABLE 9
                      PCB TEST MIXTURE COMPOSITION
     PCB                   Concentration        Percent Composition
2-MCB
2,4-DiCB
2,4'-DiCB
2,2'-DiCB
2,2',5-TCB
0.44 ug/ml
0.40 ug/ml 	 	
0.52 ug/ml
0.42 ug/ml.^— - —
0.57 ug/ml
18.7%
24.3%
The results of a 1-hour test using the volatile PCB test mixture (2.59 ug
spike) are tabulated in Table 10.  These results are reported as DCB
after perchlorination.  The collection efficiency is 90 percent based on
recovery of the spike.  Of the total PCB collected on the PUF, 93 percent
was captured on the first PUF plug.  This test showed even the most vola-
tile PCB species could be retained over a 60-minute period, and that
satisfactory results could be achieved using the perchlorination
technique.

In order to further define the behavior of the more volatile PCB species
in the system, even longer-term collection efficiency tests were run
using pure isomers of 2,2',5-TCB, 2,4'-DiCB, and 2-MCB.  The results of
these 2-hour (120 m3) tests are tabulated in Table 11.  TCB spike
recovery averaged 91+3 percent for two separate 0.48 ug spikes
(1,115 ng as DCB).  DiCB showed a 103 percent recovery for a 1.08 ug
spike (2,080 ng as DCB).  In the case of both TCB and DiCB, quantitative
capture was achieved on the first PUF plug alone.  MCB demonstrated a
107 percent collection efficiency.  Unlike the other species, however,
MCB penetrated beyond the first PUF plug to the second and the third PUF
                                    56

-------
                                TABLE 1U

       RESULTS OF INITIAL RECOVERY STUDY USING PCB TEST MIXTURE*
Sample Duration (mm)                    6U
Flow Rate (mJ./mirt)                      u.76
Volume Air Sampled (m3)                  45
Ambient Temperature (°C)                 32.5
Ambient .-tumidity ('£)                     39
Spike Concentration (ng 0 DC6)         2590

           Coreted ^
            _
  PCb (ng D'CB)

  Filter                                373
  HUT ffl                               185U
  HUF tit                                  0
  PUF #3                                147
  Total  -                               2370

Percent of Total PCB
  Col lee tea on PUF (%)

  HUF n                                 93
  PUF ff2                                  0
  PUF #3                                  7

Collection Efficiency (%)                90
*  Test mixture composition shown in Table 9.  Sample date 4/22/77,
f  Background levels are described in Table 18.
                                   57

-------
CQ

















CO

"^

Q
sz
^t

t\
CQ
O
a

»
CO
£

CU
z
H- 1
oo
:=)
i— t oo
* — i 1 1 1
LU O
1 rD
co I—
< oo
t—
>-
a:
LiJ
^fc
o
o
LU
DC

Qi
~s
0
:r
i
CM

LL.
O

00
1—

~~>
oo
LU
a:


















+—

CO
v/
•r—
o.
oo

CO
o





•H

co
^^
•r-
Q.
OO

co
o
•r—
O











*
CO
.,_
Q.
OO

CQ
0
H-









































O
s:
co
i~». in CD
r*^ r*^ c.
^-v— 1 O CM
00 1 CM ID LO LO 00 O
CMOO«— i • f-- CM ^J- 
• — • C CO 3 -i- O
•r-l — +-> >> -M CO
C E CX  ro i-
O ^ E t- -I- i- i-
•r-OO ra CO "O -t-> O
4-> E «/) Q.-I- C O
CO CO ro • — - E E CO
+J£$_ J-C03U "O

Q r~ Q •*-> < O 3
 +J O O
cocococt^cocc s-
•— r— r- ECOCOCO CD

E E E O r~ -O (^ *r- O
rOrorOi — OEEQ. ra
UO CO OO Lu > ^ ^ OO CQ


CQ
O
2:

CD
C
o o o
c a ^^ ^^ rj^
VO CO ^
t— H


CQ
O

Q

CD
C
CM
O •* O O
^-1
CM






ca
(_3
Q
CD
C

CM •*
CO O O
O





CO
o
a

CD
C

co en
•* •* o o
O^



























i- ^n CM ro
d) "^tfc; ^ft= ^^
4->
^- Lu LL. U-

Li— Q-- ^- O-









3"^
O ^ «!d- O OO
r^ c\J to oo o
^3* *— H
CVJ








5^
CM O CO
•* O O O O
T— 1 f— 1 I— 1
CM











^Q
^O O OO
co o o o cr>
O • r-l
t ^









**—*.
^

t~~-
cr> Li-
en ~~i
Q.

C
O

-o
a>
4J
o
Q)
r—
^—
O
O

CQ
O
Q_

,—
03
4^>
O
1—

4-
O

r- 4->
fQ C
4-> CO
o o
1— S-
co
a.


&^
O CTl
O O O CO
«— 4














>^
o
c
CO

o

t^.
l^
LU

c
o
^-1 CM OO -r-
^- ^2 ^.. ,^>
u
Ll_ U- Ll_ O)

0-0-0. r—
0
o




CO

•r-
O.
oo

.
o
•r-
4«>
^
JD
•o
CO

ro
••->
'Jj
C
ro
3
CO"

CQ
CO O
__J ^r
^1 ^1.
CO CM

t^ ^O
T5 C
^™ ^3

C CQ
•r- 0

T3 Q

3
O <*
4— «
CM
CO
_^^ 5^,
o
>> ^+—
fQ
E c
O
ra -T-
^_> ,!_>
ra • ra
T3 CQ 4->
O -r-
CO 1— 4->
i— 1 C
Q.LO (O
E » 3
ro - CT
01 CM
** ^s
T3 CM ro
C CO
3 CD O.
0 3
S_ CO
CDU3 r—
-V r~~ cn
(J LO C
ro * *r~
CQ O OO


-K 4-
58

-------
plugs as  indicated  in  Taole  li.   Twenty percent ot tie total MCB was
captured  on  tne  third  PDF  plug  and trie ^redt-^st amount (54 percent) was
detected  on  the  second  racier  tnan toe first PUF plug.  Apparently, the
more volatile MCB moves  through  the PUF at a significantly faster  rate
than either  DiC8 or  TCii,   The  2-hour sampling period  is, therefore, aoout
the maximum  duration tnat  can  be usea whi 1.3 achieving quantitative cap-
ture and  retention of  '1C6, out  is entirely sjitaole for the capture and
retention of PC8 species of  >)i<]h.:r substitution.  Shorter sampling
periods would ue more  suitable  for the collection of  MCB as demonstrated
by the quantitative  recovery achieved during the 2l)-minute test using
Aroclor 1221 as  described  above.

A 12-hour test was then  conducted using rlCB, DiCB, and Aroclor 1242 to
determine the maximum  length of  time that various PCBs could be expected
to De retained on tne  PUF material.  Tne results of this test are
reported  in  Taole 12.   As  expected, during this 12-hour sampling period
in which  634 m^  of air  were  drawn through the system, MCB was not
effectively  retained.   A collection efficiency of only 4 percent was
achieved  for the 'J.52  ug spike.   All  ilCB residual */as located on the
second and tnira PUF plugs,  with 54 percent of the total on the third
plug.  Ninety-tnree  percent  of  the DiCB spike (2.08 ug) was captured
during tni s  test, but  2o percent ^as located on the third PUF plug,
b3 percent on the second,  •jnd  only ly percent on t;ie  first.  It is clear
from tnis anu tr.c- .jreceeanu tlat:1. fc^ tne 20-minute and 2-hour tests,
that eacr, PC3 species  laove-s  througn the PUF in a i)3id >nuch like it would
move tnrough a crirouiat^jre^tiic  coljmi.  Tr.t lower-substituted species
(e.g., MCB,  DiCti) inovs  inore  rapidly through tne syst?n than the ninhor-
substituted  species  (e.o,, TCB).  Tir's will, tkie^eroro, li^'t tne
duration  of  sampling tnat  ":an  be o-noloy-l.

Aroclor 1242 was offecti voV ca^tu^oc: '}: t >e cirst y!JF pi j • (Tabl3 12)
during a  12-hour test  run  usny  3 10.4 uj spi'ce.  Aroclor i242 is
reported  (Taole  b) to  contain  1  percent '!C;j, 1': percent OiC? and tne
remainder TCn an-1 more  hi go ly-substi tut:c PCB sj'-iJes.  H?nc;?, only MCG
limits tne sa,:iplinu  aurati >n to  ^ ess fon 12 nou-s.

-------
















•y.
03
O
•r—
Q
C3
^g»
•

*
CM
CM
r-l
DC
O
^1
O
o
ct:
^c

CD
2Z
>— «
CM <£}
r— 1 — '
LU f
	 | UJ
CQ •— '
-
O£
LU
^^
O
0
UJ
C£
cr:
»
o
T~
I
CM
r-l

U,
0

00
t—
	 1
' )
00
UJ
a:















•K
*
C
3
CC

CD
\x
• r-
Q.
00

03
O
S
' x*
i ^
^0
CQ • T
O i —
^ CO CU
o s-
C?) T
C I -P
CT C
0 C (U
CO O ! -i-
CO«^-CM O~OO JO
•OO« CXJ'd'OO _'O^-'3" P?
O {~O LO r-l C\J OO 'O

*
jf.
C
3
0£
CU
•r*
a.
00

co
c >
•r*



CQ CQ
O O
(^ C3
cn en
c c

o o o o o
CO CO r-~ CM LO •*
OO *zt C3 OO O UO CTi
• CO- ^ « crioooooo
Q| OVOCM r-l r-l r-luDCMCTl


4_

c
3
o;

0)
JM:
•r—
a.
oo

CM
CM
r-i

i.
o

(^
o
i.
c^


CM CM
^f ^d"
CM CM
r-l r-l

S-. i-
0 O
f— !—
O <->
0 O
^ «^

o> cn
c c

o 
O
O)
r—
CQ — r—
0 O
a. u
— — *
OO O ^3
ECO) 0
-— — • O 4-> 3_
C T- U
•r- -a -P 01 i—
E OJ (O i_ fO
-^_^ ____ r {_ i _^
OO Q.-P O O
E E c o :—
— fO CU















-— ^
5"^
^^.

>^
(j
C
OJ
• r—
{^)
•r—
'4—
4-
Lul

oo o -a 4-1 c
cu c c o
•P i- O 3 r— 1 r\J OO
(O *r— C_5 O ^t= -tt: "ifc r— -P
o£  cnl u. LI_ LJ_ P a>
3 ' -^ _*: r33r> o u
i o i — T- o a_ ^_ OL. i — 5_
O
— 1 CM -O -r-
-^fc ^: -tfc -P
O
u_ u_ u- a>
1 ~~1 ^ 1 r~
"^ ^>j_ -^_ r—
i — O Q- fO O) i O

0
0
CM
CM

a>
3
^O
J,
a>
Q.
p^
O)


C
a>
JD

^j


•
r-^
1 	

uO
-. — ^
CM
r-l
C
o

-•— - •
I/) OD
a> o
•P 2:
3 1
C CM
• r—
£ "X3
C
O rtS
OJ
r-~ m
— O
r-l Q
ur> i
CTl td-
r-H *•
1 CM
r-H •
LO -•— ^ * O>
r~ r— c: c
O r— 0 ••-
*O •! — CO
t3 4- -P 3
o c  C
t. T3 •<- O
OJ i- -P -r-
Q. C -P
>7 fr^ (rj
a> > 3 -P
£T fO O"-r*
•P OJ -P
_C CD C
"O CT) — • C , S- 0 U
r- 3 0 P J*
fg "Q f—^ f^ f^J
c E a>
(O C >, Q.
3 P C
-P i- -r- i- CD
o -a 

(= •«-> C II


-------
A 6-hcur n ro\'( ry test  using Aroclor  1242  was  run  in  duj. 1 icr.to.   The
results of ihi 5  test a PC. tabulated  in  Table  13.  The  neon  collection
efficiency was <'T? percent  after  background correction.   Since  relatively
large spire conc^iitrat iors  (10.4  ug)  tu r*_  employed,  quantitafion of  this
test could be  accomplished  by  pattern  matching  methods,  as  demonstrated
in the following figures.   In  the case  of  both  spike-  samples  and the
background sample, ever C9  percent  of  the  total  PCB  Vv'as  captured on  the
first and  second PIT f,lugs.  The  third  plug  in  each  case hao  little  or no
detectable PCB.

Figure 14  is the chromatogram  of  an Aroclor  1242 standard  at  the same GC
conditions as  the recovery  sample shown in Figure  15(a), (b),  and (c).
^igure 15  (a)  is the chronatcgram of  the first  PIT plug  showing  a com-
plete Aroclor  1242 pattern.  Peaks  matching  those  in  the Aroclor 1242
standard are designated by  a "•' .   In  addition  to  the Aroclor  1242 that
was captured,  other earlier eluting compounds  are  evident.  Two  isomers
of DDE (o, p'-DDE and p, p'-DCE)  have  apparently been captured  from  the
ambient atmosphere on the first PIT plug,  although these peaks  have  not
been confirmed.  The first  PIT  plug retained a  total  of  11.6  ug  of
Aroclor 1242,  while the second  ana  thirc PIT plugs retained 0.27 jq  and
C.C7 ug, respectively.  The chromatcgrarrs  for  the  second arc!  third PUP
plug extracts  are shown in  Enure 15(b)  and  (c), respectively.   Only
traces of  PCB  remained  on the  glass fiber-  filter.   In or> r>" t^  determine
the collection efficiency  in this test,  background PCt hoc  to  be
subtracted fron the filter  and  each PUF  plug.   Chroi'.atogrdi:;s  of  the  back-
grouno sample  taken during  this test  are discussed in a  l^ter  section
(sec Figures 25 and 26).  After correction for  the background  residual
nCe, the first Plf  plug extract contained  9.08  ug  Aroclcr  1242 while the
second and thi^c! PUP pl.'C extracts  contained background  ccrn;<_ted concen-
trations of 0.17 ug and 0.07 ug Aroclor 1242,  respectively.

