Performance of RCRA
    (Resource Conservation  and  Recovery Act)  Method
    8280 for the Analysis of Dibenzo-P-Dioxins and
    Dibenzofurans in Hazardous  Waste Samples
   Lockheed Engineering and Management
   Services Co.,  Inc., Las Vegas, NV
   Prepared  for        .     '•    -

   Environmental Monitoring Systems Lab,
   Las Vegas, NV
   May 86
U.S. Department of Commerce
Natitnai Technical Information Service

-------
                                                             PB86-' S
                                                        EPA/600/A-86/021
                                                        May 1986
PERFORMANCE OF RCRA METHOD 8280 FOR THE ANALYSIS OF DIBENZO-£-DIOXINS  AND
                 DIBENZOFURANS IN HAZARDOUS WASTE SAMPLES
                                    by

       J. M. Ballard, T. L. Vonnahme, N. 0. Nunn, and D. R. Youngman
    Lockheed Engineering and Management Services Company,  Incorporated
                         Las Vegas, Nevada 89114
                        Contract Number 68-03-3249
                             Project Officer

                             Stephen Billets
                        Quality Assurance Division
               Environmental Monitoring Systems Laboratory
                         Las Vegas, Nevada 89114
               ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
                    OFFICE OF RESEARCH AND DEVELOPMENT
                   U.S. ENVIRONMENTAL PROTECTION AGENCY
                         LAS VEGAS, NEVADA 89114
                                    BT
                               NATIONAL TECHNICAL
                              INFORMATION SERVICE
                                  U.S DtP/HtlMFNT OF CGMMtOCt
                                         . v» ?j|Ri

-------
                                   TECHNICAL REPORT DATA
                            (Pteate read Instructions on the reverse before completing)
 '. REPORT NO.

   EPA/600/4-86/021
2.
                              3. RECIPIENT'S ACCESSION NO.
 ..TITLE AND SUBTITLE

 PERFORMANCE OF RCRA METHOD 8280  FOR THE ANALYSIS OF
 DIBENZO-.P-DIOXINS AND DIBENZOFURANS IN HAZARDOUS WASTE
 SAMPLES
                              6. REPORT D*TE
                                 May 1986
                              6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)

 J. M.  Ballard, T. L. Vonnahme,  N.  J.  Nunn, and
   D. R.  Youngman
                                                            8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                            10. PROGRAM ELEMENT NO.
 Lockheed  Engineering and Management  Services Company,
   Incorporated
 P.O. Box  15027
 Las Vegas.  NV   89114	
                              11. CONTRACT/GRANT NO.     «••  ...

                                  Contract  Number 68-03-3249
 12. SPONSORING AGENCY NAME AND ADDRESS
 Environmental  Monitoring Systems Laboratory - LV, NV
 Office of Research  and Development
 U.S. Environmental  Protection Agency
 Las Vegas. NV   89114         	
                              13. TYPE OF REPORT AND PERIOD COVERED
                              14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
 Project Officer  -  Stephen Billets, Environmental  Monitoring Systems  Laboratory,
 	Las Vegas, NV  89114	
16. ABSTRACT
      Further evaluation  of RCRA Method 8280 for  the analysis of chlorinated  dibenzo-
£-dioxins and dibenzofurans has been performed.   The Method has been modified  to
enable the quantitation  of total tetra- through  octa-chlorinated dioxins  and
dibenzofurans and has  been applied to six different sample matrices derived  from
industrial polychlorophenol  sources and also to  fly-ash, still-bottom, and Missouri
soil  samples.  An interlaboratory validation of  the Method has been conducted  in
two phases:  Phase  I required the analysis of spiked and unspiked clay and sludge
samples -for certain specified analytes, and Phase  II required the analysis of  10
samples of soil, sludge, fly-ash, and furans.  Method detection limits of ^C1?-
labeled polychlorinated dioxins and dibenzofurans  in seven matrices have  been
determined.  In order  to propose the most effective procedure, a comparison  was
made  of the Contract Laboratory Program carbon column cleanup (without backflush)
with  the backflush procedure  used in the proposed  RCRA Method.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                b.lDENTIFIERS/OPEN ENDED TERMS  C. COS AT I Field/Group
18. DISTRIBUTION STATEMENT
 RELEASE TO PUBLIC
                                              19. SECURITY CLASS (This Repor11
                                               UNCLASSIFIED
                                           21. NO. OF PAGES
                                                 96
                                                    RITY CLASS (Thispagef
                20, SECURITY CI.AS
                 UNCLASSIFIED
                                                                         22. PRICE
tf A F«r»2220-1 (R»». 4-77)   PMCVIOU* COITION tt OBSOLETE

-------
                                     NOTICE
     The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency under Contract Number 68-03-3249
to the Lockheed Engineering and Management Services Company, Incorporated,
Las Vegas, Nevada.  It has been subjected to peer and administrative review of
the Agency, and it has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement
or recommendation for use.
                                       n

-------
                                     FOREWORD
      On  a  molecular  basis,  2,3,7,8-tetrachlorodibenzo-p_-dioxin (2,3,7,8-TCDD)
 is  one of  the most poisonous  synthetic chemicals kncwn.1  The compound has been
 shown in animals  to  possess teratogenic, embryotoxic, carcinogenic, and
 co-carcinogenic properties  in  addition to acute toxicity.  Because of its
 chemical stability,  lipophilic character, and extreme toxicity, it presents
 potentially  severe health hazards to the human population.  Although 2,3,7,8-
 TCDD  is  the  most  toxic of the  75 chlorinated dibenzo-p_-dioxins (PCDD's), many
 of  the others (and also of  the 135 chlorinated dibenzofurans [PCDF's] which
 have  similar genesis, structures, and properties) are known to possess rela-
 tively high  toxicity to humans and animals.  For this reason, the entire spec-
 trum  of  PCDD's and PCDF's is of environmental concern.  2,3,7.8-TCDD was first
 synthesized  in 1872,2  and  only sporadic reports of the preparation of PCDD's,
 containing two, four, or eight chlorine atoms, appeared in the years 1941-1965.
 Particular interest in 2,3,7,8-TCDD, and in the PCDD's and PCDF's in general,
 increased markedly with the discovery in the early 1970's of the same terato-
 genic and toxic effects with certain commonly used herbicides, e.g., 2,4,5-
 trichlorophenoxyacetic acid (2,4,5-T), as were observed with 2,3,7,8-TCDD.
 Analysis of  116 samples of 11 different pesticides produced during the period
 1950-1970 revealed the presence of PCDD contamination (tetra- through octa-
 chlorinated) in 42 percent of the samples.3  Consideration of the chemistry of
 pesticide manufacture indicated that PCDD's could be formed in competing side-
 reactions of the polychlorophenol precursors.  The domestic use of 2,4,5-T was
 subsequently banned,  and the military use of Agent Orange (1:1 mixture of
 2,4,5-T and 2,4-dichlorophenoxyacetic acid) as a defoliant in Vietnam was
discontinued, both in the early 1970's.  Because of the widespread usage of
 pesticides potentially contaminated with PCDD's, a Dioxin Monitoring Program
was set up by the EPA in 1973 to develop an analytical method capable of
detecting 2,3,7,8-TCDD in environmental samples at the part per trillion (ppt)
 level.  This effort fonr.ed the basis of the National  Dioxin Strategy of  the
Agency.

     Although the most ubiquitous routes of non-occupational  exposure of the
general  population to dioxins have probably been via  the use  of contaminated
 pesticides and from the emissions of municipal  waste  incinerators, the most
 concentrated waste sources of 2,3,7,8-TCDD  are the tars and sludges resulting
from the commercial  preparation of 2,4,5-trichlorophenol  (2,4,5-TCP).  This
 latter fact was highlighted during an investigation in 1975-1977  of unexplained
 animal deaths at various horse arenas in Missouri.4  It was discovered that the
 sites had been sprayed with a mixture of waste oil  and distillation residues
 from the manufacture  of 2,4,5-TCP which were contaminated with 2,3,7,8-TCDD.
 Subsequent investigation of chemical  waste  dump-sites  in  New  York  State  (Hyde
 Park; Love Canal), where wastes from the manufacture  of 2,4,5-TCP  had  been
buried,  alsc revealed the presence of substantial  amounts of  2,3,7,8-TCDD.

                                      iii

-------
     Based on this experience, it was concluded by the EPA that samples con-
taining tetra-, panta-, and hexa-CDD's and CDF's are likely to exhibit in-
creased toxicity,- and a method tc analyze hazardous wastes for the relevant
PCDD's and PCDF's was included in the Resource Conservation and Recovery Act
(RCRA) requirements for hazardous waste monitoring as published in the Federal
Register.o  A single-laboratory evaluation of the RCRA Method 8280 for the
analysis of PCDD's and PCDr's in hazardous waste has been the subject of a
previous report prepared for the Office of Solid Waste.7  That report presented
results obtained with sample matrices including pottery clay, a Missouri soil,
a fly-ash, a still-bottom from a chlorophenol-based herbicide production pro-
cess and an industrial process sludge.  Major revisions to the Method as first
published in 1983 were necessary to accommodate the analysis of complex samples
such as sludge and still-bottom.

     The revised Method 8280 has subsequently undergone a period of continual
development, and this report presents results obtained during the further evolu-
tion of the Method.   New documentation in this report includes Method perfor-
mance data on complex samples from polychlorophenol  use processes, results from
an inter!aboratory study of the revised Method, and method detection limits of
selected PCDD's and  PCDF's in a variety of matrices.
                                       iv

-------
                                    ABSTRACT
     Further evaluation of RCRA Method 8280 for the analysis of chlorinated
dibenzo-p_-dioxins and dibenzofurans has been performed.  The Method has been
modified to enable the quantisation of total tetra- through octa-chlorinated
dioxins and dibenzofurans and has been applied to six different sample
matrices derived from industrial polychloropheno"! sources and alsc to fly-ash,
still-bottom, and Missouri soil  samples.  An inter!aboratory validation of the
Method has been conducted in two phases:  Phase I required the analysis of
spiked and unspiked clay and sludge samples for certain specified analytes, and
Phase II required the analysis of 10 samples of soil, sludge, fly-ash, and
still-bottom for total tetra- through octa-chlorinated dioxins and dibenzo-
furans.   Method detection limits of 13C^2-labe1ed polychlorinated dioxins and
dibenzofurans in seven matrices  have been determined.  In order to propose the
most effective procedure, a comparison was made of the Contract Laboratory
Program carbon column cleanup (without backflush) with the backflush procedure
used in the proposed RCRA Method.

-------
                                     CONTENTS
                                                                            Page

Foreword	     iii
Abstract 	       v
Tables	    vii.i
Abbreviations	       x
Acknowledgment 	      xi

     1.   Introduction 	       1

     2.   Conclusions	       3

     3.   Recommendations	       5

     4.   Comparison of CLP Carbon Column Cleanup with Backflusn Procedure
            Adapted from RCRA Method 8280	       7

     5.   Analysis of Wastes from Industrial Use of Chlorophenols.  ...      11
               Motes on Physical Characteristics and Extraction/Cleanup
                 of Samples	      14

     6.   Method Detection Limit Study  	      17

     7.   Inter-laboratory Test of RCRA Method 8280	      28
               Phase I	      28
               Comments from Laboratories on Phase I of  Interlaboratory
                 Study	      34
               Phase II	      40
               Statistical  Analysis of Data from Phase II of  Inter-
                 laboratory Study	      50

References	 .      53

Appendices

     A.  Detection and Procedure for the Determination of the
           Method Detection Limit	      54

     B.,  RCRA Method 8280 with Revisions Based on Multi-Laboratory
           Testing:  Method of Analysis for Chlorinated Dibenzo-J^-
           Dioxins and Dibenzofurans 	       58
                                      vn

-------
                                     TABLES

Number                                                                     Page

 4-1      Comparison of Carbon Column Cleanup Methods	,       8

 4-2      Percent Recovery of PCDD's and PCDF:s  from CLP Carbon Column  ,       9

 5-1      Analysis of PCP  Process  Samples Using  Method  8280	      12

 5-2      Analysis of PCP  Process  Sample (B-5) and  10 TCP Process
            Samples Using  Method 328G	      13

 6-1      Method  Detection Limits  of 13C12-Labeled  PCDD's and  PCDF's
            in  Reagent  Water (ppt)  and  Environmental  Samples  (ppb)  ...      19

 6-2      Percent  Recovery of 13C12-Labeled PCDD's  and  PCDF's  from
            Reagent Water	      21

 6-3      Percent  Recovery of 13C12-Labeled PCDD's  and  PCDF's  from
            Missouri  Soil	,      22

 6-4      Percent  Recovery of 1?C12-Labeled PCDD's  and  PCDF's  from
            Fly-Ash	      23

 6-5      Percent  Recovery of 13C12-Labeled PCUD's  and  PCDF's  from
            Industrial  Sludge	:      24

 6-6      Percent  Recovery of 13C12-Labeled PCDD's  and  PCDF's  from
            Still-Bottom 	      25

 6-7      Percent  Recovery of 13Cio-Labeled PCDD's  and  PCDF's  from
            Fuel  Oil	t	      26

 6-8      Percent  Recovery of 13C12-Labeled PCDD's  and  PCDF's  from
            Fuel  Oil/Sawdust	      27

 7-1      Inter!aborat.ory  Test  of  Method  8280, Phase  I:   Summary
            of  Analytes Reported by  Participating Laboratories  	      30
          Interlaboratory Test of Method 8280, Phase I:  Percent
           Recovery of  Internal Standard I3C1?-2,3,7,8-TCDD . .
7-2
                                                                            31
7-3       Interlaboratory Test of Method 8280, Phase I:  Quantitation
           of Analytes  in Spiked Clay Samples (ppb) 	     32
                                     VTM

-------
                               TABLES (Continued)

Number                                                                     Page

 7-4      Interlaboratory Test of Method 8280,  Phase I:   Quantitation
            of Analytes in Spiked Sludge Samples  (ppb)  	      33

 7-3      Interlaboratory Test of Method 8280,  Phase I:   Accuracy
            and Bias of Results	      34

 7-6      Interlaboratory Test of Method 8280,  Phase II:   Percent
            Recovery of Internal  Standard 13C12-2,3,7,8-TCDD  	      42

 7-7      Interlaboratory Test of Method 8280,  Phase II:   Percent
            Recovery of Internal  Standard 13C12-OCDD 	      43

 7-8      Interlaboratory Test of Method 8280,  Phase II:   Quantitation
            of Tutal  Dioxins  and  Dibenzofurans  in  Fly-Ash (ppb)	      44

 7-9      Interlaboratory Test of Method 8280,  Phase II:   Quantitation
            of Total  Dioxins  and  Dibenzofurans  in  Soil A  (ppb)  	      45

 7-10      Interlaboratory Test of Method 8280,  Phase II:   Quantitation
            of Total  Dioxins  and  Dibenzofurans  in  Soil 3  (ppb)  	      46

 7-11      Interlaboratory Test of Method 8280,  Phase II:   Quantitation
            of Total  Dioxins  and  Dibenzofurans  in  Sludge  A (ppb) ....      47

 7-12      Interlaboratory Test of Method 8280,  Phase II:   Quantitation
            of Total  Dioxins  and  Dibenzofurans  in  Sludge  B (ppb) ....      48

 7-13      Interlaboratory Test of Method 8280,  Phase II:   Quantitation
            of Total  Dioxins  and  Dibenzofurans  in  Still-Bottom  (ppb) .  .      49

 7-14      Statistical Test A:   Analysis  of Variance  of Recovery of
            Internal  Standard-;	      51

 7-15      Statistical Test B:   Laboratory Equivalency (Two-way Analysis
            of Variance on Means)	      52

 7-16      Statistical Test C:   Laboratory Equivalency (Cochran's Test
            for Equal Variances)  	      52

-------
                                  ABBREVIATIONS
CLP
DFTPP
EICP
GC
GC/MS
HpCDD
HpCDF
HxCDD
HxCDF
K-D
HDL
MID
ND
OCDD
OCDF
PCDD
PCDE
PCDF
PCP
PeCDD
PeCDF
ppb
PPt
RCRA
RSD
SD
2,4,5-T
TCDD
TCDF
TCP
Contract Laboratory Program
decaf1uorotri phenylphosphi ne
extracte  ion current profile
gas chromatofjraphy
gas chromatography/mass spectrometry
hyptachl orodi benzo-p_-di oxi n
heptachlorodibenzofuran
hexachlorodibenzc-£-dioxin
hexachlorodibenzofuran
Kuderna-Danish
method detection limit
multiple ion detection
not detected
octachlorodi benzo-£-di oxi n
octachlorodibenzofuran
polychlorinated dibenzo-£-dioxin
polychlorinated diphenyl ether
polychlorinated dibenzofuran
pentachlorophenol
pentachlorodi benzo-£-di oxi n
pentachlorodi benzofuran
parts per billion
parts per trillion
Resource Conservation and Recovery Act
relative standard deviation
standard deviation
2,4,5-trichlorophenoxyacetic acid
tetrachl orodi benzo-£-di oxi n
tetrachlorodibenzofuran
trichlorophenol

-------
                                 ACKNOWLEDGMENT
     The authors wish to acknowledge the  valuable  contributions made by the
following individuals:

     -  Dr.  Fred Shore,  Radian  Corporation  (formerly with LEMSCO),  for his
        technical  advice in the development  of the procedures needed to deter-
        mine method detection limits.

        Dr.  Gate Jenkins,  U.S.  EPA,  for providing  the complex samples and per-
        ceptive inquiries  which challenged  both the Method and the  laboratory
        staff.

     -  Ms.  Maka Grogard,  Viar  Corporation,  who coordinated the interlaboratory
        study through the  Contract Laboratory Program.

     -  Mr.  Douglas Gillard, U.S. EPA, who  helped  define the scope  of the
        interlaboratory  study.

     -  All  of  the participants in the interlaboratory study whose  comments and
        data were  used to  refine the method.

     -  Mr.  Gary Robertson,  Mr.  Jeff Wolff,  and Mr. Forest Garner,  LEMSCO, who
        audited the interlaboratory  study (Phase II) data and performed statis-
        tical analyses;  and  for providing the narrative comparison  of the
        carbon  column cleanup used in the CLP method and Method 8280.
                                      XI

-------
                                   SECTION 1

                                  INTRODUCTION


     RCRA Method 8280 for the analysis of chlorinated dibenzo-p-dioxins and
dibenzofurans, as published in thv. Federal Register in April  1983,6 revealed
the need for several modifications to allow for the determination of the target
analytes in complex matrices, such as industrial sludge and still-bottom
samples.  Details of the modifications made and of the subsequent application
of the revised Method to a limited number of samples analyzed in the course of
a single-laboratory evaluation have been reported.7

     Subsequently, the Method has been further refined in several important
areas, as was needed to meet the needs of the characterization and assessment
aspects of RCRA.  A summary of these changes is as follows:  In order to im-
prove the accuracy of quantitation of t.he hepta- and octa-CDD's and CDF's,  a
second internal  standard (13C12-OCDD) is edded together with  ll3C12-2,3,7,8-TCDD
prior to sample workup.  Some of the ions specified in the multiple ion detec-
tion (MID) descriptors have been changed so as to increase sensitivity by
monitoring the most intense ion in the isotopic cluster.  To  ensure that co-
eluting polychlorinated diphenyl ethers (PCDE's) are not contributing to the
signal  response due to PCDF's, the molecular ion of the appropriate PCDE was
included in each MID descriptor.  In addition, the criteria for the positive
identification of PCDD and PCDF isomers were made more explicit.  Instrument
tune criteria employing perfluorotri-jvbutylamine (FC-43) were substituted  for
those based on the use of decafluorotriphenylphosphine (DFTPP).  The section
on the calculation of concentrations of analytes was expanded to include a
procedure for measuring unknown PCDD and PCDF isomers.

