United States
Environmental Protection
Agency
Off ice of
Research and Development
Washington DC 2046O
EPA/600/2-81/059
April 1 981
Manual of
Analytical Quality
Control for
Pesticides and
Related Compounds

In Human and
Environmental
Samples  — Second
Revision

-------

-------
                                             EPA 600/2-81-059
                                             April 1981
WNUAL OF ANALYTICAL QUALITY CONTROL FOR PESTICIDES
               AND RELATED CQTOUNDS
      IN HUMAN AND ENVIRONMENTAL SAMPLES
  A  Compendium  of  Systematic  Procedures  Designed
     To  Assist in. the  Prevention  and  Control  of
              Analytical  Problems
                        By

                 Dr.  Joseph Sherma
               Department of Chemistry
                 Lafayette College
               Easton,  Pennsylvania

               •   Revisions by:

 The Association  of Official Analytical Chemists
                Dr. Joseph Sherma
                Dr. Morton Beroza
             Contract No. 68-02-2474

           Project Officer . - Editor


               Randall R. Watts

        Environmental Toxicology Division
        Health Effects Research Laboratory
        Research Triangle Park, N.G.  27711
        *


           Revised:  1981
       U.S.  ENVIRONMENTAL  PROTECTION AGENCY
        OFFICE  OF  RESEARCH AND  DEVELOPMENT
        HEALTH  EFFECTS  RESEARCH LABORATORY
        RESEARCH TRIANGLE  PARK,  N.C. 27711
                                           Printed on Recycled Paper

-------
                               DISCLAIMER
     This report has been reviewed ,by the Health Effects Research
Laboratory, 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 recommendation, for use.
                               ii

-------
                                  FOREWORD
     The many benefits of our modern developing, industrial  society are
accompanied by certain hazards. Careful assessment of the relative risk of
existing and new man-made environmental hazards is necessary for the estab-
lishment of sound regulatory policy.  These regulations serve to enhance
the quality of our environment in order to promote the public health and
welfare and the productive capacity of our Nation's population.

     The Health Effects Research Laboratory, Research Triangle Park, con-
ducts a coordinated environmental health research program in toxicology,
epidemiology, and clinical studies using human volunteer subjects.  These
studies address problems in air pollution, non-ionizing radiation, environ-
mental carcinogenesis and the toxicology of pesticides as well as other
chemical pollutants.  The Laboratory participates in the development and
revision of air quality criteria documents on pollutants for which national
ambient air quality standards exist or are proposed, provides the data for
registration-of new pesticides or proposed suspension of those already in
use, conducts research on hazardous and toxic materials, and is primarily
responsible for providing the health basis for non-ionizing radiation
standards.  Direct support to the regulatory function of the Agency is pro-
vided in the form of expert testimony and preparation of affidavits as well
as expert advice to the Administrator to assure the adequacy of health care
and surveillance of persons having suffered imminent and substantial en-
dangerment of their health.

     This manual provides the pesticide chemist with a systematic protocol
for the quality control of analytical procedures and the problems that
arise in the analysis of human or environmental media.
                                                F.G, Hueter, Ph.D.
                                                    Director
                                        Health Effects Research Laboratory
                                   iii

-------
                                  ABSTRACT
     This manual provides the pesticide chemist with a systematic protocol
for the quality control of analytical procedures and the problems that arise
in the analysis of human or environmental media.  It also serves as a guide
to the latest and most reliable methodology available for the analysis of
pesticide residues in these and other sample matrices.  The sections.dealing
with inter- and intra-laboratory quality control, the evaluation and stand-
ardization of materials used, and the operation of the gas chromatograph are
intended to highlight and provide advice in dealing with many problems which
constantly plague the pesticide analytical chemist.  Many aspects of the
problem areas involved in extraction and isolation techniques for pesticides
in various types of samples are discussed.  Techniques for confirming the
presence or absence of pesticides "in sample materials are treated at some
length.  This highly important area provides validation of data obtained by
the more routine analytical procedures.  The gas chromatograph, being the
principal instrument currently used in pesticide analysis, often requires
simple servicing or troubleshooting.  A section addressing some of these
problems is included.  Last, but by no means least in importance, is a short
dissertation of the value and need for systematic training programs for
pesticide chemists.
                                     iv

-------
Section
Number
TABLE OF CONTENTS

      Section
In-Section
 Page
           FOREWORD AND ABSTRACT
           GENERAL DESCRIPTION OF PESTICIDE RESIDUE ANALYTICAL
           METHODS

           A.  Sampling                                            1
           B.  Extraction Procedures                               2
           C.  Cleanup Procedures                                  3
           D.  Final Determination Methods                         5
           E.  Confirmatory Techniques                             9
           F.  Automation and Computer Processing                 10
           G.  References                                         10

           INTERLABORATORY QUALITY ASSURANCE

           A.  Quality Assurance Program of the EPA,Environ-      14
                 mental  Toxicology Division
           B.  Program Objectives                                 15
           C.  Program Activities                                 15
           D.  Types and Preparation  of  Sample Media              16
           E.  Reporting Forms                                    17
           F.  Evaluation of Reported Data                        18
           G.  Summary of Results Tables and  Results of  Check     21
                 Sample  Analyses
           H.  Relative  Performance Ranking                       37
           I.  Private Critiques                                  50
           J.  Progression of Performance                         59
           K.  Statistical Terms and  Calculations                 53
           L.  References                                         59
           M.  Additional Sources of  Information on              59
                 Pesticide Quality Assurance  Programs

           INTRALABORATORY QUALITY CONTROL

           A.  Purpose and Objectives                      ,      61
           B.  Purpose and Objectives of SPRM's                   61
           C.  Nature of SPRM's                                  62
           D.  Frequency of  SPRM Analysis                         63
           E.  Record Keeping                                     64
           F.  Quality  Control  Charts                            69
           G.  Benefits  of  the  In-House SPRM  Program             73
           H.  Analytical Balances                                74
            I.  Purity of Solvents                                 75
            J.  Distillation  of  Solvents                          79
            K.  Contamination from Reagents  and Materials         79
            L.   GC Retention  Data for Common Interferences        82

-------
Section
Number
                                    Section
                                                  In-Section
            M.
            N.
            0.
            P.

            Q.
            R.

            S.
  Cleaning  of Glassware
  Housekeeping
  Analytical Pesticide Reference Standards
  Calibration and Maintenance of the Gas
   Chromatograph and Accessories
  Adherence to Official or Standardized Methodology
  Implications of an Intralaboratory Quality
   Control Program
  References
            EVALUATION,  STANDARDIZATION, AND  USE  OF MATERIALS
            FOR PESTICIDE  RESIDUE ANALYSIS
           A.
           B.
           C.
           D.
           E.
           F.
           G.
           H.
           I.
           J.
           K.
           L.
           M.
           N.
 Adsorbents
 Introduction and Column Technology
 Column Efficiency and Peak Resolution
 Sensitivity and Retention
 Column Stability
 Resistance to On-Column Compound Decomposition
 Homemade vs.  Precoated Packings
 Packing the Column
 Column Conditioning
 Support Bonded Carbowax 20M Columns
 Evaluation of the Column
 Maintenance and Use of GC Columns
 Capillary GC  Columns
 References
           INSTRUMENTATION AND PROCEDURES FOH GAS CHROMATOGRAPHY
           A.
           B.
           C.
           D.
           E.
           F.
           G.
           H.
           I.
           J.
           K.
           L.
           M.
           N.
           0.
           P.
Temperature  Selection and Control
Selection and  Control of Carrier Gas Flow Rate
Electron Capture Detector
Microcoulometric Detector
Thermionic Detectors
Flame Photometric Detector (FPD)
Electrolytic Conductivity Detector
Other Detectors and Detector Combinations
Electrometer and Recorder
Sample Injection and the- Injection Port
Erratic Baselines
Recommended GC Columns for Pesticide Analysis
Sensitivity of the GC System
Qualitative Analysis
Quantitative Analysis
References
  84
  84
  85
 101

 103
 104

 105
 107
 113
 115
 118
 119
 119
 124
 126
 128
 131
 132
 133
 136
 138
140
143
147
158
159
164
171
178
180
181
186
187
196
197
200
215
                                   vi

-------
Section
Number
                        Section                    	

MAINTENANCE, TROUBLESHOOTING, AND CALIBRATION OF
THE GAS CHROMATOGRAPH AND DETECTOR SYSTEMS

A.  Daily Operational Considerations for Gas
      Chromatographic Instrumentation
B.  Check List When Instrument Repair is Indicated
C.  General Approach to Troubleshooting
D.  Gas Chromatograph Service Block Diagram
E.  Gas Inlet System of the Gas Chromatograph
F.  Procedure for Isolating Problems in Flow Systems
     , of Electron Capture Equipped Gas Chromatographs
G.  General Information - Flow System
H.  Temperature Control and Indication in the Gas
      Chromatograph
I.  Detector and Electrometer
J.  Observation of Problems on Chromatograms
K.  Detector Background Signal (BGS) Response
L.  Troubleshooting Coulson Electrolytic
      Conductivity System
M.  Troubleshooting the Flame Photometric, Detector
       (FPD)
N.  Epilog

OTHER CHROMATOGRAPHIC DETERMINATIVE TECHNIQUES
HIGH PERFORMANCE LIQUID COLUMN CHROMATOGRAPHY  (HPLC)
AND THIN LAYER CHROMATOGRAPHY         .
A.  Introduction to HPLC
B.  Theory  and Principles of HPLC
C.  HPLC Instruments
D.  Columns for HPLC
E.  Mobile  Phases  (Solvents) for HPLC
F.  Detectors for HPLC
G.  Practical Aspects of Successful HPLC  Operation
H.  HPLC Data
I.  Applications of HPLC to Pesticide Analysis
J.   Introduction to TLC
K.  Practical Considerations ,in TLC
L.   Quantitative TLC
M.   Thin Layer  Systems
N.   References

 SAMPLING,  EXTRACTION,  AND  CONCENTRATION PROCEDURES
 IN PESTICIDE ANALYSIS

 A.  General Considerations in Sampling
 B.  Representative vs.  Biased Sampling
 C.  Sample Containers
 D.  Sample Compositing
In-Section
   Page
                                                                    220

                                                                    221
                                                                    222
                                                                    225
                                                                    226
                                                                    228

                                                                    232
                                                                    233

                                                                    234
                                                                    236
                                                                    237
                                                                    238

                                                                    241

                                                                    243
                                                                    244
                                                                    245
                                                                    247
                                                                    249
                                                                    251
                                                                    252
                                                                    254
                                                                    257
                                                                    257
                                                                    258
                                                                    265
                                                                    268
                                                                    273
                                                                    276
                                                                     285
                                                                     286
                                                                     286
                                                                     288
                                      vii

-------
Section                                                            In-Section
Number                              Section                          Page

   8         E.   Storage  of  Samples                                   289
            F.   Sampling of Agricultural  and  Food Products           290
            G.   Sampling of Biological Materials                     291
            H.   Air  Sampling                                        292
            I.   Water  Sampling                                       293
            J.   Sampling of House Dust, Soil,  and Stream            296
                  Bottom Sediment
            K.   Marine Biological and  Wildlife Samples               296
            L.   Control  of  Procedures  for Extraction of  Residues     297
            M.   Control  of  Methodology for Concentration of          301
                  Sample Solutions.and Fractional .Column Eluates
            N.   References                                           307

  9         MOLTIRESIDUE EXTRACTION AND ISOLATION PROCEDURES
            FOR PESTICIDES AND METABOLITES AND RELATED COMPOUNDS

            A.  Tissue, Fat, 'and Food Analysis by the Mills',         312
                  Onley, Gaither Procedure
            B.  HCB and Mirex in Adipose  Tissue                      316
            C.  Human or Animal Tissue and Human Milk Analysis   .    316
                  by the Florisil Micromethod
            D.  Human Blood  or Serum                                 316
            E.  Pentachlorophenol (PGP) in Blood and Urinfc           317
            F.  Bis(£-Chlorophenyl)acetic acid (p_,p_'-DDA) in Human   319
                  Urine
            G.  2,4-D and 2,4,5-T in Urine         '                  319
            H.  Kepone in Human Blood for Environmental Samples      320
            I.  Gel Permeation Chromatography                        321
            J.  Determination of Chlorophenoxy Herbicides in         324
                  Fatty and Nonfatty Foods
            K.  Carbon-Cellulose Column Cleanup
            L.  Cleanup on Silica Gel                                325
            M.  Cleanup on Deactivated Florisil and Silica Gel       325
            N.  Low Temperature Precipitation                        327
            0.  Cleanup on Alumina                                   327
            P.  Miscellaneous Multiresidue Cleanup Procedures        329

            Organophosphorus Pesticides

            Q.  Determination of Metabolites or Hydrolysis           331
                  Products in Human Urine, Blood,  and other Tissue
            R.  Determination of _p_-Nitrophenol (PNP)  and Other      332
                  Phenols in Urine
            S.  Sweep Co-Distillation                               333
            T.  Charcoal Cleanup of Nonfatty Food Extracts          336
            U.  Acetone Extraction                                  336
            V.  Miscellaneous Multiresidue Cleanup Procedures       337
                                      viii

-------
Section
Number
                                                        In-Section
                                                         Page
  10
                        Section

Carbamate Pesticides and Metabolites and Miscellaneous
Herbicides
  W.  1-Naphthol in Urine                                 339
  X.  Analysis of Amine Metabolites in Urine              340
  Y.  Other Indirect (Derivatization) Methods of          340
        Analysis
  Z.  Direct Methods of Analysis                          341
A-,A.  Analysis of Plant and Food Materials                341
A,B.  Air Analysis                                        342
A,C.  Water Analysis                                      343
A,D.  Soil, House Dust, and Bottom Sediment               345

  Polychlorinated Biphenyls (PCBs), and Other Compounds

A,E.  Pesticide-PCB Mixtures                              348
A,F.  Appearance of TCB Chromatograms                     348
A,G.  Methods for Separation and Analysis of              352
        Pesticides and PCBs
A,H.  Determination of Polybrominated Biphenyls           359
A,I.  Separation and Determination of Polychlorinated     360
        Terphenyls
A,J.  Separation and Determination of Dioxins             361
A,K.  Determination of Ethylenethiourea (ETU)             362
A,L.  Determination of Conjugated Pesticide Residues      362
A,M.  Reviews of Analytical Methods for Pesticides,       364
        PCBs, and Other Non-Pesticide Pollutants
A,N.  References                                          364

  CONFIRMATORY AND OTHER DETERMINATIVE PROCEDURES

  A.  Requirements for Positive Confirmation of           381
        Pesticide Identity
  B.  GC  Relative Retention Times                         383
  C.  Selective GC Detectors                              386
  D.  Thin Layer Chromatography (TLC) Rj- Values           388
  E.  High Performance Liquid  Chromatography  (HPLC)       391
  F.  Extraction p_-Values                                 391
  G.  Derivatization  (Chemical Reaction) Techniques       392
  H.  Spectrometry  (Spectrophotometry)                    402
  I.  Visible, UV, Fluorescence,  and Phosphorescence      402
  J.  Infrared  (IR)                                       403
  K.  Nuclear Magnetic Resonance  (NMR)                    408
  L.  Mass Spectrometry  (MS)                              409
  M.  Quality Assurance of GC-Low Resolution Mass         425
         Spectrometry
  N.  Biological Methods                                  432
  0.  Polarography  (Voltammetry)                          432
  P.  Miscellaneous Confirmatory  Methods                  434
  Q.  References                                          434
                                     ix

-------
Section
Number

  11

  12

  13
                        Section

TRAINING OF PESTICIDE ANALYTICAL CHEMISTS

ABBREVIATIONS

QUALITY CONTROL MANUAL REVISIONS
In-Section
  Page

  448

  451

  455

-------
                                   Section 1


                    GENERAL DESCRIPTION OF  PESTICIDE

                        RESIDUE ANALYTICAL METHODS



  A pesticide residue  analysis  usually consists of five.steps:

  (1)   Sampling.

  (2)   Extraction of the residue from the  sample matrix.

  (3)   Removal of interfering co-extractives  ("cleanup").

  (4)   Identification  and estimation  of  the quantity of residues in the
       cleaned-up extract,  usually  at very low levels  (e.g., 10~" to 10~12 g
       for gas chromatography). To obtain this sensitivity, selective
       determinative methods such as  chromatography are usually required.

  (5)   Confirmation of the presence and  identity of the residues.


  The exact nature of  each of these stages is dictated by the specific
  pesticide(s) and sample substrkte involved.  A brief discussion of general
  aspects of these steps follows:


1A    SAMPLING

  The aim of sampling is to provide a reproduction of  a portion of the
  environment, on a scale that enables  the sample to be handled in the
  laboratory.  Analytical results  are meaningful only  if collected samples
  are truly representative and meet the goals of the monitoring study or
  program.  The sites, techniques,  and  frequency of sampling and the size
  and number of samples must allow the  analytical results to be statistically
  evaluated and replicated at a later time for confirmation.  If storage of
  samples before analysis is necessary,  it must be proven that alteration
  in the nature or amount of pesticide  residues  does not occur.  Samples
  may be composited or subsampled prior to analysis. . The steps in the
  analytical procedure are influenced significantly by the manner in which
  the sample is collected, preserved, stored, shipped, and otherwise
  processed prior to extraction.
                                     -1-

-------
                                                           Section IB
IB   EXTRACTION PROCEDURES
  Environmental and biological samples generally cannot be analyzed
  directly for pesticide residues because the level of the desired residue
  is too low, and the levels of interfering constituents are too high.  In
  virtually every modern method of pesticide residue analysis, the com-
  pounds of interest are separated from the bulk of the sample matrix by
  some form of extraction.  In most cases, extraction is followed by a
  cleanup procedure to eliminate, or at least minimize, interfering substances.
  In both extraction and purification procedures, the fractional recoveries
  of the compounds must .be known, and it must be possible to relate the
  amounts found in the subsequent assay to the concentrations originally
  in the sample matrix.
                            v

  A solvent or mixture of solvents should be used for extraction that is
  at least 80% efficient, selective enough to require a minimum of cleanup,
  and does not interfere with the final determination.  Simple washing of
  the whole sample may be adequate for surface residues of foliage or
  vegetables and fruits, but Soxhlet extractors, blenders,, and tumbling or
  shaking devices are used for most samples.  Hexane or hexane-acetone
  mixtures are typical solvents for nonpolar, fat-soluble organochlorine
  pesticides; and benzene, chloroform, dichloromethane, or acetonitrile
  are commonly used for the more polar compounds such as organophosphates
  and carbamates.   Acetonitrile is an excellent general solvent for preliminary
  extraction of unknown residues of a wide polarity range.   The more polar
  solvents, however, remove greater amounts of co-extractives and may complicate
  subsequent cleanup steps.  Sodium sulfate is sometimes added to help extract
  the more water-soluble compounds.  Exhaustive Soxhlet extraction with an
  appropriate solvent or mixture of solvents is the most efficient method
  for many pesticides and sample types and can be used to compare with other
  proposed procedures.

  A given extraction procedure should be validated for each type of sample
  matrix and for each class of compounds to which it is applied.   An extraction
  procedure suitable for one class of compounds in a given sample may not be
  suitable for the extraction of a different but closely related class of
  compounds from the same sample.   The nature of the sample matrix influences
  the effectiveness of the extraction procedure through the toughness,  water
  content, and lipid content.   The toughness determines the ease of finely
  dividing the sample, the water content affects the solubility of pesticides
  in the extraction solvent, and the lipid content of the sample influences
  both the amount  of solvent and the proportion of nonpolar component required.
  It is usually desirable to quantitatively extract lipids  with the pollutants
  from environmental samples for ease in reporting analytical results.   Optimum
  extraction conditions in terms of solvent polarity and in the time and
  manner of contact between sample and extraction solvent should,  therefore,
  be found by recovery studies for each analysis at several concentration
  levels.   Recovery from spiked or fortified samples may not provide valid
                                      -2-

-------
                                                            Section  1C


   information, about  the  recovery  of  endogenous material.  Extraction
   efficiency can be  checked most  accurately  if the  laboratory  is  equipped
   to biologically incorporate radioactive-labeled parent  compounds  and/or
   metabolites in the sample substrate.  When polar  compounds are  involved,
   hydrolysis to free conjugated residues must be considered before  extraction.
   However,  the conditions  must be such  that  the  compounds of interest
   survive the treatment.


1C    CLEANUP PROCEDURES

   The amount of extract  purification (cleanup) required prior  to  the final
   determination depends  on the selectivity of both  the extraction procedure
   and the determinative  method.  It  is  an  unusual situation, e.g.,  with
   some water samples, when extracts  can be directly determined without
   further treatment.  Injection of uncleaned samples into a gas chromatograph
   can cause extraneous peaks, damage to the  peak resolution and efficiency
   of the column, and loss  of  detector sensitivity.   Impure  samples  spotted
   for thin layer chromatography may  result in streaked zones or decreased
   sensitivity of visualizing  reagents,  while those  injected into  a  liquid
   chromatbgraph can  greatly shorten  the lifetime of an expensive  prepacked
   column;  Extracts  containing fatty material are especially troublesome.
   Depending "on the extent  and nature of the  co-extractives  and the  pesticide
   residue,  partition between  immiscible solvents; adsorption chromatography
   (column or TLC); gel permeation chromatography; chemical  destruction of
   interfering substances with acid,  alkali,  or oxidizing  agents;  distillation;
   sweep co-distillation; and  selective  photodegradation are most  often used
   for cleanup, either individually or in various combinations.
                       i
   Plant or crop materirl is usually  extracted with  a water-miscible solvent
   such as acetone or acetonitrile.  After  dilution  with water, the  residues
   are generally partitioned into  a solvent such  as  methylene chloride  that
   can be readily evaporated to dryness. Polar pesticides are  only  poorly
   recovered from,a surplus of water  in this  way, and the  trend is to keep  the
   residues in organic solution and remove  the water co-extracted  from the
   sample.  Evaporation of  the organic solvent yields good recoveries of
   even highly polar  compounds, which can be analyzed directly  by  GC with a
   selective detector or cleaned up further by a  multiresidue method such as
   liquid adsorption chromatography,  gel permeation, or sweep  co-distillation.
   For purification of fat  extracts,  samples are  usually partitioned between
   hexane and acetonitrile, dimethyl formamide, or dimethyl  sulfoxide.   The
   latter solvent is  used in the widely applicable Wood procedure  to elute
   chlorinated pesticides from a column prepared  from a fatty  sample mixed
   with Celite.  The lipid content of a particular  sample  has  a significant
   effect on the recovery of pesticides in solvent partitioning procedures.
   For example, DDT  is recovered more efficiently by acetonitrile  partitioning
   from pure hexane  than from a hexane solution containing dissolved fat.  This
   factor can contribute to the variability of recovery in cleanup procedures
   involving solvent partitioning.
                                       -3-

-------
                                                          Section 1C

  For adsorption chromatography,  direct extracts or extracts  purified bv
  partitioning are concentrated to a snail volume,  applied to the top of
  a Florisil,  charcoal,  alumina,  silica gel,  or mixed-adsorbent  column, and
   S?iPe!J        are 6lUted ±n fract±ons by Passage of one or more solvents
  while the co-extractives  remain on the column or  are eluted in'different
  fractions.   Elution of a  residue in a certain fraction  (selective adsorp-
  tion)  is  useful evidence  for confirmation of  identity.   The capacity of a
  column for co-extractives and the uniformity  of activity (elution pattern)
  from batch to batch are important characteristics  of adsorbents used for
  cleanup.  Methods  are  available for activation and deactivation of different
  adsorbents and  for checking  the activity  level achieved.

  Florisil  is  still  the most widely used adsorbent,  and it is  involved in
  cleanup schemes for many  fatty  and nonfatty samples.  However, Florisil
  is not available in all countries  in  sufficiently  constant quality, and
 many analysts are becoming increasingly aware  that silica gel and alumina
 are less expensive, are at least as easy  to standardize, and provide
 similarly good results.  Another trend is miniaturization of adsorbent
 cleanup columns.  These micro columns, containing, e.g., Florisil or 302
 water-deactivated silica gel, are very promising for routine analyses
 because they offer economy in solvents and materials and reduce health
 and fire hazards.  Sensitivity obtainable is,  of course, lower compared
 to corresponding large sample cleanup procedures.

 Chemical destruction methods are extremely useful, but caution must be
 used in applying them to compounds other than  those for which they  have
 specifically been validated.   For example, 2,3,7,8-tetrachlorodibenzo-p-
 dioxin i«  commonly determined following alkaline hydrolysis  of  tissue  or
 extracted  lipids, but this treatment completely destroys octachlorodibenzo-p-
 dloxin.                                                     •              •*-

 Of the procedures listed above,  liquid-liquid  partition followed by
 adsorption chromatography  is  most often applied for organochlorine pesticides
 and related compounds (e.g.,  PCBs) before GC with  the relatively unspecific
 electron capture detector.  An automated instrument based on gel permeation
 S°S?J085aphy Ja\bffn Shown t0 eff±ciently separate chlorinated pesticides
 and PCBs from the bulk  of  the lipids extracted from fatty samples, and to
 be advantageous  in  terms of convenience and  speed  of  processing laree
 numbers of samples.   When  specific GC detectors are employed, cleanup of
 extracts becomes less important.   Thus,  a  suitable  extraction procedure
 combined with a  partitioning  step is  often sufficient for determining
 organophosphorus and organonitrogen compounds  in many samples.  HPLC and
 TLC generally require more  effective  cleanup steps.

 Cleanup procedures  should be  chosen in terms of practicality, cost,  time,
 and reagent and  equipment availability.  The methods  chosen should be tested
 to be sure they allow detection  and determination of  the pesticides  of
 on'^8' *5 the desired sensitivity level, with  recovery of preferably
 80-85Z or better, and with removal  or  separation of adequate levels  of
background interferences.  Procedures  giving the highest mean recovery
of a residue may not necessarily be the best to use if there is a high
                                    -4-

-------
                                                            Section  ID

  variation  in  recovery  from sample  to  sample.  A cleanup method  giving &
  moderate but  highly  reproducible recovery may be  a better choice  than one
  giving  a high but variable recovery.   The nature  of  the lipids  in a  sample
  extract can be  important  in determining  the  choice of  an  adsorbent for
  liquid  column chromatographic  cleanup.   For  example, acidic alumina  has
  a  greater  capacity for lipids  than does  basic or  neutral  alumina.

  Concentration of solutions is  often required prior to, during,  and after
  cleanup procedures.  Great care is necessary when evaporating to  low
  volumes to avoid losses of the pesticide residue, and  evaporation to
  complete dryness is  usually inadvisable.  Kuderna-Danish  evaporators and
  micro Snyder  columns,  special  block heaters, and  rotary-vacuum  evaporators
  are recommended for  concentrating  solutions  containing pesticide  residues,
  and keeper solutions may  be added  to  retard  the loss of volatile  compounds.
  All steps, in  the analytical procedure should be checked for residue  loss due
  to volatilization or degradation   by  carrying out recovery studies on a
  spiked  control  (uncontaminated) sample at different  fortification levels.
  Since concentration  factors are often 1000:1 or more,  the possibility of
  interference  from the  solvent  itself  must be considered.

  Most of the extraction and .cleanup procedures available today will yield
  reliable and  reproducible results  when practiced  by  a  trained and competent
  analyst.   An  important precondition is that  a laboratory  gain abundant
  experience with any  method that is to be used.  The  most  important analytical
  methods are those allowing the determination of multiresidues of  pesticides
  and related compounds. Most methods  available today,  however,  are of value
  only for the  parent  pesticidal compounds and do not  include their significant
  metabolites.  A1 primary target for future research is  the inclusion  of the
  metabolites of  toxicological importance  in  the existing and new multiresidue
  schemes.   In  addition, many of the existing multiresidue  schemes  do  not
  include the many new pesticides, mainly  water-soluble  and systemic insecticides
  and fungicides, that have been introduced in the  past  few years.  There is
  little  doubt  that a  number of  analyses for  regulatory  purposes  are routinely
  performed  today for  pesticides that have been superseded  by other compounds
  which are  not detectable by the procedures  in current  usage.


ID    FINAL DETERMINATION METHODS

   Chromatographic methods are by far the most widely used for determination
   of pesticide  residues, followed by spectrophotometric  and biological methods.
   The latter include bioassay and enzymatic  techniques that are simple,  since
   they require  no cleanup,  but are  non-specific.  Enzyme inhibition, when used
   as a detection procedure after thin layer  chromatography  is a sensitive
   (low ng detection limits) and selective  method  for certain organophosphorus
   and carbamate pesticides.

   Spectrophotometric methods are generally less  sensitive and less  selective
   than gas or thin layer chromatography and are useful mainly as  ancillary
   techniques to gas chromatography  for confirmation of residue  identity  or
   for quantitation of individual pesticides.   If  selectivity and sensitivity
   are adequate, colorimetric methods can advantageously be  adapted  to  automated
                                       -5-

-------
                                                           Section  ID

 processes.   Fluorescent  pesticides  and metabolites may be determined by
 fluorometry, which is more  sensitive  than visible, UV, or IR methods.
 Since relatively few pesticides  are naturally fluorescent, fluorometry
 is  selective; however, removal of fluorescent impurities  is often necessary
 and this can be  difficult.

 Paper chromatography provided the analyst in the late 1950Ts with the first
 multiresidue method  for  separation  and identification of  pesticides.  It
 has been largely superseded by gas  chromatography as the  primary determinative
 procedure and thin layer chromatography (TLC) for screening, semiquantitation,
 and confirmation.  Compared to paper chromatography, TLC  offers generally
 increased resolution, shorter development times, and increased sensitivity.
 Mo.st pesticide analyses have been performed on 0.25 mm layers of alumina or
 silica gel, but  polyamide and cellulose are also used.  Organochlorine com-
 pounds are detected at 5-500 ng levels by spraying with ethanolic AgNO* or
 incorporation of AgN03 into the layer followed by irradiation with ultraviolet
 light.   Many organophosphorus and carbamate pesticides are detectable at low
 ng levels by enzyme inhibition techniques or at higher levels by numerous
 chromogenic reagents.  Fungicides are detectable by bioautography.  Polar
 herbicides and heat-labile,  poorly detectable carbamates, which require
 formation of derivatives prior to gas chromatography, are particularly amenable
 to analysis by TLC.

 Gas chromatography of pesticides  is  normally carried out  on 90  to  200  cm glass
 columns  packed with single and mixed organosilicone  and polyester  stationary
 phases ranging from low to high polarity.  Among the most used  phases  are
 SE-30, QF-1,  DC-200,  OV-210, OV-17,  DECS,  Carbowax 20M, and OV-17/OV-210
 SE-30/OV-210,  and DC-200/QF-1 mixtures.  The chemically stable,  low bleed
 0V series  of phases have  become quite  popular.   Samples should  be  examined
 on two or  three columns of markedly  different polarity before results are
 considered  conclusive.  A useful  series of columns with increasing polarity
 from which  to  select an optimum separation is:   OV-101 (methyl  silicone)-
 «V~JL(l?ethyl/phenyl  sil±cone>J OV-210 (methyl/trifluoropropyl  silicone);
 7"  5N^cyan°ProPy1/Phenyl/Jnethyl silicone); and Carbowax  20M (polyethylene
 glycol).

 Glass columns are preferred  because  they minimize decomposition and are
 easier to pack for optimum efficiency.  After being packed, columns are
 conditioned at an elevated temperature to minimize liquid  phase bleeding
 and to obtain reproducible chromatograms.  In some cases,  large quantities
 of the pesticides being determined are injected  initially  to improve the
 response of these compounds.  The formation of derivatives in GC residue
 analysis is a necessity when the analyte is labile or otherwise troublesome
 or is poorly detected by  selective detectors.

 Columns loaded with relatively low percentages of liquid phase generally
 give superior resolution and sensitivity but become contaminated more easily
and are more prone  to interactions between solutes and the solid support
 than more heavily coated columns.   Certain pesticides, such as DDT and
                                    -6-

-------
                                                         Section ID

ehdrin, are subject to degradation in columns under certain conditions,
and these conditions should be avoided.  Highly inert columns have been
prepared by chemically bonding Carbowax 20M to a GC support.  These
packings are used for GC directly, or after further coating with a liquid
phase.  Glass capillary columns, with their outstanding separation ability
for difficult samples, are being reported much more frequently in residue^
analysis, even though they are not as tolerant of the injection of "dirty
extracts.

Organochlorine pesticides are usually analyzed with a tritium or   Ni electron
capture detector with DC or pulsed applied voltages.  Though this detector
is less specific than the other common pesticide detectors, it can detect
as low as 10~13 g amounts of many halogenated compounds.  63Ni detectors are
operable at high temperatures (over 300°C), thus reducing possible problems
from contaminants condensing in the detector.  Tritium detectors are less
expensive, and contaminated foils can be easily changed or  removed for
cleaning.  Commercial devices are available for linearizing EC response
over a 103-105 range of concentration, and the pulsed wide-range °3Ni detector
has become especially popular because it can be used with automatic injection
systems.  Organophosphorus pesticides are detected selectively at
ca. lO-^-lO'11 g levels by the flame photometric detector  (FPD) in the
phosphorus mode (526 nm), and the FPD has overtaken the thermionic detector
as the primary detector for the determination of these compounds.  Sulfur-
containing pesticides may be selectively detected by the FPD  (394 nm) with
about one order of magnitude less sensitivity, or at the low ng level with
the S-mode of the Hall electrolytic conductivity detector.  Nitrogen-con-
taining pesticides are detected selectively with the N-mode of the Hall detector
 (ca.  107* g  sensitivity) or with  the N/P mode of the flameless N-P thermionic
detectot'; (ca. IQ-3-2 g sensitivity).  The N-P thermionic detector also has a
mode  of operation  that is selective for pg levels of only phosphorus-containing
compounds, and the Hall  detector  can be operated selectively  for organochlojrine
compounds at low ng levels.  Labile, polar carbamate pesticides or their
hydrolysis products are  often  derivatized with a halogen-containing reagent
and  the  resulting  derivative can  be sensitively detected with the  electron
 capture  detector.   Selective detectors have  the advantages  of simplifying
 cleanup  procedures and aiding  residue  identification.  The  mass  spectrometer
 is a unique  GC detector  in  that it  is  capable of almost  specific  detection
 and  identification of pesticide residues.   It  is, however,  expensive  for
 routine work.

 Samples and standards must  be  injected into  the gas chromatograph using
 a consistent and reproducible technique.   It is advisable that injected
 volumes of standards and samples be nearly equal  and represent 20-80% of
 the total volume of the syringe used.   Syringes must be well cleaned  between
 injections,  and injection port septa and liners must be changed regularly.
 Standards should be injected before and periodically during the analysis
 of a series of samples.   Cleaner samples require fewer standard injections.
 A pesticide mixture that indicates the overall performance of the GC  system
 should be injected at least once daily.
                                     -7-

-------
                                                         Section ID

Modern high performance liquid chromatography (HPLC)  is being  used
increasingly for the final,  room temperature determination of  polar,
involatile, or heat-labile pesticide residues without derivative  formation.
Analytical columns  are 10-50 cm in length and 2-6 mm  in internal  diameter.
For pesticides, they Sire -commonly packed with 5-10 ym particles of a
totally porous adsorbent silica gel to which a C^s hydrocarbon phase has
been chemically bonded (reversed phase chromatography).  The mobile phase
±s  -pumped through the column at flow rates of 1-2 ml/minute (100-200 atm.
pressure).   Most residue analyses have been carried out with detection by
UV  absorption, and  to a lesser extent by fluorescence or photoconductivity
detection.   The electrochemical detector is just beginning to  find use in
residue analysis.   Refractive index detection has been reported infrequently,
if  at all.   UV detection with a mercury  lamp at its major  emission wavelength
of  254 nm is most often used,  but use of the variable,wavelength  detector
is  growing because  254 nm or other wavelengths available from  a mercury
lamp are not optimal for many pesticides.   Fluorescence detectors have been
used in determining nonfluorescent pesticides by fluorogenic labeling
employing derivatization methods similar to those applied  earlier to
facilitate thin-layer fluorodensitometry.   The major  disadvantage of HPLC
at  present  is the poor sensitivity (ca.  10~7-10-l*gl  an
-------
                                                            Section IE
IE    CONFIRMATORY TECHNIQUES

   Three truly independent results are considered necessary for positive
   confirmation of the Identity of a residue.  Alternative methods that
   can be combined are TLC and/or paper chromatography with sorbent-solvent
   systems of different polarity or different visualization reagents, gas
   chromatography with columns of different polarity and selective detectors,
   preparation of derivatives to alter structure and volatility and thereby
   chromatographic properties, extraction pj-values, ultraviolet photolysis,
   and mass spectrometry.  Unlike conventional NMR, IR, UV, etc., mass
   spectrometry has sufficient sensitivity for general application to residue
   identification as well as for confirmation of pesticides in the presence
   of PCBs.  Thus, the directly coupled gas chromatograph-mass spectrometer
   is a powerful tool for positive identification of mixture components at
   residue levels.  The ability of the high resolution mass spectrometer to
   measure precise ionic masses has allowed individual pesticides with different
   elemental compositions to be identified in complex mixtures without prior
   separation in some cases.

   The reliable detection and estimation of pesticide residues is one of the
   most difficult and demanding tasks an analytical chemist can be called
   upon to perform.  Important commercial pesticides include insecticides,
   fungicides, herbicides, acaricides, and rodenticides.  There are many
   hundreds of these compounds with greatly differing chemical structures
   and properties (e.g., organohalides, organophosphates, carbamates, anilines,
   ureas, phenols, triazines, quinones, etc.)-  Their determination may involve
   traces of any of these materials alone or in combination in a great variety
   of matrices, each with its own peculiar problems.

   Further complications arise because metabolic degradation of certain
   pesticides produces compounds that may be more toxic and of different
   polarity than the parent pesticide.  Examples include metabolically
   derived heptachlor epoxide and dieldrin, from heptachlor and aldrin,
   respectively, and oxygen analog metabolites of sulfur-containing organo-
   phosphorus pesticides.  The analyst should be able to determine the identity
   and quantity of these metabolites and degradation products as well as the
   residue of the original pesticide, and extraction and cleanup procedures
   and chromatographic determinative conditions may have to be modified to
   accommodate these compounds.  Multi-component pesticides such as chlordane,
   toxaphene, and strobane and their metabolites pose difficult confirmation
   and quantitation problems.  Closely related, non-pesticidal compounds with
   similar analytical behavior such as PCBs or chlorinated naphthalenes may also
   be present in extracts, and the analyst must be able to isolate, identify,
   and measure pesticides of interest while simultaneously separating, isolating,
   and identifying these related compounds, if necessary.  Trace contaminants
   contained in solvents or reagents, or extracted from plastic apparatus, can
   give rise to GC peaks or TLC spots that may be confused with pesticides.
   Positive confirmation of some pesticides is especially difficult because
   of very similar chromatographic properties of compounds, e.g.,  dieldrin and
   "photo-dieldrin."

                                       -9-

-------
                                                               Sections IP,  IG

      The amount of effort expended and the choice of confirmatory tests are
      determined by the Importance of the sample, resources available,  and the
      amount of residue present.   A possible alternative to testing of  every
      residue is confirmation of  selected samples at intervals, when the same
      residues are apparently present in all samples of a group.   If sufficient
      residue is not available in individual members of a group of samples -to
      permit use of, a certain test,  purified extracts are often pooled  for
      confirmation.


 IF     AUTOMATION  AND  COMPUTER  PROCESSING

     Automation of  pesticide analyses  is presently  in  its  early stages.  Totally
     automated  procedures have been developed for analyses not requiring column
     «  ™P J°n cleanuP and  those ln wb-*<* the final determination is colorimetry
     or UV absorption.  Several microprocessor-controlled systems for automatic
     transfer of manually prepared  samples  onto a gas or liquid chromatography
     column are being marketed.  Laboratories with high sample throughput can
     find such  systems save  time and cost  in determinative steps.  A system
     for automatic  cleanup of samples by gel permeation chromatography has been
     designed, although automation of preparative and cleanup steps is not yet
     far advanced.  Data systems allow storage of large amounts of data with
     computerized printouts that increase the speed and efficiency of analyses
     and improve both quantitation and identification of residues.

     Although advances in automation are being reported at an ever-increasing
     rate, available systems are generally useful only for well-defined samples
     containing known pesticides.  A skilled analyst using conventional, non-
     an«?™!   frocedures ls still required to carry, out successfully multiresidue
     analyses of complex samples  containing an unknown variety of pesticides and
     interferences.   A proven, completely automated procedure for multiresidue
     SSSlLJ8 M  ±s1U8ually Performed (i.e., extraction,  partition and adsorption
     chromatographic cleanup, and gas or liquid chromatography)  is not  yet available.
          ? 1S Jn2odjctol£ ^tion Is intended as a broad overview of modem
            £2fr2C£ *f hod%and thelr f»Uty control, no details have teen
           ,    ,-?      foregoing material will be discussed more completely
   •p,      T ^' ?*£ 8pec±fic "ferences to relevant sections of the EPA
    Pesticide Analytical Manual or other sources will be given.  A general bibli-
    ography of recent books and reviews on pesticide analysis follows*
16     REFERENCES
    (1)  AOAC General Referee Reports:  Subcommittee E,  J.  Assoc.  Off  Anal
         Chem., 6.2, 376 (1979) (yearly reports).         - ~

    (2)  Berck, B., Analysis of Fumigants and Fumigant Residues, J.  Chromatoer.
         Sci. » 13 ,  256 (1975) .                                  - — a—

    (3)  Bontoyan,  W.  R.,  chairman of editorial board, Manual of Chemical  Methods
         for Pesticides and Devices.  Volumes I and II, U.S. Environmental  Pro- -
         tection Agency, TSD-Chemical and Biological Investigation,  Chemistry
         Laboratory, Beltsville,  MD  (updates every six  months).

    (4)   Bowman,  M.  C., Analysis  of Insect Chemosterilants , J. Chromatoqr. Sci..
         13« 307  (1975) .
                                       -10-

-------
                                                           Section 1G
 (5)


 (6)


 (7)



 (8)


 (9)


(10)


(11)


(12)


(13)


(14)


(15)


(16)


(17)


(18)


(19)



 (20)
Burchfield, H. P., and Storrs, E. E., Analysis for Organophosphorus
Insecticides and Metabolites, J. Chromatogr. Sci., 13, 202  (1975).

Cochrane, W. P., Confirmation of Insecticide and Herbicide Residues
by Chemical Derivatization, J. Chromatogr. Sci.t 13, 246  (1975).

Cochrane, W. P., and Purkayastha, R., Analysis of Herbicides by Gas
Chromatography, lexicological and Environmental Chemistry Reviews, JL,
137-268  (1973).

Dickes, G. J.i The Application of Gas Chromatography to Food Analysis,
Talanta, 26_  (12), 1065-1099  (1979).

Borough, H. W.,' and Thorstenson, J. H., Analysis for Carbamate Insecticides
and Metabolites, J. Chromatogr. Sci.. 13,  202  (1975).

Egan, H., Methods of Analysis:  An Analysis of Methods. J.  Assoc.  Off.
Anal. Chan..  60, 260  (1977).
Fishbein, L., Chromatography  of Triazines,  Chromat'ographic  Reviews,  12,
177  (1970).

Fishbein, L., Chromatographic and Biological  Aspects  of PCBs,  Journal
of Chromatography,  68,  345  (1972).

Fishbein, L., Chromatography  of Environmental Hazards III,  Pesticides,
Elsevier, N. Y.,  1975,  830  pp.

Frehse,  H.,  Problems and. Aspects  of Present Day Residue Analysis,  Pure
Appl.  Chem.. 42,  17 (1975).

Lively,  J.  P.,  editor,  Analytical Methods Manual,  Inland  Waters Directorate,
Water  Quality Branch, Ottawa, Canada,  1974  (periodic revisions).

Magallona,  E. D., Gas Chromatographic  Analysis of  Carbamates,  Residue
Reviews. 56_, 1  (1975).

Malone,  -:B., Analytical Methods for Determination of Fumigants, Residue
Reviews. 38_, 21 (1971).

McLeod,  H.  A.,  Systems for Automated Multiple Pesticide Residue Analysis,
J.  Chromatogr.  Sci.. 13, 302  (1975).

McLeod,  H.  A. ,  and Ritcey,  W. R., editors, Analytical Methods for Pesticide
Residues in Foods, Department of National Health and Welfare, Ottawa,
 Canada  (yearly revisions).   Referred to throughout this manual as the
 Canadian PAM.

McMahon, B. M., and Sawyer, L. D., editors, Pesticide Analytical Manual.
 Volumes I (multiresidues) and II (individual residues), U.S. DHEW, FDA,
 5600 Fishers Lane, Rockville, MD (yearly revisions).  Referred to through-
 out this manual as the FDA PAM.
                                      -11-

-------
                                                            Section 1G

 (21)  Oiler, W. L.,  and  Cranmer,  M.  F.,  Analysis of Chlorinated Insecticides
      and  Congeners,  J.  Chromatogr.  Sci..  13,  296 (1975).

 (22)  Ruzicka, J. H.  A.,  and Abbott, D.  C.,  Pesticide Residue Analysis,  Talanta,
      20,  1261 (1973).                                                   	

 (23)  Schooley, D. A., and  Quistad,  G. B., High Pressure,  High Resolution
      Liquid Chromatography and its  Applications to Pesticide Analysis and
      Biochemistry, Prog. Drug Res.  J3, 1-113 (1979).

 (24)  Shenna, J., Chromatographic Analysis of  Fungicides,  Chromatographic
      Reviews, 19_, 97-137 (1975).

 (25)  Shenna, J., Gas Chromatographic Analysis of Some Environmental Pollutants,
      in Advances in  Chromatography.  Volume  12,  Marcel Dekker,  IncI, 1975,
      Chapter 5, pp.  142-176.

 (26)  Sherma, J., Chromatographic Analysis of  Pesticide Residues, CRC Critical
      Reviews of Analytical Chemistry. August,  1973,  pp. 299-354.

 (27)  Sherma, J., and Zweig, G.,  Pesticides, Chapter  25 of Chromatography.
      Heftmann, E., editor,  Van Nostrand Reinhold Co., 3rd edition, 1975,
      pp.  781-814.-

 (28)  Sherma, J., and Zweig, G.,  Thin Layer  and  Liquid Chromatography and
      Analysis of Pesticides of International  Importance, Volume VII of
      Analytical Methods for Pesticides and Plant  Growth Regulators. Zweig, G.
      editor, Academic Press, N.Y.,  1973, 729 pp.

 (29)  Slade, P.,  IUPAC Commission on Development,  Improvement, and Standardiza-
      tion of Methods of Pesticide Residue Analysis, J. Assoc. Off. Anal. Chem..
      59, 894-910 (1976).                            	~	

 (30)  Smyth, M. R., and Smyth, W. F., Voltammetric Methods for the Determination
      of Foreign Organic Compounds of Biological Significance, Analyst 103  529-
      567  (1978).	

 (31)  Taylor, I.  S.,  and Thier, H. P., Status Report on Cleanup and Determination
      Procedures,  Pure Appl. Chem. 51 (7), 1605-1613  (1979).

(32)  Tindle, R.  C.,  Handbook of Procedures for Pesticide Residue Analysis.
      Technical Pap~ers of the Bureau of Sport Fisheries and Wildlife No.  65,
      U.S.  Department of the Interior, Fish and Wildlife Service, Bureau  of
      Sport Fisheries and Wildlife,  Washington, D.C., 1972, 88 pp.

(33)  Watts, R. R.,  and Thompson,  J.  F.,  editors, Analysis of Pesticide Residues
      in Human and Environmental Samples. USEPA, ETD, HERL (MD-69),  Research
      Triangle Park,  NC (yearly revisions).  Referred to throughout this
      manual as the EPA PAM.
                                     -12-

-------
                                                          Section 1G
(34)   Williams,  I.  H.,  Carbamate Insecticide Residues in Plant Material:
      Determination by  Gas Chromatography,  Residue Reviews,  38.» * <1971)'

(35)   Yip, G., Analysis for Herbicides and Metabolites, J.  Chromatogr.  Sci..
      13, 225 (1975).

(36)   Zweig, G., and Sherma, J., Gas Chromatographic Analysis, Volume VI of
      Analytical Methods for Pesticides and Plant Growth Regulators^ Zweig, G.,
      editor, Academic  Press, N.Y., 1972, 765pp.

(37)   Zweig, G., and Sherma, J., Federal Regulations, Analysis of Insect
      Pheromones, and Analytical Methods for New Individual Pesticides,
      Volume VIII of Analytical Methods for Pesticides and Plant Growth
      Regulators, Zweig, G., editor, Academic Press, N.Y.,  1976, 509 pp.

(38)   Zweig, G., and Sherma, J., Spectrophotometric Analysis and Analytical
      Methods for New Individual Pesticides, Volume IX of Analytical Methods
      for Pesticides and Plant Growth Regulators, Zweig, G., editor, Academic
      Press, N.Y., 1977, 297 pp.

(39)  Zweig, G., and Sherma, J., New and Updated Methods, Volume X of
      Analytical Methods for Pesticides and Plant Growth Regulators. Zweig, G.,
      editor, Academic Press, N.Y., 1978, 593 pp.
                                       -13-

-------
                                                Section  2
                                   INTERLABORATORY QUALITY ASSURANCE
          2A
                                 PR°GEAM °F TBE *** ENVIRONMENTAL TOXICOLOGY DIVISION
             Quality control in the context of this Manual connotes procedures taken
                                *07 ?nd P.recislon of analytical results..  Qualitative
                              determlna,tion8 fcy "siclue analysts are utilized for such
                  »          ^ surve±llance or monitoring of pesticide levels in human
             tissues, some segment of the environment, or the food supply;  if con-
             SSSLfS !ub;e^w?J f?t±ons are to be ^lid, it is vital  that the
             analytical data be reliable.  The complex nature and pitfalls of the
             !^iy i^J Procedures as outlined in Section 1 require a set  of built-in
             controls to prevent or detect incorrect results.  This Manual  is dedicated
             out flr2S?<°Mq   ^ C°?~01 thSt Wl11 ^^"^antly minimize the out-
             put of unreliable and invalid analytical data.   In a legal action,  it is
             not unusual that the testimony of the analyst is evaluated on  the strength
             or  weakness of the operating quality control program in his Lborato^y?
              bo                          °f  the ^ytical Chemistry Branch, EPA-ETD
            Laboratory  in Research Triangle Park, N.C., functions as the coordinating
            unit for a  quality  control program involving various laboratories in the
            EPA regions.  This  program was inaugurated in 1966 by the Technical
            Services Section of. the Perrine Primate Laboratory, Perrine, FL. , before
            the Laboratory was moved to North Carolina.  Originally, the program was
            25?    T° 50nmunljy Pesticide Studies, National Monitoring, ind St^te
            Services Laboratories operating under contract with the U.S. Department
            dL?S  J E,ducat*on%and Welfare, and more recently with the EPA to con-
            duct chemical monitoring for pesticide residues in man and environment.
            e^andJ"? f ?! ^f^'ratory check sample program have now been
            expanded to include other state and private laboratories cooperating with
              tt                                       d±Vlded lnt° two ^Ossifications,
            both of which will be discussed in detail in this and the followine Sections
            The Interlaboratory control program,  which was the first one formalized!
            involves analysis of uniform samples* by a number of participating  laboratories
             The terms  check sample" and "blind sample"  are used  interchangeably for the
             £2 £  Prf^red and distributed by the  coordinating  laboratory^  The former
             term should not  be confused with the other widely used meaning of "check
             sample   (or control sample), that is a sample substrate known to initially
             contain  no  pesticides, and  then spiked to evaluate recovery by a certain
_
                                               -14-

-------
                                                             Sections 2B, 2C


   in order to assess the continuing capability and relative performance of
   each.   In addition, this program indicates,  on a mathematical basis,  the
   degree of confidence that can be placed in the results of sample analyses,
   and identifies analytical areas needing further attention.  The coordinating
   laboratory receives data from the participating laboratories on a special
   report form, processes the data, ranks the laboratories in order of
   relative performance, and distributes the final results.  Details of  these
   procedures and typical sample data are given in Subsections 2D through 2J.

   The Intralaboratory control program,  which will be treated in detail  in
   Section 3, assists a single laboratory in improving the accuracy and  pre-
   cision of data produced by its personnel by providing systematic guidelines
   for top quality analytical methodology and techniques.  One feature of this
   program is the continual, periodic analysis of standard reference materials
   (SPRM's) by each analyst and recording of the results on a graphical  quality
   control chart.  This chart, which is  a plot of the analytical results, vs.
   their  time or sequence, evaluates periodic performance in terms of both
   precision and accuracy and includes upper and lower control limits to serve
   as criteria for remedial action or for judging the significance of varia-
   tions  between duplicate samples.                 "

   A "Statistical Terms and Calculations" subsection at the ,end of Section 2
   will explain some basic terms, equations, and operations used in the  quality
   control-programs for data handling and calculation and statistical evalua-
   tion of analytical results.
                                                                               »


2B   PROGRAM OBJECTIVES

   The objectives of the interlaboratory program are:  .

        a.  To provide a measure of the  precision and accuracy potential of
   analytical methods run routinely by different laboratories.

        b.  To measure the precision and accuracy of results between laboratories.

        c.  To identify weak methodology.

        d.  To detect training needs.

        e.  To upgrade the overall quality of laboratory performance.


2C   PROGRAM ACTIVITIES

   The interlaboratory program includes  the following activities:

        a.  Analysis of interlaboratory check samples by all participants.

        b.  Operation of a. repository to provide any non-profit laboratory
   with analytical grade pesticide reference standards, over 700 of which are
   now available.  These are listed in an index available from the ETD labora-
   tory.
                                         -15-

-------
                                                             Section 2D

        c.  Providing uniform, standard analytical methods in the form of
   an analytical manual also available from the ETD laboratory.

        d.  Quality control of materials of uniform standard quality such as
   precoated GC column packings, cleanup adsorbent, etc.  These materials
   are purchased from commercial suppliers under stringent specifications
   in bulk lots, and distributed in individual units to EPA laboratories  and
   other facilities under formal 'contract with EPA to conduct pesticide
   studies.

        e.  Providing abbreviated, informal, on-the-job training for specific
   requirements.

        f.  Assisting with problems relating to analytical methodology by
   phone, mail, or on-sita consultations.

        g.  Operation of an electronic facility for repair, overhaul,  design
   and calibration of laboratory instruments.
2D   TYPES AND PREPARATION OF SAMPLE MEDIA

   The check sample program is probably the most important interlaboratory
   activity because all allied activities closely depend on it.  Samples used
   in the program are mixtures of pesticides in a substrate ranging from pure
   solvent, in the simplest case, to those media routinely analyzed by the
   participating laboratories, such as fat, blood serum, gonad, brain, and
   liver tissue, water, soil, and simulated air samples.   .

   As an example, a description is given of the preparation and handling of
   a blood interlaboratory check sample by the coordinating laboratory:
   General population serum samples are obtained from a local blood bank,
   typically in 300 ml bottles.  The frozen samples are thawed in a re-
   frigerator (2-3°C), poured together into a stainless steel container  -
   (previously rinsed with acetone), and mixed well.  Approximately 4 liters
   of serum have been sufficient for the program for one year.  Experienced
   chemists analyze the pooled serum to establish the base level profile and
   to be sure no gross contamination is present.  Fart of the sample is then
   divided into small storage bottles with Teflon-lined caps and stored in
   a freezer (-18 to -23°C).  The remainder is stored in bulk in the freezer
   for later spiking.

   Sub-samples are mailed to participating laboratories to serve as their
   interlaboratory check sample and to provide sufficient intralaboratory
   standard pesticide reference material (SPRM) for six months.  Each labora-
   tory supervisor requests in advance the amount of sample required for the
   latter purpose based on his estimated routine sample load (see Subsection
   3D).  A careful study has indicated there is no need to mail the samples
   frozen because neither pesticide nor sample degradation has been observed
   in a 3-to 4-day period.  After removing the amount required for the inter-
   laboratory check sample, personnel at each laboratory sub-divide the re-
   mainder into small vials that are stored continuously in a freezer.
   Individual vials are removed as needed to provide 2.0 ml intralaboratory
   SPRM samples.

                                        -16-

-------
                                                            Section 2E
  The next  time an interlaboratory blood sample is required, the same
  pooled base sample is spiked with pesticides common to blood.  This sample
  will allow the participating laboratories to test their recoveries at high
  pesticide levels, thereby  simulating analysis of routine samples from
  individuals occupationally exposed to high pesticide levels.  Again,
  enough sample will be provided each laboratory to serve both as inter-
  laboratory sample and intralaboratory SPRM's for six months.

  The same  basic procedure is used for other check sample substrates.
  Rendered  chicken fat from  a poultry plant has been used for fat samples,
  while animal brain, gonad, and other tissue check samples have also been
  prepared.  It is anticipated that urine, milk, and soil samples for testing
  certain procedures will be supplied in the future.

  With the  check sample, each participant receives a covering letter providing
  the protocol for handling  the sample.  The time allowed for analyzing and
  reporting results corresponds to the normal time for processing a similar
  routine sample.

  Although  it is presumably  a blind sample to be analyzed along with the  daily
  work load, the interlaboratory check sample will most often be recognized
  as such t»y the chemist at  the time of analysis.  The chemist is likely  to
  give special care and attention  to this sample, and, in addition the best
  chemist in the laboratory  may be assigned the sample in the first place.
  Therefore, poor results on an interlaboratory check sample must be con-
  sidered a serious matter since they will often represent the very best  work
  the laboratory produces.

  The importance of  the interlaboratory check sample program is indicated
  by a number of actions  that were initiated toward standardization based on
  information obtained over  the years.  These include distribution of pre-
  tested Florisil cleanup adsorbent and GC column packings and frequently up-
  dated  standard analytical  methods, and a centralized calibration and
  electronic repair  facility.


2E   REPORTING FORMS

  Laboratories  are  requested to  report  their results on  special forms.  The
   forms  are designed to provide  supplemental operating data in addition to
   numerical results  of  the  analysis.  The  standard reporting form, with de-
   tailed instructions for completion on the reverse  side,  is shown as Table
   2-1.   The data and information supplied by each laboratory include the
   sample size,  extent of  concentration of  the  sample extract, injection
   volumes,  elution cuts  if column cleanup  is required, all  instrumental
   operating parameters,  identity of the GC column, and  the numerical data and
   original chromatograms  upon which all calculations are based.  The  chromato-
   grams must be clearly identified so that  they may be related  to the  data
   on the reporting forms  for easy checking of  the quantitative  results by the
   coordinating laboratory.
                                        -17-

-------
                                                              Section 2F
2F   EVALUATION OF REPORTED DATA

   When the completed reporting forms from the participating laboratories
   are received in the coordinating laboratory, the quantitative results are
   entered on a Summary of Results sheet illustrated in the next subsection.
   If any results appear obviously and grossly erroneous,  the laboratory is
   contacted at once and given a chance to review its work and change the
   report if a simple computational or transcription error is found.   After
   all results are recorded, a statistical analysis of the results is made
   and recorded on the Summary of Results sheet.   A relative performance or
   ranking table is also prepared, establishing a numerical ranking value for
   each laboratory (Subsection 2H).

   After the data evaluations and calculations are made,  the completed report
   forms and chromatograms from the laboratories  with the poorer rankings are
   subjected to detailed examination to determine, if possible,  the reasons
   for the inferior performance.   Availability of the actual recorder traces
   of the chromatograms for study is vital because they allow the coordinating
   laboratory to check such factors as column efficiency,  sensitivity of de-
   tection,  instrumental problems such as baseline noise  and improperly adjusted
   recorder gain,  proper operating parameters to  produce well-resolved peaks,
   inaccurate reference standards, and faulty quantitation procedures.   A de-
   tailed critique is then written, and in cases  of extremely poor performance,
   the laboratory is immediately  contacted by phone to apprise it of  the poor
   ranking and to make suggestions to improve its performance.
                                       -18-

-------
















LU
a!
<

b
LU
U
K
|
eo
<
C£
LU
z

u»
°
oe
O
a.
LU
at.












T-<
C*4
01
t-H
Cfl














!





'


H
M
«
I
u
•
a
0




s
a
,
I
C
s






•
s
0
1
3


5
to
3
3








p
c
I








I ,
•a •
i i
§ i
i
1
 O
-4 *J
•3 ^
e t
& »
e
.
a
|
X 0
V U
a. o
J
» S
C "^
c
M

- i
s -8
M
u
*
§

a a
s o
w
a
o t*
•«. o
f
e
**
u
•2 °
"" s'
1
U. -H
c §

= 2



J,
"*" •—
c o
1 1


1
S'
1













-





















„









-














•



s
«*
w*
|
Hodlflc







;
£
•
1

1
*
I
9
S
t
*

*d
0
s
h













'
















•g
1
1
4*
i *
1 !
j 1
o
E S'
:5 • 'f ™*
1 ,' - - •
1 .-.'I.- 1
u , v • •• u
3 .7-1
• .« i
u —
w **
-« O t»
* U * D
w > to
1 **
S £ S
S S
S £ *
U , U
i *
o
u
** c
*i %4
M K

ri n «




-19-



-------
                       INSTRUCTIONS FOR COMPLETION OF REPORT
 1.
 2.
5.
6.
7.
8.
 If we are to provide you with a critique on the analytical performance the
 data requested on the report form must be complete.  All of it is meaningful
 for a full performance evaluation whether it makes sense to you or not.

 Use one report form for.each GLC column and show under RESIDUE only those
 values you want to appear in the final summary of results.  For- example,
 if a given compound is quantitated on two columns, report only the value
 in which you have most confidence.  Do not leave the choice to us.

 Under the column headed Ml in Final Vol.. the numerical value to be placed
 here should be the final volume after concentration (or dilution).  For
 example, in handling a blood sample, an extract concentration down to 1.0 ml
 might prove necessary for the quantitation of dieldrin.  In this case, the
 figure 1.0 would appear in the column opposite dieldrin.   However, for the
 determination of £,£f-DDE, a dilution of the concentrated extract  up to 10 ml
 may be indicated.  In this case,  the figure 10 would appear opposite £,£'-DDE.

 Each chromatogram of sample and standard shall be sequentially numbered.
 These numbers are then to be written on the appropriate. lines in the column
 headed CHROMATOGRAM NUMBER.  Two  numbers should then appear opposite each
 pesticide reported,  one representing the standard chromatogram,  the  other
 representing the sample.

 Include with the chromatograms  a  standing current profile for each E.G.
 detector being used  in D.C.  mode.   Label each step with the polarizing
 voltage for that step.

 Mail only original chromatograms with your  report,  not  photocopies.   All
 chromatograms  will be  returned.  Fold chromatograms for each  column.in
 accordion fashion from one  continuous roll.   Do not slice up.  If your
 recorder runs  a, ^considerable distance beyond  the  last peak, don't  slice it
 off  right after''the peak.   Let us have" the whole  pen run.

 With your report,  include all chromatograms related to  the  sample work whether
 used for quantitation or confirmation.  However,  for those G.C. columns used
 for  confirmation  only, include all data on  the  report form except a final
quantitative value.

 On submitted chromatograms,  identify  each peak  resulting from a standard
 injection and show the amount of compound represented by the peak in pg
or ng.
                                       -20-

-------
                                                              Section 2G

   The reports from all other laboratories are then scanned to locate any
   irregularities that may lead to future problems.  A general letter is
   drafted', and a copy is mailed to all participating laboratories.  The letter
   discusses common analytical difficulties encountered by several labora-
   tories and offers suggestions that appear to have general applicability for
   improving compound identification and quantitation.  Each laboratory also
   receives a copy of the Summary of Results, with each laboratory identified
   by a code number, and a copy of the Relative Performance Table.  Finally,
   a private critique of performance is sent to each laboratory exhibiting
   special need for help (Subsection 21).


2G   SUMMARY OF RESULTS TABLES AND RESULTS OF CHECK SAMPLE ANALYSES

   Typical summary tables are illustrated as Tables 2-2 through 2-13.
   Definitions and means of calculating the items included in the statistical
   report at the bottom of each table are given in Subsection 2K;
                                        -21-

-------
                                                             Section 2G
TABLE 2-2
          INTBRLABORATORY CHECK SAMPLE NO. 26. MIXTURE OP STANDARDS IN SOLVENT
                                  SUMMARY OF RESULTS .
LAB CODE
NUMBER

45.
*7.
48.
52.
53-
5*.
66.
68.
69.
71.
72.
73.
83.
84.
85.
87.
88.
89.
90.
92.
93.
95.
96.
97.
113.
113A.
130.
135.
137.
160.
161.
162.
163.
164.
Mean
Std.Dav.
Ril.Std.
D»V. ,f
Total
Error(#

PESTICIDES REPORTED IN PICOGRAMS PER HICROLITER
jindane
10
12
10
9.7
18»
14
29*
9.4
10.1
10.4
10
13.8
12
11.4
8.5
8.5
9.5
4.0*
9.0
4.8*
9.0
10
11
12.4
9.4
9.0
9.2
11.2
9.8
11
11
9.7
10.1
14
4.5*
10.5
1.56

14.9

36

Aldrin
10
46*
11
9.2
11
14
3.0*
8.0
10
3*
5.0
8.8
12
14.1
9.6
8.3
10.6
14
8.0
6.9
10
12
8.0
12.1
11
8
7.8
11.0
9.4
10
9.0
9.6
12.2
11
4.8
9.9
2.31

23.4

4.7

Kept.
Epoxide
10
53*
11
8.7
14
14
6.0»
9.2
8.5
	
10
9.4
10
13.5
8.4
8.1
9.6
12
10
5.1*
9.0
81*
8.0
10.9
9.8
8.6
9.4
11.0
9.3
10
10
10.4
10
10
6.3*
10.1
1.61

15.9

33

P.Pf-
DDE
75
379*
88
73
29«
47


66
76
	 T
	
91
64
59
70
75
63
78
70
45
90
22»
80
119*
89
69
72
86
74
73
102
70
77
78
18*
74. C
12.<

17.*

36

Dieldrin
20
145»
	
31
16
18


14
21.6
92*
100*
19.5
. 19
29
10
18
22.8
22
10
11.5
20 •
32,
21
21.7
23-4
16.4
16.8
23.8
14
23
	
16.9
24.3
19
13.4
19.6
5.64

28.8

58

Endrin
30
87*
	
29
15
87*


30
—
—
6o»
42
27
39
15
27
32
12
14
—
—
33

35-5
33.6
29
28
29
25
29
24
28
36
—
24.9
27.7
7.82

28.2

60

o.p'-
DDT
20
32
	
24
	
	


14
65*
24
-—
20
33
32.5
11
20
24
28
23
41
46
31
-__
27.6
41
20.4
20.4
26.5
21
27
— _
18.5
22.6
36
16.2
26.2
8. 5<<

32.6

116

P.P'-
DDT
100
578*
195*
113
—
91
22*
115
99
_J_
150*
98
94
88
94
90
99
94
120
73
US
125
87
150*
110
94
96
112
104
103
97
95
101
96
56*
100
11. (

ll.<

23

Por-
thane
0



379

34



































Ethion
0





























68











o,p'-
DDD
0

































20







P.P'-
DDD
0°





4.














16


.13.








12








Mi-
rex
0





29



































DCPA
0





9.

