A final series of recovery  tests  wen?  conducted  using Aroclor  1242 to
rier,orrn'ne  if recovery was dependent on  airborne  PCE  concentration.  The
results are tabulated in Table  14.  Spike  concentrations of 10.4 ug,
                                   61

-------











CM
^-
CM
r— 1
Q;
O
	 1
c_>
o
a:

CD
i— *
OO
i
oo
i— I oo
UJ
UJ •— i
CO ra
< t—
1— 00
>_
rv^
l-lj
^»
O
0
UJ
C£

DC
l
O
1
^r>
u.
o

00

	 1
1
oo
UJ
CtL





-1-
4-
T3
C
O
s-
tn
v^
o
TO
CO
4-
CQ

C
3
a:

cu
•^
a.
oo]

LT>
r~ oo
~^^H ooocnoo r>-ooo
O I *~^ ^O O C7> CT»
CM kO O •-! r-l
^r-l CTl O O I
•— I CM Ln • VO I OO OO
i— 1 i— < OO •— 1 OO 1

LO
r^ <— i
r^ co LO
•^ •— i o ooocr>o o r^ooo oo^o «^" •
O 1 O ' — I ^0 CT^ i — I CTi CT» i — I CTt r-n i — I \ — t
CSIVO'O'vJ" i— IOO LO CMCM LOCTl
-— i CM '.n • ^ o CM CM cr> en
t— It— 1 OO O OO •— < i— 1 r- 1





^
*
O1
00



4_

<
C
3
a:

CD
vx
•r—
a.
oo



LT>
[^^ i — t OO
r^. oo •
~^^-i o OOOCMCM o r^CMr-H OOCMCM ->j- r>.
C3 1 CZJ «~^ LO r1^ P^* i—i CTi «^ f^ (^ CO 00
CM^OCM'd" 'X>C\JO CO r— 1 O
^^. T— I O^ CT> t-H ^ " ^ " ^
^-iCMtn-ooo •— i CM co 
O CM U
i- r- 1 CU

t_ f —
\ C7> O O
C r- 0


















	 .— (J
OO O =Q
ECS- O
-— — O < Q.
C -r-
•i- "O +J C7) i —
£— S QJ ^ C~ 'O
0 ^-^^ S- — -t-»
•r-OO O--P O
-X +-» E S c: TS (—
 E S- OO O C *^—
ra -r- 3 a> c: •!- o
Ol — O-Pi-O (til-!— i.\J<"O
T3-r—O +->aJ3=H=1fc r 	 +->
CU CU CU Q£ °£ O)4-> T3 C:
^-^-r— CJJ Qi — LL.LL.U. •*-> CU
a-a-rD-S'-*: -r-rDrsrD o u
E E E O — •- nLL.OLa.a- i— s_
(t3a3T5 — OQ.O CD
COOOOOU-S-OOQ. 3_
CM
CM
f — 1

I
o

o
o
s~


ccn
c~
% 	 .

CO
0
a.

•a

^_>
o
cu
i-
S-
0


~o
c~
r-( c\j ro a
^~ ^~ -f^ Q
i_
[ i 1 i [ i_ O^
r3 ^3 ^) ^
a. CL. a. o
13
22
















, —
&^
- -

>^
(,J
c:
OJ
•t^
o
•r~
4—
4-
wU

C
0
1_ — ( CM CO •<-
0) -tfr ^t= ft: . — -(->
+J re o c
Lu 0- a. .T_ I— i— ~*L
0
0























un
i-H

cu
S-
3
.,_
Li_

•r-

C
^
o
je
co

CO
E

O "t— C~
'-O -P-i-
CM C
(O C
CU 3 3
s- cro
3 _C
•»-> cento
T3 C
&« -r— CO
cu -C E
^^ (_) *^
E •»-> S-
GJ fQ O")
-*-> e o
-}->
•(-> E (0
c i- E
a> cu o
•.- -M S_
l.S§
-K +-+-

-------


JTT 	 , 	 : ., -- - - - . - . . .______.


j 	 . 	 . 	 T~~~ 	 " 	 ' 	 ' 	 - 	 ~~-~ 	 • • • — 	 	 	 - — — 	 — 	 	 	 	 	 — 	 	 -T 	 	 — 	


| 	 	 	 : 	 • 	 • 	 ~ 	 H
0 	 10 	 20 	 	 30- 	 40- 	 £0 	 60 	 70— — -80 -- - -
	 	 1— — — __ 	 	 , 	 „ 	 	 __ j-_ 	 ^_ — ,- — _ 	 , 	 - _ 	 . — . — — . — p_ — — 	 , 	 —
I ' ! '
( '

f— — • — - 	 	 > 	 — — 	 	 7— — — 	 — • — - — • — — ' 	 —~ 	 — 	 	 	 	 	 • 	 — 	 ~.~~ 	 ,--— — --
I ^ ' , • , \
j 	 __j — ___ 	 , 	 __ 	 ,. — __ 	 ___ — r 	 ^^ 	 _ 	 _ 	 . 	 . — , — 	 _ — 	


"
. i- -~-^

— — <-— 	 — 	 	



90 - -- 100
— . 	 — — -




"^
	
o

 X




 c
4->
-M




 s-

T3

 03



CM

CM
                                                                                                                                   o
                                                                                                                                   S-
                                                                                                                                  <=c
                                                                                                                                   S_
                                                                                                                                   cn
                                                                                                                                   O
                                                                                                                                   O
                                                                                                                                   S-
                                                                                                                                  CJ)

    -—:	-J-	.-	
100	90
                                    -70—
                                                 -60  ---   - 50-'  ----- 40  —   -  3D ----
 I	 .

                                                                                                                                  01
                                                                                                                                  S-
                                                                63

-------
                                                                                                                           o
                                                                                                                      i   en
                                                                                                                           O
                                                                                                                           ra
0	10	20	30	40	50—- 	60-	70 - -  -^--,-80	90-—       100
                                                                                                                          Lf)
                                                                                                                          s_
                                                                                                                          rs
                                                                                                                          en

-------
                 70	80	 -90	100
65

-------
    f^ O
    I — CM
    \  I
    CO LT>     00
    •\ <* O  r~-
    CM CO CM   '
    * — I • — i • — I  O
                                    O
                             ^O    OO
                               • m «— '
                            li— ICO   «
CM
             O
             Lf)
             00
CM
     LO    to     r-^
i    \ "^J- O  OO CM   • OO 3-I
    CM CO CM   • O •—I 00  »
    •—t «—( *—^  CD *~* *~~*     ^J"
«d-
, — i

UJ
_l
1—
















































i
OO !
UJ
1— 4
Q
ID
00 :
y_
ro
o£. 1 r-~ LT>
LL!
o
o
UJ
oz
cc.
o
1
CM

Ij
0

OO
t—
_J
=>
ul
Q£








t ^" 1 ^-
\ •-< O
CO 1 CD
CM rO CT> ^"
^ Ln o '£> oo LT> co ••
r— i in CM •cocM'd-o






. — .
CM

CM


O
r—
^J
O
L-
^_ "^
O en
0 C
. — . • — • —
cr ro • — •
; -i- E o> s-s c








— , 	 . 	 . i_ ^ 	 Qt
— C =! -r-
•i — O -4-* !>^ ^-^
C £ CL1 T3 -t-5 03
0 -^r- i. -r- S-
•i-OO O. 
•X +-> E E= 0.-I- C
nj u 03 — • 03 E E 3-'
•»-> E S- OO CU 3 U
1 03 -r- 13 O) h- ~ C
O 1— C3 •*-> i- O
1 03 -t- -(-> -1-5 O
a) a> 11 a: 
OO v/^ wO U_ ^* "^C cJI OO








0 0 O O| O r^
XI X) CTi
0 i 0










i
0 0
t__-j ^ 	 ^


Cxi
^j-
CM


S-
0
r—
U
o
,3;

en
c:
— ^
CQ
C_3
~^_

•^
CU
40
0
O)
^
^_
o


T3
c~
^3
0
s_
en
sx
u
T5
^D
























c
0
.,_
-4--)
03
^— ^ i ( ii
99 -r-
- — - 4-^
^
>•> 03
U ! Z3
C O~
OJ
•i- ^o en
o ! u cr
-1-1 N -r-
"^~ 1 ' ^i ' —
4-1 ' r- U
UJ 1 ! 03 4->
' ~ 03
C 03 —
O
^ r-l >J ,-0 -t- ' *-> C
CU*:^* ^- ->-> OS-
•!-> O3 O CD
r— U_ U. U. 4-> O 4->
•r- r> r3 ZD o i — u +->
U. ^_ X O- 1 — — • O3
O 
-------
b.U ug, and 1.1 ug were used along with a 2-hour sampling period.
Collection efficiencies were 97, 96, and 111 percent, respectively, for
the three spike concentrations after subtraction of background levels.
These  values lie within the normal range of analytical variability for
trie method,  hence, there is no apparent interrelationship between
recovery and concentration over the range tested.

The results of all the recovery tests that were conducted to validate the
sampling and analytical method are summarized in Table 15.  The following
conclusions may be drawn from these studies:

a.  There is no apparent variation in either collection efficiency or
    recovery as a result of variations in sample flow rate over the range
    of 0.6 to 1.0 m^/min.

b.  Ambient temperature and ambient humidity exerted no obvious influence
    on collection efficiency.  One test was conducted during heavy rain-
    fall with satisfactory results.

c.  The pattern matching method of quantitation is reasonably accurate
    for the analysis of Aroclor concentrations in excess of 1 ug in the
    PUF extract.  Perchlorination is required for lower concentrations.
    When the sample elution pattern does not closely resemble an Aroclor,
    the perclorination method should be routinely used to achieve
    accurate results.

d.  The various PCB species move through the PUF collection medium much
    like they would move through a chro
-------
c
0 i
I CM
0>

-(-> U-
o r:
o> o_



CTi O
1 ^ + ^- 00 + 1 oo +1
o &*




















OO
1—
OO

._
CC
Lf) LU
o
LU O
_J UJ
O •— < -—
=*»=
! 	 ) |
ra ^
•4-1 O
0
1—

S-?
c: v_~
0
£ o
o c
Ol O)

O -f—
0 <4-
LU
ro
s- &
5.
•Q
• 
•2 O)







CTi CT:
•^~ c^I5
CTi O
, — ^






,__i
r-l OO
0 +10 +1
o en
^H CM •— <
o en
^_— )





VO ^O VO CO
i — i ' — t *^~ C^i











O O 0 CO
CM CM 'O C\J
1 — (



O O '-O ^O
CO X) r^ r^
• • * *
O O 0 O







^ — 1 '>O < — I CM


-K
O)
i-
3
X
>r_
i "
cu
Sfc.
to"
r-l O OJ
3 CM r-l I—
+j CM CD v\
X •— I •— ( O O
•r- O
"^ CL>
1 \ -!->
CO LO

O) Ol
1 -r- (—
O
-r- t—
S- S- +-> 1
o o o> -n

"u "o ••->
o O c: rvj
S- S- >,

0.

v/O
I





r~-
• 1
o
0 CM O ^D 0 +|0
O ^- CO OO O O
r— 1 r— 1 O •— 1
o
, — 1







oo co
00 OO CO •*  +| +|
o cr> o * — i
• — l . — 1 i — 1 CTl i — 1
CO O



CM
o
LO <* u-i ^a- o 01
r^ oo r- oo CM oo i
|-O '*o i — • oo
OO
CO








o o o o o cy> o
CM CM CM CM CM LO CM
I*™ ( r**^ »~^ r^^ r~^- co *~~ ^

in
-Xi
CM ro CM x> •* o
'*O "O lsO CO C^ CT)
.... 1
0 O O O r-l O
CT)
'-O
C5




	 1 t— t T— 1 >— 1 r— t CM OO










CM CM CM
^" "^ '^
CM CM CM
r— 1 r-l r-l
±Q CD
0 O S- 1- 1-
•r- -r- OOO
O O '33 CD — . — i —
1 1 CJ O O O O
<* ^ S ?" O O O










































.
03
O
) —

^
CM
-o
S
CO
O
• r—
a

s-s
r»*s
UO
^
CQ
O
s^
cr\
r-H

01
c
•1^
c
•1 —
^ts
-^
r « i i s_ j_ &. ;r
CM ,"M ."NJ %) <=X «=£ f^t O


C_5
-K

-------

-------
         total airborne RGB present in the atmospheric sample, as was found to
         be the, case in the field tests discussed later, then it is estimated
         from the recovery tests that up to a 6-hour sampling period can be
         accommodated with satisfactory recoveries.  TCB and more highly-
         substituted species are retained for sampling periods in excess of
         12 hours.

     f.  The mean collection efficiency for all  the tests conducted for
         Aroclor 1016/1242 was 101 + 10 percent with sampling periods ranging
         from 20 minutes to 12 hours.

     g.  When the duration of sampling is within the limits specified in 'e'
         above, independent assessment of the amount of PCB captured on the
         third PUF plug will  allow the analyst to confirm a high collection
         efficiency for the airborne PCB without the necessity of conducting
         recovery tests.

3.2  PRELIMINARY FIELD TESTS
     During the course of development of the recommended method presented in
     Appendix A, some preliminary field investigations were necessary.
     Certain aspects of these preliminary field data have been previously
     discussed in the appropriate section of this report.  Brief mention will
     be made at this point of the findings from those preliminary investiga-
     tions which have not been discussed in an earlier chapter.

