TB85-158145
Thermally Modulated Electron Affinity
Detector for Priority Pollutant Analysis
Radian Corp., Austin, TX
Prepared for

Environmental Monitoring and Support Lab.
Cincinnati, Oil
Jan 85

-------
                                                      PB85-1561H5
                                           -EPA/600/4-.S5/009
                                          -January 1985
           THERMALLY MODULATED ELECTRON

              AFFINITY  DETECTOR FOR

           PRIORITY POLLUTANT ANALYSIS
                        by

                   R.C. Hanisch
                    L.D. Ogle
                    A.E. Jones
                    R.C. Hall
                Radian Corporation
               Austin, Texas  78756
             Contract No. 68-03-2965
                  Project Officer

                  Stephen Billets
        U.S.  Environmental Protection Agency
             26 West St. Clair  Street
             Cincinnati, Ohio  4.W68
\ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
         OFFICE OF RESEARCH AND DEVELOPMENT
        U.S.  ENVIRONMENTAL PROTECTION AGENCY
             CINCINNATI, OHIO  45268

-------
                                   TECHNICAL REPORT DATA
                            (Please read tnslruclions on iht itvrne before completing)
 I. REPORT NO.

  EPA/6007 A-85/009
12.
                              3. RECIPIENT'S ACCESSION NO.
                                 PBS 5  1 5 8 i A 5 /AS
4. TIT LE AND SUBTITLE
 Thermally Modulated  Electron Affinity Detector for
 Priority Pollutant Analysis
                              5. REPORT DATE
                                 January 1985
                              6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)       "  '                          	

 R.C. Hanisch, L.D.  Ogle.  A.E.  Jones, R.C. Hall
                              8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS

 Radiai Corporation, Austin.Texas  78766
                              10 PROGRAM ELEMENT NO
                                                           11, CONTRACT /GRANT NO.
                                                            EPA 68-03-2965
12. SPONSORING AOENCY NAME AND ADDRESS
Environmental Monitoring and Support Laboratory
26  W.  St. Clair Street
Cincinnati, Ohio  45268
                              13. TYPE OF REPORT AND PERIOD COVERED
                                  Final	
                              14 SPONSORING AGENCV CODE
                                                              EPA 600/06
IS. SUPPLEMENTARY NOTES
16 ABSTRACT
       In the  area of environmental monitoring,  a need exists for a rapid, sensitive,
   and selective  method to analyze for chlorinaced organic compounds such as
   pesticides,  PCB,  PCDD, and PCDF at trace  lovels in  complex samples.

       In response to this neeu, a program w*s conducted to determine the feasibility
   of using a  new detector concept in the gas chromatographic analysis of certain
   priority pollutants.  The concept is based on  tne thermal alteration of a
   compound's  electron affinity in a flow-through reactor, which can be used to
   modify the  selectivity and sensitivity of the  ECD to certain compounds.  The TM
   ECO consists of two ECDs connected by a temperature-controlled reactor.  Different
   classes of  organic compounds respond to the reactor conditions in different ways:
   some compounds exhibit an enhanced ECD response after passing through the reactor;
   others a diminished signal; and still others no change in the magnitude of the
   signal.  The ratio of a co^uurid's response from the post-reactor ECD to that
   obtained from  the prereactor LCD appears to be a property characteristic of each
   compound.   This  peak area ratio can be used in conjunction with its retention time
   to increase the  confidence level of the identity of a given compound while still
   taking advantage  of the excellent sensitivity  characteristics of the ECD.
17.
                                   WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              l> IDENTIFIERS'OPEN ENDED TERMS C.  COSATI field/Croup
 B. DISTRIBUTION STATEMENT

'-.Distribute to Public
                 IB. SECURITY CLASS (ThisReport 1
                   Unclassified
21. NO. OF PAGES
     56
                                                S6COHITY ClfiSS/rhi
                                                Unclassified
                                                                        22 PRICE
EPA Fwa 2220-1 (R»». 4-77)   PREVIOUS EDITION it OBIOLCTE

-------
                                  DISCLAIMER
     This report has been reviewed by the Environmental Monitoring and
Support Laboratory-Cincinnati, U.S. Environmental Protection Agency,  and
approved for publication.  Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendations for use.
                                      ii

-------
                                   FOREWORD


     Environmental measurements  are  required  to determine  the quality of
ambient waters and the character of  waste effluents.  The  Environmental
Monitoring and Support Laboratory-Cincinnati, conducts  research to:

     •    Develop and evaluate methods to measure the presence and
          concentration of physical, chemical, anu radiological
          pollutants in water, wastewater, bottom sediments, and solid
          waste.

     •    Investigate methods for the concentration, recovery, and
          Identification of viruses, bacteria and other microbiological
          organisms In water; and, to determine the responses of aquatic
          organisms to water quality.

     •    Develop and operate an Agency-wide quality assurance program
          to assure standardization and quality control of systems for
          monitoring water and wastewater.

     •    Develop and operate a computerized system foe Instrument
          automation leading to improved data collection, analysis, and
          quality control.

     Thiu report describes the development and evaluation of an analytical
system designed to measure selected priority pollutants in various
environmental media.   The ef£e. -.3 of varying caCalvst and  tennerature
to modify compound response for  an electron capture detector were
evaluated.
                                        Robert L.  Booth
                                        Director
                                        Environmental Monitoring and
                                          Support  Laboratory-Cincinnati
                                     ill

-------
                                   ABSTRACT


     In Che area of environmental monitoring, a need exists for a rapid,
sensitive, and peleccive method to analyze for chlorinated organic compounds
such as pesticides, PCfli PCDD, and PCDF at trace levels in complex samples.
Current methodologies typically employ analysis by GC-ECD, GC-HECD, or GC-MS
for identification and quantification of the compounds of Interest.  Each of
these approaches has its advantages and limitations.1»2  The ECD usually
provides the highest sensitivity for these types of analyses but this detec-
tor responds to any compound capable of capturing electrons In an ionized
field including ubiquitous contaminants such as phthalate esters.^*   The
HECD is very selective for chlorinated organic compounds when operated In the
halogen-specific mode but it sometimes lacks the required sensitivity.  Mass
spectrometry provides unequivocal Identification of compounds but also has
problems with respect to adequate sensitivity and cost of analysis.

     In response to this need, a program was conducted to determine the
feasibility of using a new detector concept in the gas chrosatographic
analysis of certain priority pollutants.  The concept is based on the thermal
alteration of a compound's electron affinity in a flow-through reactor, which
can be used to modify the selectivity and sensitivity of the ECD to certain
compounds.  "The TM ECO consists of two ECDs connected by a temperature-
controlled  reactor.  Different classes of organic compounds respond to the
reactor conditions in different ways:  some compounds exhibit an enhanced ECD
response after passing through the reactor; others a diminished signal; and
still others no change in the magnitude of the signal.  The ratio of a com-
pound's response from the post-reactor ECD to that obtained from the pre-
reactor ECD appears to be a property characteristic of each compound.  This
peak area  ratio can be used in conjunction with its retention time to
Increase the confidence level of the identity of a given compound vhlle still
taking advantage of the excellent sensitivity charcteristics of the ECD.

     This  report was submitted in fulfillment of Contract No. 68-03-2965 by
Radian Corporation under the sponsorship of the U.S. Environmental Protection
Agency.  This report covers the period from September 1, 1980 to November 9,
1961, and  work was completed as of June IS, 1984.
                                      iv

-------
                                   CONTENTS


Disclaime r	
Foreword[[[
Abstract	      lv
Figures	*i	      vi
Tables	     vil
Abbreviations and Symbols	    viii

     1.   Introduction	       1
               Background	       1
               Scope of Work	       1
     2.   Conclusions.»	       3
     3.   Recommendations	       4
     4.   Experimental.	       5
               Hardware Development and Preliminary Evaluation	       5
               Method Validation Study.^.^....,.	;	       7
     5.   Results and Discussion..-;.;.	      U
            — -Response Ratios Obtained with Argon/Methane Carrier
                    and a Gold Reactir'ii Tube	      11
               Response Ratios Obf-alr.ed with Nitrogen Carrier and
                    a Gold Reaction Tube	      15
               Response Ratio's Obtained with Hydrogen/Helium and a
                    Gold Reaction Tube..'	      15
               Response Ratios Obtainei^wich Hydrogen/Helium and a
                    Nickel ReactioTf'Tube	      21
               Response Radios Obtained with Nitrogen Carrier and a
                    Nickel Reaction Tube	      2i
               Response Rftios Obtained with Argon/Methane Carrier
                    and a Nickel Reaction Tube....'	      24

-------
                                   FIGURES


Nunber
                                                                         Page
  1  Block diagram of detector system ................................       5

  2  Chromatcgrams of chlorinated pesticides with Ar/CH4 carrier
       and gold reaction tube ........................................      13

  3  Chromatograms of Coxaphene with Ar/CH^ carrier and gold
       reaction tube
   .\
  4 • 'Chromatograms of parathions with Ar/Cfy carrier and gold
       reaction tube ................................. ..... ...... .....      16

  5  Chromatograms of phthalates with Ar/CH* carrier and gold
       reaction tube ................................................ .      17

  6  Chroma tog rams of Aroclor 1254 with* nitrogen carrier and gold
       reaction tube .................................................      19

  7  Linearity plot - commercial ECD ......................... •„ .......      ->6

  8  Linearity plot - TM ECD .........................................      27

  9  Accuracy estimates:  Phases I ar.J II ............................      32

 10  Precision estimates:   Phases 1 and II ...........................      33

 11  Percent recovery plot - commercial ECD...* .......................      34

 12  Percent recovery plot - TM ECD ..................................      35

 13  Percent recovery plot - p re-cleanup ..................... ^s .....      3g

 14 .Percent recovery plot - post-cleanup ......... .j^r. ..............      37
   *
 15  Differential  chroma tog ram of a toxapherte-contaminated
       industrial  waste ............ ^.-r;".'. ... ..........................      3e

 16  Differential  chroma tcgJMn of Sample X-2 prior to cleanup ........      $p

 17  Differential  chroma tog raa of Sample X-2 after cleanup ....... ....      41
 18   Precision estimates for toxaphene determinations in waste
       effluent s ........ . ............. . ................ .. ............       ..