     This report presents data on the performance of the Method as it was
applied to the analysis of a variety of wastes derived from the use of pcly-
chlorophenols in the wood-preserving industry.  As an additional test of Method
performance, an interlaboratory validation study was conducted in two parts.  A
two-part study was used because the Method had been extensively revised since
its publication in the Federal Register, and it was felt that participating
laboratories would be unfamiliar with some of the proposed procedures.   The
first phase was intended to allow the participants to acquire familiarization
with the Method by analyzing relatively simple matrices for a few specified
analytes which had been spiked into the samples.  The second  phase required the
total  quantitation of tetra- through octa-CDD's and CDF's in  complex samples
containing the analytes at both low and extremely high levels;  no spiking was
used for these samples.  A method detection limit study using all  available
l^Cio-labeled PCDD and PCDF isomers spiked into seven different sample  matrices
has also been performed.  A comparison of the EPA Contract Laboratory Program
(CLP)  carbon column cleanup without and with  a backflush  elution procedure  was

                                       1

-------
conducted to test the adequacy of the CLP method  for the  determination  of  total
PCDD's and PCDF's.

-------
                                   SECTION 2

                                  CONCLUSIONS


     The single-laboratory application of the Method to the determination of
PCDD's and PCDF's in complex environmental samples (e.g., fly-ash, still-
bottom, and wastes from the industrial use of penta- and tri-chlorophenol) has
routinely yielded high recoveries (60 to 85 percent) of the spiked internal
standard, *3C^2-2,3,7,8-TCDD.  jhis indicates that the extraction and cleanup
procedures are able to accommodate samples ranging from those with a high
aqueous content to viscous oils and chemical  sludges,  it can be assumed that
endogenous PCDD's and PCDF's are extracted with equal success if matrix effects
are not operating.

     In the absence of a full range of standard reference materials, the accu-
racy of the Method is rather difficult to assess.   However, data obtained from
Phase I of the interlaboratory study indicate that the Method is biased high
and that the bias appears to decrease as the concentrations of the analytes
increase.  Data from the method detection limit (MDL) study can be used as an
indicator of intralaboratory precision.  For seven replicate determinations of
a TCDF and a PeCDD in fly-ash with each at a measured concentration of 2.6
times their final calculated MDL's, the relative standard deviations (RSD's)
were 12.3 percent and 12.2 percent, respectively.   Similar determinations for a
PeCDF and a TCDD which were measured at a level 6.0 and 4.4 times their MDL's
gave RSD's of 5.2 percent and 7.2 percent, respectively.

     Encouraging results were obtained from Phase  I of the interlaboratory
study in which specific analytes spiked into clay  and sludge samples were
quantitated.

     The good overall recovery (greater than 50 percent) of the internal stand-
ard and the small differences between the spiked concentrations and the mean
measured values both indicate that the Method can  provide acceptable data in a
rnulti-laboratory program.  Phase II of the interlaboratory study which required
the quantitation of total tetra- through octa-CDD's and CDF's in 10 aliquots of
4 sample types also provided generally satisfactory results.   The internal
standards (13Cp-2,3,7,8-TCDD and 13C12-OCDD) were recovered in overall  accept-
able yields ranging from 51 to 82 percent.  However, quantitation of the
analytes was less precise than in Phase I.  Two major, probable reasons for
this are as follows:

          the complex samples themselves, some of  which contained endogenous
          amounts of the target analytes at low and at extremely high levels.
          This required a large dilution effect which minimized the value of
          the internal  standard, and

-------
          the analysis required the identification, confirmation,  and
          quantisation of an unknown number of peaks for each coroner often
          without an authentic reference material  which could be used to
          confirm the identification.

     Statistical analysis of the Phase II data revealed that:

          the recovery of the 13C12-2,3,7,8-TCDD internal  standard was a func-
          tion of sample type whereas thV: of the  13C^2-OCDD internal  standard
          was not;

          the laboratories were equivalent in accuracy for all  analytes except
          OCDD; and

          the laboratories were equivalent in precision for 31  of  the 40 pos-
          rible matrix/analyte combinations.

     The comparison of the Contract Laboratory Program (CLP) carbon column
cleanup (backflush procedure is not used) with the backflush procedure used  in
Method 8280 indicated that although the CLP cleanup as written  is  very satis-
factory for the determination of 2,3,7,8-TCDD (and possibly other  tetra- and
penta-CDD's and CDF's), it is not adequate for the determination of hexa-,
hepta-, and octa-CDD's and CDF's.  However, the combination of  open carbon
column with a backflush procedure gave acceptable  performance for  the tetra-
through octa-substituted congeners.

     Method detection limits of eight 13C12-labeled PCDD's and  PCDF's spiked
into reagent water were found to be in the low ppt range (luss  than 10 ppt); 42
of 48 values determined for 6 environmental samples were less than 5 ppb.

-------
                                   SECTION 3

                                RECOMMENDATIONS
     As a result of the experience gained during the single-  and  multi-
laboratory testing of the Method with a variety of environmental  samples,
several modifications to the Method and areas of further study are  recommended:

     1.   The Method should allow for the use of disposable,  open carbon
          columns as an option to the currently specified HPLC carbon  column
          cleanup.  This would allow for an increase in the  rate  of sample
          throughput and would also reduce solvent consumption.

     2.   The use of stacked acidic/basic silica gel  columns  instead of mul-
          tiple liquid-liquid partitioning in the extraction/cleanup proce-
          dures should be investigated.  This would eliminate the problems of
          emulsion formation currently encountered and would  also greatly
          reduce the quantities of corrosive wastes generated.

     3.   Gas chromatography (GC) conditions should be modified to  improve the
          resolution between the internal standard (l3C1?-2,3,7,8-TCDD) and the
          recovery standard (i<3C12-l,2,3,4-TCDD).  If tin's cannot be readily
          achieved, then use of an alternative recovery standard  should be
          considered.

     4.   The elution windows (defined by first and last eluting  isomers)  of
          the tetra- through octa-CDD and CDF congeners should be established
          for the GC conditions used in the Method.

     5.   Because of the known elution overlap of some tetra-substituted iso-
          mers with penta-substituted isomers (and other potential  overlaps
          between homologous groups), the multiple ion detection  (MID) descrip-
          tors should be modified to include at least one ion for each overlap-
          ping homologue.

     6.   Method 8280 should be written to require as many GC/MS  analyses  as
          necessary by using the appropriate MID descriptors  whenever an
          elution overlap is noted in a sample.

     7.   Kovats Indices should be determined for available PCDD's  and PCDF's.
          This would aid laboratories in the identification of  isomers not
          known or available to them and would be useful  in a GC  screening
          program.

-------
8.   The need to monitor for polychlorinated diphenyl  ethers (PCDE's)  in
     the final sample extract should be investigated.

9.   A source of a well-defined GC performance standard should be identi-
     fied.  Column performance guidelines should be established for a
     variety of columns.

10.  Sample reanalysis requirements given the presence of low and of very
     high levels of target analytes should be defired.

-------
                                   SECTION 4

             COMPARISON OF CLP CARBON COLUMN CLEANUP WITH BACKFLUSH
                    PROCEDURE ADAPTED FROM RCRA METHOD 8280


     The major difference between the Contract Laboratory Program (CLP)  Low
Resolution Dioxin Method and RCRA Method 8280 is one of objective.   For the
CLP Method, it is the determination of 2,3,7,8-TCDD only whereas for RCRA
Method 8280 it is the determination of total concentrations of all  tetra-
through octa-chlorinated dibenzo-£-dioxins and dibenzofurans.  This has led to
technical differences during the development of the Methods, and a  side-by-side
comparison of the carbon column cleanup employed by the two Methods is shown in
Table 4-1.  It can be seen that, once the column has been prepared, the CLP
method should be faster and consume much less solvent.  In addition, the CLP
method does not require HPLC equipment.  It was therefore of interest to deter-
mine whether the CLP carbon column cleanup, as written, was adequate for analy-
sis for PCDD's and PCDF's.

     The major operating difference between the two carbon column cleanup
methods is the use of a toluene backflush in Method 8280 to elute the analytes
of interest.  To evaluate the applicability of the CLP procedure for the analy-
sis of higher PCDD's and PCDF's, a gravity-feed carbon column using a cut-down
disposable serological  pipet (20 cm x 5 mm i.d.) was prepared and was pre-
eluted as directed.  The column was spiked with 25 \ii of a standard solution
containing 11 PCDD's and PCDF's each at 10 ng/yL and was eluted using the
specified eluents.  After collection of the toluene fraction (2 mL) which was
held separately, the column was eluted with an additional  5 mL of toluene.  The
toluene fractions were concentrated separately to 500 pL,  and each  was analyzed
by GC/ECD; the analytes were quantitated against an internal standard, endrin
ketone.  A second column was prepared, spiked, and eluted  as described above
except that it was reversed before collection of the two toluene fractions
(2 mL and 5 mL).

     The percentage recovery of each PCDD and PCDF in each fraction for  the
column elution with and without the backflush procedure is shown in Table 4-2.
The data reveal  that, for the particular isomers used, the CLP method as
written without  backflush procedure gives acceptable recoveries only for the
tetra- and penta-CDD's  and CDF's.   Recovery of the hexa- and hepta-CDD's and
CDF's is low while the  OCDD and OCDF are not recovered at  all.   The additional
5 mL toluene fraction elutes significant amounts of the higher substituted
congeners although the  total  recovery of the OCDD and OCDF still  remains low.
Results obtained using  the backflush procedure show that the 2  mL toluene
fraction eluted  very much greater  amounts of the hexa-,  hepta-,  and  octa-
isomers than did the procedure  without the backflush.   The total  recovery for

-------
            TABLE 4-1.  COMPARISON OF CARBON COLUMN CLEANUP METHODS

          CLP Method                                    Method 8280
1.  Column Preparation:

    Mix 3.6 g Carbopack C with 16.4 g
    of Celite 545 in a 40 mL vial and
    activate by heating at 130°C for
    6 hours.  Check each new batch to
    ensure TCDD recovery >^ 50 percent;
    use the low level  concentration
    standard.

    Insert a glass wool plug in a 15 cm
    by 7 mm O.D. disposable pipet.
    Apply suction to the pointed end of
    the pipet and add the prepared
    mixture to form a 2 cm column.
2.  Pre-e'lute the column with:

    2 mL toluene
    1 mL 75/20/5 (v/v) methylene
         chloride/methanol/benzene
    1 mL 50/50 (v/v) cyclohexane/
         methylene chloride
    2 mL hexane -

3.  While the column is still wet with
    hexane, add the sample  extract and
    elute with the following sequence:

    2 mL hexane
    1 mL 50/50 (v/v) cyclohexane/
         methylene chloride
    1 mL 75/20/5 (v/v) methylene
         chloride/methanol/benzene

    Discard the eluates.  Elute with
    2 mL toluene and collect the eluent,
    store until  GC/MS analysis.
 !.  Column Preparation:

    A 10 ium x 7 cm silanized glass
    HPLC column is prepared by
    mixing 5 percent (w/w) AX-21
    active carbon with 10 pm silica
    (Spherisorb S10W).  Stir and
    sieve through a 40 urn screen.
2.
4.
    Evaporate the 60 percent methy-
    lene chloride/hexane fraction
    to 400 yL and prepare for trans-
    fer to a 1 mL HPLC loop injector.
    Rinse the tube with 500 yL hexane
    and add both fractions to the
    HPLC injector loop.

    Elute the column with 30 mL
    cyclohexane/methylene chloride
    1:1 (v/v) at 2 mL/min.  Discard
    the effluent.  Next elute with
    10 mL 70/20/5 (v/v) methylene
    chloride/methanol/benzene.  Dis-
    card the effluent.  Backflush
    the column with 40 mL toluene
    collecting the effluent which
    contains the PCDD's and PCDF's.

    Clean the column with 30 mL
    methanol followed by 40 mL tolu-
    ene in the backflush position.
    Return to normal  position and
    equilibrate by pumping through
    30 mL 1:1 (v/v)  cyclohexane/
    methylene chloride.

    Check for bleed  after high
    'love!  (>500 ppb)  samples and
       1 ace as needed.
                                                Evapc Mte  the toluene fraction
                                                on  a  Mfry  evaporator  at  50°C.
                                                Transfer to  a 2.0 mL reacti-
                                                vial  using a toluene rinse and
                                                concentrate  to the desired
                                                volume using nitrogen.

-------
              TABLE 4-2.   PERCENT RECOVERY9 OF PCDD'S  AND  PCDF'S  FROM CLP CARBON  COLUMN
Method as Written (without Backflush)
Analyte
2,3,7,8-TCDF
1,2,3,4-TCDD
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
1,2,3,4,7-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3, 4,6, 7,8-HpCDF
1,2,3,4,6,7,8-HpCDD
1, 2,3,4,6, 7,8,9-OCDD
1, 2,3,4,6, 7,8,9-OCDF
2 mL
Toluene
Fraction
81.5
80.0
87.6
71.0
80.9
35.9
39.6
7.8
13.4
NO
ND
Additional
5 ml.
Toluene
Fraction
ND
ND
ND
14.0
ND
43.0
46.0
52.6
62.0
50.8
36.0
Total
81.5
80.0
87.6
85.0
80.9
78.9
85.6
60.4
75.4
50.8
36.0
Method Modified (with Backflush)
2 mL
Toluene
Fraction
83.0
80.3
87.6
85.4
87.5
74.5
80.3
54.4
57.8
48.2
45.7
Additional
5 mL
Toluene
Fraction
ND
ND
ND
ND
ND
9.8
9.8
25.5
22.6
25.8
26.9
Total
83.0
80.3
87.6
85.4
87.5
84.3
90.1
79.9
80,4
74.0
72.6
aResults  of  a  single  determination.
ND = Not  detected.

-------
each isomer was also equal to or greater than that obtained without  the  back-
flush procedure.  It is evident that while the CLP carbon column  cleanup is
satisfactory for analysis for 2,3,7,8-TCDD (and possibly other tetra-  find
penta-CDD's and CDF's), it is not adequate for the determination  of  hexa-,
hepta-, and octa-CDD's and CDF's.  However, use of an open column technique
which does not require liquid chromatography equipment should  improve  the rate
of sample throughput by enabling many samples to be processed  concurrently.
For this reason and because of the acceptable recoveries demonstrated  using  a
backflush procedure with the gravity-feed carbon column, it is recommended that
the open column cleanup described below should be included as  an  option  to the
HPLC carbon column cleanup specified in Method 8280.

                           OPEN CARBON COLUMN CLEANUP

     Prepare a standard 5-mL disposable pipet by cutting off 1 cm from the
     tip.   Insert a glasswool  plug at the 2.5 mL mark and pack with  300  mg
     of charcoal/silica gel.  Cap the packing with a  glasswool plug.   Pre-
     rinse the column with 5 mL hexane in the forward direction of flow  and
     then in the reverse direction of flow.  While still in the reverse  direc-
     tion of flow, pre-elute the column with 2 mL of  toluene,  1 mL of  methylene
     chloride/methanol/benzene (75:20:5, v/v), 1 mL of cyclohexane/methylene
     chloride (50:50, v/v), and 2 mL of hexane.  Discard all  column  rinsates.
     Still  in the reverse direction of flow, transfer the sample  concentrate
     to the column and elute with 1 mL of hexane, 1 mL of hexane, 1  mL of
     cyclohexane/methylene chloride (50:50, v/v), and 1 mL of  methylene
     chloride/methanol/benzene (75:20:5, v/v).  Discard all  of the above
     eluates.   Now turn the column over and in the direction of forward  flow
     elute PCDD's and PCDF's with 15 mL of toluene.  Proceed with the  next step
     of the analysis using this toluene fraction.

     NOTE:   The charcoal/silica gel  packing is prepared by thoroughly  mixing
            Carbopack C (3.6g)  and silica gel  (16.4g)  followed by activation
            at 130°C. for 6h.
                                      10

-------
                                   SECTION 5

            ANALYSIS OF WASTES FROM INDUSTRIAL USE OF CHLOROPHENOLS


     Eleven waste samples derived from the industrial use of pentachlorophenol
 (PCP) and 10 samples derived from the industrial use of trichlorophenol  (TCP)
were provided by the EPA.  The 21 samples together represent 6 different matrix
types, viz, sludge, fuel oil, alcohol fuel oil, soil, water, and sawdust.  The
range of matrix types encompassed is expected to be representative of those to
be analyzed under RCRA regulations and is expected to provide varying degrees
of sample complexity.

     Each sample was analyzed in duplicate for the quantitation of total tetra-
through octa-CDD's and CDF's.  Two criteria were applied to confirm that peaks
in the extracted ion current profiles (EICP's) of the quantitation ions  were
due to the targeted analytes and were not due to either interferences or spuri-
ous noise signals.  These were the following:

          signal-to-noise ratio greater than 3 to 1

          the presence of the confirmation ion (listed in Table 3 of the
          Method) such that the relative intensity of the quantitation ion and
          the confirmation ion was within the limits specified in the Method,
          Table 4 (see Appendix B).

     Signal  responses that did not meet these criteria are reported as "ND"
(not detected).  Quantitation was usually performed against 13C12-2,3,7,8-TCDD
as the internal standard, and values are corrected for the recovery using this
compound as an isotopic diluent.   However, due to the extremely high levels of
hexa-, hepta-, or octa-CDD's/CDF's present in some samples, these analytes were
quantitated against 1 C12-l ,2,3,4-TCDD or 13C12-OCDD added to the extracts
after dilution.  It is a disadvantage of the quantitation method that multiple
GC/MS analyses are therefore required for samples containing both low and high
level  analytes.  To spike the sample with the appropriate l^C-standard would
have caused an unnecessarily large expense.   Whenever possible,  sample re-run
requirements should be imposed to lessen the used for quantitation by anything
other than the isotopic diluent.   Several  characteristics are evident in the
data presented in Tables 5-1 and  5-2.   First is tl.e total  absence of detectable
levels of TCDD in all  of the 21  samples and  the occurrence of PeCDD in only 3
samples.   Second are the very high levels  of hepta- and  octa-CDD present in 10
of 11  PCP process samples;  the presence of only very low levels  of analytes in
the remaining PCP process sample  (B-5)  was not unexpected  in view of the known
low solubility of PCDD's/PCDF's  in water.   The proposed  detection limit
achieved for wastewater (0.01 ppb) is  a reflection both  of the absence of
interferences in the sample and  of the subsequent facile extraction and  cleanup

                                       11

-------
                       TABLE 5-1.  ANALYSIS3 OF PCP PROCESS SAMPLES USING METHOD 8280
ro


PCDD/
PCDF
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDF
13r
C12-
2,3,7
TCDH
cent

Sludge
B-6d
(PPb)
NDb
ND
2150
51520°
72300°
ND
ND
68
343
4100°
66.8
,8-
00.-
recovery
Fuel
oil
B-7b
(PPb)
ND
ND
2186
67176°
154000°
ND
154
2933
1342
7500°
69.0



==========
Sludge
B-8b
(PPb)
ND
ND
ND
2166°
2670°
ND
ND
ND
ND
ND
64.3




Sludge
B-12h
(PPb)
ND
ND
ND
978°
2550°
ND
ND
ND
ND
76
67.8




Fuel oil
A-2g
(PPb)
MD
ND
2079
38195°
59100°
ND
246
2852
1913
447
69.2



Alcohol
fuel oil
A-3g
(PPb)
ND
ND
762
17956°
24500°
NO
ND
76
1118
741
60.0




Sludge
A-4g
(PPb)
ND
ND
726
59600°
10*000°
ND
ND
1568
1948
3200°
62.9




Soil
A-5g
(PPb)
ND
ND
283
12945°
Ii500°
ND
ND
65
533
900°
77.0




Soil
A-6.1g
(PPb)
ND
27
730
24700°
26300°
ND
61
252
1695
3080°
75.4




Soil
A-6.2g
(PPb)
ND
ND
396
12300°
15000°
ND
ND
56
434
1690°
74.8



     aMean of duplicates; concentrations shown are for the total of all isomers within a given
       homologous series.
     bND  is below the detection limit for the sample matrix.  Detection limits are estimated as 5 ppb
       for the tetra- through hexa-isomers, and 10 ppb for the hepta- and octa-isomers.
     °Due to the extremely high levels of HpCDD, OCDD, and OCDF detected in the GC/MS analysis, the
       extracts were diluted after normal qi'antitation of the tetra-, penta-, hexa-CDD/CDF and
       hepta-CDF.  HpCDD, OCDD, and OCDF were then quantitated versus 13C12-1,2,3,4-TCDD added after
       dilution; the values are corrected for 13Cl2-2,3,7,8-TCDD recovery.