     Samples were taken of urban ambient atmosphere in Jacksonville, Florida,
     using a single polyurethane foam plug of the dimensions described in the
     method in each high-volume sampler.  Collection efficiency is, therefore,
     not known, but it is expected to be less than the laboratory studies
     where three polyurethane foam plugs were used in each sampler.  Several
     samples were collected periodically over a 24-hour period.  Sample peri-
     ods of 2 to 24 hours were selected.  The volume of air sampled and the
     analytical results are shown in Figure 16.  No PCB was detected in the
     glass fiber filters used in the sampling apparatus.  These data are
     considered to represent minimum ambient airborne PCB concentrations.
                                         69

-------
      o
      s-
ro
o
cu
      o
      o
      CM
      o
      o
      o
      o
      §•
      o
      o
      g-
      o
      o
      o
        '
      o
      o
      CM '
      o
      o
      o
      o
      o
      CM
      CM
      o
      o
      o'
      CM
      o
      o.
      oo
      o
      o
      UD '
-o ro
 £ E

ro en
LD sr

r~ un
   CM
  CO
   E

   CD
                    Sample
                  Location A
C3
O
                                                        o
                                                       -s
                                                        O


                                                       •§
                                                        O
                                                       •§
                                                        o
                                                        o
                                                       •g
                                                        o
                                                        o
                                                        o
                                                       '
                                                        o
                                                        o
                                                       ' CM
                                                        o
                                                        o
                                                        o
 o
, o
 CM
 CM
                                                        o
                                                       •g
                                                        CM
                                                        o
                                                        o
                                                        CO
                                                        o
                                                        o
                                                       ~
                                                             "3
                                                             Q
                                                                                     cu
                                                                                     E
                  Sample  Location B
                cu

                c
                o
                                                                                               c
                                                                                               cu
                ro
               -O
                S-
                (O

                c
                                                                                               to
                                                                                               c
                                                                                               o
               +J
                c:
                QJ
                O
                sz
                o
CO
o
CL.

QJ
cr
s_
o
.a
s_

ro

•»->
c
cu

-O
E
ro
                                                                                               cu
                                                                                               S-
                                                                                               cn
                                             70

-------
Values observed ranged from lb ng/m^ to 25 ng/m^ at  sample location A
and from 3 ng/m^ to 3G ng/m^ at sample location B.   At  sample  location B,
there appears to be a greater variability with time  of  day than at sample
location A.

Figure 17 illustrates a typical chromatogram obtained during this test.
There are numerous extraneous peaks on the chromatogram which  make
interpretation and quantisation difficult.  At the time this sample was
taken, precautions were not used to prevent sampler  recirculation since
the problem was not known.  It appears that the PCB  present falls in the
range of Aroclors 1242 and 1254.  A "•" indicates peaks that clearly
match corresponding Aroclor peaks.

A second preliminary field evaluation was conducted  at  sample  location B
in Jacksonville.  Three polyurethane foam plugs were used this time and
the maximum sampling period was shortened to 6 hours.   Unfortunately, all
analytical results were less than the detection limit.  Ambient levels
were apparently significantly lower at this time.

A third sampling was conducted at this location to obtain additional
field samples for evaluation.  Samples were collected at 4-hour intervals
over a 24-hour period.  The results are tabulated in Table 16.  A reason-
ably uniform PCS concentration was measured over the course of the 24-
hour period.  Concentrations varied from 4 ng/ai^ to 9 ng/m .   The
chrornatograms obtained have been discussed in a previous chapter of this
report.
                                   71

-------
                                                    ,CM
                                                     CM
                                                            
-------
                                TABLE  Ib

                    RGB  IN AMBIENT URBAN ATMOSPHERE
                                                  Total PCB
        Date                Period           Concentration (ng/m3)*
11/2U/75
11/2U/75
11/2U/75
11/20/75
11/21/75
11/21/75
^Reported as decachl
12UO-160U
1200-18UO
lbOU-2400
2UOO-24UO
24UO-0600
U600-120U
orobiphenyl .
9
8
9
7
4
6

Environmental Protection Agency (EPA) Region II personnel employed the
high-volume sampling method using polyurethane foam to assess ambient
airborne concentrations of PCB in the vicinity of a major industrial
user.  The samples collected by EPA personnel were analyzed in this labo-
ratory.  This sampling effort provided preliminary field test conditions
for relatively high airborne concentrations.

Two polyurethane foam plugs were used in the standard high-volume
sampler.  Each sampler was operated for a period of 4 hours.  Both
plugs were analyzed together, so collection efficiency cannot be
determined.
                                    73

-------
Figure 18(b) shows the chromatogram of the polyurethane plug extracts.
A standard Aroclor 1016 pattern is shown in Figure 18(a).  Those peaks
designated by a "•" match corresponding peaks in the Aroclor 1016
standard.  Those designated by an "x" do not match Aroclor peaks.  The
elution pattern closely resembles that of Aroclor 1016.  Four peaks in
the pattern do not correspond with Aroclor peaks.  Two of these resemble
PCN, but this has not been confirmed.  Quantitation of the three samples
taken in the area yielded airborne concentrations of 498 ng/m3,
400 ng/m3, and 300 ng/m3 as Aroclor 1016.

Confirmation of the PCB present was obtained by gas chromatography/mass
spectrometry (GC/MS).  An AEI Model 30 (Associated Electrical Industries)
GC/MS was used with the OV-17/QF-1 column under the operating conditions
described in the appendix, except helium was used in place of nitrogen
tor the carrier gas.  The results of GC/MS analysis are depicted in
Figure 19.  The GC/MS was operated in the peak monitoring mode at mass
values corresponding to each of the major PCB species.  Good correspon-
dence may be seen between the patterns for Aroclor 1016 and the airborne
sample for trichlorobiphenyl and tetrachlorobiphenyl.  Semi-quantitative
information can be gained by comparing peak heights of the sample and the
standard in this mode of operation.  When this is done, it is apparent
that trichlorobiphenyl is enriched in the sample relative to Aroclor 1016
and tetrachlorobiphenyl is reduced in the sample relative to Aroclor
1016.  That is, the ratio of trichlorobiphenyl concentration to tetra-
chlorobiphenyl concentration in the air  sample is significantly greater
than the Aroclor 1016 standard.  This indicates the effect of selective
transport of the trichlorobiphenyl species which are generally more vola-
tile than the tetrachlorobiphenyl  species.  GC/MS and  GC/EC analysis also
indicated the absence of monochlorobiphenyl species and the presence of
trace amounts of the dichlorobiphenyl species.

High-volume air samples were collected near an electric power substation
and transformer storage facility operated by a small electric coopera-
tive in northern Florida.  The sampler was placed approximately 20 meters

-------
(a)
(b)
                                                                      Min
Figure 13.   (a)  Standard  Aroclor  1016;  (b) Chromatograph of air sample
            taken in the  vicinity of  a  major  industrial user.
                                      75

-------
                                      (a)
Air sample;
(Mass 256)
trichlorobiphenyl
                                       (b)   Standard  Aroclor  1016;  trichloro-
                                            biphenyl  (Mass  256)
                                       (c)   Air sample;  tetrachlorobiphenyl
                                            (Mass 290)
                                       (d)  Standard Aroclor  1016;  tetrachloro-
                                           biphenyl (Mass 290)
                                     10
     15
   Min
Figure 19.  GC/MS analysis of ambient air sample taken
a major industrial user of PCB.   All  chromatograms are
monitoring mode at the mass indicated.
            in the vicinity
            taken in peak
                of
                                       76

-------
from tne transformer storage area and 1UO meters from the  substation.
Analysis was by gas chrornatography using an electron capture detector
(GC/EC) and confirmation was obtained by GC/MS.  The analytical  results
are summarized in Table 17.  Biphenyl was found to be present  at an
estimated concentration of Ib ng/m3.  This was detected by GC/MS.
Klonquantifiaole traces of monochlorobiphenyl and dichlorobiphenyl were
detected by GC/tC and confirmed by GC/MS.  Most PCb present appeared by
both GC/EC and GC/MS analysis to be trichlorobiphenyl.  A concentration
of 3U ng/m3 was estimated for this species.  Tetrachlorobiphenyl was
also present at b ng/m3.

                                TABLE 17

    RESULTS OF HIGH-VOLUME AMBIENT AIR SAMPLE TAKEN IN THE VICINITY
    UF AN ELECTRICAL SUBSTATION AND TRANSFORMER STORAGE FACILITY

Species
biphenyl
monochlorobiphenyl
dichlorobiphenyl
trichlorobiphenyl
tetrachlorobiphenyl
Analytical
GC/EC
ND
X, trace
X, trace
X, strong
X
Method
GC/MS
X
X, trace
X, trace
X, strong
X
Concentration
(ng/m3)
15 ng/m3 (est.)
1 ng/m3 (est.)
no est.
30 ng/m3
6 ng/m3
Sampleu IbO minutes, total of 185 m3 of air drawn.  No PCB was
present in substitution greater than tetrachlorobiphenyl.
 X = detected
ND = not detected
Tnese preliminary field tests provided some experimental data  that
applies to practical application of the method  under  field conditions.
The preliminary experimental field data were of value  in development of
the method suitaDle for routine field use.  These  preliminary  field data
do not, however, constitute a field test of the method  since significant
                                   77

-------
       changes were incorporated into the method as  a result of this  field
       experience.   The following section describes  valid field tests of the
       final  method for PCB in air as described in Appendix A.

  3.3  FIELD TESTS OF THE FINAL METHOD
       The sampling and analytical method, as described in Appendix A, was  field
       tested on several occasions in Gainesville, Florida, and on two occasions
       in New Bedford, Massachusetts.  The procedure was followed as  written
       without deviation.  In each sample, the glass fiber filter and each  of
       the three polyurethane foam (PUF) plugs were  analyzed separately.  Air-
       borne PCB concentrations were measured over the range of 4 ng/m^ to
       1.4 ug/m3 during these field tests under a variety of ambient  temperature
       and humidity conditions.

3.3.1  Tests in Gainesville, Florida
       Table 18 is a summary of the results of several field tests conducted
       over a 9-month period in Gainesville, Florida.  Sampling periods range
       from 30 minutes to 12 hours.  Measured airborne PCB concentrations varied
       from 11 ng/m3 to 44 ng/m3 as DCB.  On the average, 94+3 percent
       of the PCB captured from the atmosphere was retained on the filter and
       first two PUF plugs.  There are no apparent effects of ambient temper-
       ature or humidity on collection efficiency over the ranges encountered
       during these field tests.  There is also no statistically significant
       correlation between collection efficiency and sample flow rate or
       duration.

       Figures 20 through 23 illustrate the CC/EC chromatograms for the test
       conducted on July 26, 1977.  Figure 20(a) shows the chromatogram of an
       unexposed PUF plug from the set of plugs used  in the tests.  No detect-
       able PCB is present.  Figure 20(b) is a chromatogram of the glass fiber
       filter extract from the 2-hour ambient sample taken on that date.
       Perchlorination yielded a concentration of 74 ng PCB as DCR on this
       filter, which is below the detection limit by pattern quantitation.
       Figure 21 is a chromatogram of an Aroclor 1242 standard.  Figure 22 is

-------

































cxi


UJ
CO

I—

































































Q
cc
0

U-

*
uu

H— ' 1
LU
i—*
s

T*»
t— 4
*^
LU
h-
0
5
3
o
0

oo
i —
00
UJ
1—

Q

LU
r^

Lu
o
OO
1—
— 1
00
LU
o;


























*
•X LO '31 O O '-O -O
r*-- i — l LO • '%O CM LO
r-- oo oo i— i oo
^•~- f—H
O 1
CM UO
	 .— 1
r-H CM
,-H r-H
-X CO O CM UC2 OO OO
!^ CD CM 1 — 'DO CM r-.
r~ oo r— i •
r-H 1
oo oo
--.O
O r-l
r-H rH