                                     vi

-------
                                  TABLES
Number                                                                  Page

  1  Compounds Used for the Electron Affinity Study ..................      '8

  2  Gas Chromatographic Parameters for the Method Validation Study..      10

  3  Response Ratios at Various Temperatures with Ar/CH^ Carrier
       and a Cold Reaction Tube ......................................      12

  4  Response Ratios at Various Temperatures with Nitrogen Carrier
       and a Cold Reaction Tube ......................................      18

  S  Response Ratios of Chlorinated Pesticides  with Helium/Hydrogen
       Carrier and a Cold Reaction Tube at 350"C .....................      20

  6  Response Ratios Obtained with Helium/Hydrogen Carrier and a
       Gold Reaction Tube at Four Temperatures. ....... . ..............      22

  7  Response Ratios at Various Temperatures with a Nitrogen Carrier
       and Nickel Reaction Tube.. ..................... •. ..............      23

  8  Response Ratios at Various Temperatures with Ar/CH^ Carrier
       and a Nickel Reaction Tube ................... . ................      25

  9  Method Validation Study:  Toxaphene Spiking and  Recovery Data
       for Phases 1 and II ...........................................      29

 10  Statistical Summary:  Accuracy and Precision Estlmaces of
       Percent Recovery for Toxaphene Spikes in Reagent Water ........      31
 11  Toxaphene Levels in Industrial Waste Effluent
 12  Statistical Summary:  Precision Estimates of Recovery  for
       Toxaphene In Waste Effluent Samples. ........ ...... . ...........      63
                                     vii

-------
                      LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATION'S

C              ~ centigrade
ECD            — electron capture detector
CC-ECD         — gas chromatography-electron capture detector
GC-HECD        — gas chromatography-Hail electrolytic conductivity
                  detector
CC-MS          — gas chromatography-iuss spe;.trometry
L              — liter
Mln.           ~ minute
mL             — mllllllter
on             ~ millimeter
MS             —mass spect rone try
ng             — nanogram (10~9 gram)
PCS            — polychlorinated blphenyl
PCDD           — polychlorinated dlbenzo-p-dloxins
PCDF           — polychlorinated dlbenzofurans
pg             — plcogram (10~!2 gram)
TM ECO         — thermally-modulated electron capture detector
ug             — microgram (10~b gran)
uL             — mlcrollter (10~6 liter)              V
US EPA         — U.S. Environmental Protection Agency
SYMBOLS

Ar             — argon
               — methane
                                    vlil

-------
                                   SECTION'  1


                                  INTRODUCTION
 BACKGROUND
 r«   3?e.P°t       f°r enhancln* ehe  selectivity and sensitivity of the ECD
 R c  ?*}" "BP°"nd8wa8 demonstrated In preliminary 8tudle8 conducted by
 R.C. Hall at Purdue University In 1973.  Thftse studies utilized a detector
 system comprised of two ECDs and a flow-through reactor.  The detectors "

        "
 cmber       i              8eile"' bUt "•» "P^ted by a reaon
 design.          reaction tube and nitrogen carrier gas wire used In this



      It was found chat at moderate reactor temperatures (up to 800eC)  the

           i KhS 8eC°wd detector «•«  ""««<* relative to the first for
  «     f   ^"f^bon Pesticides; approximately the same for PCBs; and
 S5I ?V   Pht»'al«e e««».   At high reactor temperatures (950'C)  PCBs and
 phthalates continued to produce a significant response in the second
 detector; most chlorinated hydrocarbon pesticides exhibited Uttle or no
 response.  The response ratio  of the. detector" was also foum* to be
 compound-specific and useful  for confirmation of compound identity.


      These phenomena were  used  to enhance  detector selectivity  by directly
 Ht fn&^     response ot certain components and by differentially summing

 SS-"lTp2 T        8"   !S° ellnlnate  the  re8F°«e of stable  components.    §
 These techniques  were used to enhance  detector selectivity to chlorinated
 hydrocarbon pesticides in  the presence  of  PCBs and phthalJte..
'I!"6 P™1^"8^ results were d^couraging. this approach was

      L 3nd dld n°C "Ver a U>lde Varlety of compounds.  The
«lllv   f                                               compouns.     e
stability of the system ua, uut verlfieil ^ ..tuai samples wire not investl-

af It r?U8r Uentl,y>, fUndlng W"  sollcite" to  ^ly  investigate th« utility
0f the technique and determine the  feasibility  of constructing a commercial
detector system based on this principle.
SCOPE OF WORK
     The proposed program consisted of four phases.  Phase I was designed to
demonstrate the feasibility of the concept using existing off-the-shelf
components.  Phase II would involve the construction and evaluation of a

  Uow°in »£.!eiM'  ^"^i011 and teSCln* °f 8 P'O'OW detector would
   ^  ,!5    II    t detailed evaluation of the prototype detector would
   conducted » Phase IV.  Initial funding was sought and granted for only

-------
Che £irst phase, during which a number of specific objectives were required
Co be accomplished.   They were:

     •    construction of a detection system;

     •    deCerolnatlon of response characteristics Cor selected
          coopounds as a function of reactor temperature using at lease
          two dltferenc materials as reactor cubes;

     •    investigation of response characteristics in different
          reaction gas coaposlclons Including nitrogen, argon/methane,
          and helluo/hydrogen;

     •    determination of Che most promising reaction Woe/resetion gas
          conbination;                              '

     •    syeCetn evaluation with a Halted nutnVer of model compounds;

     •    determination of Che relation of-'detector response as •
          function of key operational parameters to reproducibility and
          signal aCabillcy for each rjmpound studied;

     •    evaluation of detector specificity; and

     •    determination of &n absolute detection level for each compound
          studied.

-------
                                    SECTION 2

                                 '  CONCLUSIONS


;—.'.__ In this Initial phase of  work,  ItJbaS^been demonstrated' cfTat &  TM-ECD
 svsteffcad. have an enhanced  select ivirjT fjr Che determination  jjt;ss»€e levels
i>i--chloriu.ifv;els of
rlTitBTferences, .-Th^se interfereng>tf"are compounds that eRhi.Ki^ifr'StD'-''
 response, sneci'ilcally PCBs  ar.u phthalato esters,-andiier^s^U^ifjaiati.'sresent at
 significant levels in_sonie_&tutt-goi.m»iugarwastewater rcdtrices.  The output
_/.rrm n 'I'M .XLi can De^ii'ferentially  amplified to^pull the  signals cue  to PCBs
 and phthalste estB»'-&and-thereby Isolate the signal due only, to chlorinated
 pesticides.  TVre oevelopcicnt. of tht^'iri ECO could eliminate  the necessity for
 sample-extract cleanup in the  screening by GC of environmental samples for
 the -priority pollutant pe.sL~ici«es. -

      The opcicuo selectivity for chlorinated hydrocarbon pesticides  was
 achieved when the TM ;ECD was operated with argon/uiethane reaction gas  and'a
 gold reaction-tub? ,sfa5iearhed ^it "350*C.  The TM BCD provides  a response ratio
 which is gene-r3tre. vmics of  ECDy/2 by  the
 refipor.se in . rSa units of* ECD/fl.  The ntsgultude of this ratio appears  to be a
 function of.-cKr.iiSca! cla^«. ^-Phthalate esters exhibit a larger ECD response
'after passing'through the reactor and yield a response ratio  greater than
-one. -jhider the conditions listed  above, PCBs display less  than 10 percent
 reduction in^jxsgpbnse.  As a result, their response ratio is  approximately
 op.
-------
                                  SECTION 3

                              -RECCMMESDATIONS-


    - The evaluation of the detector prototype has demonstrated the validity
of the TM ECD concept.  A final determination of the deteqtor's potential  for
widespread application in the area of environmeucol •ecr.itruring is not
possible without additional work.  The following areas mexit continued
Investigation:

     •    determination of the reaction products responsible for the
          post-reactor signals;         -              '             '

     •    determination of molecular positional effects on the response
          factors of various isoners;' "

    "•    utilva<.viroii of capillary columns in conjunction with the TM ECD;

     •    definition of the limits of matrix effects on TM ECD response
          ratios; and

     •    optimization of TM ECD selectivity and sensitivity for PCDD and
          FCDF.