-------
 TABLE  5-2.   ANALYSIS3  OF  PCP  PROCESS  SAMPLE  (B-5)  AND 10 TCP PROCESS SAMPLES USING METHOD 8280
PCDD/
PCDF
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDF
13r
Water
B-5
(PPb)
NDb
ND
0.072
2.5C
1.25°
0.024
NU
0.017
0.136
0.029

Sawdust
H-3
(PPb)
NO
385
2680d
2314C
1250C
3598d
1908d
11903d
1374d
94C

Soil
H-7a
(opb)
NO
ND
ND
ND
5.5
ND
ND
ND
ND
ND

Soil
H-7b
(PPb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND

Soil
H-7c
(PPb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND

Water
1-1
(ppb)
ND
ND
110d
1677C
345C
3.9d
ND
233d
108C
16°
£
Sludge
1-2
(PPb)
ND
30
2410d
42134°
14658°
201
429
5496d
2768°
239°

Sludge
I-ll
(PPb)
ND
ND
399
4404°
4080°
68
23
626
622°
151°

Soil
I-12c
(PPb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND

Soil
I-14a
(PPb)
ND
ND
ND
37
20
ND
ND
ND
ND
ND

Soil
I-14b
(ppb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND

        67.3
2,3,7,8-
TCDD
percent
recovery
                 95. 2e
71.2  76.4   72.3
             77.1
75.6
84.9
78.7
84.1
  °Mean of duplicates, except for samples B-5 and 1-1 which are the result, of single deter-
    minations.   Concentrations shown are for the total of all isomers within a given homologous
    series.
  bND is below the detection limit for the sample matrix.  Detection limits are 5 ppb for soil,
    sawdust, sludge, and 0.01 ppb for water.
c>dDue to very high levels of some hexa-, hepta-, or octa-CCD/CDF isomers, some samples were
    diluted and the PCDD's and PCDF's noted ••1D"Q «••=>««••!*<«'« ««*.<-,,c 13r __nrnr\c nr
    13C12-l,2,3,4-TCDDd; the values are cor
  eSome interference at the quantitation ion was noted.
  Across interference at the quantitation ion was noted.
were quantified versus   C12
ected for 13C12-2,3,7,e-TCDD
       -OCDDC or
        recovery.

-------
procedures.  That the extraction and cleanup procedures work well  for the
matrix types examined is demonstrated by the generally high recovery (60 to 85
percent) of the internal standard, 13C12-2,3,7,8-TCDD.  Interference at the
quantisation ion of the internal standard was noted for only two samples (H-3,
sawdust; 1-1, water) which were both TCP process samples.

NOTES ON PHYSICAL CHARACTERISTICS AND EXTRACTION/CLEANUP OF SAMPLES

Sample B-5      Clear, aqueous sample (wastewater from cooling tower).   A
                yellow aqueous phase resulted from the base wash.   A red-brown
                aqueous phase resulted from the first acid wash, and a  color-
                less phase resulted from the second acid wash.  No problems
                were encountered with emulsions.

Sample H-3      Sawdust (spray-booth filter debris).  The base wash gave a
                yellow aqueous layer.  The first acid wash gave a  dark  brown
                aqueous layer, and the second acid wash gave a colorless layer.
                The emulsion required 15 to 20 minutes to clear after the first
                acid wash.

Sample H-7a     Soil (soil core from treated-wood storage yard).  The acid wash
                was performed before the base wash.  Three acid washes  were
                required to progress from a dark brown to a colorless aqueous
                layer.  The base wash gave a cloudy aqueous layer, but  no
                emulsion.  The subsequent water wash gave a bad emulsion which
                required CJK 15 minutes to clear.

Sample H-7b     Soil (saturated soil  core with some water on top).  Three acid
                washes were required to progress from a dark brown to a color-
                less aqueous layer.  No problems were encountered  with
                emulsions.

Sample H-7c     Soil (soil core sample was very wet mud, saturated with water).
                Two acid washes were required to progress from a dark yellow
                to a colorless aqueous layer.  No emulsion problems were
                encountered.

Sample 1-1      Red, oily-lookiny water (aqueous dip tank solution).  Performed
                four extractions with methylene chloride because of very bad
                emulsion formation.

                A small  amount of red sediment was observed in the K-D  concen-
                trate after performing the solvent exchange into hexane.   The
                normal  order of base/acid washes was therefore reversed.   Two
                acid washes were required to  progress from a dark  brown to a
                colorless aqueous layer which v/as cloudy.

                The base wash resulted in yellow scum on the aqueous layer.   No
                emulsion problems were encountered.

                Concentration of the  organic  extract after the acid/hase  wash
                sequence and again after the  alumina column cleanup gave  a

                                       14

-------
                            dirty film on the sides of t!:e glass container.  This film
                            could not be re-dissolved itt hexane but was washed onto the
                            carbon column by methylene chloride/hexane (60:40 v/v).
            Sample 1-2      Oily sawdust (dip tank sludge).
                            Two acid washes were required to progress from a yellow-orange
                            to a colorless aqueous layer.  No problems were encountered
                            with emulsions.
            Sample 1-11     Saturated sludge with 50 percent standing water (from catch
                            basin sump).  The extract was extremely dirty after K-D concen-
                            tration.  The normal order of base/acid washes was therefore
             \               reversed.  Four acid washes were required to progress from a
                            black to a yellow aqueous layer.
                            The water wash immediately following the post-acid wash gave a
                            large amount of emulsion.  The base wash gave an emulsion which
                            immediately cleared.
            Sample I-12c    Very wet soil  (core from treated-wood storage yard).
                            One acid wash only was required.
                            No problems were encountered with emulsions.
            Sample I-14a    Wet clay (soil  core from treated-wood storage yard).
                            One acid wash only was required.
                            No problems were encountered with emulsions.
            Sample I-14b    Wet lumpy clay (soil core from treated-wooc! storage yard).
                            On shaking with petroleum ether/methanol, the sample dispersed
                            readily but did not appear to "turn over" well.
                            One acid wash only was required.
                            No problems were encountered with emulsions.
                            Some residue adhered to sides of the conical  concentrator tube
                            immediately prior to loading the carbon column.
            Sample B-6d     Sludge (very thick - difficult to obtain a homogeneous sample).
                            The toluene solution from the Dean-Stark separator contained a
                            large amount of finely divided solid; filtration through
                            Whatman No. 54 paper was very slow.   The acid wash procedure
                            gave a yellow aqueous phase even after four washes.
            Sample B-7b     Sludge (high oil  content).
i                                                   15

-------
                The aqueous layer was still  very yellow after four acid washes.
Samples B-8b    Sludge (high water content).
        and
        B-12h   The samples were cleaned-up readily due to low oil content.
                Only three acid washes were required to obtain a colorless
                aqueous phase.
Sample A-2g     Oil (no water content).
                The drying procedure using the Dean-Stark separator was omitted
                because of the absence of water.  The aqueous layer was still
                yellow after three acid washes.
Sample A-3g     Alcohol fuel oil.
                The sample was very viscous; the toluene extract filtered very
                slowly.  A large amount of emulsion formed during the first
                acid wash.
Sample A-4g     Sludge (high oil content).
                The drying procedure using the Dean-Stark separator was omitted
                because of the absence of water.  The acid wash was performed
                before the base wash.
Sample A-5g     Soil (contained some oil).
                The base wash gave a strong yellow aqueous layer.
Samples A-6.1g  Soil.
        and
        A-6.2g  A finely divided solid adhered to the K-D apparatus after
                concentration of the hexane solution.
                                       16

-------
                                   SECTION 6

                          METHOD DETtCT!ON LIMIT STUDY


     For  purposes of this study, the method detection limit (MDL) is defined as
the minimum concentration of a subr.tance that can be identified, measured, and
reported  with 99 percent confidence that the analyte concentration is greater
than zero and is determined from analysis of a given matrix containing the
target analyte.  The experimental design for this study (Appendix A) was that
reported  by Glaser et al.,8 and was applied using all available 13C12-labeled
analytes  spiked into seven different sample types at a concentration of twice
the estimated MDL of each analyte.  This experimental design was used in order
to obtain MDL values in each matrix of interest without spiking the matrix with
unlabeled PCDD's and PCDF's and without changing the integrity of the matrix.
In order  to establish an appropriate spiking level, the MDL was estimated as
that concentration at which the response of the appropriate quantitation ion
gave a signal/noise ratio of 3 to 1.  Statistical considerations required that
a minimum of seven replicates of each sample type should be processed through
the entire analytical method.  Two initial replicates were tested to verify the
reasonableness of the MDL estimate for each sample type.  When a reasonable
spiking level had been achieved, five more determinations were made at the same
spiked concentration.  The standard deviation (S) of the mean concentration
determined for each analyte was calculated from the seven replicate measure-
ments for each of the seven matrix types.  The MDL was then calculated from the
equation:
                             =  t(n_i, i_a = 0.99)  x (S)

where t(n_i  !_„ = o.99) ^s the Student's t value appropriate for a 99 percent
confidence level and'a standard deviation estimate  with n-1 degrees of freedom.

Therefore,

                               MDL  =  3.143(S).

The concentration (Cs) of each analyte in each sample was  determined  with
reference to the "0^-1.2, 3, 4-TCDD internal  standard, which was  spiked  after
sample workup to give a final concentration of 40 pg/uL» using  the  following
equation:

                                        AS * CIS
                                            x  RF
                                       17

-------
 where

      As    =  response (area)  of  analyte

      AIS   =  response (area)  of  13C12-1,2,3,4-TCDD

      CIS   =  concentration  of 13C12-1,2,3,4-TCDD  (40  pg/uL)

      RF    =  response factor  of  the  analyte.

 The  response  factor  (RF) used for each analyte was the mean of five determina-
 tions  calculated  from a  linear,  5-point calibration curve  (constructed by the
 use  of standards)  according to the equation:
The use of  13Ci2-l,2,3,4-TCDD as an internal standard rather than  13C12-2,3,7,8-
TCDD allowed trie determination of the MDL of the latter analyte.   The samples
included  in this study were selected to provide a representative range of
matrices  of environmental interest, and they are described below.

     Reagent water:  distilled, deionized laboratory water.

     Missouri soil:  soil blended to form a homogeneous sample.

     Fly-ash:  alkaline ash recovered from the electrostatic precipitator of a
          coal -burning power plant.

     Industrial sludge:  sludge from cooling tower which received  creosotic
          and pentachlorophenolic wastewaters.  Sample was ca. 70  percent
          water mixed with oil and sludge.

     Still-bottom:  distillation bottoms (tar) from 2,4-dichlorophenol
          production.

     Fuel  oil:  wood-preservative solution from the modified Thermal Process
          tanks.  Sample was an oily liquid (>90 percent oil) containing
          no water.

     Fuel  oil /Sawdust:   sawdust was obtained as a very fine powder from the
          local  lumber yard.  Fuel  oil  (described above) was mixed at the
          4 percent (w/w) level .

     The eight 13C12-ldbeled PCDD's and PCDF's used in this study and their
MDL's in the seven sample matrices  are listed in Table 6-1.  Several charac-
teristics  and trends are apparent in the data:   1<3C12-2,3,7,8-TCDD/TCDF  usually
had the lowest MDL values for each  sample type while 13C12-HpCDD/OCDD usually
had the highest; as might be expected,  the MDL values  for all  analytes general-
ly increased in passing from the  "clean" sample types  (reagent  water, fly-ash)
to the more complex, organics-containing matrices  (still -bottom,  industrial

                                       18

-------
                                               13r
         TABLE  6-1.   METHOD DETECTION LIMITS OF   C12-LABELED PCDD'S AND PCDF'S  IN REAGENT
                           WATER (PPT) AND ENVIRONMENTAL SAMPLES  (PPB)
13C12-Labeled
Analyte
2,3,7,8-TCDD
1,2,3, 7,8-PeCDD
1,2,3,6,7,8-HxCDU
1,2,3,4, 6, 7,8-HpCDD
OCDU
2,3,7,8-TCDF
1,2,3,7 ,8-PeCDF
1, 2,3,4, 7,8-HxCDF
Reagent
Water3
0.44
2.35
6.63
5.45
7.37
0.58
1.50
2.53
Missouri
Soilb
0.13
0.70
1.24
1.60
1.35
0.11
0.33
0.83
Ashb
0.07
0.25
0.55
1.41
2.27
0.06
0.06
0.30
Industrial
SI udgec
0.40
1.47
2.26
3.39
7.68
0.36
0.58
1.15
Still-
Bottomd
1.81
2.46
16.2
4.59
10.1
2.26
1.61
2.27
Fuel
Oild
0.75
2.09
5.02
8.14
23.2
0.48
0.80
2.09
Fuel Oil/
Sawdustb
0.13
0.18
0.25
0.49
1.34
0.04
0.09
0.17
aSample size 1,000 mL
^Sample size 10 g
cSample size 2 g
^Sample size 1 g
Note:  The final sample-extract volume was  100 yL for all  samples.

-------
sludge).  The MDL for 13Ci2-2,3,7,8-TCDD i" reagent water (0.44 ppt)  determined
in this study using Method 8280 compares well  with the value reported^ for
2,3,7,8-TCDD in reagent water (2 ppt) which was determined using Method 613
(capillary column GC/MS with selected ion monitoring).  The MDL procedure,
involving seven replicate determinations of each of the eight analytes in
each of seven sample matrices, also generated  other data (percent recovery,
precision) of interest in assessing the performance of Method 8280.   These data
are presented in Tables 6-2 through 6-8.  It can be seen that good recoveries
were obtained and that the precision at such low spike levels was acceptable.
                                       20

-------
TABLE 6-2.   PERCENT RECOVERY  OF  13C12-LABELED PCDD'S AND PCDF'S FROM REAGENT WATER
13Ci2-Labeled
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDD
~ 1,2,3,4,6,7,8-HpCOD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentration (pg/yL)
in Final Extract
Spiked
20
40
60
80
100
10
20
40
pleasured
23.2
45.8
52.7
77.8
91.3
7.70
18.8
32.4
RSD Percent
6.0
16.4
40.0
22.3
25.7
24.0
25.4
24.8
Mean Recovery
Percent
116.0
114.5
87.8
97.3
91.3
77.0
94.0
81.0

-------
13
TABLE 6-3.  PERCENT RECOVERY OF 1JC12-LABELED PCDD'S AND PCDF'S FROM MISSOURI SOIL
13C12-Labeled
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentrati
in Final
Spiked
40
80
120
160
200
20
40
80
on (pg/nL)
Extract
Measured
37.3
65.3
72.1
88.2
84.7
15.7
34.3
58.8

RSD Percent
11.0
33.9
54.9
57.8
50.5
22.5
30.1
44.9
Mean Recovery
Percent
93.3
81.6
60.1
55.1
42.4
78.5
85.8
73.5

-------
               TABLE  6-4.   PERCENT  RECOVERY  07  13C12-LABELED PCDD'S AND PCDF'S FROM FLY ASH
rv>
oo
13C12-Labeled
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDD
1,2,3,4,6,7,8-HpCOD
OCDO
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentration (pg/uL)
in Final Extract
Spiked
40
80
120
150
250
20
40
80
Measured
31.0
64.9
100.5
169.1
235.8
14.2
33.2
57.8
RSD Percent
7.2
12.2
17.4
26.5
30.6
12.3
5.2
16.3
Mean Recovery
Percent
77.5
81.1
83.8
112.7
94.3
71.0
83.0
72.3

-------
                                      13
       TABLE 6-5.   PERCENT RECOVERY OF ^C^-LABELED PCDD'S AND PCDF'S FROM INDUSTRIAL SLUDGE
13Ci2-Labeled
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDDa
1,2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentrati
in Final
Spiked
40
80
80
120
175
20
40
80
on (pg/pL)
Extract
Measured
25.3
56.7
90.6
128.6
302.5
15.6
26.4
49.5

RSD Percent
10.2
16.5
15.8
16.7
16.2
14.6
14.1
14.8
Mean Recovery
Percent
63.3
70.9
113.3
107.2
172.9
78.0
66.0
61.9
apeak shape was distorted by very high  level  of  interferent.

-------
                                            13
ro
en
            TABLE 6-6.  PERCENT RECOVERY OF ^C^-LABELED PCDD'S AND PCDF'S FROM STILL-BOTTOM
13C12-Labeled
Analyte
2,3,7,8-TCDOa
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDDa
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF&
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentrati
in Final
Spiked
40
80
80
120
175
20
40
80
on (pg/uL)
Extract
Measured
57.3
58.1
109.0
97.7
197.6
14.5
24.5
64.8

RSD Percent
10.1
13.5
47.2
15.0
16.3
49.7
20.9
11.2
Mean Recovery
Percent
143.3
72.6
136.3
81.4
112.9
72.5
61.3
81.0
   aLow  level interferent was observed; peak shape was not distorted.
   bPeak  shape was distorted by very high level of interferent.

-------
             TABLE 6-7.  PERCENT RECOVERY OF  13C12-LABELED  PCDD'S  AND PCOF'S FROM FUEL-OIL
ro
o\
13C12-Labeled
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentration (pg/pL)
in Final Extract
Spiked
40
80
120
150
250
20
40
80
Measured
29.7
56.8
110.7
118.3
232.8
12.9
27,7
57.2
RSD Percent
8.0
11.7
14.4
21.9
31.7
11.9
9.2
11.7
Mean Recovery
Percent
74.3
71.0
92.3
78.9
93.1
64.5
69.3
71.5

-------
TABLE 6-8.   PERCENT RECOVERY OF  13C12-LABELED PCDD'S  AND PCDF'S  FROM FUEL  OIL/SAKDUST
13C12-Labeled
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,6,7,8-HxCDD
1,2,3,4,6,7,8-HpCDD
~ OCOD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
Concentration (pg/uL)
in Final Extract
Spiked
40
80
120
150
250
20
40
00
Measured
32.4
57.0
92.2
120.6
241.6
13.1
28.0
52.7
RSD Percent
12.7
10.1
8.5
12.8
17.6
10.6
10.4
10.4
Mean Recovery
Percent
81.0
71.3
76.8
80.4
96. fi
65.5
70.0
65.9

-------
                                   SECTION 7

                    INTERLABORATORY TEST OF RCRA METHOD 8280
PHASE I
     Phase I was intended to allow participating laboratories to acquire famil-
iarization with the requirements of the revised RCRA Method 8280.  A familiar-
ization phase of the study was considered necessary since extensive revisions
had been made to the original Method 8280.  These included (1) changes in the
procedure for the extraction of analytes from the sample, (2) modification of
the open column alumina chromatography cleanup, (3) deletion of the HPLC clean-
up, (4) addition of a carbon column cltanup, and (5) incorporation of internal
and recovery standards into the Method.

     Five laboratories were selected by the Contract Laboratory Program to
participate in the study and were each provided witn samples, analytical stan-
dards, isotopically labeled internal and recovery standards, and the revised
Method.  Detailed supplemental  instructions which included guidance on the
sample size, the volume of the final extract, typical  MID descriptors, typical
MID and RIC chromatograms, and the reporting of data and deliverables, were
also provided.  Each laboratory was provided with the following six samples
which had bsen prepared by personnel of Lockheed-EMSCO who also analyzed a set
of the six samples.