Rejt
out)
B-BHC
0























2.5















tod as
era
                                             -22-

-------
                                                                Section 2G
CM
      I
      «



      o



      «  CO




      M  03
      \O
      CM
       i

       w
       w

       ^
       o

       s
       s
       o
       3
       &
       sz
ft*
o




cz
£
M
1-5
g
0
M
s
OS
H
PH
W
0
O
O
M
d.
2:
M
«
w
w
Q
I-l
O
w
w
w
P4


JM
OS
o
LABORA1
*g
euQ
"p.g
ftO

"«g
0|Q
c
•H
T)
W
Dieldz:ih
i
"oiw
PJQ
Q)
•O
+> X
P<0
0) ft
xw
c
h
T3
H
<
i
i
•H
(-3
*
§
O
o
0
0
o


0
(M

vn

o
C\2
xr\
c^.

o
T-)

o
T-t

O
TH


CM
r<
* . *
^O\ONO\CNJOOO«HOOVi-«<*>
CT*ONOOO^^OOONO»-iOONOvOO

ts-ts- ^ CM. \o oo oo ^ «n
C^N^J- i-l O^t C^-T-I C^(*AVO»Af^CMO\
CJCMCM C*.C^CM \O
ONOOooo3»Ac3TH«^CNoeoooscs.
•rtT-t'rtCMT-ICaCNJi-lTHCvaiHCNJ'r^T-l
»A
* • *
^l\A»T»N'«HC>-V>trvCMi}' f^^l-CMvO
{N-C>-C-OCO e>-c*-i>-p-!>-a3eoco>A
T-l
C^C^- ^-jJ-CMO. OvVNOOtn>OMO
O^ONOOOOONOOOsCJvOO Ov P»-
T-4 T-« T-l t-< T-l «-«
I
"J
c^os C^C^-OXC^-OXO^I-IP^O^CM
O O^ O ^4 O^ O^ O^ Os Os CN O O GO D~
i-4 i-l T-l — 1 i-l
•
r- O\*AT-IC>-W^OO cMen P^OO
OsOO\OOOOO\ONOOOOOvvO
i-« 1-lrH i-li-4l-«f-«
>> •
•v n
c •**
0) bfl
^ A'sOi>-ooovOi-«c\itrNJi'
1-4 i-l T-* 1-4 lr« ja TJ
a] IB
•^ 5
\ 0
i-)| "~S
•a
a> m
•P M
O  «H
"•* i-4
0) CO-P
« nj 3
* o
0
J* l>- D-
• • *
OS Jj- ^J" 0
O\ *•*

Ov
PV. SO
-3- «M i-l CO
CM iH -d"
00
>A C>- C^V
jj- T-« CM 0>-
iH i-l CM

SO
OS C*\
O\ ' «^t O V%
T-I i-l CM
c«-
«"* pw
c^. j* »A .*
p>- 1-1
CM
o\ »n <^»
Os O >r\ 1-1
1-4
0
Os so «-«
Os O so C*\
T-»
os \n w>
Os O xrv CM
»-«
•«.
^ ^ ^R-
•H O
•H Q h
CO O
•H » (4
> *r4 L»

 
-------
                                                                                      Section 2G
CM

1
W
      V\
      CM
           3
           00
                      PJ
          o  w
              Q
              O
                                          o\
                        C*~ 00
                          •   •

                        00  ON
                                                                       ONC^-OOONONOO  O
                           CM
                                                                                                  n
                                                                                                   o>

                                                                                                   o
                                                                                                   0)

                                                                                                   3?
                                                                                                  « .
                                                                                                 (U
                                                                                                 •H
                                                                                                     VO
                                                                                                  CM CO
                                                                                                      0 . ON  IN
                                                                                                     CM
                                                                                                         OJ
                                                                                                             CM
                                                                                                 M3  rH  IS-
                                                                                                     •*  CM
                                                                                                      •   •

                                                                                                     O  tH
                                                                                                            C\J;
    §  >^
    •H  
-------
CM
 o
  pen







VO OOCMOOCOOvOOOrH O\ r-ICO

ooooooomo o ovo
i cMCMcncncMcocnsj- co ^-co
o oooooomo o ovo
^^ vO ^^ ^^ r™i ^3" ^3 f^* ^f i^^ . C? *^*
in mminvomvosovo m vom

i-t
o
is
*rt • 4 A
01 4J T3
3 >> . -
to iH cj
co to eu
•* § * 3
o w 3
i— 1 "OHM
ex o o
g o a. w
CO iH 
co m VQ r>.
co o r*. CM
tH *d"
co m r-
oo <• vo r-
m c? i^ f*»
•^







S »^ t^
<0 CD O
S p b
^•4
i-H > 13 W
i-l  4J 0) O
O CO PS H
                                                                                                                                    0)
                                                                                                                                    4J
•H
 O
 O.
 a-
 to
 §
                                                                                                                                    4J


                                                                                                                                    g  .
                                                                                                                                    CO M
                                                                                                                                    
-------
C3
04
           Cft
               I  t-f
               M  C
                 •a
                 iH
                  OJ
                 •H
                                       to
                                                     in
                                                               ir>   in
                                                                                   vo
                                                                                        u  st  in-
                                       oo
                                                                                   00   -H
                                                                              a\   en
                                                      in

                                                 -*   00
                                                 CO   CM
                                                     in
                                                                    m
                                                                              oo   oo
                                                                                        a>
                                                                                                 eo
en   en   co   CM
                             m   vo   co
                                                                                                                oo
                                                                                                             rH r»- a\

                                                                                                             in o en
                                                                                                                CM

                                                                                                             o IH o
                                                                                                             CM
                                                                    CM
                                                                    CM
                                                                    O\
                                                                 CTt OO  CTk
                                                                   •   •   ^
                                                                 CO O  CM
                                                                        CM
                                                                    00
                                                                    m
                                                                 i-< t^.  m
                                                                  *   *   •
                                                                 •it O  00
   VO
                  OJ
                                      in
                                                oo   o   1-1   CM
                                                                              m  vo
                                                                                       sr   m   vo
                                                                                       CM   CM   CM
                                                                                                             CO
                                                                                                            H  >
                                                                                                            i-l  0)
                                                                                                             CO  O
                                                                                                             M
                                                                                                             CU  T3
                                                        -26-

-------
       00
       r*.

s     S
 01
      r--
       a)
      pH

      I



















S3
O
rH
pa
PL,
CA
EH
1
M
E3
E"^
PA
O
5j
PS
en
ta
M
U
H
H
W
P-i





















O
0
1

"<*
(4
e
•H
•O
rH
CU
•H
A
0)
. TJ
4J TH
0. X
u o
w

1
CO.
8
?

"p
O
PH
CJ
Pu


M
A5
s-s
t . P>4
W CO
H C
O
S3




n
CJ

0

-
^
ft
W
O
Q
T
"o.
ex



,J
(
H



























CO
o
in




.19








rH


CO
0





CD
O

0
^
• OJ
a *o
0 O
4 U

...

en
rH



O
*ri^er>
inin«3-«*m«*voin I inmtn ,.oo m m vo

•0
•ii
"

*

OOOvOOOSOOCTvrHenCTlr-tCO O\ O  f~-
rH rH rH rH rH rH rH

m * o o cr> en oo
oor-ii srsrsro-.inc^minc*! r~ jenr-»
es i^ es iH e%^e<4-*evieseMeM en e>jrH


~
•K
cnvoooooNCNOO-OQenevien omm
o\ineNoooNvomcj»r-.oooooo Ornr»oo
- . .


•^•lOP^OOO r-fCM<*^in ^v if) \Q rH O"v <*"> ^O
P^ p^ f~i ^ f«f (sj (\j t^sj pMJ rH



































en
OO CO
in o en
rH




en rH
o> m 



I*i*,
C>  W 4)
o cn oi
                                                                                                                                                  TJ

                                                                                                                                                   (1)
                                                                                                                                                  1
                                                                                                                                                   o.
                                                                                                                                                   ca
                                                                                                                                                   c
                                                                                                                                                   cu    co
                                                                                                                                                   CO    M
                                                                                                                                                   cu    cu

                                                                                                                                                   O.   rH
                                                                                                                                                   (U   4->
                                                                                                                                                   r<    3
                                                                                                                                                         0
                                                                                                                                                   CO
                                                                                                                                                   0)    ca
                                                                                                                                                   3    co
                                                                                                                                                   CO
 
-------
 g
•rl
4J
 O
 0)
CO
  CO
  d,
  9)
  »-<
  •s











55
O
J
H
1|
n
53

§
I]
prf
->
s

o
a
o

3






















H
O
R
1
ft)
M
O
*?
Pj|
fti

^
Vi
"O
rH
0)
•H
O


M
1 0
CO rH
qj O
M CO
H a
5

0)
. -a
11


u

03.

•
O
a
 O i-l S 5 ° S
i^ co ^^ ^^ vo ^o vo i^ i^^ ^i* i^^ ^i* \o vo irt r^ i


vOCNrHsrOOOOOOVfHrH iH OvOsfO'rH
srcoent>.cMcooocoooin -*<-oovorHevi

.


SlnSScovoooA^S voc*i«!n'cNcN§
iH





rHcNcnmsri-i CNCO sf t-)«Ncoi-iincrivorH
»H rH



cocT\cocNeviooco co  <.r'»voin vosrimcovoi-i
'



«^ ^/^ f«^ QQ |M{ ^v| ^^ t/1^ \f) \Q , 	 I ^n O\ ff\ 1O fVj v^
t-j rH r— 1 i— j CN CM J_| ^| fr"5 fT)










VO 00
in o m o
in rH O\ ON
r-f 00


in oo
vo in CN vo
vo «* r~- vo
i—) CM in



ON iH
i*- i-t r» in
rH CO





CN 00
r*» vo oo in
r-. rH rH m
CM sr



OO CO
vO CT> O VO
St O rH in
CN st

Cv| CT»
o\ o oo co
• • • •
vo f-i m cs
rH CH
6^ ^
&* £
at o
^J l«4
*> *T-I r?i
a) 5
Q OS H
C CO
CO tJ rH 4J
0) 4J 0) O
S en P4 H
                                                                                                                                      4J

                                                                                                                                      g
                                                                                                                                      m

                                                                                                                                      2
                                                                                                                                      a.

                                                                                                                                      a
                                                                                                                                      o
                                                                                                                                      •rl
ca
•U
(-1
<0
                                                                          -28-

-------
O
CM


 §
•H
4J
 U
 III
CO
O\
1^
OS
   cr>
   CM
   
p"







£




o
vO







00







1
1



1
1






vO
OS
CTv








vO
vO







en
CM
CO



OS
in







,_,
rH






i
1



sr
oo
•
r-»





o
u-








f:
OS







o
VO




OS
vO







CM
rH






1
1



co
•
VO





0
CM
i-l








en
r-







in
00



"vO
in







sr
*





i—

ev



m
m
,
r—





m
in
rH








en
00







CM
00



CM
in
vO







sr
eM






1
1



VO
m
•
m





vC
e>-








ao
vO
OS







3
as



. sr
m
«-







m
CM






1
1



o
•
vO





O
m
C\'








OS
CM
CO







o
m




m
en
f^-







vO
CM




























































1




O

ir





vo
v£
r-








0
0\







Sf
O



rs.
oo







rH







1




CM
c
m





o
JZ








oo
vO







vO
r-.



CO
m







vO







I




|Z]
,
vO





m
vO
OS*








sr
vO







sr
o



sr
•-







en
i— i






I




m
OS

If





oo
CM
C








co
rr
vO







S
VO



vo
rH
vO







VO
rH






I
1



co
vO

If





en
o
f—








sr
en
ao
9






CN
00



00
m
vO







oo
en





1
•
CM



^
K
r^





m
*
rH








m
OS







^
"



^.
^







CM
m











CM oo eM en
• * • •
vo o sr  rH «^" P*-
• • • •
CM CM vO OS
rH rH en








in m m
CM en vO rH
» •. • •
OO rH VO CM
rH ST







rH in en
00 CM rH r~.
CO rH ST CM
rH en



OO en rH
OS CO OS rH
VO O rH OS
rH CM


B-5 £>S
•t A
> I*
CU 0
. £\ Lj
to
> -a w
CU 4J
Q CO i-H
C CO
CO *VJ rH 4J

-------
o
01
 U
 0)

CO
    CM


     
*o
. -H
w x
a) a,
1

rH
O
2:
a)
•a
o
o

CO




o





o
CO


0
01



o
sr
o





o
00
o






o
•"••
o







t~l
co
o
o
rH
O










0
0
•
o

~ ' ,
: *

iHvOOOOvOOCOO\vOf-«C7\ll^OOI>«l^<^I^CTi



cosfsrsrsfsrsreM. 5cosrcos?co«Mcoco
OOOOOOOOOOOOO OOOO





*
r~\oovor^i^-vOr>eovor^si>t-.in«nvoih
OOrHOOOOOOOOOOOOOO

: , . -


•K
O * *

srinvovof-r^momcoi-'.covovocoi-fH
OOOOOOOOOOOOOO,OHt-(




J
•K *
o, o

COeMCOeMCOCOOeMCOCOCOOCOcMCMCNrH
OO O.O O O CDO O O O OOOO O
O OOOO' O O O O O O O
1 OrH 1 OOOOI rHOOOOOOO
OO OOOO OOOOOOOO





r*H sr in *o i^« oo ct\ rH sr in vo ^ sr tn ^o t^H co sr
1^4 fH rH rH CM* C^l C*l CO CO CO










co oo
r* co
CM C3 sT sT
i-H CO

co O
0 0
00 rH CM CO
i-H CO

0
O 00
sr o
OOOO
CM sr




OO CM
\O >-)
O O 00 VO
I-H sr





^3 i^*
vp rH
o o co sr
CM m






CS»
oo sr
es o
o o m tn
rH CO
O CM
00 CM
o o
o o oo in
CM vO
> ^£ g\g
A *
ai o
o *
^4
> tj m
Q) 4J
Q CO rH
See)
"O rH 4J

-------
 O
 CM



  O
 •H
  4J
  U
  V
••s.
 H







1
3
+1

9
M
H
2
p-t

g
61
S
W

W
§
S
M
H
PM





















1

•1
i

"eJ
2


14
o
Sot
d
H 8


V

J^
a. o
M
,
0
!c
o S
IA

n
<
,
CQ

A
o
Z

ooo o oooooooooooo
* • '
QQ f^ l/^ ^t* ^^ ^^ ^Q ^D .^^ ^3 ^D ^^ f™l "i <^ ^*
-»ooco m tnsf-*c«<»-)esp~i-ioo p o sr




1AO4 (H «^-O\ %OCO CO^OQO OO
cr, ^ ^ eo h»o«nsfcnvSH
OOO' O OOOOOOOOOO.OO


* • -, -. '
csi co ^f vH co in *0 co • oo in in o o *o\ oo tN
(MCMtM CO CSP>I £
00 a M
r-l ^4 iH i-l C4. 
-------
CM






£
HI
£
«
W
cu
M
1
5C
t-t
o
u
1
w
a
H







<
J
*
a
0
i
c
X
14
fSsr
o in
0 CM
%«
£
b
!O
7!
T<
o
?
o.
o.
w
g
"pj
£
3
JS
P-o
aid.
KU
14
45
i«
s •
-1
1
t4
o
i-C
•ss
^•S
0

0. 0 VO
o sr o»
.CM c»4 a *a* «s
0 C 0 C

•* si in \o
rH fs csl CM

9 9
o o
CO
•» . O

O fH vO CM
O in o> en co tH o
•H o o o en o» vo
en vo CA
«
sr eo ch
CM »H fH CM cn
«H IH o fHOr^cocn
O O O O O r-. co eM
CM in en
* en cn
iH O O fH CM
o in sr vOfHcncnr*
fH O 0 OOCOOlfH
• fH en o
rH
CM vO
co cn vo en vo
01 01 rH oimOfHr-
M CM CM cMOcncoen
iH •» CO
4JC
cn en in o
r— in IH co CM
O O O OOfHrHCO
O O O OOmrHCO
CM m 01
m vo m
VO fH fH ST CM
fH rH fH »HOO-»en
o o o oo^cMcn
rt cn 01
CM tO VO CM O
CM in en o» cn
in 9 o o o vo o e
O O O O O CM 00 CM
en vo 01
4JC
CM en 01
fH  -a &
fH HI U tH
o Q cn « M
133**'

















s
ss
bT4
°-3
«
rucis pec B1XJ
*Rejected as ai
                                  -32-

-------
 .o
  CM
   O
   O
   H)
  to
en
CM
I














C1
§
M

iJ
H

otf
P-c

O5
H


ft.
§


8
g
PS
3

jn
3
fn
O
g
M
S
g



0*
CO
w
0
IH
CJ
M
H
£
04



















-







































































































M
O
O u1
O C-
l-i r-l


1

1

*
«4


g
s
i

oil
L
o
rH

J 0)
>% c
t-3



8
S3




o

0>
•a
o



01

































*0
w



a)
0
vC
fl




M
O
VC
o*


a)
O

,
o



"o
en

o












oo
1 CM
i  i
• • IH CM
III II 0 i-4
O I EH EH M i
o i S So i n
o  i-4
en en
,-H o encn
cMm




H en
m t-4 .
o o vo co
CM m



en r^
en o
• •
O O CM en
CM iri



o •SI-
CM o

OOOO
CM v<1
fr-t »<
a « «
(0 > t-t
3 & S

r* > T9 W
iH  -a  « 4) O
O CO OS H
                                                                                                                                                     •O
                                                                                                                                                      01
                                                                                                                                                     1
                                                                                                                                                      01

                                                                                                                                                      §
                                                                                                                                                     •rt
                                                                                                                                                     4J  •
                                                                                                                                                      0)0]
                                                                                                                                                      0)1-
                                                                                                                                                      301
                                                                                                                                                     03 JJ
                                                                                                                                                     •U3
                                                                                                                                                     •HO

                                                                                                                                                     3 CO
                                                                                                                                                       CO
                                                                                                                                                     00
U.U
M O
OH)
Oft-i
01 4)

PS PS
                                                                             -33-

-------
                                                             Section 26


  a.  Check Samples Composed of Standards Dissolved in Solvent

      Table 2-2 shows data from a group of 34 laboratories participating in
  an interlaboratory control program for the first time as a group entity.
  The distributed sample consisted of a precise formulation of eight chlorina-
  ted pesticide and metabolite standards dissolved in pure solvent in a sealed
  ampoule; no cleanup steps were required.  The mean and standard deviation
  values were calculated after rejection of the outlying values designated  by
  asterisks.   (See Subsection 2Kf for description of fitness test).   The pre-
  cision (relative standard deviation)  was considered "good" for this type
  sample only for the compounds lindane, heptachlor epoxide, and P,pf-DDT,
  "fair" for j9,.p_f-DDE,  and "poor" for the other four.   The overall average
  RSD (relative standard deviation;  Section 2Ke)  for all compounds was 21.6Z,
  nearly double the value expected from a group of laboratories with top
  quality analytical output,  such as illustrated by Table 2-3.   Total error
  values considered "good"  include lindane,  heptachlor epoxide, and  £,£.f-DDT,
  aldrin is "fair",  and  the others "poor."  The average total error  was 52*,
  Just outside  the  "acceptable"  level of <50Z.

  Table  2-3 shows,  for comparison, results on the same  sample (except for a
  more difficult, lower endrin content)  by a group of laboratories that  (except
  for one) had been  in the  quality control program for  several  years.  The
  calculated average RSD value is  7.7% and the average  total  error is  20%,
  both "excellent" performance values.   The  average total time  spent  in each
  laboratory on the sample by this group was 1.5  days.  During  the earlier
  years of participation, the data output  of these laboratories was similar
  to that shown in Table 2-2, but continuing participation in both Inter- and
  Intralaboratory Programs resulted in gradual improvement in performance to
  the levels shown in Table 2-3.  As an example of a factor responsible for
 the poor results in Table 2-2, the 34 laboratories used 33 different GC
 columns, while the experienced group represented in Table 2-3 used only
 the optimum GC columns and the operating parameters recommended in the EPA
 Pesticide Analytical Manual and in Section 5 of this Manual.

 b'«   Blood Serum Check Samples

     Table 2-4 shows results for a blood serum check sample reported by 18
 laboratories with experience in the quality control program.  The average
 RSD of  13% and total error of 33Z are quite acceptable for this type of
 sample.

 Table 2-5 shows  results for  a second serum check sample reported by 12
 participating  laboratories.   This sample was prepared from the same base
 lot of  serum used  for a previous sample (No. 31) issued earlier.  It was
 held In a deep freeze for  the. intervening six  months.   The  formulation
 values  on the  summary sheet were regarded as approximate.   The formulation
 was based on the data profile on the sample as  it was  originally  received,
 plus data from the  laboratories analyzing the  earlier  check sample.   The values
were believed to be valid within ca + 15Z.  Precision on this sample is very
                                    -34-

-------
                                                           Section 26

good for dieldrin, £,p/-DDE, and £,£*-DDT and is fair for jo,£/-DDT.  The
total error, embracing both precision and accuracy, is highly satisfactory
for dieldrin and satisfactory for the other three spiking compounds.  Two
laboratories reported traces of heptachlor epoxide and B-BHC, both of which
compounds were probably actually present in trace quantities.

Results of a third serum check sample are given in Table 2-6.  The mean
values reported are very close for dieldrin and £,p_'-DDT (1022 recovery in
each case), high for o^o.'-DDT (130% recovery), and slightly low for £,p_'-DDE
(93% recovery).  No reported values were rejected, but the laboratories with
DDE values below 23 ppb and the one laboratory reporting 34 ppb were cautioned
to scrutinize their recoveries to bring them into a range closer to the
mean.  Interlaboratory precision for DDE and _p_,j>/-DDT was excellent, and,
considering the low concentrations present, the RSD values for dieldrin and
£,£-DDT were acceptable.  The sample was prepared by spiking serum used to
prepare an earlier check sample (No. 35).  Although the formulation values
are reported as approximate, in-house analysis of the final formulation indi-
cated that the correct values were as shown.

Table 2-7 shows the results reported by 16 laboratories for a fourth blood
serum blind sample.  The sample contained three actual residues and three
spiked residues, one of which (PGP) was added only to enrich an already-
present residue.  Residue identity and quantitation were straight-forward
because all peaks were resolved on the recommended GC columns•  A formal
laboratory performance ranking (Section 2H) was .not prepared for this sample
since only HCB and trans-nonachlor were spiked in known amounts to blank
serum.  Calculations were made, however, for review purposes based on the
known values for ECB and trans-nonachlor and the average recovery as the true
values for £,£'-DDE and p_,p_'-DDT.  On this basis, good laboratory performance
was demonstrated by participating laboratories with few exceptions.  Of the
16 reporting laboratories, 9 would have scored above 190, 3 above 170, and
4 in the 116-147 range.  The mean recovery for HCB was 99% and for £cajas.-
nonachlor 110%.  The relative standard deviation (RSD) figures as shown in
the accompanying table demonstrate good precision with the possible exception
of £,j>'-DDT.  The 31% RSD for jB,p_'-DDT appears excessive at first glance but
is certainly understandable considering that the low residue level of 2.6 ppb
is close to the method sensitivity limit.

A 50 ppb spike was added to pooled serum containing PCP in an unknown amount.
Three values were reported of 102, 142, and 190 ppb.  A fourth value obtained
by one laboratory was 180 ppb.  The true value obtained by one laboratory was
180 ppb.  The  true value was probably in the 180-190 range, since analysis of
the unspiked serum in the coordinating laboratory yielded 126 ppb.  Compared
to 180 ppb  in  the fortified sample, this gives a difference of 54 ppb, which
is in excellent agreement with the actual  spike of 50 ppb.  The higher
results obtained by  the  coordinating laboratory and laboratory No. 4 can be
explained by the fact that a revised PCP method including a hydrolysis step
to free conjugated residues was used by these laboratories but not by
laboratories No.  7 and 25.

Although performance on  sample 54 was generally good, significant quantita-
tion error  was observed  in a few instances.  The integrity of standard
                                      -35-

-------
                                                           Section 2G
solutions was suspect in some results, but poor chromatography techniques
were undoubtedly also major factors.  Some specific examples based on careful
analysis of results are:  measuring small responses like 5-6 mm peak heights;
injecting less than 3 yl; large differences in injection volumes of sample
and standard; and large differences in peak heights of sample and standard.
Careless mistakes were also evident.  One laboratory spoiled an excellent
set of data by identifying traris-nonachlor as heptachlor epoxide on the
OV-17/OV-210 column.  A re-injection on an OV-210 column was made, but
identity was evidently not checked against heptachlor epoxide standard also
on the same chart.  This would have clearly signalled an identification
problem.  Two laboratories missed jj^'-DDT.  The GC system did not appear
sensitive enough in one case.  However, from chromatograms submitted by the
other laboratory, it appears that the sample chromatogram was not allowed to
run sufficiently long to elute DDT.  These kinds of problems are discussed
in detail elsewhere in this Manual, especially in Section 5.

The results of 17 laboratories with a fifth serum blind sample are shown in
Table 2-8.  Although this sample represented a rather simple residue mixture,
the necessity for a judicious choice of GC columns is well Illustrated.
Heptachlor epoxide and trans—nonachlor are well separated on the two mixed
phases OV-17/OV-210 and SE-30/OV-210, but not on the single phase OV-210.
Dieldrin and jj.j.E.'-DDE separate on SE-30/OV-210 or OV-210 columns, but an
exceptionally efficient OV-17/OV-210 column would be necessary for accurate
quantitation.  All Office of Pesticide Programs project laboratories involved
in the study correctly identified the six serum residues.  Two non-project
laboratories missed one compound each.  Fifteen of the 17 participants had
performance ratings in excess of 190 points (Table 2-18).

Table 2-9 gives results for a serum blind sample containing mirex,  All
participants except one correctly identified five pesticide residues, while
only 3 laboratories identified mirex. . The 2.ȣ.'~DDE was not spiked but repre-
sented the actual residue in the pooled blood serum matrix.  Mirex was not
included in the scoring but was fortified at the 10 ppb level.  Low recovery
of mirex demonstrates the poor extraction efficiency of this compound in
hexane.  The results of laboratories as 24 and 52 and the coordinating
laboratory indicate an extraction efficiency of ca  25%.  This low level
placed the mirex concentration below the GC sensitivity limit for several
of the laboratories.

c.  Fat Check Samples

    Tables 2-10 through 2-12 show results for fat check samples.  For sample
56 (Table 2-11), all reporting laboratories correctly identified the seven
added pesticides.  Dieldrin was not added as a spike, but was identified by
6 laboratories in amounts of 5-10 ppb.

Sample 70 (Table 2-12) was designed to measure the proficiency of a labora-
tory in recognizing and quantitating PCS contamination in an adipose sample
containing common organochlorlne pesticide residues.  The pesticides and
fortification levels were chosen based on data from national surveys so as
to represent a realistic analytical problem.  In the summary of results table,
                                       -36-

-------
                                                             Section 2H
   laboratories  listed above the double line are Epidemiology/Human Monitoring
   Contract laboratories.   This  sample proved to be a very difficult  challenge
   for the majority of the laboratories.   The percentage total error  (TE)
   figures demonstrate a generally unacceptable level of performance  for this
   analysis.   Five of the nine TE figures were greater than 50% and,  there-
   fore,  "unacceptable."  TE figures for only the EPA contract laboratories
   were:   HCB, 51.3%; 3-BHC, 16. 4%; oxychlordane, 45.7%; trans-nonachlor,
   22.3%; heptachlor epoxide, 43.5%; p_,£'-DDEi 46.7%; p_,£f-DDT, 30.1%;  dieldrin,
   42.7%; and PCB 1254, 120%.


   d.   Water Check Sample

       Table 2-13 shows the results of an interlaboratory water round robin
   sample reported by 19 laboratories in 1977.  The sample contained  spikes
   of four parent organochlorine pesticides plus Aroclor 1254, but no partially
   altered compounds as would undoubtedly be present in routine monitoring
 •  samples.  The results .shown in Table 2-13 are acceptable considering the
   relative difficulty of the sample.  Interlaboratory precision,  as  measured
   by the relative standard deviation values, are reasonable, and total error
   values, although above the 50% level considered satisfactory for less
   complex formulations, are not too far off on this particular sample.
   Mean recovery values for all laboratories, after rejecting outliers, were
   HCB, 67%, oxychlordane, 83%;  p_,jE.'-DDE, 85%; p_,p_'-DDT, 86%; and Aroclor 1254,
   81%.  The value for HCB is not as bad as indicated because the best re- ,
   covery possible for HCB was 85% (0.25 ppb).


2H   REIATim PERFORMANCE RANKING

   a.  Original Performance Ranking Scheme

       A scheme has been used from the start of the QC program until 1980 for
   the relative ranking of laboratory results in the analysis of multiresidue
   check samples.  This scheme, described in this section, was used to calcu-
   late the rankings shown in Tables 2-15 to 2-22.  A new scoring procedure
   that has been adopted for future interlaboratory samples is described in
   Subsection 2Hb.       .

   There are three essential criteria for a high score in the performance
   ranking, namely,  correct identification of all pesticides present, correct
   quantitative assay of the pesticides, and non-reporting,of pesticides not
   present.  The ranking scheme incorporates all three criteria and provides
   a numerical score for each.

   The maximum possible score is 200 points, 100 for correct identification
   and 100 for quantitation.r A detailed explanation of the calculation pro-
   cedure  follows:

   (1)   Identification

         The  100 possible total  points divided by the number of compounds
   actually  present  yields  the  point value per  compound.  Correct identification
                                        -37-

-------
                                                           Section 2H
 of all compounds present and reporting; of no extra compounds results  in a
 total score of 100 points.   A penalty''equal to the point value per compound
 is assessed for each compound reported that is not actually present.   For
 example, if five compounds  are present in the check sample, each is worth
 20 points.   If one is missed and one extra is reported,  a penalty of
 2x20-40 points will be assessed against identification performance.  The
 score in this part would then be 100-40»60 points.

 (2)  Quantitation

      The point value per compound is again the total possible points  (ldp)t
 divided by  the number of compounds present.   Should all  reported values  *
 coincide exactly with the formulation values (an unlikely situation),  tfit
 full 100 points are awarded.   When a reported value differs from;the  formula-
 tion value,  the difference  between the two (the absolute error)^'Ijjte&ed  by
 the standard deviation (previously calculated for  each compound)" |?£yeS a
 "weighted deviation."  This value is subtracted from the point value  of  the
 compound to  give the quantitative score for that compound:
Compound Quantitative      Compound Point
Score                 "   Value
Absolute Error
Standard Deviation
The total score  for this part is the sum of the individual compound quantita-
tive scores.

An important aspect of the quantitative portion of ranking is the role
played by^the standard deviation for each compound.  If the precision of
the group for the analysis of a particular pesticide is poor, the standard
deviation for that compound will be relatively high.  If a laboratory has
a large absolute error for this one compound but an otherwise excellent
performance, division of the error by the large standard deviation will
lower the point  loss so that an unduly heavy scoring penalty is not received.

(3)  Total Score and Sample Results

     The total score for laboratory performance is the sum of the identifica- "
tion and quantitation point totals.  Table 2-14 illustrates in detail the
method of calculation for a hypothetical analysis in which one compound is
missed and one extra is reported, resulting in an inferior total score of 125.
                                     -38-

-------
                                                      Section 2H
Table 2-14
        CALCULATION OF TOTAL SCORE FOR RELATIVE PERFORMANCE RANKING

6-BHC
£,£'-DDE
Dieldrin
£,£'-DDT
£,£'-DDT
Ct-BHC
Formulation Reported Analytical
pg/Vl Values pg/vl
30 27
40 40
20 50
10 Not Reported
50 47
None 10
Standard
Deviation*
2.10
1.75
2.50
0.60
1.44
— -
*0f all data from participating laboratories
Point value for
Identification
8-BHC
£,£'-DDE
Dieldrin
O,£-DDT
£,£' -DDT
each compound is 100 -5- 5 » 20
- 20
•20
20
0
— , 2£






 Quantitation

 B-BHC


 p_,p_'-DDE


 Dieldrin


 c_,p_-DDT

 p_,£'-DDT
sum «• 80
     -20  Penalty for false identification  of oc-BHC
      60  Total identification points


             30-27
      20 -


      20 -
  2.10

40 - 40
  1.75

20 - 50
  2.50



50 - 47
  1.44
= 20


=  8


=  0

- 18
                      Tb€al quantitation points  — 65

 Total laboratory  score  60 + 65 = 125 (of a possible 200 points)
                                     -39-

-------
                                                            Section 2H

  Tables 2-15 through 2-22 show Relative Performance Rankings  for groups  of
  laboratories on check samples of different types.   Table 2-15  shows  rankings
  for the laboratories reporting the data in Table 2-3.   Laboratories  with
  scores over 190 are considered to have demonstrated generally  acceptable
  performance with some possible minor problems.   Scores  between 150 and  190
  indicate definite problems  that should be quickly  resolved.  Those with
  scores below 150 should suspend all routine pesticide analytical work
  pending the resolution of very serious problems  in both identification  and
  quantitation,  the effects of  which place in doubt  all routine  analytical
  data output of the laboratories.   The laboratories are  to initiate corrective
  action immediately based on the general and individual  critiques received
  and personal consultations  with the coordinating laboratory.   The remaining
  portions  of the original check sample can be used  as a  reference standard
  material  to internally evaluate improvement before receipt of  a new  check
  sample to again test  laboratory performance.

  Each set  of performance  ranking data must be carefully Appraised by highly
  skilled,  experienced  residue analysts  in the coordinating laboratory before
  deciding upon what, if any, action  should be taken based on. the results
  For  example, Table 2-16  shows ranking  data  for 17 laboratories analyzing
  a fat sample (results  reported  in Table 2-10) and Table 2-17 a blood analysis
 performed by 16 laboratories.  : Examination of the scores for the fat sample
 Indicates a significant breaking point between laboratories  with 185 or more
 points and those with 168 or lower.  Reference to Table 2-16 shows  that those
 below the break point had readily apparent problems, and these four labora-
 tories received corrective critiques.  To the contrary,  all  rankings  for the
 blood analysis were 192 or greater, and all laboratories were considered'to
 have turned in acceptable performances, even those with the  poorest  relative
 scores.


 Performance  rankings for laboratories participating in the analysis of some
 later blind  samples are shown  in Tables 2-18 to 2-22.  Rankings for blood
 serum sample No.  59 shown in Table 2-18 indicate  that only two  laboratories
 scored  below 190;  one laboratory scored below the 150 level indicating very
 serious problems requiring immediate resolution.  Performance on  serum sample
 No.  66  (Table 2-19)  was also generally excellent, with all but  one laboratory
 scoring above 190  points.  Performance on fat sample No. 56 (Table 2-20) was
 also excellent, with 12 of the 16  laboratories scoring over 190 points and a
 low  score  as high  as  175.9.  Since there were no missed  compounds, the scoring
 spread  reflects  entirely the ability of the  laboratories to accurately
 quantitate residues.   The lowest scoring laboratory  did not use proper
 standards  for some quantitations,  so accuracy was understandably poor.  The
 12*  recovery for heptachlor  epoxlde  by  laboratory 16 was also understandable
 because  the  analyst attempted to quantitate  a peak height of only 8  mm against
 S.,S;  T  stto?ard   Laboratories with
scores above 170 had only minor problems, if any.   Below this/there was
sharp drop-off to a score of 115.9.  The score of zero for the lowest labora-
tory, which was a new participant in the EPA AQC program, resulted because
tne sum of-the penalty points exceeded the positive points.
                                          -40-

-------
                                                    Section 2H
Table 2-15
                      RELATIVE PERFORMANCE RANKINGS




                CHECK SAMPLE NO. 26, MIXTURE IN SOLVENT
Lab. Code
Number
161.
137.
135.
162.
87.
113A.
113.
85.
48.
130.
66.
73.
72.
84.
89.
88.
83.
96.
97.
164.
68.
92.
93.
90.
53.
163.
95.
160
45.
71.
52.
47.
69.
54.
I/ Values
2_/ Total
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
1
1
1
2
2
0
3
2
3
4
4
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
1
0
0
1
0
1
0
o
1
0
0
4
No. of
Rejects I/
0
• 0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
2
0
4
1
0
2
2
1
0
0
0
6
3
2
1
2
4
Total
Score 2_/
198
198
197
197
197
197
196
196
195
195
195
194
194
192
192
189
189
187
181
169
168
168
164
159
158
157
146
133
128
127
123
115
84
25
outside confidence limits
possible score
200 points


                                        -41-

-------
                                                     Section 2H
Table 2-16
        REEATIVE 'PEKFOKMANCE RANKINGS - CHECK SAMPLE NO. 21, PAT
Lab. Code
Number
15.
16.
8.
25.
7.
4.
26.
. 33.
5.
34.
11.
9.
31.
6.
1.
14.
24.
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
f
1
1
0
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
No. of
Rejects I/
0
0
0
0
0
0
0
1
0
1
1
1
0
0
0
0
3
Total
Score 2/
198
198
198
197
195
193
191
191
191
189
188
187
185
168
168
167
165
 I/   Values outside confidence limits
 2/   Total possible score 200
                              -42-

-------
                                                      Section 2H
Table 2-17
       RELATIVE PERFORMANCE  RANKINGS  - CHECK SAMPLE  NO.  23,  SERUM
Lab. Code
Number
1.
2.
3.
4.
5. •
6.
7.
8.
9.
10.
11.
12.
13. '
14.
15.
16*
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
b
0
0
0
0
i 0 '
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
No. of
Rejects I/
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
Total
, Scorei 2/
198
198
197
197
197
197
196
196
196
196
196
195
194
193
192
192
  I/  Values outside confidence limits
  2/  Total possible, score 200
                                        -43-

-------
                                                                Section 2H
Table  2-18
                       RELATIVE  PERFORMANCE RANKINGS
                      CHECK SAMPLE NO1.  59,  BLOOD SERUM
"Lab Code No.
8
15
26
25
4.
11
14
5
1
13
7
16
6
32
12
9
34
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
False
Identifications
0 .
0
0
0
0
0
0
0
.0
0
o
0
0
0
0
0
o
No. of ., ,
'Rejects —
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
4
Total , ,
Score -'
19«.04
197.72
197.31
197.08
196.54
196.51
195.90
195.41
195.30
194.37
194.05
193.49
193.30
191.60
191.52
162.41
147.15
JL/ Rejected as outliers
2j Total possible score - 200 points
                                  -44-

-------
                                                               Section 2H
Table 2-19
                      RELATIVE PERFORMANCE RANKINGS
                     CHECK SAMPLE NO. 66, BLOOD SERUM
Lab. Code No.
52
25
12
4
8
24
38
26
7
14
1
13
5
6
16
11
Compounds
Missed
o.
0
0
0
0
0
0
0
0
0
0
0
0
0
. 0
1
- False
, Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
No. of , ,
Rejects r-
0
0
0
o
0
0
0
0
0
0
0
0
1
0
0
0
Total -,
Score -'
199.3
197.7
197.6
196.9
196.7
195,7
195.5
195.5
195.4
194.1
194.0
193.8
192.0
191.0
190.8
153.1
 I/  Rejected as  outliers
 27  Total possible score  - 200  points
                                   -45-

-------
                                                                Section 2H
 Table 2-20
                       RELATIVE PERFORMANCE RANKINGS
                         CHECK SAMPLE NO. 56, FAT
Lab. Code No.
15
38
1
25
11
8
12
4
6
14
34
7
$
26
16
31
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
I
Otj
0
0
0
No. of I/
Rejects
0
0
0
0
o-
0
0
0
1
0
0
0
0
1
2
o
Total _,
Score -'
197.60
196.29
195.42
194.87
193.67
193.17
192.59
192.57
192.19
191.89
191.63
190.54
190.29
189.09
184.69
175.87
JL/ Rejected as outliers
2f Total possible score - 200 points
                                  -46-

-------
                                                              Section 2H
  Table 2-21
                    RELATIVE PERFORMANCE RANKINGS
                      CHECK SAMPLE NO.  70,  FAT
Lab. Compounds
Code No. Missed
51*
8*
26*
38
4*
52
25*
11*
7* ,
16
14*
13
9
12*
24*
6
10*
1
0
0
0
0
0
0
0
1
0
1
1
1
1
1
, 1
2
1
2
False Identification
Identifications Score
0
0
0
0
0
0
2
0
1
0
0
0
0
1
1
0
2
3
100
100
100
100
100
100
77.78
88.89
88.89
88.89
88.89
88.89
88.89
77.78'
77.78,,
77.78li
66.67
44.44
Quantitation Total
Score Score
95.57
92.89
91.17
90.37
90.24
73.80
94.03
81.30
80.99
79.14
77.68
77.12
75.95
80.62
78.38
73.33
83.21
69.10
195.57
192.89
191.17
190.37
190.24
173.80
171.81
170.19
169.88
168.03
166.57
166.01
164.84
158.40
156.16
151.11
149.88
113.54
* Epidemiology/Human Monitoring Contract Laboratory
                                     -47-

-------
                                                                  Section 2H
  Table 2-22
                         RELATIVE PERFORMANCE RANKINGS

                          CHECK SAMPLE NO.49, WATER
Lab. Code No.
5
3
25
15
8
11
34
36
13
26
7
10
16
1
9
24
12
6
23
Compounds
Missed
—
—
—
—
—
—
—
—
2
2*
2
2
1
3
1
3
3
4
3
False No. of . ,
Identifications Rejects — '
— — , __
— —
— ' __
— __
—
— .
1
1 1

—
. — 1«
1 — —
,4
,ri
4 1
1
1 Ma^
--. -._
4 1
Total 0>
Score -7
195.72
195.37
195.14
194*31
191.87
188.93
183.37 '
171.40
115.92
115.49
114.42
94.0
80.0
78.75
74,22
73.74
56.57
39.14
0.00
*  Later reported the presence of the two compounds
_!/ Rejected as outliers
2/ Total possible score - 200 points
                                        -48-

-------
                                                           Section  2E
b.  Current Performance Ranking Scheme

    The new scheme adopted by the coordinating laboratory for ranking
laboratory performance on interlaboratory check samples differs from the
original only in the scoring of the quantitative results.  The purpose of
the change is to cause a greater point loss for laboratories with signifi
cant quantitation errors.  This, in turn, will improve the relative per-
formance of laboratories with more accurate results.

The new procedure involves dividing the absolute error by the standard
deviation (SD) to obtain the "weighted deviation" as before (Subsection
2Ha).  The score for each compound is obtained by squaring the weighted
deviation and subtracting from the compound point value.
             Weighted deviation        Point loss

          0-1 standard deviations         '0-1
          1-2                              1-4

          2-3 .                             4-9

          3-4                              9-16

          4-5                             16-25

The scoring penalty cannot exceed the point value per compound.  It is
felt that this approach more fairly penalizes large errors but is not overly
harsh for results with small errors .

As a specific example, -the quantitative scoring shown in Table 2-14, would
change in the following manner under the new scoring system.
Compound

  MHC
          Dieldrin
                       "Weighted
                        deviation"
                        1.4


                        0


                       12
  Points
Subtracted

  1.96


  0


 20
                                                            20-20   « 0
          j^'-DDT
               not
             reported
                       2.1
 4.41
                                                            20-20
20-4.41 ** 15.6
                                            Total quanti-
                                            tation points
                                                        53.6
The major difference is the loss of all quantitation points for dieldrin,
which certainly seems fair considering the deviation of 12 units.  As
before, all points are lost for compounds not correctly identified (£,jJ/-DDT)
                                      -49-

-------
                                                                 Sections 21, 2J
   Experience will have to be accumulated using the new scoring method in order
   to assess how the numerical values for satisfactory performance will vary
   compared to the scores calculated with the old formula.


21   PRIVATE CRITIQUES

   As already mentioned, laboratories with significant analytical problems receive
   added assistance in the form of a private, individual critique of their results
   reported for a check sample.  The content of this critique depends upon the
   problems that ar£ obvious from a careful analysis of the submitted results and
   might include comments on incorrect standards, instrumental factors,  calculation
   errors, poor choice of materials or parameters, etc.


2J   PROGRESSION OF PERFORMANCE.

   Buring the early years of the Interlaboratory Quality Control Program,  results
   were expectedly poor.  Methodology, equipment, and reagents were a matter of
   individual laboratory preference, and the first priority was development of
   uniform methodology and standardization among all laboratories.

   As an example of the improvement attainable from such a program, the recovery
   and precision results on interlaboratory fat check samples for a group  of
   human monitoring laboratories over a period of years are shown in Table 2-23.
   The method used in the first year was based on gas chromatography of a  concen-
   trated tissue extract without cleanup (1).  Although the method was fairly rapid
   and simple, it was discovered that the GC column and detector became rapidly
   contaminated by repeated injection of uncleaned samples, and the check  sample
   results proved the method was unsuitable for routine use by a laboratory net-
   work.  Not only was precision poor as measured by the RSD, but the spread from
   minimum to maximum recoveries for several compounds was extremely wide,  and mean
   recoveries were generally far from correct.

   Conversion was made to a procedure including cleanup of the extract by  ace-
   tonitrile/petroleum ether partition and Florisil column chromatography  (2),
   resulting in significant improvement in not  only precision (sample 9, Table
   2-23) but in accuracy as well.  After several months'  experience with the method,
   results on another check sample (sample 11,  Table 2-23) were even better, and
   with continued participation in the program, the laboratories made still further
   progress in their performance through 1974 as the figures in the Table  show.
   The 1978 fat check sample results indicate an apparent reversion to 1972
   precision levels.   Since methodology has not changed,  the results apparently
   reflect need for reestablishment of a training program for pesticide  residue
   chemists such as was once conducted by the EPA Perrine Primate Laboratory in
   Florida.  The 1979 results are also indicative of training needs plus the
   significant complications of analyzing organochlorine  pesticide  residues in
   the presence of PCBs.

   The results in Table 2-24 show progression of laboratory performance on  inter-
   laboratory blood check samples between 1968  and 1979.   In the beginning  the
                                            -50-

-------
                                                               Section 2J
 Table 2-23
                           PROGRESSION OF RESULTS
                              FAT CHECK SAMPLE
Interlaboratory
Sample Number
3
9
11
21 -1
24
28
56
70
Year
1967
1968
1969
1972-
.1973
1974
1978
1979
No. of
Labs
15
21
19
"16
14
10
16
18
No. of
Compounds
7
7
7
7
7
7
7
Q •
Average
Recoveries ,
-I/
-I/
108
89
95
96
92
91
Average
RSD,
50
38
24
19
14
12
19
23
JL/ Complete data given in Table 2-10
2J Unspiked samples for which exact pesticide  levels were unknown
JJ/ Sample contained eight single component compound plus Aroclor 1254
                                       -51-

-------
                                                                 Section  2J
  Table 2-24
                            PROGRESSION OF RESULTS
                              BLOOD CHECK SAMPLE
Interlaboratory
Sample Number Year
6
10
16
17
22
23
25
27
31
33
40
46
54
59
66
1968
1969
1970
1971
1972
1972
1973
1974
1975
1975
1976
1976
1977
1979
1979
No. of No. of Average
Labs Compounds Recoveries,
22
20
22
20
17
17
18
15
17
13
14
14
16
17
16
6 *
5 *
4 *
4 *
4 96
4 91
4 100
3 *
4 *
4 *
4 *
4 *
4 *
6 *
5 *
Average
RSD,
36
29
21
17
14
12
13
16
20
13
17
18 ;
18
20
15
* Unspiked samples for which the actual pesticide levels were not known
                                           -52-

-------
                                                                 Section 2K

   direct hexane extraction method of Dale e.t al.  (3)  was adopted but was found
   to yield poor interlaboratory precision.  Sample 10 was analyzed by a triple
   extraction modification of this procedure, which also proved inadequate.
   The later samples were done with the currently recommended Thompson and Walker
   (A) extraction method, which utilizes a constant speed mixer (Subsection 9D).
   The results of the blood check samples illustrate again the dual value of the
   Interlaboratory Control Program in upgrading laboratory performance and in
   identifying weak analytical'methodology.


2K   STATISTICAL TERMS AND CALCULATIONS

   a. Accuracy and Precision

      Precision refers to the agreement or reproducibility of a set of replicate
   results among themselves without assumption of any prior information as to the
   true result.  Precision is usually expressed in terms of the deviation, variance,
   or range.  Accuracy is the nearness of a result or the mean of a set of results
   to the true value.  Accuracy is usually expressed in terms of error, bias, or
   percentage recovery.

   Good precision often is an indication of good accuracy, but one can obtain
   good precision with poor accuracy if a systematic (determinate) error is present
   in the method used.  Systematic errors are either positive or negative in sign.

   The other general classification of errors in analysis is indeterminate  (random)
   errors.  These are errors inherent in the analytical methods because of un-
   certainties in measurements.  An example is the measurement of the height and
   width of a gas chromatographic peak with a ruler, which requires estimation
   between the mm division lines.  Indeterminate errors are random, that is, they
   are Just as likely to be positive as negative.  For this reason, the average
   of several replicate measurements is always more reliable than any of the
   individual measurements.  Although random errors are unavoidable, determinate
   errors can be corrected once their cause is located.
                                 •0
   Standards of accuracy 'and precision are not the same for a residue analysis
   as for a macro analytical method such as a titration, for which a precision and
   accuracy of 1-5 parts per thousand is usually expected of an experienced analyst.
   The analysis of technical pesticide products is also a macro method for which
   accuracy and precision are fundamental factors, and the measurement step
   (usually internal standard GC or LC) must be carried out with this in mind.
   In contrast with macro methodology, residue analysis Involves the assay of
   nanogram or lower amounts of pesticides, and with the extensive cleanup and
   great amount of experimental manipulation required, procedures are considered
   adequately quantitative when values + 15-20Z or  better are obtained on re-
   covery samples fortified at ppm levels, + 30Z at ppb levels.  One authority
   has suggested that a  rela'tive standard  deviation or coefficient of variation
   (Section 2Ke) of less than 40% is acceptable for precision between laboratories
   for a  trace analytical method.  A model has been presented (5) to analyze the
   reproducibility of results of determinations of unknown amounts of pesticides
   in relatively few samples.  The reliability of the analytical procedure, the
   influence of sampling techniques, and the number of samples that should be
   analyzed can be determined with the model.
                                           -53-

-------
                                                              Section 2K


Absolute error is the difference between the experimental result and the true
value.  Relative error is absolute error divided by the true value"and multi-
plied x 100 to yield percent relative error or x 1000 to yield parts per
thousand relative error.  As an example, an absolute 0.2 pi error in injection

of a sample for GC corresponds to    * .,*n    - 20% for a 1.0 ul sample but only
                                      1.0

            4% for a 5.0 ul sample.  It is explained later in Section 5 that
0.2 x 100
   100
low sample injection volumes are to be avoided because of high potential errors.
Bias is defined as the mean of the differences (having regard for signs) of the
results from the true value.
b.  Significant Figures

    The uncertainty of a piece of data is assumed, to lie in the last digit re-
corded, and unless qualifying information is given this last digit is assumed
uncertain by i-1.  If the height of a GC peak is reported as 10.0 cm, the
absolute uncertainty is + 0.1 cm, and the relative uncertainty is
    x 100 » 1%.  Likewise, one should always be sure to record all certain
0.1
10
figures plus one uncertain figure in a measurement, these figures being desig-
nated as significant figures.

Only significant figures should be used in recording and calculating analytical
results.  If the value 12.3 mg/g is reported for a pesticide analysis, the 12
should be certain while the 3 is more or less uncertain.  Good judgment on the
part of the analyst is required to decide on the proper number of figures so
that significant digits are not lost or non-significant ones retained.  All
numbers written after the first real number are considered significant.  The
numbers 1.23, 12.3, and 123 all have three significant figures.  Zeroes can
cause some problems and should be paid special attention.  Zeroes written
before the first real number are not significant but merely serve to locate
the decimal place,  Therefore, the numbers 0.123, 0.0123, and 0.00123 all have
three significant figures; the number 0.1012 has four significant figures since
the second zero follows the first real number (in this case 1) and is, there-
fore, significant.  All terminal zeroes following a decimal point are significant,
For example, 9.800 g indicates a weight of 9.8 grams accurate to the nearest
1 mg.  All four figures are significant.  The number 10,100 should indicate
five significant figures, but terminal zeroes in a whole number must be con-
sidered with suspicion because the proper rule is not carefully followed.  If
the value 10,100 mm indicates that measurement was made to the nearest 1 mm,
the absolute uncertainly is + 1 mm and the relative uncertainty
        x 100 « 0.01%.  If the measurement was actually made to the nearest
10,100
0.1 meter and the final zeroes only indicate the magnitude of the number in mm,
the number would better be written in exponential form,  1.01 x 10^ mm,  to indi-
cate an absolute uncertainty of i 1 x 10* mm and a relative uncertainty of
                                        -54-

-------
                                                              Section 2K
Significant figures should be properly retained when performing mathematical
operations.  Simplified rules that serve in most cases are as follows.  In
addition or subtraction, the answer has as many decimal places as the number
with the fewest decimal places.  For example:
7U. / — — —
8 PI ___
A ft CC1 ___

——-——.— ~y* i
80
_..,__..___ n fi
—_.__«___ U.D
                                        100.1


Inspection of the three numbers to be added indicates the answer can have only
one decimal place.  Each number is initially rounded off to one decimal place
and then the sum is taken.  Note that the correct answer has four significant
figures (even though each number added had only three) but only one'decimal
place.  Rounding off is done by rounding the last retained digit up if the
discarded digit is greater than or equal to 5; the last digit is retained
unchanged if the discarded digit is less than 5.  For multiplication and
division, the answer can have no more significant figures than the number with
the fewest significant figures.  For example, in calculating the ng of pesti-
cide represented by an unknown GC peak by comparison with the area of a standard

peak, the formula he  « nga  	IH.  is used if response is linear over the
                    **     s  &irs3a
range in question.  If 1.0 ng standard gives a peak of 9.0 cm height (measured
to the nearest 1 mm) and the unknown peak height is 12.0 cm, the ng of unknown is
1.00 X
       12.0
1.3 ng with only two significant figures reportable.   If an
analysis is based on peak areas calculated by the usual formula height x width
at one-half height, a width of less than 10 cm measured only to the nearest
one mm limits the area and the calculated amount of pesticide to two significant
figures.

c.  Average

    The average or mean (X) of a set of n values is calculated by summing the
individual values and dividing by n:
                                           I Xi
d.  Range

    The difference between the highest and lowest values in a group.
                                      -55-

-------
                                                               Section 2K
e.   Standard Deviation and Variation

    Standard deviation(s) of a sample of n results is calculated by use of the
 equation:                —                     ,/«
                                7    fZ y-n
                            Z xi  -  ** X:LJ
                          __        n-1

 Variance xs equal to, a .   Relative standard deviation (RSD)  or coefficient
 of variation (CV) is the standard deviation divided by the mean and multiplied
 by 100 (percentage)  or 1000  (parts per thousand),   cr is the  standard  deviation
 for a very large set of data,  calculated by the above equation with n rather
 than n-1 in the denominator.   Precision is increased (value  of s reduced) by
 increasing the number of replicate analyses, enabling one to determine with
 greater statistical  confidence that the true mean lies within certain limits
 about the experimental mean  or to reduce the interval at a certain confidence
 level.   Confidence limit or  interval is defined as:
                         U - X +
                                    ts
where y is the true mean,  X is  the experimental mean,  and  t  is  a value obtain-
able in tables for different percentages  of confidence and numbers of  trials
 (n).  Values  of  t  increase as percentage confidence  desired  increases and
decreases as  the number"of replicates increases.

f.   Fitness Test

     EPA Quality Assurance  personnel have  applied  the following test for re-
jection of "outlier" values in  check sample data, which,  if left  in, would
exert a significant effect on the overall data:

      (1)  Compute the mean  and the standard deviation of the entire data set'.

      (2)   Compute the  absolute  value of the arithmetic deviation  from  the
mean of all values in  the  data  set.   ,

      (3)   Establish the correct factor to be used in the  calculation (Step 4)
from the following table.
                                        -56-

-------
                                                              Section 2K
                Number of data points (n)
                    in the data set
                          5
                          6
                          7
                          8
                          9
                         10
                         12
                         14
                         16
                         18
                         20
                         25
                         30
                         40
          Factor
           1.65
           1.73
           1.81.
           1.86
           1.91
           1.96
           2.04
           2.10
           2.15
           2.20
           2.24
           2.33
           2.39
           2.49'
     (4)  If the absolute value of the arithmetic deviation from the mean for
any number in the data set is greater than the, factor from step (3) times the
standard deviation of the entire data set, the number is rejected as lying
outside a reasonable data set.
     (5)  The percent confidence interval for the retained values would be
given by:
                                 (1 -
-)  100
This Fitness Test has proven to be practical and reasonable over many years
with round robin interlaboratory blind sample exercises wherein proven
methodology is used.  It is based in part on Chauvenet's criterion as des-
cribed by Hugh D. Young (6).  Individual statisticians disagree on the best
test for rejection of questionable results, and no claim is made for the
rigorous statistical validity of the method described in this subsection.
                                      -57-

-------
                                                                Section 2K
 g.   Total Error
     Total error ±8 a. method proposed by McFarren et al.  (7)  for combining  pre-
 cision and accuracy in one reporting expression:
                                  Absolute Value of
                   Total Error  - the Mean Error
            +2s
                                  True Value
                x 100
where s - standard deviation.   In general,  total error values <  25% are  con-
sidered excellent, < 50% acceptable,  and >  50% unacceptable.
 Specifically,  in. the interlaboratory control program,  total error is calculated
 from the following  equation:
                            Total Error *
                                          x + 2s
where x »  the arithmetic deviation of  the overall mean obtained for a given
pesticide  from  its known formulation value  (the absolute value of the mean
error), y  •  the formulation  (true) value, and s • the standard deviation.  A
discussion of this equation  has recently been published (8), indicating it
may unnecessarily downgrade  a  considerable portion of results.  Alternative
equations  are recommended which rigorously meet the McFarren et al. 25 or 50%
criterion  with  at least 95%  confidence.  These equations are:
                                     x 4- 1.7 s
                                                  x 100
to be used when x/s > 0.3 and up to 44 results are available,

                                     x +. 1.8 s
                                T -
        x 100
when x/s «• 0.3 - 0.15 and number of results are 45-170, and
                               .T -  2_8    x 100
when x is not significantly different from zero with 95% confidence.
h.  Numerical Conversions
                               1  g
                               1 mg
                               1 yg
                               1 ng
                               1 Pg
                               1 ml
                               1-yi
1000 mg      ,
1000 yg - 10"° g
1000 ng - 10~6 g
1000 pg - 10-9
10-12 g
1000 VI - 10-3 liter
10~6 liter
                                          -58-

-------
                                                              Sections 2L, 2M
2L   REFERENCES
   (1)  Radomski, J. L., and Fiserova-Bergerova, V., Indust. Med. and Surgery,
        34, 12 (1965).

   (2)  Mills, P. A., Onley, J. H., and Gaither, R. A., J. Assoc. Off. Anal.
        Chem.. 46, 186 (1963).

   (3)  Dale, W. E., Curley, A., and Cueto, C., Life Sci., 5, 47  (1966).

   (4)  EPA Pesticide Analytical Manual. Section 5,A, (3),(a).

   (5)  Kaiser, R. E., in Advances in Pesticide Science,  Geissbuehler, H.,
        ed., Pergamon Press, N.Y., (3), 643 (1979).

   (6)  Young, H. D., Statistical Treatment of Experimental Data. Chapter 10,
        McGraw Hill  (1962).

   (7)  McFarren, E. F., Lishka, R. J., and Parker, J. H,, Anal.  Chem...  42,
        358  (1970).

   (8)  Midgley, D., Anal. Chem.. 49,  510  (1977).


2M   ADDITIONAL  SOURCES OF INFORMATION ON  PESTICIDE QUALITY ASSURANCE PROGRAMS

   (1)  Burke, J. A., and Corneliussen, P. E., Quality Assurance  inr the  Food
        and  Drug Administration's Pesticide Residue Analytical  Laboratories,
        Environ. Qual. Saf..  Suppl.. £ (pesticides), 25-31 (1975)*

   (2)  Carl, M., Internal Laboratory  Quality Control in  the  Routine Determina-
        tion of  Chlorinated Pesticide  Residues, in Advances in  Pesticide Science,
        Geissbuehler, H., ed., Pergamon Press, N.Y.,  (3), 660-663 (1979).

   (3)  Cochrane, W. P., and  Whitney,  W.',  The Canadian Check  Sample Program on
        Pesticide Residue Analysis:  Reliability and Performance, Adv. Pestic.
        Sci.. Plenary Lect. Sym. Pap.  Int. ..Congr.  Pestic. Chem..  4th 1978
        (Publ.  1979).  3, 664-667, edited by Geissbuehler, H., Pergamon Press,
        N.Y.

   (4)  Egan, H., Methods of  Analysis: An Analysis of Methods, J.  Assoc. Off.
        Anal. Chem.,  60, 260-267  (1977).

   (5)  Eiduson, H.  P.,  Applications of Tolerances, Standards,  and  Methods in
        the  Enforcement  of  the Food, Drug, and  Cosmetic Act,  J. Chem.  Inf.- Comput.
        Sci.. 17 (2),  102-105 (1977).

   (6)  Eiduson, H.  P.,  Laboratory Quality Assurance, Bulletin  of the Association
        of Food and Drug Officials, pp.  151-156 (1976).
                                         -59-

-------
                                                            Section 2M


 (7)  Elgar, K. E., The Variability of Residue Results, with Particular
      Reference to the Codex Study on Organochlorines in Butterfat, Adv.
      Peatic. Sci., Plenary Lect. Symp. Pap. Int. Congr. Peatic. Chem..
      4th 1978 (Publ. 1979). 3, 668-672, edited by Geissbuehler, H.,
      Pergamon Press.
•
 (8)  Horwitz, W., Good Laboratory Practices in Analytical Chemistry,
      Anal. Chem. 50(6)  521A-524A (1978).

 (9)  Horwitz, W., The Inevitability of Variability in Pesticide Residue
      Analysis, Adv. Pestic. ScJ.. Plenary Lect. Symp. Pap. Int. Congr.
      Pestic. Chem.. 4th 1978 (Publ. 1979). 3, 649-655, edited by
      Geissbuehler, H., Pergamon Press.     "~

(10)  Wilson, A.  L., Approach for Achieving Comparable Analytical Results
      for the Analysis of Water from a Number of Laboratories, Analyst, 104,
      273-289 (1979).                                          	JL~  ~

(11)  Youden, W.  J., and Steiner, E. H., Statistical Manual of the Association
      of Official Analytical Chemists - Statistical Techniques for Collabora-
      tive Tests, published by the AOAC, 1111 N.  19th Street,  Suite 210,
      Arlington,  VA  22209, 1975, 88 pp.
                                    -60-

-------
                                  Section 3
                       I^LABQRATORY QUALITY CONTROL
3A   PURPOSE AND OBJECTIVES              ,

   The intralaboratory control program is a continuing, systematic, in-
  .house regimen intended to ensure the production of analytical data of
 "continuing high validity.  Several of the program objectives are
   parallel to those given in Section 2 for the interlaboratory program:

   a.  To provide a measure of the precision of analytical methods,

   b.  To maintain a continuing assessment of the accuracy and precision
   of analysts within the laboratory group.

   c.  To identify weak methodology and provide a continuing source of
   research problems aimed at overcoming deficiencies.

   d.  To detect training needs within the analytical group.

   e.  To provide a permanent record of instrument performance as a basis
   for validating data and projecting repair or replacement needs.

   f.  To upgrade the overall quality of laboratory performance.


   The following subsections will treat several integral parts of a high
   quality intralaboratory quality control program, embracing such areas
   as the periodic analysis and interpretation of results of spiked
   reference materials  (SPRM's), instrumental maintenance and calibration,
   and monitoring of the quality of various materials used in the analyti-
   cal scheme.

 3B   PURPOSE AND OBJECTIVES OF SPRM'S

   In contrast  to the interlaboratory check sample program in which one
   analyst  in a laboratory will analyze a  sample occasionally sent by  the
   coordinator, the  intralaboratory  SPRM program provides a continuing
   measurement  of the performance  capability of each analyst.  Each person
   can be  constantly aware of his  strengths and weaknesses, and  corrective
   steps can be undertaken when necessary, before serious problems occur
   and erroneous  data are  reported out  of  the  laboratory.
                                     -61-

-------
                                                         Section 3C

   The program Involves continual, systematic recovery studies on prepared
   test samples of each type of substrate routinely analyzed by a laboratory.
   Each staff chemist conducting routine analyses should participate, and
   all recovery results are recorded on a table available for examination
   by the chemist's supervisors.

3C   NATURE OF SPRM'S

   One possible approach is for a laboratory-to prepare its own SPBM's.   If
   the laboratory routinely analyzes animal fat samples, an appropriate check
   sample may be prepared as follows:  Obtain a local bulk sample of 2 Ib.
   or more of fatty tissue, place in a large beaker, and warm carefully on
   a hot water bath to a temperature not above 50°C with intermittent
   stirring.  After a sufficient quantity of liquid fat has been expressed,
   filter into a second beaker through glass wool (pre-extracted with
   hexane) held in a glass or porcelain funnel.  Heat the filtered fat to
   ca 45°C, transfer about one-half to a previously tared flask with standard
   taper stopper, and reweigh to the nearest 0.1 g.  This portion is stored
   in the stoppered flask in a freezer at -18° to -23°C for later spiking.
   The remaining half is divided into individual analysis units in small
   vials or bottles that are also stored in the freezer.  The weight of each
   unit is slightly larger than the intended sample weight.  These serve as
   unspiked SPBM's.

   Sufficient analyses are made on the unspiked subsamples so as to be satis-
   fied with the reproducibility of results from the same analyst and among
   all participating analysts.   For verification, the sample may be sent to
   an outside laboratory with experience in performing the analysis in
   question.  When reproducibility is sufficient to establish a reliable
   pesticide profile in the unspiked sample, the other half is spiked to
   produce residue levels approximating or slightly exceeding the levels
   obtained in routine media.   The spiked fat is thoroughly mixed,  trans-
   ferred to small bottles, and stored in a freezer.  These spiked samples
   serve to test the capability of the analyst for recovery of higher pesti^
   cide levels.

   For both the unspiked and spiked SPKM's, at least a dozen replications of
   the analysis on the same sample should be conducted by chemists  with
   recognized competence.   From these data, the percentage relative standard
   deviation is calculated and used in construction of control curves as
   described later in Subsection 3F.

   The same basic program outlined for fatty tissue can be followed for
   other sample materials.  If the compound(s) and media are known  to be
   fully stable at room or refrigerator temperature, freezer storage is
   not required.

   The EPA-ETD Interlaboratory Program provides participating laboratories
   with a sufficient supply of each interlaboratory check sample to serve
   also as an intralaboratory SPEH for a six month period (Subsection 2D).
                                       -62-

-------
                                                         Section 3D
   Laboratories should store the excess  material in sample-size portions  in
   a freezer to be withdrawn periodically for analysis  along with routine
   samples.   The correct formulation value will be known to  the laboratory
   supervisor when he receives the interlaboratory Summary of  Results  Table
   (Subsection 26) from the coordinator, so  that he can compare the results
   of his personnel with the "correct" values.  The advantage  of this
   second approach is that a participating laboratory will have internal
   SPRM's with reliable results available to them without having to prepare
   their own samples and establish residue levels and BSD values before
   they can be routinely used.

   Because of their nature, it has not been  the practice to  treat intra-
   laboratory SPRM's as blinds in the EPA program.  A homogeneous, frozen
   fat check sample in a vial, which is  simply dissolved in  hexane as  the
   first analytical step, would be difficult to camouflage as  a routine
   fat sample, normally encountered by the chemist as a chunk  of adipose
   tissue requiring initial grinding (EPA Pesticide Analytical Manual,
   Section 5,A, (1),(a),III,3).  Likewise, routine blood samples are re-
   ceived as a whole blood rather than as the serum form'of  the check
   sample.  It would be undoubtedly advantageous to devise SPRM's that could
   be offered to the chemist as a true blind along with his  normal sample
   load, but this has proven a difficult task with fat  and blood when  it
   is necessary to prepare a homogeneous sample guaranteed to  give a con-
   sistent analysis regardless of the portion taken. It might well be
   feasible for some other sample substrate, such as urine or  water.