-------
                                    SECTION A

                                 -  EXPERIMENTAL


 HARDWARE DEVELOPMENT AND PRELIMINARY EVALUATION

      The chroaatcgraphlc system u sed in t hi a Etli'4v~consi s t ed of a "Radinn—11 OB
 Gas Chromatograph modifled  to accept an experimental detector systeja which IB
 shown  schematically  in  Figure 1.   It consists of two modified Tracor. ECDs
 with a  flow-through  reactor bei-^een L'ue two detectors.          >

      The detectors were modified  by replacing the 1/4-inch inlet fitting with
 .a  1/16-inch  fitting  and reworking the top detector biscuit to reposition, the
 exit port.   The modification Involved:

      •     removing th«.. I/4-inch in-let—a-e- the- base •of~rfie'"de"tect"or"^hd ~
            replacing  it  with a 1/16-inch bored-through male tube fitting;

      •     removing the  1/8-inch exit tube from the side of the top detector
            biscuit and plugging the resulting hole;  and

      •     machining  a female tube connection (10-32 thread with 39 degree
            'ferrule seat) in  the top face of the detector biscuit.

 These modifications  resulted in a straight flow-through carrier path of
 minimum volume.  It  also enabled  the reaction tube  and sample transfe.r inlet
 tube to be positioned flush to the detector cavity, thus eliminating
 stainless steel surfaces from the flow path.

      The reactor used for this study consisted.of &• two-hole ceramic tube
 wrapped with resistance heating wire and enclosed in ceramic insultation.
 The reaction tube was Inserted through  one hole  of  the ceramic tube and a
 thermocouple in the  other hole.   Temperatures of the reactor were controlled
 +2°C from 350°C to 900°C.'

      A  1/8-inch o.d.  glass-lined  stainless steel column (1.8 mm l.d. x 168 cm
 long) was used for all  separations.   The exit of the column was interfaced to
 the first detector via  a short piece of l/16*-inch o.d.  (0.035-inch l.d.) gold
 tubing.   Gold was used  for  this interface to prevent catalytic decomposition
 of the  sample prior  to  detection.

      All temperatures of the chromatographic  system were controlled by the
 DART computer which  is  an integral part of the chromatograph.   An inlet
 temperature  of 226°C and detector temperatures of 340°C were used.   The
" cofOmn  temperature was  varied to  achieve the desired separation.   Reactor

-------
GC Column
ECD #1
Reactor
ECD #2
Signal B
                         DART II
                         Computer
T | - •
i i .
• 1 i
Report For
Signal A
(area percents)

Reoort For
Signals A & B
(response ratios)
  Figure 1.  Block Diagram of Detector System

-------
temperatures were va"Yled from 350°C to 900°C.  Detector outputs were routed
through the DART II computer which integrated all peaks.  The Integration and
chronacographic information were printed on a Teletype printer.  A strip-
chart recorder was also employed in order to obtain a record of the
separation.

     Seventeen model compounds were selected for use in the determination of
the TM ECD's response characteristics.  The basis for their selection was the
fact chat they represent electron-capturing analytes from a variety of com-
pound classes '.ncluding chlorinated hydrocarbon pesticides, PCBs, phthalate
esters, organophosphate pesticides, chloroaromatics, nltroaromatics, and
chlorophenols.  The test compounds were grouped into mixtures each of which
contained similar compounds that could be resolved chromatographieally under
the analytical conditions employed.  The mixtures were Introduced to the
chromatograph in 5 uL injections.  The total amount injected for each test
compound and its absolute detection limit are presented in Table 1.

METHOD VALIDATION STUDY

The target compounds selected for use in the method validation study included
phthalate esters and toxaphene.  The validation study itself was based oa^the
analytical procedures employed In US EPA's Method 608 for organochlorine
pesticides and PCBs.  The study was conducted based on the assumption that
the phthalate esters were to act as interferants in the analysis of water
samples for toxaphene.

     The 200°C i so thermal column temperature required 'for US EPA Method 608
precluded utilization of all six priority pollutant phthalate esters.  At
200°C the 1.5% SP-2250/1.952 SP-2A01 mixed phase column employed in the
analysis of organochlorine pesticides would yield a c>rooatogram in which
dimethyl phthalate and diethyl phthalate'would be merged with the solvent
front ar.d di-n,-octyl phthalate would,-elute well after the  ast toxaphene
component.  As a result, the following phthalate esters were selected for use
in the method validation:  di-n-butyl phthalate, benzyl butyl phthalate, and
bls(2-ethylhexyl) phthalate.

System Linearity

     Prior to initiating the method validation, the linear range of both the
TM ECD and a commercially available linearized Ni63 gcD was established.
This was accomplished by generating a five-point linearity check using
toxaphene standards prepared at different concentration levels. - Triplicate
injections of each concentration level were made in order to ensure
statistically valid data.

Preparation of Youden Pairs

     The toxaphene spiking solutions used in the validation study were
prepared as ten Youden pairs.  Acetone was used as the solvent in order to
enhance the water solubility of the concentrates.  The spiking solutions were
prepared at toxaphene levels that would yield sample concentrations ranging
from  9.8 Ug/L to 115 ug/L when 1.0 mL of the concentrate was used to fortify
a 1 L water sample.  The phthalate ester concentrations were prepared at a
single level that would yield a sample concentration of 50 yg/L of each of

-------
TABLE 1.  COMPOUNDS USED FOR THE ELECTRON AFFIMjrY STUDY
Amount (ng) Injecte^' Absolute
Compound Into Chromatograph Detection Limit (ng]
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Lindane
Hcptachlor
Heptachlor Epoxlde
p,p'-DDE
p.p'-DDT
Aroclor 1254
Aroclor 1016
Toxaphene
Chlordane
Diethyl Phthalate
Dibutyl Phthalate
bis-(2-Ethxlhexyl^ Phthalate
Methyl Parathfott^J
Ethyl ParatSj?*5
1,2,4-Trichrorbb^^-
Nitrobenzene
2, 4-Dichloro phenol • ""
0.55
0.51
0.48
0.50
0.50
~5.2
-------
 the esters when a 1 L water sample was fortified with 1.0 tnL of the concen-
 trate.   The toxaphene concentration in each of the spiking concentrates was
 verified oy duplicate analysis against a toxaphene quality control check
 solution obtained from the Quality Assurance Branch of US EPA's Environmental
 Monitoring and Support Laboratory in Cincinnati, Ohio.

 Phase I

      The method validation study was conducted in three phases.  During Phase
 I, twenty 1 L samples of reagent water were fortified with a 1.0 n»L aliquot
 of spiking solution containing appropriate concentrations of tcxaphene and
 the three selected phthalate esters.  After spiking, the 20 water samples
 were extracted according to the protocol detailed in US EPA Method 608.
 Briefly, this Involved extraction of the sample with three 60 mL portions of
 nethylene chloride, drying of the combined extracts oa an anhydrous sodium
 sulfate column, and concentration of the dried extracts followed by solvent
 exchange into hexane.  A method blank was extracted with each set of samples
 using the protocol described above.

      After extraction, the samples were analyzed on both a conventional ECD
 (ECDll  in Figure 1) and TH ECD.  The gas chroaatographic parameters utilized
 in the  method validation study are presented in Table 2.  Quantitation of the
 samples was achieved by peak height comparisons against external standards.

 Phase II

      In Phase II, the extracts analyzed in Phase I were cleaned up on
-F3ro?£-silffiLcolumns using the procedure recommended in US EPA Method 608.  The
 6£_diethyl ether in hexane cluate fraction was analyzed under the same
 chromatographic conditions as were used for Phase I.  Prior to beginning this
 phase of the study, the Florisii<& eluticn pattern tor toxaphene was
 established using standard solutions.

 Phase"IIT ~

      During Phase III, I L industrial wastewater samples known to contain
 toxaphene as a contaminant were extracted and analyzed according to US EPA
"Method  608 procedures.  The analyses were conducted prior to and after
 Flori7ltS-Qplu:nn cleanup using both detection systems.

      A total of fo;Tr^£'\xaphene-containir.g wastewater samples were obtained
 for use in the validation^Studjr.  One sample was divided into three 1 L
 aliquots.

      Two of the-ailquots were extracted aritT-aaalyzed as previously described.
 The third aliquot* was fortified with a toxaphene-spiking solution as a method
 recovery"check. -The three remaining samples were each- treated as single
 determinations.

-------
         TABLE 2.  GAS CHROMATOGRAFHIC PARAMETERS FOR THE
                   METHOD VALIDATION STUDY
Coliimn -                168 cm X 1.8 mm I.d. glass-lined stainless
                        steel packed with 1.5% SP 2250/1.95% SP 2401
                        on 100/120 mesh Supelcoport.

Temperature Program -   Isothermal at 2003C.

Carrier Gas -           5% methane/952 Argon at 35 uL/mln.

Purge Gas -             5Z methane/95% Argon at 25 mL/min.

Detector Temperature -  ECD - 340°C.
                        TMEDC - 340°C.