     SAMPLE                                                TYPE

Clay Blank                           Pottery clay

Clay Spike No. 1                     Pottery clay,  spiked with selected PCDD's
                                     and PCDF's

Clay Spike No. 2                     Pottery clay,  spiked with selected PCDD's
                                     and PCDF's

Sludge Blank                         Sludge from industrial  PCP process

Sludge Spike No. 1                   Sludge Blank,  spiked with selected PCDD's
                                     and PCDF's

Sludge Spike No. ?.                   Sludge Blank,  spiked with selected PCDD's
                                     and PCDF's.

     The results obtained by Lockheed-EMSCO personnel  famiMar with the Method,


                                       28

-------
 although  not  obtained  as  part of the blind study, are valuable for comparison
 and  are included  in the relevant Tables.

      A very simple, qualitative measure of how well the combined extraction,
 chromatographic cleanup,  and GC/MS analysis prescribed in the Method deals with
 the  various analytes may  be obtained by noting the number of the specified
 analytes  for  which values were reported by each laboratory.  As shown in Table
 7-1,  at least one laboratory in addition to LEMSCO detected all of the ana-
 lytes, and two laboratories reported on 32 of 38.  Relatively lower reporting
 by Laboratories II and IV (total:  49 of 76) is presumed to be due to lack of
 familiarity with the extraction and cleanup procedures and may be expected to
 improve with  experience.

      A more quantitative measure of the extraction and cleanup efficiency is
 provided  by monitoring the percent recovery of the internal standard (  C12-
 2,3,7,8-TCDD) which was added to the sample immediately prior to extraction.
 Table 7-2 shows that three of the participating laboratories, apart from
 LEMSCO, obtained acceptable results (mean recovery greater than 40 percent).
 The two other laboratories reported mean recoveries of less than 30 percent.

      Individual  results for the 13 analytes spiked into Clay Spike No. 1 and
 No. 2 and for the 6 analytes spiked into Sludge Spike No. 1 and No. 2 are shown
 in Tables 7-3 and 7-4, respectively.

     The mean value for 114 determinations of 11 analytes spiked into clay at
fie 5 ppb level  was 6.02 ± 2.78 ppb.

     The mean value for 16 determinations of 2 analytes spiked into clay at
the 2.5 ppb level  was 3.56 ± 2.35 ppb.

     The mean value for 57 determinations of 6 analytes spiked into sludge at
the 125 ppb level  wa: 126.4 ± 57.9 ppb.

     The accuracy and bias of these determinations of analytes spiked into clay
at 2.5 ppb and 5.0 ppb and spiked into  sludge at 125  ppb are calculated  as
follows:
                           (Mean Measured Value\
                           	—	  x 100%
                               True Value     /
                  Bias  =  (Accuracy - 100)  %


                          I  SD of Mean Measured Value          1    \
Standard Deviation (SD)  = [	  x  	]x  100%
of Bias Estimate          \(Number of Determinations)!/^    True  Value/

and are shown in Table 7-5.


                                       29

-------
         TABLE 7-1.  INTERLABORATORY TEST OF METHOD 8280, PHASE I:
         SUMMARY OF ANALYTES REPORTED BY PARTICIPATING LABORATORIES


Sample
Clay Spike
No. 1
Clay Spike
No. 2
Sludge Spike
No. 1
Sludge Spike
Number
of
Analytes
Spiked
13

13

6

6
Number of Analytes Reported
by Each Laboratory

LEMSCO
13

13

6

6

I
13

13

6

6

II
8

8

3

3

III
12

12

4

4

IV
11

4

6

6

V
12

12

4

4
No. 2
                                     30

-------
TABLE 7-2.  INTERLABORATORY TEST OF METHOD 8280, PHASE I:  PERCENT RECOVERY
                  OF INTERNAL STANDARD 13C12-2,3,7,8-TCDD
Participating Laboratories
Sample
Clay Blank
Clay Spike
No. 1
Clay Spike
No. 2
Sludge Blank
Sludge Spike
No. 1
Sludge Spike
No. 2
LEMSCO
66.3
62.8
78.9
88.4
66.1
74.0
I
34.3
64.3
132
84.5
79.0
96.4
II
9.1
13.3
10.2
28.7
28.4
32.9
III
54
53
60
57
12
15
IV
21
26
14
34
25
36
V
67
63
58
35
32
44
Mean
42.0
47.1
58.9
54.6
40.4
49.7
RSD
Percent
57.7
46.7
76.4
48.6
64.6
60.2
                                     31

-------
           TABLE  7-3.   INTERLABORATORY TEST OF METHOD 8280, PHASE I
             QUANTITATION OF ANALYTES IN SPIKED CLAY SAMPLES (PPB)
Participating Laboratories
Analyte
2,3,7,8-TCDD

1,2,3,4-TCDD

1,3,6,8-TCDD

1,3,7,9-TCDD

1,3,7,8-TCDD

1,2,7, 8-TCDD

1,2,8,9-TCDD

1,2,3,4,7-
PeCDD

1,2,3,7,8-
PeCDD

1,2,3,4,7,8-
HxCDD

1,2,7,8-TCDF

1,2,3,7,8-
PeCDF

1,2,3,4.7,8-
HxCDF

Spike
Level
5.0

5.0

2.50

2.50

5.0

5.0

5.0


5.0


5.0


5.0

5.0


5.0


5.0

LEMSCO
4.87
5.15
3.52
3.67
2.11
2.11
2.26
2.18
4.88
5.00
4.16
4.58
4.32
4.51

4.40
4.56

5.48
5.22

5.25
5.29
4.61
4.63

4.82
4.94

4.70
4.90
I
9.54
12.4
7.88
5.58
5.40
2.18
4.0
3.92
10.8
10.8
10.3
9.74
6.83
5.29

7.89
9.61

9.45
5.58

9.58
4.83
7.80
4.87

8.07
4.76

7.83
4.03
II
4.9
5.4
6.6
9.2
2.65
6.1
6.4
9.6
12.8
19.0
4.4
4.7
ND
ND

0.3
0.3

ND
ND

ND
ND
3.3
3.1

ND
ND

ND
ND
III
4.29
4.27
4.00
4.10
ND
ND
0.54
0.68
2.61
2.77
3.90
3.86
4.26
4.17

3.20
3.29

3.65
3.88

2.64
2.80
4.17
4.29

3.88
3.65

1.78
2.13
IV
3.9
ND
4.1
ND
__a
__a
__a
__a
3.0
ND
3.8
ND
5.7
ND

6.0
4.7

8.5
6.8

6.1
29. Ob
4.8
ND

5.0
ND

5.8
9.6
V
9.3
7.7
8.3
7.5
3.0
3.8
__a
__a
9.7
7.9
6.9
7.6
7.5
8.1

6.6
7.0

7.9
10.4

7.3
9.4
9.3
8.4

8.4
8.1

8.1
9.0
Mean
6.52

5.86

3.42

3.70

8.11

5.81

5.63


4.82


6.69


5.91

5.39


5.74


5.79

RSD
Percent
42.9

97.8

45.6

82.7

63.1

41.8

26.8


58.5


34.5


42.3

39.0


33.1


47.8

 aNot determined.
 &Not included in calculation
ND = Not detected.
of mean and standard Deviation.
                                       32

-------
           TABLE 7-4.  INTERLABORATORY TEST OF METHOD 8280, PHASE I:
            QUANTITATION OF ANALYTES IN SPIKED SLUDGE SAMPLES (PPB)
Analyte
2,3,7,8-
TCDD
1,3,7,8-
TCDD
1,2,7,8-
TCDD
1,2,8,9-
TCDD
1,2,3,4,7-
PeCDD
1,2,3,7,8-
PeCDF
Spike
Level
125.0
125.0
125.0
125.0
125.0
125.0
Participating Laboratories
LEMSCO
132.1
127.9
147.3
134.5
132.2
128.5
136.4
135.6
113.3
118.8
115.1
123.4
I
62.2
48.9
50.7
29.1
85.7
38.3
60.9
33.7
67.2
42.4
48.2
33.4
II
156.4
153.2
204.2
185.1
181.6
173.5
ND
ND
ND
ND
ND
ND
III
270
220
ND
ND
50
37
160
3.40
87
68
ND
ND
IV
150
155
93
84
180
160
260
180
110
140
120
150
V
163.2
184.8
177.9
116.1
ND
ND
805.2^
205.6
ND
ND
170.9
171.7
Mean
152.0
122.2
116.7
145.8
93.3
116.6
RSD
Percent
39.4
47.8
50.8
47.3
34.9
44.4
 aNot included in calculation of mean and standard deviation.
ND = Not detected.
                                       33

-------
           TABLE 7-5.   INTERLABORATORY TEST OF METHOD 8280, PHASE I
                          ACCURACY AND BIAS OF RESULTS
Sample   Spike Level      Number of     Accuracy       Bias        ± SD of
 Type        (ppb)      Determinations   (Percent)   (Percent)    Bias Estimate
Clay
Clay
Sludge
2.5
5.0
125
16
114
57
142.4
120.4
101.1
+ 42.4
+ 20.4
+ 1.12
± 11.8
± 5.21
± 6.14
COMMENTS FROM LABORATORIES ON PHASE I OF INTERLABORATORY STUDY

LABORATORY I

     None

LABORATORY II

1.   Standards that were provided for calibration were at a concentration level
     that did not allow the highest concentration requested to be analyzed
     without further concentration of the solution.

          Analytical standards provided with Phase II of the Inter!aboratory
          Study will be at a concentration in the range 10 to 50 ng/uL.  It
          will therefore not be necessary to concentrate the standards for the
          initial calibration.

2.   Internal standard (1^Cio-2,3,7,8-TCDD)  recovery for all samples was noted
     to be extremely low.  this should be investigated before any further work
     is attempted.

          The Method has been shown to routinely provide recoveries  greater
          than 60 percent.  A summary of results provided by the other partici-
          pating laboratories is included in this report.

3.   The cleanup of samples after extraction seems tedious when compared with
     Dioxin-IFB procedures.  The use of concentrated sulfuric acid presents a
     safety problem and in the case of sludge samples required seven rinses to
     remove all  color.

          The Dioxin-IFB method requires determination only of 2,3,7,8-TCDD.
          The sulfuric acid wash is necessary for certain sample matrices,
          e.g., industrial sludges, still bottoms, sawdust, etc., which contain
          high levels of interferences.  The revised Method requires a maximum
          of four acid washes.  Appropriate  care should be exercised.
                                       34

-------
LABORATORY  III

1.   The selection of internal standards must be improved.  At a minimum, we
     feel that 13C12-TCDD, 13C12-OCDD, and 13C12-TCDF should be used.   As
     standards become more available, the goal should be to have an internal
     standard for each congener level (true isotope dilution technique).

          Agreed; the OCDD isomer is now included in the Method, and the  use of
          other isomers was referenced in previous versions as well as in the
          current edition.

2.   We feel 13C12-1»2,3,4-TCDD is a poor choice as the recovery standard
     because it can swamp the 2,3,7,8-TCDD internal standard response on  a DB-5
     column.  An isomer with better separation from 2,3,7,8-TCDD should be
     selected.

          Baseline resolution of the internal  and recovery standards would
          certainly be preferable; however, to our knowledge, no other   Cio-
          TCDD is available.  The internal  standard response can be swamped
          when poor recoveries from the cleanup step are obtained.

3.   The DFTPP requirement is totally useless; it has nothing to do with
     Cl4-Clg selected ion monitoring analyses.  As an alternative, we would
     suggest calibrating the instrument using FC-43, then using mass-ratio
     windows as acceptance of instrument tune and peak identification.

          Agreed.  The revised version of Method 8280 deletes the DFTPP tune
          requirement and substitutes criteria based on mass calibration  over
          the range m/z 222-506 using perfluoro-tri-n-butylamine (FC-43).

4.   Use tetradecane as a keeper instead of the "just to dryness" technique.

          Agreed, the use of an appropriate "keeper" would be preferable.

5.   Use tetradecane as solvent for the final  extract.  This reduces volume
     changes and makes it much more accurate to specify exact final volumes
     (e.g.,  100 uL, 50 uL, 10 pL).

          Tetradecane may be used as the final solvent, as noted in the
          procedure.  For the purposes of this study, toluene is recommended.

6.   Make the final volume for all matrix types 50 uL.  This would increase
     the absolute response of internal  standards and require less recovery
     standard.

          To require a 50 pL final volume for all  sample types may require very
          small  sample size for high-level  samples.  The response of internal
          standards is adequate provided the extraction/cleanup procedure is
          performed efficiently.   It is difficult  to concentrate accurately to
          low volumes, and it is not recommended.
                                       35

-------
*

f           7.   For highly organic sludge samples, we would suggest adjusting the extract
i                after roto-evaping to 5 mL with hexane, then proceeding to the cleanup
|                step with only 1 mL of the extract.  This reduces by 80 percent the
»                organic matrix that' tends to overload the columns and alter elution pro-
I                files to the point where internal standard recoveries are affected.
I
|                     Although this would remove much of the organic interferences, it
5j                     would also of course remove 80 percent of the analytes of interest.
I                     In addition, 80 percent of the expensive 13Ci2-labeled standards
1                     would also be discarded.  Spiking with five times the normal  level  of
*                     *3Cj2-standards would probably be prohibitively expensive.

-------
          1:3,6,8- and 1,3,7,9-TCDD were each spiked at 2.5 ppb by error;  this
          requirement has been removed from this phase of the study.

4.   A Soxlilet extractor procedure with solvents such as benzene has  been
     demonstrated to work well as the initial means of separation.  The
     petroleum ether/methanol  extraction of clay looks weak.

          A Soxhlet procedure may require as much as 16 hours per extraction,
          and benzene is not a recommended solvent because of its toxicity/
          carcinogenicity; toluene would be an acceptable substitute.   This
          procedure is included as an acceptable method of extraction.

5.   We have routinely done CDD's in a variety of samples for the past  10  years,
     This is the worst set of recoveries I have ever seen.  The step  where the
     analyte has been lost needs to be found and the procedure modified to
     correct the recovery problem.

          The Method has been shown to routinely provide recoveries greater
          than 70 percent.  A summary of results provided by  the other  par-
          ticipating laboratories is included in this report.

6.   Isotope dilution (mass spectrometry) has been used since the late  1930's,
     for inorganic compounds in more recent times.  Two complications  relative
     to this technique should be mentioned.

     a.   When the ratio of labeled compounds to analyte is large or  when  the
          ratio of analyte to labeled compound is large, the  accuracy  and
          precisions are decreased through error propagation.  The minimum
          error occurs when the ratio is 1:1.

     b.   When different types of compounds are being quantitated relative to
          the internal standard, the method is basically an internal  standard
          method.  It works only as long as pre-separation of the analytes has
          not occurred.

               a.   Agreed.

               b.   Agreed.

               We recognize that this laboratory recommends the use of  external
               standard  quantitation.  However, that procedure is not to be
               used in this method.

LABORATORY V

1.   Some of the terminology relating to the standards is confusing and, in  at
     least one instance  of what is printed in the Method, wrong.   The use  of
     the term "internal  standard" can be totally confusing.   At one point  in
     the method, the compound  relative to which response factors  are calculated
     and relative to which the quantities of analyte are calculated is  referred
     to as the "internal  standard."  Then, at a later point in the procedure,
     the "internal standard" is described as the compound added shortly before

                                       37

-------
-.—\
                   analysis and, at another point, the "internal  standard"  is the compound
                   added at the start of the extraction procedure.   In my laboratory,  we have
                   removed any confusion by referring to the "Quantitation  Standard"  and the
                   surrogate compound.  We don't use the term "internal  standard" at  all.
                   The Quantitation Standard is the basis for all  quantitative calculations.
                        The compound added at the start of the extraction procedure and
                        referred, to as the internal  standard is ^C-Z.S.y.S-TCDD.  This is
                        the same compound relative to which response factors are calculated
                        and against which analytes are quantitated.  Use of the term
                        "internal standard" is therefore accurate and self-consistent.   The
                        compound added to the extract shortly before GC/MS analysis is  not
                        referred to as the internal  standard; it is a different compound
                        (1JC]_2-1»2,3,4-TCDD) , it is referred to as the recovery standard
                        and is the compound against which the recovery of the internal  stan-
                        dard is quantitated.

                   We found the 200 ng/mL calibration standard to be below our quantifiable
                   limit.   We could not get good reproducible results for the calibration
                   standard at this level.  The extracted ion current profiles for the  masses
                   of interest are jagged peaks near the noise level  of the instrument  at
                   this concentration, and we did not feel  that we cculd validly use the
                   information from this calibration sample in performing our quantitative
                   calculations.  Perhaps the method ought to provide for  'ns contingency
                   with the option of discarding this lower point or of raising the concen-
                   tration in order to generate a valid data point.

                        Samples containing dioxins/furans at low levels need a calibration
                        standard at this level for accurate quantisation.  If 200 ng/ml is
                        too low to be accurately quantitated, then instrument sensitivity
                        should be improved by cleaning ion source or rods or both, replacing
                        the GC column, etc.

                   The criterion of 10 percent for the relative abundances of ions in any
                   given isotope cluster is not realistic.   I'm not sure just where it  ought
                   to be set (or, indeed, if this type of criterion ought to be set at  all).
                   We encountered a range of variation greater than 10 percent in going
                   through the concentration rar:ge of our calibration samples (see the  tables
                   with the clay soil  results).  This variation might be due to a number of
                   reasons, but at least in the calibration standards, one is certain of the
                   compound identifications and if the criterion cannot be met across a
                   concentration range, the criterion is unrealistic.

                        The criterion of relative ion abundances being within ±10 percent  of
                        the theoretical values is readily achieved, even with very low  level
                        calibration standards.  The  allowable error has been increased  to
                        ±15 percent in the revised Method.   A relative abundance criterion is
                        necessary because of potential,  co-el uting interferences,  e.g.,
                        chlorinated diphenyl  ethers.
                                                     38

-------
4.    Setting a criterion of satisfying DFTPP tuning requirements for this
      analysis is the height of idiocy and probably contributes significantly to
      problems No. 2 and 3.  The lowest mass of interest in this entire assay
      is mass 306.  Thus, all of the DFTPP peaks below mass 300 are irrelevant
      and tuning to have the correct mass balance for the masses below 100
      incurs substantial costs in overall sensitivity.  It would make far more
      sense to tune the quadrupole mass spectrum for resolution and optimum
      sensitivity, with good peak shape for the higher masses of the calibration
      standard.  Tuning in this manner would enhance sensitivity for the higher
     masses to lower LODs.

          Agreed.  The revised version of Method 8280 deletes the DFTPP tune
          requirement and substitutes criteria based on mass calibration over
          the range m/z 222-506 using perfluorotri-n-butylamine (FC-43).

5.    A blanket requirement to generate confirmatory mass spectra using peak top
      enhancement may also produce false negutives, since sometimes the best
      spectra are produced by enhancement and sometimes the best spectra are
      produced by manual averaging with specific background subtraction.  The
     operator should be left with discretion in this area.

          There is no such requirement in the Method.

6.   The exact technique used for final  volume adjustment of samples appears to
     be critical  in determining even whether quantisation standard will be
     observed.  There is a very real tendency for the analyte to cling to the
     glass walls, so a concentration technique which ensures that the analyte
      is not blown toward the vessel  walls is really important.  Also, our
     experience is that toluene is the solvent of choice for the calibration
     samples.  We had problems with observing the compounds in methylene
     chloride.

          The concentration step is indeed critical.  Vials should be silanized
          prior to use to minimize the tendency of analytes to cling to the
          glass walls.  Only a very gentle stream of nitrogen should be used in
          the blow-down step.   To our knowledge, no other concentration method
          is available.  Toluene j_s_ the solvent of choice:  the calibration
          standards were not provided in methylene chloride.

7.    It should be stressed in the method that it is extremely important to do a
     thorough filtration (with copious solvent rinses)  to remove particulate
     materials.   The presence of particulate materials produces terrible
     emulsions in subsequent separation  steps, with associated loss of analyte.

          Samples containing large amounts of particulate matter should have
          the acid washes performed  first.   This will  greatly reduce the
          problem of particulates in the filtrate.   Thorough  rinsing of the
          filter  with solvent  is of  course important.   It is  almost impossible
          to prevent the formation of emulsions in  very dirty samples.