+_
GO OO *— ( CM l-O CTl
r^ LO CM O*J r— ( OO OO
r~*. <^f- i — i • , — i
~~~, t— l CD
'-D 1
CM LO
^ LO
.— (

^H CO r-- LO CM 00
1^ OO CD Cn CD CM oo
r-- CM •— ' « .— i
\,-i O
r""** i
CNJ CO
^"*-» «^*
^ o
,-H




^-O Cl^ C2 C1D 'NJ C7>
r*^> co ^o • *>D ^o co
f^v. '^O < — I
""-^. i-H
CNJ 1
OO ijO
-^x, SV~)
*3- un
r— t





f~^«, ^\j C^) LO O^ C\J O^
r*^ cvj oo o^ cvi oo oo
^ *-<
CM -O ^
	 LO
<1- —1
t — 1







. 	 * — -
c oo • — •
'r— ^ ^&
£ ^-.—^^-~-~- .
••- -a o >,
c: H? cu — •*-*
•i— oo CL • "O
•(-> E= E a.-»-
 ^
fQ -^- 33 CU 1 — -*-
a l— a +-> s-
f^j ,,^ ^_> 4_>
CLI a> cu ci  1)
Q. C- C- ~Z • -r- •<-
= E Cr O -— -Q -Q
fO fO fO f~ "~ O c~~ .*""
oo ^ ^ co 1 1- ^* ^^c e^

LO O O ^f CT* O LO ^ O O
oo cr^ i — i LO oo • "^ • • • f~~-
COC--J — 1 O OO^-IC^ r-t
'-O i-O CTi




OO O O ^d" |P^ *Tl ^—1 *>J i — l 0s) • — f
r-^ '^o ^~* 3O ICM • f*^ ' — ' • • ^~^
-.0 <— i |cn r-- cyi o








^- ,^2 oo CD |r~^ O^ C5 C5 ^— < CD LO
P*- CM 'vO LO [CD • 1^- CM • • OO
r~- P— oo CTI ^H CXD CM
CM |oo en






OOOOCMOOK CM r-. i— 1 . OO
OO ^f CM <— ( r— 1
CM |OO CT>








tO CD CD CM ^O X) LO CM LO CD ^"
LD CO OO ^J < — 1 i — ' ^t" OO • ^f
^3- ^-i OO ^n LO LO
r-i CM en










LO .-H t— i oo CD LO •— i en oo ILO oo
LO'X)CMLOCM sj- VO CM <— 1 1 • ^)"
UO OO r-< CVJ jg-
1-4 '
















c
•r-

c
1
~

c:
TO
i.
CD
o
-p

^
E
o
o

>1

1 —
TO
o

00 +J
CM TO
-o 53
C -C
TO -4->
TO
CM" ~
CM -O
rS
o"-*
CM CM CT>
c:
l/» OJ •!-
O) i- JC •
^ I 1,^-^2
CO -r- U_ = --~
•a rj o U- .Q
•l> "^ o ' c: c: • — •
•CO U Lu ^B.' ••— H TO
C_5 CD TTD ' C3 4-* "~-^
CD — d. OO ' c: 3; 4->
r— ?E ~Z O TO "^
C?J O "t~ ^-^ ' O JC~ O- CM
C.5 C7I JC 00
C i— -S^ —

-O O ZD O S- S- d) LO
O) 1 — O_C O"JCTH->CM
J-> O ! O O fO
0 ^ +0
CU O
•t-J 4-> J-) t/l
TO TO *r- CU
i — S_r- ICMOO S_ — 1 CM -^O 4. O) 1 = "r +J 4-
r— CU **= ^ ^t= r— -4~* sU '^ "*^ -^ T,1 *^ ' O O C 33
O4->" T3 C+-* •»-> i- ' i- i_ -3 O1
CD ~~~ 1 i 1 i i i_ 4-J CD ~~ l i i i ; t — O ! ,CT C 73 •»"•
•i- TD n ZD O '-.> ••- ^3 ^> ZO -t- O ! O O '"D* LU
02 ij. a, o. a. i — t- U- a. CL d- u_ *•
O O) -r- -t- -K -K
a. o_ 
-------
                                                        dOIS
                                                                        0£
                                                                                 dH
                              SS  '66
                                                    8S6
                                          7. yaau
   60 7
   is •
   ib

 JOJ.S,
                                                                                             C\J

                                                                                             r^
                                                                                              c
                                                                                              o
                                                                                              10
                                                                                              (U
                                                                                              -o

                                                                                              O)
                                                                                              O)
                                                                                              -o
                                                                                               OJ
                                                                                               CO
                                                                 SO 'Z
zs -p
                                                         dOIS,
0£  '
   d
  b08££
                                                                                            en 

                                                                                           a. -I-
                   =) S-
                   Q. OJ
                      _Q
                   T3 -r-
                   o; u_
                   (/)
                   O CO
                   Q- CO
                   X rt3
                   O) i—
                   C CJ3

                      QJ
                                                                                            O O
                                   OOT
                                                                             IS
                                                                   3dAi
                                                                          dOIS ,
                    n3  03
                    S_  S-
                    CT) CD
                    o  o
                    4->  4->
                    fO  ro
                    E  E
                    O  O
                    S-  S_
                    ^:  _c
                    CJ  CJ
                                                                                            LU LU
                                                                                             ro JD
                                                                                            O
                                                                                            
-------
76 '07
                                          oe  s
                                                                                                               
 (O

 o
 s_
                                                                                                              o
                                                                                                              C3
                                                                                                              CM


                                                                                                               
-------
                                                                          QJ
                                                                           O
                                                                           s_
             16'£2
 .*-**  fO
en  -a
 i   -I—
 o  i-
 r—  O
  X r—
 CO Ll-
                                                                           C  OJ
                                                                               to
                                                                          •»->  a>
                                                                           o  £=
                                                                           re •!-
                                                                           s_   O
                                                                           X
                                                                          UJ  E
                                                                           rs r--
                                                                          i — r--
                                                                          CL. ~^.
                                                                              V£)
                                                                          U_ OsJ
       9S '
   1/1  O)
   i_  -i«£
  •i-  re
  u-  I—

   D  gj
  -G  i—
  -(->  Q.

  M-  re
   o  oo

   E  S-
   re  -r-
   S-  

   re  QJ

   o ****
                                                                           C_5 3
                                                                           UJ O
                                                                           C_>  I
                                                                           O CM
90  T
                                                                           C\J
                                                                           CM
                                                                            a>
                                                                            s-
                                                                            cn
                          82

-------
                                                                                                      u
                                                                                                      fO

                                                                                                      4J
                                                                                                      X C
D-

u.
                                                                                                         CSJ
                                                                                                         c:
                                                                                                         HI
                                                                                                         QJ
                                                                                                      QJ •—
                                                                                                     j= Q.
                                                                                                     4-> E
                                                                                                         fO
                                                                                                     -— -co
es -
                                                                                 dOlS
                                                                                                      13
                                                              sr "9
                                                               *?  5
                                                                                                         C
                                                                                                         o>
                                                                                                      OJ
                                                                                                     oo
                                                                                                      QJ
                                                                                                      (T3
                                                                                                         d)
                                                                                                      O     ro
                                                                                                         E -O
                                                                                                      E  O -r-
                                                                                                      ro  i- J-
                                                                                                      S- M- O
                                                                         9f  '
 rO  I    "
 = O OJ
 O i — i —
 S-  X r-
JC 00 -r-
C->     >
    •• CO
C_5  C O)
UJ +J C

O «=C rO
                                                            oc
  S  0
CO
C\J
                                                                                                      
                                              83

-------
the chromatogram of the extract of the first PuF plug in the sampler.
Those peals de.nctec hy a V coincide with peaks in Aroclor 1242.
Figures 23(a) and  (b) are chromatograms of the second and third PUV
plugs, respectively, from this test.  Note that all compounds eluting
after 10.6 minutes are quantitatively captured on the first PUF plug and
all those eluting  after C.2 minutes on the first two PUF plugs.  The
third PUF plug extract contains no peaks matching Aroclor peaks.  The
unidentified compounds eluting at 15.8 and 10.6 minutes are captured with
high efficiency in the system.  The unidentified compound eluting at
3.6 minutes is also effectively captured.  The unidentified compounds
eluting at 1.97, 2.5, and 3.0 m'nutes, respectively, appear to saturate
all three PUF plugs and, hence, are probably ineffectively captured.   In
this test, the first PUF plug contained 2720 ng PCE as DCB upon
perchlorination, the second 763 ng, and the third 350 ng.  The measured
airborne concentration was 35 ng/rn^ as DCB.  Most of the PCB peaks
detected were in the elution range of tri-, tetra-, and pentachlorobi-
phenyl species.  The 2-hour ambient air sample taken in the same location
on October 31, 1977, yielded the chromatogram shov;n in Figure 24(a).
Peaks in this chromatogram v,hich coincide with Aroclor peaks are
designated by a "•", but the pattern match is poor.  After perchlori-
nation of the extract, as shown in Figure 24(b), quantitation is direct
and unequivocal.   This sample contained 660 ng as OCR on the first PUF
plug, 110 ng on the second, and 84 ng on the third.  The measured
airborne concentration was 11 ng/m3 as DCB.  This is approximately
one-third the ambient level measured on July 26, 1977.

PCB collection efficiency remains high for longer than 2-hour sampling
periods.  Therefore, it may be desired to obtain a  longer time-integrated
sample.  Figures 25 and 26(a), (b), (c) show the chromatograms of the
filter and three PUF plug extracts from a 6-hour air sample.  PCB species
are found on the first and second PUF plugs (indicated by a "•"), but  are
absent from the filter and the third PUF plug.  MCB was not detected.
This sample was quantitated by pattern matching, yielding an airborne
concentration of 8.8 ng/nr as Aroclor 1242 or 17 ng/m3 as DCB.   The
pattern most closely resembles Aroclor 1242, but not all peaks are
present.  Hence, the concentration reported incorporates the uncertainty
inherent in the pattern quantitatiot: r.ethod.

-------
85

-------
Figure 25
GC/EC Chromatogram (Attn:
(PUF #1)  Taken 11/20/77  in
      )  of 6-Hour
Gainesville,  Flori
Ambient Air Sample
da.
                                      86

-------
Figure 26.   GC/EC Cnromatograms  of (a)  Second  PUF  Plug,  (b) Third  PUF  Plug
            and   c)  Glass  Fiber  Filter  (Attn:  16xlQ-9) Of  6-Hour Ambient Air
            Sample Taken  11/20/77  in  Gainesville,  Florida.
                                       87

-------
       It may be noted in Figure 25 that DDE is apparently present in this air
       sample and is quantitatively captured (cf. Figure 26)  on the first PDF
       plug, although this peak has not been confirmed by mass spectrometry.
       The sampling system would, therefore, serve as well for the measurement
       of airborne DDE concentration.

3.3.2  Tests in New Bedford, Massachusetts
       Two capacitor manufacturers are located in New Bedford, Massachussetts.
       The New Bedford municipal landfill served as the site  of the disposal of
       reject capacitors and other PCB-containing wastes from these industries
       for many years prior to 1975.  Over 22,000 kg of PCB have been disposed
       of in this landfill over the years (EPA, 1976).  PCB wastes have not been
       placed here since 1975.  Air samples were taken at this landfill site on
       June 28-30,1977, and again during January, 1978.

       The results of the first field test conducted at the New Bedford landfill
       are tabulated in Table 19.  One-hour and 6-hour samples were taken at the
       landfill site on June 28 and June 30, 1977, respectively.  In this case,
       quantitation of the PCB was easily accomplished by pattern matching tech-
       niques, since relatively large amounts of PCB were collected and the
       pattern closely resembled that of Aroclor 1242.

       Figure 27 is a chromatogram of the glass fiber filter  extract from the
       60-minute air sample at this site.  Marginally quantifiable traces of PCB
       are present on the filter.  Figure 28 is a chromatogram an Aroclor 1242
       standard at the same GC operating conditions.  Figure  29 shows the
       chromatogram of the first PUF plug extract after an 8:1 dilution.  The
       pattern closely resembles that of Aroclor 1242.  The full Aroclor pattern
       is present and is easily quantifiable in the PUF plug  extract.  The
       second and third PUF extracts, shown in Figures 30 and 31, respectively,
       appear to be devoid of quantifiable amounts of Aroclor 1242.  Collection
       of PCB on the first PUF plug is, therefore, quantitative.

       On Figure 29, the "x" denotes the retention times of o,p'-DDT and
       p,p'-DDT.  It appears, but has not been confirmed by mass spectrometry,
                                         88

-------
                                 TABLE ly

      RESULTS OF FIELD TESTS CONDUCTED IN NEW BEDFORD, MASSACHUSETTS,
                              IN JUNE, 1977
Sample Date* 6/28/77 t
Sample Time 1059-1159