Reactor Temperature -   850°C.

Inlet Temperature -     220°C.
                                  10

-------
                                  SECTION 5

                            RESULTS AND DISCUSSUN.-


     Thermal  stabilities  of the  17 model compounds  were  determined with  tha
detector configuration  phown in  Figure 1.  The  primary objective  of this work
was  to  find the most useful combination of temperature,  carrier S&B, and
reaction tube material  which can be used to differentiae between chlorinated
pesticides, rCBs,  phthalates, and chlorinated hydrocarbon^

RESPONSE RATIOS OBTAINED  WITH ARGON/METHANE CARRIER AND  A GOLD  REACTION  TUBE

     The response  ratios  determined experimentally  at various temperatures
with argon/ire thane carrier and a gold reaction  tube are  presented in Table 3. .
The  ratios ?>hovn are the  average of three determinations.  The  standard  devi-
ations  for vhe determinations are also shown.   These ratios are based on the
assumption that the  response of  ECDtfl and ECDfr'2 are equal for all compounds
with an ambient reactor temperature.  An ambient  temperature for  the reactor,
however, was  impossible to attain due to the temperature of the detectors
(340°C each).  Therefore, equal  response was assumed and a reactor temperd^
ture of 3M)°C was  used  as the minimum temperate-re.

     At 350°C, the response ratios were not equal to 1.0 for all  compounds.
This indicates that  some  rearrangement takes place  either, in the  first
detector or in the reaction tube at 350°C.  The compounds altered the most
were toxaphene, chlordane (both  chlorinated hydrocarbons), and  the
phthalat.es.

     As the temperature of the reactor was increased, the chlorinated
pesticides were degraded  tc species less responsive to the electron capture
detector.  This resulted  in response ratios less  than 1.0.  The response of
ECDfll and ECDfr'2 for  the chlorinated pesticides  at a reactor temperature  of
900°C is shown in  Figure  2.  The most thermally stable chlorinated pesticides
were p,p'-DDE and  p,p'-DDT.  Toxaphene and chlordane, both aliphatic chlori-
nated hydrocarbons, were  found to be very unstable  as shown by  the data  in
.Table 3 and the chromatograms reproduced in Figure  3.

     PCBs.. nitrobenzene,  1,2,4-trichlorcbenzene and 2,4-dichlorophenol
exhibitec1. high thenaal  stability.  The thermal  stability of PCBs  is well
documented.  Therefore, the response ratios were  expected to be close to 1.0
and  show little change  as a function of temperature.  The stability of
nitrobenzene, 1,2,4-trlchlorober.zene and 2,4-dichlorophenol was surprising.
These compounds, particularly nitrobenzene and  2,4-dichlorophenol, are
chemically reactive and were therefore expected to  exhibit thermal
instability.  -
                                      11

-------
             TABLE 3.  RESPONSE RATIOS1 AT VARIOUS TEMPERATURES KITH Ar/CHu CARRIER
                       AND A COLD REACTION TUBE

1.
2.
3.
4.
5.
6.'
?:
8.
9.
10.
11.
12.
13.
•I/..
15.
16.
17.
Compound
Llndane
llvptdclilor
; lleptaclilor Epoxldc
I>,P'-DDI£
'. p,p'-l)DT
.Aroclor 1016 *
Aroclor 1254
Toxaplicne
Clilordane
Dlothyl I'll thai ate
Ulbutyl Piithalate
l>ls(2-EU,ylhexyl)
Phthalate
Methyl Parothioi)
Ltliyl Parathlon
Nitrobenzene
1,2,4-Trlclilorobunzene
2,4-Dlclilovoplienol
Reactor Tumpurattire (°C)
350
0.9310
0.8210
0.8210
0.8UO
0. 7810
0.9810
0.9410
0.6110
0.7210
1.5410
1.6110
1.8410
0.9010
0.9210
l.OCUO
0.8210
1.0710

.01
.01
.01
.01
.01
.02
.06
.06
.01
.30
.24
.39
.01
.01
.02
.01
.01

0.
0.
0.
0.
0.
1.
1.
0.
0.
1.
1.
1.
0.
0.
0.
I.
1.
600
9J10.01
77i0.01
8210.01
8310.01
7810.05
001 0. 01
OU0.01
6410.02
6910.02
5810.10
5510.04
8010.05
8910.01
9110.01
8410.01
0210.02
0410.01
700
0.9010.01
0.4110.06
0.7710.02
0.8210.01
0.5410.03
0.9810.02
1.0110.02
0.0510.03
0.2710.10
1.8210.01
1.5710.04
1.8510.07
0.3110.01
0.2610.01
0.9010.03
0.9710.01
0.9010.05
800
0.1110.
0.05
0.1710.
0.5910.
0.2110.
0.9210.
0/.9210.
<0.01
0.(,)3
3.4810.
1.81.10.
2.04^0.
0.0910.
0.09tO.
0.841,0.
0.9210.
0.8510.
859
01

01
02
03
03
03


04
05
OS
01
01
02
01
03
0.05
0.02
0.0810
0.4810
0.2510
0.9210
0.8910
<0.01
0.01
6.5510
2.5210
2.4410
0.0810
0.0910
0.8510
0.9310
0.7510


.01
.01
.03
.01
.03


.04
.02
.08
.01
.01
.01
.01
.01
900
0.04
9.02
0.05
0.40*0.01
O.lSlv-.Ol
C.('.S10.01
0.7710.04
<0.01
<0.01
8.0910.15
3.1610.02
3.0310.12
0.0710.01
0.0810.01
0.9010.02
0.9110.01
0.7710.03
1Based on equal response of ECD tfl  and  ECD

-------

-------

-------
     Phthalates and the parathions were found to be thermally Bistable.  The
organophosphate pesticides, methyl and ethyl parathion, lere thermally
degraded to products having very little electron affinity.  As a result, the
response ratios were quite small (see Figure 4).  The phthalates were also
very unstable.  However, the products formed during the thermal degradation
of the phthalates had a greater electron affinity than the parent compounds.
Figure 5 Illustrates the threefold enhancement of bls(2-ethylhexyl) phthalate
and the eightfold enlianceoent of the diethyl phthalate at a reaction tempera-
ture of 900er.

RESPONSE RATIOS OBTAINED WITH NITROGEN CARRIER AND A COLD REACTION TUBE

     The response ratios obtained at six reactor temperatures using nitrogen
carrier and a gold reaction tube are shown in Table 4.  In general, the
results with nitrogen as the carrier were similar to those obtained with
argon/methane.  The chlorinated pesticides, the parathions, and the
phthalates were again found to be thermally unstable.  The PCBs, nitro-
benzene, trichlorobenzsne, and dichlorophenol, were found to be more
thermally stable.

     Chlorinated pesticides appear to be less stable with nitrogen carrier
than with au argon/methane carrier.  Significant degradation was observed at
700°C with nitrogen, whereas argon/methane required a temperature of 8CO°C
before significant decomposition was observed.  PCBs also appear less stable
with a nitrogen carrier.  For example, Aroclor 1254 has a response ratio of
0.77 at 900°C with argon/methane and a ratio of 0.29 at 900°C with nitrogen.
This is illustrated by the chroma tog rams presented l:i Figure 6.

     Comparison of the response ratios of phthalates with nitrogen and
argon/methane carriers show contradictory results.  The response ratio of
diethyl phthalatr. is greater in nitrogen than a.r&on/methane at 900°C (14.86
versus 8.09, respectively), but bis(2-ethylhexyl) phthalate has a greater
response ratio at 90f:°C in argon/methane (3.03 versus 1.98 for nitrogen).

     The remaining compounds, the parathions, nitrobenzene, 1,2,4-trichloro-
benzene, and 2,4-dichlorophenol, displayed similar response ratios in
nitrogen and argon/methane.  It is interesting to note that the response
ratio of nitrobenzene, gradually decreases as the temperature approaches
850°C, but then increase at a reactor temperature of 900°C for both nitrogen
and argon/methane carriers.

RESPONSE RATIOS OBTAINED WITH HYDROGEN/HELIUM AND A GOLD REACTION TUBE

     The response ratios for the helium/hydrogen gas composition were
obtained by using a carrier of 3CmL/min helium £nd a make-up gas of 30 mL/oin
hydrogen.  The make-up gas was adcisd to the column effluent immediately prior
to entering the first detector.

     Response  ratios obtained with a helium/hydrogen carrier gas did not
yield reproducible results as indicated by the data presented in Table 5.
This table shows the response ratios obtained for the chlorinated pesticides
at 350°C with different equilibration times.  The phthalates also exhibited
poor reproducibiJiricc.
                                     15

-------
                                                             B
                                               I
       V
OJ
C
O
o.
CO

&
    FIGURE  4.
Chromatograca of ParatHons with Ar/CH*  Carrier and Gold
Reaction Tube.
    Order of elution:   Methyl Parathior. and Ethyl Parathion


    Chronatogram A, ECt) 11, 500 pg each. Attenuation X10.

    Chromatogran B, ECU 92, 500 pg each after passing through a gold  reactor
                    at  900°C, Attenuation X2.
                                      16

-------
 a.