3D   FREQUENCY OF SPBM ANALYSIS

   The frequency of SPRM analysis is related to the volume of  routine  samples
   run.  Laboratories making less than one routine analysis  per week of a
   given substrate should analyze a corresponding SPRM sample  with each
   routine sample, and not less than one SPRM analysis  per month even  if
   no routine samples are encountered.  Laboratories analyzing one or  more
   samples per week should analyze at least  10 percent  as many SPRM samples
   as routine samples, with a minimum of one per week.   For  example, if
   one to fourteen samples are run per week, at least one standard sample
   should be analyzed each week.  If thirty samples are run, one corres-
   ponding SPRM sample should be analyzed for each nine samples, or a
   total of three standard samples.  The SPRM is carried through the
   analysis in parallel with a group of routine samples, giving it no
   special care or treatment.

   In laboratories where more than one chemist performs an entire routine
   analysis of a given substrate, each individual should run separate
   SPRM samples.  However, if protocol is that routine analyses are handled
   by a team, e.g., with one chemist preparing extracts and another doing
   the determination, SPRM samples should be handled in this same normal
   fashion.
                                      -63-

-------
                                                         Section 3E
3E   RECORD KEEPING
   Immediately upon completion of each analysis of an SPRM sample, results
   are recorded on an Internal Check Sample Form.  An example for blood
   serum is shown as Table 3-1.  Data are entered in legible handwriting.
   Each participating chemist should have access to this record.   If
   significant deviations from the furnished correct (mean) values occur,.,
   an investigation is begun at once to determine the reason or reasons.

   The chief chemist of each laboratory completes a quarterly-report for
   forwarding to the coordinator and includes in the confidential in-house
   section (Table 3-2) one copy of each Internal SPRM Report.   The coordi-
   nator compiles the data from all labotatories and furnishes .to each
   statistical summaries for- comparison of results.

   The following two publications by the National Enforcement Investiga-
   tions Center (Denver Federal Center, Bldg. 53, Box 25227, Denver, CO
   80225) of the EPA contain information on record keeping and reporting of
   analytical results:

   (a)   NEIC Policies and Procedures Manual, May, 1978.

   (b)   Pesticide Product Laboratory Procedures Manual,  August, 1979.

   Reference (a)  outlines legally oriented standard operating  procedures for
   EPA chemistry and biology laboratories.  Pages 11-19  to 11-29  focus on
   document control, with information on serialized documents,  project log-
   books, field data records,  sample identification documents,  chain-of-
   custody records,  analyst and instrument logbooks,  photographs,  document
   corrections,  document consistency,  document numbering and inventory,  and
   files.

   The  information in Reference (b)  is specifically for  laboratories per-
   forming analytical testing on pesticide formulations  and products to
   determine if labeling is correct.   The results of these analyses  can lead
   to a number of legal actions,  including criminal action.  Although  it
   may  be more important that records  be kept carefully  and completely in
   this situation compared to a monitoring laboratory, the same principles
   can  be applied to all analytical  laboratories.   Analysts are specifically
   referred to Section IV of Reference (b)  above for recommended procedures
   on two aspects of sample custody:   documentation and  physical sample
   security.   These  procedures are designed to ensure that collected samples
   are  not tampered  with in the event  of any subsequent  legal action.

   Section VI of  Reference (b)  describes proper record keeping.  Although
   this material  is  not uniquely applicable to SPRM samples, the points  of
   importance to  all pesticide analysts  will be summarized here.  The reader
   is also referred  to Section 50d of  this  Manual  for further information
   on reporting of results and record  keeping.
                                     -64-

-------
                                                               Section  3E
TABLE 3-1
                    RECORD OF ANALYSIS OP STANDARD REFERENCE MATERIAL
Analyst or Team
Sample
No.




















Date




















Analyst or Tear













	





















.






Aldrin




















i
3-EHC





















Hept.
Epox.



















































































Diel-
drin









































0,p'-DDT









































p,p'-DDD


•






































Media
p,p'-DDE


















•























p,p'-DDT




















































































        Reporting units should be in ppb or ppm.  Observe standing instructions for
        minimum reporting levels.
                                             -65-

-------
                                                                      Section 3E
•PABLE 3-2
                                                           (Date)
SUBJECTt  Quarterly Report, Quarter Ending
     TOj   Chief,  Quality Assurance Section, Analytical Ghent.  Branch, Health Effects
         Research Laboratory, (MD-69), Research Triangle Park, NC   27711
   PROMt   Chief Chemist
                                                            _(Laboratory)
                                                      _, (specify)
     During the past quarter we have analyzed the following numbers of
routine* samples for pesticide residues:
                Blood (multiresidue)      __________
                Blood (PCP)	
                Blood (Other)              _____
                Adipose Tissues           ______
                Other*Human Tissues       _____
                 Air                      	
                 Soils                    	
                 Stream Sediment           _____
                 Water (multiresidue)      ______
                 Water (Other)    '         	
                 Urine  (alkyl phosphate)   _,	
                 Urine  (Other)             	
                Hbusedust                 	
                Pish or Shellfish        	
                Wildlife                 	
                Other**
                                                        (specify)
                                                        (specify)
   Are any spiked  SPRM's prepared in-house? Yes  No    If Yes, 'list the
   substrates on the reverse  side of this  sheet  giving theTpiking level
   range of each compound spike.
                                                    Chief Chemist
   *The terra "routine" is intended to mean samples of local origin such as
    donors, autopsies, etc.
  **Spacify substrates if 10 or more samples were analyzed during quarter.
                                      -66-

-------
                                                      Section 3E
Detailed and specific notes should be made regarding all sample analyses,
manipulations, and'observations.  Sufficient detail should be provided
to enable the analyst or others to reconstruct the analysis step-by-step
at a later date.  All analytical work (graphs, charts, notes, etc.) should
be retained in a general laboratory locked file cabinet and identified
by sample number.  This is in addition to the individual analyst's note-
book or logbook and will assure that all primary information regarding
a sample is in one location so that there is less chance of loss.  Labora-
tory notebooks should be the "two-page" carbon or pressure-sensitive
paper type.  The originals are then removed from the notebook and re-
tained with the laboratory records.

Careful notes should be made concerning the sample as received (see
Subsection 8D), the preparation of the sample for analysis, and the actual
determination.  If & specified procedure is being followed, this should
be referenced, and any variations from the procedure must be recorded.
If the method is not specified, details of every step are recorded.
Each laboratory operation should be accurately documented as to date
performed, particularly when an analysis or several analyses of a sample
or samples extend beyond one day.  Time of starting and stopping should
be recorded for all operations when duration is a factor, e.g., extrac-
tions, separations, centrifugations, color formations, etc.

Photographs can be made when they might be useful, e*g., to record the
results of a  thin  layer chromatographic separation.  All photocopies
should be mounted  on heavy paper and identified as to sample number,
date, analyst, and subject matter.

Custody information and storage location should be documented  if samples
are  stored overnight.

Reference standard information, including source, purity, and  age  should
be recorded along  with appropriate weighing and dilution data.   If a
reference standard is  used  that was prepared  at an earlier date, then
the  original  weighing  and dilution data should be referenced.

All  instrumental conditions should be  recorded either on the worksheet
or on an appropriate  chart,  graph, or printout.  All graphs, charts,
and  printouts should be  identified by sample  number,  date, analyst,
 and  determination  number.

 Gas  chromatography data  should be recorded for each  analysis at  least
 to the following extent:
      1.  Gas chromatograph
      2,  Column
-  Make, model, and detector.
   Include designation if more
   than one of same model is
   available.

-  Source and/or date prepared
-  Length, id, od, and compo-
   sition
-  Packing (%, type, and source)
                                  -67-

-------
                                                          Section 3E
       3.  Conditions
       4.   Injection


       5.   Response
      6.  Internal standard  (if used)


      7.  Any conditioning or calibration

      8.  Recorder

      9.  Sensitivity
     Temperature of oven,
     injection port, detector,
     transfer lines, etc.
     Flow rates, composition
     and purity of carrier,
     detector, and purge gases
     Electrometer conditions such
     as range, attenuation,
     voltage, amperage, etc.

     Amount injected and size of
     syringe

     Digital integration (incl.
     make,  model,  slope sensi-
     tivity, and  other  pertinent
     parameters),  planimeter,
     peak height,  cut and weigh,
     etc.

     Identification, source, and
     concentration
 -  Make, modelr range, and speed

 —  Z response to pg, ng of a
    standard material
HPLC data to be retained for each analysis should include at least the
following:
      It  Liquid chromatograph


      2,  Detector


      3.  Column
      4.  Mobile phase
     5.   Injector
 -  Mhke, model, type, and lab
    designation

 —  Make, model, type, and wave-
    length

 -  Source and/or date prepared
 -  Length, id, od, and compo-
    sition
 -  Packing (type,  source, and
    particle size)
• -  Pre-column, if  applicable

 -  Isocratic or gradient?
 -  Name and Z of each solvent
 -  Degassed?  Filtered?

 -  Type, make, and model
 -  Amount injected
                                  -68-

-------
                                                         Section 3F
       6.  Temperature


       7.  Sample handling


       8.  Response measurement



       9.  Recorder

       10.  Internal  standard
-  Type of control and tempera-
   ture

-  Filtration?  Pore-size of
   filter

-  Digital integration (incl.
   make, model, and settings),
   planimeter, peak height, etc.

-  Make, model, range, and speed

-  Identification, source, and
   concentration
   Spectrophotometric data should be retained to the extent called for
   on the specific charts, along with any additional information as may
   be relevant to the measurement.


3F   QUALITY CONTROL CHARTS

   In addition to recording numerical results of each analysis of an internal
   SPRM, it may be desirable for each analyst or team to construct a Quality
   Control Chart.  This depends to a great extent on the number of SPRM
   analyses of a given substrate per week or month.  The purpose of this
   chart is to provide graphic assessment of accuracy and precision for
   the analysis of each substrate and instant detection of erroneous data.
   The charts allow quick and easy observation of recovery trends for a
   particular analysis and have long term value for the self evaluation
   of analytical output by staff personnel.  Another significant value of
   the charts is that of providing a laboratory administrator with a rapid
   assessment of the continuing analytical capability of the staff chemists
   as related to the output of valid analytical data.

   The first and very important step in  the  development of a control chart
   is the  determination of an appropriate value of  the relative standard
   deviation  (RSD)  (Subsection  2Ke) for  the  particular analysis.  The RSD
   value used in preparation of control  curves  should be determined as
   suggested below,  and should  be a fixed value that represents the best
   precision possible for this  particular method  and substrate.   This value,
   when once  established,  should then  remain fixed  for an  indefinite period
   of time or until a method revision  or improvement is made that would
   permit  the determination of  a lower RSD.   A  separate RSD value could
   be calculated and used for each  pesticide residue in each method-substrate
    combination.   However, this  is unnecessarily complicated for  a multi-
    residue method.   A preferred practice is  to  determine an RSD value  for
    several pesticides analyzed  by a given method, and the  average RSD  value
    which will remain fixed as previously mentioned. An example would  be  the
    10% RSD figure that is commonly accepted for all organochlorine residues
    determined by the EPA PAM procedures for blood serum or adipose tissue.
                                    -69-

-------
                                                        Section 3F

 An In-house RSD value should be determined as suggested in Section 3C
 of this Manual:  "at least a dozen replications of the analysis on the
 same sample should be conducted by chemists with recognized competence.
 From these data, the percentage relative standard deviation is calculated
 and used in construction of control curves."  As a satisfactory alterna-
 tive, two or more competent chemists of the staff analyze six replica-
 tions of the same sample, from which a reasonable value for percentage
 RSD is obtained.

 In summary,  the RSD value is a measure of the best possible precision
 obtainable with a method.  The accuracy of the method is, of course,
 not reflected in this figure.   The quality control charts,  however,
 provide a rapid assessment of both accuracy and precision.

 The preparation of QC charts is illustrated by the following Figures
 3-A and 3-B  in which results for serum intralaboratory SPRM analyzed
 over a period of three months  for £,2>DDE and £,£f-DDT by chemists in
 two different laboratories are shown.   (Several additional  pesticides
 were also found, but only two  are illustrated).  Consecutive results
 are plotted  on every second space along the X-axis.   The Y-axis contains
 zero (0),  plus (+),  and minus.(-)  lines.   The (+)  line represents two
 standard error units (comparable to standard deviations)  on the high
 side from the "correct" answer (the spiking level, or the level found
 by an experienced analyst in the coordinating laboratory), while the
 (-)  line represents  two standard error units (SEU) on the low side.
 In the case  of this  sample,  it had been previously determined that an
 appropriate  RSD value was 10Z  of the spiking level for each pesticide.

 The  known  formulation or  spiking value is  subtracted  from the experi-
mental value (obtained for an  analysis of  the in-house standard sample
 to provide a (+)  or  (-) arithmetic deviation (difference).  This
 difference is  then divided by  the  calculated standard error unit to
 give the number of standard  error  units from the correct value.  This
 is the number  plotted on  the appropriate horizontal line.

Assume,  for  example, the  first  serum SPRM analysis is run during a
quarter, and a value of 105 ppb  is obtained for the content of DDT.
The spiking  level, however, was  150 ppb.  One standard error unit (SEU)
is calculated by multiplication of the formulation value by the percent
RSD to give a standard error unit  that should be valid throughout the
life of  the specific SPRM:  150 x 0.10 - 15 ppb » one SEU.  The difference
105 - 150 - 45 is then divided by 15 to give the number of standard
error units to be plotted, in this case -3.0.  If the second result is
125 ppb, the second point plotted along the horizontal axis would be
calculated as:
                 125 - 150
                  one SEU
or
-25
 15
-1.7 SEU
                                  -70-

-------
                                                     Section  3F
 Figure  3-A.   Laboratory A  control curves for blood SPBM,
               three-month period.
R
                                     TT
  2+
  2-
                                                      -r
                                           P.P-DDE
  2+
                                  _!,_,j
                                  -r. J .;.  i.   - ••- •!-'  •!
     lijuHnH-H—r—i	f—I
                                          p,p'-DDT
Figure 3-B.  Laboratory  B control  curves for blood  SPBM,
              three-month period.
                               r . - i. r.!-''—i--1 i ~.r=»" ~ * T-™ =
       r^rF'frr''TTTTr7""i~
-------
                                                           Section  3F


 When constructed in this way,  quality control  charts readily show levels
 of accuracy and precision for  repetitive analyses  by a given analyst.
 Figure 3-A demonstrates excellent precision since  the results all fall
 along an essentially horizontal line.  Accuracy  is good because this
 line is well within the control area of +2 SEU,  all recovery values
 being slightly low, probably due to an inherent  negative determinative
 error in the procedure being used.   Figure 3-B,  on the other hand,
 demonstrates very poor analytical performance  in both accuracy and  .
 precision.  Nine of the repetitive values for  DDE  and eleven for  DDT
 are out of the acceptable control range of +2  SEU.

 Control charts also highlight  cases where errors are present exerting
 similar effects on the analyses of several pesticides.  The following
 Figure 3-C, for.example, demonstrates rather poor  precision and also
 a distinct correlation in the  configuration or shape of the curves for
 both compounds.  This signals  some common error  proportionately affecting
 both compounds, most likely the extraction step  in this blood analysis.
       Figure 3-C.   Laboratory  C control curves for blood SPBM,
                     three-month period.
             ifflnBSn|iSHMa=aaa:a«nu^ ^

                    iilil LahH
          frpr |"il;..,,-u.i. --t '•'''•• -'-T- •:i.Il:.-.i..-.: ..} .'"'-...:....7.. I....... .;".".. .'"~.1>...K^ '"i ^'!^f^frli?L'!^'I"^H^H^|d''';i;_^"!'
          BBij;j^jpn!fe?-:;isi-p^
                                                          	1
                        'tt'

          fe%ffi!:t ™ ; E-»:.' i^-teifcLf:; VSi^S-'i \  S; *£j&.*:z3?«'"-?K3s&&^"g-'fe^\'&-'>&'. r. J '&
:||?i:,%s::it;l:«; 1 :rs3t ' '• ..|::::it :jl'' i;;;:::-:J:;: ; ';:-: i.":J:::H:;.l:t:::::'::''i:::i::-:j;':.!::l:i;;-:::::-:::"i:::-f::i;!w:':.':!!. :• '. '?:I

                                           '
From time to time, the  following question is asked;  "What  is  to prevent
an analyst from 'fudging*  the control chart points so that  his curves
will appear significantly  better than they should"?  This can  and has
occurred in very rare instances.  The alert laboratory administrator,
however,  should have little difficulty detecting the doctoring of curves.
When a  chart is submitted  that is virtually a straight line such as that
                                    -72-

-------
                                                          Section 36


   for t>,£f-DDE  in Figure 3-A, his  suspicion should be  aroused  to the  extent
   of personally checking the raw data  to  either  confirm or  refute his
   apprehensions.  Furthermore,  such an apparently outstanding  performance
   will  catch the attention of other analysts  in  the  peer  group whose  data
   may look relatively poor by comparison.

   In the first  sentence of this subsection it was stated  that  "it may
   be desirable" to  prepare control charts.  One  main value  of  the charts  • '
   is to detect  trends.  Therefore, if  a given SPBM sample is analyzed on
   an infrequent basis, a chart  would serve little purpose as trends would
   not be evidenced.  On the other  hand, if a  laboratory is  monitoring a
   waterway, for example, for certain pesticides  or other  organic pollutants,
   the number of routine samples per month may be 100 samples or more.  If
   the  controlling  SPBM is analyzed a-t  the recommended  minimum  rate  of
   one  SPBM per  10  samples, this would  amount  to  at least  10 SPBM analyses
   per month, a  number sufficient to justify preparation of  the chart.


3G   BENEFITS OF THE IN-HOUSE SPBM PROGBAM

   Analyzing in-house SPBM's will require a certain amount of man hours
   during which laboratory personnel cannot accomplish routine, productive
   analytical work.  The time, effort,  and expense spent on such a program
   has proven an invaluable investment, however,  in the quality of analytical
   output in those laboratories involved.   For example, chemists from regu-
   latory laboratories are sometimes called upon to testify in a court case
   based upon their analytical results.  If a chemist is armed with high
   quality analytical assurance data, the validity of his results on the
   sample(s) in question will be much more difficult to disprove, and the
   case will be that much stronger.
                              !\
   If a laboratory has a correctly functioning intralaboratory control'pro-
   gram in  effect,  the morale of personnel is high, everyone has confidence
   in the routine data output, and interlaboratory check samples  can be
   taken  in stride  and handled with little disruption of the normal work
   schedule.  Since a higher volume of  uncontrolled analytical data is
   obviously of much  less value  than a  lower  output of  reliable  results,
   time and effort  must  be  allowed for  each pesticide analytical laboratory
   that  cares about valid  results  to conduct  a proper quality  control
   program.

   Certain minimum  requirements  are necessary for the physical plant  in which
   analyses are to  be performed.   Minimum considerations  should  include  such
   factors as safety  of personnel, reasonable temperature and  humidity
   control, an  adequate ventilation system, refrigerated  storage areas for
   samples, facilities for an assembly line layout if  large numbers of
   samples are  processed,  and an efficient glassware wash area.   In addition,
   all  necessary equipment for  safety,  sample preparation,  analysis,  and
   sample and data processing must be  available.
                                       -73-

-------
                                                          Section 3H


   The following subsections are intended to highlight a number of in-
   house factors that can lead to inaccurate analytical data in any labora-
   tory and to present guidelines for avoiding these pitfalls.  Further
   details of many of the areas mentioned will be given in appropriate
   later sections of this Manual or are covered in the cited sections of
   the EPA PAM.
3H   ANALYTICAL BALANCES

   Most laboratories contain balances of two types.   Rough triple beam or
   Dial-0-Gram balances are used for weighing approximate amounts of
  .materials to the nearest 0.01 or 0.001 g.  For example, to prepare one
   liter of a 2 percent solution of NaCl for use in  the liquid-liquid
   partitioning step of the modified Mills, Onley, Gaither Procedure
   [EPA PAM, Section 5,A,(1),(a)J the required 20 g of salt could be
   weighed out on one of these rough balances since  the concentration
   of the solution is specified to only one significant figure.

   An analytical balance is required, however, for the critical  weighing
   of primary analytical pesticide standards in preparing standard reference
   solutions.   The usual analytical balance has a capacity of 160 g and a.
   capability of weighing to the nearest HjO.OOOl g (error and uncertainty),
   the fourth decimal place being obtained by estimation and,  therefore,
   the final significant figure recordable (Subsection 2Kb).   This leads
   to a total accuracy and precision of
                               0.0002  g
                               0.0200  g
x 100 - 1%
   in weighing 20.0 mg of pesticide standard by difference  (two weighings),
   as in usually done in  preparing  primary standard solutions  (Subsection
   30).  This  value is quite  acceptable  considering the other  errors in-
   herent  in the total analytical scheme.

   The accuracy,  precision, and  sensitivity of the analytical balance should
   be checked  at least once a year  by a  qualified balance specialist, and
   the balance should be  used properly by  all personnel to insure its
   maintenance in good condition at all  times.  Since the analytical balance
   is used to  weigh standard  pesticides  for preparation of solutions upon
   which all analytical results  are based,  its Importance, and the need
   for Its care and protection,  should be  obvious.  The single pan, direct
   reading analytical balance that  weighs  by the principle of substitution
   is by far the type in  widest  use today.  As compared with the classic,
   double  pan,  equal  arm  balance, the single pan balance is more automatic,
   convenient,  and much faster (although no more accurate or precise), but
   it is still a very fragile instrument requiring certain precautions in
   its use.  These include the following:
                                    -74-

-------
                                                         Section 31


  a.  The balance should be placed on a heavy, shock proof table or
      cement block slab built up to convenient height from the floor.

  b.  The balance is preferably located away from laboratory traffic and
      protected from drafts and humidity changes.

  c.  The balance temperature, room temperature, and temperature of the
      object being weighed should be equilibrated.

  d.  When not in use the balance beam should be locked, objects removed
      from inside, and all weights released from the beam.

  e.  The inside and outside of the balance must be kept scrupulously clean.
      Never place chemicals directly on the balance pan.  Remove spilled
      chemicals immediately with a brush.

  Before using an analytical balance for the first time, the manufacturer's
  literature should be consulted or instructions obtaine.d from someone
  experienced in its proper use.                           .


31   PURITY OF SOLVENTS

   The purity of reagents, solvents, adsorbents, distilled water, etc.  is
   of extreme importance when analyzing samples for residues; In the. low ppm
   or ppb range.   The electron capture detector senses any electron capturing
   materials in the injected sample, whether they be pesticides or other
   impurities.   Quite often, extraneous contaminants will give rise to GC
   peaks that may precisely match the retention characteristics of certain
   pesticides,  even,on two or three different stationary phases.   A common
   contaminant of solvents and reagents is di-jl-butyl phthalate plasticizer,
   which can be easily confused with BHC and aldrin in GC with electron
   capture detection.  Construction materials have been suggested as the
   source of phthalic acid esters and PCBs present in laboratory air and the
   cause of solvent, reagent, and glassware contamination (1).  Sulfur and
   sulfur-containing compounds can be present in solvents and column materials,
   as well as in certain substrates (onion, cabbage, turnips), and can give
   rise to peaks easily confused with pesticides (2).  The use of plastic
   gloves, Tygbn tubing, plastic tubing, Nalgene containers, and plastic
  • screw caps without Teflon liners should be strictly avoided whenever
   contact with organic solvents is possible.

   Commercial solvents designated "pesticide grade" or "distilled in glass"
   can usually be used without further treatment, but care must be exercised
   in their storage.  For example, it was reported that photo-chemical
   reactions can produce compounds from pesticide-grade hexane that are
   detected by an electron capture detector and interfere with pesticide
   residue determinations (3).  Storage in the dark was recommended to
                                      -75-

-------
                                                        Section 31

 prevent this.  Reagent- or technical-grade solvents almost always require
 distillation by the user in an all-glass still.  In any case, each
 solvent should be checked before use for interference in the analytical
 procedure by evaporating a portion to provide as great a concentration
 factor as will ever be employed in any method for which the solvent
 will be used.

 A typical procedure is to concentrate 100 ml of solvent to 1 ml and to
 inject 5 vl into the gas chromato'graph equipped with the detector of
 choice.   Detector response is recorded for at least 20 to 30 minutes.
 No cloudiness or discoloration should be observed when the volume is
 reduced, and no GC peaks that would interfere with sample analysis should
 be produced.   Details of the test for electron capture GC are given in
 Section 3,C of the EPA PAM.

 Tests for interfering substances not detected by this procedure but
 causing pesticide degradation and loss are made by carrying known amounts
 of standards through the analytical method in the absence of any sample
 substrate (a complete reagent blank check).   Solvents containing oxidants
 are especially troublesome in causing losses of organophosphorus pesti-
 cides, most notably carbophenothion.   Acetonitrile and'ethyl ether are
 two common solvents that may require special attention.   Impure acetonitrile,
 the vapors of which will turn moistened red litmus paper blue when held
 over the mouth of the bottle, should be redistilled.   Recoveries of some
 phosphate pesticides from Florisil columns are low if peroxides are
 present  in ethyl ether eluants.   Ethyl ether is tested for the  presence
 of peroxides  by adding 1 ml  of fresh 10% KI solution  to 10 ml of solvent
 in a clean glass-stoppered flask or cylinder previously rinsed  with ether.
 Shake and let stand for 10 minutes.   No yellow color  should be  observed
 in either layer.   If present,  peroxides are removed by extraction with
 water, after  which the 2% ethanol normally present in ether and also
 removed  by the partitioning  is replaced [EPA PAM,  Section 5,A,(l),(a)].

 Reagent  grade solvents purchased in large cans  with plastic pour-spouts
 can be a significant source  of contamination.   If  these  solvents  are
 used in  the laboratory glassware cleaning routine,  the  glassware  should
 be rinsed with pesticide  grade solvent  immediately after  rinsing with  the
 reagent  grade solvent  from the can.   The  solvent from these  cans  should
 never be  used for  extraction  of  samples or dilution of samples or  standards.
Metal safety  cans  commonly used  for solvent storage can also contribute
 contamination.  Plastic snap  caps  that  seal cans of diethyl ether  can be
 a  large source  of  sample background.  These impurities begin to elute
 early in  the  chromatogram and  continue until well after the p,p'-DDT
peak (relative  retention of 4.6  on  a  1.5%  OV-17/1.95% OV-210 column).

 Solvent purity  for HPLC  (4, 5) is at  least as critical as for GC, and it
 is frequently necessary to repurify even the highest quality commercial
solvents.  Impure solvents can lead to baseline instability, spurious
peaks, variable retention volumes, impure recovered fractions, and other
problems.  Solvent purity  is more important in gradient than in isocratic
                                  -76-

-------
                                                       Section 31
elution.  This is especially true of the weaker solvent since more of it
passes through the column, and its impurities can be concentrated on the
column head.  Intentional impurities such as the ethanol stabilizer in
chloroform and the antioxidant (UV absorbing) in tetrahydrofuran (THF),
as well as HC1 or oxidation products in chlorinated hydrocarbons, benzene
in hexane, and the aforementioned peroxides in ethers, may have to be
removed if they interfere.  Water content of solvents has an important
effect on separations and must be controlled.  Tests for solvent purity
include recording the UV spectrum in a 5 or 10 cm cell versus air over
the dynamic range of the UV detector, spotting residue from evaporation
of a large volume for TLC with 12 vapor visualization, and Karl Fisher
titration for water.  Antioxidants are easily removed from THF by
distillation, but the THF then rapidly oxidizes and must be tested for
peroxides with KI as described earlier.  HC1 is removed from chlorinated
solvents and alcohol from chloroform by extraction with water.  Water
and some other polar impurities are removed from low to moderately polar
solvents by column chromatography on activated silica (heated to 175°C),
alumina (heated to 300°C), or a molecular sieve.  About 2-6 bed volumes
should be passed through the adsorbent before replacing it; low cost,
larger particle adsorbents that can be dry packed may be used.  Water
content of  solvents is best controlled by preparing dry solvent and
blending with water-saturated solvents.  Impurities in water are removed
by filtration, reverse osmosis, deionization, distillation (neat or from
alkaline permanganate), electrolysis, passage through a reverse-phase
column  (for reverse-phase separations),  or  combinations of these*

Reagent grade water, especially purified for HPLC use, is commercially
available from several sources.  Particulates are removed from solvents
 (especially those  cleaned up on an adsorbent column) prior to use in
HPLC by passage  through a solvent-resistant 0.5 V membrane filter, and
dissolved gases  are removed by heating,  stirring under a vacuum, or ultra-
sonic agitation.   The composition of solvent mixtures can be altered by
prolonged heating  or exposure to vacuum.  Table 3-3 summarizes some
aspects of  solvent purity in HPLC as outlined by one instrument manu-
facturer.
                                   -77-

-------
       3-3
                                                                                 Section  31
                                  JSOWZNT TOKCTI IN LC
 Particulat*
 •»tt*r
                    Possible Source

                    During transfer.
                    unclean vessels.
                    Glasaware, solvent
                    ^preparation or
                    manufacture
 Sffeet

 May block in-line
 lodge in pump  seals,  or.ac-
 cumulate at  column head.

 Tariable column activity.
 k'  variation,  stability! of •
 silieata ester bonded
                                                    R«nov«l

                                                    Filtration through m«abtan«
                                                    filter.
                                                    Drying over molecular sieve
                                                    anhydrous sodium sulfate.
 tlcohal
 Hydrocarbons
 {in water)
Peroxide*
(in ethors)
 KC1,  KBr
 (halogenated
  solvents)
 Stabiliser in        Similar to water.
 chloroform, impurity
 in 'hydrocarbons
                                                                       Tron hydrocarbon*, pas* through
                                                                       activated silicai froa CHC1,
                                                                       extract with HgO, dry with
 Organic natter
 Degradation
                                         B«»«lin« inatabllity
                                         during graditnt tlution.
                                         Oxidation of bonded phaa*
                                         («.fc..  -NH2 to -«02X,
                                         ftaotion with sample,
                                         column  deactivation or
                                         degradation (polyetyrcn*-
                              Passage through•porous
                              polymer column or C^a bonded
                              phase.

                              Distillation or passage;
                              through activated silica.
                              gel or
 Degradation
                                         Column degradation esp,        Passage  through activated.
                                         bonded phases,  UV abaorbance  silica or calcium carbonate
                                         (bromide),  stainless steel    chips.
                                         attack.
BHI *              Antioxidant in THF

Dissolved oxygen   Solvent preparation
                                         UV absorbing.
                                                    Distillation.
Unknown
t!Y-absorbin<
Kith boiling
compounds

Algae in water-
Froa aanufaeture
Degrades polystyrene-based    Degas solvent with vacuum
packing, oxidizes 8,0*-oxy-   or heat.
dlpropionitrile, may react
with sample.

Baseline instability.or drift  Try activated silica ot
during gradient elution, high  aluainai  distillation for
detector background.           organic* and recrystalli-
                              cation or passage  over ion—
                              exchange, eoluan for inorganics.
Frn« solvent
Growth daring
prolonged storage
                                        Contaainates collected
                                        sample in preparative 1C.

                                        Can plug in-line' filters,.
                                        ooiuan entranoe frits.
                                                   Distillation.
                                                                      Distillation tre» alkaline
                                                                      permanganat* or discard.
*  Butyl  hydroxytoluene
                                                    -78-

-------
                                                         Sections  3J,  3K
3J   DISTILLATION OF SOLVENTS (6, 7)
   Distill reagent grade acetonitrile over reagent grade AgN03 (3 g/1) with
   an all-glass fractionating column equipped with a water cooled condenser.
   Discard about the first 10% of the distillate and leave the last 20%
   of the solvent in the flask..  Rinse the flask and use fresh AgN03 and
   boiling chips for each distillation.  Test the distillate for interference.
   Alternatively, to 4 liters of acetonitrile add 1 ml of 85% I^PO^ 30 g
   ?2Q5» an<^ boiling chips.  Allow to stand overnight and then distill from
   all-glass apparatus at 81-82°C (do not exceed 82°C), discarding the first
   and last 10% of distillate.  Distill acetone, hexane, benzene, carbon
   tetrachloride, chloroform, ethyl ether, isopropanol, methanol, methylene
   chloride, isooctane, petroleum ether, and ethyl acetate from all-glass
   apparatus.  A technique for recovery of reusable solvent from Kuderna-
   Danish evaporators has been described (8).

3K   CONTAMINATION FROM REAGENTS AND MATERIALS

   Any other reagents used in the extraction or cleanup steps are also
   potential sources of contamination.  These reagents, such as sodium
   sulfate  (Na2S04>, glass beads, sodium chloride, and glass wool, should
   be pre-extracted with the solvent to be used in the analytical method
   or another solvent known to remove the potential interferences.  For
   example, Na2S04 is extracted in a reserved Soxhlet apparatus, the
   thimble of which is pre-extracted before the first use.  Methanol
   followed by hexane or petroleum ether are cycled for several hours each,
   after which the Na2S04 is dried and stored in a glass container with a
   glass cap at 130°C in the oven used to dry Florisil and other adsorbents.
   Plastic fiber pack liners have been found to contribute PCBs and phthalates
   to Na2S04 that must be removed by this procedure.  Phthalate esters are
   also removed from sodium sulfate by heating at 600°C for 4 hours in a
   muffle furnace (FDA PAM, Section 121).  Impurities in batches of silicic
   acil that interfere with separations of pesticides from PCBs were re-
   duced by extraction of the adsorbent with solvent (9),  Adsorbents
   that are activated and stored in an oven that is not cleaned at least  -
   yearly will absorb vapors from the oven.  These impurities may be eluted
   along with pesticides in cleanup procedures and could interfere in .the
   later determinative step.

   Filter paper and other reagents and apparatus should be checked by
   washing  through the solvent to be used and injecting a sample, after
   concentration, into the gas chromatograph.  No peaks should appear.
   Impurities from filter paper were the cause of interfering signals in
   the GC-alkali flame ionization detector determination of pesticide
   residues in plants; Soxhlet pre-extraction of the paper with acetone
   was recommended  (10).  Teflon and aluminum foil should be rinsed with
   an appropriate solvent.   Solvents in polyethylene wash bottles can
   become contaminated with  electron capturing and TJV absorbing species
   and  should be tested  for  impurities.  Better still, avoid the use of
   plastic  wash  bottles  and  use  Teflon or all-glass  ones.   Glass wool was
                                       -79-

-------
                                                        Section  3K


 shown  to  contain hydrocarbons,  phthalate esters,  and unesterlfied acids,
 among  other compounds  detected  by  GC.   The most efficient way of  elimina-
 ting these impurities  was  to  treat the  wool  for a few minutes with
 hydrogen  chloride vapors,  followed by continuous  Soxhlet extraction
 for  24 hours with methylene chloride  (11).   Losses of 2,4-D caused
 by a glass wool  plug have  been  reported (12).

 Distilled water  can be troublesome, particularly  in  a procedure where
 a large volume is used.  Such a procedure is the  Mills, Onley,  Gaither
 cleanup method for adipose tissue  [EPA  PAM,  Section  5,A,(l),(a)]  where
 700  ml of water  is partitioned  with acetonitrile,  the latter being finally
 concentrated to  5 ml (a potential  contaminant concentration factor of
 700/5  s 140).  Since the source of contamination  in  laboratory  water
 is organic in nature,  distillation will not  be sufficient cleanup if
 the  organic material co-distills with the water.   An activated  charcoal
 filtration prior to the distillation procedure has been found to  signifi-
 cantly improve water quality.   If  deionization through a column of ion
 exchange  resin is carried  out,  an  activated  charcoal filter should be
 installed between the  column  and the distillation equipment to  trap
 any  organic impurities eluted from the  resin before  the' water enters
 the.  still.   For  analyses at ppb and ppt levels, distilled and deionized
 water  should be  further purified by a double extraction with an approp-
 riate  immiscible solvent,  e.g., benzene or isooctane, followed  by boiling,
 if necessary, to remove the residual solvent.  Aqueous salt solutions
 such as 2% NaCl  or saturated  NaCl  used  in some isolation procedures are
 prepared  from properly purified salt and water and then solvent extracted
 as a further precaution.   Contamination can  result from Teflon  or Tygon
 lines  and/or plastic resin or charcoal  cartridges.   Water samples from
 systems containing these elements  must  be analyzed at the level of sensi-
 tivity necessary for the analysis  prior to use of  the water.

Materials  in which the initial  sample is  stored must be given considera-
 tion.  Polyethylene bags are  totally unsuitable for  samples to be
 examined by electron capture  GC or TLC  because of  trace contaminants
 that may be  present.   As an example, it has been reported (2) that poly-
 ethylene contains a contaminant that reacts with AgN03 chromagenie
 reagent, giving  a TLC  spot close to that  of £,2_'-DDE and having similar
 GC retention times to £,_p_'-DDE  and £,£f-DDE.  Figure 3-D shows a  gas
 chromatogram of  a hexane extract of the cardboard  liner from a  common
 type screw-cap bottle.   The peaks were  found by GC-MS to be due to
various phthalates.  Although these plasticizers would not interfere
with GC if a halogen-selective  detector was used,  they are a potential
 interference for electron  capture-GC and  GC-MS analysis (13),  Glass
 containers with  solvent-rinsed  aluminum foil or Teflon-lined caps are
 generally  acceptable as sample  containers and for  storing purified
 reagents.
                                   -80-

-------
                                                       Section 3K
figure 3-D.  Gas chromatog-
raphy of hexane extract of
cardboard bottle cap liner
on 10% OV-3 column with a
hydrogen flame ionization
detector at 16 x 10~H
amp/mV sensitivity.  Amount
injected represents 1/100 of
the extract from one 3/4 inch
diameter liner.
Other examples of problems with reagent contaminants have appeared in
the literature.  Bevenue et_ al_. (14) reported on the "contribution of
contaminants by organic solvents, glassware, plastic ware, cellulose
extraction thimbles, filter paper, and silica gels to water samples
causing interference with subsequent GC analysis in the ppb range.
Prior to their use, heat treatment of glassware and the silica gels
was recommended to eliminate contaminants, while plastic ware and
filter paper were excluded from the procedure.  Levi and Nowicki (15)
found that cloth bags contained residues that were absorbed by cereal
grains stored in these bags and gave spurious GC peaks with electron
capture detection.  The same workers (16) found that Na2S04, filter
papers, solvents from wash bottles, Teflon gaskets, and glass wool
•produced interfering EC-GC peaks and gave methods for their elimination.
Bevenue and Ogata (17) reported on the contribution of extraneous
components by high purity, analytical grade basic reagents used for
adjustment of pH during isolation steps in the analysis of chlorophenoxy
acid esters and ethyl or methyl derivatives of hexachlorophene and PCP
in plant and animal tissue and water samples.  Baker et^ ai_, (18)
found contamination of acetone with an impurity corresponding to CCl^
and interfering in the analysis of the latter pesticide (fumigant) by
EC-GC.  It was shown that this contamination could be caused by CCl^
in the laboratory atmosphere, possibly arising from the use of aerosol
propellant cans for spraying thin layer chromatograms.  Trotter and
Young  (19) found that impurities in SbCl- reagent caused erratic re-
coveries of PCBs in perchlorination procedures.
                                   -81-

-------
                                                            Section 3L
   The chromatogram in Figure 3-E is of material extracted from disposable
   vinyl laboratory gloves.  A chemist wearing these gloves had touched
   the lip of a concentrator tube with the fingertip of his gloved hand.
   One of the extraneous peaks produced coincided exactly with the com-
   pound (TCDD) that was being determined.

   In view of these problems, it is mandatory that reagent blanks be run
   constantly for each analytical procedure, with final extracts being
   reduced to the same concentration level normally used for the sample
   material.  A reagent blank involves repetition of the entire procedure
   without including the sample itself.
        Figure 3-E.
Electron capture gas chromatogram of material
extracted with ja-hexane from outside surfaces
of disposable vinyl plastic laboratory gloves.
                                        468
                                       TIME/min.
3L   GC RETENTION DATA FOR COMMON INTERFERENCES

   Table 3-4 contains relative retention data for common contaminants on
   several GC liquid phases used in EPA and FDA laboratories (Section 5L).
   These compounds will be eluted at the same positions as certain pesti-
   cides (EPA PAM, Section 4) and will, therefore, interfere in the
   analysis of the pesticides or be mistaken for them.
                                     -82-

-------
 c
 o
*1
•p
 o
 CO
OQ
     co
     cd
     EH
           w
           M
           O
           M
           EH
           W
           W
           g
           M
           OS
>H
»J


O
o
o

C/3
Q


1
P-

O
O

OS
o
PH

<
EH
<
Q

"Z.
O
M
EH
•Z.
w
EH
W
OS

O
C5
        0)
        g
                   C
                   o
                  •H
                  •P

                   CO
                  •P
                   CO
                  OS
       •H
       •P
        cd
       rH
        0)
       OS
            0
             §
 CO

 1
•H
 O
O
                O
                O T-»
                CM  I
                 I  fe
O
CM
 I
O
Q
                                   °  **

                CM
                           fc
                C^CM
                    I
       O
       rH

    C^CV


     I  ^
                            ic

                           -d" -H


                          , X'X
                 CO
                 ra
                 cd

                CU

                •o

                '3
         CO CO

         3 O -P
         •P-H cd
         cd co os
         ^ C

         P,g o
         6-HrH
         co Q Pt,
                                    G  e
                                    CO  3
                                    > rH
                                   o  o
                                      o
                                                       vO
                                                       CM
                                       000
                                       CM CO
                                                            O
                                                            OVO
                                                                                                              CM
                                                   COVO

                                                   o o
                                                                                                   J-
                                                                                                     •
                                                                                                   o
                                CM CMVO O cn   -d-vOCO
                                U>CN-O VTVd- CM  •  «  •
                                  ...  .  .  .rHOOtv-
                                O O CM «H i
                                             cn-d|

                                             C5 O «H O
                                                                                                              vr>.
                                                                                                              -d"
                                                                             CM
                                   CMOO

                             O O ••«-< O CM
                                                                          cnvo
                                                                                           vo
                                                                                                              co
                                                                                                              -d-
                                                                                                            eo
                                                                                CM
                                                                                                           CO


                                                                                                           CO
                                                                 a>
                                                   CD
                                                   •P
                                                                          cd
                                                                                      CQ
                     •H
               CO     Cd
co    co a    -p    2:
•p    -p +5    cd
co    cd cd    o     g
CO    P,rH    Cd     O
                                                                                                             cu
                                                                                                            13
                                                                                                           o
                                                                                                           •p
                                                                                                           o
                                                                                                           CO
                                                                                                           •p
                                                                                                           CO
                                                                                                           •o

                                                                                                           CO
                                                                                                           f-l

                                                                                                           •p

                                                                                                           &
                                                                                                           o
                                                                                                                       •p
                                                                                                                        o
                                                                                                                        CO
                                                                                                                       rH
                                                                                                                        CO
                                                                                                                                  .H
                                                                                                                       •H
                                                                                                                       •P
                                                                                                                       •H
                                                                                                                        CO

                                                                                                                        §
                                                                                                                        (0

                                                                                                                        -p
    CO rH
 co -a  o  co
 >     m  >

'•p  o  CO'-P
 cd-H 45  rt
•H -P -PrH
 co  cd     co

   •H -P
 C  C-H  C
 O  O  5  O
•rH *rH    «rH
-P     CO -P
 C  co  co  C
 CO  g -P  CO
•P  rt  3 -P
 CO rH rH  CO
K TTi TT1 f*^
                                                                                 83

-------
                                                           Sections 3M,  3N
 3M   CLEANING OF GLASSWARE
    The residue analytical chemist must be sure his glassware is entirely
    free from contamination.   The cleaning operation generally includes:

    a.   Soaking and  washing in a high temperature (50°C)  bath of synthetic
        detergent  (e.g.,  Alconox) in water.

    b.   Rinsing with tap  water.

    c.   Rinsing with distilled water.

    d.   Rinsing with acetone.


    Cleaning  of glassware used to concentrate  samples  (e.g.,  K-D flasks or
    evaporative concentrator tubes)  should include  a soak for at least 15
    minutes in  hot (40-50°C) chromic acid  cleaning  solution  (observe rigid
    safety precautions) after  the tap water rinse to remove all  traces of
    organic material.  This soak is  followed by  thorough  rinsing with tap
    and distilled water and then with acetone  and hexane.  Pipets are washed
    in  the same way, preferably  using a  commercial  automatic  or  semiautomatic
    self-contained washing unit.

    Large glass items such as beakers and  flasks  are inverted and suspended
    to  dry in metal racks.  Small items  such as glass stoppers and bottle
    caps are wrapped in aluminum foil, dried in an  oven, and  stored in foil.
    Fipets are wrapped in  bundles  in aluminum  foil  and oven dried.

    Clean, dry  glassware is stored in a  dust-free cabinet.  (Stainless steel
    storage tubes are available  for pipets).  As an  extra precaution, each
   piece should be rinsed with  the solvent to be employed in the analysis
   immediately before use.  As  soon as possible after a piece of glassware
   has come in contact with a sample containing pesticides, it should be
   rinsed with acetone to remove surface residues.   If this is not done,
   the subsequent soak bath of detergent will pick up the pesticide and
   may then serve to contaminate all other glassware placed therein.
   Details for cleaning glassware are given in the EPA PAM, Section 3,A.


3N   HOUSEKEEPING

   Good general housekeeping procedures aife important in the analytical
   laboratory.   Benches should be neat, labels legible,  and files orderly.
   Certain contaminants such as cleaning agents and dust are impossible to
   exclude,  but others should not be deliberately introduced, such as by
   eating or smoking in the laboratory.  To reduce  possibilities of errors
   and cross contamination,  food, beverages,  or snacks should not be stored
   in a refrigerator used to store samples.
                                      -84-

-------
                                                          Section 30
30   ANALYTICAL PESTICIDE REFERENCE STANDARDS
   It has often been noted when evaluating chromatograms from interlabora-
   tory check samples that reference standards used in certain laboratories
   reporting rejected results were undoubtedly inaccurate.  (This can be
   determined by the coordinator by comparing the peak height ratios in
   the chromatograms from the check sample of precisely known composition
   against the same ratios from the laboratories1 internal standards).
   The proper preparation and storage of analytical standard solutions is
   of utmost importance.  Since the working, diluted standards may be in
   use for up to six months, any mistakes in preparation of the concentrated
   stock solutions or in their dilution would be reflected in the accuracy
   of analytical results for this entire period.  Incorrect standards will
   result in correspondingly incorrect analytical data even though first
   class technique is thereafter employed and all laboratory instruments
   are in perfect operating condition..  Even including improperly operated
   equipment, the greatest single source of quantitative error in GC
   analysis is undoubtedly inaccurate standard solutions.

   Identification and record keeping of reference standard solutions are
   activities that often receive too little attention in some laboratories.
   Its importance cannot be overemphasized, particularly in a laboratory
   concerned with law enforcement.  Therefore, the protocol should be
   formalized and standardized for all staff chemists within the laboratory
   group.  By so doing, it should be possible for any other staff chemist
   or a supervisor to consult, a given chemist's reference standards log-
   book years after an analysis was conducted and readily determine the
   precise identity and concentration of any standard .used in an analysis.-

   The logbook should reflect a complete record of each prepared reference
   standard solution, starting with the pure .primary standard and ending
   with the final working standard solution.  Data that should be docu-
   mented include weight of primary standard, concentration of all subse-
   quent serial dilutions, and the dates of preparation of all dilutions.

   In multiresidue analysis, it is common practice to prepare final working
   standards as a mixture of pesticides of interest to the laboratory, this
   subject to be treated in some detail later in this section.  Such a
   mixture should be assigned an identification number and so documented
   in the logbook.  The same number should be printed on the bottle label
   of the mixture and should also be used to identify all reference standard
   chromatograms during the life of the mixture.

   In one possible coding system, each standard is assigned a number pre-
   ceded by the 'letter C for "concentrated", I for "intermediate", and
   W for "working".  Referring to Figure 3-?I, the concentrated stock solu-
   tions could be given the numbers Cl (lindane), C2  (aldrin), C3 (dieldrin),
   C4 (o_,2/-DDT) ,and C5 (p_,2.f-DDT);. The number would represent the compound,
   and the prefix the stage of concentration.  After dilution, the intermediate
                                     -85-

-------
                                                       Section 30


stock solutions would be designated II, 12, 13, 14, and 15, respectively.
The final working standard mixture prepared from these solutions could be
designated Wl-5, or it can be given a totally new number such as W6.  The
latter is probably less awkward in certain situations, e.g., if the final
working standard mixture is remade using solutions 11-14 and a later-
prepared standard of p_,£T-DDT (perhaps designated 110).  The new working
standard would be designated Wl-4,10 with the former system, but could
be numbered more simply as ^7 if a unique sequential system number is
given to each solution.  It is likely that each laboratory can devise a
numbering system to suit its needs.  The important point is to use some
dear and consistent system to designate standards and to have records
fully describing the preparation of each numbered solution.  Sample
sheets for maintenance of the reference standards logbook are shown in
Figures 3-F, 3-G, and 3-H.  These forms are in routine use at the EPA
laboratory in Research Triangle Park, NC.

Organic compounds are subject to a wide variety 6f oxidation, hydrolysis,
isomerization, and polymerization reactions.  Instability of organic
standards is, therefore, often a problem.  Storage and use conditions
should be those that retard degradative processes, and purity should be
periodically rechecked.

Details for the preparation, storage, and use of pesticide analytical
standards are given in Section 3,B of the EPA PAM.  Some important  •
considerations as they pertain to quality control and identification
of potential trouble spots are outlined below.

a.  Primary Standards

    There are no officially recognized pesticide "primary" standards,
although in the parlance of the pesticide chemist, analytical grade
standards of 99% or higher purity are referred to as primary standards.
Purities of standards are commonly greater than 99% and seldom less than
95% but may be lower in some cases.  For example, chlordane and toxaphene
are available in technical grade with 60-70% purity.  The percentage of
purity must be known in order to apply a correction factor in weighing
out the standard for subsequent dilution.

There are several sources of pesticide standards.  Most manufacturing
companies will supply the analyst with technical grade pesticides and
in some cases a small amount of a more highly purified grade.  The
technical material may be purified by repeated recrystallization and
checked for purity by at least two analytical criteria such as elemental
composition; IR, NMR, or mass spectrum;  melting point;  GC trace;  or TLC
spot pattern.  The EPA Quality Assurance Program maintains a pesticide
calibration and reference materials repository at its Pesticide Labora-
tory at Research Triangle Park, NC.  This laboratory supplies 100 mg or
less of standards of certain pesticides, metabolites, and derivatives,
on a discretionary basis as time and resources permit,  to nonprofit,
government, and university laboratories.  EPA publication EPA-600/9-78-012
                                   -86-

-------
Figure 3-F
                        PREPARATION OF CONCENTRATED STOCK STANDARDS
                                                                                  Section 30
       final Gross Wt
            «Tare  Wt
              Net  Wt
        **Adj. Net Wt_
                            Date
                                      Lot Ho.
                                                  Chemist
                                               Solvent
                             Purity_
                   Dilution Vol._

                   Concentr.
                                                                              ng/yl
    No	
    Conipound_
       Final Gross Wt
             •Tare Wt
               Net Wt
        **Adj. Net Wt_
Date
                                                  Chemist
          Lot No.
Purity
       mg
                   Solvent

                  Dilution Vol._

                  Concentr.  	
       Final Gross Wt_

             *Tare Wt_

               Net Wt
        ••Adj. Net Wt
                            Date
                                                  Chemist
                                      Lot No,.
                                   mg
                   Solvent
                             Purity
                   Dilution Vol._

                   Concentr.
                    ml
                                                                             ng/Wl
       Final Gross Wt_

             •Tare Wt
               Net Wt_

        ••Adj.  Net Wt	
                            Date
                                                  Chemist
                                      Lot No.
                                                         Purity
                   Solvent

                  Dilution Vol__

                  Concentr.
       rag
                    ng/pl
    •If weighing into a  bcak«r,  this is tha empty beaker weight.  If weighing from a
     dropping bottle,  this  is the initial weight of bottle and contents.
   ••Corrected for purity of primary standard.
                                                 -87-

-------
Figure 3-G
                                                                   Section 30
                  PREPARATION OF STANDARDS OF INTERMEDIARY CONCENTRATION
    NO,.
     Compound
  Date   /   /
                                                       Chemist
                       Strength of Concentrated .Stock




                       Aliquot of Concentrated Stock	



                       Dilution Volume




                       Dilution solvent .  .  .
                                ng/M.
                                ml
                                ml
                       Final Concentration
                                                                 _ng/M.
   NO,.
    Compound_
 Date	/
 Chemist
                       Strength of Concentrated Stock




                       Aliquot of Concentrated Stock



                       Dilution Volume



                       Dilution solvent
                               ng/M.
                               ml
                               ml
                      Final Concentration
                                                                 ng/Pl
  NO,.
   Compound
 Date	/
                                                      Chemist
                      Strength of Concentrated  Stock




                      Aliquot of Concentrated Stock



                      Dilution Volume




                      Dilution solvent
                              ng/Ul
                              ml
                              ml
                      Final Concentration
  NO,.
   Compound
Date   /   /
Chemist
                      Strength of Concentration Stock_




                      Aliquot of Concentrated Stock



                      Dilution Volume




                      Dilution solvent
                             _ng/Pl
                              ml
                              ml
                     Final Concentration
                                                               ng/wi
                                                   -88-

-------
                                                                  Section 30
   Figure 3-H
                    PREPARATION OF FINAL WORKING STANDARD SOLUTIONS
No,.
    Date
Chemist^

Solvent
 1.
 2.
 3.
 4.
 5.
 6.
 7.
 8.
           Compound
             Cone,  of
Parent Sol.  Parent Sol.   Aliq. Vol.  Dilution  Final  Cone.
  Number        ng/til         ml      Vol.  (ml)    pg/ul
 No,.
    Date
Chetnist_

Solvent
 1.
 2.
 3.
 4.
 5.
 6.
 7.
 8.
           Compound
            Cone,  of
Parent Sol. Parent Sol.  Aliq.  Vol.   Dilution  Final Cone.
  Number       ng/ul     	 ml     - Vol. (ml)
No,
 1.
 2.
 3.
 4.
 5.
 6.
 7.
 8.
                                 Date
           Comoound
                                                         Chemist

                                                         Solvent
            Cone, of
Parent Sol. Parent Sol.  Aliq.  Vol.
  Number       ng/ul          ml
         Dilution  Final Cone.
         Vol.  (ml)    pg/ul
                                                  -89-

-------
                                                        Section 30

 lists available standards and supplemental data.   Purified standards
 can be purchased from a number of U.S.  companies  handling chromatographic
 equipment and -supplies and from the National Physical Laboratory,
 Ministry of Technology, Chemical Standards,  Teddington, Middlesex,
 England.   Purities of these standards are not always  what they are
 advertised to be,  so the chemist should always verify the purity  and
 repurify if necessary and practical.

 Concentrated stock standard solutions are conveniently made up at a
 200 ng/yl concentration by weighing 20.0 mg  of pure standard  and  diluting
 to  100 ml.   If  the primary standard is  given as 99.0  percent  pure, weigh
 20.0   or 20.2  mg; if the purity is given as 90.0%, the weight will be
 0.990
 20.0
 0.900
22.2 mg.
Toxicity levels and relative  stabilities  are  important  factors that
dictate the methods of handling  and  storing various pesticide standards.
Highly toxic pesticides  (low  LD50 values) require special precautions
such as wearing disposable rubber or plastic  gloves and avoiding inhala-
tion of vapors.  The stable organochlorine compounds may be stored at
room temperature in tightly sealed containers, while organophosphates,
which are subject to a wide variety  of oxidations, rearrangements and
hydrolytic reactions, should  be  desiccated in a refrigerator and allowed
to come to room temperature in the desiccator before use.  If standards
are stored in a freezer, containers  are not opened until warmed to room
temperature, or condensed water  vapor will be introduced.


b.  Concentrated Stock Standards

    Secondary standards are liquid solutions  of the primary standards.
The final concentration df working standard will depend upon its use,
e.g., pg range for electron capture  GC, ng range for TLC and other GC
detectors, and yg range for IR spectroscopy.

For electron capture GC, usually three dilutions of the primary standard
are made to arrive at the working standard.   An analytical balance
capable of weighing to at least  0.0001 g and  scrupulously clean glass-
ware are employed.  Stable, low  toxicity pesticides may be weighed into
a small beaker or cupped aluminum foil, transferring solid compounds to
the balance with a stainless  steel micro spatula and liquids with a pipet
or dropper.  Crystalline standards weighed on aluminum foil are transferred
dry through a small glass funnel into a volumetric flask, the foil and
funnel being carefully rinsed with solvent.   Standards weighed into
beakers are completely dissolved (observe carefully)  in a small volume
of solvent and transferred quantitatively by  rinsing with the rest of
the solvent through the funnel into  the volumetric flask.  Liquid primary
standards can alternatively be transferred to a dropping bottle with
ground-in stopper; the bottle containing the standard is weighed,  an
estimated amount of standard  transferred directly into a volumetric flask,
                                   -90-

-------
                                                       Section 3O


and the bottle reweighed, the loss in weight representing the net sample
weight.  This closed dropping bottle technique is mandatory for high
toxicity liquid pesticides.  Solid primary standards may be weighed
(10.0 mg) directly into 50 ml volumetric flasks, which will fit onto
the pan of most one pan analytical balances.  A procedure for storage
and transfer of degradable pesticides under inert atmosphere is given
in the FDA Pesticide Analytical Manual, Volume I, Section 132.

It is difficult with any of these techniques to weigh exactly the pre-
determined amount to obtain all solutions of 200 ng/lil.  It is seldom
necessary to take the trouble' to attempt this, in any case, since the
important thing is to obtain a formulation near to that which is desired
and to know its exact value.  This is calculated by dividing the known
weight by the flask capacity.  A possible procedure for preparing
standards of exactly a certain concentration is to weigh the solid and
then add only enough solvent (e.g., the s.olvent is measured from a pipet,
or the solid is weighed into' a graduated centrifuge tube and solvent
added to the appropriate line) to give the desired concentration.


c.  Intermediate Concentration Standards

    It is usually necessary to prepare standards of intermediate concentra-
tion by dilution of the concentrated standards and then to prepare working
standards by dilution of the intermediate standards.  It is impractical
and hazardous to prepare the final solution from the concentrated stan-
dard in one dilution or to prepare an original secondary standard at a
concentration low enough to allow only one subsequent dilution.  Some
analysts have attempted to make this enormous dilution by aliquoting
microliter volumes with a syringe into a volumetric flask.  This is
extremely poor technique, however, since an error of as little as 0.2 yl
in a 5.0 pi transfer will be grossly magnified when a 5 to 10 Vl injection
of the resulting solution is chromatographed.

Separate solutions of each compound or a standard mixture can be prepared
at this point.  The concentration level depends on the response of the
detection mode of the analytical procedure in which the standard will
be used.

All solutions should be equilibrated to room temperature before any
pipeting or diluting is carried out.  Volumetric transfer pipets should
be used where available, or a Mohr-type measuring pipet in other cases.
Be sure to note whether the pipet is calibrated "to deliver"  (TD) or
"to contain"  (TC) and use accordingly.  The accuracy of well cared for,
properly cleaned commercial Class A pipets and volumetric flasks is such
that  calibration is not required in order for potential errors from this
source to be  insignificant.

Pipets calibrated to deliver  their stated volume should be used if
possible.  Measuring pipets are  calibrated, like a buret, but do not
deliver a volume of liquid as  accurately or reproducibly as volumetric
pipets.  The  latter are  recommended whenever possible  for analytical
                                   -91-

-------
                                                         Section 30

 work.   Pipets  are filled by use of a rubber suction bulb  rather than
 the mouth.  After filling and  dropping the meniscus to  the  etched line,
 no  air  bubbles should be evident anywhere  in the  pipet.   The  outside
 of  the  pipet tip  is wiped free of liquid and the  tip then placed against
 the inside wall of the vessel  to which the solution is  to be  transferred.
 The liquid is  discharged,  keeping the tip  against the inside  for 20
 seconds after  the pipet  has emptied.   The  pipet is removed  from the
 side of the container with a rotating motion to completely  discharge
 any drop on the tip.  The  small quantity of liquid inside the tip is
 not to  be blown out; the pipet has been calibrated to account for this.
 Only properly  cleaned and  dried pipets can be inserted  into the  solution
 container without fear of  contamination or dilution.

 d.  Working Standards

     Working standards are  generally made up as mixtures,  the  actual
 combinations being dependent upon the compounds of interest and  the
 ability of the analytical  method to resolve them.   Care must  be  exercised
 that these mixtures do not contain compounds  that  may react with  each
 other.  Each working standard mixture should  be made  up at two or even
 three concentration levels, depending on variations in pesticide  concen-
 trations in routine,samples.  No  compound  should be present in such
 concentration  that when  injected  into  the  gas chromatograph the linear
 range of the detector will be violated.  If p_,£'-DDT  is present in a
• standard mixture, neither p,j>'-DDD nor j>,p'-DDE should be present since
 these compounds are breakdown products pf^DDT and  their presence or
 absence is useful for monitoring this  degradation.  All compounds present
 in  each mixture should be  resolved by the  working  GC  columns  used in the
 laboratory.  Suggested mixtures  and concentration  levels  of common chlorin-
 ated pesticides for,laboratories  analyzing tissue  samples by  EC-GC with
 the recommended coltjmns  (Subsection. 5L)  are given  in  the  EPA  PAM, Section
 3,B.  A typical mixture, diagrammed in Figure 3-1, may be prepared as
 follows:  weigh 20.0 mg  each of primary  standard lindane, aldrin, dieldrln,
 o.»Z'-DDT»  sad .g^'-DDT int° separate 100 ml volumetric flasks to prepare
 concentrated stock solutions of 200 ng/yl  each.   Transfer, respectively,
 0.5, 0.5,  1.0, 2.0 and 2.0 ml of each of these by  separate pipets to
 individual 100 ml volumetric flasks to prepare intermediate stock standards,
 Transfer 2.0 ml of lindane, aldrin, and dieldrin and 1.5 ml each of
 o^'-DDT and £,£f-DDT to the same 100 ml flask to. prepare a final working
 standard mixture containing, respectively,  20, 20, 40, 60, and 60 pg/yl.
 For other than EC-GC,  stock standards of 0.5 mg/ml and working standards
 from 50-100 to 0.5-1 ng/yl are typical.  The procedure outlined in Figure
 3-1 is but one option for preparing the required solution; other approaches
 requiring less glassware are possible.
                                   -92-

-------
                                                         Section 30
   Figure 3-1.  Serial dilutions of pesticide standard mixture
                PREPARATION OF WORKING STANDARD A.

           Undone    Aldrin    Bieldrin    o,p'-ODT   p,p'-DDT
                    O       C3      D       D   Primary Standards

                                  '     ''
                                               Cone. Stock Stdt.
                                                (200ng/^leadO
                                          Ar*"1***
                                          G5_V  Irrterm. Slock Stds.
                                 Final Working Standard Mixture
e.  Choice of  Solvents and Storage of  Standard Solutions

    The EPA ETD laboratory has evaluated storage conditions for analytical
reference standard solutions to minimize the decomposition of the
reference standard in the solution and the evaporation of the solvent
from the solution.

(1)  Solvent Evaporation Control

     The evaporation of the solvent  containing the standard is easy , to
evaluate and observe.  Under almost  all conditions, some solvent evapora-
tion can be measured.  This solvent  evaporation must be minimized  to
maintain the concentration of the standard solution for a reasonable
length of time.  The study of solvent  evaporation centered around  the
following factors:

            choice of solvent

            solution volume

            choice of container

            storage temperature

 (a)   Choice of solvent

      The  chosen solvent should easily dissolve the reference  standard,
not  chemically alter  the reference  standard, be compatible with the
                                    -93-

-------
                                                       Section 30

method of analysis, and not evaporate very rapidly.  Solvents that
evaporate rapidly include diethyl ether, petroleum ether, pentane, and
acetone.  These solvents are very poor choices for the long-term storage
of reference standards.  Desirable solvents from the standpoint of
evaporation are isooctane, isopropanol, and toluene.
The use of hexane is very popular in residue laboratories.  However,
hexane evaporates 2.6 times more rapidly than isooctane at ambient
laboratory temperatures from closed volumetric flasks with glass
stoppers.  The evaporation rates of different test solvents under these
conditions and the predicted placement of- other commonly used solvents
that were not tested are listed in Table 3-5 in order of decreasing
evaporation rate.               '
                              Table 3-5
             EVAPORATION RATES OF ORGANIC SOLVENTS FROM
                  GLASS STOPPERED VOLUMETRIC FLASKS
       Solvent
       Pentane      :
       Diethyl ether
       Methylene chloride
       Acetone
       Hexane
       Chloroform
       Benzene
       Methanol
       Ethyl acetate
       Acetonitrile
       Ethanol
       Isooctane
       Heptane
       Isopropanol
       Toluene
       Decane
Evaporation Rate
   (ml/wk)
     0.634
     0.254
     0.221
     0.158

     0.096
     0.086
     0.061
     0.045
                                  -94-

-------
                                                       Section 30
(b)  Solution volume
     Volumetric flasks of four different sizes were evaluated as to the
effect of solution volume on the solvent evaporation rate.  The absolute
evaporation rates of hexane from the flasks (5 ml, 10 ml, 50 ml, and
100 ml) were 0.167, 0.155, 0.113, and 0.100 ml/week, respectively.'  The
decreasing evaporation rates with increasing solution volume in itself
makes the larger volume flasks more desirable.  The relative evaporation
rates (percentage of the solution volume evaporated per week) of the
four flasks were 3.34, 1.55, 0.226, and 0.10% of the container volume
evaporated per week, respectively.  As GC measurements are sensitive to
3% volume changes, the use of the 5 and 10 ml volumetric flasks for
more than one week is not recommended.  The use of .the larger solution
volume increases the useful lifetime of the reference standards.

(c)  Choice of container

     Several different types of glassware were evaluated at ambient
temperature in an attempt to find a container that is easy to use and
will retard solvent evaporation.  The containers tested included:

     Volumetric flasks with glass S stoppers
     Screw-cap prescription bottles with cardboard lined plastic caps
        (Brockway Glass Co., Inc., Brockway, PA)-with added Teflon disc
        cap liners  (A. H. Thomas, #2390H)

     Multivials  (Supelco, Inc., #3-4579), 10 ml size
     Serum bottles  (Wheaton Scientific  #223739) , with Teflon faced septa
        (Wheaton  Scientific #224167) and metal seal  (Wheaton Scientific  #224182)

     Small volume  flat bottom and  conical vials from various sources
        with appropriate  cap.3  (Supelco,  Inc., #3-3291, 3-32.33, and
        3-3300; Wheaton Scientific  #225170 and 224882), 0,3, 0.6, and
        1.0 ml sizes

Figures 3-J and  3-K illustrate  these  containers.  Both hexane and isooctane
were used as  solvents in this evaluation, the results of which  are summar-
ized in Figure 3-L.