Sample Duration (iriin) 60
Wind Direction WSW
Wind Speed (mph)
Sample Flow Rate (m^/min)
Vol. Ai r Sampled (nr)
PCB Collected (ug Aroclor 1242)
Filter
PUF ,fl
PUF #2
PUF f3
Total
Percent of Total PC6 Collected
Filter
PUF #1
PUF ifZ
PUF #3
Filter + PUF #1 + PUF #2
Airborne PCB Cone, (ug/m^
Aroclor 1242)
Biphenyl Collected (ng/m^)
10
1.02
61
0.2
53. y
<0.2
<0.2
54.1

0.4
99.6
<0.4

-------
Figure 27.   GC/EC  Chromatogram of  the  Filter  Extract  (Attn:  16x10   )
            from the 60-Minute Ambient Air  Sample  Taken  at  New  Bedford
            Landfill.
                            90

-------
       rdtzrt:
       J-4
Figure 28.  GC/EC Chromatogram (Attn:  16xlO"9)  of  an  Aroclor
            1242 Standard.
                                91

-------
Figure 29.   GC/EC Chromatogram of the  First PUF  Plug  Extract  (Attn:  15xlO~9
            Diluted to 1/8) of the 60-Minute Ambient Air  Sample  Taken at
            New Bedford Landfill.
                                        92

-------



:!
        " —41-

                                          ______ _(,__

Figure 30.  GC/EC Chromatogram of the Second PUF Plug Extract  (Attn:  16x10"
            of the 60-Minute Ambient Air Sample Taken at New Bedford  Landfi
                              93

-------
                                                  r-EE






31.   GC/EC Chromatogram of the Third PUF Plug Extract (Attn:  16 x 10"9)
     of the 60-Minute Ambient Air Sample Taken at New Bedford Landfill.
                           94

-------
that DDT is also present in these air samples and is quantitatively
captured on the first PUF plug.

The extracts from the 6U-minute air samples were then perchlorinated for
confirmation of the PCB.  Tne results were as follows:
                          filter     173 ng DCB
                          PUF #1  72,800 ng DCB
                          PUF #2     129 ng DCB
                          PUF #3      43 ng DCB
PCb is confirmed and the high collection efficiency is validated by
perch!orination.  The total amount of DCB in the filter and first PUF
plug extracts (73 ug) can be converted to equivalent Aroclor 1242 using
the conversion factor in Table 1 of Appendix A.  Applying this conver-
sion, an equivalent of 0.62 ug/m^ of Aroclor 1242 is calculated.  This
compares to 0.89 ug/m^ determined by pattern matching.  The 30 percent
discrepancy in the amount measured by the two methods in this case may be
attributed to two factors:  (a) the large amount of dilution required to
quantitate the first plug extract in effect results in a multiplication
of analytical variability; and (b) it is assumed, both for pattern quan-
titation and for conversion of perchlorination results to equivalent
Aroclor 1242, that Aroclor 1242 is present alone and in unmodified form.
These necessary assumptions limit the degree of agreement that may be
expected between pattern quantitation and perchlorination results.  It
would not normally be necessary to perchlorinate such high level samples,
however, for purposes of quantitation.

The samples taken in the vicinity of the New Bedford landfill  were also
checked tor the presence of biphenyl (BIPH) using the analytical method
described in Appendix A (Section 12.0).  3IPH was absent from all 1-hour
and 6-hour filter and PUF plug extracts.  Hence, BIPH was collected in
concentrations below the detection limit of 0.5 ng/nr, if at all.
Figure 32 is a GC/FID (flame ionization detector) chroinatogram of a BIPH
standard.  Figure 33 is a chromatog ram of the first PUF plug extract from
the 60-minute sample.  Mote the absence of a BIPH peak in Figure 33.

-------
_.	;__ L~~


 1	

Figure  32.
GC/FID Chromatogram  of Biphenyl
at 2.8 ng.
                    96

-------
             1
                               ^£
Figure 33.   GC/FID Chromatogram of the First PUF Plug
            Extract from the 60-Minute Ambient Air Sample
            Taken at New Bedford Landfill
                         97

-------
Since relatively high airborne PCB levels were detected over the Mew
Bedford landfill, a second sampling trip was scheduled in January, 1973,
in an attempt to determine if the landfill itself was the source of the
airborne PCB, or if one or more of the other potential sources in the
area were emitting significant amounts of PCB.  The samping was conducted
by EPA Region I personnel using the procedure in Appendix A.  Sair.pl ing
was conducted upwind and downwind of the two PCC user facilities in the
area, the municipal sewage sludge incinerator, ano the landfill.  Sam-
pling duration was 3 hours in each case.  The ground was frozen at the
time of sampling and a light snow was falling.  The analytical results
are summarized in Table 20.  One may conclude from these data that the
landfill is a possible low-level PCB emitter during the winter months as
compared to the sunnier months.  Some PCC  is possibly emitted from the
municipal sewage sludge incinerator, but PCB user A is clearly a major
emitter in the area.  PCB user B apparently had no significant emissions
at the time of sampling.

Figure 34 shows the GC elution pattern for the sample taken over the
landfill.  The pattern closely resembles Aroclor 1016/1242.  Each peak
designated by a "•" matches a peak in Aroclor 1242.  At lower levels such
as this, it is not possible to unequivocally distinguish between these
two similar Aroclors.

Figure 35(a) shows the air sample taken upwind of PCB user A, v/hile
Figure 35(b) shows the sample taken downwind.  The downwind sample
clearly contains a substantial amount of Aroclor 1016.

The field tests described above demonstrate the practical use of the
method described in Appendix A for the sampling and analysis of ambient
airborne PCB.  The tests covered a wide range of ambient concentrations.
In the Gainesville location, the PCB was  not generally recognizable by
pattern matching and perchlorination was  required for quantitation,
while at the New Bedford location, the PCB pattern was clear.  In all
field tests conducted, an average of 96 + 4 percent of the PCB collected

-------
                                 TABLE 20

      RESULTS OF FIELD TESTS CONDUCTED IN NEW BEDFORD, MASSACHUSETTS,
                             IN JANUARY, 1978
Date
  Site
1/17   landfill
1/24


1/27


1/19
PCB user A


PCB user B
Location    Concentration (ng/m3)     Aroclor
                upwind
                on site
                downwind
sewage sludge   upwind
incinerator     downwind
upwind
downwind

upwind
downwind
  8.5
 21
 13

  4.3
 13

  5.6
490

 19
  5.1
1242/1016
1242/1016
1242/1016

1242/1016
1242/1016

1242/1016
1016 only

1242/1016
1242/1016

-------
Figure 34.   GC/EC Chromatogram of  Air  Sample Taken  over New  Bedford,
            MA,Landfill.   0.5  ul  Injection, Attn:   64  x 1Q-9
                                  100

-------
                                            O  I
                                            ••- o
                                            o
                                            O)  X
                                          - C sj--
                                            •p- UD
10
              20
                            30
                                                                                                                                re

                                                                                                                                T3
                                                                                                                                Q.
                                                                                                                                cu
                                                                                                                                CO +->
                                                                                                                                O) -i-
                                                                                                                                E  o
                                                                                                                                re  re
                                                                                                                                co H-

                                                                                                                                i.  i.
                                                                                                                               • i-  O)
                                                                                                                                ro  to
                                                                                                                                O CO
                                                                                                                                   o
                                                                                                                                CO D_
 o  o
-t->
 ro TD
 E  c
 o -r-
 S-  3
JC  C
 O  2
    O
O 13
                                                                                                                                CIS
                                                                                                                               en
                                                                                                                               ro
                                                                                                                                O)

                                                                                                                                :3
                                                                                                                                en
                                                    101

-------
was captured on the first and second PUF plugs.  Less than 4 percent was
detected on the third PUF plug.  Several unidentified, non-PCB compounds
were also trapped with a high degree of efficiency while others passed
through the collection medium.  Interferences were not, however,
encountered which inhibited quantisation of the PCB by perchlorination at
the lower levels or by pattern at the higher levels.  Sample handling
techniques as described in Appendix A have been demonstrated to be ade-
quate to avoid unwanted PCB contamination in the field.  Furthermore, it
has been demonstrated that the sampling procedure can be followed by
persons other than the authors with satisfactory results.
                                   102

-------
.0   CONCLUSIONS  AND  RECOMMENDATIONS
    A  method  has been  developed  for  the  sampling  and  analysis  of polychlo-
    rinated biphenyls  (PCBs)  in  air.   An easily constructed,  high-volume
    sampling  system  is employed  with  porous  polyurethane  foam  as the  collec-
    tion  medium.   The  sample  is  collected at the  rate of  0.6  to  1.0 m^/min.
    The method is  described  in detail  in Appendix A.

    Laboratory procedures  have been  developed,  and are described in  this
    document, which  permit the quantitative  analysis  of even  the most vola-
    tile  PCB  species in an air sample,   A perchlorination technique  for  the
    quantitative analysis  of  PCB has  been adapted for use with air  samples.
    The technique  is shown to convert even the  most volatile  PCB species to
    decachlorobiphenyl  for simple and direct quantitative analysis.   Data  is
    presented to show  conversion efficiencies of  a variety of  PCBs  to deca-
    chlorobiphenyl of  101  +  6 percent over the  range  of 0.103  to 10.0 ug.  A
    ruggedness test  was conducted which  indicates the proposed perchlorina-
    tion  technique can yield  reliable interlaboratory results.  The  perchlo-
    rination  technique is  a  recommended  optional  method of quantitation  of
    PCB when  the commonly  used pattern matching techniques are found  to  be
    unsuitable as  a  result of either  low quantities of collection or  poor
    Aroclor pattern  recognition.   It  is  the  opinion of the authors  that  the
    use of the perchlorination technique is  generally necessary  for  the  anal-
    ysis  of low  (i.e., less  than 25  ng/m^)  airborne levels of  PCB.   The
    analytical metnod  is effective for the analysis of airborne  PCB  levels
    within at least  the range of 1 ng/m^ to  50  ug/rn .

    The total sampling and analytical  method was  subjected to  rigorous spike
    recovery  tests under a variety of ambient conditions.  The mean  collec-
    tion  efficiency  for Aroclor  1016/1242 was 101 + 10 percent for  sampling
    periods of 20  minutes  to  12  hours and air volumes of  16 to 720  cubic
    meters.   It  is demonstrated  tnat  monochlorobiphenyl is the most  difficult
    species to capture. A 2-hour sampling duration is the maximum  that  may
    be employed  to quantitatively capture monochlorobiphenyl.
                                       103

-------
Dichlorobiphenyl and more highly substituted species may be quantita-
tively captured using up to 6-hour sampling periods.  Data is presented
to show that the analysis of the final one-third section of the poly-
urethane foam collection medium is a direct and reliable method of
estimating system collection efficiency under specified system operating
conditions.

Field tests of the method were conducted under a variety of ambient con-
ditions including both low and high PC8 levels.  Typical ambient airborne
PCB data is presented and discussed.  In most locations the PCB that is
collected does not closely resemble a single or a simple mixture of
Aroclor;-hence, perchlorination is required for quantisation.  In loca-
tions near known or suspected sources of PCB, where high ambient levels
are detected, Aroclor patterns are generally discernable and pattern
matching techniques of quantitation may be employed.

The proposed method is recommended for the analysis of ambient airborne
PCB.  Both laboratory and field tests have demonstrated that the analyst
may apply this method with a reasonably high degree of confidence in the
resulting data.  The accumulation of additional ambient data and addi-
tional experience in practical application will provide for a continuing
evaluation of the method under an ever expanding range of conditions.
                                   104

-------
6.0  REFERENCES

     Armour, J.S.   1973.   Quantitative Perchlorination of Polychlorinated
          Biphenyls as a Method for Confirmatory Residue Measurement and
          Identification.   Journal  of the Association of Official  Analytical
          Chemists, 56(4) :987-993.

     Bidleman, T.F. and Olney, C.E.  1974a.   Chlorinated Hydrocarbons in the
          Sargasso Sea Atmosphere and Surface Water.   Science, 183:516-518.

     Bidleman, T.F. and Olney, C.E.  1974b.   High-Volume Collection of
          Atmospheric Polychlorinated Biphenyls.  Bulletin of Environmental
          Contamination and Toxicology, 2:5.

     Giam, C.S.,  Chan, H.S., and Neff, G.S.   1975.   Rapid and Inexpensive
          Method  for Determination  of Polychlorinated Biphenyls and
          Phthalates in Air.  Analytical  Chemistry,  47(13) :2319-2320.

     Harvey, G.R.  and Steinhauer, W.G.  1974.  Atmospheric Transport of
          Polychlorobiphenyls to the North Atlantic.   Atmospheric
          Environment, 8:777-782.

     Hutzinger, 0., Safe,  S., and Zitko,  V.   1974.   The Chemistry  of PCB's.
          CRC Press, Cleveland, Ohio.

     Lewis,  R.G.  and Zimmerman, N.J.  1976,   Danger  of Re-Circulation in
          Hi-Vols.  U.S.  Environmental Protection Agency Environmental
          Monitoring and  Support Laboratory,  Cincinnati, Ohio.  Analytical
          Quality  Control  Newsletter #28.

     Lewis,  R.G.,  Brown,  A.R., and  Jackson,  M.D.  1977.  Evaluation of Poly-
          urethane Foam for Sampling of Pesticides,  Polychlorinated Biphenyls
          and Polychlorinated Nephthalenes in Ambient Air.  Analytical
          Chemistry, 49(12) :1668-1672.

     Margeson, J.H.  1977.   Methodology for  Measurement of Polychlorinated
          Biphenyls in Ambient Air  and Stationary Sources—A Review.  U.S.
          Environmental  Protection  Agency, Environmental Monitoring and
          Support  Laboratory, Research Triangle Park, North Carolina.

     Mieure, J.P., Hicks,  0., Kaley, R.G., and Saeger, V.W.  1975.
          Characterization of Polychlorinated Biphenyls.  Paper presented at
          National Conference on Polychlorinated Biphenyls, November 19-21,
          1975, Chicago,  Illinois.

     Murphy, D.G.   1972.   Sulfuric  Acid for  the Cleanup of Animal  Tissue for
          Analysis of Acid-Stable Chlorinated Hydrocarbon Residue.   Journal
          of the  Association of Official  Analytical  Chemists, 55(6):
          1360-1362.