 §
'CU
 ca
       O-
       JC.1
                                            o
                                            V)
     f •/

                                                                   B
                                           Time


     FIGURE   5.   Chromategrams of ?hthalates with A.r/CH* Carrier and Gold

                  Reaction Tube.



     Order of elution:  ^Diethyl, Dibutyl,  and bis(2-Ethylhexyl) Phthalate.


     ChromatosjrJm A, ECD #1, 5 ng of each. Attenuation X10.


     Ghromatogram B, ECD #2, 5 ng of each after passing through a gold reactor

                     at 900°C, Attenuation X20.
                                        17

-------
    TABLE 4. ' RESPONSE  RATIOS1  AT VARIOUS TEMPERATURES WlTH^NITROGEN CARRIER AND A'
              C'0(,D  REACTION  TUBE
•a K c: «w«s

L.
2.
'3.
4.
i.
h.
7.

8.
9.
io.
11.
12.

13.
14.
15.
16.
17.
a. »j=»^j^aa- «a^j-j-^=a=a..,-yia.-»._.j.i^ j j A
Compound
L Ind Jim
llcptaclilor '
lleptaclilor Cpoxtde
p.p'-DDli
p.p'-DDT
Aroclor 1016
Aroclor 1254

Toxaphune
(Milordane
Uieihyl Plulialate
Dibutyl Phtlmlate
bis(2-Ethylhoxyl)
I'lilhalate
Methyl Parathlon
Ethyl ParatliJon
Nitrobenzene
1 , 2 , 4-Tr Icli lorobenzene
2,4-DJchlorophcnol
Keaetiir Tempera cure (°C)
350 .
0*9310.01
0.4710.03
0.7610.01
0.7510.01
0.5110.01
0.9710.01
0.95A0.04

0.6110.01
0.7310.01
2.0910.09
1.6410.08
s2. 0710. 12
.
0.8610^.01
0.8810.01
0.7810.01
0.9810.01
l.U.'»10.04
600
0.9110.01
0.3810.01
0.7310.01
0.7310. 01 «
0.4410.04
0.9510.01
0.8910.03

0.5110.02
0.7110.01
2.0910.07
1.7810.06
1.9310.07

0.6810.02
0.00 10. 02
0.8410.01
0.99*0.01
0.9810.03
700
0.12^0.01
0.041
0.1610.01
0.4910.01
0.1810.0^'
o.83to. 01 ;.
0.7310.02 '

0.0910.01
0.03
2.8510.23
1.8510.14
2.2010.02

0.1710.01
0.1810.01
0.9210.03
0.9810.02
0.6910.03
'800
0.03
0.02
0.05
0.1710.01
0.0610.02
0.7210.02
0.4410.02

0.04
0.01 '
5.7810.74
1.9410.17
1.8810.32

0.1010.01
0.1210.01
0.8810.02
0.9210.03
0.5710.03
850
0.02
0.01
0.02
0. 1010. 01
0.0510.01
0.6310.01
0.34±0.03

0.0?
<0.01
6.9610.15
2.1710.28
1.8810.20

0.0710.01
0.09:4.01
0.8810.01
0.9010.01
0.6010.02
1
900
0.02
0.01
0.02
0.0910.01
0.03
0.6010.01
0.2910.01
i
0.02
<0.01
14.8610.72
3.0210.57
1.9810.14

0.0710.01
0.0810.01
0.9710.02
0.9010.01
0.7210.02
on equal response of BCD ffl and LCD

-------
0*
CO
C
o
a*
V)
CJ
PS
                                                                B
                                                I
                                                 o
                                                CO
                                       Time


    FIGURE  6.    Chromatograms  of Aroclor 1254 with Nitrogen  Carrier and Gold
                 Reaction Tube.


    Clircmatogram A, ECD #1, 5.2 ng of Aroclor 1254, Attenuation X10.

    Chromatograin B, ECS #2, 5.2 ng of Aroclor 1254 after  passing through a gold
                    reactor at 900°C, Attenuation X10.
                                        19

-------
-'TABLE 5.  RESPONSE RATIOS OF CHLORINATED PESTICIDES WITH HELIUM/
           HYDROGEN CARRIER AND A GOLD REACTION TUBE AT 350°C

                    RESPONSE RATIOS  (ECD 02/ECD
Comoound
Llndane
Heptachlor
Heptachlor Epoxide
p.p'-DDE
p,?'-DDT
Held at 350°C
for 1 Day
0.14
0.30
0.39
0.80
0.23
Held at 350°C
for 2 Davs
0.90
0.43
0.88
C.81
0.45
Reactor Heated to 900°C
and Cooled to 350 9C
0.01
0.08
0.09
0.31
0.07

-------
     Reproducible results were obtained for the PCBs, 2,4-dichlorophenol, and
l»2,4-trichlorobenzene.  Toxaphene, chlordane, nitrobenzene and the para-
thions had very low response ratios at all temperatures, as indicated in
Table 6, and were not studied in great detail.

     Heating the reactor to temperatures above 800°C caused an apparent
"activation" of the reaction tube.  The reactor would require several days to
return to its original level of activation once the temperature had been
reduced.  This "activation" caased very poor reproducibility, especially for
the chlorinated pesticides and phthalates.  The reactor "activation" was
postulated to be due to a temperature dependent reaction between the hydrogen
and some substance coating either the inside of the ECDs or the reaction
tube.  Removal of the material would then leave an active surface which could
degrade compounds until a new "coating" was built up at lower temperatures by
column bleed or reaction products.  Regardless of the-cause, this problem
makes helium/hydrogen an unlikely choice for the carrier gas except for the
thermally stable species.

     Detector linearity with a helium/hydrogen gas composition was determined
with the geld reaction tube at 600°C using standards spanning three orders of
magnitude.  The observed detector response was not linear over the complete
range of concentrations.  Similar results were obtained with dibutyl
phthalate and p,p'-DDE.  In addition, the response ratios determined were not
consistent over three orders of magnitude for these compounds.  The non-
linear response observed is not surprising since the detector's relative
pulss width was not adjusted for this gas composition.
   /
RESPONSE RATIOS OBTAINED WITH HYDROGEN/HELIUM AND A NICKEL REACTION TUBE

     Replacement of the gold reaction tube with a nickel tube did not
alleviate the problems observed with gold.  All compounds, except diethyl
phthalate, were completely destroyed at a reactor temperature of 900CC.   The
"activation" of the reactor was again observed at high -emperatures.  After
reducing the reactor temperature, several days were required for restoration
of the original activity level.

     The temperature at which the reaction tube was "activated" was deter-
mined to be approximately 860°C.  Below 860°C, most of the compounds
responded on ECD#2.   Above 860°C, all of the analytes except nitrobenzene and
diethyl phthalate were destroyed.  The  poor reproducibility and lack of
discrimination in destroying the various compounds makes helium/hydrogen a
poor gas composition for use with a nickel reaction tube.

RESPONSE RATIOS OBTAINED WITH NITROGEN CARRIER AND A NICKEL REACTION TUBE

     Response ratios determined with a nitrogen carrier and a nickel reaction
tube at six different temperatures are shown in Table 7.  When compared  to
the results of both argon/methane and nitrogen with a gold reaction tube,  the
response ratios of the nickel tube with a nitrogen carrier were found to be
lower at 350°C for every compound except the phthalates and nitrobenzene.
The phthalates were observed earlier to have larger response ratios due  to
the formation of a degradation product with a greater electron affinity.   The
increased degradation of all compounds at 350°C with-nitrogen/nickel sug-
gested that catalytic reactions were occurring (i.e.,  the  gold tube is more


                                    21

-------
    TABLE 6.   RESPONSE RATIOS OBTAINED WITH HELIUM/HYDROGEN CARRIER
              AMD A COLD REACTION TIIBF AT FOUR TEMPERATURES
Teroeracure (°C)

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Compound
Lindane
Heptachlor
Ueptachlor Epoxide
p.p'-DDE
p,p'-DDT
Aroelor 1254
Aroclor 1016
Toxaphene
Chiordane
Diech/1 Phthalate
Dibutyl; Phchalate
bis<2-Ethylhexyl) Phthalate
Methyl Parathion
Ethyl Parathion
Nitrobenzene
1,2, 4-Trichloroben zane
2,4-Dichlorophenol
- 350'
0.14
0.30
0.3S
0.80
0.23
0..75
0.76
0.12
0.33
3.76
7.01
2.26
0
0
0.08
0.92
0.76
600 	
<0.01
0.02
0.02
0.18
0.09
"0™. >6 .
0.75
0.04
0.02
7.03
2.7,4
2.05
0
0
0
0.96
0.75
800
<0.01
<0.01
0.01
0.08
0.07
0.57
0.76"
<0.01
<0.01
_J.2.86"
2.84
2.18
0
0
0
0.91
0.61
9UO
0
0.02
0.01
0.09
0.07
0.77
0.77
0
0
20.29
3.93
2.36
0
0
0
1.00
0.64
1Reactor held at 350°C for one day prior to analyses.
                                      22