The evaporation rate of  hexane  from the volumetric  flasks was the fastest
found  in this study.   The use of volumetric flasks  to store hexane solu-
tions  is discouraged.  Isooctane also evaporated  quite  rapidly  from volu-
metric flasks.  The evaporation, rates of hexane and isooctane are sig-
nificantly  reduced in screw-cap prescription bottles when compared to
 the volumetric flasks.   The evaporation rates  of  the solvents  from the
prescription bottles was approximately 0.025 ml solvent evaporated per
week.   This rate is low enough to allow use of these containers.
                                    -95-

-------
                                                                 Section 30
           Figure  3-J.   Serum vial with Teflon-lined septum
Figure 3-K.  Other containers used in evaporation study.  Back row (left to
             right):  prescription bottle, volumetric flask, laultiviai.
             Front row.  flat bottom vial, conical vial.;. conical vial.
                                      96

-------
 c
 o
•H-
-P
 O
 CO
00
        ra
        M
        CD
        c
        •H
        OJ
        +?
        C
        o
        o
        0>
       •H
       •a
        ca
        -p
        C
        CD

        rH
        O
        to
        cu
       -p
        rt
        C
        o
       •H
       -P
        aJ
        ^
        o


        I
        CD
        f-<

        bfl
       •H

—



•™"
_ ..
_ .

—

m


— •


__



-r


—
UJ UJ
2 2
< <
X 1-
UJ CJ
X 0
O t*
"1 vi
CO —
^/ ,1^
CO t-
< C
_1 CC
"• 2
O C
cr H
H- 0.
111 5

-J LL
o cc
> o-




—




—
— .

—

- — _





^~




«•• '
II
*
VI
WJ
V
•1
•
—


—



il
o
r—
CJ
IT}
C3
ro
o
^
*"i LU
0 £
cc
2
2 O
o —
° <
cc
0
Q.
r*^ 
O 1 1 1
LO
O
0
CO
0
o
o
o
ujoi ^ii 5 5 • UJ ui
| < | | It 1|
' 2. 0 22 co w x 0
1 8 • "i a • - 2 s °
CO < < CC CC -
_l -J 13 13
O C3 1— 1—
•••' • 3 3- § §
3 3 UJ => 3
2 ' OL j o- °- M
s s H o ^ 3
_J < < H- > & <
<< ^. O & uj —
^ O g CO uj 2 >
: ^ 5 rii ^ 2 < _j
i r^ "^ **• ^™ -»< sx ^i
! H uj 2 D < >< ^
1 ^ 2 < CC X u^ <
: i < x "4 x i
^ ^v tu C/3 J- t/3





UJ X
X
-97-

-------
                                                               Section 30

     The use of the multivials, septum bottles, and small vials are all dis-
     couraged due to problems of handling or large relative evaporation rates.
     In the case of multivials, the evaporation rate of solvent from the con-
     tainer when used as a screw cap vial was high (0.071 ml/week) unless the
     extra glass below the snap-off score line of the ampule seal was removed.
     The multivials also contaminate the solution in the vial with torn up
     pieces of septum cap seal if the ampule glassware in the vial neck is
     not removed.  In the case of the serum bottles,  the evaporation rate of
     the solvent was 0.094 ml/week after the bottle seal had been punctured
     with the beveled needle of a standard 10 Hi Hamilton syringe.  The rela-
     tive evaporation rates of the small vials were all greater than 5% of
     the solution per week, making them useless for long-term storage of
     standards.

     (d)  Storage temperature

          The evaporation rates of hexane from 10 ml  volumetric flasks at
     ambient temperature, +3°C, and -15°C were 0.155, 0.0575,  and 0.0226 ml
     per week,  respectively.   This represents a reduction in evaporation by
     a factor of 2.6 when the solution is stored in the refrigerator (+3°C)
     and 6.8 if the solution is stored in the freezer (-15°C),  compared to
     room temperature.   Storage under refrigeration or in the  freezer when
     not in use significantly increases the useful life of standard solutions.

(2)   Standard Chemical Stability Control

          The chemical  stability of organochlorine, organophosphate,  carbamate,
     and triazine reference standards in solution was evaluated for one year
     under four  different storage conditions:   at ambient temperature  (23-24°C)
     out on the  bench top in natural and fluorescent  light;] at  ambient tempera-
     ture in the dark;  in the refrigerator at +3°C; and in!the  freezer at
     -15°C.   The results  were as follows:

     (a)   Organochlorines

          All of the 28 organochlorine  compounds  tested (Table  3-6) were stable
     in isooctane solution under the four test  conditions.

     (b)   Organophosphates

          All of the 20 organophosphate  compounds  tested  (Table 3-6) except
     disulfoton  were stable in isooctane solution  under all of the  four test
     conditions.  Disulfoton was not stable under  any of the test conditions.
     Solutions of disulfoton  should  be replaced monthly if not stored in the
     refrigerator or freezer when not in use.  Solutions of disulfoton that
     are properly stored  in the  refrigerator or freezer should be monitored
     for decomposition  bimonthly.
                                       -98-

-------




D
*"%

C
o
H
0
0)










S
OS
§
^J
C-4
M
w
rj
|— |
O
HI
E-i
CO
P«
o
M
I-1 ,

S
•f~ 4-i
M %-^


V)

4-*  CJ
o o
JC 1 ,
CL O
O 
C •!-
(O
cner
S* •(••



















in ' — .
(U 01
CZ C

'si 4->
£8
JC 1
u o
o 
w
o> c:
o-~*





O) OJ C OJ C
c c c: T- -r- o t. •!- c s-
>> O -r- N N +J .4J N -r- 4J
^_ 4^ f^4 (O fO ^^ QJ (& f*4 ^^
OJI-i- 3 tO O O 3 C
OOn3£-l(*--ClO''— X, 0 C "-
^ «r- l/> O JCZ .
4J J= O -r- ' 4J '•
ft! 4-> J= -C (a-f-.4-> CX"OOCLOO-I->
"a. o" £" r25 lo)^ o O.-P §"5'>,>c'5 S 1o H'gj
•r- S- r— U. O SO«1-Ej=§ Q-r^ •*•>'> S- 0> O O 'g '
<
o o fc -o ai .£ • ^
rt> fll ' -»••» " •»•• £Z OJ i» **™^
4-> 4J ^-..c P* x a> -a . o j=
-_'~^^ ai <-> o o M s- +-»o
COvi •— • Q. C O T-OJ.
vo «^ns-4-» i—  i— s_ OJE+J
r— ui T3 — • -Q -c o s- c: o — '
oevj c cc s_i-ot>r— oms-
I— r— -r- OJ 0) OJ (O(O O O S- >)-C r— "O O OJ
,_icc:c H- a uj H- c <«-»«- r— ^-oxo j=i-i— c
•r-i— r— CJCJS-S-S- 1 1 > IT3COU>-F-«0(00-aOXC:[cnjQ.

•o o o en co •— •— •— ex. o.|O.|O.|Qja> ~a -o T3 cxcLxar-t-* s-2; >>c x
i— S. slii-c^r^icj »*•• « •>•!— -c c c 4> v ai i a>v-ixa>o
-c_>ooo o |Q- ^i.|O.|O uj LU LU re 3; 3: i— SS^OQ-H-



































-P
T— 1

ON
ON
>. ^

O
rt
x:
0)
e
1
0)
C
0)
C
0)
.a

^M
•rH
"D
a>
rH
O
CO
CO
•H
Q
*
-99-

-------
                                                        Section 30
 (c)  Carbamates

      Eight of the 13 tested carbamates (Table 3-6)  were stable in toluene
 solution under all four test conditions.   Standards-of CDEC and butylate
 decomposed under all four storage conditions.   Butylate decomposed
 approximately 50% per year under all four storage conditions.   CDEC
 decomposed between 50 and 98% per year depending on the storage conditions.
 Carbaryl,  methiocarb, and carbofuran decomposed 38,  48,  and 17%,  re-
 spectively, when stored at ambient temperature with exposure to natural
 and fluorescent light.

 Solutions  of  CDEC and butylate should be  stored in  the freezer when not
 in  use and replaced monthly.   Solutions of carbaryl, methiocarb,  and
 carbofuran should be  stored away  from light and replaced after six
 months.

 (d)  Triazines

     Half  of  the  10 triazines  tested  (Table 3-6) were  stable to decomposi-
 tion of the reference standard material in ethyl acetate-solution under
 all four test conditions.  Prometryn, prometon, atrazine, and ametryn
 decomposed between 12 and 17%  at ambient temperature with light exposure.
Atrazine and atraton degraded  approximately 15% at ambient temperature
when stored in the dark.

 Recommendations for Storage of Pesticide Analytical Reference Standards

     The following recommendations are made concerning the factors that
affect analytical reference standard solution integrity:

     Select a  solvent, such as isooctane or toluene, that will  dissolve
     the standard material and evaporate as slowly as possible.

     Store the standards in relatively large  volume  solutions (50-100  ml)
     to  reduce the percentage volume losses to acceptable levels.   Monitor
     the solvent loss  by placing an indelible mark on the side  of the
     solution  container when the container is  filled and then discard
     the solution when 3-5% of the solvent has evaporated.

     Select a  container,  such as a -screw-cap prescription bottle or a
     large  volume volumetric flask, that 'will  not allow rapid solvent
     evaporation.

     Store  the standards  away from light in a refrigerator when not  in
     use  to reduce evaporation and reference material decomposition.

    Replace the  easily decomposed reference standard solutions  at the
    recommended  intervals  (e.g.,  monthly for CDEC, butylate, and
    disulfoton).
                                  -100-

-------
                                                          Section 3P


        Periodically check standard solutions by comparison against fresh
        dilutions of the stock solution.to prove that the solutions are
        still valid for qualitative and quantitative use.-
                                        \,
        Do not store any standard solutions for longer than one year.

3P   CALIBRATION AND MAINTENANCE OF THE GAS CHROMATOGRAPH'AND ACCESSORIES

     It is essential that the entire gas chromatograph be maintained in top
   operating condition if high quality analytical data are to be produced.
   In appraising results of interlaboratory check samples, it is clear
   from data and chromatograms that this is not the case in some labora-
   tories.  Section 5 will present details of proper operation of a gas
   chromatograph.  This section will offer guidelines for making routine,
   periodic checks of equipment to insure continued good operation and
   minimal down time.  Correct procedures for the operations mentioned
   (e.g., silylation and conditioning of columns, obtaining background
   profile) will be described infections 4, 5, and 6 of this Manual.

   Certain checks should be made daily, others on a weekly or monthly basis.
   Table 3-7 outlines the suggested frequency of such instrumental checks
   for a chromatograph equipped with an electron capture detector.

   It is suggested that a written log be maintained for  each instrument,
   recording the following data:

   a.  Date of installation and serial number of each detector installed
        (this will also serve as a record  for Atomic Energy Commission
       inspection).

   b.  Background current  (BGC) profile furnished with the detector under
       the EPA Interlaboratory Quality Assurance Program or from the
       commercial manufacturer.

   c.  Your  own  BGC  profiles obtained at  time of installation of each
       detector  and  subsequent profiles  (column identity notations  should
       be made).

   d.   Date  of change  of pyrometer batteries,  if used.

   A record  should also be kept  of each GC column  packed and  installed in
   an instrument,  logging such information ast

   a.   Assignment of a column number.

   b.   Date  of packing column.
                                                        •
    c.   Liquid phase identity and lot number of precoated column packing.

   d.  Conditioning temperature,  flow rate, and number of hours.

    e.   Length and shape of column.
                                    -101-

-------
                   Section 3P





i






-
1
u
UJ
i
i

§
5 • *"

s? a
fH 21


t>
3
X
1
0
i
ee
s
i
tn
H



fi
I .
••"9
i/»«3
8§
If U

C


o*5
r»



44
I*

^»
§

VI «J *
SS
63 e

SlJ-
58

*~


if
<* 0.
««
•sS
ss
ll
^i S?
Ca-3
^

T9
e
:«
r"5
S3
** u
19 
8 41 §
C «-»*—
Srt '

4)
cncr c
c fa o
•a «u cj
5-Gii:
*
CM






ml
•
VI
1
s-
VI

en
e
19
5

CvJ


e
Ir
ll
sH
ni

?

U1
1
V)
U •
U*—
£ §

j^ O


u «S
CO

u S
^* c
g v^
0. U
•M 41
Sen
rtS

L. r«


C C
ll^l
M3 O .

c-b u
tt i 3

CO


Check recorder electrometer
zero and noise level at
operating attenuation.
CO



s.
!?
S!
*O)3

^ VI

til m
64!
*
P» •
A JE **•
| SS
Sw" ?
e Q
A> Of
42 S*.2 S,
1. *-»
C O> « «O


w 2 G"O
++ u u a.

•*'


«
2
si
sle
ffte
^



s^
«j t-M p


U ^< -M ** •
at u c 'r~ c*j
6 a.S o>^

in


Check temperature of detector.
inlet, transfer 'and colunn oven. \J
in



|| c|
ill «s.
£ " oj 2
** W» *3 Ck
II § is
£L> 4J O
01 ni C w
•M C 3 f— Of
SSS -31= .
•f O C
^£41 AC U 0
U M 01 (J 4-» «f-
OI.O^E 01 C -M

• *
10 1-



















Chronatograph standard £»£' -DOT
on each working column used at a
concentration and attenuation that
will produce a peak of 50 to 60
percent full scale deflection (fsd). 2/
•a



*s
S3
OH- §
t- 1/1-0
0-33
• Is-!
E o §

(££ jj Q

01 "O 1
.c -o c:
»— « o
2*1
f
0 S
vi "5
3 •
4t (U C
j; j"| Ql
E 41 •*•

vi 0 4»
« u »-
*pg

U Vt VI
fit


0^-0)
^"o*?

??l


Oven temperature should be
nonitored by an outboard
thermometer.
>,
n H
* O

'-'2
liss


*~ £"
§,2 «

5i


The formation of £.£'-000 and/or
£.£'-DOE indicates on-colunn breakdown.
Si
-102-

-------
                                                          Section 3Q
   f.   Background current obtained on newly installed column and subsequent
       background current profiles during the life of the column.
   g.   Date of each silylation of column.
   h.   Compound conversion data, with dates monitored, and percentage of
       compound breakdown.
   i.   Monthly, chromatograph the following special column evaluation
       mixture, recording absolute and relative retention data and
       efficiency based on the £,£f-DDT peak.
                   Chlorinated Pesticide Mixture for GC
                            Column-Evaluation
              Compound
Concentration, ng/yl
             O-BHC
             P-BHC
              Lindane
              Heptachlor
              Aldrin
              Eept. Epoxide
              _p_,j3'-DDE
              Dieldrin
              Endrin
               ,2.' -DDT
     0.010
      .040
      .010
      .010
      .020
      .030
      .040
      .050
      .080
      .080
      .080
      .090
      .100
3Q   ADHERENCE TO OFFICIAL OR STANDARDIZED METHODOLOGY
   If reproducible and corresponding data are to be produced on both routine
   samples and interlaboratory check samples by a group of different labora-
   tories, it is essential that uniform standard analytical methodology be
   used by all.  In the EPA program, this methodology is developed, tested,
   and collected in the Analytical Manual by the coordinating laboratory
   of the quality assurance program, in close cooperation with the EPA
                                    -103-

-------
                                                          Section 3R

   methods development section.  Individual laboratories in the multi-
   laboratory system are encouraged to suggest improvements in existing
   procedures, but at no time should any individual summarily introduce
   method revisions, changes in GC columns, alterations in instrumental
   parameters, etc., without consultation with the coordinator or authors
   of methods. , Past experience clearly indicates that the vast majority
   of poor analytical performances on interlaboratory check samples were
   performed by laboratories deviating in some way from the standard
   procedures.  It is important, therefore, that laboratories adhere to
   standard analytical methods, but also that they report any problems
   with them to the coordinator so that these methods can be further re-
   searched and improved as experience dictates the necessity.  A standard
   procedure is generally not circulated until such time that reproducibility
   and precision have been well established.  Chemists having troubles with
   some phase of a standard procedure should search internally for the cause
   of the difficulty rather than making revisions in the method that cannot
   be fully studied and statistically evaluated by the individual.


3R   IMPLICATIONS OF AN INTRALABORATORY QUALITY CONTROL PROGRAM

   An intralaboratory quality control program such as described in the
   preceding pages requires a good deal of time and effort and does not
   come cheaply.   It is a conservative estimate that around 15% of the
   typical analytical laboratory's resources should ideally be channelled
   into quality control.  The questions often arise, particularly in a
   smaller laboratory, "Is such a program worth all this effort and expense?
   What is the return on the investment?"

   Each laboratory administration officer must resolve the answers to these
   questions in light of the impact of his ultimate analytical data.  If
   his laboratory is regulatory in nature, would he feel comfortable going
   to court to defend the validity of his analytical data?  Would his
   control program hold up under a barrage of cross-examination questions?
   If the laboratory's work is primarily of a monitoring nature, would he,
   for example,  feel fully confident in advising his superior officials that
   a given waterway is carrying a pollution load of x micrograms per liter
   of PCBs?

   From observations in the EPA interlaboratory quality assurance program
   (Section 2),  it can be stated without reservation that laboratories
   lacking a systematic internal control program more than likely will do >
   very poorly in the analysis of a blind sample.  In numerous Instances,
   laboratories  joining the program and analyzing a blind for the first
   time have performed rather badly in contrast to the peer laboratories
   that have been practicing rigid internal quality control.   The practical
   implication of this, of course, is that analytical data from such loosely
   controlled laboratories are simply unreliable.
                                                             .1
   To cite a specific instance, one laboratory Joining our program and re-
   porting the results of their first analysis of a spiked water sample
   reported the presence of £,£*-DDE, p_,j>/-DDT, o_,p_'-DDT,  heptachlor
                                     -104-

-------
                                                           Section 3S

   epoxide, o^'-DDE, and dieldrln.  The actual spiking  composition was
   HCB, oxychlordane, £,£f-DDE, p_,p_'-DDT, and Aroclor  1254 (PCS).   In other
   words, the analyst found two of the compounds  that  were actually present,
   four that were not present, and missed three that were  present.

   It takes no .great stretch of the imagination to assess  the  reliability
   of routine analytical data from this laboratory.  Such  data would  do
   far more harm that good.

   Unfortunately, laboratory administrators are sometimes  inclined to regard
   analytical data as a production commodity, expecting  x  numbers  of  analyses
   to be completed in y length of time with little thought to  such ancillary
   factors as quality control or specific analytical problems  related to
   certain samples.  We have no great quarrel with output  norms,  provided
   that quality control activities are built into the  norms.   When they are
   not, analytical data such as those described above  should not  be regarded
   as unusual.

3S   REFERENCES

   (1)  Singmaster, J. A., and Crosby, D. G., Bull. Environ. Contam.  Toxicol..
        16_, 291 (1976).

   (2)  Ruzicka, J. H. A., and Abbott, D. C., Talanta, 20_, 1282  (1973).

   (3)  Williams, I. H., J. Chromatogr. Sci., 11, 593  (1973).

   (4)  Saunders, D. L., J. Chromatogr. Sci., 15, 372  (1977).

   (5)  Majors, R., Varian Instrument Applications, 10 (3), 8  (1976).

   (6)  Analytical Methods for Pesticide Residues in Poods, Department of
        National Health and Welfare, Canada, Section 12.1  (b)  (1973).

   (7)  Fi)A PAM. Section 121.

   (8)  Wanchope, R. P.. Anal. Chem., 47, 1879 (1975).

   (9)  Huckins, J. N., Stalling, D, L., and Johnson,  J. L., J. Assoc. Off.
        Anal. Chem.. _59_, 975 (1976).

  (10)  Kirchoff, J., Dt. Lebensmitt. Rdsch.. 7J), 284  (1974).

  (11)  Schwartz, D. P., J. Chromatogr.. 152, 514 (1978).

  (12)  Osadchuk, M., Salahub, E., and Robinson,  P., J.  Assoc. Off. Anal.
        Chem., 60, 1324 (1977).

  (13)  Oswald, E. 0., Albro, P. W., and McKinney, J.  D.,  J. Chromatogr.»
        98_, 363 (1974).
                                     -105-

-------
                                                        Section 3S


(14)  Bevenue, A., Kelley, T. W.,  and Hylin,  J.  W.,  J.  Chromatogr.. 54,
      71 (1971).                                                     ~"

(15)  Levi, I., and Nowicki, T. W.,  Bull.  Environ. Contain.  Toxicol.. 7,
      133 (1972).                    '	~  -

(16)  Levi, I., and Nowicki, T. W.,  Bull.  Environ. Contain.  Toxicol.. 7,
      193 (1972).                                                     ~

(17)  Bevenue, A., and Ogata, J. N.,  J.  Chroinatogr.. 61,  147 (1971).

(18)  Baker, P. B., Farrow, J. E., and Hoodless,  R.  A., Analyst.  98,
      692 (1973).
  •
(19)  Trotter, W. J., and Young, S.  J. V., J.  Aseoc. Off. Anal.  Chen..
      58, 466 (1975).
                                  -106-

-------
                                   Section 4

             EVALUATION, STANDARDIZATION, AND USE OF MATERIALS FOR
                           PESTICIDE RESIDIE ANALYSIS
4A     ADSORBENTS

   Cleanup and preliminary fractional:ion of sample extracts are most often
   accomplished by chromatographic elution through a column of an active
   adsorbent.  Florisil, a synthetic magnesium silicate, is most widely used
   for this purpose;  but the latest trends indicate that partially deactivated
   silica gel and alumina, charcoal, and adsorbent mixtures, as well as gel
   permeation chromatography (1), are becoming increasingly popular.

   The activity (adsorptive strength) of adsorbents can be checked by elution
   of standard dye materials (2), lauric acid, or standard pesticide mixtures
   through a prepared column.  Host materials may be activated by strong
   heating, and some may be activated for a particular purpose by pretreat-
   ment with an acid or base (e.g., alumina) or an organic solvent (e.g.,
   charcoal).  Deactivation of a polar adsorbent to a desired level has been
   achieved by addition of a certain percentage of water.

   Florisil has proven to be nonuniform in elution characteristics (3,4) and,
   therefore, each batch requires careful pretesting of the adsorptive
   properties prior to use.  One activity of the EPA Health Effects Research
   Laboratory, Environmental Toxicology Division, Analytical Chemistry Branch
   Interlaboratory Quality Control Program (Section 2) is the furnishing of
   uniform, standard quality Florisil to other EPA laboratories and to labora-
   tories with direct contracts to conduct environmental monitoring.  Pro-
   cedures specified by the program coordinator and other available standardiza-
   tion methods will be described in the following subsections.

   a.  EPA Procedures for Handling and Evaluation of Florisil

       Details are given in Section 3D of the EPA Pesticide Analytical Manual.
       Especially high quality lots of Florisil (calcined at 1250°C) are
       purchased from the manufacturer in 200-400 Ib quantities after favorable
       evaluation of an advance sample by the coordinator of the interlaboratory
       program.  When the entire lot is received, another evaluation is made
       on plugs taken from each polyethylene-lined fiber shipping drum by
       means of a grain trier.  If satisfactory, adsorbent is transferred
       from the drums to specially cleaned wide mouth glass jars with foil-
       lined metal screw caps and a capacity for 2 Ib of Florisil.

       Evaluation of Florisil for use in a modified Mills, Onley, Gaither pro-
       cedure is made by heating Florisil in an Erlenmeyer flask overnight or
       longer at 130°C in a clean oven that is preferably dedicated to this
       sole purpose.  Heated Florisil is stored at 130°C in the oven with the
       flask covered by aluminum foil or glass stoppered.  Three columns
       (Kontes 420530, size 241, 25 mm od x 300 mm length) are packed with
       4 inch beds of activated adsorbent topped with Na2S04 immediately prior
                                       -107-

-------
                                                      Section 4A

 to use,  as  described on page  6 of  the EPA PAM, Section 5,A,(1),(a).
 Alternatively,  the  columns may be  prepacked, activated, and stored with
 aluminum foil covers in the oven,  and withdrawn a few minutes prior to
 use.

 Two standard mixtures containing, a total of 17 organochlorine and
 organophosphate insecticides  are prepared at levels of 20-250 pg/pl;
 5 ml of  each is added to separate  columns and 5 ml of hexane is added
 to the third as a control.  Elution is carried out with 200 ml of 6%
 diethyl  ether in petroleum ether in two 100 ml portions, and similarly
 with 15% and finally 50% ether-petroleum ether.  The six eluates are
 concentrated and injected for analysis by gas chromatography with an
 OV-17/QF-1  column capable of  resolving the mixtures of pesticides in
 the fractions.

 The percentage  recovery for each compound is calculated from the chroma-
 tograms  of  the  eluate increments and the original standard mixture.
 Results  are recorded on the standard form shown as Table 4-1.  The
 Florisil is evaluated on the  basis of the elution pattern and recovery
 of the pesticides of interest.  All chlorinated insecticides should
 be recovered in the range 90-105% with the possible exception of
 aldrin,  for which recoveries  may be low.  Some organophosphates, such
 as carbophenothion, may also  yield low recoveries.  Ethyl ether should
 contain  2%  v/v  ethanol as commercially supplied, or if absolute ether
 is used,  exactly 2% v/v ethanol should be added to obtain the correct
 polarity which  results in the compound elution pattern shown in Table
 4-1.  The effects of the ethanol constituent may be observed in the
 following Figure 4-A, wherein three identical mixtures of seven com-
 pounds were eluted through three separate but identical Florisil
 columns.  Petroleum ether with no ethanol was used in one column,
 petroleum ether with the correct 2% ethanol in the second column, and
 petroleum ether with 4% ethanol in the third column.               ,:

A copy of the elution pattern is enclosed with each shipment of Florisil
 to qualified field laboratories, which should attempt to verify the
 results.  Changes in local conditions, such as packing procedures,
 temperature, and humidity, can affect the amount of adsorbent or the
nature (polarity) of the solvent required for proper elution.  Although
 the method  outlined evaluates Florisil for use with certain pesticides
 in a specific procedure, similar methods can be used to pretest different
 adsorbents  for  any residue analysis.
                               -108-

-------
  §
              co   B-e
                    m
              CO
         in
                             rH
                              a
                             CO

                             H
S
w
                              H
                              EH.
                              W
                                     4J
                                     U
                                     •C8
       O
       m
        C
        o
       • *(H •
        4J
        U
        0)
                                     in
   CO
                                      g
                                      9-S
                                      vO
 (1)
rH


I
A
&
1
u
Q>
PA
0
o
SO
1
O
0
m
o
o
in
i
o
0
•JT
O
O
v
o
§

-------
                                                             Section 4A
   Figure 4-A.
The effects of polarity variation  of  eluting solvent in
Florisil partitioning of  7 pesticides.  Absolute  ethyl
ether mixed with 0, 2, and 4% absolute  ethanol.
Elution Fraction*
Kept. Epoxide
Dieldrin
Endrin
Diazinon
Methyl Parathion
Ethyl Paralhion
Motathton^
No EthanoL
I
100






n

87
100
100
»
16

m

13


100
84

               'Eluting mixtures:
                  Fract. I -6% Et,O in pet. ether
                  Fro«t.ll-lS%  	
                  Frect.lll-50% ••  - •• ••
b.  Florisil Standardization by Laurie Acid Adsorption

    For details, see Section 121.3  of  the FDA Pesticide Analytical Manual.
    Florisil is activated and stored as described in Subsection 4Aa.   As
    an alternative, stoppered containers of activated Florisil may be stored
    in a desiccator at room temperature and the adsorbent may be reheated
    at 130°C (unstoppered) after two days.

    Standardization by weight adjustment based on adsorption of lauric acid
    was originally described by  Mills  (5).   An excess of acid solution
    (400 mg in 20 ml in hexane)  is added to  2.00  g of Florisil in a flask, and the
    amount not adsorbed after shaking  for 15 minutes  is measured by alkali
    titration of an aliquot removed from the flask.   The weight adsorbed is
    used to calculate by proportion equivalent quantities of Florisil batches
    having different adsorptive  capacities:
          Equivalent quantity
          of Florisil batch
          required per column
          Lauric Acid Value
                           110
                   Lauric Acid Value of
                   batch
                                           20  g
                   mg lauric acid/g Florisil
                   200 - (ml required for titration
                         x mg lauric acid/ml 0.05N NaOH)
   This gross method gives no real indication of the elution pattern to be
   obtained  from a  column containing the standardized Florisil.
                                    -110-

-------
                                                           Section 4A

    To verify the value obtained by the lauric acid method and to test for
    proper elution of organochlorine and phosphate insecticides, 1 ml of a
    standard mixture containing 1-15 pg of eight compounds is applied to a
    22 mm id column containing 4 inches of Florisil (or the weight deter-
    mined by the lauric acid method) and eluted with 200 ml portions of 6,
    15, and 50% diethyl ether in petroleum ether.  The three fractions are
    concentrated prior to gas chromatography on an appropriate column.
    Heptachlor, heptachlor epoxide, ethion, and carbophenothion should
    elute with good recoveries in the 6% fraction; parathlon, dieldrin, and
    endrin in the 15% fraction; and malathion in the 50% fraction.  This
    mixture is recommended for routine testing since it contains pesticides
    that give indication of improper elution, poor Florisil, and impure
    reagents.

c.  Deactivated Florisil

    Water-deactivated Florisil is required for the Osadchuck et al. multi-
    residue screening procedure for foods (Subsection 9M).   Preparation
    and standardization is carried out as follows for this method (6):

    (1)  Deactivation

         Heat 1-2 kg of Florisil in a one gallon jar at 300°C for 8 hours and
         cool overnight.  Add 2% (w/w) distilled water and place a screw cap
         lined with aluminum foil on the bottle.  Place the jar in a rotary
         mixer, tumble for 1 hour, and allow Florisil to stand for 24 hours
         after mixing.

    (2)  Standardization

         A mixture of dieldrin, malathion, and azinphosmethyl.  is added to a
         2.5 cm id tube filled with 15 cm of deactivated adsorbent.  The
         column is eluted successively with 300 ml portions of 30% methylehe
         chloride in hexane, 10% ethyl acetate in hexane, and 30% ethyl
         acetate in hexane.  Dieldrin should elute in the first fraction,
         malathion in the second, and azinphosmethyl in the third, with all
         recoveries greater than 90%.  Late elution, especially of malathion,
         which just barely elutes with 10% ethyl acetate, indicates insufficient
         deactivation and the need for more polar solvents.  Early elution
         indicates over-deactivation, requiring less polar solvents for
         chromatography (i.e., lower percentage of methylene chloride or
         ethyl acetate).

         Comparable standardization is carried out for other methods employing
         deactivated Florisil.

d.  Deactivated Silica Gel and Alumina

    Silica gel deactivated with various percentages of water has been
    successfully used for cleanup and fractionation in many residue deter-
    minations.  Preparation of 20% deactivated adsorbent on a small scale
    has been conveniently and successfully carried out as follows (7):
                                     -111-

-------
                                                           Section 4A
    (1)  Activate Woelm silica gel 48 hours at 170-175°C.
    (2)  Add 2 ml of water to 10 g of adsorbent in a tightly capped Teflon-
         lined screw top vial.
                                         tnjr
    (3)  Mix on a rotary mixer (Roto-Rack  ) for 2 hours at setting 8.

         Silica gel prepared in this manner can be stored in the capped vial
         for at least one week with no change in adsorptivity.  Standardiza-
         tion is carried out, as above, by packing the required column,
         adding an aliquot of standardization solution containing the pesti-
         cides of interest at a level providing adequate detector response,
         eluting with appropriate solvents, and examining fractions of
         eluate by gas chromatography.

         Alumina deactivated with water is used in conjunction with silica
         gel in the Eolden and Marsden cleanup procedure (Subsection 90}
         and its various modifications (8).  This may be prepared in a
         similar manner by addition of the required percentage of water to
         alumina previously activated at 800°C for 4 hours.

e.  Celite 545

    Electron capturing"impurities are removed from Celite 545 as follows:
    Slurry with 6M HCl while heating on a steam bath, wash with water until
    neutral, wash with several solvents ranging from high to low polarity,
    and dry.  Impurities interfering with phosphorus-selective detectors
    are removed by heating Celite at 600°C in a muffle furnace for a minimum
    of 4 hours (FDA PAM, Section 121).

f.  Charcoal

    Charcoal adsorbent is purified as follows:  Slurry 200 g with 500 ml of
    concentrated HCl, and stir magnetically while boiling for 1 hour.  Add
    500 ml of water, stir, and boil another 30 minutes.  Recover the char-
    coal by filtering through a Buchner funnel, wash with water until
    washings are neutral, and dry at 130°G.  (FDA PAM, Section 121).  As an
    alternative procedure (9), add 225 g of charcoal to 1.2 liters of
    ethanol-conc. HCl-water (50:10:40) and reflux for 1 hour.  Collect the
    charcoal on a Buchner funnel and wash with distilled water until pH test
    paper shows only a trace of acid to be present.  Further wash the char-
    coal with acetone and aspirate until nearly dry.  Air dry until odorless
    (2-3 days) and finally dry in a porcelain dish at 130°C for 48 hours.  .,
    Store in a tightly stoppered bottle.

g.  Magnesium Oxide (Sea Sorb 43)

    Slurry 500 g with enough distilled water -to cover it in a 1 liter
    Erlenmeyer flask, heat with occasional shaking for 30 minutes on a~
    steam bath, and filter with suction.  Dry for 12-24 hours at 105-130°C
    and pulverize to pass a No. 60 sieve.  About 10% water is adsorbed in
    this procedure.  Store in a closed jar (FDA PAM, Section 121; 9).
                                      -112-

-------
                                                             Section 4B
  h.  Packing and Elution of Adsorbent Columns
      Pack the adsorbent in glass chromatographic columns containing a loose
      plug of glass wool (coarse porosity fritted glass discs as support are
      not recommended because of the difficulty of keeping them clean).
      Columns 300 mm x 22 mm id with or without a Teflon stopcock (e.g.,
      Kontes 420530, size 241, or equivalent) have been, widely used for larger
      scale cleanup, and 7 mm id columns (e.g., Kontes size 22 Chromaflex
      columns, or equivalent) for small scale chromatography.  Add the re-
      quired amount of dry column packing ,in increments and gently tap to
      settle after each addition; then add a layer of granular sodium sulfate
      (ca. 0.5 inches) on top of the adsorbent.  Prewash the column with
      hexane or petroleum ether, bring the level of liquid to the top of the
      bed, add the sample, and wash it into the bed with several small portions
      of the first eluant.  Collect the various fractions in separate containers.

      Carry out the elution with a series of solvents and solvent mixtures of
      increasing polarity.  Select the polarity of the solvent series con-
      sistent with the activity (polarity) of the adsorbent and the polarity
      of the sample.  Use the least polar solvent that will elute the pesti-
      cides from the adsorbent to minimize co-elution of polar impurities.

      The order of polarity for several common solvents is .as follows:

           hexane (petroleum ether) - least polar
           benzene
           ethyl ether
           methylene chloride
           ethyl acetate
           acetonitrile
           methanol - most polar


                          GAS CHROMATOGRAPHY PACKINGS

4B    INTRODUCTION AND COLUMN TECHNOLOGY

  It  is appropriate to reiterate that the column is the "heart of the gas
  chromatograph."  Even though all other modular components of the instrument
  may be  functioning perfectly, a bad column will cause the entire gas chro-
  matographic output to be correspondingly bad.  In this subsection, a number
  of  practical operational problems will be discussed; many of these problems
  have come  to light in the  interlaboratory quality control program described
  earlier in Section 2.   Some of the operational instructions, fully covered
  in  Sections 4,A- 4,D of the EPA PAM, will be briefly reviewed in this sub-
  section, but only as they  relate strongly to the success or failure of the
  gas chromatographic performance.

  A column for gas  liquid chromatography  consists of a tube filled with a
  powdered support  on which  is uniformly  coated a liquid, stationary phase.
  When a  mixture  of compounds is injected into the gas chromatograph, each
  compound is swept through  the  column at a rate that is determined by the
                                        -113-

-------
                                                            Section 4B

 interaction of the compound with the stationary phase under the given
 operating parameters such as temperature and flow rate of carrier gas.  If
 the phase and the parameters are properly chosen, the different compounds
 will migrate through the column at different rates, and separation will be
 achieved as diagrammed in Figure 4-B (10).


             Figure 4-B.   Schematic diagram for elution analysis.
             SAMPLE (A+8)-
                              'COLUMN
                                            DETECTOR
                                                 , CHROfvWOGRAM
                                             3-D
                                             ra
                                              HZS
Most applications of residue  analysis are  carried  out with packed  columns
of this type rather than wall coated or  support  coated  capillary (open
tubular) columns, although  the latter are  being  used now with  greater
frequency  (see Subsection 4M).  GC tubes are usually made of borosilicate
glass.  Copper and stainless  steel are best avoided because both can
cause decomposition of compounds unless  special  precautions are taken.
Commercial solid support materials are usually composed of flux-calcined
diatomaceous earth that may be treated by acid- or base-washing or silaniza-
tion.  Firebrick, glass beads, and Teflon are other support possibilities.
A good support material should be available in narrow and uniform ranges of
particle (mesh") size and have a minimum of active adsorption sites for inter-
action with injected compounds passing through the column, high surface area
per unit volume, good thermal stability, and mechanical strength.  Although
greatly improved in recent years, various supports and different lots of the
same support are not necessarily equal in surface area or inertness.  Ad-
sorption or degradation of a pesticide on the support can affect the relative
retention time and response of the compound.  It is important to select the
most inert solid support possible for pesticide analysis, with additional
special treatment being desirable for columns used to determine certain
sensitive compounds (Sections 4F and 41).  Chromosorfa W is among the least
active diatomaceous earth supports commercially available.  As a general
rule, column efficiency increases as the particle size of the support de-
creases, but a greater carrier gas pressure is required to maintain a given
                                     -114-

-------
                                                             Section 4C


   flow through columns  with smaller mesh sizes.   For most pesticide work,
   supports  with mesh sizes of 80-100 or 100-120  will be satisfactory.   The
   presence  of very fine particles,  those above the upper limit of each
   individual mesh range, may cause  column inefficiency.  If it is likely that
   particles have been broken during shipment or  use, thus increasing  active
   sites and exposing untreated surfaces, check to determine whether the
   mesh size of the solid support is completely within the expected range.

   There are a great number and types of liquid phases commercially available.
   The choice of liquid  phase is usually made on  the basis of the polarity of
   the compounds to be separated. Phases recommended for general use  in pesti-
   cide analysis are described in Section 5L.  Recently, liquid phases have
   been marketed that are purportedly "equivalent" to previously available
   phases but with greater thermal stability.  It is important to determine
   whether they provide  the same relative retention times.

   Important column considerations include efficiency and resolution capa-
 ,  bility, sensitivity (in relation to the detector), retention, compound
   elution pattern, stability to heat and injection loading, and freedom from
   on-column compound decomposition.  These will  be discussed in light of their
   effect on day-to-day  operation of the column -


4C     COLUMN EFFICIENCY AND PEAK RESOLUTION

   Figure 4-C shows the  equations used for calculating column efficiency (in
   theoretical plates) and the resolution (R), or degree of separation between
   peaks, from a chromatogram.  A numerical value for efficiency, in itself,
   is of little practical import.  However, efficiency is generally synonymous
   with peak resolution, and^ this is of considerable importance to the chroma-
   tographer.  Figure 4-D, for example, shows superimposed chromatograms of
   standard chlorinated pesticide mixtures on two separate 6-foot columns of
   2% OV-1/3%QF-1, one  (A) with very poor efficiency (740 total plates) and the
   other  (B)~"with high efficiency (4,530 plates).  It will be observed that'on
   column JB, all seven peaks give baseline separation, whereas on the low  ,'.
   efficiency column A,  poor separation is evident for four of the peaks.

   A column efficiency value of 500 theoretical plates per foot for £.,£.' -DDT
   is considered to be of minimal acceptability in terms of the generally
   expected peak resolution.  A 6-foot column of 3,000 plates will usually
   provide acceptable resolution of mixtures encountered in residue analyses.
   Since  the absolute retention  time of the peak used for measurement has an
   effect on the calculated N, it is necessary to choose a standard peak such
   as £,£_'-DDT for comparison of column efficiency.  Column efficiency as
   measured by this equation is affected by noncolumn factors such as dead-
   volume in the instrument construction or by any gas  leaks.
                                         -115-

-------
                                                         Section 4C
  Figure 4-C.   Calculation of column' efficiency  and resolution
              Efficiency: N = 16(y)1
              Resolution: R =
Figure 4-D.   Effect of column  efficiency on pesticide  resolution
                                  -116-

-------
                                                           Section 4C

Column factors that influence efficiency are the particle size of the
support (small particles lead to higher efficiency), uniform coating, care
in handling and packing the coated support, column diameter and length
(longer columns provide more total plates), and operating parameters such
as temperature and flow rate, particularly the latter.  These parameters
must be optimized in relation to the liquid phase loading and the analysis
time.  In general, lower temperatures and flow rates and low liquid phase
loading beneficially affect efficiency.  Figure 4-E illustrates the advantage
of low loading (column A) by comparison of resolution and elution time for
two columns of nearly equal polarity operated at similar temperatures.  A
pitfall of low-loaded columns, however, is easier degradation and/or adsorp-
tion of certain susceptible pesticides, affecting both the retention time
and the apparent response of these compounds.  The minimum coating that can
be used is limited to the amount for complete coverage of the support,
usually. 1-3%, and also by the reduced capacity for sample components.
   Figure 4-E.  Effect of stationary phase loading on column efficiency.

            A:  Temp. 187 C,  70 ml/minute, effic. 4550 theor. plates.

            B:  Temp. 190°C, 100 ml/minute, effic. 2600 theor. plates.
                                                       s*/.. De-20p/7.5 v.Qp-1
                                                                I
                                                                 [
                                                                ?
                                    -117-

-------
                                                              Section 4D
4D
SENSITIVITY AND RETENTION
   The same principal factors influence the sensitivity and retention of the
   column:  type and loading of the liquid phase, carrier gas flow rate,
   column temperature, column length, and particle size of the support.
   These column parameters influence the sensitivity in that any change in-
   creasing the peak height for injection of a given amount of pesticide will
   thereby increase detector response.  The columns recommended in this Manual
   (Subsection 5L)  are designed for adequate resolution consistent with
   practical elution times, and an absolute retention -of 16-20 minutes for
   2.»JB.'-M>T has been found to approximate these characteristics for a column.
   This retention range can be obtained by operation of lower load columns
   (3-6%)  under such conditions that will produce maximum efficiency.   Higher
   load columns must be operated at elevated temperature and flow rate, and
   therefore decreased efficiency,  to obtain this elution time.   Relative
   retention times  are affected only by the nature of the liquid phase and the
   column  temperature.   That is, at a constant temperature,  the percentage
   loading of a particular liquid phase can be varied without changing the
   relative retention of two or more pesticides.

   The following "bar graph, Figure 4-F, provides comparative  sensitivity data
   on  eight GC columns  using the 3% OV-1 column as unity for reference purposes,
   Each column included in the study was operated at its optimum parameters in
   terms of the achievement of maximum response,  efficiency,  and a practical
   retention time.
e
a '
CJ
0
-i
Of
1"
o
-4-rT
— =*
— . — U-i
HP
t r
1 '* > '
i : T
it
t --
; =:^
: —
rig*
iff
_r.
ra 4*£r
On
•
to
CO
opi
:rrfcr
. . .4. ._
—r|"
•p«rat
SLC CO
tte r
3* OV
L« o.
ElM.
^
Lw r<
ZtlMU
•/•r*i
-1. mi
pArmm*
— ^~
rrrt:::
* • * * * i
_4-- '
— 4 "H"
rapons
a* r«
«« eo
th «•
d «t
jj'l1'
::r:
:_IL
• • 't
I'^I"
• of
:»t«d
ch
Its
-
— 4-
T"
1—
-*-r-,
rT
TT4
201
_i.
-».
H-
j.
j No. LMdia
i 4.
- 1.9/1.
r 3/7.J
:SE'3bfciri:
r"-: ' : :£
•• i — '
B« 1
IS
)V-"l7
- - - 4 -'- .
Optima I
C
tap. t°C.
1*0
1*0
200
3OO
200
300
<2F-i::?^-
1 DC*2C
«rrl«r Flow
45
60
TO
6O
TO - BO
TO - »O
60
120
i i • ' 	 	
| 'T- • * ••

                                     -118-

-------
                                                             Sections 4E, 4?
4E
   COLUMN STABILITY
  It is desirable to use columns that are heat-stable or "bleed" resistant
  and  that continue to function properly under  injection loading with dirty
  extract.  Liquid phase bleed is evident from  a persistently  drifting base-
  line and the  inability to  obtain  a normal  level of standing  current (Sub-
  section 5C) from an electron capture  detector.  Minimum baseline  noise
  and  drift are achieved with a relatively lightly  loaded column containing
  a stable liquid phase of low volatility.

  When a succession of "dirty" extracts are  passed  through  the system, the
  column performance is usually affected.  The  most prevalent  symptoms of
  injection overloading are  depressed peak height response, lowered efficiency
  and  resolution, on-column  breakdown of pesticides, erratic recoveries, and
  unsymmetrical peaks  (see Figure  4-J). Columns with  low liquid phase load-
  ing  are more  susceptible to  injection overloading.


 4F     RESISTANCE TO ON-COLUMN  COMPOUND  DECOMPOSITION

  Uniess  a  column is properly  prepared, conditioned, and maintained, it  can
   cause  such compounds  as endrin and/or _p_,£'-DDT to undergo some degree  of
   decomposition.  The main symptom of endrin decomposition is  a greatly re-
   duced  endrin peak with the formation of one or two additional peaks arising
   from decomposition  products.   p_,_p_'-DDT decomposes to £,j3_'-DDD and, in extrei
          to _p_,£f-DDE.
                                                                      extreme
cases,
   Newly packed columns should be specially treated with a silanizing agent such
   as Silyl 8 to reduce the number of active adsorption sites that can cause
   decomposition of endrin.  The beneficial effect in improving response and
   minimizing conversion of endrin to breakdown products is illustrated in
   Figure 4-G.  Chromatogram A was obtained for an aldrin-endrin mixture
   immediately after heat conditioning and equilibrating a column of 1.5%
   OV-17/1.95% QF-1.  It exhibits a small endrin peak and two breakdown peaks.
   (In principle, endrin could be quantitated using.the sum of these three peaks;
   however, the final breakdown peak elutes very slowly and would cause the
   analyst to waste considerable time.)  After treatment with Silyl 8, the same
   amount of the same mixture was injected, and Chromatogram B shows significant
   improvement in the endrin response and complete disappearance of the two
   breakdown peaks.
                                         -119-

-------
                                                        Section 4F

Figure 4-G.   Reduction in breakdown of endrin  resulting from
              column silanization
                            BEFORE SILYLATION
            2
                               AFTER SILYLATIQN
                               2b
                               -120-

-------
                                                          Section 4F

Sllanization does not always provide such dramatic results.  Cases have
been noted when no endrin response whatever, either in the form of a main
peak or breakdown peaks, was obtained, and silanization did not improve
the situation.  On the average, however, silanization clearly improves the
gas chromatographic behavior of endrin.

DDT breakdown is manifested by the appearance of £,£f-DDD and/or £,£f-DDE
on the chromatogram resulting from the injection of pure analytical grade
£,j3'-DDT that is known to be free of these metabolites as impurities.
This problem is associated with overloading of the column packing adjacent
to the front glass wool plug, the plug itself, or the glass insert if off-
column injection' is used, with contaminants from dirty extracts.  Figure
4-H illustrates the DDT breakdown phenomenon.  Chromatogram A is an aldrin-
DDT mixture on an SE-30/QF-1 column with no decomposition, while B shows
another column containing the same phase (operated with somewhat different
parameters) that caused a total of 25% decomposition of the DDT peak to its
two metabolites.  This chromatogram was obtained in a laboratory where the
injection insert had not been changed for three weeks.


     Figure 4-H.  Breakdown of £_,£_'-DDT on 4% SE-30/6% QF-1 column
                 A - No Conversion
                                             B - Maximum Conversion
 Figure 4-1 is a similar illustration of deterioration of column performance
 with age or with heavy use for dirty samples.   The older column (B)  is pro-
 moting degradation of DDT to DDD (peak 5 to peak 4),  and retention times
 have lengthened.  These chromatograms point out the importance of frequent
 analysis of GC standards that are representative of those compounds  that are
 most frequently analyzed.
                                    -121-

-------
                                                            Section 4F


  Figure 4-1.   Electron capture gas chromatograms of DDT and metabolites on
               a 4Z SE-30/6% QF-1 column, 180 cm x 4 mm id,  at 180°C.   (A)  New
               column,  (B) column after 2 months use for "dirty" samples.
               Compounds:   (1) £,£'-DDE; (2)  o,£>DDD;  (3) o,p«-DDT;  (4)
               £,£'-DDD; and (5)  2,2'-DDT.                 ~
                                                B
                         12   15
                                                                12    15
18
                                  TIME (mini
 Figure 4-J Illustrates an extreme case of overloading of a column of 2Z
 OV-1/3Z QF-1.   Chromatogram A Is from a standard mixture of seven pesti-
 cides on a freshly prepared column.  The column was then disconnected from
 the detector so the exit end vented Inside the oven.  Eighteen consecutive
 injections were then made of fatty tissue extract after elution with 15%
 diethyl ether-petroleum ether through a Florisil column, each injection
 containing the equivalent of 25 mg of fat.   After 30 minutes the column was
 reconnected to the detector, the system was equilibrated,  and an identical
 volume of the  same standard mixture was injected.   Chromatogram B shows
 the results of column overloading:  depressed peak heights,  peak tailing,
 peak broadening, and conversion of  £,£'-DDT to 2.,£>DDD (in actuality,  the
 ratio of these changed from 8:10 to 4:10).   A clean Vycor glass insert  was
 then installed in  the injection port,  the system was  re-equilibrated for
 30 minutes,  and another equal volume of standard mixture was  injected.
 Chromatogram C shows  the dramatic recovery  of the  system after  this single
 step.  Finally, Chromatogram D  indicates  a  complete rejuvenation of the
 system when  the same mixture was  injected after overnight purging at normal
 operating temperature and carrier flow  parameters.

This series of chromatograms is striking  evidence that damaged columns can
often be salvaged by changing the injection Insert, forward glass wool plug,
and perhaps the first one-half or one inch of column packing.  More Import-
antly, properly maintained and monitored  columns should provide Cop perform-
ance without problems for many thousands of injections.
                                        -122-

-------
                                                  Section 4F
Figure 4-J.  Chromatograms illustrating column overloading

             and subsequent rejuvenation
           i
'0.
       g
       a

       •I
                            -123-

-------
46
                                                              Section 4G
HOMEMADE VS. PRECOATED PACKINGS
    The decision whether to make column packings or to buy them precoated con-
    fronts every laboratory conducting GC analyses.  Prior to 1969 the answer
    to this question was easy.   The precoated supports available from commercial
    suppliers were so poor in quality that it was necessary to hand-coat  packings
    to obtain satisfactory materials.   Since that time,  however,  several  commercial
    firms have developed the capability to produce high quality packings.  Not-
    withstanding,  anyone purchasing this material should do so on specification.
    As broad guidelines, the following quality criteria are presented:

    a.   Must meet  a column efficiency  of a minimum of 3,000 theoretical plates
        for a column of 183 cm (6  ft)  x 4 mm (5/32 in.), computation  being made
        on the basis of a peak for _p_,_p_'-DDT.

    b.   A specific pattern of compound elution and peak  separation.

    c.   An absolute retention time range for  the elution of _p_,£f-DDT  using
        specified  parameters  of column temperature and carrier gas velocity.

    d.   No appreciable  decomposition peaks  to  result  from the. injection of pure
        standard endrin or 2.>Z'-DDT

    e.   Final acceptance of each lot purchased  to  be  based  on  buyer's evaluation
        at time of delivery.


   The  final decision on whether  to purchase or prepare column packing may de-
   pend on the situation in a given laboratory.  The successful formulation of
   column packing  in small batches requires a degree of expertise somewhat
   beyond the purely scientific.  The procedure has been described as 50%
   science and 50% art.  If some particular individual on a laboratory staff
   has developed the expertise to produce good column packing in small lots, it
   may prove advisable to prepare the material on an in-house basis.   This is
   somewhat cheaper and far more convenient in terms of immediate availability.
   On the other hand, if no individual on the staff has this "knack"  and  the
   laboratory has no appropriate equipment for the task, it may prove advisable
   to rely on a commercial supplier.

   There are a number of methods available for the preparation of column
   packing.  The simplest probably is the "beaker technique" wherein  the  liquid
   phase or phase mixture is dissolved in an appropriate solvent in a beaker
   the support is added, and the mixture is stirred while evaporating the
   solvent under a stream of air or nitrogen.  The strong disadvantage is that
   the constant hand stirring tends  to fracture the support particles.

   An extension of the beaker technique is known as the ^'filtration technique."
   The slurry in the beaker comprised  of liquid phase,  support,  and solvent  is
   removed by drawing air through the  layer of packing on the filter  paper by
   means of a side arm flask connected to a vacuum source.
                                      -124-

-------
                                                         Section 4G


The "fluidization technique" is a more sophisticated extension of the beaker
technique.  The slurry in the beaker is transferred to a fluidizer cylinder
(Applied Science Laboratories, Catalog Number 13994) so constructed that a
high volume of nitrogen can be blovn up through the packing from the bottom
of the cylinder, while heat is applied by an element at the base of the
cylinder.

In the "rotary vacuum technique" the liquid phase or mixture is dissolved
in an appropriate solvent in a small beaker and transferred to a Morton
flask (Kontes No. K-295900) with indented sides.  The support is added and
the flask is placed in a variable heat water bath and connected to a rotary
evaporator  (Rinco).  Mixing and solvent evaporation are carried out by
rotating  the flask under vacuum with applied heat.

Because no  preparation technique is presented in the EPA PAM, one method is
offered below for the benefit of laboratories that may like to prepare their
own packing.  While other methods may be equally satisfactory, the rotary
vacuum method as detailed here has proved very satisfactory for the pro-
duction of  small batches of GC column packing.  The batch size described
will provide enough packing to fill three 183 cm x 4 mm columns.

a.  Based on a  21 g total batch  size, compute the amount of liquid phase(s)
   • to weigh in 30 ml beaker(s)  on an analytical balance.
 b.
 c.
 d.
Weigh out liquid phase to two-place accuracy.  If making mixed-phase
packing, weigh each liquid phase in a separate beaker.

With a 25 ml graduated cylinder, transfer 15 ml of the appropriate
solvent into each beaker.  Stir with a 3 inch glass rod until the liquid
phase is completely dissolved.

through a glass funnel, transfer each liquid phase solution into one
300 ml Morton flask.  Note;  From this point on, all solvent used for
rinsing beaker(s) and funnel(s) will be measured so that the final
solvent volume in the flask will be just sufficient to produce a slurry
of about heavy cream consistency when the support is added.  This is a
somewhat critical point because too little solvent does not permit
adequate mixing for uniform support coating, and too much solvent
involves an excessive evaporation time for the solvent.  A 10 ml Mohr
pipet works nicely for adding  and measuring  the applied solvent.  The
beaker(s)  should be rinsed with four consecutive applications of 7-9 ml
of  solvent, the exact amount depending on the appropriate solvent/support
ratio.

After  the  liquid phase  transfer into the flask, place a powder funnel in
.the flask  and add  the  support. Note;  The amount of  support to weigh
 out for a  21 g batch  is  the difference, in grams, between the total
 amount of  liquid phase weighed and 21  g.  For example, with a 21 g  batch
 of packing of 4%  SE-30/6% QF-1:
                                      -125-

-------
                                                               Section 4H
               SE-30	0.040 x 21.0 - 0.84 g

               QF-1  	0.060 x 21.0 =» 1.3  g
                  Total liquid phase:              "JTlg

                  21.0 - 2.1 - 18.9 g of support
     f.  Attach the flask to a rotary  (Rinco) evaporator.
     g-
     h.
     Mix slowly for 10 minutes at room temperature with just enough vacuum
     applied to hold the flask on the evaporator.

     Advance the hot plate control sufficiently to increase the temperature
     of water in the beaker to 45°C in ca. 20 minutes.  Increase the vacuum
     slightly at the start of heating and continue increasing, a little at
     a time.  Notes;  (a) By the time the temperature reaches 45°C, the
     vacuum should be such that the slurry is at a'near-boil.  This con-
     dition should be maintained throughout, until all visible solvent is
     removed,   (b)  After the 10 minute initial mixing period, the flask is
     rotated very slowly.  This is a very critical point.   It is generally
     not possible to slow the power stat or Variac sufficiently to com-
     pletely accomplish this, and it is necessary to brake further by hand.
     This requires continuous attention by the operator throughout, really
     a small time investment in light of the importance of good column
     packing and the length of time good columns should give service.

 i.   Advance heat gradually to 55°C,  applying as much vacuum as possible
     just short of flushing liquid solvent out of the flask.   Remove all
     visible solvent at  this temperature.

 j.   Advance heat to produce 65°C,  applying all vacuum available and rotating
     very slowly and intermittently.
    fc.
    When all evident solvent is  removed,  release  the vacuum carefully and
    shut down the  assembly.   Transfer  the flask of packing to an oven and
    hold at  130°C  at least 2 hours,  or overnight.
4H
Alternative pan  coating and  filtration coating procedures are described in
the FDA PAM, Section 301.5.

Once a column is prepared, the actual weight percent loading can be de-
termined, if required, by exhaustive Soxhlet extraction in glass thimbles
or standard low  temperature  or thermal ashing procedures.


   PACKING THE COLUMN

Columns for pesticide analysis are generally 4-7 feet (120-210 cm)  in length
and 1/8 or 1/4 inch (0.32 or 0.64 cm) od metal or glass.   Aluminum columns
have been found  suitable for chlorinated pesticides, but  glass is usually
                                        -126-

-------
                                                          Section 4H

preferred to prevent degradation often associated with metal columns.   U-
shaped, 6-foot, glass columns are used in the Tracer MT-200 gas chromato-
graph that is standard throughout the EPA network of laboratories (Section
5).  These are cleaned before packing by scrubbing with soap and water and
a pipe cleaner, rinsing with water and acetone or anhydrous methanol,  and
drawing vacuum to dry.  Glass columns' should be silanized prior to packing
for chromatography of especially labile compounds.  See also Subsection
5J for information on silane treatment of glass injection inlets.

There are several methods for packing a column, e.g., hand vibration,
mechanical vibration, and vacuum.  The method of choice may be dictated
by the configuration of the column.  Thus, vacuum is about the only method
for packing a coiled column.  A U-shaped column may be packed by any of the
three methods.  In general, the aim is to pack the coated support tightly
to increase efficiency, with the least amount of particle breakage possible
to decrease adsorption/degradation problems.  The recommended method is
hand vibrating, which has produced columns of consistently high quality.

a.  The operator should be sure that the column, if intended as a 6-foot
    column, is really 6 feet in total length, and not some lesser length.
    Efficiency and retention time are both reduced in a shorter column.
    For off-column injection in some chromatographs such as the MT-220,
    the inlet end of  the column should be 1  inch shorter than for on-
    column injection.

b.  On each column leg place a mark at a point on  the glass that will be
    just visible at  the Swagelok nut when the column is installed in the
    oven.

c.  Through a glass  funnel  attached to the  column, pour ca. 6  inches of
    packing into each leg.

d.  Repeatedly tap  the U-bend  of  the  column on the floor  for ca.  30 seconds.
    Hote;   The glass is fragile and it is,  therefore,  advisable to  place
     some type of padding  such  as  a magazine on the floor.

 e. Repeat this operation,  adding ca.  6  inches at a time  to  each column
     leg.   It  is advisable to vibrate  additionally with a  wooden pencil,
     running it up  and down the length of the packing.

 f.  Continue adding packing and vibrating until  the pencil marks are reached
     and the packing will not vibrate below the marks.   This  should  be  done
     with great care, tapping the column a sufficient length  of time to be
     certain that no further settling is  possible by manual vibration.  The
     use of mechanical vibration is not advised as the packing  may be packed
     too densely, thus introducing the possibility of excessive pressure
     drop when carrier gas is applied.

 g.  Place plugs of ca. 1 inch length of silanized glass wool in each  end,
     just tightly enough to prevent dislodging when carrier gas is applied
     but not so tight as to impede gas sweep through the column.  If glass
                                     -127-

-------
                                                              Section 41

         wool is packed by hand, the hands should be carefully washed with soap
         or detergent, rinsed, and dried to minimize skin oil contamination of
         J?6 Sr?!/??1*  ?J-SSS W°01 Can be s±lanize
-------
                                                         Section 41
                Figure 4-K.  Column conditioning (11).
The following schedule of heat conditioning is recommended for some EPA
(Subsection 5L) and FDA prescribed GC columns:
"' • 1 '
Phase
••MMauBMMW
4% SE-30/6% OV-210
1.5% OV-17/1.95% QF-1
3% DECS -
10% OV-210
10% DC-200 -
10% DC-200/15% QF-1 (1:1)
15% QF-1/ 5% DC-710 (2:1)
o I/
Oven Temp . , C —
245
245
235
245
250
250
240

Minimum Time, hour
72
48
20 &
48
16
72-120
120
— Carrier  gas  flow 60  to  70 ml/minute.
—• Shown for  information only.   Column not  recommended  for routine  use.
3/
— Do  not exceed  this time period.
—• DC-200 columns are significantly improved if conditioning  is  carried
   out without  carrier  gas (12).
                                         129

-------
                                                           Section 41
 In general, it is desirable to heat cure the column at a temperature
 ca. 20° below its maximum useable temperature with a. normal flow of oxygen-
 free carrier gas in a leak-tight GC system.  An alternative, more gradual
 approach considered preferable by some laboratories is to use a temperature
 program starting at 50°C for 30 minutes and increasing at about 5°C/minute
 up to the desired maximum.  Details for the connection of the inlet column
 leg (which is 1 inch shorter for off-column injection) to the inlet port  of
 a MT-220 chromatograph through a special Swagelok attachment are given in
 the EPA PAM Section 4,A,(2),IV,1.  The column exit is vented inside the
 oven and not connected to the detector.  The outlet ports leading to the
 transfer line are sealed off with Swagelok nuts to prevent traces of
 column effluent from seeping through to the detector.  Particular caution
 is needed when preparing mixed columns with different, but supposedly
.equivalent, liquid phases.  Use of one or more of the newer, stabilized
 liquids (e.g., 0V silicones, SP products, silars, etc.)  may give a column
 with an altered phase ratio after conditioning because of increased
 temperature stability.  These more stable columns still require condition-
 ing before use, but shorter times will be.necessary.   To determine the
 proper time,  the column should be cooled and connected to the detector after

a reasonable  conditioning period (e.g.,  2-3 hours)  and the  baseline should
be checked at the sensitivity to be used for the  analysis,  or slightly
higher.   If necessary,  conditioning is repeated until stability  is  satis-
factory.   Capillary columns,  which are also made  by evaporation  of  a
solution of a partition liquid,  should be conditioned the same way  as a
packed partition column.   The maximum  temperature may be lower than for
the same liquid in a packed column,  however,  due  to the  weaker attraction
of the liquid to the column wall.                           •

As mentioned  previously,  column  efficiency  and  response, especially  the
response of endrin,  would slowly improve as new columns  become "seasoned"
with use,  but silanization is  a  means  of rapidly  conditioning  the column
to full endrin response.   After  heat curing and with  the column still
isolated  from the detector, the  oven temperature  and  carrier gas flow rate
are adjusted  to  the approximate  recommended operating  conditions for the
column of  interest (Subsection 5L).  Four consecutive  injections of 25 yl
each of  Silyl 8  (Pierce Chemical Co.)  are made, spaced 30 minutes apart.
Following  the final injection, about 3 hours  is allowed for all traces of
the silanizing material to  elute from  the column.  The syringe used for
these  injections  should be  used  for no other purpose and should be rinsed
immediately after  use to avoid damage.  The  effects of silanization do not
persist indefinitely, and repeat  treatment about once a month is recommended.
Silanization  is particularly useful when low loads of liquid phases are
used.  Columns to be used with flame photometric or thermionic detectors
for detection of organophosphorus pesticides should not be silanized but
rather Carbowax-treated.

Generally, Carbowax-treated columns are much more responsive and capable of
higher peak resolutions for organophosphate pesticides than columns that
are untreated.  Depending on the specific compound and column, increases
in response have ranged from 10 to 200%, with a 100% increase, of doubled
response, being most likely.  Silyl 8 conditioning has no beneficial effect
                                    -130-

-------
                                                             Section 4J
4J
on organophosphate response, and silylated columns should definitely not
be used with the flame photometric detector since bleed will cause excessive
fogging of the heat shield.  Details of the treatment and a special Swagelok
assembly used in the MT-220 chromatograph are given in Section 4,B,(2),IV of
the EPA PAM.  This is a modification of the method reported by Ives and
Giuffrida to bleed Carbowax from a 2 inch 10% precolumn heated in the
chromatograph oven at 230-235°C for 17 hours with a carrier gas flow of
20 ml/minute. (See also Section 4J concerning a different type of Carbowax
column treatment.)

The response characteristics of the column should be monitored with a
standard mixture of organophosphorus pesticides immediately after treatment
to serve as a reference point for later checks on the longevity of the
beneficial effects.  Response will sometimes drop rapidly for several days
after treatment and then stabilize, usually at a level well above that for
the untreated column.  Carbowax-treated glass wool may also be less
adsorptive than the silanized wool usually used.

Following the silanization or Carbowax treatment and with the oven tempera-
ture and carrier gas flow rate adjusted to the approximate operating levels
for the particular column, several successive injections of a pesticide
priming mixture in the microgram range are made onto the column with enough
time between injections for all compounds to elute.  Injection of priming
standards each morning will help assure consistent peak response for
working standards throughout the day.  With some easily degraded compounds
such as underivatized monocrotophos, the column is primed before every
analysis.  Other difficult pesticides that may not chromatograph well
unless the column is aged and primed include perthane, methoxychlor,
dicofol, tetradifon, chlorobenzilate, Prolan, captan, esters of 2,4-D,
malathion, azinphosmethyl, coumaphos, and PGP.


   SUPPORT BONDED CARBOWAX 20M COLUMNS

Section 4,A,(7) of the EPA PAM describes the preparation of highly inert
GC columns by chemically bonding Carbowax 20M to diatomaceous earth GC
support.  The Carbowax is coated (using a 5% solution) on acid washed
Chromosorb W, 80-100 mesh, and after heat conditioning at 270-280°C, the
nonbonded phase is removed by solvent extraction.  A thin layer of liquid
phase remains unextracted, bonded to the support surface.  Columns packed
with support prepared in this way, or purchased commercially prepacked
under trade names such as Ultrabond (Supelco) or Fermabond (Dow), have
been used without further treatment or after being conventionally coated
with another liquid phase for the separation of chlorinated, phosphate,
carbamate, and triazine pesticides and chlorinated phenols.

Support bonded columns have been used with electron capture, Hall electro-
lytic conductivity, and N-P thermionic detectors (see Section's 4,A,7; 4,C;
4,D; and 12,A of the EPA PAM).  The great advantage of these highly de-
activated columns appears to be the ability to directly chromatograph
polar and unstable compounds without derivatization and to achieve sharp,
symmetrical peaks.  Support bonded columns allow lower operating tempera-
tures and provide minimal column bleed, longer column stability, and high
efficiency and sensitivity  (13-16).
                                        -131-

-------
                                                              Section 4K.
    Aside from Carbowax 20M, polyester phases have been evaluated for the
    preparation of support bonded column packings (17).  In some cases, double
    support bonding was advantageous.  This involves coating the original heat
    treated and extracted support with an additional 5% of the same liquid phase
    and repeating the heat treatment process.
4K.     EVALUATION OF THE COLUMN

    Unfortunately, many chromatographers, after packing and conditioning the
    column, proceed immediately to use it without making the effort to system-
    atically determine whether it is good or bad.  Considering the fact that
    the column, if properly prepared and maintained, may be in constant use for
    a year or more as the most vital component of the gas chromatograph, the
    2 or 3 hours spent conducting a systematic evaluation is time well invested.