                                     105

-------
Rhoades, J.W., Johnson, D.E., and Lewis, R.G.  1977.  Evaluation of
     Collection Media for Measurement of Pesticides and Polychlorinated
     Biphenyls in Ambient Air.  173rd National  Meeting, American
     Chemical Society.  New Orleans, Louisiana, March 25, 1977, Paper
     No. 77.

Rice, C.P., Olney, C.E., and Bidleman, T.F.  1976.  Use of Polyurethane
     Foam to Collect Trace Amounts of Chlorinated Hydrocarbons and Other
     Organics From Air.  Paper Presented at World Health Organization
     Meeting, Gothenburg, Sweden, October, 1976.

Snyder, D. and Reinert, R.  1971.  Rapid Separation of Polychlorinated
     Biphenyls from DDT and Its Analogs on Silica Gel.  Bulletin of
     Environmental Contamination and Toxicology, 6(5):385-390.

Trotter, W.J. and Young, S.J.V., 1975.  Limitation on the Use of
     Antimony Pentachloride for Perchlorination of Polychlorinated
     Biphenyls.  Journal of the Association of Official Analytical
     Chemists, 58:3.

U.S. Code of Federal Regulations.  1975.  Procedures for the Measurement
     of Total Suspended Airborne Particulate.  Title 40, Part 50,
     pp. 12-16.

U.S. Environmental Protection Agency.  Region I.  1976.  New England PCB
     Waste Management Study.  Boston, Massachusetts.

Youden, W.J. and Steiner, E.H.  1975.  Statistical Manual of the Associ-
     ation of Official Analytical Chemists:  Statistical Techniques for
     Collaborative Tests; Planning and Analysis of Results of Col-
     laborative Tests.  Association of Official Analytical Chemists,
     Washington, D.C.  88 pp.

Zitko, V., Hutzinger, 0. and Safe, S.  1971.  Bulletin of Environmental
     Contamination and Toxicology, 6:160.
                                106

-------
                   APPENDIX A




METHOD FOR POLYCHLORINATED BIPHENYL (PCB) IN AIR

-------
                   METHOD FOR-



     POLYCHLORINATED BIPHENYL (PCB) IN AIR
                  Prepared by



  ENVIRONMENTAL SCIENCE AND ENGINEERING, INC.



       P.O. BOX 13454, UNIVERSITY STATION




          GAINESVILLE, FLORIDA  326U4








                      for



            QUALITY ASSURANCE BRANCH



ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY



       OFFICE OF RESEARCH AND DEVELOPMENT




     U. S. ENVIRONMENTAL PROTECTION AGENCY



       RESEARCH TRIANGLE PARK, NC  27711



                      and



           OFFICE OF TOXIC SUBSTANCES



        ENVIRONMENTAL PROTECTION AGENCY



            WASHINGTON, D.C.  20460
              Contract 68-01-2978



                 October, 1977

-------
The mention of trade names of commercial products and suppliers in
this procedure is for purposes of illustration only, and does not
constitute endorsement  or recommendation  for  use by  the U.S.
Environmental  Protection Agency.

-------
                METHOD FOR POLYCHLORINATED BIPHENYL (PCB)  IN AIR

1.0  Principle of the Method
     Airborne polychlorinated biphenyl  (PCB)  is collected  on porous poly-
     urethane foam by use of a modified high-volume (hi-vol) air sampling
     device.d)   The PCB is extracted from the polyurethane with an organic
     solvent.  Chlorinated pesticides and other interferences are removed by
     column chromatography prior to gas chromatographic (GC) analysis using
     an electron capture (EC) detector.   When a typical  PCB pattern is not
     recognizable, analysis is performed by perchlorination of all  PCB to
     decachlorobiphenyl  (DCB) and subsequent quantisation  as DCB.  When
     biphenyl is present, or to demonstrate the absence of biphenyl, GC/FID
     (flame ionization detector) analysis is required to  correct the
     analytical  results  for this non-chlorinated species.   Figure 1
     summarizes the steps this method entails.

2.0  Range of the Method
     A detection limit of 1 ng/in^ of PCB reported as DCB  is achievable
     using a four-hour sampling period.  The upper limit  of the method is
     in excess of 50 ug/m^ for short sampling periods (e.g., 20 minutes).

3.0  Interferences
     Solvents, reagents, glassware, and other sample processing hardware
     may yield discrete  artifacts causing misinterpretation of gas
     chromatograms.  All of these materials must be demonstrated to be free
     from interference under the conditions of this analysis.  Specific
     selection of reagents and purification of solvents by distillation in
     all-glass systems is required.

3.2  Toxaphene, phthalate esters, DDE,  and polychlorinated naphthalenes can
     cause interferences with the determination of PCB if they are present.

-------
                         CLEANUP OF POLYURETHANE
                    FOAM PLUGS, GLASS FIBER FILTERS
                        AND SAMPLING APPARATUS
                            Field Sampling
                         Soxhlet Extraction of
                        Polyurethane Foam Plugs
                        and Glass Fiber Filters
                              Silica  Gel
                              Separat ion
  GC/EC Analysis
  of PCB Mixture
GC/FID Analysis
 of Biphenyl
  PCB  Quantitation
by Pattern Matching
                                                              Pesticides
                                                               Fraction
Perchlorination
                           GC/EC Analysis  of
                           Decachlorobiphenyl
           Figure 1.   Analytical  Scheme for PCB in Ambient  Air

-------
  4.U  Precision and Accuracy

  4.1  Analytical.   A test of the  perchlorination  methodology using both
       Aroclor lU16a and a PCB mixture containing  monochlorobiphenyl,
       dichlorobiphenyl, and trichlorobiphenyl  was conducted over the  range of
       0.103  ug/ml  to 10.0 ug/ml.   Eighteen analyses were performed.   The  mean
       percent conversion of PCB  to DCB was 100 +  4 percent over this  range.   A
       ruggedness test(2) Of the  perchlorination method indicated inter!ab-
       oratory results obtained by competent trace analytical chemists should
       be readily comparable using this method.

  4.2  Sampling.  The recovery during tests of the method using Aroclor 1016
       and Aroclor 1242 spiked onto the glass fiber filter that preceeds  the
       polyurethane foam (PUF) collection media in the sampling device (see
       Figure 2), averaged 101 +  12 percent over the range of 1.25 ug  to  112
       ug employing sampling periods between 20 minutes and 12 hours.  Three
       PUF plugs were used at an  average air flow  rate of 0.8 m^/min.

  5.0  Apparatus, Materials, Instrumentation, and  Reagents

  5.1  Sampling Apparatus and Materials.

5.1.1  Modified hi-vol air sampler described in Section 6.3 (Figures 2 and 3)^.

5.1.2  Variable voltage controller to adjust flow  rate of hi-vol sampler.

5.1.3  Rotometer for measuring hi-vol flow rate0.
       a. Aroclor 1016® is a commercial  product containing dichlorobiphenyls
          trichlorobiphenyls, and tetrachiorobiphenyls.
       b. Basic unit may be obtained from General Metals Works, Cleves, OH.
          This unit is recommended for total  suspended particulate sampling
          (Code of Federal Regulations,  Title 40, Part 50, pp. 12-16,
          July 1, 1975).
       c. May be obtained from:  Dwyer Instruments, Inc., P.O. Box 618,
                                 Marietta, GA  30061

-------
                                              Faceplate
                                               Stainless Steel Throat Extension
                                                       Polyurethane Foam
                                                       Plus Location
                                              Throat Extension
                                                    Wire
                                                    Retainer
                                                     Motor Unit
                                                    '\daptcr
                                       Exhaust Duct
                                       (3 in minimum  length)
Figure 2.   Assembled sampler and shelter with exploded view of the filter holder,

-------
                                   Throat Modification
Figure 3.  Exploded view of hi-vol sampling apparatus
           designed for use with a rotometer showing
           modified throat to accoraodate polyurethane foam
           plugs.

-------
5.1.4  Glass fiber filters (Gelman Type A, 25 cm x 20 cm).

5.1.5  Polyurethane foam.  Three inch thick sheet.  Density of 0.025 g/ccd.

5.1.6  Glass jars (e.g., one pint fruit jars) suitable for  storage and
       transport of the polyurethane foam plugs.

5.1.7  Collapsible exhaust ducting.

  5.2  Analytical Apparatus and Materials.

5.2.1  Soxlet extraction apparatus (5.5 cm x 20 cm) equipped with a 300 ml
       florence flask and a reflux condenser (ground glass  joint)— Kontes6
       item #K-585000, or equivalent.

5.2.2  Hot plates or heating mantle.

5.2.3  Perchlorination reaction vial--16 x 100 mm glass screw top culture
       tube with teflon-lined screw cap is modified by drawing the bottom
       into a cone shape as depicted in Figure 4.  This may be easily
       accomplished by heating the bottom of the culture tube and pulling it
       out to a uniformly tapered tip.

5.2.4  Disposable Pasteur pipets--5 3/4 inch long.

5.2.5  Glass boiling beads--2 mm diameter.

5.2.6  Aluminum block tube heater—capable of maintaining a temperature of
       160 + 3°C.  Suitable to receive reaction vials (Section 5.2.3).
       d. May be obtained from:  Flexible Foam Products, Grand Sheet
                                 Metal Products Co., 1101 W. Wisdom St.,
                                 Chattanooga, TN  37406, Product # 1636
       e. Kontes Glass Co.
          Vinland, NJ

-------
                        Teflon-lined cap
               fl6 mm
Figure 4.   Perchlorination reaction vial,

-------
5.2.7  Gas flow meter--range of 50-500  ml/min.

5.2.b  N2 organic removal  filter—organic  contamination  of compressed  N£  has
       been shown to be a  problem.   A filter constructed as shown  in Figure 5
       has been found to be effective in  removing these  organic  contaminants.

5.2.9  Syringe--! to 10 ul.

5.2.10 Hypo-vials®—10 ml  capacity  with Teflon  crimp-on  septa.   May be
       obtained from Piercef (item  # 12903).

5.2.11 Kuderna-Danish evaporating apparatus with modified micro-Snyder
       condenser-Kontes item #K-569250, or equivalent.

5.2.12 Chromatographic column (13 mm OD x 250 mm) with  200 ml  reservoir and
       teflon stopcock (Kontes item # K-420280) for preparation  of silica gel
       column (10.21).

5.2.13 Kuderna-Danish evaporating apparatus with three-ball macro-Snyder
       column--Kontes item # K-570000,  or equivalent.

5.2.14 Hengar® granules, micro for  boiling chips.

5.2.15 Miscellaneous laboratory glassware.

  5.3  Analytical Instrumentation.

5.3.1  Gas chromatograph equipped with an electron capture detector.

5.3.2  Potentiometric strip chart recorder compatible with the gas Chromato-
       graphic detector system.
       f. Pierce
          Box 117
          Rockford, IL  61105

-------
                                                                         . Hood
                        Nitrogen
                Polyurethane
                    Foam
                     Activated
                     Charcoal
                     Polyurethane
                       Foam

                    Drying Agent
                      (Dri-Rite)
                      Glass Wool
                                          Variable
                                          Voltage
                                          Control
                                           Heating
                                            Tape
                                      Glass  Tubing
                                      120cm  x  10  cm
Polyurethane
Foam Plug
                                                               *   •
rigure  5.  Apparatus  for drying polyurethane foam plugs.

-------
5.3.3  The GC column used for both  identification  of  the  PCB  isomers  present
       and pattern quantitation  is  1.5%/1.95% OV-17/QF-1  on Chromosorb W-HP,
       8U/10U mesh in glass 180  cm  x 2  mm ID  column.

5.3.4  A suitable confirmatory column is  4%/6% SE-30/OV-210 on  Chromosorb
       W-HP, 8U/100 mesh in 180  cm  x 2  mm ID  glass column.

5.3.5  The GC column used for quantitation of DCB  after perchlorination  is  3
       percent SP-2401 on 80/100 Supelcoport® in glass  column 90  cm  x 2  mm
       ID.

5.3.6  For biphenyl analysis a 10 percent SP-2100  on  80/100  Supelcoport®,
       180 cm x 2 mm ID glass or stainless steel column may  be  used.

  5.4  Reagents.

5.4.1  Antimony pentachloride, SbCl5--reagent grade SbCl5 from  certain  sources
       has been found to contain DCB and bromonanochlorobiphenyl  (BNCB).  The
       DCB results in a positive blank value.  The BNCB interferes with  per-
       chlorination of the lower-substituted PCB isomers. The SbClg  selected
       for use must first be analyzed for these contaminants.  Reagent  grade
       SbCl5 manufactured by Cerac Pure9 (stock no. A1147, lot 15-26A)  has
       been  found  to be suitable.

5.4.2  Hexane—nanograde or pesticide-grade solvent,  distilled in glass.

5.4.3  Hydrochloric acid—reagent grade.

5.4.4  Sodium sulfate—reagent grade, anhydrous granular".  Heated to
       300°C prior to use in order to remove any  interfering organics.
       g.  Cerac Pure
           P.O. Box 1128
           Milwaukee, WI  53201
       h. Mallinckrodt, Inc.
          St.  Louis, MO  63147
                                        10

-------
5.4.5  Compressed nitrogen--water pumped and pre-purified to 99.99% pure.

5.4.6  Decachlorobiphenyl  (DCB)—pure DCB isomer for quantitation purposes may
       be obtained from Analabs, Inc.1 (Stock #RCS-051) or from RFR
       CorpJ (Stock ffRCP-28).

b.4.7  Chrornatographic grade silicic acid (10U mesh)—Woelm^ silica gel
       .activity grade I.

5.4.8  Aroclor standards and biphenyl—may be obtained from Analabs, Inc. or
       from EPA Pesticide Repository (MD-69), Research Triangle Park, North
       Carolina  27711.

  6.0  Preparation of Materials and Equipment

  6.1  Polyurethane foam absorbent media.

6.1.1  Circular plugs of diameter 8.9 cm are cut from 7.6 cm thick sheet poly-
       urethane foam.  These plugs will fit snuggly into the throat of the
       modified hi-vol sampler (Figure 2).  A cutting device as depicted in
       Figure 6 has been found suitable for cutting highly uniform polyurethane
       foam plugs from the stock material.  This device was manufactured by
       removing the internals and bottom of an automobile oil filter of the
       correct dimensions and buffing a cutting edge.  A bolt is attached as
       shown in Figure 6 to allow attachment to a drill chuck.  Using a reason-
       ably slow drill speed, this device is used to cut through the sheet
       polyurethane foam.   Lubricate the cutting edge constantly during cutting
       with a fine stream of water.
       i.  Analabs, Inc.                       j.  RFR Corporation
           80 Republic Drive                       1 Main Street
           New Haven,  CT   06473                   Hope, RI  02831
       K.  May be obtained from:  ICM Pharmaceuticals, Inc.
                          Life Science Group
                          26201 Miles Road
                          Cleveland, OH  44128
                                         11

-------
                                     i.
Figure 6.   Polyurethane Foam Cutting Device
                      12

-------
fa.1.2  To remove contaminating organics, extract each polyurethane foam plug
       for twelve hours with hexane in a soxhlet extraction apparatus
       (Section 5.2.1).

6.1.3  Dry the extracted polyurethane foam plugs with organic-free N^-  A
       glass cylinder (10 crn diameter) wrapped with heating tape arid fitted
       with a N^ inlet and exhaust as depicted in Figure 5 has worked well
       for the drying of many plugs at one time.

6.1.4  Avoid exposure of cleaned foam plug to laboratory atmosphere to
       prevent contamination with PCB.

6.1.5  Store each polyurethane foam plug in a wide-mouth, sol vent-rinsed
       glass container with pre-cleaned aluminum foil between jar and cap.
       The aluminum foil is heated to 200°C for at least one hour before use
       to remove organics.

  6.2  Glass Fiber Filters.

6.2.1  Gelman Type A (20 cm x 25 cm) glass fiber filters are heated to 200°C
       for several hours to remove contaminating organics.  They are then
       wrapped in pre-heated aluminum foil until use.  After use they are
       returned to the aluminum foil wrappers for transport to the
       laboratory.

  6.3  Hi-vol Sampling Device.

6.3.1  Standard hi-vol air samplers are modified by the addition of a 30 cm
       stainless steel extension to the throat section of the assembly as
       shown in Figures 2 and 4.  Air flow rate of this device can be varied