8.    It is also critical  to pre-test the column chromatography with standards
      in order to  monitor exactly which fraction to  analyze.   The method states

                                       39

-------
     that a certain elution fraction will contain the analyte, but exactly
     which fraction contains the analyte and whether it is appropriate to
     combine certain fractions in order to obtain the t :t yield of surrogate
     can only be determined for a given batch of Mumina by performing a
     preliminary test.

          Of course the column chromatography should be pre-tested using tetra-
          through octa-calibration standards, as specified in the Method
          (Section 6.1).  To combine fractions from the alumina column defeats
          the purpose of the column-  Alumina which is kept properly condi-
          tioned shjuld require very little change in the comprsition of the
          methylene chloride/hexane fraction.

PHASE II

     The objective of Phase II of the interlaboratory study was to test the
applicability of the Method to the analysis of samples which wer? much more
difficult and complex than those used for Phase I.  Significant revisions to
the Method were made between completion of Phase I and initiation of Phase II.
These revisions were made to several critical areas, including:

     -  Multiple ion detection descriptors

     -  Carbon column cleanup

     -  Sample/final extract size

     -  Analytes to be quantitated

     -  DB-5 GC column requirement

     -  Deletion of DFTPP tune requirement

     -  Insertion of perfluoro-tri-n-butylamine (FC-43) tune req"irement

     -  Sample reanalysis requirements based on recovery criteria for the
        internal standards.

Further, tht-.se revisions were based both on our own laboratory experiences and
on the comments/suggestions provided by the participating laboratories at the
conclusion of Phase I.   The same five laboratories which took part in Phase I
also collaborated in Phase II, and each was provided with identical  packages of
samples, analytical  standards and documentation etc.,  similar to those provided
for Phase I.   Each laboratory was requested to analyze a total  of 10 samples of
4 different sample types.  As in Phase I,  duplicate samples of each  sample type
were provided to check  accuracy and precision.  In addition, a third different
sample of each of two sample types was also to be  analyzed.  The greatest dif-
ference between the requirements for Phase I and Phase II  was in which analytes
were to be quantitated.   Whereas Phase I required  the  determination  of certain
selected CDD's and CDF's, the Method was tested under  Phase II for the quanti-
tation of total  tetra-  through octa-CDD's  and CDF's.   This change as specified


                                       40

-------
j;            in the Method presented some difficulties in view of the use of  different  MID
«i            descriptors at different points along the GC elution profile.   It  is  appreci-
'            ated that not all  of the isomers within a given series  may be detected  using
|            this procedure.  For instance,  at least five early-eluting PeCDD isomers will
I            probably overlap the late-eluting TCDD/TCDF isomers.  Similar overlap may  occur
\            between the penta-, hexa-,  and  hepta-congener groups.   However,  the chromato-
»            graphic windows prescribed  here were developed within the limitations of the
1            PCDD/PCDF standards available to this laboratory during the course of this
,|            study.  It is expected that, as additional  reference standards become avail -
\            able, chromatographic conditions will be refined accordingly.
'1
ij                 However, to ensure standardization amongst the laboratories for  the pur-
;            poses of this study, explicit instructions regarding change-over points of
            the MID descriptors were provided.

i                 Each laboratory was provided with a set of the following 10 samples which
•            had been analyzed  by Lockheed-EMSCO.  The samples were  all  hazardous  waste
;            samples provided by the EPA and hod not been spiked with PCDD's/PCDF's  in
            contrast to the samples used for Phase I.

            Sample No.         Type                   Sample Mo.         Type

              1,2       Fly-ash duplicates                6         Sludge A

               3        Soil  A                          7,8        Sludge B  duplicates

              4,5       Soil  B duplicates               9,10       Still-bottom duplicates

                 Data packages from the six laboratories were audited with an  emphasis on
            Method performance.  The data v/ere  entered into a spreadsheet to facilitate
            comparison and evaluation.   In  those instances where a  large difference
            occurred among the values reported  for a particular sample, the  raw data were
            examined to try to determine if the difference was a Laboratory  problem or a
            Method problem.  Laboratory II  evidenced analytical  problems due to errors or
            not following instructions  which affected their results for all  of the  sam-
            ples.  Only one-fifth the required  amount of internal  standard was added;  this
            resulted in extremely low internal  standard responses  and biased quantisation.
            Because of these problems,  it was recommended that the  data from this labora-
            tory be excluded from the statistical  analysis as not  being representative of
            Method performance.  Problems with  individual  data points were also examined
            and were traced to a variety of causes.   Calculation and data  transposition
            errors were found; other discrepancies were traced to differences  in  instrument
            sensitivity, to differences in  retention-time windows  scanned, and to isomer
            peaks which met identification  criteria  in one laboratory and  not  in  others.
            The data reported  are summarized in Tables 7-6 through  7-13; although the  data
            from Laboratory II were excluded from the statistical analysis for the  reasons
            cited above, they  are included  in these  Tables for completeness.

                 In general, the Method performed well  when the  laboratories followed  the
            protocol.   A visual examination of  the data showed that, approximately 85 per-
            cent of the values reported by  the  5 laboratories and used  in  the  statistical
            analysis were consistent among  the  laboratories.

                                                   41

-------
               TABLE  7-6.   INTERLABORATORY TEST OF METHOD  8280,  PHASE  II:  PERCENT  RECOVERY
                                 OF  INTERNAL STANDARD  13C12-2,3,7,8-TCDD
Participating Laboratories
Sample
Fly-Ash*
Soil A
Soil B*
Sludge A
Sludge B*
Sti 1 1 -
Bottom*
LEMSCO
89.8
81.1
74.4
67.2
62.2
64.9
53.8
41.9
72.6
84.8
I
103
98
42.3
54.3
54.3
77.7
78.7
84.1
118
23
11
98.1
102
53.6
50.0
62
40.6
40.9
48.9
61.9
78.3
in
59
60
53.6
51
45
18
69
51
74
69
IV
64
56
46
42
46
33
72
58
46
53
V
101
109
75
90
82
74
72
90
104
90.5
All Results
Mean
85.1
F7.5
58.8
51.4
63.4
72.9
RSD Percent
23.5
24.3
25.2
47.3
26.1
35.2
Excluding
Lab II Results
Mean
82.1
58.3
59.4
53.5
67.1
73.5
RSD Percent
25.2
26.7
27.1
49.6
23.1
38.1
*Blind duplicates.

-------
OJ
                  TABLE 7-7.  INTERLABORATORY TEST OF METHOD 8280, PHASE II:  PERCENT RECOVERY
                                         OF INTERNAL STANDARD 13C12-OCDD
3=3=========:
Sample
Fly-Ash*
1
Soil Pi
Soil 8*

Sludgfe A
i
Sludge B*
i
1
Still-
Bottom*
Participating Laboratories
LEMSCO
75.5
78.1
46.2
68.1
67.3
55.5

47.3
39.2

96.4
125
I
98
119
53
87
67
75

145
132

180
30.8
II
90
50
65
3.2
5.8
oa

23
24

oa
oa
III
34
38
29
14
29
37

50
28

32
33
IV
64
56
43
95
58
40

75
48

58
46
V
102
83
112
150
119
46

74
47

55
39
Al
Mean
74.0

58.0
63.6

50.7

61.0


69.5

1 Results
RSD Percent
33

49
71

30

65


71

.5

.9
.6

.2

.4


.3

Excluding
Lab II Results
Mean
74.8

56.6
75.4

50.7

68.6


69.5

RSD Percent
36.7

56.8
53.1

30.2

57.8


71.3

   aNot included in calculation of mean and standard deviation.
   *B1ind duplicates.

-------
TABLE 7-8.  INTERLA80RATORY TEST OF METHOD 8280, PHASE II:  QUANTITATION OF
              TOTAL DIOXINS AND DIBENZOFURANS IN FLY-ASH (PPB)
Participating Laboratories
Analyte
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDF
LEMSCO
106
97
136
108
215
217
116
102
55
50
215
203
138
130
65
89
24
22
7
6
I
87.3
78.7
196
212
300
136
128
250
209
159
73.9
89.6
148
129
53.6
44
53.2
49.8
ND
17.7
II
51
44.6
14.4
58.1
170
176
135
415
191
137
ND
ND
14.4
45.3
24.8
15.2
47.2
97.4
16.4
14.6
III
150
116
201
94.4
391
324
417
295
160
124
281
219
223
86.6
148
107
119
82
26.6
19.3
IV
66
66
55
16
230
170
150
180
100
92
150
140
96
67
50
37
380
130
16
9
V
81
148
231
248
330
323
262
377
186
574
115
301
175
201
113
123
67
79
5
20
All
Mean
91
131
249
236
170
179
121
73
96
14
Results
RSD
Percent
37.6
64.8
32.8
50.1
80.8
43.3
51.6
58.7
99.7
48.7
Excluding Lab
II Results
Mean
100
150
264
228
171
179
139
83
101
14.1
RSD
Percent
30.7
52.9
30.9
48.7
88.5
43.3
35.8
46.2
103.4
53.8

                                    44

-------
en

Analyt
TCDD
PeCDt
TABLE 7-9
. INTERLABORATORY
D10XINS
TEST CF METHOD 8280, PHASE II: QUANTITATION
AND DIBENZOFURANS IN SOIL A (PPB)
Participating Laboratories
,e LEMSCO
395
ND
HxCDD ND
HpCDD ND
OCDD 5
TCDF 31
PeCDF ND
HxCDF ND
HpCDF ND
OCDF ND
I
712
ND
ND
ND
3.51
31.6
ND
ND
ND
ND
II
470.6
ND
ND
20.7
141.1
38.4
ND
ND
2.6
7.1
III
674
ND
ND
ND
ND
49.3
4.1
ND
ND
ND
IV
480
14
ND
27
140
32
ND
ND
ND
ND
V
764
ND
ND
ND
ND
54
4
ND
ND
ND
All Results
RSD
Mean Percent
583 26.1
14
ND
24 18.7
72 108.7
39.4 25.3
4.1 1.7
ND
2.6
7.1
OF TOTAL

Excluding Lab
II Results
Mean
605
14
ND
27
50
39.6
4.1
ND
ND
ND
RSD
Percent
26.3
—
--
—
158.3
28.2
1.7
—
--
—

-------
TABLE 7-10.   INTERLABORATORY TEST OF METHOD 8280,  PHASE II:  QUANTITATION OF TOTAL
                     DIOXINS AND DIBENZOFURANS IN  SOIL B (PPB)
Participating Laboratories
Analyte
TCDD

PeCDD

HxCDD

HpCOD

OCDD

TCDF

PeCDF

HxCDF

HpCDF

OCDF
===333====:
LEMSCO
2
2
ND
ND
265
330
2,974
2,989
7,550
7,229
15
22
19
ND
510
423
725
706
523
506
I
ND
ND
7.17
2.17
427
456
3210
3930
ND
ND
ND
0.85
51
50.3
333
409
627
578
626
772
II
ND
ND
4
2.3
230
277
194
224
94.2
88.5
0.6
0.4
53
50.4
344
331
67.2
78.9
9.4
7.9
III
ND
ND
4.6
ND
430
298
9480
3990
6880
6310
6.5
5.2
66
60.2
777
553
2750
1100
1140
557
IV
ND
ND
5.1
2.3
360
560
2450
4500
6800
8500
9.8
7.3
120
715
630
715
620
1000
470
800
V
ND
ND
ND
ND
333
389
3254
4365
45174
33726
ND
8
53
62
475
308
612
794
1166
1205
Excluding Lab
All Results II Results
Mean
2

3.9

363

3463

12235

7.6

118

484

805

649

RSD RSD
Percent Mean Percent
0 2

47.1 4.3

26.0 385

68.2 4114

121.7 15271

89.9 9.3

169.0 133

32.5 513

85.0 951

61.5 777

0

48.6

22.5

48.6

99.8

69.9

165.3

30.5

68.9

37.7


-------
TABLE 7-11.  INTERLABORATORY TEST OF  METHOD 8280,  PHASE  II:   QUANTITATION  OF  TOTAL
                    OIOXINS AND DIBENZOFURANS  IN  SLUDGE  A (PPB)
=238=====
Analyte
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDF
Participating Laboratories
LEMSCO
6
ND
1741
15743
19013
23
1
52
400
60
I
ND
181
2150
12500
18400
7.26
67.1
99.3
174
88.6
II
ND
ND
986
337
42.1
ND
ND
47.8
4.4
ND
III
ND
431
6970
14500
12900
ND
ND
622
304
110
IV
ND
77
3520
2200
14000
ND
33
270
300
100
V
ND
ND
768
6348
52912
ND
ND
48
ND
271
:==================
All Results
RSD
Mean Percent
6
230
2689
8605
19545
15
34
190
237
126
--
79.2
86.1
76.3
90.7
73.6
98.1
120.2
64.5
66.0
================
Excluding Lab
II Results
RSD
Mean Percent
6
230
3030
10258
23445
15
34
218
295
126
—
79.2
79.7
56.3
71.2
73.6
98.1
111.5
31.4
66.0

-------
               TABLE 7-12.   INTERLABORATORY TEST OF  METHOD 8280,  PHASE II:  QUANTITATION OF TOTAL
                                   DIOXINS AND DIBENZOFURANS IN SLUDGE B (PPB)
00
=03======
Analyte
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDF
Participating Laboratories
LEMSCO
10
ND
ND
ND
353
356
1582
1494
2588
2911
69
78
6
28
491
471
457
411
98
94
I
ND
ND
ND
ND
444
310
1590
450
3800
3080
82.1
ND
56.7
35.6
9.67
237
130
295
ND
97.9
II
ND
ND
ND
ND
234
216
155
228
253
386
45.1
9.3
21
ND
179
134
446
40.8
7.8
8.2
III
ND
ND
42.5
30
604
607
2400
2460
3160
3290
93.3
77.7
120
148
694
744
716
781
145
147
IV
ND
ND
ND
ND
320
306
110
3100
2950
5300
ND
ND
NO
ND
380
410
360
830
76
260
V
ND
ND
NO
ND
708
326
1748
1797
11403
5774
27
64
54
82
552
523
627
508
31
36
All Results
Mean
10
36
399
1426
3741
61
61
402
467
91
RSD
Percent
—
24.3
39.7
69.7
77.4
45.7
77.6
55.9
52.4
81.8
Excluding Lab
II Results
Mean
10
36
433
1673
4426
70
66
451
512
109
RSD
Percent
—
24.3
34.7
53.4
60.3
30.2
72.6
47.3
44.0
63.7


-------
TABLE 7-13.  INTERLABORATORY  TEST OF  METHOD  8280,  PHASE  II:  QUANTITATION OF TOTAL
                  DIOXINS AND DIBENZOFURANS  IN  STILL-BOTTOM  (PPB)

                                                                         Excluding Lab
              Participating Laboratories                  All Results        II  Results
Analyte L
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDF
appm
EMSCO
1210
1443
ND
ND
1771
ND
ND
ND
ND
ND
2693
2453
1813
166^
163
163
ND
ND
ND -
ND •

I
3280
ND
ND
ND
ND
ND
ND
ND
ND
ND
2463
3103
3373
1433
2320
1160
5450
3410
2800
810

II
553
948
17.7
ND
ND
ND
ND
ND
ND
ND
112a
1443
3993
703
237
253
119
456
54.1
162

III
1300
1790
330
311
ND
ND
ND
ND
ND
ND
2793
2603
1913
1783
133
123
8650
8100
1850
1700

IV
640
640
ND
460
95
ND
ND
440
760
2860
2403
1803
1853
1413
143
848
4720
4070
1050
720

V
1600
2160
ND
ND
ND
ND
ND
ND
ND
ND
294a
275a
3143
2223
263
143
11245
4940
ND
840

RSD
Mean Percent
1415 56.7
280 66.7
933 127
440
1810 82
238a 25.7
2113 44.4
9.73 83.7
5116 68.3
1110 78.6

Mean P
1563
367
933
440
1810
2603
2063
11.53
6.33
1.43

RSD
'ercent
51.9
22.1
127.0
—
82.0
13.8
32.7
69.2
42.9
54.7


-------
 STATISTICAL ANALYSIS OF DATA FROM PHASE  II OF  INTERLABORATORY STUDY

      The  following three statistical tests were applied:

 Test  A    to  see  if the mean recovery of each  internal standard is a function
          of  sample type

 Test  B    to  see  if the laboratories are equivalent in accuracy (mean analyti-
          cal value) for any of the 10 analytes

 Test  C    to  see  if the laboratories are equivalent in precision (variance of
          mean analytical  value) for any of the 10 analytes

      The  data used were those reported by five laboratories (LEMSCO, I, III,
 IV, V).   Data from Laboratory II were excluded because the internal standards
 were  added at only one-fifth of the level required; this resulted in very low
 area  counts.  All calculations were performed on a IBM PC/XT computer using a
 Lotus-123 worksheet written specifically for tin's application.

      Test A was performed using one-way analysis of variance, and the results
are presented in Table 7-14 for 1JCi2-2,3,7,8-TCDD and   C12-OC[)D.  The results
                                  ffect due to sample type tor 13C
	— — .„.    -^ ._„„.  ... _ther words, !3C^2-2,3,7,8-~
depends on sample type whereas 13C12-OCDD recovery does not.
show that there appears to be an effect due to sample type for   Ci2-2,3,7,8-
TCDD but not for 13C-OCDD.  In other words, 13C12-2,3,7,8-TCDD recovery
     Test B was performed using two-way analysis of variance.  Analytical
values reported as "Not Detected" were considered to be zero for the purposes
of this statistical analysis and were disregarded.  The results, presented in
Table 7-15, show that significant differences among laboratory means were found
only for OCDD.  The other analytes do not exhibit significant evidence of
interlaboratory variation of the mean.

     Test C was performed using Cochran's test for homogeneity of variance.
The data used were the analytical values reported for the blind duplicate
samples (soil, fly-ash, still-bottom, and sludge), and the results are pre-
sented in Table 7-16.  As the Table shows, 9 of the 40 possible matrix/analyte
combinations exhibited significant evidence of unequal variances among the 5
laboratories.  The other 31 combinations showed no such evidence, but it should
be noted that several had insufficient data due to "Not Detected" analytical
values.

     It should be further noted that equivalency of variances was assumed in
Test B.  In view of the results of Test C, the significance level indicated in
Test B should be considered approximate only.  Fortunately, each of the F-
ratios in Test B was either not significant (probability > 0.05) or was highly
significant (probability < 0.01), with none being moderately significant (0.05
> probability > 0.01),  Thus the results of Test B are valid.
                                       50

-------
TABLE 7-14.   STATISTICAL TEST A:   ANALYSIS OF VARIANCE
          OF RECOVERY OF INTERNAL STANDARDS
= =:= = = = = = = = = = = = = :
Source of
Variation

Samples
Error
Total
*Significant at

Sampl es
Error
Total
Degrees of
Freedom

3
46
49
5 percent level (

3
46
49
Sum of Mean
Squares Square F-Ratio
13C12-2,3,7,8-TCOD
3,905.8 1,301.9 3.02*
19,816.2 430.8
23,722.0
a = 0.05).
13C12-OCDD
919.3 306.4 0.22
65,054.2 1,414.2
65,973.5
                          51

-------
             TABLE 7-15.   STATISTICAL  TEST B:   LABORATORY  EQUIVALENCY
                     (TWO-WAY  ANALYSIS OF  VARIANCE  ON  MEANS)
Analyte
TCDD
PeCDD
HxCDD
HpCDD
OCDD
TCDF
PeCDF
HxCDF
HpCDF
OCDK
*Significant

Samples
9
9
9
9
9
9
9
9
9
9
at 1 percent
Degrees of Freedom
Labs. Error
4 36
4 36
4 36
4 36
4 36
4 36
4 36
4 36
4 36
4 36
level (a = 0.01).
F-Ratios
Total Samples Labs.
49 753* 0.65
49 197* 1.95
49 485* 0.95
49 1,006* 1.38
49 426* 4.39*
49 17,877* 1.30
49 4,595* 1.21
49 868* 1.94
49 692* 2.44
49 409* 1.58

TABLE 7-16. STATISTICAL TEST C: LABORATORY EQUIVALENCY
(COCHRAN'S TEST FOR EQUAL VARIANCES)
Matrix
Soil
Soil
Soil
Soil
Fly-Ash
Fly-Ash
Fly-Ash
Fly-Ash
Sludge
Analyte
HpCDD
ocno
PeCDF
HpCDF
OCDD
TCDF
PeCDF
HpCDF
HpCDD
Cochran's
G-Statistic
0.84*
0.98*
1.00*
0.94*
0.97*
0.89*
0.91*
0.98*
0.87*
Laboratory with
Largest Variance
III
V
IV
III
V
V
III
IV
IV
*Significant at the 5 percent level (a = 0.05).
Note:  This Table lists only those matrix/analyte combinations with significant
       G-statistics.  All other combinations were not significant.