    In fact,  learning immediately whether the quality characteristics  are
    sufficiently good to justify placing the column on-line as  a  working tool
    could result in a considerable overall time saving.

    Full details of the evaluation procedure are included in Section 4,A of the
    EPA PAM.   The following material provides highlights  of the procedure.

    After completion of conditioning steps,  the oven and  carrier  flow  are shut
    down, and the column is connected to the detector.  A clean glass  injection
    insert and septum are also installed.   The oven temperature and carrier
    flow are  then increased to their operating values.  When the  proper  oven
    temperature is reached, the carrier flow rate is carefully  tested  with  a
    soap bubble device and adjusted.   (Subsections 5A and 5B discuss the proper
    performance of temperature and flow rate measurements.)   At least  1  hour,
    or preferably overnight,  is allowed for the chromatograph to  equilibrate.
    The temperature and flow rate are rechecked after equilibration.   Before
    making any injections,  a background (standing) current  profile  is  run at
    the normal operating parameters for the specific column being tested if an
    electron  capture detector is- used.   The polarizing voltage  is set  at its
    proper value.   These operations are further discussed in Subsection  5C  of
    this Manual.

    A complex chlorinated pesticide mixture is now chromatographed  to  evaluate
    efficiency,  resolution, compound stability,  and response characteristics.
    The mixture described in Section III,  C,5 of the EPA  PAM is useful for
    this purpose since it contains compounds that give a  number of  very  closely
    eluting peaks  on the recommended pesticide GC columns.   If  the  mixture  is
    prepared  in isooctane and stored tightly stoppered in the deep  freeze,  it
    is useable for a year or more for column evaluation (but not  quantitation).

    Fr.om the  chromatogram of this mixture,  one can calculate the  column
    efficiency based on the peak from £,jg_'-DDT.   For successful pesticide
    analyses,  this should be at least 500  plates per foot, or 3,000 plates  for
    a 183 cm  (6  foot)  column,  as  calculated from the equation shown in Figure
    4-C.   The relative retention time for jDjJE.1-MDT will indicate  the actual
    column temperature (Subsection 5A of this Manual and  Section  4,A of  the
    EPA PAM)  and serve as  a check on the instrument  pyrometer readout.
                                            -132-

-------
                                                            Section 4L
   The absolute retention time of the £,_p_'-DDT peak should be 15 to 18 minutes,
   or the operating parameters are incorrect,  the column is not the correct
   length, or it is not properly packed.   Too  low an absolute retention indi-
   cates too high an oven temperature or  carrier gas flow, too short a column,
   packing which is too loose, or a combination of two or more of these.  A
   high retention time would indicate the possibility of opposite causes.

   If column efficiency and resolution are favorable, compound breakdown is
   evaluated by injection of £,jo'-DDT and endrin.  Columns indicating poor
   resolution, efficiency, and/or retention characteristics that cannot be
   corrected by slight parameter adjustments should not be further used.  On
   the other hand, satisfactory columns will often improve or "season" with
   use, especially as cleaned-up sample extracts are injected onto the column.
   The percentage composition of the liquid phase undoubtedly changes with age
   for most columns as well./
  •
   Pure analytical standard £,j3/-DDT and endrin are  injected in turn in
   sufficient  concentration  to result in a total peak height of 50-60% full
   scale  recorder deflection.  Breakdown, as indicated by appearance of peaks
   in addition to the main pesticide peaks, should not exceed 3% for DDT and
   6% for endrin of  the amounts  injected.  The breakdown percentage is the
   value  of all peaks on  each chromatogram divided into the total peak area
   value  for  the breakdown peaks x 100.  Similar procedures are used for
   other  pesticide classes with appropriate standard mixtures.
   Reproduclbility of the size of peaks when a compound is injected re-
   petitively should be <2-3%.  Poor reproducibility can be due to breakdown
   or adsorption of the compound on the column or to extra^column causes such
   as faulty syringe or syringe technique  (Section 5J), a leaky septum (Section
   5J), or detector malfunction.  Reproducibility should be checked with those
   compounds that are possible to chromatograph successfully but that can
   break  down or be adsorbed  (e.g., endrin).  Priaing  injections of large amounts
   of a difficult compound, as mentioned earlier, may  allow maintenance of re-
   producibility for an adequate period of time for an analysis.  Difficult com-
   pounds should also be checked for linearity of response (Section 50d) since
   one cause of non-linearity may be on-column breakdown or adsorption.


4L    MAINTENANCE AND USE OF  GC COLUMNS

   Table  3-3 of Section 3 outlines a recommended maintenance program for a gas
   chromatograph with an electron capture  detector in  monitoring laboratories
   in which biological media  are predominant samples.  A properly cared-for
   column should provide service for many  months.  Off-column  injection of
   biological samples will  enhance column  life  (Subsection 5J); frequent (daily)
   changing of the  injection  insert and septum helps ensure continuing good
   performance.  Weekly, bi-weekly, or monthly, depending on the number and
   types  of samples  injected, the silanized  glass wool plug at the column inlet
   should be replaced.  This  is mandatory  when  injecting biological samples
   directly on-column.  If  the  glass wool  plug becomes contaminated by extraneous
   material, chromatograms  showing excessive DDT breakdown, peak tailing, and
   depressed peak height response will result.  Changing the glass wool
   regularly will usually restore proper performance.
                                            - 133  -

-------
                                                           Section  4L
The column packing near the inlet must also be replaced with fresh, con-
ditioned packing if it becomes contaminated.  Contaminated packing can be
removed without removing the column from the instrument by applying gentle
suction through a long-tipped disposable pipet inserted into the column.
The column interior should be swabbed where the packing was removed to
eliminate fatty deposits on the glass wall.  Elimination of the glass wool
at the column inlet has been recommended (EDA PAM, Section 301.9) for
minimizing fatty extract buildup at the top of the column by permitting the
extract to spread over the top portion of adsorbent.  This adsorbent* which

will trap  or degrade pesticides  less  readily  than contaminated  glass wool,
is  regularly replaced.  Daily monitoring of DDT breakdown is important
for early  indication of contamination of the  injection port and/or column.
Improved cleanup  of dirty  extracts  prior to gas chromatography  is  an
obvious  aid in maintaining good  column performance.

The effects of silanizing  conditioning do  not last indefinitely, and
breakdown  of endrin should be monitored weekly to determine if  and when
the treatment must  be repeated.  The  effects  of Carbowax  treatment appear
to  persist for at least three months  under normal use.  The operator should
watch for  a slow  decrease  in the response  of  organophosphorus pesticides
as  compared to that produced by  the column immediately after the initial
conditioning.  A  repeat Carbowax treatment of the same column appears to
rejuvenate the response, but may cause a shift in some retention values
relative to parathion.  Repeat treatments  are, therefore, not recommended
since consistent  relative  retention values are important  for tentative
peak identification (Subsection  5N).

When the column is  idle overnight or  weekends, a  low carrier flow  of
ca.  25/ml minute  is maintained through the column.and a simultaneous purge
flow of  25-30 ml  through the detector.  When  an instrument has multiple
columns  connected to  a single EC detector,  a  carrier flow just high
enough to provide positive pressure is maintained through the unused
column(s).   In a  series of observations with  a pair of nearly identical
lowload  columns having the same  70  ml/minute  flow through each, the peak
height response for aldrin was reduced ca.  25% compared to when the off-
column had  a very low carrier flow.   If the column not in use is of a
highly stable  liquid  phase such  as  OV-1, OV-17, etc., the carrier flow on
this  "off"  column may be reduced to zero with no  ill' effects, thus allowing
for  full response from the column in  use.

Columns removed from  an instrument  are tightly capped and are reconditioned
if out of the  instrument for more than a few days.  A flow of 60 ml/minute
carrier gas  for several hours at a  temperature ca. 25°C above the prescribed
operating temperature (venting into the oven) is  used for this operation.

Erratic and  noisy baselines frequently indicate leaks in  the column
connections  or some other  point  in  the flow system between the inj ection
port  and the detector inlet.  If the  chromatograph oven can accommodate
two or more  columns but only one is installed, the unused transfer line to
the detector must, of course, be plugged to prevent a massive leak.
                                   -134-

-------
                                                          Section 4L
Further details of instrument maintenance, troubleshooting, and calibration
are given in Section 6.

a.  Carrier Gas

    Impure carrier gas can often virtually and irreversibly destroy a
    column.  The main manifestations of this are evident in the inability
    to obtain an adequate background current profile, and low or zero re-
    sponse upon injection of standard solutions.  Every effort should be
    made to avoid installing a new column for evaluation at the same time
    a new tank of gas is placed on-line.  With this situation, the chroma-
    tographer cannot be sure whether he simply has a bad column or a bad
    tank of gas.  If the problem is traced to a bad tank of gas, the
    molecular sieve filter at the inlet of the flow system should also be
    replaced as experience has indicated that the contamination of the
    molecular sieve will perpetuate the problem even after a fresh column
    and good tank of gas are installed (Subsection 5C).

b.  Erratic Baselines

    This phenomenon may be caused by a number of instrumental factors, and
    these will be treated in detail in Subsection 5K.  The contribution
    of the column to this problem is largely one of loose joint connections,
    allowing air to seep into the carrier system.  Special care should be
    taken to ensure that both column joint nuts are tight.  One common
    occurrence is this:  The chromatographer connects the freshly con-
    ditioned column to the detector and makes certain that the Swagelok
    nuts are tight.  After about two days of operation, the oven door should
    be opened and the nuts should be tested with a wrench.  In almost all
    cases, it will be found that the nuts are no longer tight, sometimes
    requiring as much as a half turn for retightening.

c.  Accuracy of Oven Temperature and Carrier Gas Flow Velocity

    Information gleaned from the interlaboratory check sample program de-
    scribed in Section 2 has clearly indicated that in many laboratories
    the chromatographer does not really know his true column temperature
    or carrier gas flow velocity.  In most such cases, full reliance is
   , being placed in the accuracy of the instrumental pyrometer and ball
    rotameter, both of which may be grossly inaccurate.  These subjects
    will be discussed in Subsections 5A and 5B but are highlighted here
    because of the profound effects on the day-to-day operation of GC
    columns.  Figure 4-L is presented as an illustration.  A temperature
    of 200°C is recommended as optimum for the 1.5% OV-17/1.95% QF-1 column.
    At this temperature, the separation between p_,£*-DDE and dieldrin is
    normally as shown in Chromatogram A.  One laboratory reported operation
    at 200°C, but their chromatogram was that shown in B.  Subsequent
    Investigation revealed that the actual oven temperature was 185°C, or
    15°C at variance with the value given by the instrument pyrometer.
    Resolution or quantitation of either p_,;p_'-DDE or -dieldrin would not
    be possible in Chromatogram B.

                                     -135-

-------
        Figure 4-L.
                                          Section 4M
Effect of temperature on  resolution,  1.5% OV-17/1.95%
QF-1 column.
                  Actual Own Tana.   20!'C
                  SteNd Carrifffc*  55ml.
                  Computed Hfickwcy  I960 IK
                                         Vain.
                                        23mill.
                          Actual Own 1«m.   185'C
                          Statad Carrier ffow  61ml.
                          Computed Efficiency 3,310 Tl?
    d.  Sources of Supply of Blank Glass  Columns

        This subject is mentioned  here  only by reason of a very significant
        variation in prices between various suppliers for the same commodity.
        Price markups in excess of 700% are not uncommon, so it behooves the
        laboratory purchasing group to  do a little shopping to achieve the
        appreciable savings possible on quantity lots.

        The cited subsections of Section  5 treat these problems in greater
        detail as they relate to overall  operation of the gas chromatograph.


4M     CAPILLARY GC COLUMNS (see also Subsection 5L in Section 5)

   The bulk of  the material in this chapter concerns traditional packed GC
   columns,  which are  predominantly used today in residue analysis.  However,
   applications of capillary GC have increased greatly in recent years.

   Coating  a capillary column requires the deposition of a uniform 0.1-1.5 ym
   film of  liquid phase onto the walls of the glass tubing, generally
   10-60 m  x 0.25-0.50 mm id.   Coating techniques for wall coated open tubular
   columns  can  usually be fitted into one of two general methods, termed
   dynamic  and  static.   The dynamic method consists of forcing a solution con-
   taining  approximately 10% liquid phase in a suitable low boiling solvent
   through  the  column  under closely controlled flow conditions.  Usually the
   coating  solution is  applied as  a single, coherent slug occupying from 2 to
   15 coils  of  the column.   The slug  is forced through the column at a velocity
   of ca. 1-2 cm/second with nitrogen pressure.  Some workers utilize a single
   application while others prefer two  or three consecutive coating treatments.
   Several formulas have been  proposed for calculating the final thickness of
   film deposited by the dynamic technique.
                                        -136-

-------
                                                           Section 4M
In the static technique, the column is completely filled with a dilute
solution (3-10 mg/ml) of liquid phase in a low boiling solvent, and one
end is carefully sealed.  The filled column is placed under vacuum, and
solvent is evaporated under quiescent conditions leaving a thin film of
liquid phase.

A discussion of these techniques, as well as methods for preparing support
coated open tubular (SCOT) and porous layer open tubular (PLOT) columns
is contained in the book by Jennings (18).  SCOT columns have the liquid
phase deposited on a surface covered with a porous layer support material
such as diatomaceous earth.  PLOT columns have the liquid phase deposited
on a surface extended by substances such as fused silica or elongated
crystal deposits.

The methods of Grob &t_ al^. are probably the most followed by analysts
attempting to prepare their own capillary columns.  The procedure involves
treatment of the glass surface with barium carbonate, deactivation with
Carbowax 20M and Emulphor ON 870, and static coating of nonpolar phases
and dynamic coating of polar phases (19).  The same workers have described
a standardized quality test for capillary columns (20).

Onuska and Comba have described the preparation of surface modified wall
coated open tubular columns for specific application in pesticide analysis
(21).  A borosilicate glass column (20 m x 6.24 mm id) was treated with
NH4HF£ to form filamentary crystals on the inner wall.  After heating, the
column was washed with 10% EC1, methanol, acetone, and ether, followed by
deactivation with a 1% (w/v) solution of Carbowax.  The column was then
heated to 290°C and dynamically coated using a mercury plug method with a
4% (w/v) solution of OV-101 in ii-hexane.

Because of the difficulties in achieving reproducible surface preparation,
deactivation, and coating, most workers purchase capillary columns pre-
dated from commercial sources (e.g., Supelco, Applied Science Laboratories).
Single phases with a range of polarities are currently supplied.  Test
chromatograms are usually supplied with the columns, and efficiency is
guaranteed at a certain level.  A typical value is 2500 plates per meter
for a 0.25 mm analytical column.

The stability of capillary columns depends on the liquid phase, the technique
of the coating, and the temperature of operation and time of use at that
temperature.  Some workers have observed that columns last longer if they
are maintained at the operating temperature than if they are frequently
cooled and heated.  The use of dry carrier gas is important, especially when
flowing through a cool column.  Coated columns store best if they are filled
with dry, inert gas and flame sealed.  The size and composition of injected
samples affect column life.  Large injections may have a scrubbing effect
that displaces some liquid phase.  Some solvents, e.g., C&2, are especially
efficient at displacing liquid phases.
                                    -137-

-------
                                                              Section AN

    If capillary columns are not used above 260°C, excellent, low dead-volume
    connections can be made with 30 gauge heat shrinkable Teflon tubing.  The
    glass capillary is carefully butted against the connecting line, and a
    butane micro torch is used to shrink the covering Teflon tubing and seal
    the junction.


4H     REFERENCES

    (1)  Tindle, R. C., and Stalling, D. L., Anal. Chem., 44, 1768  (1972).

    (2)  Brockmann, H., and Schodder, H., Chem. Ber., 74.* 73  (1941).

    (3)  Bevenue, A., and Ogata, J. N., J.  Chromatogr..  50, 142  (1972).

    (4)  Hall, E. T., J. Assoc. Off. Anal.  Chem.,  54., 1349  (1971).

    (5)  Mills, P. A., J. Assoc. Off. Anal. Chem.. 51,  29  (1968), FDA PAM. Sec. 121.32.

    (6)  Analytical Methods for Pesticide Residues in Foods.  Health Protection
         Branch, Dept. of National Health and Welfare,  Ontario,  Canada,
         Procedure 12.2  (a).

    (7)  Sherma, J.,  and Shafik, T. M.,  Arch. Environ.  Contain. Toxicol.,  3_,
         55  (1975).

    (8)  Zitko, V., and  Choi,  P. M. K..  Technical Report No.  272,  Fisheries
         Research Board  of  Canada, Biological  Station,  St.  Andrews, N.  B.,
         p.  27,  (1971).

    (9)  Committee for Analytical Methods  for Residues  of Pesticides in Food
         Stuffs of the Ministry of Agriculture, Fisheries,  and Food, Analyst,
         102,  858  (1977).
    (10)  Figure from Dal Nogare, S.,  and Juvet, R. S.,  Gas-Liquid Chromatography,
         p.  15, Interscience Publishers, N.Y., 1962.

    (11)  Thompson, B., Varian Instrument Applications,  12_(3), 13 (1978).
               i
    (12)   Ives, N.  F., and Giuffrida,  L., J. Assoc. Off. Anal. Chem., 53_, 973
          (1970).
    (13)  Hall, R.  C., and Harris,  D.  E., J. Chromatogr.. 169, 245  (1979).

    (14)  Winterlin,  W. L., and Moseman, R. F., J. Chroma togr.^ 166, 397  (1978).

    (15)   Moseman, R.  F., J. Chromatoer., 166, 397 (1978).

    (16)  Crist, H. L., and Moseman, R. F., J. Chromatogr., 160_, 49  (1978).
                                          -138-

-------
                                                           Section 4N

(17)   Edgerton, T. R., and Moseman, R. F., J. Ctiromatogr.  Sci.,  18,
      25 (1980). .

(18)   Jennings, W., Gas Chromatography with Glass Capillary  Columns,
      Academic Press, N.Y., 1978, 184 pp.

(19)   Grob, K., Jr., and Grob, K., Chromatographia,  10,  250  (1977).

(20)   Grob, K., Jr., and Grob, G., and Grob, K., J.  Chromatogr., 15_6,
      1 (1978).

(21)   Onuska, F. I., and Comba, F. E., J. High. Resolut.  Chromatogr.
      Chromatogr. Commun.. 1(4), 209-210  (1978).
                                   -139-

-------
                                  Section  5
                      INSTRUMENTATION AND PROCEDURES FOR

                             GAS CHROMATOGRAPHY
   During the extended period of operation of the interlaboratory check
   sample program described in Section 2, a significant,number of analytical
   "bloopers" have been attributable to improper operation of the gas
   chromatograph.  In many such cases the operator had no idea that any-
   thing was wrong, primarily because no systematic guidelines were followed
   for monitoring the instrumental performance.  This section will present
   such guidelines for the proper operation of the gas chromatograph in
   pesticide residue analysis.  Some of the material repeats the instructions
   outlined in the EPA Pesticide Analytical Manual, but because of its
   importance in analytical quality control, it is worthy of reemphasis.
   Section 4 should be consulted for material on evaluation, standardization,
   and maintenance of GC columns and Section 6 for details of instrumental
   troubleshooting and calibration.

   Since the gas- chromatograph is the instrument in most widespread use in
   the pesticide residue laboratory, its proper maintenance and use is of
   primary importance.  Failure of any of the components, such as the oven,
   gas flow system, detector, electrometer, or recorder,  to function at
   optimal potential can markedly distort the overall .Instrument performance
   and the resulting qualitative and quantitative data.  Table 3-7 in Section
   3 outlines a series of periodic checks recommended for insuring a con-
   tinuing high level of chromatograph performance.  Figure 5-A, appearing
   on the next page shows the MT-220 (Tracer, Inc.) gas chromatograph, which
   is in widespread use throughout EPA laboratories.  This is a floor model
   chromatograph that features four vertical U-columns, on- or off-column
   injection, and simultaneous installation of up to four different detectors.

5A   TEMPERATURE SELECTION AND CONTROL

   Proper adjustment of the column oven temperatu-re and the carrier gas flow
   rate (Subsection 5B) will have a great influence on the caliber of per-
   formance of the entire chromatographic system.  Improper selection and
   control of these parameters may result in poor column  efficiency with
   concurrently poor resolution of peaks, inaccurate relative retention
   values, depressed peak height response (poor sensitivity), elution times
   that are either too fast to yield adequate peak resolution and reliable
                                      -140-

-------
                                                       Section 5A
peak identification or too slow to be practical, or rising or erratic
baselines.  Impaired resolution may preclude accurate quantitation of
two important pesticides that, are not adequately separated, while
inaccurate retention values will make proper residue identification
difficult.
            Figure 5-A.  Gas Chromatograph, Tracer  MT-220
The temperature regulation and readout systems of the column oven,
detector, and injection port of the gas chromatograph are critical for
obtaining reliable analytical results.  Accuracy of the pyrometer read-
outs must be established and maintained to prevent occurrences such as
electron capture detector tritium foil vaporization due to excessive
temperature or an injection port or column significantly higher or lower
in temperature than desired.

A properly operating temperature programmer will maintain the column oven
temperature without appreciable deviation (±0.1°C), provided that room
temperature fluctuations are minimal.  Excessive temperature fluctuation
will lead to erratic baselines and retention measurements.  Pyrometer
batteries (if your instrument is so equipped) should be checked monthly
to determine if they are delivering full voltage under load.  A hint of
                                   -141-

-------
                                                       Section 5A
inaccurate pyrometer operation is obtained by switching to an unused
sensor and observing the readout.  A value more than 5 C from room
temperature suggests faulty operation.   In addition, the oven temperature
must be monitored by means other than the built-in instrument pyrometer.
A pre'calibrated pyrometer with leads inserted through the oven door or
a. mercury thermometer placed down through an unused injection port is
recommended.  The instrument pyrometer must not be relied upon as the
only means of monitoring column temperature.

Injector temperature is determined by the nature of the sample, the
identity of the pesticide, and the volume injected.  An excessive in-
jection port temperature may lead to decomposition of heat-labile pesti-
cides, stripping of the partition liquid from the front end of the
column resulting in peak tailing, and increases septum bleed (leading
to spurious peaks) and reduced septum life.  A temperature lower than
the optimum may cause slow or incomplete sample volatization.  The
detector temperature should be 30-50°C above that of the column (50°C
above the final temperature when programming is used) to prevent the
possibility of condensation of sample components or liquid bleed from
the column.  An excessively high detector temperature can result in
reduced sensitivity and/or increased noise level.  Inaccurate column
temperature can affect peak retention times and resolution and may alter
the elution pattern of certain pesticidal compounds that may be present
in a sample, sometimes to the extent that two compounds that completely
separate at a given temperature may completely overlap at some other
temperature.

Column temperature may be checked by computing the relative retention
ratio for j3,£f-DDT (or another convenient pesticide) compared to aldrin
as follows:  divide the distance in mm on the recorder chart between the
injection point and the peak maximum for DDT by the distance between the
injection point and the aldrin maximum on the same chromatogram.  There
is a linear relationship between column  temperature and relative retention
values for organochlorine compounds (not organophosphates) so that com-
parison of this computed value with those available for over 50 pesti-
cides on the recommended pesticide columns between 170°C and 204°C [EPA
PAM, Subsection 4, A, (6), Tables 2(a) - 2(c)] should provide a check of
the actual column temperature.  Selected values for £.,£.'-DDT are shown
in Table 5-1 as an example.  A computed  relative retention value much
below the given value in the table at the selected oven temperature
indicates a temperature that is actually higher while a value much
higher than the chart denotes a low oven temperature.  Relative retention
ratios are also a function of the type and proportion of the component
liquid phases in the column packing, so  preparation of the column and
packing should also be carefully checked if retention values are discrepant.

Surveys of the data and chromato grams submitted by laboratories newly
participating in the EPA Interlaboratory Control Program (Section 2)
indicated that a high proportion of gas  chromatographs were operating
with column oven temperatures deviating  significantly from the supposed
                                  -142-

-------
                                                          Section 5B
   TABLE 5-1
                   RETENTION TIMES FOR
                                         RELATIVE TO ALDRIN
     Liquid Phase
                    170
174
178
                Temperatures, °C

               2    186    190
                                                            194
                                 198
                                                                         202
1.52 OV-17/1.95Z QF-1
4.0Z SE-30/6.0Z QF-1
5Z OV-210
5.57
4.04
4.47
5.
3.
4.
39
92
31
5.20
3.80
4.15
5.01
3.67
3.98
4.83
3.34
3.82
4.
.3.
3.
64
43
66
4.46
3.30
3.49
4.27
3..18
3.33
4.09
3.05
3.17
   values.  These erroneous temperatures resulted from inaccurate instru-
   ment pyrometers and a lack of alternate temperature monitoring devices
   and procedures.  As an example, one laboratory using a  column of  1.5%
   OV-17/1.95% QF-1 indicated an absolute retention time of  26  minutes and
   a relative retention ratio of 4.87 for p_,£'-DDT at  a temperature  of 200°C
   and flow rate of 65 ml/minute.  Under these stated  conditions the re-
   tention time for DDT should have been 18-20 minutes, and  reference to
   Table 5-1 shows the true oven temperature was ca 185°C, fifteen degrees
   less than the pyrometer indicated.  See Figure 4-L  jbi Section 4 for
   illustration of the effects of inaccurate column temperature on peak
   resolution.

   A discussion of the importance of the GC oven to chromatographic  per-
   formance and suggestions for simple evaluation techniques for thermal
   variables have been published (1).

5B   SELECTION AND CONTROL OF CARRIER GAS FLOW RATE

   The exact carrier flow system depends on the chromatograph in use.  A
   common arrangement is for the gas to flow from the  cylinder  through a
   two stage regulator to a filter-drier element, branching  thereafter to:
   (a) a purge line running through the purge rotameter, flow controller,
   and detector,'and (b) the carrier gas flow line running through the
   rotameters, the flow controllers, the column, and finally through the
   transfer line into the detector.  If temperature programming is used,
   differential flow controllers should be installed in the  carrier  gas
   line to prevent a decrease in flow as the pressure  drop increases across
   the column due to increasing temperature.

   The choice of carrier gas is dictated mainly by the requirements  of the
   detector being employed.  Nitrogen is required for  the  usual pesticide
                                     -143-

-------
                                                       Section 5B


detectors, except  that  the pulsed mode  of  the electron capture detector
may employ argon with 5% methane.  Flame detectors require gases such
as hydrogen,  oxygen, and air  for combustion.  N/P detectors require
helium.

Gases should  be obtained in the highest possible purity and gas cylinders
equipped with dual stage regulators.  "Prepurified" nitrogen is required
for the DC mode and, in some  cases, the constant current pulsed mode of
electron capture detection.   A gas that is 99.998% pure has an impurity
level of 20 ppm, and at least this purity  should be employed for the
carrier and auxiliary gases in trace analyses.  Each gas supply is
filtered through a filter-drier cartridge  connected at the regulator
output of the cylinder.  A filter containing Linde 13X (1/16 inch)
molecular sieve pellets will  remove water, most hydrocarbons, and CC^.
Before the filter-drier is charged with the fresh molecular sieve, the
interior of the cartridge is  acetone rinsed and heated at 130°C in an
oven for at least  one hour.   The bronze frit is acetone rinsed and flamed.
After filling, the unit is heated at-350°C for four hours with a nitrogen
flow of ca 90 ml/minute passing through the unit.  If activated units are
to be stored  for a period of  time, the  ends must be tightly capped.  The
filter unit should be replaced with a fresh one in the rare event one
discovers that a contaminated tank of gas  has been used.  Oxygen removal
requires a special scrubber or a molecular sieve filter immersed in liquid
nitrogen.  Gas cylinders should always  be  replaced before they are com-
pletely empty.

It is essential that no leaks exist anywhere in the flow system.  Even a
minute leak will result in erratic baselines with the % or 6%i electron
capture detectors.  If  the baseline has been stable but becomes erratic
upon installation  of a  new column, a loose column connection is indicated.
Leaks are detected by application of "Snoop" or some similar product at
all connections in the  flow system'from the injection port to the.de-
.tector, provided the connections are at room temperature.  Do not^attempt
to use "Snoop" on  a hot column.  Commercial high temperature liquid leak
detectors are also available  for high temperature connections such as
around injection ports.  These bubble-type leak detectors should be used
with caution  since the  solutions can be drawn into the GC system at the
leak source or at  a checked source once it is broken.

Another means of detecting leaks when using an electron capture detector
is by spraying connections with Freon MS-180 with the instrument operating
and observing any  recorder response.  Short sprays are applied close to
the connection, but not around the injection port or the detector.

The carrier gas flow velocities are checked using a soap bubble flow-
meter, which  can be purchased commercially or easily constructed by
attaching a sidearm near the  bottom of  a 50 ml buret [Subsection 4, A,
 (6), Figure 4(a),  EPA PAM].   Since rotameters are installed ahead of the
columns, they cannot be relied upon when adjusting the carrier flow as
they may be in error.   It is  necessary  to  check the flow rate at a point
                                   -144-

-------
                                                        Section 5B


after the column because the pressure drop across columns will vary
somewhat from one  column to another.  An equilibration period  of only
a few minutes with normal operating parameters is required before the
flow rate check is made.

If two or more columns  are connected to the same detector via  a common
transfer line, the carrier flow to the column(s) not  in  use  is shut off
while the flow rate  through the column in use is being measured.  Like-
wise, the purge gas  is  shut off.  If flow in all columns is  shut off,
the purge gas flow through the detector can be measured.  The  flow
through unused columns  is also shut off while determining the  background
current of an electron  capture detector,  The head pressure  gauge on
some commercial instruments allows continuous monitoring for problems
upstream of  the column, such as a leak in the carrier gas lines, as well
as determination of  minimum regulator pressure, changes  in column head
pressure during temperature programming, and long term column  changes.
If available, head pressure monitoring can accomplish some' of  the same
results as checking  of  the flow rate.

Carrier flow rates in excess of recommended values lead  to lowered
absolute retention times and compressed chromatograms while  rates that
are too low  will have the opposite effect.  Relative  retention values
reflect only the  operating temperature of the column  (Subsection 5A),
while absolute retentions indicate either or both carrier  gas  flow rate
and temperature.   Other effects of excessive flow rate may  include
depression in peak height response and poor column efficiency.  Figure
5-B illustrates  chromatograms of identical pesticide  mixtures  from the
same OV-17/QF-1  GC column operated with approximately the  recommended
conditions  (B) and then with too" rapid a flow ratep(A).

           Figure 5-B.   Effect of flow rate on GC resolution
                          1.5%OV-17/1.95% QF-1
                                     CoUam Oven Temp., "C      197
                                     Retent. Time p,p'-DDT, min.    21
                                     Carrier gas flow, ml/min. " .   55
                                     Oiart Speed, in./nin,      !j
                                     Efficiency, total theor. plates 3,560
             _ Injection point
3
-,.. g
= -fr
3 Q.
-., '
g
4.

I §
.2 1
Q -Q_

•§-•?•?
-S 0. Q.
                                    -145-

-------
                                                       Section 5B
                             GC Detectors
GC detectors for residue analysis must be sensitive to minute amounts
of the pesticides sought, but selective enough not to detect reasonable
amounts of co-extracted substrate material.  Despite this selectivity,
it is necessary to protect  the  total gas chromatographic system,
including the  detector, by  purifying the extracts.  This will reduce
the amount of  co-extracted  material in the  final solution to a level
that will not  be detrimental to the chromatograph or to the quality of
the separation, identification, and measurement.  Nonselective detectors,
such as the flame ionization detector, produce very complex chromatograms
with peaks from pesticides  as well as from  co-extracted compounds; these
detectors are  not selective enough to be practical for the quantitative
analysis of residues  of only a  limited number of pesticides of certain
classes.  With detectors  that are less selective  (or less specific) for
pesticide(s) of interest, more  effort is required in sample preparation,
in avoiding reagent  contamination, and in residue identification.

For pesticide  analysis, sensitivity of a GC detector has been  traditionally
designated as  the amount  of pesticide that  will provide a peak whose
height  corresponds  to some  percentage of the  full scale recorder de-
flection  (usually 10 or 50%).   Minimum detectable amount has been that
quantity of pesticide giving a  signal at least  four times the  background
noise (random fluctuations) at  baseline.  Detector sensitivity and minimum
detectable level are now  generally not differentiated  and are  reported
by instrument  manufacturers and many  chromatographers  in units of weight
 (Vg.  ng, Pg)  per ml for  concentration-sensitive  detectors  (e.g.,  electron
 capture GC detector, UV HPLC  detector) and  in weight per second  for mass
sensitive  detectors (FPD, N-P  GC detectors).  These numbers  designate
 the  concentration or flow rate that will produce  a  signal level  that  is
 some multiple of the noise (usually  2X).   If  the,sensitivity of  a .compound
 is  stated  as  16 ng/ml, the flow rate  1.5 ml/minute,  and  the  peak width
 20  seconds,  the absolute amount detectable is calculated as  1..5  ml/minute
 x 20/60 minutes x 16 ng/ml = 8 ng.   If,  for a mass  sensitive detector,  the
 minimum detectable level is 12 pg/second and the peak width  is 5 seconds,
 12 x 5 » 60 pg can be detected.  Both systems of stating detector sensi-
 tivity are used in this chapter.

The reader is  directed to references  (2-5)  for general reviews of the
element selective pesticide detectors.   A quality control program for  GC
detectors has been initiated by Agriculture Canada.   Twenty-seven labora-
tories were supplied with chlorpyriphos standard solutions  (chosen because
this pesticide contains Cl, S,   P, and N atoms) for the determination of
linear range and minimum detectable amounts.  Results have  been reported
(6)  for 23 FPDs in the P-mode,  18 FPDs in the S-mode,  28 electron capture
detectors,  and 20 linearized electron capture detectors.   This information
will be of interest to any  laboratory wanting to compare, its  detector
operating conditions and performance with the experience of others, or
those wishing  to set up a program of continuous detector monitoring.
                                   -146-

-------
                                                          Section 5C
5C   ELECTRON CAPTURE DETECTOR

   a.  Principles and Operation

       The electron capture (EC or electron affinity) detector is widely used
   for sensitive detection of halogenated pesticides or other classes of
   pesticides, often after derivatization with halogen-containing reagents.
   The detector consists of a radioactive source which emits low energy
   g-particles (electrons) capable of ionizing the carrier gas to produce
   secondary electrons.  A voltage is applied, causing a steady stream of
   these secondary electrons to flow from the source (cathode) to a collector
   (anode) where the amount of generated current is fed to an electrometer
   and recorded on a recorder..  Thus, a standing current or background
   signal is produced.

   When an electronegative species is introduced into the detector, a
   quantity of electrons will be captured and the current reduced.  The
   negative signal is in contrast to the positive current produced in
   detectors such as the flame ionization detector.  The magnitude of
   standing current reduction, which depends upon both the number of electron
   capturing species present and on their electronegativity, is measured on
   the recorder and indicates the amount of material capturing electrons.
   After the component passes through the detector, the standing current
   returns to the original value, and a characteristic GC peak is shown on
   the recorder, provided that the radioactive foil is not overly contaminated.
   The.exact theory of operation of the EC detector is still unresolved (7-9).

   The EC detector is selective in principle for highly electronegative com-
   pounds, but in practice it is the least selective of the widely used pesti-
   cide detectors.  Rigorous cleanup of pesticide extracts  is required to
   eliminate extraneous peaks due to compounds containing halogen, phosphorus,
   sulfur, nitrogen dioxide, lead, and some hydrocarbons.   Its sensitivity,
   however, is the highest of any contemporary detector, with many halogenated
   compounds being detectable in low pg  (10~12g) amounts.  Advantage is taken
   of this sensitivity by preparing halogenated derivatives of compounds
   (e.g., carbamate insecticides) normally not detected well by EC.  The
   response of EC detectors has been studied and guidelines presented for
   predicting which derivatives might best increase sensitivity  (10).

   Sources of  3-radiation have usually been either tritiated titanium on
   copper or  a 63Ni foil.  The latter is more expensive but can be used
   at  temperatures above  250°C, which would damage the tritiated detector
   (maximum temperature ca 225°C).  The nickel detector can be used safely
   to  400°C without appreciable loss of  radioactive material.  The higher
   operating  temperature  reduces  the possibility of  contaminating the de-
   tector with extract  impurities or the bleed from  GC liquid phases.   It
   also extends  the number of compounds  that  can be  detected and greatly
   reduces detector maintenance.  The tritium source is more sensitive  than
   nickel for a  short period of time, reaching maximum sensitivity after a
                                       -147-

-------
                                                       Section 5C
few days of operation.  Then there is typically a constant loss in sensi-
tivity, requiring frequent recalibration and eventual foil replacement.
The sensitivity of the ^^Ni cell is reputedly less than that of a tritium
cell, but it remains relatively constant and may equal or surpass the
sensitivity of a tritium cell after a period of use.  Some compounds
show increased sensitivity at the higher temperatures possible with the
Ni cell than with a new tritium cell (11).

The EC detector is used with either a constant negative DC voltage or
an intermittently pulsed DC voltage (constant frequency or "plain pulsed"
mode) imposed across the anode-cathode.  The former mode requires nitro-
gen carrier gas, while argon plus 5-10% methane is used with pulsed
voltage.  The argon-methane can be added to the chromatographic system
as a make-up gas or as the carrier gas.  When added as a make-up gas,
introduced after the column but prior to the ionization portion of the
detector cell, nitrogen or helium can Be used as the carrier gas, and
simultaneous dual detector operation is possible.  The pulsed and DC
modes provide approximately equal sensitivity and linearity, but advantages
have been claimed for the former in terms of freedom from anomalous re-
sponses (11), reproducibility of response, independence of response to
voltage, and operation with somewhat dirty samples.  However, DC operation
has proven entirely adequate for routine analyses in the.EPA Laboratories
when properly cleaned-up samples and low-bleed columns -are employed.

Constant current (pulse- or frequency-rmodulated or variable frequency
mode) operation is a third mode of EC detection.  A standing current is
again achieved by applying voltage pulses, but in this case the pulse
sampling frequency is varied by a servo-mechanism closed loop control
circuit that maintains the standing current constant even when an electron
absorbing compound enters the detector.  'Pulse frequency is.converted to
a DC signal that is monitored in the usual way to provide a chromatogram.
The basis of quantitative measurement is the relationship between the
change of pulse frequency and the concentration of electron capturing
substance.  This mode of operation provides an increased linear response
range without loss of detectability and a high degree of baseline stability
(12-14).
           63
Linearized   Ni constant current EC detectors are available from several
commercial sources.  They allow detection of low pg amounts of chlorinated
insecticides with isothermal or temperature programmed operation and
have a linear dynamic range of lO^-lO-*.  in one study (15), 27 laboratories
reported an average of 1.5 pg of chlorpyriphos required to produce a
readily discernable peak.  The compound-independent, extended linearity
is of great benefit for automated analyses where a wide concentration
range of samples can be analyzed without dilutions or reruns.  Those
detectors with small cell volumes (ca 0.3 ml) are well suited to capillary
column GC.  Most commercial linearized constant current EC cells can be
operated with either argon-methane or nitrogen carrier gas; the linear
range may be one decade higher with the former (15).  Detector sources
are either **-%i or tritiated scandium.  The latter has been found to have
a significantly greater linear range and similar sensitivity (16).
                                    -148-

-------
                                                       Section  5C
Figure 5-C illustrates the linearity
detector equipped with an electronic
[EPA PAM, Section 4,A,(3),IV].  This
of 104 with either nitrogen or argon-
the linearity of the Tracer detector
                          of the Tracor   Ni electron capture
                          linearizer (frequency modulated mode)
                          detector offers a linearity range
                          •methane carrier gas.  By comparison,
                          in the DC-pulse mode is 102-103 (17).
          Figure 5-C.  Response of the Tracor linearized I
                       detector from 5 x 1CT12 to 5 x 10~*
                       using argon-5% methane carrier gas.
                                                grams
               9.0 r
2
o>
8
OL
w
0)
Q.

I  8.0
            2
I
CO
o
Q.
CO

-------
                                                        Section 5C
 few times into a 300 C system-employing an empty column.  Purging the de-
 tector at 400°C overnight may also be helpful.  It has been reported (7)
 that the major contamination of the EC detector occurs by deposition of
 material on the anode surface, causing a significant reduction in
 efficiency of collection of electrical charge, and that performance can
 be restored by cleaning only the anode without disturbing the other parts
 of the cell.

                                         63,
               Figure 5-D.  Tracor, Inc.   Ni EC Detector
              """" ~    "
Response of EC detectors depends upon  temperature  (18); type, flow rate,
and pressure  (19) of the carrier gas;  cell and electrode configuration
and dimensions; electrode positions; amount of radioactivity;,contact
potentials caused by adsorption of sample components on electrode surfaces;
space charges of slow moving ions surrounding the electrodes; and applied
potential.  The adverse effects of even slight scoring on the EC collector
probe have been described (20).  The unpredictable nature of these
parameters causes anomalous responses, drifting baselines, variable
sensitivity, and a limited, variable dynamic range in the DC mode.
Operating parameters must be optimized for each manufacturer's detector.

b.  Background Current Profile.

    Measurement of the background current profile (recorder response vs.
voltage) should be made regularly to evaluate the performance of the
detector as influenced by the condition of the foil or other factors
such as column bleed or contaminated carrier gas.  At maximum voltage
and an attenuation setting of 12.8 x 10~9 A.F.S.  when using a 1 mv re-
corder, a good detector should produce a response of 60-80% full scale
deflection.  With aging this will approach 30%, when the foil should be
                                    -150-

-------
                                                       Section 5C

replaced.  A profile that drops drastically in a period of one or a few
days indicates problems with the detector itself or an adverse influence
by the column.  Detailed instructions are given in Section 4,A, (3) of the
EPA PAM for obtaining a BGC profile with a Tracer MT-220 chromatograph,
and typical profiles, are illustrated in Figure 5-E.  Operational details
for obtaining background current vary somewhat from one instrument to
another, and each particular instrument manual should be consulted for
recommended column parameters for making this test.  Some commercial
detectors regrettably do not provide for easy variation and readout of
the potential.  In general, more significant information is obtained by
determining the background current at the normal operating parameters
for the column being used.
            Figure 5-E.
Typical electron capture detector
background current profiles.
                  . GOOD .     FAIR
 From the background current  profile,  the optimum polarizing voltage  for
 operation may be estimated.   If the detector foil  is new and  the background
 current is high, it is usually acceptable practice to  simply  set the
 polarizing voltage at such a value to give 85% of  the  total profile  with
 the 3H detector or at 92% with the 63Ni detector.   If  operation above
 this range is attempted,  anomalous results can occur.   Figure 5-F  shows
 an obvious case in which  anomalous results were produced by operating
 above 99% of the maximum  current with a DC mode tritium detector.
                                     -151-

-------
                                                         Section 5C
                                 situatlon'  *» "PPer  chromatogram in
                              over-current can result from cleans ing of


A more reliable method,  especially for older, partially dirty detectors

iYn^Ar1^2?8 -^Se/response curve Is described^ Section  '

4,A, (3) of the EPA PAM.   Selection of the proper polarizing voltage is

very important so  as  to  (a)  produce maximum peak height (rfsponsefwJth
         Figure 5-F.  Normal electron capture  response  (A)' to

                      chlorinated pesticide mixture  and response

                      (B) resulting from operating at an

                      excessively high applied voltage.
           ui
           tn

           O

           %
           ui
           cr

           cc
           01
           Q
           ce
           o
           u
           UI
           cc
                                                    B
                                              JL
9    12    15    18


  TIME (min)
                                                   21
24
                               -152-

-------
                                                  Section 5C
Figure 5-G.
Chromatogram of. standard chlorinated pesticide
mixture.  Column:  1.8 m x 4 mm id glass  packed
with 10% DC-200 on a silane-treated support.
Column temperature:  200°C.   Detector:  electron
capture at 1 x 10~^ AFS.  (A) detector  voltage
10V, (B) detector, voltage 30V.   Broken  line on
chromatogram A indicates apparent baseline for
chromatogram B.  (Courtesy of Applied Science
Laboratories, Inc.)
                     Ul
                     Z

                     Q
                 Id
                    B
                             6    9    12

                                TIME(Min)
                                               18
                                                   21
                               -153-

-------
                                                        Section 5C

 minimum electrical overshoot on the backside of the peak (Figure 5-A,H),
 and (b) ensure maximum possible efficiency and peak resolution.  The
 polarizing voltage must be adjusted to accommodate a slowly deteriorating
 background current, so frequent profiles must be run to keep a check on
 this.   An article published in Gas-Chrom ®  Newsletter, March/April 1973
 treated this subject so well that a reprint is presented on the following
 two pages, courtesy of Applied Science Laboratories, State College, PA.

 As shown in Figure 5-E, DC mode detector profiles are plots of detector
 current along the Y-axis versus polarizing level in 5 volt steps.   In
 the pulsed mode,  profiles are plots of response current versus the
 frequency of the  polarizing pulse at baseline, i.e., with no electron
 capturing material present other than that over which there is no control,
 such as column or septum bleed.  Pulse profiles are normally generated
 using  variable frequency steps of 1, 2, 5, 10, 20,  50,  100, and 200 thousand
 cycles per second (kilohertz).  ^Figure 5-H shows a  comparison of profiles
 made with argon-methane carrier'gas and pulse polarization and profiles
 made on the same  detector-with DC polarizing voltage and nitrogen carrier.
 The left hand pair show a typical clean detector installed on a gas
 chromatograph.  Note that IgAT (argon-methane) is 6.4 x 1Q-9 aTnps versus

 an ZSAT 
-------
                                                                                      Section  5C
 Chromatographer,   Beware   of   Thy   Detector!
  W« ail know that the performance of'the same tyO«s of GC
columns can vary with the quaiitv of the packings and the
columns themselves. Figures 1 and 2 an examples of extreme
differences in column  performance whtn used  for pesticide
analysis. Both columns ir» 6  ft x 4 mm 10 glass U-tubas
packed with 10 wt % OC-200 on'asilane trMttd support.  Both
run* were mad* at tht  same operating conditions. Resolution
in Figure 1  is good,  but th«  resolution  in Figure  2 is far
superior because of the unbelievable 13.000 theoretical plates
obtained. Ytt, the separation  factors are the same in  both
    i (e.g.. 1.12 for endrin/dicldrin).
           L_jU
                    S    9    12   15
                     TIME (Minum)
                                      18   21
 figure 1. Qironutoorim of ftandard ehlorinattd pitticid* mixtun.
 Column: S ft * * mm IO gl«* p«efc«d with 10* OC 200 an •
 «H«n«-triMrtd «uooort. Column timptrnura: 200°C. DiMcter: £l«von
 ciptur* n 1 x 10 3 APS.

*)J
3






°
i 	 1
tf . 1
1 t
|
«




r*





X
.S
2
-





«
e
r*





	 J'1-'1— '
1 1
I) 38
N
K
i r
s i-
2 3
1 3 -

c
V


* I
Ul B
e ^
«. &
* c
j '
^ — P^

» 13 IS 18 21
TIME {Mi.iutml
 F^ura 2.  Chromiio^rim  o( itandxd ehiormand p«ticid« mixfun.
 Column:  8 ff « «  mm  IO gloii u-:k«i «nn  10% OC 20O on i
 ulxi*-tr»lM luooort. Column »mp«rolT>q> tor in €C dctKtor,

   At higner voltages, the rssponsa osccmes non-linear and the
 response-to-concentration slope increases with increasing con-
 centration. This non-linearity becomes ex treme on the plateau
 of  the current  vs.  voltage curve. Here, the response to
 concentration sloo«  is very small at low  concentrations and
 increases rapidly at  high concentrations.  This results  in an
 extreme contraction of the lower part  of 3 GC peak and an
 extension  of the upper part of the peak. When this occurs. A
                                                        -155-

-------
                                                                                             Section  5C
 chrorrutogram like  (ht one in Figure  2 is produced. Figure 1
 cm be convened to an approximation  of Figure 2. as shown in
 Figure 4. A superficial bavelme has been drawn which GUIS out
 tht bottom part of the peaks.  The similarity between Figures 2'
end 4 is obvious If we extended the upper part of the peaks in
Figure 4, the ehromatooram would resemble that  in Figure 2
still more closely.  This may appear extreme, but notice that in
Figure 2 we have lost not only the endnn decomposition
product, but also aU the smell impurity peaks that are seen in
Figure 1.
           Figure 4. Sam« chromatogfam a* in Figure 1.
                      At voltage* below the optimum range, the reverse occur*.
                   The response to concentration slope is high  at low concentra-
                   tions and decreases with increasing concentration. In this case.
                   'the lower part of a peak  will be extended and the upper part
                   contracted.  This is observed as a  wider  peak,  giving  a low
                   theoretical  plate calculation, and the peak maximum will tend
                   to be rounded. Also, sma.ll peaks will be overemphasized.
                      The graphs in Figures 5 to  8 summarize the effect of EC
                   detector voltage on GC results. These figures show the  effect
                   of voltage on various peak height ratios and  on the theoretical
                   plate calculation.
                      Another complication  is that the current vs. voltage curve
                   and the optimum voltage are not  always the same, but vary
                   with factors such  as  detector  cleanliness  and  liquid  phase
                   bleed. A dirty detector or a high liquid phase bleed will cause
                   the plateau  in the  current vs. voltage curve to shift to lower
                   currents and voltages,  as  shown by the broken  line curve in
                   Figure 3.
                      This  problem  of  variable  non-linear  response with  EC
                   detectors complicates quantitative analysis  and is the reason
                   why frequent and  careful use of  calibration standards is so
                   important  in  pesticide   analyses.  However, when  one  is
                   interested in determining column efficiency,  the effect  of
                   non-linear detector response on the theoretical plate response
                   has not been so obvious: This effect can also be observed with
                   argon ionization detectors, where  applied voltage also affects
                   linearity. Years  ago we found we were consistently obtaining
                   about  150  more  theoretical  plates  per  foot  from  argon
                   ionization detectors at voltages above the optimum than from
                   flame ionization detectors, which have good linear response
                   characteristics.
                      To be fair to EC detectors, there are now one or more on
                   the market which operate at a fixed voltage and are claimed to
                   have good linear characteristics over a wide dynamic range.
                      Let it  suffice to say: "Chromatographer, beware of  thy
                   detector! Also, know thy detector!"
      9-
      "

  Is-
  Ul
  S  4-
  V
  2  3-
      2-
      1-
      O
                                         O 0.4
                                         £  0
           ,10       30
            VOLTAGE (Vain)
                                                                           30
Figure •. flat of paak haifhi ratios of
•ndrin to dialdrin (a/dl and p.p'-ODT to
           t n.  EC dattctor valla**.
                 10       30
               VOLTAGE (Volti)
                                   30
   F»fure 5. Plot of padk height ratio of
   rtaptachlor to dialdrm (h/d) ««. EC da-
   Wclor voltae*.
              10       20

            VOLTAGE (Vdu)

Figure 7. Plot of paak haight ratio of
andrtn dacomp'oatioh product todialdrm
(E/d) v*. EC (tettclor voltaga.
             10        20
           VOLTAGE (Volts)
                                                                                                                   30
Figure «. Plot ol ealeulalad thaoratieal
platH for dialdrin v*. EC datactor volt-
                                                   -156-

-------
o
o
•H
-p
o
CO
CO
 o
 CO
•H
 J-l
 03
 ft

 o
 o

 0)
•H
•H
                                               tfe
                                               ccz
                                                  ZUJ

                                             • o^t
                                                             UJ
                                                             o
                                                 cc   o ^. .o
C O CD — <« CD LU
^UJ
5£l-UJ^K-UJ_J
  i UJ C/5O 3 tO ZD
                                                                      «M
                                                                                  -L
      ft

      J-l
      O
      •H
      O
      CD
      -P
      0)
      13
      O
     -p
      o
      0)
      CD
     13
      O
      6

     o
     o
     •O
      (1)
      CO
     OH


      •

     w




      0)
      W)
     •H
                         (9
           O

           ?5   I-

       SK5   S
                CE
                                              oi
                                            "*
                                                       ro          cvj

                                                         y010313Q
                                                     -157-

-------
                                                          Section 5D

   Oxygen is a frequent  contaminant  in nitrogen carrier gas, and the EC
   detector responds exceptionally well to traces of oxygen.  A background
   profile should be made after changing the tank of carrier gas and allow-
   ing at least one hour for the system to equilibrate.  A suitable oxygen
   scavenger and a clean chromatographic system are most important for good
   performance (13, 21).

   Certain liquid phases tend .to bleed in varying degrees at normal operating
   conditions, even after conditioning for extended periods of time.  DC-200,
   DC-550, DECS, and QF-1 are such phases and should be avoided where possible.
   The high temperature  0V silicones with low  (1-5%) phase loadings produce
   very favorable columns.  The background current determination is par-
   ticularly important with a new column in the instrument because background
   current that cannot be brought up to the expected level indicates the
   probability of a bleeding column requiring additional heat treatment to
   vaporize off the volatile impurities.

   Solvents and monomers can bleed from the septum and be swept through the
   column into the detector.  Glass inlet liners used for off-column injection
   should be changed frequently.  Proper handling of septums and maintenance
   of the injection port are discussed in Subsection 5J.

   Contamination from dirty tubing or other system components prior to the
   inlet can be caused by a bad tank of carrier gas containing grease, oils,
   or water vapor.  Use  of a molecular sieve filter-drier will usually pre-
   vent this problem.  These adsorbent traps must be regenerated regularly.
   If moisture has accumulated in the tubing and flow controllers : from a'
   bad tank of gas, simply changing the tank may not solve the problem.
   The entire system would have to be flushed out with a low boiling solvent.

   A good rule is to introduce only one variable at a time into the GC-EC
   system and to run a background profile just before and just after changing
   the variable.   For example, a column and a tank of gas should not both
   be changed at the same time.  This rule makes the isolation and correction
   of problems a much easier task.

   A review of the operation and principles of the electron capture detector
   for pesticide analysis (22) and a review of its theory and characteristics
   (11, 23)  have been published.

5D   MICROCOULOMETRIC DETECTOR'

   The original detector for the specific detection of organochlorine pesti-
   cides was the microcoulometric (MC)  detector.   The detector operates on
   the following principle:   column effluent is mixed with oxygen,  the organic
   compounds are combusted in a furnace,  and the formed HC1 is titrated in an
   automatic cell with internally generated silver ions.   The MC detector has
   now been largely replaced by the electrolytic conductivity detector
   because of its greater sensitivity and easier operation and maintenance.
                                     -158-

-------
                                                          Section 5E


   The MC detector can also be operated to be specific for pesticides con-
   taining S, P, or N, but the FPD, N-P, and electrolytic conductivity de-
   tectors are preferred for the detection of these compounds.  One advantage
   of the detector for some applications is that the amount of ions used in
   the cell reaction can be related theoretically by Coulomb's law to the
   absolute amounts of pesticides passing through the GC column.  Although
   the detector can still be purchased from Dohrmann-Envirotech and is
   occasionally reported in the literature (24), the absence of wide use
   in EPA laboratories and pesticide analytical laboratories in general has
   dictated that detailed material on this detector be deleted.  Interested
   readers are referred to earlier revisions of this Manual and the FDA
   PAM (25).

5E   THERMIONIC DETECTORS

   The original alkali flame ionization detector was described by Karmen
   and Giuffrida in the early 1960's.  They found that if a crystal of an
   alkali salt is placed over a flame and a collector electrode above the
   alkali source, a very enhanced response to phosphorus-'-containing species
   could be obtained.  A variety of designs of the alkali flame ionization
   detector (AFID) were reported, the mechanism of operation was widely
   studied, and further modifications made the detector sensitive to nitrogen
   as well as phosphorus compounds.  The AFID is described in detail in the
   FDA PAM, Section 313 and in Section 5E of earlier revisions of this QC
   Manual and is reviewed in references (26) and (27).  The alkali flame,
   Coulson, and flame photometric detectors have been compared and evaluated
   (28).

   New developments in the 1970's, pioneered mostly by Kolb et al. (29), have
   led to a thermionic detector involving an electrically heated bead, re-
   sulting in more reliability and an extended linear range.  This detector,
   termed the nitrogen-phosphorus or N-P detector, is described in Section
   4,D of the EPA PAM and Section 316 of the FDA PAM and will be Siscussed
   in the rest of this section.  N-P detectors are supplied by different
   manufacturers (see below).  A schematic diagram of the Perkin-Elmer
   detector incorporating a rubidium silicate bead is shown in Figure 1
   of Section 4,D of the EPA PAM.  In the usual mode of this detector, shown
   in the center of this Figure 1, the source is electrically heated, and
   the detector is sensitive to both nitrogen- and phosphorus-containing
   substances, but more sensitive to P than to N.  The linear range is over
   five orders of magnitude, and sensitivity is in the pg range.  In the
   P-mode, shown on the right of Figure 1, the source is heated by a high
   energy flame and the jet is grounded.  The response to phosphorus is the
   same as in the N-P mode, whereas the N response is suppressed ca 50 fold
   and the linear range is reduced to 10^.

   Compared to the KC1 AFID, the N-P detector operates with reduced flow of
   hydrogen gas and temperature.  This causes an increased response to
   nitrogen while maintaining high sensitivity to phosphorus.
                                      -159-

-------
                                                        Section 5E

 The principle of operation of the N-P detector is as follows:   GC column
 effluent is ionized at or near the electrically heated alkali  source in
 the presence of a relatively low temperature plasma around the bead.
 P- and N-containing species are preferentially ionized and drawn to  the
 collector electrode, where the resulting change in current is  amplified
 and recorded.

 For detection of N-containing pesticides, the N-P detector is  less se-
 lective than the N-mode of the Hall electrolytic conductivity  detector.
 However,  the former is more stable, sensitive,  and easier to operate.
 Sample extracts can often be examined after  a minimum of  cleanup using
 the N-P detector (see the FDA PAM,  Sections  232.4 and 242).  Extracts
 containing*traces of acetonitrile cause  a large detector  response that
 obscures  early-eluting pesticide peaks.   Since the usual  mode  of the
 N-P detector detects phosphorus and nitrogen compounds simultaneously,
 a  single  extract can be conveniently examined for both types of  compounds.
 For examination of P compounds only,  the P-mode of the FPD is  the usual
 best choice.   Some models of the detector do not tolerate injection of
 halogenated solvents without deterioration of the alkali  source.  -This
 should be determined prior to use.

 N-P detectors  are available from Perkin-Elmer,  Hewlett-Packard,  Varian,
 and Tracer.  Although each is basically  a flame'ionization detector to
 which an  electrically heated source has  been added between the jet and
 ion collector,  they differ in design  and operation and in the exact
 nature of the  alkali source (see the  EPA PAM,  Section  4,D,II).   Installa-
 tion,  operation,  and maintenance instructions  should be carefully followed
 for each  of  the detectors.   The detectors include a power supply  for
 heating the  source and for maintaining source bias  voltage.  An electro-
 meter as  used  for an FID  is required.  High  purity hydrogen and air
 should be used for the detector,  and  nitrogen or  helium for the carrier
 ga's.

 The  following  are general characteristics  of the  various N-P detectors:
 selectivity to' nitrogen against both  hydrocarbon  and phosphorus increases
with  decreasing hydrogen  flow, but  selectivity  to P, vs. hydrocarbon is not
 greatly influenced by flow rate.  The electrical potential difference
between the jet and  collector  can also affect selectivity, but this effect
varies with individual  detector  designs.   Selectivity factors range from
 10,000-100,000:1  for  N vs  C and  75,000-200,000:1 for P vs_ C.  Under de-
 tector conditions  providing a  1/2 fsd response  for 2 ng of parathion, pg
amounts of hydrocarbons or  S-  and halogen-containing compounds should cause
no response.  The  N/P response ratio  is 10-20:1.

Sensitivity is dependent  upon  the temperature of the source, which is a
function of the source heating .current and the flows of air and carrier
gas, and to the flow of hydrogen gas.  Manufacturer's optimal specifica-
tions are ca HP*3 grams N/second and lO"1^ grams P/second.  Greenhalgh
and Cochrane (17)" reported a minimum  detection level of 0.12 or 0.13  pg
                                  -160-

-------
                                                       Section 5E
chlorpyriptios for three different N-P detectors , which represented a ten
fold better sensitivity than the FPD (P-mode) .  The FDA PAM reports
practical detection limits of 1 ng for most OP pesticides and 5-30 ng
of most nitrogen pesticides.  Sensitivity for nitrogen pesticides varies
with the number of N atoms in the molecule and its structure.  Compounds
containing a P=0 group are about twice as well detected as the analog
with a P=S group (17) .  Detector linearity is at least three orders of
magnitude and as high as five orders.

It has been noted by analysts that differences among alkali sources are
common, even as supplied by one manufacturer.  Recommended gas flows
are:  air - 30-200 ml/minute, with 100 ml/minute being usual; less than
10 ml/minute is required if He rather than N2 is the carrier gas; hydrogen -
1-5 ml/minute, usually ca 3 ml/minute; carrier gas - 15-30 ml/minute.
Sensitivity is lower at carrier flows above 40 ml/minute.  To accommodate
this lower flow rate, 2 mm id packed columns rather than 4 mm id columns
can be used.  Similar chromatograms are produced in a 2 mm id column
with a 15 ml/minute flow and in a 4 mm id column at 60 ml/minute, with a
constant operating temperature.

The body of the detector is normally at 250°C and the source at 700-900°C.
The temperature is chosen by adjusting the potentiometer to produce a
baseline current as . recommended by the manufacturer, usually 1.5 x 10"^
amps or 4 x 1071.2 .amps,  or, the temperature can be set to produce a
specific response for a standard amount of pesticide, e.g. , 1/2 'fsd for
2 ng of parathion.

To reduce contamination of the detector, low bleed septa should be employed,
columns should be properly' conditioned, and injection of "dirty" samples
should be avoided.

Figure 5-1 shows a photograph of the Tracer N-P detector, and Figure 5-J
is a schematic of the mounting detail for connection to a gas chromato-
graph.  Figure 5-K shows a typical chromatogram using this detector.
The selectivity for the N- and P-containing pesticides compared to
eicosane  
-------
                                                     Section 5E

 Figure 5-1.   Tracer Model  702  N-P detector and control module
Figure 5-J.  N-P detector mounting detail for Tracer Model 560
                    rrn
:MOUNTING PLATE
                                          CERAMIC WASHER
                                               INSULATING TAPE
                                               OVEN WALL
                                      76I83-O334
                                -162-

-------
                                                    Section 5E

Figure  5~K.  Typical  chromatogram of pesticides  with  the
              N-P detector.

•"• v
	



UJ
_J
O
u,
a.
2
To
(0





































































— c
































































I—

UJ
— ~3

^








































— uj



— ffy

i
	 i

1 — u.







• ~" -
- 'o

	 X


CO



k
\
\
-v-
^«
, 	





'*"'-
C/5
2>
<.

01





x
"


Lu
^
M



















































































































^t
°f
-r



- 2it
- off

• ~~ir
i—

• cc.
^j
. 0_

1

- x
. h-
. ,g


























(














J=

e
•k
~ '
: (
: (

I





_

I































	 1
i
4-
T
i
11 , ,
-Jl
1 1
1
•
\
f v

SAMPLE: 1 x
5 )
1 >
SAMPLE SI
:OLUMN:6'
on
:OLUMN TE
NLET TEM
DETECTOR
:ARRIER F
3LASMA G/
DETECTOR
1.



















IMIN




•^"~






0
~~- n

2,
—

— o

*•

                                         I0"9g//tl Atrazine
                                          lO'^g/^J Methyl Parathion
                                          IO"6 g/fj.1 Eicosane in Hexane
                                          E = 4.0/iI
                                       6'x l/4"Glass-3%OVIOI
                                       on Chromosorb W-HP(80-90mesh)
                                           'ERATURE: 200 °C
                                           RATURE:2I5°C
                                           MPERATURE=2500C
                                           N- 25cc/min Helium
                                           FLOWS; 2!5cc/min Hydrogen
                                                  I20cc/min Air
                                           :KGROUND CURRENT:
                                       1.5 x 10"" amperes
                                -163-

-------
                                                          Section 5F
5F   FLAME PHOTOMETRIC DETECTOR (FPD)
   This detector operates by. monitoring HPO and S2 emission bands, which
   result from burning the column effluent in a cool, hydrogen-rich flame,
   at 526 nm (P-mode) or 394 nm (S-mode) using a combination of a narrow
   band-pass interference filter and a suitable photomultiplier tube.
   Samples require relatively little cleanup because of the selectivity
   of the detector for pesticides containing P or S.  Applications to the
   detection of certain other elements (e.g., Ti, As, Zr, B, Cr)  have also
   been made with limits of ca 10~7 - 10""1! grams.

   Figure 5-L shows the external appearance of the FPD.  The carrier gas
   exit line is seen on top and the 02/air inlet connection and hole for
   the heater on the bottom side.   The mirror lies behind the circular
   bulkhead seen covering the burner chamber on the front,  and the filter
   lies behind the screw seen on top of the photomultiplier housing.   The
   hydrogen gas inlet is on the opposite side, and column effluent enters
   underneath the burner chamber.   Signal and polarizing cables are
   attached at the back end of the PM tube.   A cross section, view of  the
   FPD is shown as Figure 5-M.   Figure 5-N pictures the dual FPD  which
   in principle allows simultaneous monitoring of sulfur and phosphorus
   output from a single injection,  as well as normal flame  ionization out-
   put if it is of interest.   In practice,  differences  in sensitivity of the
   P  and S modes make dual operation impractical for analysis of  low  amounts
   of residues  where maximum sensitivity is  sought.   That is,  if  the  P-mode
   is optimized,  the S-mode will not be sensitive enough to  be of use.
   Figure 5-A shows  a Tracer  FPD mounted  to  the  MT-220  gas chromatograph.
   At least three other companies produce FPDs that  differ in a number
   of construction aspects  and  performance.

              Figure 5-L.  Tracer fi?me photometric detector
                                                             fit
                                     -164-

-------
                                                      .Section 5F
Figure  5-M.  Cross section of a flame photometric detector
      Photomultiplier tube
                                                    Sw'agelcck
                                                     fitting
                                                  02 in
                                              ^•Column effluent
                                                   '(N2)
        Figure 5-N.  Dual flame  photometric detector
                                -165-

-------
                                                        Section 5F

 In the original operating configuration of the FPD,  oxygen (± air)  is
 nixed with the nitrogen carrier gas and pesticide at the entrance to
 the detector;  hydrogen is brought directly to the burner.  Components
 are burned in  a hollow tip that shields the flame from view by the
 photomultiplier (PM)  tube.  Emission occurs above the flame tip,  and
 the light is transmitted to the PM tube through a filter that transmits
 a specific wavelength of the element to be monitored.  A potential  is
 applied to the PM tube, and its output is amplified by the electrometer
 and read-out on a recorder.

 When the FPD is operated in this manner, as little as 3 pi of solvent
 in the injected sample will extinguish the flame unless modification is
 made to vent the solvent.   A Valco 4-part switching  valve (No.  CV-8HT),
 silylated before installation, was recommended for. this purpose.  Re-
 versal of the  hydrogen'and air/02 gas -supply lines at the detector
 inlets has been shown to give a "hyperventilated" flame (30)  that allows
 injection of up to 25 pi of solvent with no flame blowout and similar
 or better sensitivity,  baseline stability,  and linearity, but an
 approximately  20-fold loss in selectivity for most detectors.   In this
 "reverse configuration",  effluent is premixed with hydrogen..