-------
                TABLE  7.   RESPONSE RATIOS AT VARIOUS TEMPERATURES'WITH A NITROGEN CARRIER AND
M
                           NICKEL REACTION TUBE
                                                          Ruatitor Tcnruraturu  (°C)
                                                               tni\   i      onn    "
	Compound	       3r>0         'fi'lO         70?)	      800         850      	9(10.	
    	" * ' •—••                                  f                                                     -i
 1.  Llndanu                   0.8510.01    0.84*0.01   0.66+0.05',  0.35+0.04   0.03            0
 2.  llupcuclilor                0.6110.01    0.^310.01   0.1710.02   0.09*0.01   0.02            0;
 I.  lluptarhlor Exoxlde        0.6110.01    0.5BJ0.01   0.3310.03 \ 0.1610.01   0.03  ,          0'
 4.  p.p'-DDE                  0.5410.UI   ;0.i3iO.()l   0.4510.01   0.36i0.01   0.17+0.01   0.02  •
 5.  p,|>"-l)l)T                  0.5Ji0.01    0.4810.02   0.2410.01   0.2010.02   O.lOlO.Ol       0
 6.  Aroclor 1016             0.7:»iO.Ol/  0.7410.01   0.7110.01   Q.63-«0.01   0.05        0.0.1
 7.  Aroclor 1254             0.7iiO.O?    0.6510.00   0.60i0.02   O'.44i0.02   0.1210.01   0.01
 8.  Toxaphone                 0.47J0.02    0.4710.02   0.3510.02   0.1210.04   0.0)            0
 9.  Chlorddiic                 O.n3'0l0l    O.VJ10.01   0.49+0.01   0.1810.02   0.01            0
10.  Dicthyl Phtlialate         2.05i6.03    2.4610.12   1.8410.29  10.0011.64   0.3510.01   0.1210.01
11.  Dlbucyl I-liLliulale         2.1bJ0.02    2.6710.03   2.95+0.10   3.52+0.07   0.1U0.01   0.03
12.  bis(2-Erhylliexyl)         3.5910.15    4.3610.17   4.38t0.17   4.1810.19   2.0310.29   0.4010.05-
     riithuluic
13.  Methyl  Purathion          '0.7210.01    0.69J0.01   0.57i0.03   0.02'.           0           0
14.  Ethyl Paratliion           0.74i0.01    0.69JG.OI   0.4510.01  <0.01 '           0           0
15.  Nitrobenzene   •         1.4910.02    1.4910.02   1.3110.01       000
lb.  1.2.4-Trlclilcrobtinzcno   0.7510.01    0.7310.01   0.7110.01   0.05+0.01   0.01            0
17.  2,4-Uicliloropliuiml        0.7110.02    0.65i0.04   0.5)10.04   0.04        0.02        0.01

-------
inert Chan the nickel Li be).  The nickel tube displayed another difference
from the gold tube with nitrogen.  At temperatures >900°C all compounds were
degraded including the normally thexnally stable PCBs, chloroaromatlcs and
nitroaromatics.  In ad lit ton, the response ratios of the phthalates were
smaller suggesting that the intermediate species formed by  the thermal
degradation of tiie phLhalates were degraded further to species which have
small electron  ifiinities.  Due to the.se difficulties, this system was not
investigated further.

RESPONSE RATIOS OBTAINED WITH ARGON/METHANC CARRIER AND A NICKEL REACTION
TUBE

     Response ratios obtained using an ar^on/cethane carrier and a nickel
reaction tube at six different temperatures aru shown in Table 8.  At 3SO°C,
Che response characteristics for the argon/methane carrier  are very similar
to those observed for a nitrogen carrier when a nickel tube is employed as
the reactor.  The reactivity of the nickel tube appears to  increase more
rapidly with a corresponding Increase in temperature when argon/ex* thane is
used in place of nitrogen as the carrier.

     This trend is evident when the values at a reaction temperature of 8tiOcC"
are compared for the two systems.  The response ratios of all the compounds  -
except the phthalates are extremely small for the argon/methane system at
800°C whereas ~he response ratios of the chlorinated compounds and the
phthalates resulting from the nitrogen system are significantly greater at
the same temp'srature.  The response ratios for the organophosphate pesti-
cides, the chloroaromatics, and the nitroaromatics are very similar for both
systems at this temperature.

     In the case of the chlorinated pesticides, these data  cay be indicative
of the occurrence of free radical formation in the argon/methane atmosphere
within the aickel reactor followed by subsequent recombination reactions in
which-species are-formed that have lower electron affinities.  The nickel
tube appeals to act as a catalyst under these conditions since the test
compounds exhibit much greater thermal stability when argon/methane is used
in conjunction with a gold reaction tube.  Because of these characteristics,
this syst-*m was not selected for use in the validation study.

METHOD VALIDATION STUDY

System Linearity Check

     In order to determine whether the prediction equation  fits the
five-point calibration data, measures of the adequacy of the fit were
examined.  In particular, residual plots were constructed and used as
graphical aids in visually inspecting the fit of the prediction equati'on.

     The plots of amount (ng) versus the response (mm) for  both the commer-
cial ECB and the TK ECD presented in Figures 7 and 8 respectively, indicate
that a linear a<>£o?iatlon between the two variables Is the  dominant
relationship.  Because of the minimum nvnber of points, it  is difficult to
say whether other components, peihcps a quadratic term, should be added to
the prediction equation.

-------
                 TABLE 8.  RESPONSE RATIOS AT VARIOUS TEMPERATURES WITH Ar/CH.,  CARRIER AND
                           A NICKEL REACTION TUBE
10
Reactor Temperature °C

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Compound
Llndane
Haptachlor
HepCnchlor Epoxide
p,p'-DDE
p,p'-DDT
Arocloi 1016
Aroclor 1254
Toxaphene
Chlordane
Diethyl Phthalate
Dlbutyl Phthalate
bis(2-Ethylhexyl)
Phthalate
Methyl Parathion
Ethyl Parathion
Nitrobenzene
1,2,4-Trichlorobenzene
2 , 4-Dichlorophenol
350
0.84+0.01
0.65;:0.01
0.62+0.01
0.5 /tO. 01
0.53+0.02
0.79+0.01
0.85+0.02
o.4:»+o.n
0.6(1+0.02
2.29+0.09
2.15+0.03
3.49+0.27
0.80+0.01
0.85+0.03
1.64+O.OJ.
0.79+0.0?.
0.79+0.0-4
600
0.82+0.01
0.52+0.01
0.59+0.01
0.57+0.01
0.46+0.02
0.74+0.01
0.67+0.08
0.34+0.0*
0.49+0.0)
2.67+0.41
2. 35 H), 08
3.54+0.60
O.;8f_0.04
0.51+0.03
1.2040.01
'0.76^0.01
0.74+0.01
700
0.73+0.01
0.34+0.01
0.48+0.01
0.50+0.01
0.37+0.01
0.. 32+0. 05
i ^
0.43+0.10
0.06+0.03
(),25K).02
1 -6.73+P.38
', 2. 78+0, 14
.3.85+0.29.
A. 08+0. 01
10.11+0.02
1 0.02
0.46+0.03
p.d6_H).oi
800
0.01
<0.01
0.01
0.01
0.03
0.06+0.01
0.04+0.01
<0.01
0.04+0.02
2.56+1.90
1.06+1. OR
2.37+0.78
0.02
0
0
0.14+0.02
0.06+0.02
850
0
0
0
0.01
0
0.07+0.01
0.05+0.01
0
0
0.97+0.10
0.13J.0.01
0.09+0.02
0
0
C
0.12+0.02
Q. 05

900
0
0
0
0
0
0.02
0.01
0
0
0.31+0.20
0.03
0
0
0
0
0.07+0.01
0.02

-------
10        M
o^        ™
                 I
            1250 *
                 I
                 1
            i:co *
                 i
         R       I
         E       I
        . S       1
        ' P  750 *
         '"       I
         t)       I
         S.       I
         r       i
            500 *
         M       I
                 I

                 I
            250  *                   X
                 I

               *  '
               .  I          V
                 I      X
               0  *
                                   -4*--4«— .4-— .4---4 — 4 — 4---4— «4— «-4 — -* — 4 — 4— »4-

                                   2P  ?«   ?0  IF   40  *5   5C  55   60  6F.   7f  7F   PC  «"5
                                    Figure; 7.  Linearity Plot - Commercitl\ BCD
                                                                        •,   \

-------
        I
  1250  *
        I
        I
        I
        I
  1000  »
        1
R       *
r       l
s       '
P ^750  +
0       f
N       I
S       I
r     t  I
   50J  •»
•»       I
M       IN

     ,   I
   ?fO  •»
        I
        r         x
        i  .   *
     o  *

         0   5  1,"  IF   ?0  2?  30  3F   ^0  45  5fl   5S   69  6F   7S   75  flO

                                      AMOUNT (NO

                            Figure 8.  Linearity Plot - TM BCD

-------
     In Investigating the possible need to include a quadratic terra or some
other component in the prediction equation, residuals of the' linear fit were
plotted against the amount (ng) for each of the two ECOs.  The results did
not indicate a need for some transformation of the predictor variable as the
residuals appear to be randomly centered around the value zero.