                                       52

-------
                                   REFERENCES



1.   Poland, A., and A. Kende.  Fed. Proc. 3_5, 2404 (1976).

2.   Merz, V., and W. Weith.  Ber. St 460 (1872).

3.   Fishbein, L.  The Science of the Total  Environment.  4, 305 (1973).

4.   Shea, K. P., and B. Lindler.  Environment. 17, 12 (1975).
     U.S. Environmental Protection Agency.  Hazardous Waste Disposal Damage
     Reports.  Document No. 2.  EPA/530/SW-151.2 (1975).
     Commoner, B., and R. E. Scott.  Accidental Contamination of Soil with
     Dioxin in Missouri:  Effects and Countermeasures.  Center for the Biology
     of Natural Systems.  Washington University, St. Louis, Missouri (1976),

5.   Federal Register 40 CFR 261:1978,  January 14,  1985.

6.   Federal Register 40 CFR 65:14514,  April  4, 1983.

7.   U.S. Environmental Protection Agency.  Single-Laboratory Evaluation of the
     RCRA Method for Analysis of Dioxin in Hazardous Waste.  EPA-600/4-85/082
     (1985).

8.   Glaser, J. A., D.  L. Foerst, G.  D. McKee, S. A. Quave, and W.  L. Budde.
     Environmental  Science and Technology, 15, 1426 (1981).
                                       53

-------
                            Appendix  A

     Definition and Procedure for the Determination
                of the Method Detection Limit
The method detection limit (MOD is defined as the minimum concentration of a
substance that can be identified, measured and reported with 99% confidence that
the analyte concentration is greater than zero and determined from analysis of a
sample m a given matrix containing analyte

Scope and Application
This procedure is designed for applicability to a wide variety of sample types
ranging from reageni (blank) water containing analyte to wastewaier containing
analyte The MDL for an analytical procedure may vary as a function of sample
type  The procedure requires a complete, specific and well defined analytical
method  It is essential that all sample processing  steps of the analytical method be
included in the determination of the method  detection limit

Tne MDL obtained by this procedure is used to judge the significance of a single
measurement of a future sample

The MDL procedure was designed for applicability to a broad variety of physical
and chemical methods  To accomplish this, the procedure was made device- or
instrument-independent

Procedure
1.  Make an estimate of the detection limn using one of the following

   (a) The concentration value  that corresponds to an instrument signal-'noise
       ratio in the range of 2 5 to 5  If the criteria for qualitative identification of
       the analyte >s based upon pattern recognition techniques, the least
       abundant signal necessary to achieve identification must be considered m
       making the estimate
   (b) The concentration value  that corresponds to three times the standard
      deviation of replicate instrumental measurements for the analyte m
       reagent water
   (c) The concentration value  that corresponds to the region of the standard
      curve where there is a significant change  in sensitivity at low analyte
      concentrations, i e . a break m the slope of the standard curve
   (d) The concentration value  that corresponds to known instrumental
       limitations

   It  is recognized that the experience of the analyst is important to this process
   However, the analyst musttnclude the above considerations in trie estimate
   of the detection limit

2. Prepare reagent (blank) water tha; is as free of analyte as possible Reagent or
   interference free water is defined as a water sample in which analyie and
   mterterent concentrations are not detected at the method detection limit of
   each analyte of interest Interferences are defined as systematic errors m the
   measured analytical signal of an established procedure caused by the
   presence of interfering species (mterferent). The interferent concentration is
   presupposed to be normally distributed m representative samples of a given
   matrix.

                                     54

-------
                               3.  (a) If the MDL is to be determined m reagent water (blank), prepare a
                                      laboratory standard (analyte in reagent water) at a concentration which is
                                      at least equal to or in the same concentration range as the estimated MDL
                                      (Recommend betwean 1 and 5 times the estimated MDL.) Proceed to Step
                                      4.

                                   (r>) If the MDL is to be determined in another sample matrix, analyze the
                                      sample. If the measured level of the analyie is in the recommended range
                                      of one to five times the estimated MDL. proceed to Step 4.

                                      If the measured concentration of analyte is less than the estimated MDL.
                                      add a known amount of analyte to bring the concentration of analyte to
                                      between one and  five times the MDL. In the case where an interference is
                                      coanalyred with the analyte:

                                      If the measured level of analyte is greater than five times the estimated
                                      MDL. there are two options.

                                      (1) Obtain another sample of lower level of analyie in same matrix if
                                        possible
                                      (2) The sample may be used as is for determining the MDL if the analyte
                                        level does not  exceed 10 times the MDL of the analyie in reagent
                                        water. The variance of the analytical method changes as the analyte
                                        concentration  increases from the MDL. hence the MDL determined
                                        under these circumstances may not truly reflect method variance at
                                        lower analyte concentrations.

                              4.  (a) Take a minimum of seven aliquots of the sample to be used to calculate
                                     the MDL and process each through the entire analytical method. Make all
                                     computations according to the defined method with final results in the
                                      method reporting  units  If blank measurements are required to calculate
                                     the measured level of analyte. obtain separate blank measurements for
                                     each sample aliquot analyzed The average blank measurement is
                                     subtracted from the respective sample measurements.
                                  (b)  It may be economically and technically deirable to evaluate the estimated
                                      MOL before proceeding with 4a.  This will. (1) prevent repeeting this entire
                                     procedure when the costs of analyses are high and (2) insure that the
                                     procedure is being conducted at the correct concentration. It is quite
                                     possible that an incorrect MDL can be calculated from dan obtained at
                                     many times the real MDL even though the background concentration of
                                     snalyie is less than five times the calculated MDL. To insure that the
                                     estimate of the MDL is a good estimate, it is necessary to determine that a
                                     lower concentration of ana.yte will not result in a  significantly lower MDL.
                                     Take two aliquots  of the sample to be used to calculate the .'/DL and
                                     process each through the entire method, including blank measurements
                                     as described above m 4a. Evaluate theso data:
i
j                                     (1) If these measurements indicate the sample is in the desirable range for
                                        determining the MDL. take five additional aliquots and proceed. Use
                                        all seven measurement* to calculate the MDL
{                                     (2) If these measurements indicate the sample is not in the correct range.
|                                        reestimate the MDL. obtain new sample as in 3 and repeat either 4a or
                                        4b.
                                                                   55

-------
 I
!!
                                  5.   Calculate the variance (S2) and standard deviation (S) of the replicate
                                      measurements, as follows:
                                                           = (S2)
                                     where: the x.. i = 1 to n are the analytical results in the final method reporting

                                     units obtained from the n sample aliquots and ^   x,2 refers to the sum of
                                     the X values from i = 1 to n.                 '  = ^

                                 6.  (a) Compute the MDL as follows.

                                                          MDL = t,n-,, l-o. e»i(S)


                                     where

                                            MDL = the  method detection

                                     tm-i  1-0 . 9». = the  students' t value appropriate for a 99% confidence
                                                   level and a standard deviation estimate with n-1 degrees
                                                   of freedom. See Table

                                               S = stendprd deviation of the replicate analyses.

                                     (b) The 95% confidence limits for the MDL derived in 6a are computed
                                        according to the following equations derived from percentiles of the chi
                                        square over degrees of freedom distribution (X'/df) and calculated as
                                        follows

                                         MDLLCL = 0.69 MDL
                                         MDLuci. =  1-92 MDL

                                        where MDLi.ei and MDLoci. are the lower and upper 95% confidence limits
                                        respectively based on seven aliquots.

                                 7.  Optional iterative procedure to verify the reasonableness of the estimated
                                     MDL and calculated MDL of subsequent MDL determinations.

                                     (a) If this is the initial attempt to compute  MDL based on the estimated MDL
                                        in Step 1. take the MDL as calculated in Step 6. spike in the matrix at the
                                        calculated MDL and proceed through the procedure starting with Step 4.
                                     (b) If the current MDL determination is an iteration of the MDL procedure for
                                        which the spiking level does not permit qualitative identification, report the
                                        MDL as that concentiation between the current spike level and the
                                        previous spike level which allows qualitative identification.
                                     (c) If the current MDL determination is an iteration of the MDL procedure and
                                        the  spiking level allows qualitative  identification, use S2 from the current
                                        MDL calculation and S2 from the previous MDL calculation to compute  the
                                        F ratio.
                                                                      S'
                                                                    if  ff < 3.05
                                                                      SB
                                                                    56

-------
        then compute the pooled standard deviation by the following equation:
                 [651*651"] ' 2
                 L  12    J
         SA
        if r? > 3.05. respike at the last calculated MDL and process the samples
         OB
        through the procedure starting with Step 4.
     (c) Use the SP
-------
                                   APPENDIX B

       RCRA METHOD 8280 WITH REVISIONS BASED ON MULTI-LABORATORY TESTING:
              METHOD OF ANALYSIS FOR CHLORINATED DIBENZO-^-DIOXINS
                               AND DIBENZOFURANS


1.       Scope and Application

1.1      This method is appropriate for the determination of total  tetra-,
penta-, hexa-, hepta-, and octachlorinated dibenzo-£-dioxins and dibenzofurans
in chemical wastes including still  bottoms, filter aids, sludges, spent carbon,
fly ash, reactor residues, soil, and water.

1.2      The sensitivity of this method is dependent upon the level  of inter-
ferents within a given matrix.  Target quantification levels of individual
analytes were 1 ppb in solid samples and 10 ppt in water.

1.3      This method is recommended for use only by analysts experienced with
residue analysis and skilled in mass spectral  analytical techniques.

1.4      Because of the extreme toxicity of these compounds, the analyst must
take necessary precautions to prevent exposure to himself, or to others, of
materials known or believed to contain PCDD's  or PCDF's.  Typical infectious
waste incinerators are probably not satisfactory devices for disposal  of
materials highly contaminated with  PCDD's or PCDF's.  Generators of 1  Kg or
more of dioxin wastes must register as a generator.  A laboratory planning to
use these compounds should prepare  a disposal  plan to be reviewed and  approved
by the Dioxin Task Force of the EPA (Contact Conrad Kleveno, WH-548A,  U.S. EPA,
401 M Street S.W., Washington, D.C. 20450).  Additional  safety instructions are
outlined in EPA Test Method 613.

2.       Summary of the Method

2.1      This procedure is an extraction, cleanup, and high resolution capillary
column gas chromatography-low resolution mass  spectrometry method using inter-
nal standard techniques which allow for the measurement  of PCDD's and  PCDF's
in the extract.

2.2      If interferes are encountered, the  method provides selected cleanup
procedures to aid the analyst in their elimination.  The analysis flow chart is
shown in Figure 1.

3.       Interferents
                                       58

-------
                       Sample
                             (1)  Add Internal  Standards:  13C12-2,3,7,8-TCDD
                                  and 13C12-OCDD ,13C  _2 3 7 f 	
                       Sampl e
                       Extract
                             (1)
                             (2)
                             (3)
                             (4)
                             (5)
                             (6)

                             (7)
                             (2)  Perform matrix-specific extraction
                          t
Wash with 20% KOH
Wash with water
Wash with cone.
Wash with water
Dry extract
Evaporate to near dryness and
redissolve in hexane
Alumina column
                   60% CH2Cl2/hexane
                      Fraction
                             (1)   Concentrate to 400 uL
                             (2)   Carbon column cleanup
                             (3)   Add recovery standard 13C12-1,2,3,4-TCDD
                   Analyze by GC/MS
Figure 1.   Method 8280 flow chart for the  analysis  of PCDD's  and  PCDF's.
                                   59

-------
3.1       Solvents, reagents, glassware, and other sample processing hardware
may yield discrete artifacts or elevated baselines or both which may cause
misinterpretation of chromatographic data.  All of these materials must be
demonstrated to be free from interferents under the conditions of analysis by
running method blanks.  Solvents distilled in all-glass systems are required.

3.2       Interferents co-extracted from the sample will vary considerably from
source to source and will depend upon the industrial  process being sampled.
PCDD's and PCDF's are often associated with other interfering chlorinated
compounds such as RGB's and polychlorinated diphenyl  ethers which may be found
at concentrations several orders of magnitude higher than that of the analytes
of interest.  Retention times of target analytes must be verified using
reference standards.  While certain cleanup techniques are provided as part of
this method, unique samples may require additional cleanup techniques to
achieve the method detection levels stated in Table 9.

3.3      Resolution of the 2,3,7,8-TCDD isomer from other closely eluting
TCDD's must be used to establish column performance criteria.  High resolution
capillary columns are used to resolve as many PCDD and PCDF isomers as
possible; however, no single column is known to resolve all  of the isomers in a
complex mixture.

4.       Apparatus and Materials

4.1      Sampling equipment for discrete or composite sampling.

4.1.1    Grab sample bottle—amber glass, 1-liter or 1-quart volume.  French  or
Boston Round design is recommended.  The container must be acid washed and
solvent rinsed before use to minimize interferences.

4.1.2    Bottle caps—threaded to screw onto the sample bottles.  Caps must be
lined with Teflon.  Solvent washed foil  used with the shiny side toward the
sample may be substituted for Teflon if the sample is not corrosive.  Apply
tape around cap to completely seal  cap to bottle.

4.1.3    Compositing equipment—automatic or manual  compositing system.  No
tygon or rubber tubing may be used, and the system must incorporate glass
sample containers for the collection of a minimum of 250 ml.  Sample containers
must be kept refrigerated after sampling.

4.2      Water bath—heated, with concentric ring cover, capable of temperature
control (+2°C).  The bath should be used in a hood.

4.3      Gas chromatograph/mass spectrometer data system.

4.3.1    Gas chromatograph:   An analytical  system with a temperature-
programmable gas chromatograph and all  required accessories  including  syringes,
analytical columns, and gases.

4.3.2    Fused silica capillary columns are required.   As  shown in  Table  1,
four columns were evaluated  using a column  performance check mixture
containing 1,2,3,4-TCDD, 2,3,7,8-TCDD,  1,2,3,4,7-PeCDD, 1,2,3,4,7,8-HxCDD,

                                       60

-------
 1,2,3,4,6,7,8-HpCDD, OCDD, and 2,3,7,8-TCDF.

 The columns include the following:  (a) 50 m CP-Sil-88 programmed  60°-190°  at
 20°/minute, then 190°-240° at 5%ninute; (b) 30 m DB-5 programmed  170°  for  10
 minutes, then 170°-320° at 8°/minute, hold at 320°C for 20 minutes;  (c)  30  m
 SP-2250 programmed 70°-320° at 10°/minute; (d)  30 m DB-225 programmed  70° -
 230° at 10°/minute.  Column/conditions (a) provide good separation of  2,3,7,8-
 TCDD from the other TCDD's at the expense of longer retention  times  for  higher
 homologs.  Column/conditions (b) and (c) can also provide acceptable separation
 of 2,3,7,8-TCDD.  Resolution of 2,3,7,8-TCDD from the other TCDD's is  better on
 column (c), but column (b) is more rugged and may provide better separation
 from certain classes of interferents.

 4.3.3    Mass spectrometer:  Capable of scanning from 222 to 506 amu in  not
 less than 5 seconds, utilizing 70 volts (nominal) electron energy  in the
 electron impact ionization mode and producing a mass spectrum  which  meets the
 criteria in Table 2 when a mass calibration using perfluoro-tri-n-butylamine
 (FC-43) is performed.  The system must also be  capable of selected ion  moni-
 toring (SIM) for at least nine ions simultaneously, with a cycle time  of 1  sec
 or less.  Minimum integration time for SIM is 50 ms.  Acceptable selected ion
monitoring is verified by injecting 0.15 ng of  native TCDD to  give a minimum
 signal-to-noise ratio of 5 to 1 at m/z 320.

4.3.4    GC/MS interface:  Any GC-to-MS interface that gives an acceptable
calibration response for each analyte of interest at the concentration  required
 and achieves the required tuning performance criteria (see Sections  6.1-6.3)
may be used.  GC-to-MS interfaces constructed of all glass or  glass-lined
materials are required.  Glass can be deactivated by silanizing with dichloro-
dimethylsilane.  Inserting a fused silica column directly into the MS  source is
 recommended.

 4.3.5    Data system:  A computer system must be interfaced to the mass  spec-
 trometer.  The system must allow for the continuous acquisition and  storage on
machine-readable media of all mass spectra obtained throughout the duration of
 the chromatographic program.  The computer must have software  that can  search
 any GC/MS data file for ions of a specific mass and can plot such  ion
 abundances versus time or scan number.  This type of plot is defined as  an
 Extracted Ion Current Profile (EICP).  Software must also be able  to integrate
 the abundance, in any EICP, between specified time or scan number  limits.

 4.3.6    HPLC pump with loop valve (1.0 ml) injector to be used in the  carbon
 column cleanup procedure.

 4.4      Pipettes-Disposable, Pasteur, 150 mm long x 5 mm ID (Fisher Scientific
 Company, No. 13-678-6A, or equivalent).

 4.5      Amber glass bottle (500 ml, Teflon-lined screw cap).

 4.6      Reacti-vial  1 mL, amber glass (silanized)  (Pierce Chemical  Company).

 4.7      500 ml Erlenmeyer flask (American Scientific Products Cat.  No.  f4295-
 SOOfO) fitted with Teflon stoppers (ASP No. S9058-8, or equivalent).

                                       61

-------
4.8      Wrist Action Shaker (VWR No. 57040-049, or equivalent).

4.9      125 mL and 2 L Separatory Funnels (Fisher Scientific Company,  Mo.
10-437-5b, or equivalent).

4.10     500 mL Kuderna-Danish fitted with a 10 ml concentrator tube and
3-ball Snyder column (Ace Glass No. 6707-02, 6707-12, 6575-02, or equivalent).

4.11     Teflon boiling chips (Berghof/American Inc., Main St., Raymond,
New Hampshire 03077, No. 15021-450, or equivalent).  Wash with hexane prior to
use.

4.12     300 mm x 10.5 mm glass chromatographic column fitted with Teflon stop-
cock.

4.13     15 mL conical concentrator tubes (Kontes No. K-288250, or equivalent).

4.14     Adaptors for concentrator tubes (14/20 to 19/22) (Ace Glass No. 9092-
20, or equivalent).

4.15     Nitrogen evaporator (N-Evap No. 1156, or equivalent).  Teflon tubing
connection to trap and gas regulator is required.

4.16     Microflex conical vials (Kontes K-749000, or equivalent).

4.17     Filter paper (Whatman No.54, or equivalent).

4.18     Carbon Column:  A silanized glass HPLC column (10 mm x 7 cm),  or
equivalent, prepared by mixing 5 percent (by weight)  active carbon AX-21,
(Anderson Development Co., Adrian, Michigan), washed  with methanol and  dried j_n
vacuo at 110°C, and 10 urn silica (Spherisorb S10W from Phase Separations, Inc.,
Norwalk, Connecticut).  The mixture must be stirred and sieved through  a 38 pm
screen (U.S. Sieve Designation 400-mesh, American Scientific Products,  No.
S1212-400, or equivalent) to remove any clumps.JV

4.19     Dean-Stark trap, 5 or 10 ml with T joints, (Fisher Scientific  Company,
No. 09-146-5, or equivalent) condenser and 125-mL flask.