 The minimum detectable  quantities of the elements S  and P reported  in
 different sources  are about 40 pg-1 ng and  10-100 pg,  respectively.  In
 routine operation,  2.5  ng  of ethyl parathion should  yield a peak  height
 equal  to 1/2 fsd,  although sensitivity can  usually be improved well
 beyond this in most analyses by careful adjustment of operating parameters,
 In a comparative study  (15),  the limit of detection  for chlorpyriphos was
 115 pg and 174 pg  in  the P-mode for normal  (flame-out)  and reverse gas
 flow configurations,  respectively,  and 167  and 87 ng in the S-mode with
 the two flow configurations.

 The degree of  sensitivity  of the FPD relates  to  at least  four factors;
 condition of the PM tube;  voltage applied to  the PM  tube;  flow rates
 of H2,  02,  and air; and the condition of the  viewing system (optical
window and filter);   Each  of  these factors  should be  checked and  opti-
mized  for each installation.   A drastic reduction in  the  peak height of
malathion can-be an indication of a poor column,  provided the rest of
 the system is  known to  be  operating properly.  Equal amounts of malathion
 and ethyl parathion normally  give a peak height  response  ratio of about
 0.70 on a good column.

To  obtain optimum flow  rates,  set hydrogen  flow  at 150-200 ml/minute,
obtain maximum response for an injected,  early eluting phosphate pesti-
cide (e.g.,  ronnel  or diazinon) by varying  oxygen  flow with zero air
flow,  then maximize response by varying  the air  flow with oxygen set at
its optimum value.  Some detectors may show best response with no air
flow.  Maximum response is  indicated by a large signal to noise ratio;
an  increase  in  flow rates may  increase peak height while also causing
an  increase  in baseline noise.  Generally, a high flame temperature
resulting  from too much oxidizer  and too little fuel will give the poorest
                                    -166-

-------
                                                       Section 5¥


sensitivity.  Typical operating parameters for the FPD in the P-mode
are in Table 5-2.  These values are for the Tracor FPD, based on recommenda-
tions in the manufacturer's detector manual and experience in EPA labora-
tories.  Table 5-3 lists FPD parameters used in analytical laboratories
participating in the Canadian Check Sample Program on Pesticide Residue
Analysis (6).  These values were chosen to produce in the S-mode an
exponential response factor of 1.9, and to give maximum response in
the P-mode.  The Tracor company (Personal Communication, 1980) recommends
the introduction of 60-100 ml/minute of hydrogen through the air/02 inlet
and 80-150 ml air into the port marked for oxygen.  The use of oxygen is
not considered necessary by Tracor with this reversed plumbing arrange-
ment .

The temperature gradient between the column and detector is kept as low
as possible, and the detector is always heated before the column when
starting up a cold system.  The detector base is maintained at about
210°C.  The nitrogen to oxygen ratio should be ca 4:1, the carrier and
purge flow rates should be equal, and the total flow of air, oxygen,
and carrier should not exceed 200 ml/minute.  A lower total (100-150 ml/
minute) usually produces the most favorable signal to noise ratio.

Although the FPD is not as sensitive to gas leaks as the electron capture
detector, the flow system of the chromatbgraph should be tight.  Leaks
in the hydrogen, oxygen, or air supply can be hazardous from an explosion
standpoint*  A common cause of sensitivity loss in the FPD is air leaking
into the burner chamber causing a change in flame characteristics.  This
is remedied by replacing the gaskets (probably discolored and brittle)
with new, bright, and flexible ones.  Stability of the FPD is promoted
by having a very stable flame.  This is best accomplished by using high
quality flow controllers for the detector gases.

Optimum response voltage for the PM tube is determined using a variable
power supply that allows the voltage to be increased with little increase
in electronic noise.  Raising the voltage from the electrometer will
increase electronic noise inordinately.  With optimum flow rates, the
power supply is set at 750 V, and a sample of diazinon is injected to
give 30-60% fsd.  The injection is repeated at 850 V and at voltage
increments between and around these values until the point of maximum
signal to noise ratio is determined.  It may be necessary to attenuate
to keep on scale during this determination, so the linearity of the
electrometer must be known.  Different PM tubes require different voltages
for best performance, a value of about 850 V being typical.  A suspect PM
tube may be checked with one of known sensitivity to give indication of
its condition.  Satisfactory operation of the FPD over its full dynamic
range requires both a highly stabilized 750 V power supply plus an
electrometer with a bucking capability at least 1 x 10~" amps.  PM tubes
are heat sensitive and should be well insulated from sources of heat in
order to keep sensitivity from being lost.  The tube can be kept cool by
blowing air over it or circulating cooling water through a copper tube
wrapped around it.  In a modified FPD design that protects the PM tube
from heat radiation, emitted light is carried to the phototube by a light-
pipe.
                                    -167-

-------
                                                            Section 5F
Table 5-2
         OPERATING PARAMETERS FOR THE P-MODE OF THE TRACOR FPD
           Temperatures ( C)
Flow Rates'(ml/min.)
Column
Injection Block
Detector**
Transfer Line
Switching Valve*
200
210-225
165-250
235
235-240
Purge*
Carrier
Hydrogen
Oxygen
Air
70-80
70-80
50-200
10-30*** .
0-100
      *  With Valco switching valve
     **  High temperature model is never heated above 250 C, low
           temperature model never above 170°C
    ***  To ignite the flame, an oxygen flow of 50 ml/minute or
           more may be required, depending on the detector
  Table 5-3


           GAS FLOW PARAMETERS SUGGESTED FOR OPTIMUM RESPONSE

        WITH THE, MELPAR (TRACOR) FLAME PHOTOMETRIC DETECTOR (6)
Old configuration

H_ (ml/mdLn)
Alr( " )
o2 ( " )
02/H2 ratio
P-mode
200
80
10
0.13
S-mode
70
30
10
0.22
New Reverse
P-mode .
200
30
15
0.11
Configuration
S-mode
50
50
10
0.4
                                   -168-

-------
                                                       Section 5F


Interference filters may be changed at any tike, i.e., it is not necessary
to shut down the instrument to do this.  The power to the PM tube must be
turned off when removing it from the flame base.  Excess light will damage
or destroy the sensing element when the tube is connected to the power
supply.  Light leaking into the PM tube during operation of the chromato-
graph will increase the noise ,level and decrease sensitivity.  One remedy
is to tighten all connections having gaskets between the burner and the
mounting of the PM tube.  A second remedy is to construct an opaque fiber
or plastic tube to slip over the connection of the PM tube and the metallic
link with the cooling fins, or a light-tight box to completely house the
burner and PM tube.  Ignition of the flame may be detected by observation
of recorder pen deflection up or down scale, hearing the "pop" of the
hydrogen gas, or deposition of moisture on a cold, flat metal surface
held near the exhaust tube.  After several months of operation, the quartz
window in the FPD burner becomes pitted and "fogged" or opaque.  This loss
of transparency can cause a decrease in sensitivity that can usually be
restored by polishing the window with carborundum or jeweler's rouge and
a polishing mat commonly used by infrared spectroscopists for salt windows.
If the pitting or fog is too deep and cannot be polished out, the window
should be discarded and replaced.

The FPD response to P is linear over a concentration range of about 3-5
decades, e.g., 0.4-400 ng for parathion and 0.14-400 ng for chlorpyriphos
(15) in the P-mode. "Nonlinear response of the FPD (526 rim filter) to
oxygen analogs of OP pesticides is often noted and is thought to be caused
by degradation of these P •= 0 compounds.  GC columns should be optimized
for separation of these compounds without breakdown, and metal transfer
lines between the column and detector should be as short as possible and
preferably made from Teflon or glass-lined metal tubing.  The S-mode is
inherently less sensitive than the P-mode, and response for compounds
containing a single S atom is nonlinear starting in the 1-10 n£ range.
The response increases very,roughly as the square of the concentration
of sulfur, so standard curves are plotted on semi-log paper for S-mode
quantitation.  Quantitative evaluation of chromatographic data from the
nonlinear S-mode FPD has been theoretically :and experimentally studied
(31).  A linearizing amplifier option is available for commercial detectors
that electronically transmits the square root of the detector response to
the recorder so that plots of peak height or area vs concentration are
linear within ± 5%.  This linear response facilitates easy interpretation
and allows electronic integration and data acquisition not possible with-
out the square root function.  The potential errors involved in the use
of these commercial linearizers, if response is not actually proportional
to the square of S concentration, have been evaluated and recommendations
made to minimize the error (32).

The unique square-law sulfur response of the FPD can be used to help
distinguish sulfur pesticide peaks from interfering peaks due to large
concentrations of nonsulfur compounds, such as hydrocarbons, that can also
be detected.  Because the peak height of sulfur compounds varies as the
square of the sulfur atom flow into the FPD, peaks due to sulfur compounds
                                   -169-

-------
                                                          Section 5F

 tend to be narrower than those of nonsulfur compounds with  comparable
 retention times.   Therefore,  visual inspection can often identify these
 unusually narrow peaks.   A more definite identification will be  achieved
 if the volume of sample  injected is increased.   Peak heights of  sulfur
 compounds will increase  as the square of the injected volume while peak
 heights of nonsulfur compounds will increase linearly.   As  a result, the
 sulfur compounds effectively  rise up out of the background.   If  the
 hydrocarbon peaks are not well resolved from the sulfur compounds, hydro-
 carbon quenching of sulfur light emission may diminish the  advantage of
 this square-law response.


 In order to operate in the dual mode, it is necessary to optimize com-
 bustion gas flows for the S-mode and to have sufficient sulfur to detect
 in this mode.   This combustion mixture is not necessarily the optimum
 for best phosphorus response.   Optimum conditions will vary from de-
 tector to detector.  If  enough residue is present to detect in the S-mode,
 attenuation must be used to keep the P response on scale with the S re-
 sponse.

 The proper attenuation for a  given sample will depend upon  the sensitivity
 achieved,  but,  in general,  it  is best to operate near the maximum
 and to dilute  the sample as necessary.   Selectivities for P and  S are
 about 10,000-25,000 or more:l   compared to nitrogen,  carbon,  hydrogen,
 and oxygen.  Large amounts  of  sulfur impurities give a response  in the
 P-mode (P:S response ratio  4-25:1 at 526 nm)  whereas phosphorus  impurities
 cause negligible response in  the S-mode (S:P response ratio  100-1,000:1
 at  394 nm).  As  the degree  of  sulfur oxidation in the molecule increases,
 there is usually a decrease in sulfur response.   Factors  affecting se-
 lectivity of the FPD have been studied (33).

 Maximum utility  of the FPD  is  afforded by the dual photomultiplier arrange-
 ment (Figure 5-N)  whereby P and S are simultaneously monitored on a dual-
 pen recorder.  This arrangement informs the analyst whether  a compound
 contains only P  or S, or both,  and the P/S  ratio  (the P-response  divided
 by  the square root of the S-response)  is  important information for con-
 firmation  of residue identity.   The.response  ratio  (xlO3) ranges  from
 5.0-5.8  for  PS compounds, 2.5-3.4 for PS2 compounds,  and  1.6-2.4  for
 PS3  compounds  (34).   As mentioned earlier,  dual operation will not be
 practical  for analysis of low  amounts  of  residue barely detectable by
 the  P-mode of an FPD optimized for this mode  because  of the much  lower
 sensitivity  of the S-mode under these conditions.

 Errors have  been noted (35) in quantitations with the FPD in  the P-mode
when automatic integration  is  applied.  The detector  response passed
 through  a minimum  after the solvent peak  and  then gradually rose to the
baseline without passing through  a maximum  to stop the integration before
 the  first pesticide peak.   This was overcome by adding a low boiling
organophosphate  (e.g., tributyl phosphate) that eluted after the solvent
but before the pesticide peak  (malathion was studied).  The FPD has been
 coupled with a capillary GC column for analysis of OP pesticides  (36).
                                   -170-

-------
                                                          Section 5G
   The FPD has proven to be a versatile, sensitive, selective, and reliable
   means of analyzing not only pesticides and metabolites containing P and
   S atoms, but also for compounds such as carbamate insecticides in the
   form of derivatives containing these elements.  The FPD has advantages
   over the normal flame thermionic detector for routine analysis in terms
   of ease of operation, better stability, less maintenance time, inde-
   pendence of response to gas flow rates, and need for less frequent
   injection of standards.  Sensitivity of the FPD is about one order of
   magnitude less for P compounds than with a fully optimized flame or
   flameless thermionic detector.  Applications and limitations of the FPD
   in atmospheric analysis have been reviewed (5).

   Varian has developed a new FPD with dual flame design that is reportedly
   (37) superior, but it has not yet been carefully evaluated for routine
   residue analysis.  Two hydrogen-air flames are used .to separate the
   regions of sample decomposition and emission, so that the upper emission
   flame is more efficient , and sensitivity is improved compared to the
   single-flame FPD.  The major claimed advantage of this construction is
   the reduction of the effect of hydrocarbon background quenching of the
   light emission from S- or P-compounds, because much of the C-H emission
   takes place in the lower oxygen-rich flame, while only the upper hydrogen-
   rich flame is viewed by the photomultiplier.  Reported selectivities are
   10* grams C/gram P and 10^ grams - 10^ grams C/gram S, and response is less
   affected by the compound structure because of more complete breakdown into
   82 and HPO species.  Up to 200 Ul of solvent can be injected without
   extinguishing the flame, and a pushbutton JLinearizer for the exact
   quadratic response of the S-mode' is included.
5G   ELECTROLYTIC CONDUCTIVITY DETECTOR

   This detector operates by mixing of the column effluent with oxygen or
   hydrogen reactant gas j followed by oxidation or reduction in a furnace
   containing certain catalysts.  In the original Coulsrn detector, ionizable .
   species emanating from the combustion zone are contacted with deionized
   water, and the carrier gas is separated from the liquid in a separator.
   The conductivity of the water is changed due to the presence of the
   ionized species, and the change is measured and displayed on a recorder
   in the form of usual GC peaks .  Table 5-4 shows the various modes of
   operation of the Coulson conductivity detector as described by Cochrane,
   The conditions and selectivity would be similar for the Hall detector.
   Selectivity and sensitivity in these modes are governed by the furnace
   temperature, nature and flow rates of the reactant and carrier gases,
   flow rate of water through the cell, and proper choice of catalysts and
   scrubbers.  Each analyst must optimize his conditions for the compound
   in which he is interested.  As a result, the minimum sensitivity values
   reported by different workers for compound classes have varied quite
   widely.  A general review of the electrolytic conductivity detector has
   been published (38).

   Gas chromatography with the Hall electrolytic conductivity detector (HECD)
   is described in Sections 4,C,(l)-(5) of the EPA PAM.  Included are discussions
                                      -171-

-------
                                                            Section 5G
sr
to
I














^•^
4J
•H
45

•H
CO
B
PJ
W
















^
jj
T*1
•5
4-1
U
cu
rH
CU
CO














CO
0
•H
4-1
•H
*T3
B
0
CJ






eu
o
SC




• M
• CO
co eu •
ca I 4j B
cu P ca o
P ca s fi
3 ej ca -H
,0 N
• 60 p CB
4J B  eu o 4J
S > 3
0 0 TJ >>
*-• S O rH
M cu p B
CO p IX 0

TJ
eu p
, g°
o cu
cj iw re
o
o en i
in re
co is -f^
i i re
55
ca 4-1
ca ca co
6C >> ca co
fj H) 13 ^0
o u cu
•H CB 4-1 p
4J o o eu
o eu -H
3 (U 4J p
TJ p CU p
cu -H TJ eo
P 3 u
TJ
CM'H B P
jn ss CB <






52;





B
CO
4)
TJ
•H
O
•H
4J
CJ
eu
CO
B
•H

rH
CJ

60
B

0
-a-
vO

• *
rH
CJ

60
CX
0
O
m



*
TJ ^
CB  O
CO O
Ss ca m
4J 60 CO
S CM P
CU 33 O





«J
CJ





0)

ca
60

CO •
TJ S
B O
3 iH
o m
ex
S ^t
0 O
o
B
2 o

60 60
B B
•rf
O TJ
0 B
CN CO
CM D.
1 eu
O TJ
ro
*
• n T3
CO CO

B'CM
*»*^
. •-< H • .

1 P B
1 -H eu cu
•H CJ ,0 CU
» CO 4J CO 3 JS JZ
tH eu 3 B TJ 3
(U <8 4J CO P CJ CO
•« P eu -H « ex co tH

5 cj CJ o P O
O B >. cB CO
4-1 6C-H4J o CUTJCJ  B ca i
TJ 1 CU -H 4-1 <1) CM
B .0 4J O p. >< >sQ
O 3 43 CJ CU$<4J4JCO
(X 3 0) O. 0) -H -H TJ
COTJPrHCO >tHPCU

Pl3cOCOPCB4->rnl4-t3

O I
CM 4-1
O TJ
CO TJ 4-1 CU
CU P- CO
O B 4J 3
4J -r( P 1
. CU CU
TJ a > co f>
eu ca B ea CU
N •> P
•H p 0 cB 3
TJ 4J p B 4J
•H CO tH CU C8
X CJ -H 4J P
O CM P CO CU
O 1 P ^ (X
CO CB rH Q
TJ B eu cj ca cu
B -H rf> 4J 4J
3 3 (Nfd
O ro 4J p; o cu
txo u
S to >, I eu ca cj
o 4J N a o
O TJ CX rH 3 PO
1 B Q CJ cB 3 m
CO CO CU 33 60 Pn CO






CO






«l
ca TJ
CU B
B cB
•H
rH CO

•• 60


60 O
B O
iH
rH 1
vO

rH •-
CJ CO

M Of.
O G ' 1
•••. CO
TJ o eu
B O TJ
CB O -rl
rH O
CO I -H
CM 4->
60 ca
a. eu eu
0 B30*
O CO rH
m P cj


CO

M .
rH
CJ1

«
H^J

d ,
4J

ca
TJ
§
CX
ca
eu
p








CO
•H
CO

rH O
O O
p in
SCO
ex
p
cu o
,0
3 0
4-1 O
O
•" *i
1-8
cu .

CtTiH
•rl ^-^
CO
tH ey\
CB rH
B
4J
B «
ca •

3 «
rH U
o

i*l
d td
*^
§.
«
-H C
CO O
O P
O* »H

& CJ
CO W
S1^
rH •
•> B

(0
jS »M
y«k^
4J cn

TJ iH
eu *-*
4-1

CU O
ca CM
cu
p «

Jv,
M
• A
*** p
• «
& 0
•> ca
cu B
G 0
CB P
U
O •
CJ H,

• •
B
O
p
(K
                                        -172-

-------
                                                       Section 5G


of the GO instrument, choice of columns, methods of quantitation and
interpretation, GC data and chromatograms, and a complete description
of the principles and details of operation of the various modes of the
Tracer Model 700 detector.  The FDA PAM also contains material on the
Hall detector, in Section 315.

The current commercial version of the Hall detector is shown in Figures
5-0 and 5-P.  Figure 5-0 displays the basic components of the 700A de-
tector and conductivity cell, while Figure 5-P shows the appearance
of the commercial detector.  The Model 700A is available only with the
Tracer Model 560 gas chromatograph.  The Model 700 is"similar to the
700A and can be connected to other gas chromatographs.  The Model 700A
features precise electrolyte flow regulation, a microreactor furnace,
extremely low dead volume, improved scrubbers, and automatic solvent
venting.  A new differential conductivity cell design combined with a
bipolar pulse cell excitation system provides increased sensitivity
compared to older models.                             •

A comparison of the Hall and Coulson detectors has been carried out.  An
approximate 7-foid increased sensitivity was found for the Hall detector
relative to the Coulson detector for nitrogen-containing pesticides.
Values obtained on a 4% OV-101/6% OV-210 column -at 205°C were as follows
(40):
          Atrazine
          Bladex

          Chloropropham
          Diazinon

          Ramrod

          Parathion
ng for
Coulson
7
15
75
25
50
150
1/2 fsd
Hall
1.1
1.1
6.0
4.5
6.5
20
The Tracer company reports sensitivities of 40 ng of atrazine (N-mode)
and 40 ng atrazine and 20 ng aldrin (Cl-mode) for 30-60% fsd peaks with
<1% fsd background level for the 700/700A detectors under the following
typical operating conditions:  1.8 m x 6.4 mm 3% OV-1 column, 200°C,
helium carrier 50 ml/minute, hydrogen reaction gas 50 ml/minute, 50%
n-propanol in deionized water electrolyte flowing at 0.8 ml/minute,
850°C furnace temperature (41).  The improved sensitivity of the Hall
                                   -173-

-------
                                            Section 5G
Figure 5-0.
Cross section of HECD microreactor
and conductivity cell (39)
 Figure 5-P.
 Tracer Model 700A Hall electrolytic
 conductivity detector
                         -174-

-------
                                                        Section 5G


detector  is seen by  comparison  of  these values with Table 5-4.  As
seen in Figure 5-R,  experimental detection levels  are often well below
these reported amounts.
       Figure 5-R.
Chroraatogram obtained with the Hall electrolytic
conductivity detector in the Cl-mode (2)
                                   §?
                         60
                                     a. e

                                      3
                                02    4
                                  TIM6, minutaft
                                fi       '''  "5              5
Selectivity values are Cl/C > 10 , N/C >UO , and S/C > 10 , and
linearity for Cl is 105-106, for N 104, and for S > 104.  Figure 5-S
demonstrates the sensitivity and selectivity of the N-mode, while
Figure 5-T shows the chromatogram of spiked soil extracts (20 grams
soil/40 ml methanol) injected without cleanup.  Figure 5-U shows the
sensitivity and selectivity for sulfur detection with the catalytic/
oxidative mode; the pyrolytic/oxidative mode can also be used for
S-detection with about one order of magnitude superior selectivity.
Figure 5-V shows the S-mode analysis of lettuce extract (acetone)
without cleanup.
                                    -175-

-------
    Figure 5-S.
HECD 700A selectivity
  in the N-mode (39)
                                                              Section 5G
90





IU
i
a
Ul
tc









0

j|
I
t>
i
"i
<









i



rjl









eft
0
1 g
o
2
S3
1
| Figure 5-T.
g Soil extract (A and B)
J with atrazine added, and
~*~\j ""' (C) control (39)
• it , ,
048
.TIME, minutes /\


10
0


10




10
0


Soil
} +0.1 ppm
Atrazine
f^^^-^——
"~K
h—INJECT
B
Soil
+O.05ppm
| Atrazine
_ i
r<_. . .
. ^
p- INJECT
C
Soil Control
— , J 	
— L/v,
N-INJECT
1
t |
Q g 7ftl»3-0<09
TIME,, minutes
-176-

-------
                                                            Section 5G
 Figure 5-U.   HECD  response  to hydrocarbon and  sulfur compounds  C39),
                     so
                   id
                   
                    LU
  Lettuce
+0.05ppm
Malathion
                                                       j
                                   48      12     16

                                    TIME, minutes   "1M-°'0>
                                       -177-

-------
                                                          Section 5H

   EPA experience has been that detection sensitivity of the Hall detector
   is lost in some laboratories in  the analysis of sample extracts com-
   pared to results with standards.  Good results have been, obtained when
   gel permeation chromatographic cleanup of sample extracts is combined
   with the Hall detector.  QF-1 or OV-210 fluorinated GC liquid phases
   may not be used with the detector in the N- or halogen-modes (42).

   The following are some operating characteristics and maintenance
   instructions for the Hall detector as outlined by Bayer  (43):  Cleaning
   requirements are minimized by disconnecting the furnace  to cell transfer
   line, leaving the furnace on, and turning the pump off at the end of
   the day's analyses.  Build-up of carbonaceous residues in the quartz
   tube is alleviated by running the furnace at high temperature in the
   oxidative mode.  Siliceous deposits resulting from silicon column
   bleed or silyl derivatives.can be removed with 10% HF.  Alternatively,
   the quartz tube can be replaced.  Small variations in the conductivity
   solvent flow rate will change the detector response, so the flow rate
   should be set to a constant value each day.  The recommended 1-5 cc/minute
   hydrogen flow rate through the standard 2 mm id quartz tube is very
   difficult to achieve in the reductive mode with the supplied needle
   valve, but variations have only  minor effects on detector sensitivity.
   The maintenance and cleaning required depend on the type of samples
   analyzed.  Weekly or more frequent cleaning may be required if dirty
   samples are commonly analyzed.   The procedure, requiring less than one
   day, involves disassembling the  unit and replacing the quartz, tube,
   Teflon transfer line, ion-exchange resin, and solvent.  The needle valve,
   its filter, and the conductivity cell are cleaned in an ultrasonic bath..
   Baseline noise can be caused by  air bubbles or residue trapped in the
   conductivity cell and ionic species in the solvent.  Bubbles are removed
   by rapidly turning the solvent pump off and on.  Residues are removed by
   disassembling and cleaning the cell in an ultrasonic bath, and ionic
   species are minimized by using high purity solvents and water and routinely
   changing the ion-exchange resin.

5H   OTHER DETECTORS AND DETECTOR COMBINATIONS

   The sulfur-phosphorus emission detector (SPED) is similar to the FPD
   except that fiber optic bundles  are used to transmit light from the
   flame to the PM tube and that the chemiluminescence of the HPO and
   82 species are monitored at different heights above the flame (the
   viewing port for P is 6 mm above the port for S).   This detector has
   been evaluated for pesticide analysis with the following results (28):
   response was similar to the normal FPD (linear in the P-mode,  squared
   in the S-mode, and quadratic for compounds containing both P and S);
   linearity for three standards (Ro-neet, DEPPT, and DEPP)  in the P-mode
   was 102-103, and the minimum detectable amounts were 5 x 10"11 to
   2.3 x 10~13 g/second.

   The photoionization detector (PID)  is in principle a. flame ionization
   detector in which the ions are created by UV photons instead of a flame.
                                      -178-

-------
                                                       Section 5H
Sensitivity is 10 to 50 times greater and the linear dynamic range is
10-100 fold greater than the FID.  The PID is sensitive to inorganic
compounds such as NHg, PH3, and AsH3 to which the FID does not respond.
The detector (44) has a sealed UV source focused directly into the
ionization chamber.  The UV energy photoionizes the various compounds
eluting from the GC column but not the helium carrier gas.  Organic
compounds with ionization potentials greater than the 10.2 ev energy
of the source, such as C^-C^ hydrocarbons including many common pesti-
cide solvents (methanol, me thy lene chloride, carbon tetrachloride ,
acetonitrile) , do not respond.  The increased sensitivity is apparently
due to the increased ionization efficiency of photons compared to a flame
and operation of the PID in an oxygen-free environment * thereby eliminating
free-radical quenching.  The linear dynamic range has been reported as
10 ^ - 10**, and the minimum detection level was <2 pg for benzene.  The PID
response to carbon is proportional to the carbon number as is the FID.
Because it responds to 'the sample concentration,' maximum sensitivity is
obtained at low flow rates (1-10 to 10-100 ml/minute).  The maximum
operating temperature is 315°C.  The nondestructive detection of ng levels
of organophosphorus pesticides has been demonstrated (45) .

Principles of the use and details of the .arrangement of multiple detectors
for efficient GC determination and confirmation of pesticides of different
chemical types in a single sample are discussed in Section 320 of the FDA
PAMS and a specific description of the combination of electron capture
and thermionic detectors for simultaneous determination of OC1 and OP
pesticide residues is given in Section 321 of the same manual.  Mien a
combination of detectors is used, the chosen cleanup procedure must be
suitable for the least selective of the detectors, enough residue must be
injected to meet the minimum sensitivity of both detectors, and the nature
and amount of injected solvent must be compatible with both.  Figure F-W
illustrates the application of dual detectors to facilitate identification
and quant itat ion of pesticides.  The left chromatogram shows the electron
capture response to a mixture of six N- and/or P-containing pesticides to
which PCBs were added.  The six pesticides, the positions of which are
indicated, are obviously overlapped by the PCBs to a degree that makes
analysis very difficult.  The right chromatogram shows the selective
response of the N-P detector to the same mixture, with no detection of the
PCS peaks.  If the N-P detector were used alone, the presence of the PCBs
would not be ascertained.  Without the N-P detector, the pesticides could
not be properly determined.  Therefore, both detectors are useful for this
sample (2) .     .

A GC system with one column, a three-way effluent splitter, and five
different detectors [electrolytic conductivity (N-mode) , FPD (FID, S-, and
P-modes) , and electron capture] operating simultaneously was described
(46) .  A computer program for evaluating data from this system was later
published.  The program calculates retention times relative to two internal
standards as references and peak areas corrected for baseline drift (47) .

The mass spectrometer  (Section 10L) can be used either as a universal
detector or the ultimate specific detector for gas chromatography.  For
                                -179-

-------
                                                     Section 51


   the latter purpose, specific ion monitoring at a single mass number or
   simultaneous monitoring of several selected characteristic fragment ions
   is carried out  (48).  Linearity of response generally extends over
   several orders  of magnitude.  Picogram sensitivity has been achieved,
   even at high resolution (49).
   Figure 5-W.
Analysis of a pesticide mixture in isooctane solution to
which a small amount of polychlorinated biphenyl mixture
was added, by gas chromatography with two different de-
tectors.  Column:  1.8 m x 2 mm id glass, packed with
3% OV-101 on Gas Chrom Q, 80-100 mesh.  Column temperature:
190°C.  Sample volume:  1 yl containing each of the six
pesticides in the amount of 5 ng.  Peaks:  1-di-syston,
2-methyl parathion, 3-malathion, 4-parathion, 5-methyl
trithion, 6-ethion (2). •
                !mV«5l2
                                        ImV 1400
                                                         NPD
                                                              IINUTES
51   ELECTROMETER AND RECORDER

   The electrometer is primarily a device for amplifying the electrical signal
   from the detector prior to its introduction to the recorder.  Units may be
   single channel, designed to operate with one detector and recorder, or dual
   channel.  Customary controls on the electrometer include input and output
   attenuators, output polarity, and controls for the recorder zero and
   bucking current.  Servicing of electrometers is generally a function of
   a trained electronic technician or representative of the company manu-
   facturing the chromatograph.
                                  -180-

-------
                                                     Section 5J

   To check electrometers on the Tracer MT-220  chromatograph, set attenuators
   to the off position and zero the recorder.   Set the output attenuator at
   xl and record the baseline.   A steady  baseline with less than 1% noise
   should be obtained.

   Recorders may be of single or dual channel design, the latter being
   capable of receiving two separate voltage signals supplied from the
   electrometer to two pens that trace separate chromatograms on opposite
   sides of the same chart paper.  Electronic controls on a recorder
   usually consist of a pen zero and a signal gain adjustment.  Most
   chromatographs require a recorder with a full scale sensitivity of 1 mv
   and a full scale response of one second or less.

   Proper adjustment of the recorder gain control is extremely important.
   Some analysts, upon observing excessive baseline noise, erroneously
   conclude that this should be eliminated by lowering the gain.  When the
   gain is set too low, however, the resulting  chromatograms appear
   "terraced" with a stepping-stone effect in the baselines.  In extreme
   cases, peaks have jagged and flat rather than pointed tops.  When this
   is evident, correction can usually be  achieved by advancing the gain
   control to a point just short of pen chatter.


5J   SAMPLE INJECTION AND THE INJECTION PORT

   a.  On-Column and Off-Column Injection

       Some gas chromatographs have injection ports designed to accommodate
   either on-column or off-column injection.  The former entails insertion
   of the syringe needle directly into the glass wool inlet plug of the
   column.  For off-column injection, some type.of glass or metal insert
   is installed in the injection port, and injection is made into this
   insert where the sample is flash vaporized and swept into the column
   by carrier gas.  In practical operation in a pesticide laboratory
   that  is injecting  a heavy volume of biological extracts, off-column
   injection through  a glass insert is preferable.  A significant amount
   of extraneous material that would otherwise be injected directly into the
   column is trapped  by  the insert.  If "your chromatograph does not provide
   the option of off-column injection, it is mandatory to frequently change
   the glass wool inlet  plug.  The frequency of, change is determined by
   daily monitoring of the extent of £,£f-DDT conversion  (see 4F of Section 4)
   The plug  is  changed when the  combined areas of the breakdown peaks  (DDE
   and/or DDD)  exceeds 3 or 4% of the sum of the areas of the £,p_'-DDT
   and the breakdown  peaks.

   An earlier discussion of some problems associated with injection of
   unclean  samples and maintenance of the injection  sleeve was presented in
   Subsection 4F.  Glass injection sleeves are  cleaned in chromic acid
   cleaning  solution, rinsed with water  and acetone, and  dried.  A final
   silanization treatment  of  the clean injection sleeve with Supelco's
   Sylon-CT  has also  provided a  dramatic solution to £,D_'-DDT and endrin
   breakdown problems.   The label  instructions were  followed.
                                  -181-

-------
                                                   Section 5J

 There are known Instances where changing the insert, glass wool plug, and
 even the first 1/2 inch of packing have not diminished conversion of DDT
 resulting from massive injections of uncleaned samples without proper
 on-going maintenance.  Final correction required disassembly of the entire
 injection port and wire brush cleaning of all metal parts to remove
 encrusted filth.  Following this, each part was further cleaned in an
 ultrasonic cleaner in alcohol KOH and finally acetone rinsed.  Some
 analysts recommend the use of a small plug of quartz wool in the exit
 end of the injection insert to act as a further trap for extraneous
 contaminants.

 b.  Septa

     A large number of different-types of septa are available commercially
 including inexpensive silicone rubber designed for low temperature,
 routine GC; high temperature silicone rubber; and expensive, layered or
 sandwich types.  Catalogs "of the different suppliers should be consulted
 for the specifications of the available products.

 The septum chosen for residue analysis should not produce significant,
 bleed (ghost peaks) that can affect identification of quantitation of
 residues under the conditions used for gas chromatography.   Bleed
 generally increases as the inlet temperature increases and diminishes
 with the length of time the septum has been installed.  A leaking septum
 may cause a number of problems,  including baseline drift, loss of sensi-
 tivity or erratic sensitivity, increase in peak retention times because
 of loss of carrier gas, or column deterioration due to entering air.

 Septum leaks can be caused by loss of elasticity as a function of tempera-
 ture and time of use, injection with a large diameter or damaged (bent
 or burred)  syringe needle, or incorrect tension of the septum nut.   It  is
 preferable to change the septum before a leak develops to prevent the
 production of incorrect analytical results.  Change might be made on at
 least a daily basis if the instrument receives heavy use.

 Replacement at the end of 'the working day is convenient since this  will
 allow the septum to condition during the night and be ready for use the
 next morning.   A needle guide on the syringe or an injection port with  a
 small diameter hole can prolong  the life of the septum by causing the needle
 to penetrate the septum at the same point,  thereby minimizing coring  and
 tearing of  the septum.   A needle guide also helps to maintain the
 integrity of the needle itself.   Overtightening of the septum nut will
 tend to extrude the septum and increase the amount of septum bleed.
 Undertightening can reduce the ability of the septum to seal.

 The septa in widest use for residue analyses are the high temperature
 silicone rubber septa such as blue HT. (Applied Science),  white HT (Alltech
•Associates), Thermogreen LB-1 (Supelco);  perfluoroelastomer type (Pyrosep,.
 Supelco); and  layered septa with  Teflon or  polyimide faces  that are
 placed against the injection port (LC Company).   In one comparative
 study (50),  the blue HT tested best for high temperature use and long
 life.   After baking at  300°C in an unused injection port overnight
                               -182-

-------
                                                  Section 5J
(15 hours), this septum gave good service for 50 injections using a
26-gauge needle, without leaking or producing ghost peaks.  The Teflon
faced septum did not produce ghost peaks initially but did Bleed after
a few injections had ruptured the face and exposed the silicone rubber.
After puncturing, the Teflon septum performed more poorly than the blue
HT.

In another comparative study (51), the rubbery Teflon perfluoroelastomer
septum gave the least bleed of all septa tested, including silicone
rubber and Teflon-faced septa.  Reports from another laboratory indicate
that pblyimide-faced septa are superior to those that are Teflon coated
in terms of bleed at high temperature, but that the unlayered high
temperature rubber type is still preferable.
            »
The method of septum preconditioning depends upon the temperature to be
employed.  For use below 250°C, rinsing with acetone, wiping with a
Kimwipe tissue, and air drying may be sufficient for HT silicone rubber
septa.  For higher temperature use, overnight baking at 300°C (50) will
probably be required.  Each laboratory should evaluate its septa (see
below), and precondition and replace them as required for its applications.
Septa should never be handled with the fingers, but rather with a tissue
or clean forceps.

Although different laboratories have individual methods for changing a
septum, the following considerations are appropriate.  It is undesirable
to expose a GC column to air while it is hot.  This can cause oxidation
of the stationary phase and column deterioration.  Also, quick removal
of the septum nut while the carrier gas is flowing (column under pressure)
can cause the column packing to shift or be blown from the column.  To
avoid these problems while changing the septum, reduce the column tempera-
ture and shut off the carrier gas flow when the column is cool.  When the
gas flow has ceased, remove the nut and insert the new septum.  Resume
the carrier gas flow, allow the column to flush for a few minutes, and
reheat the column.

The following procedure allows evaluation of septa for high temperature
applications:  place a clean metal disc in the injection-port nut and
install a short  (e.g., 46 cm) nonpolar column such as Dexsil 300 on
80-100 mesh Chromosorb ¥-HP in the gas chromatograph.  Heat the injection
port to 300°C  (or the temperature of interest), set the attenuator to
a  sensitive setting, and program the oven at 20°C/minute from 50°C to
the maximum temperature of the column.  No peaks should be detected.  If
the instrument has a septum purge, turn this off.  Cool the oven to 50°C
and replace the  disc with the septum to be tested.  Wait 10 minutes and
then temperature program the column as above.  If peaks are produced, a
preconditioning  step such as baking must be used to eliminate volatiles
from the septum.  Perform this preconditioning on the septum and reevaluate
it.
                                    -183-

-------
                                                   Section 5J

 c.  Injection Techniques

     (1)  Handling the Syringe

          When a sample is injected into the chromatograph,  it is  essential
 that it be entirely vaporized without loss.  Injections are usually  made
 using a 10 yl syringe for the electron capture,  thermionic, and FPD  de-
 tectors or a larger capacity if required for other less sensitive de-
 tectors (e.g., microcoulometric).   Automatic injection devices are
 available for use with some chromatographs and detectors.   (See Section
 50k).

 Samples are injected from a microliter syringe by inserting the needle
 through the septum as far as possible, depressing the  plunger with the
 thumb  or finger,  then immediately  withdrawing the needle (keeping the
 plunger depressed)  as rapidly and  smoothly as possible..   Some analysts
 prefer a delay of 1-2 or up to 5-10 seconds before withdrawal of  the
 needle.   When initially filling the syringe,  air is expelled by repeatedly
 drawing liquid in and rapidly expelling it with  the needle  tip still
 under  the liquid  surface.   The volume of sample  to be  injected is exactly
 adjusted by drawing up a couple of yl more than  necessary into the barrel.
 Hold the syringe  vertically with the needle pointing up, put the  needle
 through a tissue  to absorb expelled liquid,  and  push the plunger  until it
 reads  the desired value.   The excess air should  now have been expelled.

 There  should be no delay between filling and injection of the sample.
 After  injection,  the'syringe is rinsed clean by  filling with and  expelling
 5  portions of ethyl acetate or acetone,  and the  syringe  is  pre-rinsed with
 the  next sample to  be injected in  the same manner.   Be sure to follow
 carefully all manufacturer's suggestions for  proper use  of  each particular
 syringe*

 When a sample is  injected  in this  normal manner  from a 10 yl  syringe, the
 needle will retain  ca 0.2-0.3  yl of sample.   It  is  usually  safe' to ignore
 this volume since  standards  are injected for  comparison,' and  the  errors
 due  to retained volumes will cancel out  if equal volumes are  used and
 concentration differences  are  negligible.  Alternatively, the syringe may
 be filled by drawing the entire sample into the barrel, noting the final
 volume by reading each end of  the  column of liquid.  After  injection,
 the plunger is  pulled  back and  the small volume of  retained solvent now
 in the syringe  barrel  is read  and  applied  as  a correction.  This will
 correct  for nonreproducible  injection  technique but not, however,  for
 the error encountered  if the retained volume has a  composition different
 from the original sample,  as would happen  if nonuniform distillation had
 occurred in the needle.  Then  the  remaining liquid would be richer in high
boiling  sample  components.

This can be  overcome by using  the  solvent flush injection technique,  the
most reliable and reproducible method available.   About 2 yl of solvent
 is first drawn  into  the syringe followed by a 1-2 yl air pocket and then
                                -184-

-------
                                                  Section 5J


the required volume of sample,  the sample is brought completely into the
barrel so its volume can be read.  On injection, the flush solvent behind
the sample ensures injection of the entire sample without loss due to
hang-up.  Whatever method the analyst chooses to employ, he must be as
consistent as possible in his injections of standards and sample.  It is
critical that the solvent chosen for injection of the sample completely
dissolves the residues of interest, and the same suitable solvent should
be drawn first into the syringe for the flush technique.  The suitability
of the solvent should .be verified by obtaining reproducible peaks from
repeated injections of a sample dissolved in the solvent [see Subsection
5Jc(3)]>

It is good practice to reserve one syringe only for electron capture work.
If a series of concentration levels. is to be injected, the more highly
concentrated solutions should be injected last. , If the complete freedom
of a presumably clean syringe from pesticide traces is suspect, pure
solvent should be injected and any peaks would indicate contamination
and need for further cleaning.  Dirty syringe plungers and needles should
be wiped with lint-free wipers dipped in an appropriate solvent  (e.g.,
ethyl acetate), and the barrel should be cleaned by ^drawing solvent
through the needle and out the top with a vacuum.

    (2)  Preferred Volume Range

         Injection of 1-3 yl samples from a 10 yl syringe is not
recommended because of the large increase in error probability resulting
from these small volumes.  For example, a typical absolute injection error
                                                0 2
of 0.2  yl in a 1.0 yl injection would produce a
                                                    x 100 = 20% relative
error, while the same 0-.2 yl error in a 5 ul injection volume reduces the
relative error to a tolerable 4% level.  The analyst is strongly urged to
inject volumes between 3 and 8 yl (5-8 yl optimum) from a 10 yl syringe
for analyses with electron capture detectors.  A syringe filled close to
capacity is more difficult to manipulate.  Proper dilution of standards
and samples will provide on-scale peaks upon injection of optimum volumes.
Standard and sample solutions are prepared so that peaks of approximately
the same area are produced (Subsection 50).  With use of proper techniques,
a capable analyst should be able to reproduce a series of 3-8 yl injections
to within 1-5% of average peak area or height when response is ca 1/2 fsd.

The preceding paragraph describes the conventional wisdom concerning normal
use of a common 10 yl GC syringe.  Data have been presented, however,
leading to recommended volumes between 2-4 yl.  Below 2 yl, the error of
injection increased above the ±4-5% range.  Reproducibility decreased for
samples greater than 4 yl, supposedly due to the difficulty in quantita-
tively transferring the total volume from the syringe because the piston
sealed poorly and allowed the liquid to be forced back or leak through
the back of the syringe  (52).
                                  -185-

-------
                                                      Section  5K
        (3)   Injection  Solvent

             The  choice of  injection  solvent has been shown to affect quantita-
   tion of polar organophosphorus pesticides.  Hexane solution aliquots con-
   taining 4 ng  of dimethoate and 8  ng of  8-phosphamidon gave GC peak heights
   only 64%  and  43%, respectively, of those from corresponding aliquots of
   acetone solutions.   The low values with hexane apparently were caused by
   adsorption of the compounds in the syringe needle (53).  This error was
   overcome  By using acetone as the  flushing solvent in the solvent flush
   injection technique [Subsection (1) above] or, preferably, using acetone
   to prepare all standard and sample GC solutions.  Since this solvent
   effect is probably  a general occurrence in the analysis of polar pesti-
   cides and metabolites,  careful consideration should always be given to the
   choice of an  appropriate GC solvent.  Other workers have also recognized
   that .the  injection  solvent can affect the precision and accuracy of GC
   analyses  (54).                            '

   d.  Capillary Columns

       An inlet  system for use with  glass  capillary columns in trace analysis
  ' has been  described  by Spencer (55).  The system has a combination splitter
   and splitless  design, the latter  being most useful for' trace pesticide
   analysis.  In the splitless mode, microliter samples can be directly in-
   jected and the inlet backflushed  to purge residual solvent from the
   vaporization  chamber after the sample has entered the column.  Relatively
   large samples can be injected without overloading the system and causing
   band spreading.
5K   ERRATIC BASELINES

   If all modules of the GC system are functioning properly, baseline noise
   should be below 1% fsd.  When noise exceeds this level with the electron
   capture detector, all analytical work should be suspended until the
   cause is isolated and corrections made.  A poorly regulated current supply
   or column liquid phase bleed can cause an erratic baseline.  The slightest
   leak anywhere in the flow system may permit the entry of air and can be
   another cause of a noisy baseline.  The most common points of leakage
   are probably septa that are not changed often enough or loose column
   connections.

   Currently, three types of ferrules are most highly recommended for glass
   column pesticide work, and the choice appears to be mostly a matter of
   personal preference.  Teflon, graphite, and Vespel polyimide-graphlte-
   combination ferrules are all available in one piece design for use with-
   out a metal back ferrule or 0-rings.  Reducing ferrules for use without
   metal reducing unions and capillary ferrules are also available.
   Temperature limits are 250°C for Teflon, 450°C for graphite, and 350°C
   for Vespel.  The three types of ferrules are reuseable at least several
   times if carefully removed from old columns.
                                    -186-

-------
                                                     Section 5L


   Overtightening  of  ferrules  causes  deformation and  limits  the possibility
   of  reuse.   Expensive and valuable  glass  columns  can also  be broken.   Under-
   tightening  of ferrules  causes  leaking  and possibly allows the column to
   be .pushed out and  broken when  it is  subjected to increased back pressure.
   Most  analysts eventually develop a feel  for  the  correct ferrule torque.
   A typical procedure  is  to tighten  the  Swagelok fitting until finger
   tight and then  a 1/4 to 1/2 turn with  a  wrench until leak tight.  More
   precise  tightening without  column  breakage is facilitated by use of  a
   commercial  torque  wrench  (available  from, e.g.,  Siipelco)  that allows only
   the correct amount of force to be  applied to each  type of ferrule before
   slipping.   The  need  for further tightening of connections should be
   .checked  by  opening the  oven after  the  initial overnight heating period of
   the column.

   The column  should  be pushed into the fitting and then pulled back. 1/16 inch
   before the  ferrule is tightened.   Because of different construction
   materials,  the  instructions supplied with each type of ferrule should be
   consulted before applying torque to  obtain a leak-free connection.   The
   threads  on  the  instrument must be  in good condition to allow the nut to  be
   properly tightened,  so  a rethrtad  device should  be periodically used to
   clean the injector and  detector fitting  threads.  Leaking ferrules are
   located  by  use  of  a  liquid  leak detector around  the top  of the connecting
   nuts  or  by  monitoring pressure readings  of the head pressure gauge.

   When  temperature programming is employed to  facilitate complex separations,
   dual  column GC  operation will compensate for the baseline of the analytical
   column.  The dual  columns contain  the  same liquid  phase but need not be
   the same length.  To'set up the desired  baseline,  the recorder and de-
   tector are  first zeroed.  The columns  are then heated to  the upper
   temperature limit  of the program where the bleed from the columns will be
   greatest.   The  resultant baseline  is adjusted to the desired baseline by
   varying  the flow rate of  the balancing column.  Another  approach for "high
   bleed" analytical  columns  is to place  a short, "low bleed" scrubber  column
   (e.g.,  a low loaded silicone) at  the analytical column exit.

5L  ', RECOMMENDED GC COLUMNS FOR PESTICIDE ANALYSIS

   a.   Column Selection

       A number of important factors  must be considered in the choice of a
   column or combination of  columns most suitable for a particular laboratory.
   Some of these factors are the following:

        (1)   The selected columns should be capable of separating the largest
   number of pesticides of interest with a minimum number of overlapping
   peaks.   For example, 10% DC-200 or 3% OV-1 non-polar methyl silicone
   columns are of limited value to the analyst determining the more common
   chlorinated insecticides in environmental or animal samples.  Partially
   or completely overlapping peaks are obtained for several pesticides
   generally detected  in these sample substrates, e.g., p_,p_*-DDE and dieldrin
   £,jp_'-DDT and p^'-DOD^  and the isomers of BHC (Figure 5-X, part A).
                                      -187-

-------
                                                    Section 5L

      C2)   A high, efficiency column is desirable if injected extracts
  contain extraneous materials and detection of low pesticide concentrations
  is  required.   This will provide sharper separation between the peaks  of
  interest and  extraneous peaks of biological origin.

      (3)   Retention time or elution speed may be a primary consideration
  if  the  analyst is  concerned only with quantitation of  certain,  specific
  pesticides.   For example,  in a project where the sole  interest  is to
  routinely determine residues of a late-eluting pesticide  such as
  methoxychlor,  the  column selection and operating parameters would be
  tailored to elute  methoxychlor in a minimal time period consistent with
  'its  separation from any extraneous peaks.

      (4)   Pairs of  working  columns should be selected to be of dissimilar
  polarity and  therefore  provide different elution patterns  ("fingerprints")
  (Subsection 5N).

      (5)   Shorter columns may be adequate for  chromatography of certain,
  late-eluting pesticides, but for multiresidue analysis of unknown samples,
  a 6-foot  column is recommended  to  obtain optimum efficiency and peak
  resolution.

 b.  Phases Used in the EPA Laboratory Network

     After a careful comparative  study of many GC columns with the above
 factors in mind, the four liquid phases listed in Table 5-5 were chosen
 as working and confirmatory columns for the routine analysis of. organo-
 chlorine insecticides in human tissues.  These columns will.efficiently
 separate the principal compounds of interest (DDT, DDD, DDE, BHC isomers,
 heptachlor epoxide, and dieldrin) in a reasonable time, have low bleed,
 and give long service when properly prepared, used, and maintained,  as
 described in Section 4 of this Manual.  They have also proven to be
 excellent columns for general use in the determination of many pesticide
 classes in various substrates.  The SE-30/OV-210 and OV-210 columns  are
 especially recommended for separation of organophosphorus pesticides  to
 be detected with the FPD.

 Each of these  phases has its own peak elution pattern for the compounds
 of a given mixture.  An efficient column of the mixed phase OV-17 (phenyl
-methyl silicone) with QF-1  or OV-210 (trifluoro propyl  methyl silicone)
 separates all  usual tissue  peaks completely except for  a ca 75%  separation
 between _p_,£?-DDE and dieldrin.   Higher load mixtures must  be operated  at
 very high temperature and carrier gas flow  velocity to  avoid  slow elution
 and  are not recommended. The SE-30 methyl  silicone/OV-210 column gives
 no separation  between lindane and g-BHC but good separation between
 dieldrin and £,£f-DDE on an efficient column.

 The  single phase OV-210  gives a full separation of the  common BHC isomers,
 but  only fair  separation between the compound pairs  of  heptachlor,
                                   -188-

-------
                                                                                     Section 5L
m













CO
H
cc
J?
5
W
a
M
H
H
CO
W
PU
^y*
O
"
CO
w
CO
(=)
s
c
M
a
W
ti2
RECOMM1














CO

frf
O
&*^
c*"»

0
rH
*!*
•^
O
m





* 1

•^ o
C rH
CO fj
tz3 ^>
CO O
V-S B-5
*5T \O •

.**'

O
-«. rH
rH 1
° ^
u-i a\
rH rH








c*
•>
^ E
J3 >-l CO
M .C 
in I
cr\ o
tH r»





0
^•^ . vp
O |
CO in
t-t **

X X C
O O O -H
WO WE
p^ ^, ^**^
< H" < e
                                                                            o
                                                                            ex

                                                                            §•
                                                                            CO

                                                                           •o
                                                                            0)
                                                                            co


                                                                            o
                                                                            o
                                                                            o
                                                                            o
                                                                           •H
                                                                           •o
                                                                            
-------
                                                   Section 5L

 epoxide/£,£>DDE and £,p>DDD/£,£>DDT.  The single polyester phase DECS
 gives excellent separations of BHC isomers, complete peak separations of
 all compounds usually in tissues, and an unusual peak .sequence (8-BHC
 after £,p.'-DDT and p_,£»-DDT before £,.p_'-DDD) that makes it useful for
 confirmation of peak identities.  The DECS column, however, bleeds and
 degrades easily and has a relatively short column life.  It is, there-
 fore, not recommended as a routine working column, but only as a special
 purpose identification tool.  An excellent pairing of columns for
 analysis and confirmation of residues in samples containing OC1 pesticides
 and PCBs is either of the mixed phases in Table 5-5 and the OV-210 column.
 Unfortunately, pairing these columns entails GC runs at 200 and 180°C
 and necessitates either two gas chromatographs or one instrument with
 a change of column temperature and a rerun of sample extracts and
 standards.   For example, 1.5% OV-17/1.95% OV-210 separates oxychlordane
 from Aroclor 1254, while OV-210 resolves Aroclor 1254 from p,p'-DDT,
 heptachlor  epoxide and trans-nonachlor.   This pair is,  therefore, an
 excellent choice for analysis of samples containing oxychlordane and
 £.»£. -DDT, among other OC1 insecticides,  plus PCBs.

 Chromatograms of standard chlorinated pesticide mixtures  on these columns
 and a single phase nonpolar DC-200 column are shown in  Figure 5-X.
 Relative retention times of over 60 chlorinated and phosphate pesticides
 on OV-17/QF-1*,  SE-30/OV-210,  and OV-210 columns between  170 and  204°C
 are listed  in Subsection 4,A, (6)  of the  EPA PAM. Use of  temperatures
 other than  those listed  in the tables in the EPA PAM (preferably  the
 temperatures indicated with an arrow) is not recommended  because  of the
 greater  difficulty in comparing experimental RRT values to  the tables.
                  t                            _
 Carbowax 20M modified supports (Section  4J)  have been used  successfully
 for pesticide separations,  either directly or after  coatijng  with  a
 liquid phase.  Th-2 columns  have been used with electron capture,  Hall,
 and N-P  detectors.   Retention  data and chromatograms  are presented for
 a variety of pesticide classes in Sections  4,A, (7);  4,C,(5); 4,D; and
 12,A of  the  EPA  PAM.


 c.   Other Pesticide  Columns

    Many other phases besides  those  recommended  in Subsection b are
 available commercially.  Some  of  these may be entirely satisfactory
 for  residue  analysis while  others  are outdated or unsatisfactory for
 the  task.  As suggested in  Subsection a, a wide  range of factors must
be  considered in making the  column selection, and a -column or columns
wholly suitable for one laboratory may. be completely unsuited to the
typical work of another.

In the early years of GC analysis, only single, nonpolar phases were
utilized  for the separation of the nonpolar pesticides then important,
 equivalent to OV-210
                                    -190-

-------
                                                                     Section 5L
    Figure 5-X.   Peak elution patterns  of 13 pesticides on  five columns.
                              10% DC-200_
                                                            5V. OV-2JO
                           3% PEGS
4% SE-30/6?iOV-2]0
I
                                     . OV-17/1.95*/. Qf-1
                                         i  ls  Hil  i
                                         .1  «! 1   - •••  -v
                                               -191-

-------
                                                    Section 5L

 and  a  survey  of  the literature indicates  that  these  are  still widely
 used today.   However, mixed  phase columns that combine polar and  nonpolar
 t 2U ™  ln vary±nS  de§rees and newer  single  phases with  varying.polarities
 ttne 0V  series)  have become  increasingly  important as the range of pesti-
 cide types has drastically grown.   Some of the more  widely used additional
 phases include nonpolar  SE-30,  DC-200, DC-11,  and Apiezon L; intermediate
 ?niSrw 7 QF;1> OV-17, XE-60, Reoplex  400,  and  DC-550; and polar Carbowax
 20M, Versamid 900,  NPGS, butanediol succinate,  and NPGA.   Common  supports
 besides  Chromosorb  W or  Gas-Chrom P include  Gas Chrom Q,  Anakrom  Q or
 ABS, Supelcoport, and Diatoport S.  In a  particular  analytical situation,
 any one  of these or some other  column might  possibly be  equal or  even
 superior to one  of  those recommended  previously.

 The U.S. FDA continues to recommend 10% DC-200 and 15% QF-1/10% DC-200
 columns at 200°C with a  carrier flow  of 120 ml/minute in  their multi-
 residue determinative methods for  foods (FDA Pesticide Analytical
 Manual, Chapter 3), even though other lower-bleed liquid phases used
 with lower loading and slower flow rates provide greater response and
 resolution.   Relative retention data for over 300 compounds are listed
 in an appendix of the FDA PAM,  arranged according to both FDA standard
 number and to retention on DC-200.  Also listed are sensitivity data for
 the electron capture and thermionic detectors,  eluates for Florisil
 cleanup columns,  and recovery through FDA fatty and nonfatty food  methods
 (see Section 9Ab  of this Manual).

 FDA's primary mission is that of testing for  compliance with established
 tolerance, generally expressed  in terms  of parts per million (ppm).   In
 view of these relatively high concentration levels,  the use of  highly
 responsive and efficient columns is not  as critical as in the case of
 laboratories  testing in  the ppb and ppt  range.   Chapter 3 of  the FDA
 PAM also  contains extensive data on columns containing 2% DECS  (200°C
 60 ml/minute), 15'% QF-1/5%  DC-710  (2:1)  (200°C, 100-200 ml/minute),
 15%  OV-210 (190°C,  80 ml/minute),  and  10%  OV-101 (200°C,  120  ml/minute).
 Other recommended liquid phases include  SP-2100 (silicone), SP-2401
 (50%  trifluoropropyl substituted silicone, similar  to QF-1 and OV-210)
 HI-EFF-1BP (similar  to DECS), and  OV-11  (35%  phenyl substituted silicone,
 similar to 50% phenyl substituted  silicone DC-710).

 The Canadian Department  of National Health and  Welfare (56) specifies
 1.8 m x 6.4 mm columns of the following single  phases coated  at a  3%
 level on  Chromosorb  W, AW, or HP for their multiresidue monitoring pro-
 cedures:  OV-1 (nonpolar), OV-17 (slightly polar), OV-225  (medium  polar),
 ethylene  glycol adipate  (polar), and DECS  (very polar).  The relative
 polarities were calculated from McReynolds  constants  (57).  The 4%
 SE-30/6%  QF-1 mixed  phase is also  recommended.  A particular phase is
 chosen according  to  the polarity of the pesticide(s)  of interest.
Relative  retention times  are listed  (58) for over 100 pesticides on
 OV-1, OV-210 (intermediate polarity)., DECS, and mixed phase columns.

The mixed phases OV-l/OV-17, OV-210/OV-17, OV-225/OV-17,  OV-l/OV-25
OV-210/OV-25, and OV-225/OV-25 have been recommended for separation'of
                                   -192-

-------
                                                  Section 5L


organochlorine insecticides, and tables of relative retentions were given
for 14 compounds (59).  A column packed with three silicone stationary
phases, namely 2.5% QV-11 + 1% QF-1 +0.5% XE-60 (phases premised and
pan coated on Chromosorb W HP), was shown to resolve a 14-component
OC1 insecticide mixture in less than 19 minutes; retention data were
compared to other common pesticide columns (60).  Other extensive
compilations of pesticide relative retention data appear in References
(61-63).

Comparison of the separating characteristics of different GC phases and
selection of new, higher purity, and more stable phases to replace older
ones are facilitated by tabulations of McReynolds Constants.  These data
rate liquid phases according to polarity and selectivity and allow pre-
diction of similarities and differences in the ability of different
columns to produce a given separation (64).


d.  Capillary Columns (see Section 4M)

In addition to the packed GC columns already discussed in this subsection
5L, the use of capillary columns is growing rapidly in analytical situa-
tions where high resolving power is required.  The advantage of capillary
column GC in separating components of complex mixtures such as toxaphene
and PCBs is obvious if the chromatograms in Figure 5-Y are compared.
However, a mass spectrometer (Section 10L) is necessary if this improved
resolution is to be fully exploited for "real-world" samples.

The more common glass capillary column is the wall coated open tubular
type (WCOT), where the liquid phase is distributed as a thin film on the
inside wall surface without employing any support.  Columns are generally
25, 50, or 100 meters in length x 0.25-0.75 mm id.  The smallest diameter
gives the best efficiency but lower sample capacities (typically 1-50 ng
per component).  Because of the low sample capacity, injection is often
accomplished by an injector-splitter where typically 1 yl is injected,
0.01  yl enters the capillary column, and 0.99 yl is vented.  Sample
splitting is not generally employed in trace analyses.  Carrier gas flow
through such columns is ca 1 ml/minute, and a make-up gas system is re-
quired to sweep any void volumes and optimize detector flows.  The major
advantage of capillary columns is the high total number of theoretical
plates obtainable (plates per meter length are comparable with packed
columns) with these long, high permeability (low back pressure), open
tubes, leading to tremendous separation efficiency for complex environ-
mental samples.  The thin liquid film thickness provides fast analysis
times, often at relatively low temperatures, and sharp peaks.  To maintain
efficiency, it is of utmost importance to have a clean-cut, blunt column
end and a butt-to-butt Connection to the inlet splitter tip assembly.
Heat sljrinkable Teflon can provide an essentially zero dead volume seal
at this point and at the detector connection'.
                                 -193-

-------
                                                                  Section 5L
            g
          to
       OJ V) O
X
a)
                                               -194-

-------
                                                          Section  5L


Also used are  support-coated open tubular  (SCOT)  columns,  where a layer
of support  (e.g.,  Celite)  is adsorbed on the  tubing wall and a liquid
phase is adsorbed  on the support.  SCOT columns have increased capacity,
wider tubing (ca 0.02 inches), and faster  flow rates (4-10 ml/minute),
and dead-volume connections are less critical than  in a WCOT column.
Sample spliting is often used but not required (sample normally.
<0.5 ul).   Capillary columns are expensive and require good technique
and instrumentation, but they are invaluable  for  separations requiring
a large number of  theoretical plates.  Figure 5-Z shows a  separation
of 30 pg levels of chlorinated insecticide standards on a  commercially
coated glass.capillary column (Supelco).

See Reference  (65)  for a review of capillary  GC.  Capillary columns
have been mostly used in pesticide analysis for the" GC-MS  identification
of PCBs and chlorinated pesticides (Section 10L).   Retention data were
reported for 60 organophosphorus pesticides and for 27 chloro-, bromo-,
and nitrophenols on SE-30 capillary columns (36).   The combination of
an OV-101 capillary column with a pulsed mode EC  detector  was evaluated
for quantitation of lindane, and a minimum detectable amount of 1 pg
and linear  response to 2.4 ng were found (66).
     Figure 5-Z.
Capillary column chromatogram of OC1 pesticides.   Column:
SE-54, 30 m.   Column temperature:  2 minutes initial
hold, 200°C to 270°C @ 8°C/minute, final hold @ 270°C.
Linear velocity:   42 cm/second, hydrogen.  Detector:
electron capture.   Attenuation:  128.  Range:  1.
Sample size:   0.1  yl.   Split ratio:  67:1.  Sample
weight:  30 pg each.
                                    1011
                                      i4ls I6
I   I   I
468
    Mln.
                                     I
                                    10
                     I    I
                     12  14
                             1. or-BHC
                             2. 0-BHC
                             3. T-BHC
                             4. 5-BHC
                             5. Heptachlor
                             6. tt-Endosulfan
                             7. Aldrin
                             8. Heptachlor epoxide
                             9. /3-Endosulfan
                            10. o,p'-OOE
                            II. Dieldrin
                            12. Endrin
                            13. p.p'-ODD
                            14. Endrin aldenyde
                            15. p,p'«DOT
                            16. Endosulfan sulfate
                                          -195-

-------
                                                     Section 5M

   Sixty meter wall coated SE-30 columns have been used for the routine
   analysis of low- and sub-picogram levels of o'rganochlorine pesticides
   in river and drinking water.  The high resolution allowed analysis of
   extracts without prior column liquid chromatography cleanup.  Over a
   three month period, 60 mixed standards and 220 sample extracts were
   injected into one column without a loss in column efficiency (67).
   Organophosphorus pesticide and metabolite residues in spinach extracts
   were separated on a 25 m x 0.3.mm column with diethylene glycoladipate
   (68).  s-Triazine residue mixtures in environmental samples were separa-
   rated and determined more successfully on a Carbowax 20M capillary
   column than on packed columns.  The detection limit was 50-70 pg with
   an alkali flame ionization detector (69).
5M   SENSITIVITY OF THE GC SYSTEM

   For analysis of pesticides in environmental media, concentration of
   residues  is  commonly in the ppb or ppt range.  High sensitivity of
   GC detection is, therefore, an obvious requirement.  This is not usually
   so important a factor in a laboratory primarily oriented to enforcement
   of statutory tolerance levels in agricultural commodities or foodstuffs,
   since tolerance levels are usually set in the ppm range.  For analyses
   of environmental media, the various electrical, gas flow, and tempera-
   ture parameters must be optimized to produce a peak at least 20-50%
   full scale deflection (fsd) (with minimal baseline noise) from injection
   of 50 pg aldrin on one of the four recommended columns (Subsection 5L)
   connected to an electron capture detector.  Other sensitivities
   (1/2 fsd) should be approximately as follows:  0.5-1.0 ng ethyl parathion
   for the FPD (P Model), 25 ng diazinon and 35 ng parathion for the Coulson
   conductivity detector (N Mode).

   The foregoing subsections survey sensitivities reported for the other
   pesticide GC detectors.   When adjusting parameters to achieve.optimum
   response, it should be recalled  that signal to noise ratio is a more
   meaningful definition of sensitivity than is peak height alone.   It
   has been found in many instances that a significant improvement in
   sensitivity (and concurrently in column efficiency)  can be achieved
   by simply lowering the carrier gas flow rate.

   A sample extract volume of 10 ml from a 5 gram sample contains  the tissue
   equivalent of 0.5 mg/yl.   A 5 yl injection of this extract (2.5  mg of
   sample)  into  an electron capture detector should easily produce  quanti-
   fiable peaks  at pesticide concentrations of at least  0.1 ppm provided
   sensitivity is adequate and attenuation is appropriately adjusted.  The
   high sensitivity capability of the chromatograph should be utilized by
   optimization  of parameters to permit operation at low output  attenuation.
   It is poor practice to operate the electrometer at high attenuations
   (10  x 32 or 10 x 64 on the Tracer MT-220)  while adjusting standard and
   sample concentrations  to  fit  this attenuation  range.   With a  new detector
   foil, high attenuation may be necessary,  but in general this  practice
                                      -196-

-------
                                                     Section 5N


   although giving chromato grains with a stable baseline,  requires injections
   of relatively high sample concentrations to produce quantifiable peaks.
   This leads to more rapid contamination of the column and detector than
   would result from injection of less sample material, a consideration
   that is particularly important when injecting the 15% ether-petroleum
   ether Florisil column eluate from a fat sample (Subsection 9A) .  If the
   instrument is functioning properly, it should be possible to have a
   noise level not exceeding 2% full scale at a low signal attenuation
   (10 x 8 or 10 x 16).