       from approximately 0.5 m^/min to 1.5 m^/rnin by use of a
       potentiometric speed controller.
                                        13

-------
6.3.2  A wire or screen retainer as shown in Figure 2, is placed between the
       polyurethane foam plugs in the throat of the apparatus and the motor
       unit to prevent movement of the plugs during sampling.

6.3.3  To avoid the possibility of contamination of interfering organic
       substances, the neoprene seal around the perimeter of the filter
       faceplate is removed.  A seal is maintained by wrapping the faceplate
       with Parafilm®.

6.3.4  An exhaust duct 5 meters or more in length, of collapsible ducting
       material is placed on the hi-vol exhaust and is positioned downwind
       during sampling to avoid recirculation of air through the sampler.

  7.0  Sample Collection

  7.1  A 20 cm x 25 cm glass fiber  filter (Gelman Type A) is mounted in the
       high-volume sampling apparatus.

  7.2  Three foam plugs (Section 6.1.1) are inserted into the 8.2 cm diameter
       throat of the high-volume sampling apparatus, taking care to avoid
       empty spaces between the plugs and cylinder wall.

  7.3  The hi-vol air flow rate is  adjusted to 0.8 + 0.2 m^/min using  the
       rotometer (see Figure 3).  Flow is measured at the beginning and end
       of each sampling period and  the volume of air  sampled is calculated  as
       follows:
                             Qn- + Qf
                         V - —	 x t
                                2

       where,     V  - air volume  sampled (uncorrected), m^
                  Q-j = initial air  flow rate, m^/min.
                  Qf = final air flow  rate, m3/min.
                  t  = sampling  time, min.

       Rotometer calibration procedures are described in (3).

-------
  7.4  A sampling period is selected which  is commensurate with the  expected
       airborne PCB level.   Under normal circumstances, a two-hour period  is
       appropriate for airborne levels of 2 ng/iu3 or greater.   If levels
       are expected to be lower (i.e., less than 2 ng/n:3), a  six-hour
       sampling period is recommended.  If  the sampling period  is too  long for
       higher airborne levels, PCB may pass through the collection media with
       consequently lower collection efficiency. y If the sampling period is not
       long enough for the ambient levels present, detectable quantities may
       not be collected.  Tests have demonstrated that all PCBs, including
       monochlorobiphenyl ,  are effectively  retained over a two-hour  sampling
       period and that dichlorobiphenyl is  effectively retained over a six-hour
       sampling period.  Aroclor 1242 is retained in excess of  12 hours at
       normal sampling rates.

  7.5  The sampling apparatus is rinsed between sample runs with hexane to
       remove any PCB contamination.

  8 . 0  Sample Extract Preparation

  8.1  After sampling is completed, the filters are removed,  and refluxed  for
       three hours with 300 ml hexane in a  soxhlet extractor.   The first and
       second polyurethane foam plugs, as located in the sampler with  respect
       to air flow, are extracted for three hours (18 cycles) in a soxhlet
       extractor (Section 5.2.1) with 200 nil hexane.  The extracts are combined
       and concentrated to 5 ml in a Kuderna-Canish apparatus (Section 5.2.13).

  8.2  Chrornatographic Separation of Chlorinated Pesticides.

8.2.1  A silica gel column is prepared by wet packing in hexane about 3 g  of
       the dry activated absorbent previously cleaned in benzene and dried to
       200°C for four hours.  Draw off excess hexane used to  make this slurry
       and wash the column with 50 ml of hexane.  A 250 mm x  13 mm column  is
       used (Section 5.2.12).
8.2.2  Place 2 g of anhydrous Na2S04 on top of the cilica  gel column to
       remove any water from the sample extract.
                                          15

-------
8.2.3  The hexane extract of 5 ml  is quantitatively transferred to the
       column, rinsing the Kuderna-Danish receiver three times with a small
       amount of hexane.

8.2.4  Allow the sample to drain through the column until the liquid level is
       about 0.5 cm above the top of the
8.2.5  Elute the PCB fraction with 50 ml of hexane, collecting all hexane in
       a clean, graduated Kuderna-Danish receiver (Section 5.2.13).  Most
       chlorinated hydrocarbon pesticides (except DDE) are retained on the
       column and may be eluted with benzene if analysis is desired.

  8.3  The eluate is concentrated to 5 ml in the Kuderna-Danish apparatus
       (Section 5.2.13).  A modified micro-Snyder column is then attached to
       the receiver (Section 5.2.11) and the solution is evaporated to less
       than 1.0 ml using the graduations on the evaporator receiver.  Care
       must be taken not to evaporate to dryness or the more volatile PCB
       species will be lost.

  8.4  Bring the volume up to the 1.0 ml graduation mark on the Kuderna-
       Danish evaporator receiver with  fresh hexane.  The extract  is now
       ready for either GC/EC analysis  and quantitation by Aroclor pattern
       matching or derivitization to DCB followed by GC/EC quantitation.

  8.5  If the sample is to be stored at  this point, the extract should be
       quantitatively transferred to a  Hypo- vial without a significant
       evaporative loss of solvent.  Close the Hypo-vial with  teflon septum
       and aluminum cap (Section 5.2.10).
                                        16

-------
9.0  Analysisand Quantitation of PCB
     Two techniques are available for the quantisation of PCB; the Aroclor
     pattern matching technique and perchlorination.   The analyst must
     decide which of these techniques to use based on the appearance of the
     chromatograms.  Airborne PCB may or may not exhibit a GC elution
     pattern that matches an Aroclor elution pattern.  Given a source of
     Aroclor, the relative volatilities of the various species will  vary
     widely resulting, generally, in a bias of the GC pattern toward the
     more volatile species as compared to an Aroclor standard prepared in
     solvent.(4)  This will  be particularly true of ambient airborne
     samples taken at a considerable distance from the source or sources,
     since differential rates of transport among the various species can
     exert a significant effect.  Most ambient air samples exhibit a large
     amount of apparent PCB  which is early eluting.  It is extremely diffi-
     cult to match these early eluting peaks to a standard Aroclor elution
     pattern.  These early eluting peaks may represent a highly significant
     amount of PCB.  Hence,  it is recommended in most cases that the per-
     chlorination method be  used for PCB quantitation and confirmation.  If
     it is of interest to know the PCB species present, the sample should
     be analyzed by comparison with Aroclor or specific isomer PCB stan-
     dards before perchlorination.  Biphenyl, if present, may be analyzed
     by GC/FID.   In those cases where the chromatogram closely matches an
     Aroclor pattern, the pattern matching technique is recommended.
     Experience has shown this is a suitable method for the higher
     ambient airborne levels observed near known sources of PCB.
                                      17

-------
10.0 Identification and Quantitation of PCB Mixture by Pattern Matching

10.1 Analyze the hexane extract on a 1.5/1.95% OV-17/QF-1 column.   The
     following conditions have been found to be suitable:
             Glass column, 180 cm length, 3 mm ID
             1.5%/1.95% OV-17/QF-1 liquid phase
             Chrom W-HP, 80/100 mesh support
             N2, 35 psig inlet pressure
             Detector temperature--220°C
             Column temperature--200°C
             Injector temperature--245°C
             Detector—electron capture

10.2 The analyst may qualitatively assess the species distribution of the
     PCB present by comparison of the GC elution pattern with that of an
     Aroclor or a series of various Aroclor standards, or, if desired, with
     individual PCB species.

10.3 Once the elution pattern is subjectively matched to a standard Aroclor
     pattern, quantitation is conducted  by integration of peak areas or
     comparison of peak height.  The methodology for quantitation by pat-
     tern matching is described by Chan  and Sampson (5) and by the U.S.
     Environmental Protection Agency (6)(7).  Care must be taken to insure
     that the sample chromatogram and the standard chromatogram are close
     to the same concentration (i.e., they may differ by no more than a
     factor of  2).  This  is necessary to assure  the EC detector is
     operating  in a linear range.

10.4 For purposes of confirmation of the PCB  present, an SE-30/OV-210
     column may be used with  the  following operating conditions:
             Glass column 180 cm  length, 2 mm ID
             4X/6X SE-30/OV-210 liquid  phase
             Chromosorb W-(HP), 80/100  mesh support
             N2, 38 psig  inlet pressure
             Detector  temperature--225°C
             Column temperature--200°C
             Injector  temperature--240°C
             Detector—electron capture
                                      18

-------
  11.0 Method for Quantitation of PCB  by Perchlorlnation

  11.1 Principal.  Perchlorination of  a  PCB  mixture  using antimony penta-
       chloride (SbC^)  as the active  reagent results  in conversion of
       all  PCB present to DCB which is a single isomeric, fully-substituted
       species.(8)  This species has a high  electron capture (EC)  detector
       response.   It can also be more  readily separated from other compounds
       by GC techniques  than can the complex PCB mixture.  Quantitation  can
       be accomplished by measurement of a single peak.

  11.2 Sample Preparation.  The sample extract (1.0  ml) prepared in the
       method described  above is quantitatively transferred to a Kuderna-
       Danish apparatus  with micro-Snyder column (Section 5.2.13).

  11.3 Removal of Solvent.  The extracted PCB must be  quantitatively
       exchanged  from the hexane solvent to  chloroform.  All residual  hexane
       must be removed from the extract prior to perch!orination.   Even  small
       amounts of residual hexane will result in the formation of a black
       solid residue upon the addition of SbC^.  This severely reduces
       PCB recovery.  The hexane is removed  by azeotrophic  evaporation from
       the hexane/chloroform mixture.

11.3.1 Add 3 ml of nanograde chloroform to the Kuderna-Danish receiver con-
       taining the sample extract in hexane  and concentrate, using a micro-
       Synder column, by slow boiling in a tube heater or water bath to  about
       U.2 ml.  Do not allow to evaporate to dryness.

11.3.2 Repeat step 11.3.1 three additional times in  order to remove all  resi-
       dual hexane.  Rinse tne micro-Snyder  apparatus  with a minimum amount
       of chloroform.  Final volume should be approximately 1.0ml.

11.3.3 Quantitatively transfer to a reaction vial (Section 5.2.3)  using  three
       chloroform rinses (Total rinse volume about 2 ml).

11.3.4 Add two micro-Hengar boiling chips and immerse reaction vial upright
       in a 70°C water bath to a depth of 6 + 2 cm.
                                        19

-------
11.3.5 Increase water bath temperature slowly until the solvent begins  to
       boil.  Boiling temperature should be 72-76°C.

11.3.6 Concentrate slowly to a volume of approximately 0.1 ml.  Under no cir-
       cumstances should the water bath temperature be permitted to exceed
       76°C or the solvent will be evaporated to dryness.  If either of these
       happen, PCB will be lost by volatilization and consequent recoveries
       will be low.  The final volume (0.1 ml) may be determined with suffi-
       cient accuracy by comparison of solvent level with another  reaction
       vial containing 0.1 ml of chloroform.

11.3.7 When a volume of 0.1 ml is achieved, cap the reaction vial  immediately
       and allow to cool.

  11.4 Sample Perch!orination.

11.4.1 To the concentrated sample extract  in the reaction vial add 0.2  ml  of
       SbCls (Section 5.4.1) and immediately re-seal the vial tightly
       with the Teflon-lined screw cap.

11.4.2 Place the reaction vial into a preheated (160 + 3°C) aluminum block
       heater for a period of 15 hours.

11.4.3 After the reaction period, remove the reaction vial from the aluminum
       block heater and allow to cool to room temperature.  Then cool to 0°C
       in an ice water bath.

11.4.4 Cautiously vent pressure from the vial in a  fume hood, directing away
       from the analyst.  Add 1 ml of 6 IN  HC1 to the cool reaction vial,
       replace the cap tightly and shake for 30 seconds.  The HC1  stops the
       perchlorination reaction.  CAUTION:  IF THE  REACTION VIAL  IS NOT COOL,
       THE ADDITION OF HC1 MAY CAUSE DANGEROUS SPLATTERING OF THE  REAGENTS
       FROM THE CONTAINER.

11.4.5 Add 1 ml hexane, shake vigorously for 30 seconds and carefully draw
       off the hexane layer with a disposable pipet.
                                        20

-------
11.4.6 Place this hexane extract on the top of a 6 mm x 12 cm disposable
       pipet packed with 2 g of anhydrous Na2$Q4.  ^nis co^umri is prewashed
       with hexane.

11.4.7 Repeat steps 11.4.5 and 11.4.6 five times to assure complete
       extraction of all DCB from the reaction vial.
11.4.8  Pass two 1 ml  portions of .fresh hexane through the Na2$04 column
        and collect all  fractions in a 10 ml  graduated Kuderna-Danish receiving
        vial (Section  5.2.11 ).

11.4.9  Connect a modified micro-Snyder column to the Kuderna-Danish apparatus
        (Section 5.2.11), add one Hengar boiling chip, and evaporate in a water
        bath (70°C)  to less than 0.5 ml.  Care must be taken to avoid bumping
        and loss of sample.

11.4.10 Cool the apparatus to room temperature and remove the micro-Snyder
        column.  Rinse the micro-Snyder column with sufficient fresh hexane to
        bring the volume up to 1.0 ml as indicated on the graduated Kuderna-
        Danish receiving vial.  Remove the micro-Snyder column and mix by gentle
        swirl ing.

11.4.11 Transfer the extract immediately (before significant solvent
        evaporation) to a properly-labeled Hypo-vial.

11.4.12 Close the Hypo-vial with a Teflon septum and aluminum cap (Section
        5.2.10).  Ml  PCB present in the sample has been converted to DCB in
        1.0 ml hexane.  The sample is now ready for GC analysis.

  11.5  Analysis

11.5.1  Analyze the DCB present in the perchlorinated extract by GC/EC.  The
        following column conditions have been found to be suitable:
               Glass column:  length 6 ft., ID 1/8"
               1.5% OV-17/1.95% QF-1 liquid phase
               Chromosorb W-HP, 80/100 mesh support
               N2> 35 psig inlet, approximately 40 ml/min
                                       21

-------
               Column temperature--220°C
               Injection port temperature—240°C
               Detector Temperature--200°C
        Under these conditions,  DCB will elute in approximately 20 minutes.
        Figure 6 is a typical  chromatogram.

11.5.2  A procedural  blank value must be determined with which to correct low
        level analytical results.  Select two PUF plugs and a glass fiber
        filter that have been cleaned-up and handled in the same manner as
        the materials used for air sampling.  Extract these PUF plugs and
        filter in the same manner as described above for the samples.  Per-
        chlorinate the extracts in the same manner as an air sample.  The
        resulting blank concentration, if significant, must then be used in
        the calculation of airborne PCB concentration.  It has been found
        that the lowest procedural blank value achievable is approximately 50
        ng of DCB after perchlorination.  This amount is readily detectable.
        The minimum detectable airborne concentration should, therefore, be
        considered to be in the range of 100 ng as DCB.

11.5.3  Quantitate as DCB by comparison of the peak area with that of a known
        concentration of pure DCB (Section 5.4.6) taking into consideration
        all concentration factors.  Care must be taken to assure the sample
        concentration and the standard concentration are near the same value
        so the EC detector is operating in the linear range.  Peak areas of
        the standard and the sample should not differ by more than a factor
        of two.
                               A - B
                 ng DCB/m3  =     S    x C
                                    V
        where  A = area of the DCB peak of the sample
               B = area of the DCB peak of the procedural blank (Section
                   11.5.2)
               S = area of DCB standard peak
               C = concentration of the DCB standard in ng