5.       Reagents

5.1      Potassium hydroxide-(ACS), 20 percent (w/v)  in distilled water.

5.2      Sulfuric acid-(ACS), concentrated.

5.3      Methylene chloride, hexane, benzene, petroleum ether, methanol, tet-
radecane, isooctane, toluene, cyclohexane.   Distilled in  glass.


\J  -    The carbon column preparation and  use is adapted from W.  A.  Korfmacher,
         L. G.  Rushing, D. M. Nestorick, H.  C.  Thompson,  Jr.,  R.  K.  Mitchum,
         and J. R.  Kominsky, Journal  of High Resolution  Chromatography  and
         Chromatography Communications,  jj,  12-19 (1985).

                                       62

-------
5.4      Prepare stock standards in a glovebox from concentrates or neat
materials.  The stock solutions are stored in the dark at 4°C and are checked
frequently for signs of degradation or evaporation especially just prior to
the preparation of working standards.

5.5      Alumina, neutral, Super 1, Woelm, 80/200 mesh.  Store at room tempera-
ture in a desiccator with CaS04 drying agent.  Oven drying at 600°C overnight
is acceptable, but alumina so processed should be checked for contamination by
solvent rinsing and GC/ECD analysis.

5.6      Prepurified nitrogen gas.

5.7      Anhydrous sodium sulfate (reagent grade).  Extracted overnight with
hexane using a Soxhlet extraction apparatus and dried at 100°C.

6.       Calibration

6.1      Before using any cleanup procedure, the analyst must process a series
of calibration standards (Section 11) through the procedure to validate elution
patterns and the absence of interferents from reagents.  Both open column and
carbon column performance must be checked.  Routinely check the 8 percent
CH2Cl2/hexane eluate of environmental extracts from the alumina column for
presence of target analytes.
                                                  21
6.2      Prepare multi-level  calibration standards"  keeping the recovery
standard (13C12-1,2,3,4-TCDD) and the internal standard (13C12-2.3,7,8-TCDD) at
fixed concentrations of 500 ng/mL.  A second internal standard, I3Cj2-OCDD, at
a fixed concentration of 1000 ng/mL is recommended for use when quantification
of the hepta- and octa-isomers is required.  The use of separate internal
standards for the PCDF's is also recommended.  Recommended concentration levels
for standard analytes (Section 11.1.3) are 200, 500, 1000, 2000, and 5000
ng/mL.  Calculation of response factors is described in Section 11.2.  Stand-
ards must be analyzed using the same solvent as used in the final extract;
toluene is recommended.  A wider calibration range is useful for higher level
samples provided it can be described within the linear range of the method.

6.3      Establish operating  parameters for the GC/MS system; the instrument
should be tuned to meet the isotopic ratio criteria listed in Table 2 for
FC-43,  By injecting calibration standards, establish the response factors of
standards vs. the appropriate internal standard.  (PCDF response factors are
established vs. 13C12-2,3,7,8-TCDF if this standard is used).  Response factors
for the hepta- and octa-chlorinated CDD's and CDF's are to be calculated using
the corresponding 13C12-octachlorinated standards as described in Section 11.2.
2J   -   *3Cj2-labeled analytes are available from Cambridge  Isotope  Labora-
         tory, Woburn, Massachusetts.   Proper standardization  requires  the  use
         of a specific labeled isomer  for each congener  to  be  determined.   When
         labeled PCDD's and PCDF's of  each homolog are available,  their use
         will be required to be consistent with the technique  of isotopic dilu-
         tion mass spectral analysis.

                                       63

-------
6.4      An adequate instrumental detection limit should be verified by
injecting 0.15 ng of 13C12-2,3,7,8-TCDD which should give a minimum signal  to
noise ratio of 5 to 1 at m/z 334.  GC column performance should be checked  for
resolution and peak shape daily using a mixed standard such as the GC column
performance check mixture described in Section 4.3.2.

7.       Quality Control

7.1      Before processing any samples, the analyst must demonstrate through
the analysis of a distilled water method blank that all  glassware and reagents
are interferent-free at the method detection limit of the matrix of interest.
Each time a set of samples is extracted, or there is a change in reagents,  a
method blank must be processed as a safeguard against laboratory contamination.

7.2      Standard quality assurance practices must be used with this method.
Field replicates must be collected to validate the precision of the sampling
technique.  Laboratory replicates must be analyzed to determine the precision
of the analysis.  Fortified samples must be analyzed to establish the accuracy
of the analysis.  Field blanks must be collected to verify that sample
collection processes are free from cross-contamination.

8.       Sample Collection, Preservation, and Handling

8.1      Grab and composite samples must be collected in glass containers.
Conventional  sampling practices must be followed.  The bottle must not be pre-
washed with sample before collection.  Composite samples should be collected in
glass containers.  Sampling equipment must be free of tygon, rubber tubing  and
other potential sources of contamination.

8.2      All  samples must be stored at 4°C, extracted within 7 days, and
completely analyzed within 30 days of collection.

9.       Extraction and Cleanup Procedures

9.1      Internal standard addition.  Use a sample aliquot of 0.1-10 g (typical
sample size requirements for each type of matrix is provided in Section 9.2) of
the chemical  waste or soil to be analyzed.  Transfer the sample to a tared
flask and determine the weight of the sample.  Add an appropriate quantity  of
13C12-2,3,7,8-TCDD and any other material which is used as an internal stand-
ard, (Section 6.2).  All samples should be spiked with at least one internal
standard, for example, 13C12-2,3,7,8-TCDD, to give an approximate concentration
of 500 pg/yL in the final concentrated extract.  As an example, a 10 g sample
concentrated to a final  volume of 100 uL requires the addition of 50 ng of
13C12-2,3,7,8-TCDD at 100% recovery.

9.2      Extraction

9.2.1    Sludge.  Extract aqueous sludge samples by refluxing a sample (e.g., 2
g) with 50 ml of toluene in a 125 mL flask fitted with a Dean-Stark water
separator.  Continue refluxing the sample until all  the water has been removed.
Cool the sample, filter the toluene extract through Whatman No. 54 filter paper
or equivalent into a 100 ml round bottom flask.  Rinse the filter with 10 ml

                                       64

-------
 of toluene,  combine  the  extract  and  rinseate.  Concentrate the combined solu-
 tion to near dryness using  a  rotary  evaporator at 50°C.  Use of an inert gas to
 concentrate  the extract  is  also  permitted.  Proceed with step 9.2.4.

 9.2.2    Still  bottom.   Extract  still bottom samples by mixing a sa.nple (e.g.,
 0.1  g)  with  10  mL  of toluene  and  filtering the solution through Whatman No. 54
 filter  paper (or equivalent)  into a  50 ml round bottom flask.  Rinse the filter
 with 10 ml of toluene.   Concentrate  the combined toluene solution to near
 dryness using a  rotary evaporator at 50°C.  Proceed with step 9.2.4.

 9.2.3    Fly ash.  Extract  fly csh samples by placing a sample (e.g., 10 g) and
 an equivalent amount of  anhydrous sodium sulfate in a Soxhlet extraction
 apparatus charged  with toluene; extract for 16 hours using a three cycle/hour
 schedule.  Cool  and  filter  the toluene extract through Whatman No. 54 filter
 paper (or equivalent) into  a  500  ml  round bottom flask.  Rinse the filter with
 5 ml of toluene.   Concentrate the combined toluene solution to near dryness
 using a rotary  evaporator at  50°C.   Proceed with step 9.2.4.

 9.2.4   Transfer  the residue to  a 125 mL separatory funnel using 15 mL of
 hexane.  Rinse the flask with two 5 ml aliquots of hexane, and add the rinses
 to the  funnel.   Shake 2 minutes with 50 ml of 5% NaCl solution, discard the
 aqueous layer, and proceed  with step 9.3.

 9.2.5    Soil.   Extract soil samples by placing the sample (e.g., 10 g) and an
 equivalent amount  of  anhydrous sodium sulfate in a 500 mL Erlenmeyer flask
 fitted  with  a Teflon  stopper.  Add 20 mL of methanol  and 80 mL of petroleum
 ether (in that order) to the flask.  Shake on a wrist-r.ction shaker for two
 hours.   The  solid  portion of sample should mix freely.  If a smaller soil
 sample  is used,  scale down  the amount of methanol proportionally.

 9.2.5.1  Filter  the  extract from  Section 9.2.5 through a glass funnel fitted
 with  filter  paper  (Whatman  No. 54, or equivalent) and filled with anhydrous
 sodium  sulfate into  a 500 mL Kuderna-Danish (KD)  concentrator fitted with  a 10
 mL concentrator  tube.  Add  50 mL of petroleum ether to the  Erlenmeyer flask,
 restopper the flask, and swirl the sample gently; remove the stopper carefully
 and  decant the solvent through the funnel  as above.   Wash the sodium sulfate on
 the  funnel with  two  additional 5 mL portions of petroleum ether.

 9.2.5.2  Add a Teflon boiling chip and a three-ball  Snyder column to the KD
 flask.   Concentrate  in a 70°C steam bath to an apparent volume of 10 mL.
 Remove  the apparatus from the steam bath,  and  allow it to cool  for 5 minutes.

 9.2.5.3  Add 50 mL of hexane and a new boiling chip  to the KD flask.   Concen-
 trate in a steam bath to an apparent volume of 10 mL.   Remove the apparatus
 from the steam bath, and allow it to cool  for  5 minutes.

 9.2.5.4  Remove and  invert the Snyder column and  rinse it down  into  the  KD  with
two  1-mL portions of hexane.  Decant the contents  of  the  KD and concentrator
 tube  into a 125 mL separatory funnel.  Rinse the  KD with  two additional  5-mL
 portions of hexane; combine these rinsates  in  the  separatory funnel.   Proceed
with step 9.3.


                                       65

-------
 9.2.6    Aqueous samples.  Mark the water meniscus on the side of the 1-L
 sample bottle for later determination of the exact sample volume.  Pour the
 entire sample (approximately 1 L) into a 2-L separatory funnel.  Note:   A
 continuous liquid-liquid extractor may be used in place of a separatory funnel.

 9.2.6.1  Add 60 mL methylene chloride to the sample bottle, seal, and shake 30
 seconds to rinse the inner surface.  Transfer the solvent to the separatory
 funnel, and extract the sample by shaking the funnel for 2 minutes with peri-
 odic venting.  Allow the organic layer to separate from the water phase for a
 minimum of 10 minutes.  If the emulsion interface between layers is more than
 one-third the volume of the solvent layer, the analyst must employ mechanical
 techniques to complete the phase separation.  Collect the methylene chloride (3
 x 60 mL) directly into a 500 mL Kuderna-Danish concentrator (mounted with a 10
 mL concentrator tube) by passing the sample extracts through a filter funnel
 packed with a glass wool plug and 5 g of anhydrous sodium sulfate.  After the
 third extraction, rinse the sodium sulfate with an additional  30 mL of methy-
 lene chloride to ensure quantitative transfer.

 9.2.6.2  Attach a Snyder column, and concentrate the extract until the apparent
 volume of the liquid reaches 5 mL.  Remove the K-D apparatus,  and allow it to
 drain and cool for at least 10 minutes.  Remove the Snyder column, add 50 mL
 hexane, re-attach the Snyder column, and concentrate to approximately 5 mL.
 Rinse the flask and the lower joint with 2 x 5 mL hexane and combine rinses
 with extract to give a final  volurr  .•  about 15 mL.

 9.2.6.3  Determine the original sample volume by refilling the sample bottle to
 the mark and transferring the liquid to a 1000 mL graduated cylinder.  Record
 the sample volume to the nearest 5 mL.  Proceed with Step 9.3

 9.3      Partition the solvent against 40 mL of 20 percent (w/v) potassium
 hydroxide.  Shake for 2 minutes.  Remove and discard the aqueous layer
 (bottom).  Repeat the base washing until no color is visible in the bottom
 layer (perform base washings a maximum of four times).

 9.4      Partition the solvent against 40 mL of distilled water.  Shake for 2
minutes.   Remove and discard aqueous layer (bottom).

 9.5      Partition the solvent against 40 mL of concentrated sulfuric acid.
 Shake for 2 minutes.  Remove and discard the aqueous layer (bottom).  Repeat
 the acid washings until  no color is visible in the acid layer.  (Perform acid
washings a maximum of four times.)

9.6      Partition the extract against 40 mL of distilled water.  Shake for 2
minutes.   Remove and discard  the aqueous layer (bottom).   Dry  the organic  layer
by pouring through a funnel  containing anhydrous sodium sulfate, wash with two
 b mL portions of hexane, and  concentrate the hexane solution to near dryness
 with a rotary evaporator (35°C water bath); make sure all  traces of toluene
 are removed.   (Use of an inert gas to concentrate the extract  is also
 permitted).

 9.7      Pack a gravity column (glass 300mm x 10.5mm) fitted with a Teflon
 stopcock in the following manner:

                                       66

-------
 Insert a glass-wool plug into the bottom of the column.  Add a 4 gram layer
 of  sodium sulfate.  Add a 4 gram layer of Woelm super 1 neutral  alumina.
 Tap the top of the column gently.  Woelm super 1 neutral  alumina need not  be
 activated or cleaned prior to use but should be stored in a sealed  desiccator.
 Add a 4 gram layer of sodium sulfate to cover the alumina.  Elute with 10  ml
 of  hexane and close the stopcock just prior to the exposure of the  sodium
 sulfate layer to air.  Discard the eluant.  Check the column for channeling.
 If  channeling is present, discard the column.  Do not tap a wetted  column.

 9.8     Dissolve the residue from 9.6 in 2 ml of hexane  and apply  the hexane
 solution to the top of the column.  Elute enough hexane (3-4 ml) to complete
 the transfer of the sample cleanly to the surface of the  alumina.  Discard the
 eluant.

 9.8.1    Elute with 10 ml of 8 percent (v/v) methylene chloride in  hexane.  As
 a quality assurance step, check that no PCDD's or PCDF's  are eluted in this
 fraction.

 9.8.2    Elute the PCDD's and PCDF's from the column with 15 ml of  60 percent
 (v/v) methylene chloride in hexane and collect this fraction in a conical
 snaped (15 mL) concentrator tube.

 9.9      Carbon column cleanup.

         Prepare a carbon column as described in section  4.18.

 9.9.1    Using N£, gently concentrate both fractions from the alumina column
 (Section 9.8) to about 1 ml.  Wash the sides of the tube  with a small volume
 (1-2 mL)  of hexane and reconcentrate to about 1 mL.  Save the 8 percent frac-
 tion for GC/MS injection to check for any bleedthrough of PCDD's and PCDF's (a
 quality assurance step).  Evaporate the 60 percent CH2Cl2/hexane fraction  to
 about 400 \il and transfer to a HPLC injector loop (1.0 mL) for carbon column
 cleanup.   Rinse the centrifuge tube with 500 \il hexane, and add both fractions
 to  injector loop.

 9.9.2    Elute the column at 2 mL/minute, ambient temperature, with 30 ml  of
 cyclohexane/methylene chloride 1:1 (v/v).  Discard the eluant.  Next elute the
 column with 10 ml of ChfcC^/MeOH/Benzene 70:20:5 (v/v).  Discard the eluant.
 Backflush the column with 40 mL4toluene to elute and collect PCDD's and PCDF's
 (entire fraction).  The column is cleaned by pumping an additional  30 mL
methanol  followed by 40 mL of toluene in the back flush position.  After
 returning the column to the original  position, 30 mL of cyclohexane/methylene
 chloride 1:1 (v/v) is pumped through the column to re-equilibrate it in prepa-
 ration for the next sample.   The column must be replaced  following  the analysis
 of  high level  extracts (>500 ppb).

 9.9.3    Evaporate the toluene fraction to about 1 mL on  a rotary evaporator
 using a water bath at 50°C.   Transfer to a 2.0 mL Reacti-vial  using a toluene
 rinse and concentrate to the desired volume using a stream of f^.   The final
 volume should be 100 yL for soil  samples and 500 pL for sludge,  still  bottom,
 and fly ash samples; the correct volume will  depend on the relative concen-
 tration of target analytes.   Extracts which are determined to be outside the

                                       67

-------
calibration range for individual analytes must be diluted, or a smaller sample
must be re-extracted.  Gently swirl the solvent on the lower portion of the
vessel to ensure dissolution of the PCDD's and PCDF's.

9.10     Approximately 1 hour before HRGC/LRMS analysis, transfer an aliquot  of
the extract to a micro-vial.  Add to this sufficient recovery standard  (   Cjo-
1,2,3,4-TCDD) to give a concentration of 500 ng/mL.  (Example:  36 yL aliquot
of extract and 4 ul_ of recovery standard solution.  Remember to multiply  final
result by 10/9 to correct for this dilution.  Inject an appropriate aliquot (1
or 2 uL) of the sample into the GC/MS instrument by using a syringe.
10.      GC/MS Analysis

10.1     When toluene is employer! as the final solvent, use of a bonded phase
column from Section 4.3.2 is recommended.  Solvent exchange into isooctance or
tridecane is required for other liquid phases on nonbonded columns.

10.2     Calculate response factors for standards relative to the internal
standards, 13C12-2,3,7,8-TCDD and 13C12-OCDD (see Section 11).  Add  the
recovery standard to the samples prior to injection.   The concentration of the
recovery standard in the sample extract must be the same as that in  the cali-
bration standards used to measure the response factors.

10.3     Analyze samples with selected ion monitoring by using all  of the ions
listed in Table 3.  It is recommended that the GC/MS run be divided  into five
scan-monitoring sections, namely:  (1) 243, 257, 304, 306, 320, 322, 332, 334,
376 (TCDD's, TCDF's, 1<3C1? internal and recovery standards, HxCDE) ;  (2) 277,
293, 338, 340, 342, 354/356, 358, 410 (PeCDD's, PeCDF's, HpCDE) ; (3) 311, 327,
372, 374, 376, 388, 390, 392, 446 (HxCDD's, HxCDF's,  OCDE) ; (4) 345, 361, 406,
408, 410, 422, 424, 426, 480 (HpCDD's, HpCDE's, NCDE), and (5) 379,  395, 442,
444, 458, 460, 470, 472, 514 (OCDD, OCDF, 1JC12-OCDD, DCDE).   Cycle  time "not
to exceed" 1 second/descriptor.  HxCDE, HpCDE, OCDE,  NCDE, DCDE are  abbrevi-
ations for hexa-, hepta-, octa-, nona-, and deca-chlorinated  diphenyl ether,
respectively.

10.4     Identification criteria for individual PCDD's and PCDF's.

10.4.1   The retention time of the chromatographic peak (relative to that of
I3C^2-2,3,7,8-TCDD) in the sample must match that in  the standard mixture of
available isomers within 0.01 units.

10.4.2   All of the characteristic ions, i.e., quantitation ion, confirmation
ions, and the [M-COC1]+ ion listed in Table 3 for each class  of PCDD and PCDF,
must be present.

10.4.3   The maximum intensity of each of the specified characteristic  ions
must coincide within ±1 scan.

10. 4. £   The relative intensity of the selected, isotopic ions within the
molecuiar ''.•-• cluster of a homologous series of PCDD's of PCDF's must lie
within the range specified in Table 4.


                                       68

-------
 10.5     Quantitate the PCDD and PCDF peaks from the response relative  to  the
 appropriate internal standard, 13Ci2-2,3,7,8-TCDD or 13C12-OCDD.   Recovery of
 each  internal standard (13C12-2,3,7,8-TCDD and 13Ci2-OCDD7 vs.  the recovery
 standard 13Ci2-l,2,3,4-TCDD must be greater than 40 percent.   Samples with
 recoveries of less than 40 percent or greater than 120 percent  must be
 reextracted and reanalyzed.

 10.5.1   In those circumstances where these procedures do not yield a defini-
 tive conclusion, the use of high resolution mass spectrometry or  HRGC/MS/MS is
 suggested.