    "As a further check on the adequacy of the linear '^t, a quadratic tetm
was assumed as part of the true underlying codel:

                  Response • a •+• bi (amount) + b2 (amount}2 + e

where a, bi, and 02 are the unknown model parameters and e is the error
tetm.

     The least-squares coefficient estimates exhibited below indicate that
amount is the most influential determinant of the response for both
detectors.

                                Intercept           Coefficient Estimates

        Detector                    a               f>\                €2


     Commercial ECD               -27.5            12.8              0.01
         \
     TH ECD                       -34.5            13.3              0.005

     The hypothesis that 02 « 0 was tested at 0.25 significance level (a
very conservative test) with an individual t-test for each of the ECDs.
-Since the significance probabilities were greater than 0.25, the hypothesis
of a zero 'regression coefficient (02) was not rejected.

     It appears from the above analyses that the linear  fit  is appropriate
for both the commercial ECD and the TH ECD.  In addition, there does not
appear to be a statistically significant difference in the linear reRpor.se
characteristics of the two detector? for toxaphene over  the  range of 5 ng to
80 ng.

Phases I and II

     A phased approach was utilized in the method validation study in order
to determine the effects of Florisll® column cleanup and sample matrix on the
precision and accuracy %>f the analytical methodologies.   The sample fortifi-
cation and  recovery data for Phase I of the method validation study are
presented in Table 9.  It should be noted that multiple  data entries at a
given concentration level are indicative of replicate analyses.

     In addition, the data listed under the "ECD" category In Table 9 are the
result of the chromatograms generated from the response  of ECDtfl (Figure 1).
The data tabulated under the "TMECD" heading were obtained from the chromato-
grams generated by the differential amplifier.  This device  subtracts the
signal produced oy ECD02 from the signal generated by ECDffl.  With this con-
figuration, compounds with low  response ratios «1.0) yield  positive peaks on
the chromatogram while compounds with high response  ratios (>1.0) produce


                                     28

-------
TABLE 9.  METHOD VALIDATION STUDY:
          PHASES I AND II
TOXAPHSNE SPIKING AND RECOVERY DATA FOR
Analysed Concentration (UR/L)
Phase I (Pre-Cleanup)
Youden Pair
1
2
3
4
5
b
7
8
9
10
11
12
13
14
15
16
17 v
18
19
20

True Value (ug/L)
9.8
12.5
i5.5
17.3
19.6
25.0
24.5
30.2
33.0
38.0
50.3
57.3
66.0
76.0
76.1
86.3
80.6
99.5
100.6
114.6

ECD
7.8, 7.5-
11.1
11.5
14.8
17.3
16.0
2J.O
30.0
33.6. 33.0
38.7
44.5
52.8
63.6
74.0
75.9
85.6
68. C
87.0
85.0
95.0, 100.
97.7
1MECD
7.8, 8.0
11.0
11.2
14.9
17.9
16.2
22.6
31.9
31.7, 31.0
35.6
47.8
57.6
67.1
77.8
82.0
90.0
74.0
93.7
87.8
103. 106.
103
Phase II (Post-Cleanup)
ECD
7.6
9.6
11.6
13.0
11.3
15.6
20.0
26.0
27.2
3-'.. 5
40.0
50.2
55.7
64.5
70.5
84.9
65.0, 65.0
84.0
80.0
91.0, 91.9

TMECD /
3
7.5
9.4
11.0
12.7
11.4
15.0
21.4
26.0
27.0
33.3
40.0
51.6
60.0
67.0
73.0
83.5
76.0, 73.0
SI. 7
84.8
98.0, 96.7


-------
negative chroaatographic peaks.  Under the conditions  used  In this  study,
toxaphene pvoducec positive peaks on the differential  chronatogram  while  the
phthalate esters exhibited a negative response.

     For the recovery Jata presented in Table 9,  the standard deviation
appeared to depend on the toxaphene concentrarlon, with constant proportion
over ell mean con-rx-nt rat ions.  Therefore, the percent  relative standard
deviation was judged to be the nost useful measure of  variation.

     A statistical sunnary of percent recovery of toxaphene  spiked  in  reagent
water is presented in Table 10 for both the commercial FCD and the  TMECD.
The intralaboratory percent relative standard deviation (precision  estimate)
which indicates che repeatability of an Individual laboratory or analyst  when
using the method i.s computed by:,-''

                                       Sr
                      RSD	x 100
                            seat percent recovery


where the single analyst standard deviation, Sr, was calculated from the
Youden pair data and deflr.cc by Youden6 as
            -D)2
    \/	-
S,
                                   /E(Di -
                              Jr -V  2(n -
where n  •" cumber of paired observations,

      D£ • the difference In percent recovery between observations for a
           sample palr,~-

      D  - the average value Cor Dj.

     A comparison of the two detectors in terms of accuracy and precision  for
both the pre- and post-cleanup aethods can he made using the statistical
suanary given In Table 10.  A graphical presentation of the accuracy and
precision estimates are presented In Figures 9 and 10 respectively.  It Is
apparent that the commercial ECD and the TM ECD work equally well for the'"
analysis of toxaphene in fortified reagent water samples.  This is.trfie for
both pre--aod post-cleanup methods.

     Additionally, it is observed that while the variability of recoveries
using the post-cleanup method is significantly, smaller than that for the pre-
cleanup method, recoveries using the post-cleanup method are consistently
lower.

     Figures 11 through 14 siaphically summarize the recoveries for the two
types of detectors and-cleanup cethods used.  The concentration level does
not appear_tp af/cct the recovery of toxaphene.

     A paired comparisons t-test~-*a«-._run to statistically test for dif-
ferences between the two detectors and the pre-_and post-cleanup recoveries.
The results of this test were<"*as foiiO«*s:
                                      30

-------
          TABLE 10.  STATISTICAL SUMMARY:  ACCURACY AND PRECISION  ESTIMATES OF:PERCENT

                     RECOVERY .FOR TOXAPHENfe SPIKES IN REAGENT WATlvR
Method
Pre-cleanup
Commercial ECD
TM ECD
Post-cleanup
Commercial ECD
x TM ECD
Mean
Estimate
89.1
92.2
80. '8
82.7
Pc-.rclpnt Recovery
Approximate
95% Confidence Interval
\
i •
(83, 95)
(86, 93)
i(79, 83)
(80, 85)
% Relative
Estimate
fe!e
3.0
,3.5
Standard Deviation (RSD)
Approximate
95% Ccu fide rice 'Interval
(5.7, 15)
(6.1, 16)
\
(2.0, 5.4)
(2.4, 6.4)
!„ \     _i          Measured  concentration ., . j,
 Percent Becovery •= —~	—:	X-10(
    r      i     '     True concentrations.       '
              Measured concentration

                                 ""'\

    :   sr                                 _.-.._    ,   .  ••

rrN—sr-r	X 100 where S,. is the intralaboratoiry standard deviation
Mean >i Recovery              r     .  .    n

       •I         .  :                    .                     '

-------
U)
1 VV

P 95
E
R
C
F 90
L-
N
T
85
R
E
C
0 .88
V
5 75
Y

-._

•
i™"
i 	 1


1
_
e

N*
_


'• «


i

i 	
\
1

\
\
i


.




_i


( 	 ,

i • 	 i
.
-




EC
D ' TM
ECD

f V L_^
:
>

i
1 •"*' '
ECD TM ECD
1

Figure 9. Accuracy Estimates: Phases I and II (Estimate <> 95% Confide

-------
w
w
1 /

E 15
E
L
A 13
T
I .,
V H
E
9
S
T
D 7
E 5
V
3



r
f-


-


m





i_
•
-i








i

i_
"1







i

1

_i
_i


El
-
:D TM E
:CD

•
i











'

1 1
i
ECD TM ECD

                   Figure 10.   Precision Estimates: ^.,'phape I and II  (Estimate
35% Confidence Interval)

-------


\100




1' 'JO
t
c
r
N fiO
T

h
r.
C 7C
^J _
^ L
V
t
K
y 60
i



' 50


\
\ i; L«
*••*****•***••**•«»** [<*«»*»***«***«***«»**<«|, »***>]*••***••••••*•»**•*•••
\ I' A
\ H
\
\ h A
•• s • A
\ I M ti A U
\ n T. A
\ \ . A A H A
\ > A A 11
* U N A A A A
t
\ A A '
\ A
\
*
\
^ B » Pre-cleanup
\ n
^ * x A « Post-cleanup
* "
\ A
•J- v
V l N
\
•1 . \
\ \
\ \
I'   .10   2u    30    <<0    b(l    t.O     7U    t!0N   
-------

ion

•
!•
F '•
i. mj
c
r
M
T
HO
B
t
C
w °
u- V 70
L
l«
Y

60



. ,0
* , ''
\ ('
\ "

V
\ H H / A
\ 1- A L
\ h I' !• A
+ A A o
\ ,- A A A L1
\ h A A
\ A
\ *
* b A
\
\ AA
\ A
'A • Pre-cleanup
••A •
* B « Post ^cleanup
\
\ »
\
+ A
\ A
\
\
\ '
10    ?o    30    <«n   "so    i-o    7o .  fcii    -
-------
      \

  ico «* *...»•,•*..•***••**•••••*••••••••«•»••••••••"•*•••••••••••••••••*•*•••
      \                                                  • L      i
      X                                              T     T