               V = volume of air sampled  in m3.

11.5.4  To convert DCB values to approximate equivalent PCB values, Table 1
        may be employed.
                                        22

-------
                                                                   O
                                                                   CXJ
                                                                       l/l
                                                                       O)
                                                                       E
                                                                   O
                                                                       CD
                                                                              s
                                                                              c
                                                                              c

                                                                              s
                                                                              c
                                                                              4-
                                                                              CC
                                                                              c:
                                                                              o_
                                                                              O
                                                                              E
                                                                              ra
                                                                              s_
                                                                              en
                                                                              o
                                                                              6
                                                                              o

                                                                              Q
                                                                             OJ
                                                                             i.
                                                              to   O
23

-------
Table 1.  Factors to Mathematically Convert Decachlorobiphenyl  to an
          Equivalent Amount of Aroclor (8)
Aroclor
1221
1232
1242
1016
1248
1254
1260
1262
DCB
Av. No. Cl*
1
2
3
3
4
5
6
7
10
MWt
188.5
223
257.5
257.5
292
326.4
361
395.3
499
X**
0.38
0.45
0.52
0.52
0.59
0.65
0.72
0.79
1.00
*  Average whole number of chlorines calculated from percent chlorine
   substitution for a specific Aroclor

t  Molecular weight of Aroclor based on the average whole number of
   chlorines calculated from percent chlorine substitution

** X = molecular wt Aroclor/molecular wt DCB (499).  To convert ppm
   DCB to ppm of a specific Aroclor, multiply ppm x DCB by X for the
   Aroclor.
                               24

-------
11.5.5  Confirmation may be accomplished by:  a) gas chromatographic/mass
        spectroscopy if sufficient DCB is present, or b) analysis on an
        alternate GC column as suggested in Section 10.4.
           *
  12.0  Analysis of Biphenyl

  12.1  Biphenyl is converted to DCB during the perchlorination step.  If a
        significant amount of biphenyl is present in a sample, the analyst
        may want to correct the results for this.  A separate GC analysis and
        quantitation using a flame ionization detector (GC/FID) is required.

  12.2  Biphenyl may be analyzed on an SP-2100 column using the following
        operating conditions:
               10% SP-2100 on 80/100 Supelcoport
               180 cm x 2 mm ID stainless steel or glass column
               Carrier gas:  N2 @ 27 ml/min
               Detector:  flame ionization, \\2 @ 22 psi
               Temperature program:  134°C for 8 min. then increase to 150°C
                                     at 4°C/min.  Elution temperature about
                                     146°C.  Elution time 10 min

  12.3  If it is desired that the perchlorination result (total DCB) reflect
        only  PCB in a sample  that contains biphenyl  as well as PCB, the
        analyst must correct the result as follows:
                                                          w
           Corrected ng DCB/m3 - calculated ng OC8/m3 -  Q 3(Jy(ng BIPH/rn3),

        where Y is the conversion efficiency of biphenyl (BIPH) to DCB in the
        perchlorination procedure.  Tests (N=1G) nave shown the value of Y is
        0.443 + O.o8b using the perchlorination technique described above.

  13.0  Quality Control

  13.1  The collection efficiency of the sampling system should be checked
        by independent extraction and analysis of the third PUF plug (the
        one farthest from the sampler intake).  Comparison of the PCB
        collected on this plug to the total on the other two plugs
        provides an estimate of the collection efficiency of the sampling
        system if the sampling guidelines described in Section 7.4 are
        fol1 owed.
                                       25

-------
13.2  Approximately 10 percent of the sample extracts are analyzed in
      duplicate.  Frequent material  and reagent blank values are
      determined.

13.3  Extraction technique may be evaluated by spiking and recovery of PCB
      from the polyurethane foam.

13.4  The procedural blank value must be frequently determined (Section
      11.5.2) and must be consistently low with respect to the PCB
      collected during sampling.

13.5  Standards are frequently injected as a check of the stability of the
      operating conditions, detector, and column.
                                      26

-------
                              REFERENCES

(1)  Bidleman, T.F.  and Olney, C.E.,  1974.   High-Volume Collection of
     Atmospheric Polychlorinated Biphenyls.   Bull.  Environ.  Contam.
     ana Tox., 11:5;  442-450.

(2)  Youden, W.J.  and Steiner, E.H.,  1975.   Statistical  Manual  of the
     Association of  Official  Analytical  Chemists.

(3)  Code of Federal  Regulations,  Title  40,  Part 50,  pp.  12-16,
     July 1, 1975.   Procedures for the Measurement  of Total  Suspended
     Airborne Particulate.

(4)  Mieure, J.P.; Hicks,  0.;  Kaley,  R.G.;  Saeger,  V.W.,  1975.   Char-
     acterization  of Polychlorinated  Biphenyls.   Presented at National
     Converence on Polychlorinated Biophenyls Nov.  19-21,  1975,
     Chicago, IL.

(5)  Chau, A.S.Y.  and Sampson, R.C.J., 1975.   Electron Capture  Gas
     Chromatographic  Methodology for  the Quantitation of Polychlorin-
     ated Biphenyls:  Survey and Compromise.   Environmental  Letters
     8:2;89-101.

(6)  U.S. Environmental  Protection Agency,  1974.  Method for Poly-
     chlorinated Biphenyls  (PCB's) in Industrial  Effluents.  Environ-
     mental  Monitoring and  Support Laboratory,  Cincinnati, OH.

(7)  U.S. Environmental  Protection Agency,  1974.  Manual  of  Analytical
     Methods for the  Analysis  of Pesticide  Residues in Human and
     Environmental Samples.  Environmental  Toxicology Division,
     Research Triangle Park,  NC 27711.

(8)  Armour, J.A., 1973.   Quantitative Perchlorination of Polychlori-
     nated Biphenyls  as a  Method for  Confirmatory Residue Measurement
     and Identification.   JAOAC, 56:   4; 987-993.
                                 '27

-------
                                     TECHNICAL INFORMATION CLEARANCE
  1. DATE PREPARED
  January 30,  1978
                                                I -GENERAL	
                                                      3. POLICY ISSUES
             for  Sampling  and Analysis of
 Polychlorinated Biphenyls (PCB's)  in Ambient
 Air.
                                                     DYES (Attach)   KX.
                          4. PUBLIC AFFAIRS
                             DYES    0
                                                   6. AUTHOR, ORGANI2,'   ." AND ADDRESi.
                                                  Charles  L.  Stratton, Stuart A. Whitlock  and
                                                  J.  Mark  Allan, Environmental  Sci.  & Enq.,  INC.
                                                  P.  0.  Box  13454,  Univ. Stat., Gainsville,  FA 32
                                       II - EPA TECHNICAL REPORT SERIES
                                               9 SERIES
                                                      10.
                                                                       TYPE OF REPOFIT
                                                         a. PROJECT (-sOne)
                                                                               b. SPECIAL
                DISTRIBUTION (Check blocki)
 30 ^PROJECT OFFICER
                               OPTIONAL (Specify keys)
                                                            IN-HOUSE
                                                                               c. OTHER (Specify)
                                                            EXTRAMURAL
                                                              PROJECT OFFICER/IN-HOUSE AUTHOR
 25)STORAGE'
                            NTIS (Justify)'"
    ADDITIONAL*
                               UNPUBLISHED ONLY (Justify)*
                                                                                           b. DAT,
  'Indicate number of copies
 "Attach complete lustification
                         COMMENT
                          EPA will  print the
                          f-eport
 ftnDtri?aArNgDe?unTtr.S.  EPA
OAB/EMSL, MD-77, RTP,  N.C.   27711
                                                                                               . TELEPHONE
                                                                                              541-2196
                                        III - OTHER TECHNICAL MATERIAL
 13. TYPE OF MATERIAL
I  I
Q
D
      JOURNAL PUBLICATION
      ORAL PRESENTATION
      PAPER FOR CONFERENCE
      PROCEEDINGS OR MEETING
                                D
                                                      14. NAME/LOCATION/DATE OF JOURNAL, CONFERENCE, ETC.
                                  OTHER /Specify!
 15 SPECIAL INSTRUCTIONS
                                                       3. SIGNATURE
                                                                       PRINCIPAL EPA AUTHOR
                                                       C. TYPED NAME AND ADDRESS
                                                                                              b. DATE
                                                                                              d. TELEPHONE
                                        IV\REVIEWS (Attach all comments)
 17 TECHNICAL   [XJYES
                                   IN/A
  a. TECHNICAL REVIEWERS NAME AND ADDRESS
                                                   18 EDITORIAL
                                                   DYES Q^IO.
                Il9. FORMAT
                [DvES DNO DN/A
20. POLICY
    LJYES  UNO
    Dr. R.  G.  Lewis,  ACB/ETD
    Dr. William .1. Mitchell,  QAB/EMSL/RTP, MD-77
                                                      21 COMMENT
                                                                TECHNICAL INFORMATION COORDINATOR
   Mr.  Thomas Hartledge, EMB/EMSL/RTP, MD-76
                                                         bWAME A'ND ADDRESS
                                                     Seymour Hochheiser,  E'1SL/RTD, MD
                                                                                      75
                                                                                              d. TELEPHONE
                                                                                               541-2106
                                               V - APPROVALS
                LABORATORY SIGNATURES
                                                                     HEADQUARTERS SIGNATURES
                 IMMEDIATE SUPERVISOR
                                                                       IMMEDIATE SUPERVISOR
 Dr.  J.B.  Clements,Chief.  QAB/EMSL/
                                         DATE
                                         ./n/78
                                                      SIGNATURE
                                     TELEPHONE
                                       541-2195
                                     ^TP 5/9/78
                                                      TYPED NAME AND ADDRESS
                                                                                              TELEPHONE
                          I DIRECTOR
                                                                         DIVISION DIRECTOR
 SIGNATURE
                                                      SIGNATURE
 TYPED NAME AND ADDRESS                  TELEPHONE

 Dr.  T. R.  Hauser,  DIRECTOR,  EMSL/RfTP 541-2106
                                                      TYPED NAME AND ADDRESS
                                                                                              TELEPHONE
                  LABORATORY DIRECTOR
                                                                  DEPUTY ASSISTANT ADMINISTRATOR
 SIGNA-TUR* '
                                                      SIGNATURE
EPA Form 5320-2 (11-761

-------
                                  TECHNICAL REPORT DATA
                           ffleait read Inscrucnoru on the revene be/ort complennfl
 REPORT NO.
  EPA-600/4-78-048
             3. RECIPIENT'S ACCESSION NO.
. TITLE AND SUBTITLE
A Method  for the Analysis of Polychlorinated  Biphenyl
(PCB)  in  Air
                                                          8. REPORT DATE
                                                            August 1978
             8. PERFORMING ORGANIZATION COOE
. AUTHOR(S)

Charles  L.  Stratton, Stuart A. Whitlock,  J.  Mark Allan
                                                          8. PERFORMING ORGANIZATION REPORT NO.
i. PERFORMING ORGANIZATION NAME AND ADDRESS

Environmental  Science and Engineering,  Inc.
P.  0.  Box 13454, University Station
Gainesville, Florida  32604
                                                           1O. PROGRAM ELEMENT NO.
                  1HD621
             11. CONTRACT/GRANT NO.
               68-01-2978
12. SPONSORING AGENCY NAME AND ADDRESS
Quality Assurance Branch, Environmental  Monitoring and
Support Laboratory, U.S. EPA, Research Triangle Park,
NC   27711,  and Office of Toxic Substances^  U.S. EPA,
Washington, D.C.  20460
             13. TYPE OF REPORT ANO PERIOD COVERED
               Final  8/75-12/77	
             14. SPONSORING AGENCY COOE
                  EPA-ORD
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
A method  has  been developed for the sampling  and  analysis  of polychlorinated biphen-
yls  (PCBs)  in  air.   An easily constructed, high-volume  sampling system is em employed
with porous polyurethane foam as the collection medium.  The sample is  collected  at
the  rate  of 0.6  to 1.0 m3 per minute.  Laboratory procedures described in this
document  permit  the quantitative analysis of  even the most volatile PCB species  in an
air  sample.   A perchlorination technique for  the  quantitative analysis of PCB has
been adapted  for use.   The technique is shown to  convert  even the most volatile  PCB
species to  decachlorobiphenyl for simple and  direct  quantitative analysis.  Data  is
presented to  show conversion efficiencies of  a variety  of  PCBs to decachlorobiphenyl
of 101 _+  6  percent over the range of 0.103 to 10.0 ug.  A  ruggedness test was con-
ducted which  indicates the proposed perchlorination  technique can yield reliable
interlaboratory  results.  The perchlorination technique is generally necessary for
the  analysis  of  low (i.e., less than 25 ng/m3)  airborne levels of PCB.  The ana-
lytical method is effective for the analysis  of airborne PCB levels within at least
the  range of  1 ng/m3 to 50 ug/m3.  The mean collection  efficiency for Aroclor
1016/1242 was  101 _+ 10 percent for sampling periods  of  20  minutes to 12 hours and  air
volumes of  16  to 720 cubic meters.  Field tests of the  method under a variety of
ambient conditions are described.
                               KEY WORDS ANO DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.lOENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Croup
 Polychlorinated biphenyl,  PCB,  high-volume
 sampling, environmental  contamination,
 environmental studies
 Methods Evaluation
 Methods Development
 Ambient Air
14B
13B
IB. DISTRIBUTION STATEMENT


 Unlimited   Release  to  public
19. SECURITY CLASS IT*ii Rtportf
   Unclassified
                                                                         21. NO. Of PAGES
20. SECURITY CLASS iThit pagt>
   Unclassified
                           22. PRICE
EPA Perm 2230-1 (»-7J|

-------