 11.      Calculations

 11.1     Determine the concentration of individual isomers of tetra-, penta-,
 and hexa-CDD/CDF according to the equation:

                                                Qis x AS
                      Concentration, ng/g  =
                                              G x Ajs x RF

Where:

Qis  =  ng of internal standard ^C^-Z.SJ.S-TCDD* added to the sample before
        extraction.

  G  =  y of sample extracted.

 As  =  area of quantitation ion of the compound of interest.

Ais  =  ar.ea of quantitation ion (m/z 334)  of the internal  standard,
        13C12-2,3,7,8-TCtu.

 RF  =  response factor of the auantitation ion of the compound  of interest
        rel?t-we to m/z 334 of 13C12-2,3,7,8-TCDD.

11.1.1   Determine the concentration of individual  isomers  of  hepta-CDD/CDF  and
the concentration of OCDD and QCDF according  to the equation:

                                                Qis * As
                      Concentration, ng/g   =
                                              G x  Ais  x  RF

Wnere:

Q.js  =  ng of internal  standard ^3C^2-OCDD,  added  to the sample before
        extraction.

  G  =  g of sample  extracted.

 AS  =  area of quantitation ion of the  compound of interest.
                                       69

-------
Ais  =  area of quantitation ion (m/z 472) of the internal  standard,  13C12-OCDD.

 RF  =  response factor of the uuantitatioti ion of the compound of interest
        relative to m/z 472 of 1<3C12-OCDD.

Note:  Any dilution factor introduced in section 9.10 should be applied to this
       calculation.

11.1.2   Response factors are calculated using data obtained from the analysis
of multi-level calibration standards according to the equation:
                                RF

Where:
                                      As x C-js

                                      Ais x cs
As  =  area of quantitation ion of the compound of interest.

                                   the appropriate internal  standard (m/z 334
 AJS  =  area of quantitation ion of the appropriate int
         for 13C]2-2,3,7,8-TCCD-, m/z 472 for ljC12-OCDD)
 Ci,.  =  concentration of the appropriate internal  standard,  ^C1?-2,3,7,8-rCDD
         or 13C12-OCDD.

  Cs  =  concentration of the compound of interest.

11.1.3   The concentrations of unknown isomers of TCDD shall  be calculated
using the RF determined for 2,3,7,8-TCDD.

         The concentrations of unknown isomers of PeCDD shall  be calculated
using the RF determined for 1,2,3,7,8-PeCDD or any available  2,3,7,8,X-PeCDD
i somer .

         The concentrations of unknown isomers of HxCDD shall  be calculated
using the RF determined for 1,2,3,4,7,8-HxCDD or any available 2,3,7,8,X,Y-
HXCDD i somer.

         The concentrations of unknown isomers of HpCDD shall  be calculated
using the RF determined for 1,2,3,4,6,7,8-HpCDD or any available 2,3,7,8,X,Y,Z-
HpCDL) i somer.

         The concentrations of unknown isomers of TCDF shall  be calculated
using the RF determined for 2,3,7,8-TCDF.

         The concentrations of unknown isomers of PeCDF shall  be calculated
using the RF determined for 1,2,3,7,8-PeCDF or any available  2,3,7,8,X-PeCDF
i somer.

         The concentrations of unknown isomers of HxCDF shall  be calculated
using the RF determined for 1,2,3,4,7,8-HxCDF or any available 2,3,7,8,X,Y-
HxCDF i somer.

                                       70

-------
         The concentrations of unknown isomers of HpCDF shall  be calculated
using the RF determined for 1,2,3,4,6,7,8-HpCDF or any available 2,3,7,8,X,Y,Z-
HpCDF isomer.

         The concentration of the octa-CDD and octa-CDF shall  be calculated
using the RF determined for each.

         Mean Response factors for selected PCDD's and PCDF's  are given in
Table 5.
11.1.4   Calculate for each internal standard    C^-Z^.y.S-TCDD and
^-      the recovery, R^s, ir. the sample extract, using the equation:
                                                 x 100%
                                Ars x RFr x Qis

Where:

Ars  =  Area of quantisation ion (m/z C34) of the recovery standard,
        il3C12-l,2,3,4-TCDD.

C" c  =  ng of recovery standard, ^Cio-l,2,3,4-TCDD, added to extract.
 I J                                 1. Lf

The response factor for determination of recovery is calculated using data
obtained from the analysis of the multi-level  calibration standards according
to the equation:
                                       AT c X Cpc
                                       Vs x ^1$

Where:

Crs  =   Concentration of the recovery standard, ^C^-l,2,3,4-TCDD.

11.1.5    Calculation of total  concentration of all  isomers  within  each
homologous series of PCDD's and PCPF's.

      Total  concentration  =  Sum of the concentrations  of  the  individual
      of PCDD's or PCDF's     PCDD or PCDF isomers

11.2     Report results in nanograms per gram; when duplicate and  spiked
samples are reanalyzed, all data obtained should be reported.

11.3     Accuracy and Precision.  Table  6 gives the precision data for  revised
Method  8280 for selected analytes in the matrices shown.   Table  7  gives
recovery data for the same analyses.  Table 8 gives the  linear  range  and varia-
tion of response factors over  the range  for selected analyte standards.  Table
9 gives estimated detection limits as measured in specific  sample  matrices.
                                       71

-------
    TABLE 1.  REPRESENTATIVE GAS CHROMATOGRAPH RETENTION TIMES* OF ANALYTES
Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,4-TCDD
1,2,3,4,7-PeCDD
1,2,3,4,7,8-HxCDO
1,2,3,4,6,7,8-HpCDD
OCDD
50 m
CP-Sil-88
25.2
23.6
24.1
30.0
39.5
57.0
—
30 m
DB-5
21.6
22.2
22.1
24.3
26.4
28.6
31.2
30 m
SP-2250
26.7
26.7
26.5
28.1
30.6
33.7
—
30 m
DB-225
42.5
37.3
37.6
NM
NM
NM
—
* Retention time in minutes, using temperature programs shown below.

NM = not measured.

Temperature Programs:

                           60°C-190°C at 20°/minute; 190°-240° at 5°/minute.
CP-Sil-88

30 m DB-5


SP-2250

DB-225
                           170°, 10 minutes; then at 8°/minute to 320°C, hold
                           at 320°C 20 minutes (until  OCDD elutes).

                           70°-320° at 10%ninute.

                           70°-230° at 10°/minute.

                              Column Manufacturers
CP-Sil-88
DB-5, DB-225
SP-2250
        Chrompack Incorporated, Bridgewater,  New Jersey
        0 and M Scientific,  Incorporated,  Rancho Cordova,  California
        Supelco, Incorporated, Bellefonte, Pennsylvania
                                       72

-------
   TABLE 2.  CRITERIA FOR  ISOTOPIC RATIO MEASUREMENTS FOR FC-43 CALIBRATION9


          Selected  Ions (m/z)                        Intensity Ratio
              232/231                                   3.8 -  7.0
              265/264                                   3.8 -  7.0
              315/314                                   4.5 -  8.4
              415/414                                   6.0 - 11.2
              465/464                                   6.8 - 12.6
              503/502                                   6.8 - 12.6
a Scan from m/z 222 to m/z 506; nominal scan time 5 sec.
  TABLE 3.  IONS SPECIFIED9 FOR SELECTED  ION MONITORING FOR PCDD'S AND PCDF'S
                Quantitation       Confirmation
                    Ion                Ions                     M-COC1
PCDD'S
13C12-Tetra
Tetra
Penta
Hexa
Hepta
Dcta
1
-------
    TABLE 4.  CRITERIA FOR ISOTOPIC RATIO MEASUREMENTS FOR PCDD's AND PCDF's

                    Selected ions (m/z)              Relative intensity


PCDD's

Tetra                    320/322                         0.65-0.89
Penta                    358/356                         0.55-0.75
Hexa                     392/390                         0.69-0.93
Hepta                    426/424                         0.83-1.12
Octa                     458/460                         0.75-1.01

PCDF's

Tetra                    304/306                         0.65-0.89
Penta                    342/340                         0.55-0.75
Hexa                     376/374                         0.69-0.93
Hepta                    410/408                         0.83-1.12
Octa                     442/444                         0.75-1.01
                                       74

-------
      TABLE 5.  MEAN RESPONSE FACTORS OF CALIBRATION STANDARDS
Analyte
2,3,7,8-TCDD
1,2,3, 7,8-PeCDD
1,2,3,4,7,8-HxCDD
l,2,3,4,6,7,8-HpCDDb
1.2, 3,4,6, 7,8-OCDDb
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,4,6, 7,8-HpCDFb
1,2,3,4,6,7,8-OCDFb
13C12-2,3,7,8-TCDD
13C12-1,2,3,4-TCDD
13C12-OCDD
RFa
1.01
0.96
0.80
1.08
1.30
1.51
1.48
1.29
1.57
1.19
1.00
0.74
1.00
RSD%
(n = 5)
32
10.6
10.8
6.6
7.2
3.9
13.8
13.4
8.6
3.8
-
10.9
-
Quantitation Ion
(m/z)
322
356
390
424
460
306
340
374
408
444
334
334
472
a The RF value is the mean of the five determinations made.  Nominal
  weights injected were 0.4, 1.0, 2.0, 4.0 and 10.0 ng.
b RF values for these analytes were determined relative to   Cio-
  OCDD.  All other RF's were determined relative to 13Cir-2,3,7,8-
  TCDD.

jnstrument Conditions/Tune- GC/MS system was tuned as specified in
                            Section 6.3.  RF data was acquired under
                            MID control, as specified in Section 10.3.

GC Program- The GC column was programmed as specified in Section 4.3.2(b).
                                  75

-------
TABLE 6.  PRECISION DATA FOR REVISED METHOD 8280
Compound
2,3,7, 8-TCDD




1,2,3,4-TCDD




1,3, 6, 8-TCDD




1,3,7,9-TCDD




1,3, 7, 8-TCDD




1,2, 7, 8-TCDD




1,2,8,9-TCDD




Matrix1
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
301 1
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
Analyte Level
Native
ND2
378
ND
ND
487
ND
ND
ND
38.5
ND
ND
ND
ND
19.1
227
ND
NO
ND
58.4
ND
ND
ND
ND
16.0
422
ND
ND
ND
2.6
ND
ND
ND
ND
ND
ND
(ng/g)
Native
+ Spike
5.0
378
125
46
487
5.0
25.0
125
84.5
2500
2.5
25.0
125
65.1
2727
2.5
25.0
125.0
104.4
2500
5.0
25.0
125
. 62.0
2920
5.0
25.0
125
48.6
2500
5.0
25.0
125
46
2500
N
4
4
4
2
3
3
4
4
4
2
4
4
4
2
2
4
4
4
2
2
4
4
4
4
2
4
4
4
3
2
4
4
4
2
2
Percent
RSD
4.4
2.8
4.8
-
24
1.7
1.1
9.0
7.9
-
7.0
5.1
3.1
-
-
19
2.3
6.5
-
-
7.3
1.3
5.8
3.5
-
7.7
9.0
7.7
23
-
10
0.6
1.9
-
-
                                                    (continued)





                       76

-------
TABLE 6.  (Continued)
Compound
1,2,3,4,7-PeCDD




1,2,3,7,8-PeCDD




1,2,3,4,7,8-HxCDD




1,2, 3,4,6, 7,8-HpCDD




1,2,7,8-TCDF




1,2,3,7,8-PeCDF




1,2,3,4,7,8-HxCDF




Matrix
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge-^
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom^
clay
soil
sludge
fly ash
still bottom
Analyte
Native
ND
ND
NO
25.8
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
8966
ND
ND
ND
ND
ND
7.4
ND
ND
ND
ND
ND
25600
ND
ND
13.6
24.2
ND
Level (ng/g)
Native
+ Spike
5.0
25.0
125
71.8
2500
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
5.0
25.0
9091
-
-
5.0
25.0
125
53.4
2500
5.0
25.0
125
46
28100
5.0
25.0
139
70.2
2500
55SSSSSSSJ
N
4
4
4
4
2
4
4
4
2
2
4
4
4
2
2
4
4
4
-
-
4
4
4
4
2
4
4
4
2
2
4
4
4
4
2
Percent
RSD
10
2.8
4.6
6.9
-
25
20
4.7
-
-
38
8.8
3.4
-
-
4
30.6

-
-
3.9
1.0
7.2
7.6
-
6.1
5.0
4.8
_
-
26
6.8
5.6
13.5
_
                                       (continued)
          77

-------
                             TABLE 6.  (Continued)

                                         Analyte Level (ng/g)
                                                    Native             Percent
Compound              Matrix           Native       + Spike      N       RSD


OCDF                  clay              ND             -
                      soil              ND             -
                      sludge           192           317         4         3.3
                      fly ash           ND             -
                      still  bottom      ND             -


1 matrix types:

  clay:  pottery clay, Westwood Ceramic Supply Co., City of Industry, California.

  soil:  Times Beach, Missouri, soil  blended to form a homogeneous sample.  This
  sample was analyzed as a performance evaluation sample for the Contract
  Laboratory Program (CLP) in April  1983.  The results from EMSL-LV and 8
  contract laboratories using the CLP protocol were 305.8 ng/g 2,3,7,8-TCDD
  with a standard deviation of 81.0.

  fly ash:  ash from a municipal incinerator; resource recovery ash No. 1.

  still bottom:  distillation bottoms (tar) from 2,4-dichlorophenol production,
  obtained from Arthur D. Little, Inc., 1983.

  sludge:  sludge from cooling tower  which received both creosote and penta-
  chlorophenolic wastewaters.

  Cleanup of clay, soil, and fly ash  samples was through alumina column only.
  (Carbon column not used.)

2 ND - not detected at concentration  injected (final  volume 0.1 ml  or greater).

3 Estimated concentration out pf calibration range of standards.

4 Not determined.
                                       78

-------
TABLE 7.  RECOVERY DATA FOR REVISED METHOD 8280
Compound
2,3,7,8-TCDD




1,2,3,4-TCDD




1,3,6,8-TCDD




1,3,7,9-TCDD




1,3,7,8-TCDD




1,2,7,8-TCDD




1,2,8,9-TCDD




Matrix1
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
Native^
(ng/g)
ND
378
ND
ND
487
ND
ND
ND
38.5
ND
ND
ND
ND
19.1
227
ND
ND"
ND
58.4
ND
ND
ND
ND
16.0
422
ND
ND
ND
2.6
ND
ND
ND
ND
ND
ND
Spiked3
Level
(ng/g)
5.0
-
125
46
-
5.0
25.0
125
46
2500
2.5
25.0
125
46
2500
2.5
25.0
125
46
2500
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
Mean
Percant
Recovery
61.7
-
90.0
90.0
-
67.0
60.3
73.1
105.6
93.8
39.4
64.0
64.5
127.5
80.2
68.5
61.3
78.4
85.0
91.7
68.0
79.3
78.9
80.2
90.5
68.0
75.3
80.4
90.4
88.4
59.7
60.3
72.8
114.3
81.2
                                                    (continued)
                       79

-------
TABLE 7.  (Continued)
Compound
1,2,3,4,7-PeCDD




1,2,3,7,8-PeCDD




1,2,3,1,7,8-HxCDD




Matrix1
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
1,2, 3,4,6, 7,8-HpCDD clay




2,3,7,8-TCDD
(C-13)



1,2,7,8-TCDF




1,2,3,7,8-PeCDF




soil
sludge^
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
Native?
(ng/g)
NO
ND
NO
25.8
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
8966
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
7.4
ND
ND
ND
ND
ND
25,600
==================
Spiked3
Level
(ng/g)
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
5.0
25.0
125
-
-
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
5.0
25.0
125
46
2500
Mean
Percent
Recovery
58.4
62.2
79.2
102.4
81.8
61.7
68.4
81.5
104.9
84.0
46.8
65.0
81.9
125.4
89.1
ND
ND
-
-
-
64.9
78.8
78.6
88.6
69.7
65.4
71.1
80.4
90.4
104.5
57.4
64.4
84.8
105.8
-
                                       (continued)
          80

-------
                             TABLE 7.  (Continued)

Compound
1,2,3,4,7,8-HxCDF




OCDF





Matrix1
clay
soil
sludge
fly ash
still bottom
clay
soil
sludge
fly ash
still bottom
Native^
(ng/g)
NO
ND
13.6
24.2
ND
ND
ND
192
ND
ND
Spiked3
Level
(r.g/g)
5.0
25.0
125
46
2500
_
-
125
-
"
Mean
Percent
Recovery
54.2
68.5
82.2
91.0
92.9
_
-
36.8
-
"
1 matrix types:

  clay:  pottery clay, Westwood Ceramic Supply Co., City of Industry, California.

  soil:  Times Beach, Missouri  soil  blended to form a homogeneous sample.   This
  sample was analyzed as a performance evaluation sample for the Contract
  Laboratory Program (CLP) in April  1983.   The results from EMSL-LV and 8
  contract laboratories using the CLP protocol were 305.8 ng/g 2,3,7,8-TCDD
  with a standard deviation of 81.0.

  fly ash:  ash from a municipal  incinerator; resource recovery ash No. 1.

  still bottom:  distillation bottoms (tar) from 2,4-dichlorophenol  production,
  from Arthur D. Little, Inc.  (1983).

  sludge:  sludge from cooling tower which received both creosote and penta-
  chlorophenol wastewaters.

  The clay, soil, and fly ash samples were subjected to alumina column cleanup,
  no carbon column was used.

2 Final volume of concentrate 0.1 ml or greater, ND means below quantification
  limit, 2 or more samples analyzed.

3 Amount of analyte added to sample, two or more samples analyzed.

4 Estimated concentration out of  calibration range of standards.
                                       81

-------
            TABLE 8.  LINEAR RANGE AND VARIATION OF RESPONSE FACTORS
  Analyte
Linear Range Tested (pg)
Mean RF    %RSD
1,2,7,8-TCDF*
2,3,7,8-TCDD*
2,3,7,8-TCDF
50-6000
50-7000
300-4000
8
7
5
1.634
0.721
2.208
============
12.0
11.9
7.9
 * Response factors for these analytes were calculated using 2,3,7,8-TCDF as
   the internal standard.  The response factors for 2,3,7,8-TCDF were
   calculated vs. 13C12-1,2,3,4-TCDD.
** Each valfj of i- represents a different concentration level.
                                       82

-------
            TABLE 9.  DETECTION LIMITS (ppb) FOR RCRA METHOD 82801'2


Analyte Class       Clay        Soil      Fly Ash       Still Bottom3   Sludge
TCDD
TCDF
PeCDD
PeCDF
HxCDD
HxCDF
1.0
0.5
1.5
1.0
2.0
1.5
5.0
2.5
7.5
5.0
10
7.5
1.0
0.5
1.5
1.0
2.0
1.5
500
250
750
500
1000
750
25
12
38
25
50
38
1  The analytes of the class indicated were not quantified below this value.
   The instrument detection limit (S = 3 x Noise) for 2,3,7,8-TCDD in standards
   is 0.5 ppb when extrapolated for a 10 g sample concentrated to 100 uL.

2  Matrix types:

Clay:  Pottery clay, Westwood Ceramic Supply Co., City of Industry, California.

Soil:  Times Beach, Missouri, soil blended to form a homogeneous sample.  This
sample was analyzed as a performance evaluation sample for the Contract Labora-
tory Program (CLP) in April 1983.  The results from EMSL-LV and 8 contract
laboratories using the CLP protocol were 305.8 ng/g 2,3,7,8-TCDD with a standard
deviation of 81.0.  The 90 percent window was 143 to 469 ng/g.

Fly Ash:  Ash from a municipal  incinerator; resource recovery ash No. 1.

Still Bottom:  Distillation bottoms (tar) from 2,4-dichlorophenol production,
from Arthur 0. Little, Inc.  (1983).

Sludge:  sludge from cooling to-.'er which received both creosote and penta-
chlorophenolic wastewaters.

Cleanup of clay, soil and fly ash samples was through alumina column only, the
carbon column not used.

3  The still bottom samples were not tested below this level  due to high analyte
   levels found.
                                       83

-------