   It is important to distinguish between the terms sensitivity and limit
   of detection.  Sensitivity is the amount of compound necessary to obtain
   a certain response from an instrument under a given set of instrument
   parameters.  At maximum useable sensitivity, the response (e.g., peak
   height) for the compound should be at least twice the response value
   of the noise (70).  Sensitivity can be expressed as the absolute amount
   of compound providing the defined response or in relative terms, such
   "as peak height or area for a given weight of compound.  Limit of de-
   tection is the concentration of pesticide above which a given sample of
   material can be said, with a high degree of confidence, to contain the
   chemical being analyzed by a definite, complete analytical procedure
   (71).  The value depends upon the pesticide and the substrate and is
   expressed in relative units such as ppm or ppb (see also Section SB).
5N   QUALITATIVE ANALYSIS

   In. analyzing a sample extract, the first step, after appropriate cleanup
 •  and concentration, is to run a preliminary chromatogram.  Assuming. the
   chromatography system is operating under the type of control already
   discussed (e.g., the actual column temperature is known from the RRT^
   £,£*-DDT), relative retention| data can be related to tables [EPA PAM,
   Subsection 4, A, (6)] for the particular column and temperature to make
   tentative peak identifications.  If data indicate  one or more probable
   pesticide peaks, proper standard mixtures are selected and quantitation
   is carried out as described in Subsection 50.  Confirmation of peak
   identity is obtained by chromatography on alternate columns and /or an
   alternate selective detector, or by another chromato graphic (e.g., TLC)
   or non-chromatographic procedure (Section 10).  In order that some com-
   pounds are not missed, it is obviously important to allow the chromatogram
   to run for a sufficient -time for all possible pesticide peaks to elute
   and be detected.

   Both absolute and relative retentions have been used for qualitative
   analysis of pesticides.  Absolute retention is the actual time between
   the injection of the sample and the elution of the peak.  On a chromato-
   gram, the measurement is usually made in millimeters between the
   injection point or the front of the solvent peak to the maximum of the
   peak of interest (distance x, Figure 4-C, Section 4).  Conversion of
   retention to minutes is easily made if the chart speed is known.  With
                                       -197-

-------
                                                   Section 5N

 detectors such as the microcoulometer that do  not  respond to the  solvent,
 the injection point must be manually or electrically marked to  serve as
 a reference point,  and this must be done with  accuracy.

 Relative retention ratio is the ratio of the absolute  retention of  the
 compound of interest to that of a reference compound,  most commonly
 aldrin or ethyl parathion.   For peaks that elute before  the reference,
 the relative retention time will be less than  1.0;  for those that elute
 after  the reference,  the relative retention time will  be greater  than
 1.0.   When reporting relative retention data,  the  absolute retention time
 and relevant instrumental parameters should be given.

 The relative retention ratio is far more reproducible  than the  absolute
 retention value since only  the column temperature will influence  the
 former.   Absolute retentions can vary slightly from day  to day  or even
 from hour to hour.   The reference pesticide may be  chromatographed  just
 before or just  after the sample,  or it can be  added to the sample so
 that its  peak will  appear on the same chromatogram.  This  latter  approach
 is  preferred if the sample  is known to contain no compounds  producing a
 peak with the same  retention time as the added reference compound.  In
 addition  to relative  retention,  peak geometry  (shape)  is often  an additional
 useful aid in comparison of sample and standard chromatograms.

 Although  confirmation will  be treated in detail in  Section 10,  some
 comments  pertaining to  compound identification will be made here.  The
 most common single  factor in failure to properly identify  a pesticide
 is  the use of only  one  GC column-.   It is  impossible to be  sure  a  given
 column has separated -all pesticides  present in an unknown mixture, and
 if  this does  occur  it is the result  of an extreme case of  good  luck.
 Reliance  on a single  column is  totally unacceptable and will usually
 lead to worthless analytical data, both qualitative and quantitative.
 If  two columns  are  to be used,  they  should  be judiciously chosen to be
 entirely  different  in their elution  patterns.  Complementary pairs of
 columns include OV-17/QF-1 with OV-210, and SE-30 with DECS.

 Elgar  (72)  ingeniously illustrated this point by demonstrating that when
 two similar columns are  used and  the  relative retention ratios for a
number of pesticidal  compounds are plotted  on respective axes, the
 points fall on a relatively  straight  line with little scatter in evidence.
 Conversely, when two  dissimilar columns are used, the plotted points
 show a wide scatter,  enhancing the probability of reliable identification.
Figure 5-A,A  shows the plots of three  column pairs for 17 pesticidal
 compounds detected by electron capture.  A is the plot of 10% DC-200 vs.
5% DC-200/7.5% QF-1, B is 10% DC-200 against 3% DECS, and C is 5% OV-llO
against 1.5%  OV-17/1.95% QF-1.  It will be observed that the RRT points
plotted in A  cluster to an extent that a fairly straight line is repre-
sented by the plot.  Plots B and C, on the other hand,  show a! very wide
scatter, indicating that either of these two pairs is an excellent choice
for complementary columns.
                                   -198-

-------
 2
•H

 &
1-1
 o
 u

 a>

 8
                                                                                Section 5N
 u
 id

s
 o
•H
4J
 CO

 &
O

 *

c
                                                                               (•3«OM.)tt2-AO«S
 r-l
 m

 a
 8
 ca
 o
 •H
 •H
 •P
 s
 01
 -P
 o
 i-H
 ft)
                (-D.OOZ) oozoa °/0 oi
01
                                                       -199-
EH

-------
                                                     Section 50


   High column efficiency is a distinct advantage for compound identifica-
   tion in that pesticides will be well resolved from each other and from
   non-pesticide artifacts coextracted from the sample substrate.  In
   addition, operating parameters must be adjusted to produce the most
   decipherable chromatograms.  For the columns recommended in Subsection
   5L, the oven temperature should be set so that £,£'-001 elutes in 16
   to 18 minutes with a carrier gas flow of 50-80 ml/minute.  The recorder
   chart speed is set to permit adequate peak spacing and a total retention
   distance such that an absolute measurement error of 1.0 mm will corres-
   pond to an insignificant relative error.  These precautions will help
   assure good peak resolution and precise retention measurements.  -

   A computer-plotting program has been described that can serve as an
   aid in qualitative analysis of pesticide residues (73).  Chromatograms
   are reproduced with corrected baseline drift and solvent peak elimination,
   and two or more chromatograms can be presented in a three-dimensional
   view to facilitate rapid visual comparison to determine whether there
   are differences in the characteristics of individual peaks (sample or
   standards) between chromatograms (see Section 50k).


50   QUANTITATIVE ANALYSIS

   a.   Introduction•

       Quantitation of pesticide residues known to be present in the sample
   from relative retention data and various confirmatory procedures is
   carried out by comparison between the size (height or area)  of the peak
   for each pesticide in the sample and the size of a peak from a similar,
   known amount of each compound injected under the same GC conditions
   just before and/or after the unknown sample.  Only one standard con-
   centration is required for each unknown if injections are made at concen-
   tration levels providing linear detector response.   This procedure is
   known as the external standardization method.

   The exploratory chromatogram of the sample extract  used to obtain
   relative retention data will provide a tentative indication to the
   analyst of the proper standard mixture to be used.   This mixture should
   contain the pesticides of interest  at proper concentration levels  to
   fall within the linearity range of  the detector and  also to  produce
   peaks comparable in size to  those obtained from the  sample chromatogram.
   Injection of the standard mixture may show that  additional dilution of
   the sample extract is required to produce peaks  of  the higher concentra-
   tion pesticides comparable to those from the standard mixture.   If
   several standard mixtures are available at different concentration levels,
   selection of one closely approximating the unknown will facilitate the
   analysis (Subsection 50g).   It should be emphasized  again that  accurate
   quantitation is not possible unless standards  are prepared and  maintained
   properly and replaced on schedule.
                                  -200-

-------
                                                   Section  50


The quantisation'of complex commercial mixtures  such  as  the pesticides
chlordane and toxaphene is a difficult problem because of  the  inability
to obtain a match between the chromatograms of real samples arid
standards.  This problem is discussed for PCB  quantitation in  Section 9A,Gc.
In technical chlordane, over 45 components have  been  identified in the  .
electron capture gas chromatogram  (Figure 5-A.B)  (74).   Toxaphene,
which is one of the most widely used pesticides  in the USA, is even
more of a nightmare since it contains over 175 components ! (Figure 5-A,C),
Both of these pesticides are more  complicated  than FCBs  in that they
contain compounds with different skeletal structures  and are more
susceptible to environmental and biological alteration.
     Figure 5-A,B.
Reconstructed total ion chromatogram of
technical chlordane.
         1.00
                        2.00             3.00            4.00

                   RELATIVE RETENTION TIME (TO ALOHIN)
                                  -201-

-------
                                                  Section 50
     Figure 5-A,C.  Electron capture gas chromatogram of toxaphene.
b.  Comparison of External and Internal Standardization

    Internal standardization is a widely used, general analytical and gas
chromatographic technique which, however, is not recommended for multi-
residue pesticide determinations.  Since multiresidue methods can detect
and measure a large number of different compounds, choice of a suitable
standard with appropriate structural and chromatographic properties in
terms of all compounds to be quantitated would be an impossible, or at
least a very difficult task.  Response calibration for all compounds
of interest vs. the internal standard would be a lengthy process and
would require frequent checking.  To determine the amount of internal
standard to add, a preliminary analysis of samples with unknown histories
and compositions would be necessary.  Many samples require gas chroma-
tography at several dilutions to quantitate all residues, so different
quantities of internal standard would be required.  Detector response to
sample coextractives further complicates the choice of an internal
standard.  These and other disadvantages dictate against the use of
internal standardization except in special cases, such as in the determina-
tion of residues of one or a small, definite group of pesticides.
                                  -202-

-------
                                                  Section 50


External standardization has the advantage that calculations are based
on a comparison of the same compound in the standard solution and in
the unknown, and no response or correction factor is required.  Accuracy
and precision depend upon the ability to inject exact amounts of samples
and standards reproducibly, having all instrumental parameters under
tight control so that data are comparable from run to run and determina-
tions are conducted within the linear concentration range.

A recent study (52) concluded that generally unrecognized systematic
errors were inherent in the accepted procedures of bot^i external (direct)
and internal standardization GC.  For example, it was found necessary to
consider,both the volume injected and the concentration of the standards
in the direct method; plotting peak area vs. quantity (gram, mole) is
not sufficient unless the concentration is stated and the volume is kept
constant.  The internal standard method was found to be not necessarily
independent of the volume injected, concentration of standard, or the
effects of temperature and gas flow on instrumental sensitivity.  The
relevance of these conclusions to pesticide analysis has not been studied,
and the procedures recommended in this section are based on the best
current practice of experienced residue analysts.


c.  Calculation Procedure

    The calculation method for any GC analysis where an unknown peak is
compared to a peak resulting from injection of a standard of known con-
centration is given below.  This method is equally applicable to external
standardization procedures based on comparison of standard and unknown
liquid chroma tography peaks or thin layer chromatography spot sizes.
The equations used are:
(eq. 1)

ng or pg of residue
represented by
sample peak
 sample peak size
standard peak size
x ng or pg standard injected
 (eq. 2)
       residue concentration
              (ppm)
      ng in sample peak
      mg sample injected
                           or

       residue concentration =
              (ppb)
      pg in sample peak
      mg sample injected
                                   -203-

-------
                                                  Section 50

The following is a specific example of a calculation based on typical
data on a Report of Interlaboratory Check Sample (Table 2-1):
DATA:
     Sample'Extract Data ,
     gms in Final Vol.
     ml in Final Vol.
     yl Injected
     mg Injected
     Peak Ht  or Area (mm)
     Attenuation
                       3.0
                      25
                       5.0
                       0.60
                     145
                       102 x 32
     Reference Standard Data
     yl Injected                 6.0
     ng Injected                 0.0.60
     Peak Ht  or Area (mm)     120
     Attenuation                 10  x 32
CALCULATIONS:
     wt standard injected = 6.0 yl x 0.010 ng
                                                      ng
from  (eq. 1)
   145 mm
   120 mm
x 0.060 ng - 0.072 ng in sample peak
   sample extract concentration
                                  3.0
                                     -_
                                0.12 g/ml or mg/yl
     weight of sample injected = 0.12 mg. x 5.0 yl = 0.60 mg
                                      yl

 (The actual chromatogram should be labeled:
          3 g/25 ml x 5 yl s 0.60 mg injected)
from  (eq. 2)
   on
   0.60 mg
              0-12 ppm residue  (result)
                   r
                                    -204-

-------
                                                  Section 50


Note that the analytical data indicate a constant attenuation so that
sample and standard peak sizes do not have to be corrected.  When
calculating the sample extract concentration, careful consideration
must be given to whether the final solution contains the entire
original sample or whether one or more aliquots that were taken during
the sample preparation procedure have to be accounted for by a dilution
factor.

The results of most actual residue analyses are not corrected for the
percentage recoveries determined from spiked samples analyzed along
with samples, although this can be done.

Some'general comments on calculations are in order in a quality
control manual.  All mathematical operations should be checked at least
twice, whether they are done with or without an electronic calculator.
It is very easy to occasionally press the wrong calculator key or not
to press a key hard enough to register.  If something appears wrong with
the results of an analysis, the first thing the analyst should do is to
check calculations, and then ask an independent person' to go over them.
It is not uncommon for a person to make the same simple calculation
error twice.  If the calculations are correct, the next most profitable
action is to prepare new standard solutions and a new standard calibra-
tion curve for the determination in question.

d.  Reporting of Results (see also Section 3£)

    The method for reporting analytical results will often differ from
laboratory to laboratory, but in general, the following should be stated:

    (1)  The compounds or classes of compounds being sought.

    (2)  Other related or important compounds of interest that were
         detected or found absent.

    (3)  The limit of detection for each pesticide, as well as its
         degradation products and metabolites.

    (4)  Recovery values and whether results were corrected for
         recovery.

    (5)  The basis for selection of the analytical procedure and any
         modifications of an accepted procedure.

    (6)  Confirmatory methods.

Results should be reported in appropriate ppm, ppb, or ppt units, and
the basis for reporting should be clear, i.e., dry weight, wet weight,
or fat- or extractable-lipid basis.  Any drying methods should be
                                    -205-

-------
                                                  Section 50


described.  If replicates are run, the individual results, the mean, and
a statistical treatment of precision  (Section 2K) should be presented.

Pesticide residue analytical data are generally reported as ppm (parts
per million), ppb (parts per billion), and ppt (parts per trillion).
Converting these terms to weight expressions, we have

          ppm  s  micrograms per gram or nanograms per milligram

          ppb  »  nanograms - per gram or picograms per milligram

          ppt  **  picograms per gram


Residues in water are quite commonly expressed as micrograms per liter,
which is equivalent to ppb.  On rare occasions, a laboratory may choose
to express a water residue result in grams per liter, but the value
becomes quite cumbersome, i.e., 5 x ,10"7 grams per liter as opposed to
the more convenient 0.5 micrograms per liter.
                                       10~3 liters

                                       10~6 liters
The following is a summary of units frequently used in pesticide analyses:

     Vg  »  10   grams          ml

              -9
     ng  -  10   grams          yl

              -12
     pg  *  10    grams

     ppm «  parts per million =• yg/g, yg/ml, ng/mg, or pg/yg

     ppb »  parts per billion = yg/1, ng/g, ng/ml, or pg/mg

     ppt «  pg/g, pg/ml (ppt is used frequently in other books to mean
                        "parts per thousand").

e.  Detector Linearity

    Linearity may be defined as the range of concentration over which a
detector maintains a constant sensitivity.  If a detector has a
linearity of 103 and the detector sensitivity for a certain pesticide
is 1 pg, the upper limit of analysis is 1 ng.  If the detector sensi-
tivity is 0.1 pg, the pesticide can be determined only up to 100 pg.
Sensitivity is affected by the molecular structure and retention time
for a particular pesticide under given GC conditions.
                                   -206-

-------
                                                  Section 50


Quantisation must be performed within the linear response range of the
GC detector.  Each detector has its own characteristic linear range under
the prevalent conditions of operation.  For a given detector, the linear
range varies somewhat.between pesticides.  For example, the isomers of
BHC exhibit a more restricted EC linear range than £,£f-DDT.  The nickel
EC detector operated in the DC mode exhibits a far more restricted
linear range than the tritium detector.  Lindane concentrations above
600 pg may result in nonlinearity with the tritium detector, whereas
the linear cut-off for this compound may occur at approximately 250 pg
for the nickel detector.  Section 4,A,(3),III of the EPA PAM compares
typical linearity curves for various pesticides with these two de-
tectors.

Before any attempt is made to try quantitation with a new or newly
renovated detector, linearity curves should be constructed for the *
pesticides of interest under the prevalent operating conditions.
Frequent checks should be made to insure continued operation within
acceptable concentration ranges.  Knowledge of the linear cut off
point will preclude such error as injecting 1 ng of aldrin and expecting
it to fall within the linear range of the Tracer EC detector in the DC
mode.

Calibration curves are constructed by injection of serial amounts of
a pesticide and calculation of the peak height of each peak.  Peak
height is the perpendicular distance from the peak maximum to the
baseline (Figure 5-A,D, distance CD).  The linear range is observed
by plotting height vs. amount of pesticide injected.

    Figure 5-A,D.  Quantitation by Peak Height Method
                                 Psolt Height -CD
                                    -207-

-------
                                                  Section 50
f.  Sensitivity Control
    When analyses are performed in the ppb or even ppt  (parts per trillion)
range, electrometer attenuation is required to attain maximum sensitivity
consistent with an acceptable baseline noise level.  Electrometer operation
at high sensitivity levels is good practice even when substrates contain
high concentrations of pesticides.  By serially diluting the final extract
and operating at high sensitivity, the possibility of exceeding the
linear range of the detector is greatly reduced and, therefore, the
quantitative error possibility is reduced.  Occasional  instances have
been observed in the EPA Quality Control Program where  operators have
set attenuation controls at low sensitivity and Injected media containing
massive concentrations of pesticides.  This was readily discernable by
malformed chromatographic peaks.  Had attenuation been  set for high
sensitivity, chromatographic peaks would have gone off-scale, requiring
serial dilutions of final extract to a pesticide concentration within
linear boundaries.

Electron capture electrometer attenuation should be adjusted to obtain
a minimum sensitivity level equivalent to a 50% fsd peak from the
injection of 50 pg or less of aldrin.

There is no objection to using different instrumental attenuation
settings for standards and samples provided that concentrations are
within the linear "detec.tion range and checks are made to insure that
the attenuator is truly linear.  A sample should produce the same peak
when diluted by 10 as if the original sample were run at an attenuation
increased by 10.  An output attenuation setting of x!6 on the Tracer
MT-220 chromatograph electrometer is convenient to assure operation
within the range of the detector.


g.  Injection Volumes and.Standards

    As described in Subsection 5J, injection of small volumes such as
1-3 yl can lead to large relative errors and should be avoided.   A
common reason for low injection .volumes is to provide on-scale peaks of
sample for reference against peaks from a particular single standard
or standard mixture.  To circumvent this problem,  standards should be
made up at several concentrations, each succeeding level being twice
the previous concentration.  For example, a typical set of three
standard mixtures in pg/yl for electron capture GC might be:
                               -208-

-------
                                                  Section 50
                 Mixture
Al
Mixture
A2
Mixture A.
Lindane
Aldrin
Dieldrin
£,£'-001
p,p'-DDT
5
5
10
10
20
10
10
20
20
40
20
20
40
40
80
Injection volumes should be controlled within 3-8 yl, and sample and
standard injection volumes should agree within ±25% (i.e., within
2-3 pi from a 10 yl syringe).
                                        r ••   •  ..
Working standards must, of course, include all compounds of interest to
a particular laboratory.  It may be necessary to run several sample
extract chromatograms of various concentrations and/or injection volumes
to achieve reasonable concurrence of peak size with those of the
standard mixture(s) if pesticides are present in the sample at widely
different levels.  Peaks of standards should never be distorted.  If
this occurs, the injection must be repeated and the cause of distortion,
if it persists, must be determined and corrected.

It should be unnecessary to reiterate that accuracy of analysis is
limited by the accuracy of the standard quantitating solutions.  Con-
sistently high recovery values on an interlaboratory check sample
strongly indicate that weak standard -solutions are being used by the
laboratory in question, while consistently low values indicate the
probability of overconcentrated standards.


h.  Optimum Peak Heights

    The ideal range of peak height response is 20-60% fsd, with a minimum
acceptable height of 10% fsd.  Peak heights of the sample and standard
should vary by no more than 10% for highest accuracy and at most by
25%.  If all GC modules are operating properly and parameters are
optimally set, the 10 or 20% fsd minimum peak requirement will cause
no problem in terms of attainable sensitivity when standard procedures
and concentration steps as given in the EPA PAM are followed.  If
sample peaks are too low (<10% fsd), the solution should be further
concentrated or a larger amount injected.  Injection of samples that
are too large can cause loss of compounds in the solvent peak in some
cases, e.g., HCB on an OV-17/OV-210 column.

The requirement of referencing samples against standards differing by
no more than 25% in peak height causes no inconvenience when the concept
of different standard concentrations (Subsection 50g) is followed.  This
point is important, even though one is working within the linear range
of the detector, because of minor variations in GC response, primarily
                                 -209-

-------
                                                    Section 50


 arising from instrumental  sources  and/or from small injection errors.
 Figure 5-A,E illustrates the  potential error that is possible from a
 peak height variation of as little as 3 mm,  when attempting to quantitate
 a 13 mm sample peak against a 130  mm standard peak.  This is shown in
 (A) on the left.  The total deviation results in a relative error of
 23%.  Oh the other hand, when the  extract is further concentrated down
 to an assumed final volume of 400  vl, the. height of the sample peak
 is increased by a factor of 10 to  130 mm as. shown in (B)  on the right.
 The same 3 mm response variation at this level will result in a final
 relative error of only 2.2%,  a very acceptable value.
 Figure 5-A,E.
Illustration of potential hazard of quantitating by
comparison of small sample peak against  large  standard
peak.  A 3 mm peak height shift is assumed.  Initial
sample size 1.67 grams; injection volume 5 yl.
                      Standard
                       0.2 ng
                                     Samp. ln|Kt. 2: p.p-b.3 0.2Kl33x«OO x „ 4 _ , g,
                                                     Deviation  0.21
                                                      ar 2.2% error
i. Standard Curves

   GC calibration curves prepared by injection of standards are  of  little
 direct use in residue quantitation.  Such curves are not valid  for
 extended periods of time, as is the case for other analytical methods
 such as spectrophotometry, and so their preparation is an unnecessary,
 time_consuming chore.  A GC standard curve constructed for an EC
 detector on a given day at 9 a.m. may well be worthless by the  afternoon
 of the same day, or even the same morning, if high lipid extracts
 causing a rapid depression in peak response are repeatedly injected.
 If such peaks are referenced against a standard curve prepared  when
 response was normal, erroneously low results will be obtained.
                                -210-

-------
                                                  Section 50


The proper procedure for quantisation is to intersperse standard mixture
injections throughout the workday with sufficient frequency to signal
the onset of response fluctuations, and quantitative referencing is
made against the interspersed standards.  For maximum accuracy, injection
of an unknown sample would be bracketed between standard injections
made immediately before and after the sample.


j.  Methods of Peak Measurements

    Both peak heights and peak areas are extensively used for calculations
of residue amounts.  The preferred method of calculation depends on the
shape of the chromatographic peak.  Peak height is recommended for measure-
ment of very small peaks or tall, symmetrical, fairly narrow peaks (<10 mm
at the base) that have no obscuring overlaps.  These are characteristic
features of most pesticide peaks from an efficient GC column, especially
those that elute early.  Accurate calculation of the area of such peaks
would be difficult because the slightest measurement errors in the narrow
width would be magnified in the subsequent area calculation.  Peak area
as estimated from peak height x width at half height is recommended for
separated, symmetrical, and fairly wide peaks.  Tfiangulatjon is used for
separated, unsymmetrical peaks or peaks on a sloping baseline.  Triangula-
tion should never be used on very narrow peaks.  Extreme care must be
taken in the construction of inflectional tangents in all measurements.
Measurements with a mm ruler containing fine division markings can be
made to the nearest 0.1-0.2 mm if care and patience are exercised.

When peak heights are used, the assumption is necessarily made that
operating parameters are closely controlled and retention times are very
reproducible.  Two consecutive injections of the same amount of compound
should ideally result in two peaks with exactly the same retention time,
width, and height.  If chromatographic parameters (particularly column
temperature) are not under strict control, the second peak may instead
elute later or earlier than the first, resulting in a wider or narrower
peak.  However, the peak areas should be the same in both cases.  For
this reason, peak area or peak height x retention time is considered by
many operators to be more reliable than peak height alone since slightly
shifting peak positions will not be so important.

Figures 5-A,D, 5-A,F, and 5-A,G illustrate the peak height, peak height x
width at half height, and triangulation measurement methods, respectively.
The two right-hand peaks in Figure 5-A,D are measurable by the peak height
method because their overlap does not obscure the height of either
peak.  Peaks on a sloping baseline but too narrow to be triangulated can
be measured by the peak height method.  The distance KI would be used as
seen in Figure 5-A,G.  Peaks can be widened by. using a faster recorder
chart speed.  Use of the planimeter is an alternate method for measuring
unsymmetrical peaks, peaks on a sloping baseline, or total area of a
series of incompletely resolved peaks.  Precision will be improved by
tracing the peak at least twice and taking an average value.
                                  -211-

-------
                                                                 Section  50
Figure  5-A,F.   Quantitation by Peak Area Method


   FIGURE 5-Q QUANTITATION BY PEAK AREA METHOD
                   Figure 5-A,G.  Quantitation  by
                                   Triangulation Method
                                                              FIGURE 5-R  QUANnTATION BY TRIANGUIATION
                                                                               METHOD
Quantitation of peaks indicating heavy electrical overshoot (Figure 5-A,H
part A)  or nonlinear response (part B) will lead  to unreliable quantitation.
Peak overshoot is influenced  by foil contamination and by improper EC  detector
polarizing voltage (Subsection 5Cb).


     Figure 5-A,H.  Examples  of Gas Chromatographic Peaks
                 FIGURE 5-S   EXAMPLES OF GAS CHROMATOGRAPHIC
                                        PEAKS
                    Distorted
                     Peaks
Non-linear
 indication
 Linear
Indication
                                           -212-

-------
                                                  Section 50


Automatic (disc) integration is a convenient, accurate procedure that
can be used in place of manual procedures whenever baselines are steady,
but it is less reliable and simple when sloping baselines or peaks with
shoulders occur.  This method has been mostly used for calculation of
late eluting peaks and multicomponent chemicals that elute over a long
period of time (Strobane, PCBs, toxaphene).  In the absence of an inte-
grator, chromatograms, especially of these complex mixtures, have been
quantitated by cutting out the peaks on .the recorder chart and weighing
the paper.  This method, although time consuming, can yield excellent
results if care in cutting is taken and if the paper is uniform.  Since
Xerox paper is especially uniform, recorder charts can be copied and the
copy cut and weighed.

Gaul (75) compared five methods for quantitation of-aldrin, heptachlor
epoxide, and dieldrin with a tritium electron capture gas chromatograph.
The methods were disc integration, triangulation, peak height times width
at half height, retention time multiplied by peak height, and peak height.
No significant differences were found among the five methods in this
study.  The same author described methods for properly placing baselines
for typical overlapping and unsymmetrical gas chromatographic peaks,
and suggested procedures for quantitating multipeak chrpmatograms of
pesticides that are mixtures of isomers, e.g., DDT, BHC, chlordane,
and toxaphene.  Poorly resolved peaks and sloping baselines present the
greatest challenge in terms of accurate quantitation, and an experienced
analyst must exercise judgment to quantitate the peaks properly.  If
necessary, improved resolution of peaks and flatter baselines may be
sought, through the use of other cleanup procedures, GC columns, or changes
in operating conditions.

k.  Integration and Automation

    Acquisition and interpretation of data are the final steps in qualita-
tive and quantitative analyses.  How this is done can in large measure
affect accuracy.  The trend today  is more and more towards automatic
or computer assisted data acquisition and treatment.  Computer acquisition
is almost a necessity in mass spectrometric analyses, but it is only
beginning to make in-roads into the areas of gas and liquid chromatography.

Digital computer and integration systems are available today that perform
baseline detection, baseline correction, area integration, area allocation
of fused peaks, and postrun calculations.  They can store and retrieve
GC data and visually display the chroma tograms.  Expansion or contraction
of either the attenuation or time axis is sometims possible after the
chromatogram is taken.  This is very useful with a wide-range linearity
detector, as it can replace multiple injections of different sample con-
centrations.  These systems range in price from approximately $10,000 to
over  $100,000 per unit.  When properly operated they provide very fast
and accurate quantitation of chromatograms, but results may not be reliable
with  complex chromatograms having narrow peaks, merged peaks, variable
                               -213-

-------
                                                   Section 50

 retention times, unidentified compounds,  and/or noisy and variable base-
 lines.   Despite the publication of computer methods tolerating these
 problems (76),  the good judgment of an experienced analyst may be better
 in such instances.  The increased use of  computer systems in the future
 is expected as  technology is improved and prices are decreased.

 All systems use logic to interpret the chromatograms.   The unique logic
 incorporated by each manufacturer in its  system must be understood if
 the operator is to properly instruct the  system how to treat the data
 and then properly understand the data obtained.   The system can only
 follow  the instruction's by the operator,  and if those instructions are
 incorrect,  the  data returned to the operator will be useless.

 The capabilities of these systems and the means of handling the data
 differ  widely,  but, in general, the systems perform the same functions.
 The systems with built-in-printer-plotters plot the chromatogram using
 the digital data generated for the integration system.   Generally,  these
 digital chromatograms include peak absolute retention times and system
 notations to help understand the system's interpretation of the data.

 All systems interpret the digital data, evaluating when the chromatograms
 are in  the baseline br on a peak'and resolving fused peaks.   Once the
 peak is confirmed,  the raw area counts and/or peak height and peak re-
 tention time are- stored for later processing.   Other logic and instructions
 from the operator are used to "draw" the  baseline under these peaks and
 discard the area or height below the chromatographic peak.   Usually these
 processed data  are the only data stored by the system for use in later
 calculations.   The system calibrates the  peak areas and/or peak heights
 using the values assigned as the standard by the operator.   Then all
 quantitation calculations are performed using the specified calculation
 method  (area, area percent^internal standard, or external standard).

 The report  formats  of the different  systems  are  just as  different as
 their logic.  In general,  however, the elution time, peak area  and/or peak
 height,  and the concentration of the compound are reported along with its
 name, if known.   Naming of peaks is  generally performed  using the relative
 retention time  system specified by the operator  (e.g., aldrin » 1.0Q_,
£.,£.'-DDE »  1.00,  or ethyl parathion  * 1.00).

 The microprocessor  and minicomputer  systems  usually also have the capa-
 bility of handling  a programming language  such as BASIC.  The added
 capabilities of these systems  are numerous.  Postrun calculations can
 be  tailored to  the  needs  of  the data user.   Some  examples of the use of
 programming languages  include:   (a)   doing statistical calculations on
 the results  of  several runs;  (b) modifications of report  format;  (c)
 analysis of data for a particular criterion (i.e., is the £,j3'-DDT level
 greater  than tolerance?);  and  (d) the  handling of standards calibration,
 and other tasks,  on automated  runs.   The limits of  the systems with
 language programming are  set by the  capabilities of the programmer
 using the system.
                                  -214-

-------
                                                     Section 5P


   The pitfalls of these systems are established by the experience and under-
   standing of the operator.   Accurate analyses of simple chromatograms are
   not difficult to obtain.   Accurate analyses of complex chromatograms are
   quite often very difficult to obtain.   One of the best ways to test the
   system and evaluate the usefulness of  the analytical data is to closely
   examine the chromatograms  and reports  from the system.  The analyst
   should establish if the system drew the baseline in the proper location.
   If not, modifications of the system's  instructions are required.  The
   same is true for the examination of the performance of all parameters
   in the instructions.  If the instructions are not getting the desired
   results, the analyses must be run again after the appropriate instructions
   are modified.  It is important _to double check the system's analyses to
   be sure they are what is desired by the analyst.

   The various approaches for automation  of sample introduction in gas
   chromatography have been reviewed and  a typical autosampler described
   in detail (77).  Besides the advantages of automation and unattended
   operation for large numbers of samples, automated injection systems will
   normally give more precise injection volumes than most operators can
   achieve manually.  Losses of up to 50% of aldrin and dieldrin in a
   commercially automated dry capsule injector were reduced by silanization
   of the capsules (78).

   Although a reliable, automated system applicable to entire pesticide
   analytical procedures at common residue levels has not yet been perfected,
   some progress has been made on automation of residue analyses.  The use
   of Technicon auto-analyzer modules for the automation of extraction and
   cleanup, followed by GC or HPLC determination, has been described and
   tested using alfalfa and string bean samples fortified with organophosphorus
   and carbamate insecticides (79).  A mechanized system for 4-nitrophenol
   and some other phenolic pesticide metabolites in urine was reported (80).
   This system performs acid hydrolysis,  steam distillation, and liquid
   chromatography separation combined with UV absorption determination,
   analyzing one sample every 24 minutes  at 1-6 ppm levels.


5P   REFERENCES

   (1)  Welsh, P., Amer. Lab., p. 36, May (1977).

   (2)  Ettre, L. S., J. Chromatogr. Sci., 16, 396 (1978).

   (3)  Sevcik, T., Detectors in Gas Chromatography, Elsevier Publishing Co.,
        New York, 192 pp. (1976).

   (4)  Hall, R. C.. CRC Critical Reviews in Analytical Chemistry. 7/4),
        323-379  (1978).

   (5)  Farwell, S. 0., and Rasmussen, R. A., J. Chromatogr. Sci., 14, 224
        (1976).
                                  -215-

-------
                                                     Section 5P


  (6)   Cochrane, W. P., Maybury, R. B., and Greenhalgh, R.  G.,  J.  Environ.
       Sci.  Health. B14(2), 197 (1979).                         ~~~~	

  (7)   Karasak,  F.  W., and Field, L. R., Research/Development.  p.  42,
       March (1977).                                      	

  (8)   Siegel, M. W., and McKeown, M., Research/Development. p. 101,
       July  (1977).	

  (9)   Sliwaka,  I., Lasa, J.,  and Rosiek, J., J. Chromatogr.. 172,  1  (1979)

 (10)   Sullivan, J. J., J. Chromatogr.. 87. 9 (1974).

 (11)   Burgett,  C.  A.,  Research/Development« 25. 28  (1974).

 (12)   Maggs, R. J.,  Joynes, P.  L., and Lovelock, J. E., Anal.  Chem., 43,
       1966  (1971).                                      	:	  —

 (13)   Aue,  W. A.,  J. Assoc. Off.  Anal.  Chem.. 5_9_, 909 (1976).

 (14)   Aue,  W. A.,  J. Chromatogr.  Sci..  13, 329 (1975).'

 (15)   Cochrane, W. P., and Whitney, W.,  Advances in Pesticide Science.
       Geissbuehler,  H.,  ed.,  (3)664 (1979)~1

 (16)   Hanish, R. C'., and Lewis,  R.  G., Na.tl.  Tech.  Infrom. Serv. PB-276.
       990,  35P  (1978).                ~~~	;—	

 (17)   Greenhalgh,  R.,  and Cochrane, W. P., J.  Chromatogr.. 188. 305 (1980).

 (18)  Poole, C. F.,  J.  Chromatogr.,'l 118,  280  (1976).           ,

 (19)  Kapila, S.,  and  Aue, W. A.,  J_,  Chromatoer. . 118, 233 (1976).

 (20)  Holden, E. R., and Hill, K.  R., J.  Assoc.  Off. Anal. Chem..  57,
      1217(1974).                                  ;	—~', — ,.

 (21)  Van de Wiel, J.  H., and Thommassen,  P.  J.,  J.  Chromatogr.. 71,  1
       (1972).                                     	— —

 (22)  Aue, W. A.,  and  Kapila, S.,  J.  Chromatogr.  Sci.. 11, 255  (1973).

 (23)  Pellizzari, E. D.,  J. Chromatogr..  98,  323  (1974).

(24)  Currie, R. A., Kadis, V. W.,  Breitkreitz, W.  E., Cunningham, G.  B.,
      and Burns, G.  W. > Pestic. Monit. J.. 1.3(2), 52 (1979).

(25)  FDA Pesticide Analytical Manual. Vol. I., Section 312.
                               -216-

-------
                                                  Section 5P

(26)   Brazhnikov, V, V., Gurev, M. V., and Sakodynsky, K. I., Chromatogr.
      Rev.. 12, 1 (1970).

(27)   Aue, W. A., Advances in Chemistry Series, No. 104. ACS, Washington,
      D.G., p, 39 (1971).

(28)   Cochrane, W. P., and Greenhalgh, R., Chromatographia, £,  255  (1976).

(29)   Kolb, B., Auer, M., and Pospisil, P., J. Chromatogr.. 134,  65 (1977);
      J. Chromatogr. Sci., 15, 53  (1977).

(30)   Burgett, C. A., and Green, L. E., J. Chromatogr. Sci..  12,  356 (1974);
      Spectrochim. Acta B, 30, 55  (1975).

(31')   Attar, A., Forgey, R., Horn, J., and Corcoran,  W.  H., J.  Chromatogr.
      Sci.. 15_, 222  (1977).

(32)   Burnett, C. H., Adams, D. F., and Farwell,  S. 0.,  J,  Chromatogr.
      Sci.. 15_, 230  (1977).                           .

(33)   Sevcik,  J., and Thao, N. t.  P..  Chromatographia.  8 ,  559  (1975).

(34)   Bowman,  M. C., and Beroza, M.,  Anal. Chem.,  40. 1448  (1968).

(35)   Zehner,  J. M., and Simonaitis,  R. A., Anal.  Chem.. 47.  2485 (1975).

(36)  Krijgsman, W., and Van De Kamp, C.  G.,  J.  Ghromatogr..  117, 201
      (1976)  and 131,  412  (1977).

(37)  Varian  Instrument Applications. 11  (1), 15 (1977); Patterson, P.^.,
      Anal.  Chem..  50.,  339 and 345 (197,1).

(38)  Pope,  B. E.,  Rpdgers, D.  H., and Flynn, T. C.,  J.  Chromatogr.. 134.
      1 (1977).

(39)  Anderson,  R.  J.,  and Hall,  R.  C.,  Amer. Lab.,  p. 110, February (1980).

(40)  Wilson, B.  P., and Cochrane, W. P., J.  Chromatogr.. 106. 174  (1975).

(41)  Tracor Instruments Model 700 Hall Electrolytic Conductivity Detector
      Operation Manual 115008A.  pp.  23-24.

 (42)  Moseman, R.,  and Thompson,  J.  F. , personal communication.

 (43)  Bayer, F.  L., J. Chromatogr. Sci.. 15_, 580 (1977).

 (44)  Driscoll,  J.  N., J.  Chromatogr.. 134. 49 (1977).

 (45)  Driscoll,  J.  N., Ford, J.,  Jaramillo, L., Becker, J. H., Hewitt, G.,
      Marshall,  J.  K., and Onishuk, F., Amer. Lab.,  p.  137, May  (1978).
                                   -217-

-------
                                                   Section 5P
 (46)


 (47)


 (48)
 *


 (49)

 (50)

'(51)


 (52)


 (53)


 (54)


 (55)

 (56)


 (57)

 (58)


(59)


(60)


(61)



(62)


(63)
 McLeod, H. A., Butterfield, A. G., Lewis, D.,  Phillips,  W.  E.  J.,
 and Coffin, D. E., Anal. Chem.. 47_,  674  (1975).

 McLeod, H. A., and Lewis, D., J. Assoc.  Off. Anal.  Chem.,  61.
 18 (1978).	 —-

 Baldwin, M. K., Bennett, D., and Benyon, K.  I., Pestic.  Sci.,  8.
 431 (1977).                                     	 —

 Gough, T. A., Chromatographia. 12, 195 (1979).

 Olsavicky, V. M., J. Chromatogr. Sci.. 16_, 197  (1978).

 Ottenstein, D. M.,  and Silva, P. H., J.  Chromatogr. Sci.. 17,*  389
 (1979).                              	fi	   —

 Purcell, J. E., Downs, H. D., and Ettre, L. S., Chromatographia. 8_,


 LeBel, G.  L., and Williams, D. T., J. Assoc. Off. Anal.  Chem., 62,
 1353  (1979).    •	  *—
 Grob,  K.,  Jr.,  and Neukom, H. P., HRC & CC, J. High Resolut.
 Chromatogr.  Chromatogr.  Commun., ,2_, 15 (1979).

 Spencer,  S.  F.,  Amer.  Lab., p. 69, October (1977).

 Analytical Methods for Pesticide Residues in Foods. Department of
 National  Health  and Welfare, Ottawa, Canada, Section 12.3.

 McReynolds,  W. 0., J.  Chromatogr. Sci.. 8_, 685 (1970).
        *
 Analytical Methods for Pesticide Residues in Foods. Department
 of National  Health and Welfare, Ottawa, Canada, Table 8.1.

 Suzuki, M.,  Yamato,  Y.,  and Watanabe,  T., J.  Assoc. Off. Anal. Chem.,
 58_, 297 (1975) ; J59,  1180 (1976) .               	"	

 Rus, V., Funduc,  I., Crainiceanu, A.,  and Trestianu, S., Anal. Chem..
 49_, 2123  (1977).                                          	

 Zweig, G., and Sherma, J.,  Analytical  Methods for Pesticides and
Plant Growth Regulators,  Volume VI,  Gas Chromatographic Analysis,
Academic Press, N.Y.,  1972.

Zweig, G., and Sherma, J.,  editors,  Handbook of Chromatography.
Volume I,  CRC Press, Cleveland, Ohio,  1972.

Watts, R.  R., and  Storherr,  R.  W.,  J.  Assoc.  Off.  Anal.  Chem., 52,
513 (1969).                        	~"~~	 —
                                    -218-

-------
                                                  Section 5P

(64)   Thompson, B., Varian Instrument Applications. 13_(1), 13  (1979).

(65)   Hammarstrand, K., Varian Instrument Applications, 11(2),  8  (1977).

(66)   Rejthar, L., and Tesarik, K., J. Chromatogr., 131» 404  (1977).

(67)   Brodtmann, N. V., and Koffskey, W. E., J. Chromatogr. Sci..  17,
      97 (1979).

(68)   Hild, J., Schulte, E., and Thier, H.-P.,  Chromatographia. 11.  397
      (1978).

(69)   Matisova, E., Krupcik, J., and Liska,  0., J.  Chromatogr.. 173_,
      139  (1979).                            *

(70)   Skogerboe,  R. K., and Grant,  C. L.,  Spectroscopy Letters. 3_ (889),
      215-220  (1970).

(71)   Sutherland, G.  L., Residue Reviews.  10,  85-96 (1965).

(72)   Elgar, K. E., paper presented to joint meeting  of ACS  and CIS,
      Toronto,  Canada, May, 1970; published in Advances in Chemistry
      Series No.  104. ACS, Washington, D.C., Chapter  10,  p.  151 (1971).

(73)  Glaser,  E.  R.,  Silver, B., and Suffet, I. H., J. Chromatogr. Sci..
      15_,  22  (1977).

(74)  Sovocool,  G.  W.,  Lewis,  R. G.,  Harless,  R.  L.,  Wilson,  N. K., and
      Zehr, R.  D.,.  Anal. Chem..  49_, 734  (1977).

(75)  Gaul, J.  A. J.  Assoc. Off. Anal.  Chem..  49_, 389 (1966); Section
      300.6,  FDA Pesticide Analytical Manual.

(76)  Bacon,  G.  D., J.  Chromatogr.. 172. 57 (1979).

(77)  Baumann, F.,  Pecsar, R.  E.,  Wadsworth, B.,  and Thompson, B.,
      Amer. Lab.. p.  97,  October (1976),

(78)  Lines,  D. S., Brain,  K.  R.,  and Ross, M. S. F., J. Chromatogr.. 117.,
      59 (1976).

(79)  Getz,  M. E., Hanes,  G.  W., and Hill, K.  R., NBS Spec. Publ. (U.S.).
      519 (Trace Ore. Anal.;  New Front.  Anal.  Chem.). 345-53  (1979).

 (80)  Ott, D. E., J.  Assoc.  Off. Anal. Chem.. 62, 93  (1979).
                                     -219-

-------
                                    Section 6
                MAINTENANCE, TRDLIBLESHOCfTING, AND CALIBRATION OF

                  THE GAS CHROmTOGRAPH AND DETECTOR SYSTEMS


   The EPA Pesticide Analytical Manual, Appendix I, contains comments on
   the maintenance and repair of instruments primarily intended for labora-
   tories that are part of the EPA or have contractural arrangements with
   EPA allowing them to make use of the electronic repair facility located
   at Research Triangle Park, NC.  The Instrument Shop in RTP is equipped
   to handle repairs, modifications, and calibrations on various gas-
   chromatographs, recorders, and GC detectors.

   A detailed treatment of instrumental servicing and calibration is beyond
   the scope of this Manual.  Some general comments and a few selected
   topics of special interest will be covered, however.


6A   DAILY OPERATIONAL CONSIDERATIONS FOR GAS CHROMATOGRAPHIC INSTRUMENTATION

   (a)  Is the proper carrier gas connected into the system?

        (1)   Is the tank capable of maintaining the desired flow for an eight
   hour work period without going below 500 psi tank pressure?

        (2)   Is the tank output pressure normal (40-50  psi recommended)?

   (b)  What is the detector condition,  temperature,  flow, background current
   and voltage profile?   Is detector set at optimum operating voltage?

   (c)  Is the electrometer operating properly,  and  is  it  zeroed properly?
   Is bucking ability adequate?  What is  the noise  level?

   (d)  Has  the programmer temperature remained constant?   Is the temperature
   limit switch in  a safe position to avoid accidental  overheat?

   (e)  Does the purge system operate smoothly?  Will it be used on  this day's
   operation?   When was  it last checked for leaks?

   (f)  Is the oven damper closed?  Does  it function?

   (g)   Are  all temperature set controllers functioning properly and is  the
   voltage to  the load (heaters)  stable?  Is the oven at proper  temperature?
                                       -220-

-------
                                                           Section 6B


        Has the recorder been checked for proper speed, zero, gain level,
   dead band,  ink supply, and sufficient paper?  How long has it been since
   calibration?

   (i)  Are the septums, 0-rings, and glass injection inserts in good
   condition?   When were they last replaced?

   (j)  Is the pyrometer reading correct and are the compensator mercury
   batteries good?  When were they replaced last?


6B   CHECK LIST WHEN INSTRUMENT REPAIR IS INDICATED

   A systematic check-out routine is recommended to determine whether instru-
   mental service or repairs can be completed in-house during normal instrument
   off-time, or if outside help is required.  Results should be written down
   as the check-out is completed so that full information can be transmitted
   for service.  Serial numbers and EPA numbers of the equipment involved
   should be recorded as part of this information.  Although the check-out
   presented is specifically for a Tracer MT-220 gas chromatograph, the steps
   are illustrative of the kind of routine that can be established for any
   analytical instrument.

   Erratic operation of  the instrument in day-to-day use is often an indica-
   tion that serious trouble is imminent.  Keep in mind the type of detector,
   column, carrier gas,  and temperature range being used.  Recall, or pref-
   erably, consult the instrument log to determine when the instrument was
   last serviced in-house  (e.g., detector or column change) and if the
   difficulties arose directly or shortly after such service.
        • '        '    t                                        '

   Check List

   (a)  Is there proper  insulation packing in  the detector compartment?
   Improper packing can  lead to variation in signal due to ambient tempera-
   ture changes.

   (b)  Are any wires exposed, shorted, or loose?

   (c)  Check  proper location and readings of  thermocouples  and resistance
   thermometers; are they fully  seated?

   (d)  Plumbing leaks  should be checked  for at  full operating pressure  before
   heat is applied, and again at operational temperatures.

   (e)  Is carrier gas  pressure  correct?  Tank pressure  should be  greater than
   500 psi and output pressure  40-50 psi.

    (f)  Are flow indicators functioning smoothly? Are they  steady?  Does the
   float  go smoothly  through its entire range?
                                       -221-

-------
                                                            Section 6C

    (g)   Are panel indicator lamps  correct?   Temperature programmer indicators
    in particular should be observed for smooth  transition  from heat to cool
    cycle.

    (h)   Are switches  set properly,  i.e., oven damper in closed position for
    heating  modes?

    (i)   Does the recorder  respond  correctly  to  attenuation or input changes,
    bucking,  zero,  heat  rise or  cycling?
    (J)  Has signal to noise level increased?
    grams obtained at like settings.
Refer to previous chromato-
    (k)  Has the EC background profile decreased appreciably?

    (1)  Is the oven door secure?  Also check rear access door.

    (m)  Are all units, such as the EC power supply, recorder, electrometer,
   etc., plugged into the proper power source?

    (n)  Check that thermometers and thermocouples in the detector compartment
   are attached to the proper temperature set controllers and pyrometer points
   if the detector has been recently serviced.

   (o)  Is the column properly conditioned and is it the best one to be used
   in terms of stationary phase selection, efficiency, and freedom from bleed?

   (p)  Check color coding of circuits for accuracy.  The correct color codes
   should always be maintained to facilitate servicing.  Further tests may be
   recommended by service personnel to confine the problem to a particular
   section of the instrument.  The various modules are interconnected by cables
   and wires, and these should be disconnected in progression from the de-
   tector toward the recorder.   A volt-ohm-milliammeter with a loading factor
   of 20 K ohms per volt is a necessary item to have on hand when checking
   and servicing electronic and electrical components.


6C   GENERAL APPROACH TO TROUBLESHOOTING

   The following pyramid approach represents a logical, general method for
   instrumental troubleshooting:
                                        -222-

-------
                                                        Section 6C
                        8 /   Fix/replace
                                Use
                            alternative teats'

                            Pause  and think


                          Isolate  problem area


                            VOM-Meter checks


                            Front  panel checks


                            Define the problems
.8
H

-------
                                                         Section 6C
            Normal
                                                              Abnormal
                                   or
    Normal
                                                     Abnormal
                                                     Output normal
 Linear path - make your check in the middle of a bracketed linear path
 "Half Split" Rule.
    Normal
   Normal
Rotameter
                                                             Abnormal
                   Observe  and measure  for
                   normal/abnormal readings
Flow Meter
Filter
                                            Injection
                                              point
                                         Column
                                                          IjDetectorl
                                                            7      '
                                                          Abnormal
Bracket Placement
e.g., no flow:  Rotameter indicates flow
                                     -224-

-------
                                                           Section 6D
   (e)   Pause and think - is there only one malfunction?  What was changed
   or done recently that could cause the problem?  Check with other personnel
   using the instrument.

   (f)   Verify by alternative tests.

   (g)   Speed, availability, and complexity determine replacement or repair
   procedure.


8D   GAS CHROMATOGRAPH SERVICE BLOCK DIAGRAM
       Gas  Flow
        control
         a
Inj ection
Ports
b
^s
— »
Column Oven
c
\ '
*
^
-*«
-*•
^
Support Electronics
g
Detectors
d
^^
  Detector
Electronics
    e
   (a)   Gas flow control:   Purifies and dries the carrier gas,  splits
   carrier gases to the columns and detectors.   Controls and regulates
   the  gas flow.

   (b)   Injection block:  Vaporizes the sample,  introduces the  sample into
   the  carrier gas stream.                              •

   (c)   Column oven:   Houses columns,  provides a dynamic constant or
   programmable temperature environment.

   (d)   Detectors:  Equalize gas temperature, detection of gas  stream com-
   ponents, exhaust gases.

   (e)   Electrometer:   Conditions the  detector signal and attenuates the
   output for transmission to the recorder.

   (f)   Recorder:  Displays an analog  signal (chromatogram).

   (g)   Support electronics:  Controls the temperature of the injection
   ports, column oven, detectors; indicates  temperatures, indicates control
   voltage.
                                        -225-

-------
                                                           Section 6E

6E   GAS INLET SYSTEM OF THE GAS CEROMATOGRAPH

   Components:

   (a)'  Carrier gas cylinder

   (b)  Two stage gas regulator

   (c)  Molecular sieve drying trap

   (d)  Copper tubing to the instrument and rotameters

   (e)  Rotameters

   (f)  Differential pressure flow controllers

   (g)  Tubing to the inlet

   (h)  Inlet/transfer block

        (1)  Thermocouple

        (2)  Resistance thermometer

        (3)  Injection port:   septum nut, septum, septum washer, insert
            retainer,  glass  demister trap.

  Comments:

  (a)   The carrier gas cylinder used in gas chromatography is generally size
   A  dry pumped nitrogen. 'When the EC detector is operated in the pulsed
  mode,  the carrier gas ±a then 5 or 10 percent methane in argon. .The
  carrier gas must be dry and contain less than 5 ppm 03.  Contamination of
  the carrier will severely affect performance.

  (b)  A two stage gas regulator should be used to reduce and regulate the
  carrier gas pressure.  The first stage gauge indicates the cylinder
  pressure while the second stage indicates the reduced pressure to the
  chromatograph.  A diaphragm valve on the regulator allows control of the
  output pressure, which should be 40-50 psi.

  (c)  Molecular sieve drying traps should be  installed between the regulator
  and  the chromatograph to prevent water and hydrocarbon contamination from
  entering the gas chromatographic flow system.  Molecular sieve 13X (1/16 inch)
  pellets have been found satisfactory for the filter load and should be baked-
  out  at 350°C in a nitrogen stream for four hours  prior to use and capped
  off  for storage.  It is advisable to do this  with every cylinder change.
  Occasionally the dryer body should be cleaned with hexane-acetone before
  reloading.   It may also be necessary to flame the dryer frit to  expel all
  contamination.
                                      -226-

-------
                                                        Section 6E


(d)  Copper tubing, 1/8 inch,instrument grade is used between the filter
dryer and the instrument.  This tubing should be rinsed with methylene
chloride and then acetone before installation.  If the old tubing is used,
it should be also flamed.  It is also advisable to clean all Swagelofc
fittings before installation.  Swagelok nuts should be placed on the
tubing for use before being placed on the instrument.  This insures
proper Swage connection and reduces the possibility.of damage during
installation.

(e)  The rotameters generally used in the carrier stream are the Brooks
Sho Rate 150.  They are calibrated for the pressures encountered in GC
work.  Charts of the calibration curves are readily available.  Contamina-
tion of the rotameters may cause the float to stick.  Moisturfe in the
tube will appear as a ring around the float.  The tube must be removed
by loosening the hex screw at the top of the rotameter body..  Do not
attempt to clean the tube in the rotameter body since solvents will attack
the 0-rings or seals causing them to swell.  It is also possible that the
0-rings will absorb the solvent and bleed vapor into the system.  The tube
should be cleaned with hexane, acetone, and Freon MS-18Q.  A flow meter
should not be placed downstream of the flow controller but rather between
the carrier gas inlet and the flow controller.

(f)  The differential flow controllers (Brooks) are composed-of a needle
valve/seat assembly and a diaphragm, preferably Teflon, to maintain a
constant flow of carrier through the column, even though the pressure
drop across the column changes.  The controller requires at least 25 psi
for proper operation and may flutter under lower pressures.  It is
recommended that 40-50 psi be used to operate them properly.  In programming,
it is recommended  that 60 psi be used to insure proper response of the
system.

The needle valve/seat assembly may occasionally stick.  This occurs when
the needle becomes lodged in the seat.  The valve must be taken apart
and the needle and seat cleaned.  Experience has indicated that when this
occurs, a new needle valve/seat assembly is advised.  The main cause of
damage to the assembly is through improper operation:  never close down
forcibly, never open wide past the point that the float rests in the
upper part of the  flow tube.  Since these are differential controllers,
there will still be slight gas flow when the rotameter float is at zero.
This is not uncommon as a flow controller is not a  shut-off valve!

It is advisable to install a small frit spring-loaded filter on the outlet
of the flow controller.  This is a Brooks type 8501/8502 unit that will
protect the controller from  flashback and trap solid contaminants from
the carrier gas, or filter.

(g)  Copper tubing 1/8 inch  od is used from the flow controller and is
joined to the 1/8  inch od stainless steel tubing of the inlet assembly
with 1/8 Swage to  1/8 Swage  unions.  It is advisable to replace these
                                     -227-

-------
                                                         Section 6F

   unions with a Swagelofc slide valve, 1/8 inch.  This will aid in checking
   the system for leaks up to this point and may enable operation of one
   column while repairs are being carried out.  By closing off the slide
   valve, the 'float should drop to zero; if it does not, there is a leak
   up to that point,.  The 1/8 inch stainless steel line is welded into
   the inlet, which is secured by screws into the inlet heater block.

   (h)  The inlet block is an aluminum block with facility for inserting
   the inlet thermocouple and a heater cartridge.  There is also an
   orifice for inserting a resistance thermometer.  With the addition of
   other detectors in the head compartment, this resistance thermometer
   may not be used in many cases and the inlet is heated (225-230°C) by
   controlling voltage through a Variac voltage regulator.  This inlet
   tube assembly-has a septum retaining nut which holds the septum in
   place and has a small orifice for insertion of the syringe needle into
   the port.  The septum washer (stainless steel) is placed under the
   septum and above the insert retainer to allow for a flat surface above
   and below the septum.  The glass insert retainer should be turned down
   until it comes in contact with the glass insert and then backed off
   1/8 turn.  The Vykor glass insert should be removed daily-and replaced
   with a clean one while the used insert is cleaned in the prescribed
   manner (Subsection 5J in Section 5). '

   Care should be taken to insure that the thermocouple and heater leads
   do not become frayed.  A small amount of insulating material or. glass
   tape should be used between metal surfaces and these wires.  Avoid
   sharp bends close to the element.


6F   PROCEDURE FOR ISOLATING PROBLEMS IN FLOW SYSTEMS OF ELECTRON , CAPTURE
     EQUIPPED GAS CHROMATOGRAPHS

   (a)  Allow the column oven to cool to ambient temperature,  set all
   rotameters to zero, and close off all oven exit ports with 1/4 inch
   Swagelok plugs.

        (1)   If the EC detector is suspect because of poor total response
   regardless of the amount of polarizing voltage applied to the detector,
   install an EC detector with a new foil.

        (2)   Prepare a line filter filled with 13X (1/16 inch) molecular
   sieve pellets and activate as previously described.  Place  it at the
   dual-stage regulator output of the carrier gas tank.   The dual-stage
   regulator should be set to deliver 40 psi to the system.

        (3)   Insure that a cylinder of carrier gas which is  known to  be
   free of contaminants is being used in the system.

        (4)   Attach the detector purge line to the proper purge connection
   and set the rotameter for a flow of approximately  90  ml/minute.
                                      -228-

-------
                                                      Section 6F


     C51  Adjust the detector temp-set controller to maintain a de-
tector temperature of approximately 200°C.  If a faulty temp-set is
encountered, substitute a Superior type B yariac.

     (6)  Obtain a profile for the background current signal (BGS) by
increasing the polarizing voltage in five  (5) volt increments, noting
on the chart the step characteristics and  the maximum deflection voltage.

     (7)  The resulting profile should approximate the profile pro-
vided with the detector in characteristics and maximum deflection
voltage although slight variations may occur.  Particular attention
should be paid to the step increase from 0-5 volts and from 5-10 volts.
A leak or contaminant in the flow system are common causes of a poor
profile (Figure 5-E in Section.5).

(b)  Corrective measures for locating faults in the purge loop.

     (1)  Check for leaks at all fittings  in the purge loop with "Snoop",
beginning at the cylinder.  At this point, also use "Snoop" to check for
leaks on the valve control stems.

     (2)  Change the detector purge line to another rotameter on the
purge module.

     (3)  Clean all lines in the purge loop or replace with new ones
that have been cleaned and flamed.  All tubing that is changed should
.be replaced with instrument grade tubing,  and it is recommended that
this tubing also be cleaned prior to  installation.

      (4)  If the condition of  the carrier  gas was not known at the time
the  above tests were made, a new cylinder  of  carrier gas  should be
tested.  All new cylinders should be  checked  for leaks at their welds
and  outlet valves.

 (c)  After  obtaining  a satisfactory profile with purge,  place  1/4 inch
Swagelok plugs  on  all four  column inlet  fittings inside  the oven.

      (1)  Replace  or  remove  the glass demisters  in the inlets.

      (2)  Replace  the inlet  septums with new and preconditioned  septurns.

      (3)  Adjust  the column #1 flowmeter to deliver approximately 90
ml/min of carrier gas.   The rotameter float should rise and very shortly
 should drop to zero if there are no leaks in the column #1 flow system.
 Repeat this on the other carrier systems.   If none of the floats fall
 to zero, the leak is usually common to all ports.   The most likely area
 for this type of leak is at the pigtail fitting on the instrument rear,
 although there have been instances of such leaks occurring due to a
 cracked inlet block.
                                     -229-

-------
                                                       Section 6F


 (d)   Corrective measures  for locating leaks  in the carrier flow system.

      (1)   Open the 1/8  to 1/8 Swagelok unions  on the  rear of the
 chromatograph where the copper lines  meet  the  stainless  steel lines  and
 cap  off the copper line at this point.   Upon.applying carrier flow as
 stated in  (c)  (3),  the  float should drop to  zero.   If the float does
 not  fall,  the leak is in  the flow controller or rotameter.   It is
 possible to tighten slightly the.flow tube in  the rotameter by the hex
 adjustment located on the top of the  rotameter housing.

      (2)   Check all lines with "Snoop"  from  the "pigtail" to the Swagelok
 plug and observe  for leakage.

      (3)   Observe the rotameter float for  "bounce" or rapid but slight
 fluctuations;  this  will usually indicate a faulty diaphragm.   Do not
 attempt to repair the diaphragm but replace  the total unit  with a  new
 flow controller.


 (e)   Install in all ports empty glass columns  that have  been thoroughly
 cleaned, taking care that they are properly  seated,

      (1)   Allow the system to remain  under carrier flow  for at least 30
 minutes to evacuate air introduced into  the  loop  during  installation of
 the  columns.
port,
     (2)  -Follow the procedures outlined in  (a)  (6) and  (a)  (7) for each
     (3)  If an acceptable BGS profile is obtained, the flow system is
free of leaks or contamination under ambient temperatures.

     (4)  Slowly increase the temperature of the inlet, oven, and transfer
lines to their operating levels and obtain a BGS profile for all ports.


(f)  Corrective measures for locating faults in the carrier flow loop.

     (1)  Test for septum leaks, if hew septurns were not used, by rapidly
cooling down the septum nut with water and allowing a small quantity to
remain in the septum nut depression.  Observe this water-filled depression
for bubbles that would indicate a leak.

     (2)  Insure that the columns are seated and sealed properly.  It is
usually advisable to tighten the columns an additional 1/4 turn after
heat has been applied to the oven.

     (3)  Tighten the transfer line fittings, but take care not to strip
the fitting threads.

     (4)  Inspect for a cracked block.  By turning up the carrier flow to
a high level, it is sometimes possible to hear the escaping gas.
                                    -230-

-------
                                                      Section 6F
     (5)  If the steps taken do not correct the problem, it may be
assumed that there is contamination in the carrier loop.

          i.  Remove all fittings in the flow loop and clean or
replace them.

         ii.  Clean the inlet and outlet ports while heated and under
carrier flow with Pre-Post 1001 cleaner.  Use a pistol cleaning brush.

        iii.  Flame all lines under carrier flow where possible.

         iv.  Remove and clean the flow controllers and dry thoroughly
with carrier gas.  Be sure to remove all moisture from the controller
diaphragms.

          v.  Add an in-line filter loaded with 13X molecular sieve at
the junction of the copper and stainless steel tubing until an accept-
able BGS is obtained.  If,. after the addition of the in-line filters,
a proper BGS is obtained, it may be assumed that the problem is in the
rotameter or flow controller area.  Do not operate with* these filters
permanently.  Install a new carrier module and replace all tubing from
the bulkhead fittings to the in-line filters.

        vi.  Install well conditioned columns and allow the system to
equilibrate.  If in-line filters remain in the carrier loop, allow
additional equilibration time because of the greater volume in the loop.
Flow control changes will take approximately 30 minutes to equilibrate
with these filters in the system.  Obtain a BGS profile from all ports.
Insure  that the columns are not filled to the point where the packing
will come in contact with the metal inlet and outlet fittings.  The
higher  temperature at these points may cause the column packing to bleed.
Slowly  raise the oven temperature to 100°C and obtain a BGS profile.
If it is satisfactory, raise the oven temperature to full operating
temperature and obtain a BGS profile.  If it is again satisfactory, the
instrument should now be in full operating condition.

It has  been the experience of the RTP Laboratory that in-line filters
loaded  with 13X molecular sieve are superior in performance to those
loaded  with type 5A.  It is, therefore, recommended that all filters
used be charged with 13X and conditioned as prescribed.  The addition
of in-line filters at the rear of the instrument between rotameters and
column  inlets is not meant to be a permanent change.  The installation
of these filters allows operation of the instrument until it is convenient
to obtain materials to make repairs.  As an additional means of rapidly
checking the system for leaks, it is suggested that the unions on the
instrument  rear where the copper tubing meets the stainless steel tubing
be replaced with Swagelok slide valves  (#200-1/8 SV-6).  These valves
will enable  the chromatographer to totally shut off any Individual
carrier loop so that the float in the rotameter, flow controller, and
                                    -231-

-------
                                                        .Section 6G

   Tines can be checked for leaks.  These valves nay also  be used if re-
   pairs and operation of the instrument are to be carried on simultaneously.

6G   GENERAL INFORMATION - FLOW SYSTEM

   Faulty flow controllers will cause the flow to change from time to time.
   The controller may also exhibit short term fluctuations and may completely
   open or close.  Care must be taken in operating flow controllers;
   they are never opened past the full scale indication of the rotameter
   or closed down in an attempt to completely shut off gas flow.   Always
   maintain a cylinder pressure in the 40-50 psi range. Always change the
   carrier gas cylinder when it depletes to 500 psi cylinder pressure.   Check
   flow through the system at the detector effluent line weekly with a
   bubble meter.   This will indicate the proper function of the instrument
   and condition of the column packing to some degree.   As the instrument
   vibrates, the columns may tend to pack down tighter causing a  decrease
   in flow and may also affect the retention time.   To-check for  a worn
   or bad flow controller diaphragm, operate the unit at 40 psi,  noting a
   setting on the rotameter.   Increase the pressure to 50  psi and note the
   rotameter setting.   If it shows an increase of 4 or more divisions, the
   controller is faulty,   A Brooks filter,  #8501, may be used at  the  outlet
   of the flow controller to protect the system from particulate  contamina-
   tion and to some extent from flashback.                   •

   When installing columns in the instrument,  they can be  set-up  as shown
   in the diagram following.    The lower 0-ring is not  critical and is  used
   simply to support the Swagelok nut.   A clearance of  at  least 1/8 inch is
   recommended between the column packing and the fittings after  installation
   is completed.
                1=4-
Teflon, Graphite, or
Vespel Ferrule

1/4" Swagelok Nut

0-Ring (nut
retainer)
                                                    Glass Wool Plug
                                       -232-

-------
                                                         Section 6H


iH   TEMPERATURE CONTROL AND INDICATION IN THE GAS CHROMATOGRAPH

   The inlet and transfer system may be controlled by a feedback bridge
   type SCR controller.   This controller uses a resistance thermometer
   and a potentiometer for control.   It has been found advantageous to
   utilize variac control of the inlet and transfer where possible (cost-
   and trouble-free operation).   Variac control is possible When ambient
   conditions are stable.

   The electron capture detector temperature should always be controlled
   by the feedback bridge type SCR controller, as minute changes here may
   cause cycling or shift due to small changes in ambient temperature or
   Improper insulation in the detector cage.  There should be no detector
   body area exposed, as slight air currents may cause cycling or drifting
   baselines.