                                                T     r             '
?  90 ;                  T       i          i         i

P     N                     L                   LCI
C     \                                         L     C             »  ,

r     N                  c    L'                      '   ,           r       r
r;  so *                                *                L          t,      C
T     \          t
      \          U   C
K     \              CT
r     \
C  70 *              T
C      V                                               c . Comaercial ECD
V     \                                                     '
C     N                                                T - TM ECD
K      N                 E
Y.  (>0  «                 T
 \     \               »-
       \
       \
       \                               ^
   •JO  «
       \
       N                                               '        .,.«.._.»,	
       "     *,"""lo"""iO    il    ''0   5iO    tO    70    I-C    SO    IOC   110   1*0

                                UUL LONCF.MKA1 ION  (i'l !•)
                 Figure 13.  Percent Recovery Plot - Pre-Cleanup

-------
\
\
\
V
[
P
C
f
K.
T

P
I
c
f
V
C
K
V
    \
 9'J *
    \
    V
.    \
    \
 HO +
    \
    \
    v
    \
 70 *
    \
    \
    \
    \
 60 *
    \
    \
    \
    \
                                                                 ,
                                                                 C
                                                      E - Conunercial  ECD

                                                      T • TM ECD
                  lc    2-0    JO    10
                                       SU
                                         (.6
                                                      7.'p    » C
                                                               ^0
                                                                 IOC-   110   i:-0
                            V-l'L  COUCCMRA1 lOJ-  It Phi
             Figure  14.  Percent Recovery Plot -\Post-Cleanup

-------
     •    [The commercial ECD and  the TM  ECD are  significantly  different
           (at the 0.05 significance level) with  respect  Co  percent
           recovery  for the precleanup method—TM ECD  recoveries-are
           closer to the true value; and

     •     The pre-cleanup and post-cleanup methods are significantly
           different (at the 0,05  significance level)  with  respect to
           percent recovery for* both ECD-detectors—pre-cleanup recovery
           results are closer to true value.

Phase III

     The purpose of this phase of the method validation  study  was ro
determine  the effect of sample matrix on the precision and  accuracy  of the
analytical methodologies.  Four Industrial waste effluents  known to  contain
Coxaphene  were ajialyzed using both detection systems,.prior  to  and after
cleanup.   One 'sample was of sufficient volume to be-divided into three
aliquots for a check on analytical precision.  One of these allquots was
fortified  with a toxaphene spiking solution at a level which would yield an
equivalent sample concentration of 86..3  pg/L-  The quantitation of the
samples was achieved by peak heigr.t comparison against external standards.
For all samples except X-2, a minimum of three peaks  were used to quantitate
the toxaphene concentration. ' The level  of phthalate  interference prior  to
cleanup was so great for sample X-2 that the quantitation was  based  on a
single peak in the chromatogram.   Figures 15 and 16 are  illustrative of  the
differential chromatograms obtained for  a sample containing only toxaphene
and sample X-2 containing both toxaphene and phthalates  respectively.  The
differential chromatogram of cample X-2  after cleanup is presented in Figure
17.  It snould be noted that the  same peaks v.-ere used in the quantitation of
the toxaphene concentrations for  both detection  systems.  The  analytical data
for the industrial waste  samples are summarized in Table  11.

     A statistical  summary of recovery precision for  toxaphene in effluent
samples is given in Table 12 for  the two detectors.   As  was illustrated  for
clean samples, the TM ECD appears to be  equivalent to the  commercial ECD in
this particular application.  Figure 18  Is a graphical presentation  of the
precision  estimates for toxaphene in waste effluents.  The  method detection
limit for  this type of sample is  estimated to be 5 pg/T"  The  results of the
method recovery check listed In Table 13 indicate quantitative recovery  of
toxaphene  from the  fortified industrial  waste sample.

Phthalic Anhydride Experiment

     An additional series of analyses using the  TM ECD vere performed in an
attempt to identify the mechanism responsible for the enhanced phthalats
response.  It was hypothesized that phthalate esters  were being converted to
phthalic anhydride within the heated reactor and that the enhanced signal was
the result of the relatively greater electron affinity of the  anhydride.  To
test this  hypothesis, a 5 ug/pL phthalic anhydride in acetone  solution was
analyzed on the 110 GC equipped with the TM ECD.  The analysis was performed
at two different reactor temperatures:   650°C and 850°C.  The  response ratio
for this compound was calculated  and found to be 0.8? for both reactor
temperatures.  This indicates that the phthalic  anhydride is thermally stable
•under the conditions employed.  These conditions included a gold reactor tube


                                    38

-------
01
§ 3
 -
OS
                                       Time
                Figure 15.   Differential Chromatogram of a Toxaphene-
                            Contaminated Industrial Waste (128  pg/L)
                                         39

-------
a,
c
                                            Time
                                                             I
            Figure 16.  Differential Chromatogram  of  Sample X-2  Prior to
                        Cleanup  (43.2  pg/L)
                                            40

-------
I
.1
I
                                                Time
                   Figure 17.  Differential  Chromatogram of Sample X-2 After

                               Cleanup  (AC.7 ug/L)
                                               41

-------
TABLE 11.  TU.XATHESE LEVELS IN INDUSTRIAL WASTE EFFLUENT  „
Concentration (ug/L)
Sample ID
X-1A
X-1B
X-1C
X-2
X-3
X-4
Spiking Pre-Cleanup
Level (Ug/L) ECD TKECD
123
112
126
86.3 235
43.2
158
35.0
33.4
108
105
123
195
63.2
112
38.3
37.3
Post-Cleanup
ECD THE CD
115
121
194
199
39.9
147
150
39.0
111
106 "
184
182
40.7
130
127
38.4

-------
  TABLE 12.   STATISTICAL  SUMMARY:  PRECISION ESTIMATES OF RECOVERY
      I       FOR TOXAPHENE  IN WASTE EFFLUENT SAMPLES

                                                  Intralaboratory
                                          % Relative Standard Deviation

                                                        Approximate
	Method	Estimate 95%  Confidence Limit

 Pre-Cleanup
   Commercial ECO                  .        6.8           (3.7,  26)
   TM BCD                                 12.7           (7.2,  47)

 Post-Cleanup
   Commercial ECO                          1.8           (1.0.  6.8)
   TM ECD                                  1.8           (1.0,  6.8)

-------
   50

R  45
E
L  40
A
T  35
I
V  30


s25
D15

V  10

%   5
 I
ECD
    I
TM  ECD
                                                      ECD
                                                     TM ECD
                ^ Prc-Cleanup
                                         Post-Cleanup -  - - - -
Figure  18.  Precision Estimates for Taxaphene
  1         Determinations in Waste Effluents (Estimate
                                                                      952 Confidence Interval)

-------
                      TABLE 13.   METHOD RECOVERY
                                                Soriceritration (
" _-• - - => Pro-Cleanup
             Parameter
                                                              .  Post-Cleanup
                                                                     '"TMECD
 Average toxaphene cone. (Samples
-X1-A and Xl-B)
 Toxaphene spiking level
 Predicted toxaphene
 (yg/L)         -""  .
 Analyzed toxaphene concentration
 Olg/L)

 Z Toxaphene recovery
                                     _86.
235   195

ItS     (8
                              i9.e-- 183

                              .95     94

-------
 and Ar/CH4 carrier gas.'  An attempt was made to compare the pose-reactor
 response of an eqjji'valent amount of di-n-butyl phthalate to that of the
 phthallc anhydride.

      The phthalic anhydride chromatographic peak tailed so badly, however,
 that it was impossible to get an accurate comparison.  Consequently, It was
 not possible to conclusively identify the formation of phthalic anhydride as
 the'underlying mechanism for the phthalate ester response characteristics.
 Nevertheless, it was found that phthalic anhydride was more sensitive than
-the phthalate ester.  The difference between the sensitivities of the
 anhydride and the ester was approximately the same as the elevated response
 observed for the"^srer after passing through the reactor at S50°C.  Thus, the
 formation of phthalic anhydride is a plausible" .reaction mechanism.

-------
                                  REFERENCES
1.   Aue, V.A. , Detectors for Use ic GC Analysis of Pesticides.  J.
     Chromatogr. Sci. , 13:  329-333, 1975.

2.   Budde, W.L. and J.W. Eichelberger, The Mass Spe&nrog^ter as a
     Substance-Selective Detector in Chroma tog raphy.  J. Chroma cog r. 134:
     147-158, 1977.
                                                           *3r
3.   Aue, V.A. , and S. Kapila, The Electron Capture Detector— Controversies ,
     Comments, and Chromatograms.  J. Chromatogr. Sci. 11:  255-263, 1973.

4.   -David, D.J. , Gals Chroma tog raphic Detectors.  John Wiley & Sons, Inc..,
     NY, 1974.  295 pp.  """" — —-^

             Register. Vol. 44, No. 233, Monday, December 3, 1979, p. 69501.
          n, W.J. , "Statistical Technique for Collaborative Tests,"
         elation of Official Analytical Chemists, Washington, DC, 1967.
                                    "47

-------