   A pyrometer (0-500°C) is generally used to indicate the temperature of
   various thermocouples in the instrument.  A switch located adjacent to
   the pyrometer is used to connect various thermocouples into the circuit
   to be monitored.  The thermocouples used in the chromatograph are usually
   terminated in a compensator, which is a cold junction reference bridge
   circuit that compensates for ambient room temperature.  This circuit is
   dependent upon thermistors in the bridge circuit and mercury (RM-12)
   batteries that should be checked monthly with a battery checker or.when-
   ever temperature indications appear faulty.  (Usually one thermocouple
   is placed in the open air to rough-check the pyrometer against ambient
   temperature).  Always be sure that battery contacts are clean.

   Oven heating and control are'obtained by:

   (a)  Two resistance wire heaters in the oven walls secured to plates
   attached to the walls and wired in series with a limit switch.

   (b)  Two thermocouples  (metal sheathed).  T/C is used with older type
   programmer units.  Ribbon Resistance Thermometers  (50 ohms) will be used
   in newer types for temperature sensing.

   (c)  Fan motor and squirrel  cage blower assembly.

   (d)  Damper system.

   The oven is generally a stainless  steel unit insulated with micro  fiber
   insulation cover.  The  design permits rapid heating and cooling dynamically
   to desired temperature  equilibration.

   The  top sheathed thermocouple is used  as  the temperature  indicating unit,
   and  the bottom sheathed thermocouple  is  used  for  programmer control.
   This placement is non-critical.

   The  programmer contains an initial control circuit which  is used mainly
   in isothermal control of the oven and  a final  control  circuit  activated
   for programming.   The programmer circuitry may be  used  to raise the
   temperature  at a specific rate  and automatically return to a set
                                      -233-

-------
                                                         Section 61

   temperature by proper use of the adjustments on the front panel.  The
   oven temperature controller will not be activated until the fan/blower
   switch is turned on.                                 •

   The onset of pyrometer problems is often insidious, and problems may be
   prevalent for some time before the operator becomes aware of them.  One
   simple and inexpensive means of monitoring this, if the oven design will
   permit, is to drill a 1/8 inch hole in the oven door at a point opposite
   the center of the columns; insert through the hole the 8 inch stem of
   a dial face bimetallic thermometer (Weston model 4200,  0-250°C).  While
   these thermometers are not highly precise, they are sufficiently accurate
   to provide the operator with an indication of trouble in the pyrometer
   network.  If the dial face thermometer is reading 185°C and the pyrometer
   readout is 200°C, a problem is indicated not only in the column oven
   temperature but in the other heating modules in the instrument, i.e.,
   Injection port, detector, transfer line, etc.


61   DETECTOR AND ELECTROMETER

   The detector and electrometer are integral parts of the gas chromatograph.
   They are connected ttigether by BNC to BNC Teflon coaxial cables, one for
   the polarizing voltage and the other for the detector signal.   The
   electrometer supplies a negative DC voltage to the detector at a regulated
   constant rate.  The radioactive detector source is encased in a Teflon
   (3H) or ceramic (6%i) cylinder.   This cylinder, in turn,  is encased in
   a stainless steel block which serves as a heat sink heated by a 50 or 100
   watt heat cartridge.   When a 3H (tritium) detector is used,  an adjustable
   limit switch in series with the heat cartridge prevents the detector from
   being heated above 225°C.  If the temperature is allowed to exceed 225°C,
   excessive losses of tritium from the foil or damage to  the Teflon parts
   will result.   Such temperatures create no problems with the ^3N1 detector,
   thus a limit switch is not needed.

   The electrometer input attenuator is comprised of high  resistance glass
   resistors forming a decade stepping switch.   These resistors are affected
   by dust,  temperature,  and light.   They must  be maintained extremely clean
   and never hand touched.   The highest attenuation available is  5  x 10~13
   amps.   This is obtained at the 0.1 setting,   A minimum  attenuation of
   5 x 10~8 amps is obtained at the 10*> position.

   The output attenuator is a binary resistance switch that enables further
   attenuation in a 1 to  256 range.   A potentiometer in the output  section
   of the unit adjusts zero balance of the electrometer amplifier to permit
   adjustment of the output voltage for zeroing the strip  charter recorder.
   A bucking control utilizing two glass resistors for normal or  high bucking
   is located on the electrometer rear apron.   This bucking control is  the
   coarse bucking allowing either 10"8 for electron capture operation or
   10-10 for flame.   A 10 turn potentiometer is located on  the  front panel
   for fine bucking control.
                                      -234-

-------
                                                      Section 61


Internally, the electrometer contains two plug-in printed circuit boards,
one for power conditioning and the other for amplification, zero, and
balance.  This is in addition to the input and output attenuators.  It
is important, due to the extreme sensitivity, to operate the electro-
meter with its cover correctly in place, otherwise noise will be
excessive due to stray field pick up.


Electrometer Problems:

(a)  Cannot zero and/or cannot buck out - possible cure:

     (1)  Check zero and bucking pot on front panel for continuity.

     (2)  Check 1 percent resistors on amplifier board for continuity,,

     (3)  Check 2N699 transistor on amplifier board for leakage.

     (4)  Check 4 mfd @ 50V non-polarity capacitor on amplifier board.

     (5)  Check all zener diodes on power supply board.


(b)  To adjust trim pots on amplifier board:

With output attenuation on 1, turn input attenuator to off and turn zero
pot on front panel five turns from either extreme.  Then adjust zero
balance trim pot until recorder reaches zero.


Electrometer Drift Check:

(a)  Set master switch to off and completely disconnect the electrometer
from all test equipment.

(b)  Connect recorder signal cable to the electrometer.

(c)  Set input attenuator to 0.1 and output attenuator to 1.

(d)  Adjust bucking potentiometer for recorder indication of SO divisions.

(e)  Set recorder to slow speed.

(f)  Allow units to run in this condition until recorder does not
deviate more than 5 divisions per hour.

(g)  Set output attenuator from position 1 to position 256 slowly, noting
pen deflection.  This should not exceed 0.25 chart division through the
total range.
                                    -235-

-------
                                                         Section 6J
   Solid State Linearizer:

   A recent Innovation introduced for use with electron capture detector
   systems in the constant current pulsed mode is a solid state linearizer.
   The linearizer enables the chroma to grapher to operate the electron
   capture detector over a dynamic range of at least 20,000:1.

   The linearizer requires a warm-up period of at least 12 hours after
   installation or after any flow interruption.  Argon-5% methane is
   the preferred carrier gas; however, a 10,000:1 dynamic range may be
   obtained using high purity nitrogen.

   Malfunctioning of the solid state linearizers has been observed where
   the recorder will suddenly go off scale and remain so until the unit
   is shut down for a period of time.  Upon reactivation of the system,
   the unit appears to function normally.  If this occurs, refer to the
   operations manual schematics- and:
(1)  Change R2 and &3 to 1K%W
(2)  Change VR£ to 7.5 or 8.2V

(3)  Add 330 mfd at 10 VNP across R14

(4)  Add 1 mfd at 50V in series with a 10 ohm
                                                    across CR5.
   These changes will improve the operation of the linearizer and reduce
   noise to an acceptable level.  It may be necessary to re-zero the unit
   after these changes.  The Tracer service manual procedure "2" states
   "Remove clip lead."  This is incorrect.  The clip lead should be re-
   tained.  In procedure "3" , - do not clip a shunt lead from E^ to ground
   and do not adjust to zero but to 30 MY-
                                         j
   Printed circuit board 1700 - 504400H cannot be repaired or aligned in
   the field and must be returned to the factory for replacement.

   All printed circuit boards in a linearizer should be removed annually
   and spray cleaned with Freon MS-180.


6J   OBSERVATION OF PROBLEMS ON CHROMATOGRAMS

   (a)  Peaks return below baseline:  dirty or partially depleted detector
   foil - clean or replace foil.

   (b)  Peaks have flat tops:  check for proper oven and detector temperature
   and recorder gain control.

   (c)  Insufficient peak height:  check for proper attenuation settings,
   proper amount of injection, recorder response.

   (d)  Tailing peak:  check for proper operating temperature and gas flow.
                                      -236-

-------
                                                         Section 6K



   (e)   Stepping baseline (may be observed on peaks):  check for dirty
   recorder slide wire, line voltage changes, recorder drive, recorder
   gain control.

   (f)   Noise level:  check for approximately 1 division of noise @
   10 x 8 attenuation.

   (g)   Spikes:  check external polarizing voltage unit, line noise, noisy
   temperature set controller, dirty system, regulation diaphragm.

   (h)   Rapid cycling:  check oven temperature programmer, oven control
  . thermocouple, compensation circuit, temperature set controllers, limit
   switches.

   (i)  .Excessive noise in baseline:  check for module noise by elimination
   or substitution, ground loops, recorder gain "(properly set?), cable
   connections (coaxial), leaks in flow system.

   (j)   Noise with erratic spikes:  check for proper carrier, clean carrier,
   leaks, ground loops, column bleed.

   (k)   Slow cycling baseline:  check oven limit switch, damper operation,
   control thermocouple, thermocouple compensator.


6K   DETECTOR BACKGROUND SIGNAL (BGS) RESPONSE

   (a)   Normal detector response - has good maximum signal (BGS).

   (b)   Abnormal detector response - has poor stepping, does not saturate
   at approximately 25 volts (Figure 5-E, Section 5).


   Possible Problems:

         (1)  Moderately contaminated carrier gas.

         (2)  Bleeding or unconditioned column (absorbs BGS).

         (3)  Positive voltage on detector.

         (4)  Leak in system.

         (5)  Detector in heating cycle (wait until pyrometer stabilizes).

         (6)  Reversed coaxial leads from electrometer to detector.

         (7)  Contaminated radioactive source in detector.

         (8)  Contaminant flowing from previous injections (residue bleed).

         (9)  Dirty, bleeding, torn septum.
                                      -237-

-------
                                                          Section 6L

   Troubleshooting:

   (1)  Leave column cold.  Eliminates problems .(2) and  (8).

   (2)  Check system from tank to detector fittings.  Eliminates  (4).

   (3)  Observe voltage with VOM.  Eliminates  (3).

   (4)  Check coaxial connectors.  Eliminates  (6).

   (5)  Check detector temperature for stability.  Eliminates  (5).

   (6)  Observe standard solution injection -  if peaks are not below base.
        Eliminates  (7).

   (7)  Replace septums with conditioned ones.  Eliminates  (9).

   (8)  Change carrier gas ^tank.  Eliminates  (1).

6L   TROUBLESHOOTING COULSON ELECTROLYTIC CONDUCTIVITY SYSTEM


   (a)  High background.

   Symptom:  Unable to zero system with bridge.  Bridge  attenuator x8 or
   below.

   Probable cause:

        (1)  Recorder zero inaccurate.

        (2)  Water contaminated.

        (3)  Ion exchange capacity exhausted.

        (4)  Gas contaminated.

        (5)  Bleeding column.

        (6)  Leak in gas system.

   Troubleshooting procedure:

        (1)  Attenuator at short, zero recorder.

        (2)  Change water.

        (3)  Change ion exchange bed.

        (4)  Check background with the cell disconnected from  furnace,
   change gas.

        (5)  Replace column  with glass jumper.

        "^  Test system for leaks.
                                      -238-

-------
                                                       Section 6L
(b)  Low sensitivity.
Symptom:,  noticeable loss from previous response.
Probable cause:
     (1)  Leak in flow system.
     (2)  Contaminated pyrolysis tube.
     (3)  Contaminated- Teflon transfer line.
Troubleshooting procedure:
     (1)  Test system for leaks.
     (2)  Replace  or 're-cure pyrolysis  tube.
     (3)  Replace  Teflon transfer line.
 (c)  Noisy  baseline.
 Symptom:  baseline noisy, 3 percent or more at x2 on attenuator.

 Probable cause:
      (1)   Recorder gain too high.
      (2)   Bridge not properly grounded.
      (3)  Ion exchange capacity exhausted.
      (4)  Bridge defective.
      (5)  Dirty cell.
      (6)  Improperly cured  column.
 Troubleshooting procedure:
      (1)  Set bridge attenuator  to x2.   Adjust recorder gain.
      (2)  Connect a jumper  from bridge white terminal to recorder ground.
      (3)  Change  ion exchange  bed.
      (4)   Substitute bridge.
       (5)   Clean cell with 10 percent solution of HF, rinse with distilled
  water.
       (6)  Recondition column.
                                       -239-

-------
                                                               6L

(d)  Loss of gas flow.

Symptom:  bubbles not present in cell.

Probable cause:

     (1)  Gas tank empty.

     (2)  Broken pyrolysis tube.

     (3)  Broken column.

     (4)  Valve blocked.

     (5)  Broken flow control.

Troubleshooting procedure:

     (1)  Check tank  pressure.

     (2)  Remove and  inspect pyrolysis  tuoe.

     (3)  Remove and  inspect column.

     (4)  Check flow  through valve  and  clean, if necessary.

     (5)  Check output from flow control.

(e)  Loss of furnace  heat.
Symptom:  pyrometer does not read 820°C or  the  set  temperature.

Probable cause:

     (1)  Heat control off.
                   i )
     (2)  Thermocouple open.

     (3)  Furnace heater open.

     (4)  Heat control defective.


Troubleshooting procedure:

     (1)  Push heat control knob to turn heat on.

     (2)  Visually check furnace.   Insure it  is red on inside,  check
thermocouple with ohm meter.
      (3)   Check resistance of furnace heater with an ohm meter.
   I/
  —  (4)   Check variable 1-110 VAC output of heater control with volt
meter.
   Caution:   use only a volt-ohmeter.   Do not use a vacuum tube volt meter.
                                   -240-

-------
                                                         Section 6M

   (f)  Loss of inlet block heat.

   Probable cause:

        (1)  Heat control off.

        (2)  Thermocouple open.

        (3)  Block heater open.

        (4)  Heat control defective.                  •      '


   Troubleshooting procedure:

        (1)  Push heat control knob to turn heat on.

        (2)  Check thermocouple with an ohm meter.

        (3)  Check resistance of block heater with an 'ohm-meter.

      — (4)  Check variable 0-110 VAC output of heater control with volt
   meter.


6M   TROUBLESHOOTING THE FLAME PHOTOMETRIC DETECTOR (FPD)
                       2/
   (a)  Noisy baseline —

   Probable cause:

        (1)  Detector temperature too high.

        (2)  750 volt power supply noisy.

        (3)  Noisy electrometer.

        (4)  Damaged photomultiplier tube.

        (5)  GC column bleeding.

   Troubleshooting procedure:

        (1)  Lower temperature to 160-170°C.
   — See footnote on page 21.
   2/
   •—In any case of noisy baseline, make certain the recorder gain is
     properly adjusted and the slidwire is clean.

-------
                                                      Section 6M


     (2)  Disconnect cable from end of PM tube and observe base-
line.  If noise continues, 750 volt power supply may be the cause.
Repair or replace,

     (3)  Continuing noise if cable from back of electrometer is
disconnected indicates bad electrometer.

     (4)  Continuing noise with extinguished flame indicates damaged
PM tube if electrometer and power supply have checked good.  Replace tube.

     (5)  Recondition or replace column.

(b)  Low sensitivity

Probable cause:

     (1)  Dirty filter.

     (2)  Dirty window.

     (3)  Damaged photomultiplier; low sensitivity will probably
be accompanied by excessive noise.

     (4)  Improper polarizing voltage.

     (5)  Light leaks.

     (6)  Improper flow rates.

     (7)  Loose cables.•'
                                ,                      -   i  ,

Troubleshooting procedure:                              ,

     (1)  Remove filter and clean with lens tissue.  Be sure to turn
off polarizing voltage before removing PM tube.

     (2)  Remove PM tube and filter and look into back of detector.
Replace window if coated with gray deposit.

     (3)  Replace PM tube.

     (4)  Perform voltage/injection profile.

     (5)  A shift in baseline will occur by shading the detector burner
housing.

     (6)  Insure all flow rates are correctly set.

     (7)  Check that all cables are tight.
                                   -242-

-------
                                                         Section 6N
6N   EPILOG
   The reader is cautioned against immediately assuming that the source of
   an operating problem is an instrumental malfunction when* in fact it
   may be something else entirely.  For example, if the operator makes
   a series of injections of a relatively "dirty" extract for electron
   capture detection, such as the 15 percent ethyl ether-petroleum ether
   cleanup.fraction of fat, certain symptoms may appear on the chromato-
   gram which could suggest electronic problems.  Peak height response
   may be greatly depressed, and significant peak tailing may occur.  How-
   ever, these manifestations are simply electronic symptoms and not causes.
   The cause of the problem in this case would be overloading and contamina-
   tion of the GC. column (see Figure 4-J in Section 4).  Because of the
   visible similarities on the chromatograms of electronic vs. other problem
   sources, the operator should not proceed posthaste to disassemble the
   instrument without first checking out the possibility of other problem
   sources.  In general, if the instrument has been performing satisfactorily
   up to the time of starting chromatography on a new series of samples,
   it would seem probable that the problem may reside in the sample extract
   rather than in the gas chromatograph.
                                      -243-

-------
                                 Section 7


                OTHER CHROMATOGRAPHIC DETERMINATIVE TECHNIQUES

              HIGH PERFORMANCE LIQUID COLUMN CHROMATOGRAPHY (HPLC)
7A   INTRODUCTION TO HPLC

   High performance liquid chromatography is becoming increasingly important.
   as a powerful technique for the separation and analysis of pesticide resi-
   dues.  HPLC is a very gentle technique that commonly operates at ambient
   temperature.  It has advantages over gas chromatography where compounds
   are not naturally volatile and derivatization is difficult or unsatis-
   factory, and for polar or thermally labile compounds.   Pesticides that
   may not gas chromatograph well unless derivatized (e.g., phenoxy acid
   herbicides, which require methylation prior to GC) often give excellent
   liquid chromatograms (1).  The most widely used detector in LC,  the
   ultraviolet (UV) photometer, is nondestructive to the sample, making
   fraction collection a routine matter.  The separated compounds emerge
   from the chromatograph dissolved in solvent that can easily be collected
   and the solvent removed to recover the compound.  A greater variety of
   separations of more complexity can be achieved by LC because of  the active
   role played by the mobile phase as contrasted to carrier gas in GC.   HPLC
   has been used for cleanup of pesticide extracts prior to GC determinations
   (2-4) as well as for the final determination itself.

   Disadvantages of LC are that detector sensitivity is not comparable to
   that obtainable with GC detectors, especially electron capture,  and a wide
   range of element selective detectors is not yet available.   In general,
   present commercial LC detectors have sensitivities in the 10" 5 to 10~9 g
   range.   In one comparative study (1), detection limits by electron capture
   GC were 100-1000 times better for DDT and 2,4,-D than with an LC photometric
   detector, 5 ng of each pesticide being detected with the latter  at 210 and
   278 nm, respectively.   However, the poorer sensitivity could be  overcome by
   injection of large sample volumes (e.g.,  50-100 Ul) without loss of
   linearity or peak symmetry.   The ability to introduce  large volumes  in LC
   can sometimes make sensitivity between the two methods comparable.   Deriva-
   tization methods can increase the sensitivity of detection,  e.g.,  by forma-
   tion of UV-absorbing or fluorescent derivatives,  but only at the cost  of
   more complicated sample preparation.   HPLC is still relatively new,  and
   as improved equipment is developed and analysts obtain a better  knowledge
   of HPLC,  this technique will be taking its place beside GC  and TLC as  an
   important tool for pesticide residue determination.

   Two terms that have evolved  from the early traditions  of liquid  chroma-
   tography should be defined.   In "normal phase" LC, the stationary  phase
                                       -244-

-------
                                                           Section 7B


   (an adsorbent or partition medium)  is more polar than the mobile phase,
   and the least polar sample components will elute first.   In "reversed
   (or reverse)  phase" LC, the stationary phase is less polar than the
   mobile phase, and the most polar sample components will  elute first.
7B   THEORY AND PRINCIPLES OF HPLC

   The principles and theory governing GC and LC are very similar, but
   the: presence of a moving liquid instead of a gas gives far different
   separation characteristics to LC.  The choice of the carrier gas in GC
   is primarily dictated by the type of detector used and has little
   influence on the separation achieved in a given column.  In LC, the
   composition of the mobile phase is of prime importance in the thermo-
   dynamic distribution process.  The other significant differences
   between GC and LC are that, in the latter, solute diffusion in the
   mobile phase is extremely low and temperature effects are of only
   secondary importance.  Low diffusion in the mobile phase (a factor of
   '10* less than in GC) is the key reason why HPLC is possible to perform.
   Instead of plate height becoming increasingly larger as the -carrier
   liquid flow increases as in GC, it becomes asymptotic to a limiting
   value.  In practical terms, this means that the mobile liquid phase
   velocity can be increased without the same great loss in efficiency
   (increase in plate height) and loss of resolution that occur in GC.
   Temperature changes can differentially alter the relative retention of
   two similarly retained solutes by the effect on solubility, mobile phase
   viscosity, mass transfer effects, etc., but .these are only indirect
   effects, and most LC separations are carried out at ambient temperature.

   The concepts of retention time and resolution are the same in LC as in
   GC.  Good resolution requires that peaks be narrow and the distance
   between peak maxima be great enough to allow the trace to return as
   nearly as possible to the baseline.  Peak width is a function of the
   column efficiency (number of theoretical plates), while peak separation
   is a measure of column selectivity (the ability of the packing to
   differentiate between two solutes).
    The  basic  equations  for HPLC are  the following:


              Retention:  k1
Vl  -  Vo
    where V]_ is the retention volume for the peak of  interest  from  the point
    of injection,  and Vo is the retention volume of a non-retained  peak
    measured at the peak apex.  Times or distances measured  along the re-
    corder chart can be used more conveniently than volumes  if the  flow rate
    is constant.
                                       -245-

-------
                                                        Section 7B
           Selectivity for
           compounds 1 and 2:
                           fc*
                           K2
V,, -
                                                  - v
           No. of Plates:   N
 where W-  is the width of the peak for component 1 (see Figure 4-C)
 in terms  of V.
Resolution:  R
                         - f(  " ^ 1 )  (  VST    )   (  3. +'k. )

                           selectivity     efficiency     capacity
 The k'  term or capacity factor measures  retention  in  column volumes;
 it  is affected by the  strength (e.g. polarity) of  the solvent and
 strength (e.g.  retentivity) of the  column packing.  The optimum value
.of  fcr is ca.  2-6.  d is the separation or selectivity term, which is
 affected by the chemistry  of  the entire  system, including the function-
 ality of the sample components.  Values  of 1.1-2 are typical in HPLC.


 N is a measure  of  band broadening;  typically, a highly efficient small-
 particle, porous 25 cm silica gel column will show ca. 15,000 plates
 for various compounds.   The resolution equation combines terms associated
 with selectivity,  efficiency,  and capacity.

 If  the resolution  of two components with k* = 2 or less is unsatisfactory,
 there are three different ways to try to improve the separation.  The
 solvent  can be  changed (solvent strength decreased), to give k* values between
 2 and 6,  the  column can be changed  to increase N and give narrower
 peak widths,  or the solvent can be  changed to give increased selec-
 tivity  (a).   As a  specific example, an inferior reversed phase separation
 on  a 37-75  urn bonded  phase €13 column with methanol/water solvent might
 be  improved by  changing to a  lower percentage methanol (increase k')>
 changing to a. 10 pm column (increase N) , or changing to acetonitrile/
water solvent (increase o) .

 In  general, the best order for developing a separation is the following:
 try to dissolve the compound(s) of interest in a series of solvents
 ranging  from  hexane to water.  If the only solubility is in the hexane
 end of the  series, choose a silica gel column; if the only solubility
 is  in the water end, use a C^g bonded column.  If there is intermediate
                                 -246-

-------
                                                          Section 7C
   solubility,  one has a choice of either column type.   If the compounds  to
   be separated are relatively polar and have functional group differences,
   silica gel is recommended.   If the compounds are relatively non-polar
   and differ mainly in the hydrocarbon skeleton, a Cj^g column is
   recommended.  Use the best  available column to increase N,  choose a
   solvent mixture and vary its proportions to alter kf, and,  finally,
   change the solvent composition while keeping the same strength to
   increase a.

   Doubling column length doubles the number of theoretical plates,  but
   the separation time will also double if flow rate is kept constant.
   Increasing pressure with a  constant column length will increase the
   •speed of separation but reduce resolution.  A simple means  of in-
   creasing the plate number is to reduce the solvent flow rate with a
   constant column length, but again we pay for this by increased separa-
   tion time.  Many operators  seek maximum resolution by using the longest
   column and highest flow rate feasible at the pressure limit of the
   available instrument.  Decreasing the eluting strength (e.g., polarity)
   of the solvent will usually increase resolution but will also increase
   the analysis time.  Column  efficiency is only marginally affected by
   column diameter:  there is  a small increase with increasing diameter,
   but diameter is principally important to sample loading capacity (sample
   size is proportional to the grams of active stationary phase available
   in the column).  Doubling column diameter will approximately increase
   capacity by four, but four  times as much solvent must flow through the
   column in a given time to maintain efficiency and velocity.  A guide
   to selecting the best experimental conditions for high resolution in
   HPLC with large- and small-particle columns has been published (5).
7C   HPLC INSTRUMENTS

   The basic elements of a complete, automated HPLC instrument include a
   solvent reservoir and gradient forming device, high pressure pump,
   injection system, column, detector, and recorder (Figure 7-A).  The
   instrument components must be joined by tubing that,is as short and as
   narrow in bore as possible with low dead-volume fittings and valves so
   as to minimize extra-column peak spreading.  The gradient device mixes
   various solvents to produce a continuous or stepwise change in chemical
   composition (e.g., polarity or pH) of the mobile phase during the
   elution.  Gradient elution is analogous to temperature programming in
   GC since both are used to speed and optimize complex separations.  In
   adsorption and partition LC, the gradient usually involves an increase
   in solvent polarity; in reversed phase partition LC, solvent polarity
   is progressively decreased.
                                    -247-

-------
                                                       Section  7C

    Figure  7-A  High Performance Liquid Chromatograph Including
                Two Model  6000 Pumps, A Model 660 Solvent Programmer,
                and Model  440 UV Absorbance and R401 Differential
                Refractometer Detectors.  Waters Associates.
The pumping system must provide the pressure required to achieve the
correct flow rate through the column.  Although most instruments per-
mit pressures up to at least 5000 psi, the vast majority of analytical
separations can be done at pressures ranging from a few hundred psi
to about 1200 psi.  EPLC pumps fall into two categories, namely,
continuous displacement (e.g., gas displacement, gas amplifier, and
syringe types) and intermittent displacement (e.g., peristaltic,
diaphragm, and reciprocating piston).  Most, users and manufacturers
today are emphasizing variations on the reciprocating pump, while other
pump designs are fading into the background.  The newest models
feature highly precise flow, no significant pulsation of the final flow
to the detector, and automatic correction to provide accurate flow
under different operating conditions.  The increasing use of micro-
processors, which is a definite trend in HPLC equipment, allows
improved operation of pumps and additional options such as low pressure
gradient systems that use a proportioning valve in front of a single
pump.  Micro-processors also allow increasing automation of the entire
LC system, including sample injection and data handling.

Injection is carried out in one of three ways in different commercial
LC instruments.  Stop-flow injection involves shutting off the flow of
solvent in the column, (either by stopping the pump or by using a shut-
off valve), removing a cap from the head of the column, and injecting
the sample directly on top of the column.  Because diffusion in the
liquid mobile phase is negligible compared to gaseous diffusion in GC,
the cap can be resealed and chromatography can be resumed without
significant loss of efficiency.   Two different injector systems allow
sample introduction from a syringe without stopping solvent flow.   With
                                  -248-

-------
                                                              Section 7T3


  a septum injector, the sample is injected directly into the flowing solvent
  with a technique similar to that used in GC.  These injectors are most
  suitable for work at lower pressures.  Septum deterioration from solvent
  attack and coring of septa during injection against high pressure are
  common problems with this type of system.  With the septumless injector,
  the sample is introduced at atmospheric pressure into a loading loop that
  is being bypassed by the pump^to-column stream; by switching to the
  "inject position", the loop becomes part of the main solvent stream and
  the sample is immediately swept onto the column.  The third type of
  injection system is a rotary injection valve in which an external loop
  is completely filled with sample, and the loop is then inserted into the
  flowing stream.  It differs from the septumless syringe injector because
  the precise loop volume determines the amount of sample introduced rather
  than a syringe.  A number of automatic devices for sequential introduction
  of multiple samples is also commercially available (see, for example,
  reference 6).                                         ,

  HPLC modules are  connected  together  almost  universally using 316  stainless
  steel.  Although very hard  and  durable,  these  fittings can be  damaged  by
  too much  tightening or many openings  and closings of  the liquid  seal.
  With a new ferrule and nut, a  leak-free  seal is readily achieved  because
  of  the clean, polished,  and level metal  surfaces.  After use, more  torque
  is  required  for sealing,  and ferrules  and  fittings become  distorted.
  When a fitting  is replaced, the ferrule  connected to  the tubing  should
  also be changed.  The  ferrule  is best  removed  by  (a)  scoring the  1/16  inch
  tube behind  the ferrule  with a sharp  triangular file;  (b)  gripping  each
  side of  the  scored tubing,  2 mm from the score, with  pinch-nose pliers;
  and (c) bending the pliers  toward and away  from the score, through  small
  arcs, until  the tubing breaks  at  the score. This procedure assures  the
  capillary bore  of the  tubing is still completely open.  Any burrs should
  be  removed with a file  from the outer tubing edges before  placing the  new
  ferrule and  nut on the  tubing.


7D  COLUMNS FOR HPLC

  Column efficiency is  increased by using  columns that  are densely  packed
  with uniform,  small particles.  Particles with an average  size down  to
  30-40 Urn  can be successfully  dry packed  in the laboratory  while microparticulates
   (5-10 ym) must  be slurry packed.  Because^of the difficulty of this
  operation, commercial, pre-packed  columns  are  usually used.

  Pellicular packings have a  thin layer or shell of stationary phase
  bonded  to a  solid glass  core.   The active  layer can be silica, alumina,
  an  ion exchange resin, or silica gel to  which  a "liquid" phase has been
  bonded  (bonded phase  partition packings).   The thin layer  of stationary
  phase  provides  good mass transfer  (efficiency) with a particle diameter
   that allows  dry packing  and low inlet pressure operation.   A disadvantage
  of  the  thin  active surface  is  reduced sample capacity.  Totally porous
  microparticulates have very high efficiency because of their small  average
   diameter (5-10  ym) and also have higher  capacity  than pellicular  packings.
  However,  they require high  inlet pressure  for  acceptable flow rates.
                                      -249-

-------
                                                          Section 7D

 For liquid-solid adsorption chromatography, microparticulate silica gel
 of spherical shape is recommended.  Alumina or other packings are used
 occasionally.  Bonded phase packings are used for normal and reversed
 phase liquid-liquid partition separations.  Reversed phase chromatography
 on a monomeric or polymeric phase consisting of GIS linear hydrocarbon
 covalently bonded to silanol groups of silica gel particles is by far the
 most widely used LC mode.  Fully aqueous or aqueous-organic solvent
 mixtures are used as the mobile phase.  Other reversed phase bonded
 packings have phenyl and cyclohexyl groups, while normal phase bonded
 packings contain polar functionalities (e.g., nitrile, attine).   Anions
 are separated on silica-based or resinous anion exchange columns and
 cations are separated on cation exchange columns.  Instead of ion exchange,
 ionic compounds may be better separated by ion pair chromatography.   In this
 technique, ionized samples interact with an oppositely charged counter
 ion in the mobile phase to produce a neutral ion pair, which is select-
 ively retained by a Cj.8 reversed phase column.   Gel permeation  or ex-
 clusion chromatography is used mostly for the cleanup  of pesticide
 extracts.   High molecular weight lipid impurities are  eluted as a group
 before the smaller pesticide molecules.

 Reversed phase HPLC on- bonded Cig packings owes its great popularity to
 its ability to separate nonionic, ionic,  and ionizable substances in
 partition, ion suppression,  or ion pairing (7)  modes;  the stability  of
 the bonded phase columns (if properly used);  and the simple,  inexpensive
 solvent  systems utilized such as methanol/water or acetonitrile/water
 mixtures.   Disadvantages of  reversed phase LC on bonded packing include
 unreacted  silanol (Si-OH)  group.s that can lead  to peak tailing  of polar,
 and particularly basic,  substances.   This is  overcome  by adding a com-
 peting base to  the mobile  phase  or employing  ion pair  chromatography.
 Another  limitation is  the  lack of clear understanding  of the  retention
 of  polar and nonpolar  solutes  on these  columes.  A detailed discussion
 of  HPLC  columns and column technology has  been published (8).

 Microparticulates  of 5 ym  and  10 ym average particle diameter are the
 currently  preferred packing for  HPLC.   The  10 pm particles provide
 adequate resolution for many separations, while 5 ym particles are
 recommended  for the most demanding  separations.  In general, 5 ym particles
 give 2-3 times more theoretical  plates  than 10 ym particles packed in the
 same length  column, but  backpressure is 3-4 times higher.  Therefore,
 lower flow rates are recommended for smaller particles.  For example,
 2 ml/minute  is a typical flow rate for a 4 mm id column packed with 10 ym
particles» while 0.5-1 ml/minute would be used for the same column packed
with 5 ym particles.  Pellicular packings are mostly useful for guard
 columns to protect the analytical column from contamination (see Sub-
 section 7G below).  For this purpose, pellicular packings can be easily
dry packed.                           '

Prepacked columns are available with inner diameters ranging from 2 mm to
8 mm.  For most analytical and small-scale preparative work, a 4 mm id
column is recommended.   For 5 ym particle diameter packings, 15  cm, 25 cm,
                                   -250-

-------
                                                           Section  7E


  and 30 cm column lengths are available.  The 15 cm column is recommended
  for simpler separations requiring only a few thousand  theoretical plates.
  It uses  less  solvent,  gives twice the separation speed, gives  slightly
  higher pressure drop,  has half  the  capacity, but gives about the  same
  number of plates as  a  30 cm column  packed with 10 ym particles.   Compared
  to 30 cm, 5 urn particle columns, 15 cm, 5 ym columns give half the number
  of plates, pressure  drop, separation time,  sample capacity, and solvent
  usage.   A good 15  cm,  5 ym particle column  should produce ca 9,000 plates,
  a 25 cm  column 15,000  plates, and a 30 cm column 18,000 plates.

  An important  consideration for  trace analysis is that  much larger sample
  volumes  can be injected for HPLC compared to GC.  Injection volumes  of
  50-100 yl are not  uncommon for  an LC column:  This may eliminate  the
  sample concentration step often required prior to GC,  and can  help to
  offset the lower sensitivity of LC  detectors.  Wider bore columns are
  better for the injection of larger  samples  since the allowable sample
  volume is proportional to the  square of the tube diameter.  In general,
  the  sample should  be injected  in the mobile phase or in a solvent that
  is significantly weaker than the mobile phase so that  preconcentration
  of the solute occurs at the top of  the column.  For  gradient elution,
  injection of  larger volumes of a dilute solution in  the initial  (weaker)
  solvent  can be made, or a very low  volume of sample  dissolved  in  the
  final solvent of  the gradient.


7E   MOBILE PHASES  (SOLVENTS)  FOR HPLC (see also Subsection  31)

   Solvents chosen  for adsorption HPLC should  have a  low viscosity  (for high
   efficiency  and low backpressure),  low boiling point  (to facilitate  sample
   recovery),  adequate purity,  low toxicity and odor, reasonable  cost,  and
   detector compatibility (low UV cutoff for  the popular  UV  absorption de-
   tector)  .  A widely versatile set  of solvents with  a  range of chromato-
   graphic properties is hexane,  methylene chloride,  diethyl ether,  ace-
   tonitrile and methanol (most polar — strongest  eluant  for normal  phase
   adsorption chromatography).   These solvents are usually used  in mixtures
   with a solvent composition and strength that optimize  capacity and se-
   lectivity (Section 7B).  Solvent mixtures  for adsorption  HPLC  of  polar
   samples often contain at least a small concentration  (0.01-1%) of a polar
   modifier (e.g.,  water, alcohol, acetonitrile)  (9).   Dissolved  oxygen in
   solvents can have a variety of undesirable effects on UV and fluorescence
   detectors.   Helium sparging of solvents is an effective way to remove
   oxygen  and eliminate artifacts it can cause (10).

   Many of the  same considerations apply to the selection of solvents for
   the most important HPLC mode, reversed phase chromatography on chemically
   bonded  packings.  Because the phases are reversed, water, the most polar
   solvent, is  the weakest eluant, while neat methanol and acetonitrile are
   the strongest eluants used in most applications.   The most polar solutes
   are the least retained on the column.  Eluants of intermediate polarity
   are usually  obtained  by mixing methanol or acetonitrile with water or an
    aqueous acid, base, or buffer solution (11).  Ternary solvent mixtures
                                       -251-

-------
                                                             Section 7F
    of varying composition, e.g., water, methanol, and acetonitrile or
    tetrahydrofuran (12), allow control of selectivity of solutes with
    different functional groups and provide the usual discrimination of the
    nonpolar portions of the molecules typical of reversed phase LC.   In
    general, silica gel bonded reversed phase packings are stable only with
    solvents in the pH range of 2-7.5.  At low pH, attack of the Si-C bond
    is possible,  whereas at high pH, the silica matrix may be attacked,
    particularly in salt solutions.   In both adsorption and reversed phase
    partition HPLC, gradient elution can optimize both resolution and speed
    for complex samples with components that cover a broad range of polarity
    [but resolution of any given compound pair is actually reduced compared
    to isocratic (constant solvent composition)  elution (13)].

    Solvents for separation of ionic compounds (e.g.,  pesticides with acidic
    or basic groups)  by ion exchange chromatography include aqueous acids,
    bases,' or buffers that allow the solutes to  possess full or'partial
    electronic charge and to be more or less attracted to the ionic groups
    of the stationary phase.   As an  alternative,  the pH of the mobile phase
    can be adjusted with acids,  bases, or buffers to reduce or eliminate the
    degree of ionization of acidic or basic  groups and allow separations by
    a  partition mechanism on a. CJ_Q bonded phase  column.   This method  is
    termed ion suppression.   A third mode for the separation of-  electrolytes
    is ion pairing.   The mobile  phase is at  a pH where the solute is  charged,
    and it also'contains a pairing agent that conjugates with the solut.e to
    form a hydrophobic,  uncharged species  that is selectively retained  by a
    GIS  bonded phase.   Typical pairing agents are a quaternary amine  for weak
    acids  and an alkyl sulfate or sulfonate  for weak bases.   The  chain  length
    of the counter  ion controls  the  hydrophobicity of  the  final ion pair and,
    therefore,  the  extent of  retention by  the column.

    Solvent  selection  for a particular separation is aided by two fundamental
   parameters, namely solvent strength (for  control of k1) and solvent
    chemistry (for  control of  a).  Solvents have  been  tabulated according'to
    their  strength  (polarity index)  and  chemistry  (solvent group) for easy
   comparison  (14).  A rational  approach  to  the  selection of mobile phases
   for all  forms of HPLC  has  been presented  (15).

   When changing from one solvent to another, time must be allowed for the
   column to become fully equilibrated with  the new solvent.  For bonded
   phases this will require only about 5 column volumes, but for adsorbents
   and some  ion exchange  resins, the equilibration volume can be large
    (ca 30-50 column volumes).


TS   DETECTORS FOR HPLC

   The most common detector for HPLC is the ultraviolet (UV) absorbance
   detector.  The original UV detector used a low pressure mercury lamp
   with a filter and emitted light of very high intensity predominantly at
                                      -252-

-------
                                                         Section 7F
a wavelength of 254 nm.  Many aromatic- compounds absorb strongly at or
near this wavelength and so can be detected with good sensitivity
(ca 10-3-0 g/ml) with this detector.  It is advantageous, however, to have
a variable wavelength spectrophotometer-type detector so that the analyst
is able to work at the wavelength of maximum sensitivity for each com-
pound, increase selectivity of detection of the analyte over interferences,
and use absorbance ratios at several wavelengths to improve identification
of peaks.  Sources for variable wavelength detectors include a phosphor-
coated low pressure mercury arc, a medium- or high-pressure mercury lamp,
or a xenon-arc lamp, each combined with interference filters or a mono-
chromator for isolating the desired wavelength.  Since the intensity out-
put of the variable wavelength detector is lower than the mercury arc,
sensitivity  of detection is somewhat lower for  those compounds with a
strong absorptivity at 254 nm.

Refractive index  detectors are either the optical deflection or  the prism
 (Fresnel) type.   These are universal, relatively insensitive (10-6-10~7 g/ml)
detectors that require close  temperature control  (i. 0.5°C) and cannot be
used  with solvent programming.  They have had little or no use in pesticide
residue  analysis.

Fluorescence detectors are highly sensitive  (10~10  g/ml) and selective
because  of the choice of excitation and emission wavelengths.  An
 especially promising  approach for trace analysis  is the use of high
 intensity laser  sources, with which  750 fg of aflatoxins have been de-
 tected (16).  Derivatization (pre- or postcolumn) of the pesticide of
 interest is  often required  for  increased  selectivity or sensitivity  of
 detection.

 The first electrochemical  detector forlEPLC utilized polarography, but
 recent models have employed amperometry,  coulometry, or conductivity.
 Reference (17) reviews the theory, applications,  advantages, and dis-
 advantages of these detector types.   The 'amperometric  detector  as
 developed by Kissinger and co-workers seems especially promising for
 pesticide analysis, having picomole sensitivity and selective  response
 (18).  Seven halogenated aniline metabolites of carbamate,  urea, and
 anilide pesticides have been separated without derivatization on a CIQ
 bonded phase column and detected at levels of 0.23-0.38 ng (to give a
 response that is twice the noise level) using a commercially available
 electrochemical  detector (model LC-2A, Bioanalytical Systems,  Inc., West
 Lafayette,  IN) operated with a + 1.1 volt oxidation potential (19).   This
 detector cannot  be used with flow or solvent programming, and waiting
 periods of  10 minutes to several hours are required for any changes in
 conditions  (flow rate, applied voltage, different  solvents) or for initial
 start-up each day.  The sensitivity of the detector varies drastically with
 applied voltage, and voltage can be varied to  obtain selectivity for
 analytes over interferences.  Both increased flow  rate and an increase in
 the volume  of sample  injected decrease detection sensitivity.  Detector
 response was .linear from the detection limit up to ca 50 ng injected.
                                   -253-

-------
                                                            Section  7G


   The traveling wire flame ionization detector gives promise of universality
   and sensitivity like its GC counterpart,  but it is at present mechanically
   complex and cumbersome.   The electron capture detector has been adapted
   for HPLC using mobile phases that do not  give responses.   The column
   effluent is vaporized directly into the detector,  in an atomized  form,
   by means of a heated transfer tube located in an oven. Nitrogen  is used
   as the purge gas to sweep the vapors out  of the detector  (20).  LC-EC
   systems are commercially available with sensitivities listed as 10"1" g/ml
   for common chlorinated pesticides.  The FPD (22) is similar to  the de-
   tector used in GC but utilizes a special  burner assembly  to handle the
   total liquid effluent of the column, which is nebulized and directed
   into the cool hydrogen-nitrogen flame. Emission is measured by a simple
   bandpass photomultiplier system using the usual S- and P- wavelengths.
   Sensitivity is limited by the quenching effect of  the organic solvents
   used as the mobile phase.  Practical analyses of pesticide residues have
   not been reported with any of these detectors.


7G   PRACTICAL ASPECTS OP SUCCESSFUL HPLC OPERATION

   All new columns should be tested with a standard mixture  at standard
   chromatographic conditions to compare with the manufacturer's guarantee
   or previously used columns, and as a reference point for  monitoring
   column changes with use.  New columns must be fully equilibrated  with
   the solvent, and the column connections must have  zero dead volume if
   constant retention times and high efficiency are to be achieved.  The
   test mixture should contain pure compounds, one of which  is nonretained
   plus at least two others that have k1 <10 and are  well resolved so that
   as the column slowly degrades they will not overlap.  The test  mixture
   can contain pesticides to be analyzed in  real samples with the  same solvent
   system that is to be employed, or it can  be a mixture specified by the
   column manufacturer so the performance data supplied with the  (pre-packed)
   column can be verified.   The concentration levels  should  be comparable to
   those to be used in the actual analysis.

   Parameters monitored include absolute (k1) and relative (a) retention,
   plate number (N), asymmetry (tailing), void volume, and pressure  drop.
   Small differences in a and k1 usually reflect normal differences  in
   solvent composition, but large decreases  in these  parameters or an increase
   in asymmetry are indications of column degradation or deactivation.  Both
   channeling and compression of the packing can also increase.  If  the void
   volume decreases, channeling may be occurring or the packing pores may
   contain gas bubbles or immiscible liquid.  Changes in pressure  drop
   indicate channeling, plugging, or leaking.  Buildup of impurities from
   the solvent or samples will eventually cause loss  of column efficiency,
   which can :i?ually be restored by regenerating the  column  with a series of
   solvents of increasing eluting strength (adjacent  solvents must be mlscible),
   The solvent sequence is then reversed, and each solvent is followed by a
   weaker one.  A possible solvent sequence  for regenerating adsorption
   (normal phase) columns is methylene chloride-methanol-water-methanol-dry
                                    -254-

-------
                                                         Section 7G


methylene chloride-dry hexane (20 column volumes each).   Acetonitrlle
containing 1% DMSO or DMF (20 column volumes) is an effective solvent
for regenerating reversed phase bonded packings.  A column volume or
dead volume is 50-60 ml for a typical 4 mm id x 25 cm HPLC column.  The
pump and connecting tubing are prewashed with each new solvent that is
put through the column, a flow rate of ca 2 ml/minute is used, and the
detector is left connected, if possible, to also clean it.  Regeneration
of HPLC columns has been described in detail (23).

The inevitable, permanent column degradation that occurs with prolonged
use can be retarded if proper precautions are taken for sample cleanup,
solvent preparation, periodic column regeneration, and storage. The
manufacturer's literature should always be carefully studied and recommenda-
tions should be faithfully followed.  Prevention of plugging is probably
the one most important precaution that must be exercised to prolong column
permeability and efficiency.  Removal of particles from solvents is dis-
cussed along with other aspects of solvent purity in Section 31.  Sample
extracts or solutions should also be free of insoluble particles (24) and
should be filtered, if necessary, with a hypodermic syringe fitted with a
Swinny-type filter  (0.5-1 Vm).  Irreversibly sorbed compounds can irrepair-
ably damage the column and, if present, can be removed on a short (5-10 cm)
guard column located between.the injector and the analytical column.  In
order not to sacrifice separation efficiency, the guard column should be
of the same diameter and packing as the main column.  Less expensive,
easily dry-packed guard columns can be prepared from ==40 Jim pellicular
sorbents, but  some  efficiency may be lost if the analytical column contains
microparticulates.  In addition to the guard column, a silica precolumn
 (25) should be placed between the pump and the injector to presaturate the
mobile phase with silica gel and retard the  dissolving of HPLC columns.
 Solubility of  silica gel leads to sunken beds with skewed surfaces, re-
 sulting  in distorted peak  shapes or increased backpressure.  The extent
 of dissolving  is a  function of the type of column packing and the exact
nature  (pE,  concentration) of the mobile phase.  When placed before the
 injector, voids  do  not contribute to band broadening and  lose of efficiency,
 so  that  large  particle silica gel is perfectly  adequate.

 Columns  damaged  by  plugging, bed compression, or  irreversibly sorbed
 material can sometimes be  returned  to  original  efficiency by  removing the
 column inlet fitting and frit and replacing  the discolored packing and
 deposited material  with  fresh packing.  The  same  packing  material  should
 be added by the  appropriate dry  or  slurry  packing procedure,  or,  alter-
 natively,  a methanol  slurry of  the  packing can  be added  dropwise  and
 allowed to  settle  into place.   The  end frit  Is  cleaned before replacement
 by immersion in an ultrasonic bath  for a  few minutes.

 A properly packed and  cared-for column should be stable  for 3-6 months or
 more with continual use.   A poorly  packed column can settle with  use,
 creating a void at the top that leads to  broad  peaks with poor symmetry.
 Purchase of a good commercial column or properly performed slurry packing
 followed by column compaction via pressure pulsing (26)  should virtually
                                   -255-

-------
                                                         Section 7G


 eliminate  this problem.  HPLC columns should never be bumped, dropped,
 jarred, bent, tapped, or vibrated.  All connection fittings mu'st be
 clean  (use an ultrasonic bath).  Fittings are never over-tightened or
 they will  become distorted and eventually leak.  Columns should be stored
 tightly capped in a compatible solvent.  Silica-based bonded reversed
 phase  packings are used between pK 2 and 7.5 and are never stored in
 aqueous solution but are flushed and stored in methanol or acetonitrile.
 I£  the column has been used with a buffer, it is flushed with water and
 then with  the organic solvent.  Aqueous solutions can be left flowing
 slowly overnight (5 ml/hour) for use the next day, but the column should
 not remain static in aqueous solution.

 Some common problems in adsorption LC and possible means for their
 correction (9) follow.  Baseline drift can be caused by strongly ad-
 sorbed peaks eluting from an earlier run.  Such drift is remedied by
 pumping through the column at the end of each run several column volumes
 of  strong  solvent (isocratic elution) or solvent of higher final strength
 (gradient  elution).  Baseline drift, can also be caused by incomplete
 system equilibration in switching from one solvent to another.  This
 problem is most acute with the RI detector.  Spurious peaks can be caused
 by  bubbles in the  detector or impurities in water or other solvents.
 Bubble formation is avoided by prior solvent degassing or installation of
 an  in-line backpressure valve generating 50 psi to keep all gases in
 solution.  Peak tailing is more common in adsorption HPLC and is often
 caused by  insufficient adsorbent deactivation.  Use of a modifier in the
 solvent can correct this problem.  Partial ionization of the sample can
 cause  tailing that ca:i be suppressed by changing the solvent pH or ionic
 strength.  Injection of the sample in a solvent stronger than the eluant
 can also cause tailing; a solvent weaker than the eluant, or more pre-
 ferably the eluant itself, should always be used, if possible, or the
 sample may be injected in a very low volume of a stronger solvent.  In
 reversed phase systems, peaks can also be broadened or even split when
 the sample is injected into a mobile phase that is either more or less
 polar  than the solvent in which the sample is dissolved (27).  Drifting
 retention  times can be caused by differences among solvent batches, changes
 in  composition of a batch of solvent on standing, changes in temperature,
 or  inconsistent adsorbent activity.  Adsorbent activity is maintained
 constant by using clean samples, pure solvents with an adequate level of
 modifier,  and frequent column regeneration.  Poor reproducibility of re-
 tention times and peak areas in gradient elution may arise from inadequate
 column regeneration between runs or poor mixing in the mixing chamber.
 Other  causes of nonreproducible retention times can be a nonconstant re-
 corder drive or slipping chart paper; a nonconstant flow rate of mobile
 phase  caused by nonreproducible pump delivery or a leak at any fitting
 throughout the system or in the injector or pump; or a contaminated or
 "coated" column that is irreversibly retaining a portion of the sample
 and thereby changing the partition coefficients of some of the components.

Dirty detector flow cells may. be cleaned by using a 10 ml syringe to rinse
 the cell successfully with methylene chloride, methanol, and water.  The
 cell is then filled with 50% nitric acid and allowed to stand for 30
minutes,  followed by flushing with filtered, distilled water (28).  If
                                   -256-

-------
                                                            Sections 7H, 71


   this operation dec:, -«». reduce the background to an acceptable level,
   either other problems are involved (e.g.,  cell misalignment or an impure
   or inappropriate solvent) or the detector  must be disassembled and
   cleaned further in accordance with the manufacturer's instructions.

   Quantitation in HPLC with UV detection is  best carried out using peak
   heights if solvent composition can be maintained precisely but flow
   control is poor.  Peak areas are used when flow rate is stable but the
   composition of the mobile phase might vary (as would be common in
   adsorption chromatography where traces of  water and polar contaminants
   are difficult to control) (29).                            ,

   Readers using HPLC for analysis of pesticide residue^ are strongly urged
   to study the excellent discussion of many  practical aspects of the field
   given in reference (9), from which much of the material in this subsection
   was taken.


7H   HPLC DATA

   HPLC data of 166 pesticides in the form of elution volume or capacity
   factor (k1) have been= tabulated (30).  Information has been included on
   the column packing and dimensions, mobile phase composition, detector,
   and sample substrate.  Data for 26 urea herbicides on silica gel 60 with
   hexane-methylene chloride-ethanol (20:79:1 v/v) were also reported (31).


71   APPLICATIONS OF HPLC TO PESTICIDE ANALYSIS

   The material presented on HPLC has been expanded in the present, revision
   of this manual because-of the increasing importance of the technique in
   residue analysis.  Even greater coverage is anticipated in future re-
   visions as HPLC becomes more sensitive, practical, and reliable in the
   multiresidue analysis of field samples.  For further information on HPLC,
   readers are referred to reviews describing LC detectors (32-34), column
   packings  (8), and general principles and equipment  (35,36), and to books
   covering  theory, principles, and practice (37,38).  A scheme has been
   published for isolating and troubleshooting instrument and column problems
   in HPLC utilizing a, glass-bead column and two simple electronic checking
   devices (39),

   A novel method with great promise for the simplified monitoring of residues
   in water  samples down  to ppt levels has been termed "trace enrichment".
   The procedure combines concentration, separation, detection, and quantita-
   tion of nonpolar to moderately polar impurities.  The water to be analyzed
   is pumped through a C^g reversed phase column until a sufficient quantity
   of impurities has been deposited at the top of  the  column.  (In a reversed
   phase  system, water is the weakest possible eluant, so the organic com-
   pounds will be  concentrated in a tight band at  the head of the column.)
   A gradient elution from  100% water to 100% methanol or acetonitrile  is
   then performed, during which the organic impurities are eluted sequentially
   in order  of their polarity  (most polar is first eluted) and detected with
   a UV absorption detector  (40).               .
                                       -257-

-------
                                                            Section 7J

          •
   Table  7-1 contains some other recent applications of HPLC to pesticide
   analysis, selected to illustrate the range of pesticide types and samples
   that have been studied.  Most analyses to date have been developed for
   one or a few specific residues in food or crop samples, and many analyses
   of formulations and technical material now involve HPLC as the determina-
   tive step.  Review of applications of HPLC to pesticide analysis are cited
   in References (1, 41-44).

   An investigation was carried out to assess the usefulness of reversed
   phase  HPLC with DV detection and gradient elution for the determination
   of residues of pesticides included in the European Economic Community
   directive on fruits and vegetables.  It was shown that most of the 42
   compounds, comprising pesticides of many different chemical classes, could
   be separated, and detected, but that sensitivity was not sufficient to
   detect some compounds at or near the EEC limit (45).  The best application
   of HPLC was considered as an adjunct to established GC procedures for •
   multiresidue screening.  Applications of HPLC in the environmental analysis
   of water were reviewed (46).  Preconcentration and cleanup of residues on
   small, disposable Sep-Pak cartridges (Waters) allowed the determination
   of 20 pesticides at 20.ppb levels in surface water.

   Attempts have been made to combine HPLC and mass spectrometry for the
   direct analysis of column effluents.  Approaches to this direct coupling
   have included (a) use of atmospheric pressure ionization; (b) enrichment
   of reversed phase effluents using a dimethyl silicone membrane interface;
   (c) chemical ionization using a small fraction of the carrier solvent as
   reagent gas; (d) transport of solute through differential vacuum locks on
   a wire or a metal or plastic ribbon; (e) reduction of solute to hydro-
   carbon and sebsequent FID and MS analysis; and (f) formation of a molecular
   beam by laser vaporization of solvent.  At present, (c) and (d) are the
   most popular methods, but none is free from major disadvantages and none
   has yet been tested for routine pesticide analysis.  The various methods
   for on-line and off-line coupling and some applications of HPLC-MS have
   been reviewed in detail (47-49).
                          THIN LAYER CHROMATOGRAPHY
7J   INTRODUCTION TO TLC
   The first multiresidue method available to the pesticide analyst for
   identification and estimation was based on paper chromatography (94-96).
   Paper chromatography has now been largely replaced by thin layer chroma-
   tography (TLC) since the latter will generally give faster and more
   efficient separations with better spot definition and greater sensitivity.
   The use of paper chromatography in pesticide residue analysis has been
   reviewed (96-99).

   TLC is used mainly for confirmation of residues following initial
   screening and quantitation by GC.  Confirmation by TLC,  which is based
   on comparison of migration distances of the pesticide of interest with
   authentic standards run on the same layer, is covered In Section 10E of
                                      -258-

-------
a
£>





i
B


01
B O

•H Vl
!• •*•*
M
Vi Vl
a o
> 0
O U
u
rH iH
IH T)
Vl 0}
U VI
^00
2 S
(U W
•5 §
•o
01
n
Vl
£
'1 

•






i
00 rH
•H H
r-j 0

4J ."2 4J
a o ws
O i-l O.
fj g T3

il *O is 01
is o s -a




8
i
oo
u
2
A
•8
o. a
a vi
•O 01
§s
» Vl
3.
a
•o
Vl



a


41
"
2
i
•s
3





B
01
t)
15
«H
U
Z
S
•H
                                                         X
Cu
O.

rH
I
H
O

O
                                                         o
                                                         o>
                                                                    a
                                                                    o.
                                                                   I
                                                                   A
                                                          H
                                                          a
                                                          B.
fluorescence

after dynaol

hydrolysis a

•labeling
                                                                    grH
                                                                    •H -H
                                                                    W U
           K 0

           8.8
                                                                                       «r
                                                                                       in
                                                                                       CM
                                                                    § 0.
                                                                    g.§
                                                         "O
                                                          a
                                                          U 01
                                                            a
                                                          00 B
                                                                      01 t>
                                                                    in oo o
                                 •c
                              o 01
                              rH O
                               I  Vl
                              rH 01
                              •H >
                               Q 01
                               8 H 01
                              u   a
                               vi to a

                              fl§£

Carbamate

Insecticide

metabolites
                                                                           0 r
                                                                    •  •  •§ |
                                                                    33312,38.
                                                         -259-

-------
 a
 o
Tt
4J
 O
  3
                                               B  IH i-i
                                               a  u o
                                               O  0)  i-l
                                              •H -a  -H
                                              W  I  0)
                                              •H  U  Tl
                                              U   in u-i a
B w o -a
                                                             -o
                                                            > CO
                                                            5 CM
                I  91
               I-l  4J
                o  ca
                B  f
                c  o-
               A  U r
          I   O  O
          g  B  B

          a  a. x
         o o  i
         ta to  u
                                             o  a.
                                             o  o
                                             a  u
8
a,
                                S< O  O
                                00 C  O
                                                                  •8

                                                                ,  3
                                                               iH O
                                                               o o
                                                               a o ^
                                                               

                               2
                               o
                                                              M  S a)
                                                              cu  (0 c
                                                              >  CD vl
                                                             •H  i-l U
                                                             iH  O. 3
                                                                                       Si-i
                                                                                        c9 i
                                                                                                      w eg  o


                                       "5 o S
                                           31  a*
                                           co g
                                                                         e
                                                                         2
                                                                                          8B
                                                                                          •H
                                                                                                      3
                                                                                                      to
                                                                          •i
                                                            -260-

-------
                                                                 o
                                                                 VO
 C
 O
•H
4J
 CJ
 0)
CO
             Cu
       I

      o
 o.
 a.
m
o
                                   o
                                   o
I
cu

a
o
                                                                 1
                                                     &
                                                                             II
             I
            e*>
            at
                                   O
                                                           I
                                                          o
                                                          00
                                                                                       s
                                                                                       CO
0)
- 00
53
•n a
w u
U X
O 01
h
s§
01 i-t



1

rH
8



II

3 rH
rH O
a u
and
allicic
•a
I §
u a
.33
rS S
J? 3
IH rH
a a
a

rH
O
O
•O
•H
U
a
llation.
•H
U
00
•ft
•a

4J
a
action
01
§:
•H a
*j o
•H -H
W S
M 4>
a -a
CU 0
H
o

to CJ
•rt M
^4 4)
0) *" tfl
si M a
O.O J3
CO O CU
41
•o
M
o
rH
•i-S

B 01
a c
CU 41
O rH *-s
n >>r»
O.J2 CTv
O U «•
co 4) en
•rt B1—
O
rH
1
rH
U rH
0>
•SM
g-s
U -H
0 rH
SM
to
                                                          CO

                                                           i
                                                           41
                                        £  5
                                            rnl
                                           -J2 rH
                                          10 -rt O
                                          3 4J O
                                          o  a
                                          •rt  O.X
                                          rJ  S <-»
                                          a  o -»H
                                          r>  U S
                     g    §
                    •H -a -H o
                     jj QJ   ^Q

                       &CO  " C!
                       P U CJ
                     O  €0
                                                                                             M  > O
                                                                                             a  4i js
                                                                                             P4  >4 O.
   co

   U

rH M     -O


& S     <2
                                                             o
S
             e
             1
 I
                       Ps S".

                       g g  41*
                     CO JS A rH *-*
                     41 cu eu o B
                     •O    -H  N 1H
                     •H  «,0  a M
                     tj rH   TJ B
                     •H rS  • B -O
                     00 8 rH  41 C
                     B O O A 01
                     3 B B  a rO
                     pu 4) 01 -H H
                       .a .c .c a
                                                                                        •o

                                                                                        a
          s a

          f*S
                                                                                                      I
                                                      -261-

-------
                            in

                            CO
 O

•H

4J

 O

 Q)

CO
               o.
               O.
 I
 00

 O

 O
               I
               o

               o
                         s
                         a.
                         c.
                                                  e
                                                  o.
                                                  a.
                                          o
                                           i
                                                00
                                                                                          •a
                                                                                           o.
                                                                                           o.
                            OS
                             I
                                    -I
                                    O -H
                                    C CO
                                                             g
                                                             o
                                                                                                     u  >
                                                                                                     O -H
                                                                                                     a  ao
              a
              p.
                             H

                             a >
                             o.

                             0)

                             CO
                            •a
                            •H
                            O
                                     •H 81
                                      a 1-1
                        o u
                        u  co •a
                        •a -H a)
                                        a  a
                                        a  s
                                                                                                        4
                                                                                                     •H "O
                                                                                             01
                                                                                        B
                                                                                        O
                                                                                          «"•§.
                                                                                        W 00    CO
                                                                                        o d t3  B
                                                                                        a 3 u  I
                                                                                        M .c to  9
                                                                                        JJ O H iH

                                                                                        ssgs
                             O g


                             S B
                            (S 0)
                            OJ M
                               o
                                               00
                                               in
                                               ts
   •* w

 a x 
M 4)
M
•O
C T3
U 01 01
•a co
g§5
3-J3 P.
s
•^1
•?>
T<
o
CO
gg
J!
00 CJ
iH-ri
SS
•O
0)
CO
W



*
* 01
00 n
 
-------
                                     CO
                                     CO
§
Tt
4J
O
 S O  4J
                                                                                 ^H P.O  OJ
                                                                                 W O.  >
                                                                                 •H   Oh
                                                                                 am    41
                                                                                   So -a  «
                                                                                     • s  a)
O
o
rH

O
CT>
                                      O
                                      Oi
                                      A
                                   B
                                   a
                                 u >
                                 O -rt
                                 C DO
A
lO
                                                                               g
                                                        I
                                                       S
                         O
                         H
                         §•
                        •a
                                              •o o
                                               0 o
                                   §0
                                   60

                                 *» a
                                 •rt O
                                 4J 1-1
                                 H iH
                                 a -H
                                 o. a
                                   •o o
                                    §°
                                      tH
                                    a 4i
                                    O 60
                                    (0 -H
                                    B. a
                                                          H 3     41

                                                          5-S:    §
                                                          (a O     B
41
t-i
•g-5
•H 00
gg
   41
                                      o
                                      r»
                                      CM
                                 •O

                                  §


                                 S g
                                 CM S

                                  ••*

                                 S CM
                                                                                   O
                                                                                   O
                                                                                   ft
                                                                                                          S
                                                                                                           U
                                      41

                                   S:s
           fo,
                                      41  B
                                      U  4>
                                         4>    rH VO  C
                                        •H    -H ••  O
i!
              4> a
              S 3
o a
a B. 41
             •S5
                                 M n 41 u
                                 •H^^ O B
                                 B   B 41
                                 OH    It
                                 W 41 X T3
                                 41 4J O a
                                 o a o n
                                 a » fH 60
I
  §^^
  m
 a CM
fi "
 u 
                                                        o
                                                        B
 a m
 o 
-------
                                                                        cs
                                                                        vo
 §
•H
-U
 G
 01
CO
             a
             o.
            r-l

            O
              •

            O
                                                                        60
                                                                        3.
                                   .a
                                    no


                                   in

                                   o
                                               m >

                                               CM cj
                                                                       g
                                                                       o
                 0)
                 B
                           •O
                            B
                            Q
            •ri -H a
            u u 3
            U O rH
            a rH O
            p.ta o
                                               
                I
                 o
                 E
                 13 ^N
                j: rt
            1
            a.
            2  g
            frfl
                                I  U rH
                               t-i a ••
                                o o H

                                g «~
                                                                       I
                                                                      •o
                                                01
                                             U  rH
                                             in  -H
               01


               x PI


              •o <-*
                                                B  rH
                                                O  PI
                                                                                   o
                                                                                   0)
                                                                                  •a
                a -a
                o o
                   ea
                o u
                a a
                •o >

                  •
                           a
                           u  B
                          •H  o
                                             a
                                             u
                                            •rl
                               •8
                               a,
                               o o
                                                a.

                                               •a
                                                «
                                                to
                           0) 4J
                          rH O
                          J3 01
                           a .c
               i-l IS
               •rt  O
                a  o
                a M-I
 u
•H
ja

 o
                a
                0)
                u
               =1
                                              ea m
                                              o -a
                                                        ••H  CO
                                                      •O i-l  O
                                                       o     K
                                                       o -a  o

                                                      2  §1
                                                                                  -83
                                                        •H   01

                                                         g   «"
                                                         Lj   M

                                                         s   S  "S

                                                         «•   3  11
                                                         «  13   n
                                                         r<   B  4   u
                                    P  B
                                                                 01
                                                                 
-------
                                                            Section 7K
   this  Manual.   In addition,  TLC may be used as  a screening procedure
   followed by confirmation and quantitation using GC,  or the quantitation
   can be carried out by TLC if a gas chromatograph is  not available or if
   the pesticide of interest is unstable during GC.   Extraction,  cleanup,
   and concentration steps normally precede TLC determination.  Often more
   stringent cleanup is required for TLC than for GC if streaked  zones are
   to be avoided.  For example, the 15% diethyl ether fraction from the
   Florisil column cleanup of  a fat sample contains a large amount of lipids.
   Although adequate for GC, further cleanup prior to TLC is required (EPA
   PAM,  Section 12,B,V).  TLC  has also been occasionally used for cleanup
   of extracts prior to determination by GC (100).

   Major advantages of TLC are simplicity, rapidity, and low cost.  Sensi-
   tivity ranges from about 5-500 ng for most pesticide detection methods.
   Rapid semi-quantitative estimation can be achieved by visual comparison
   of sample and standard spot sizes and/or intensities, and more accurate
   quantitation can be carried out by in situ scanning  of spots with a
   spectrodensitometer.

   This  section will briefly survey important aspects of TLC for  screening
   and quantitation of pesticide residues.  General techniques of TLC were
   described in detail in an extensive treatise (101),  while specific
   procedures for pesticide TLC were covered in several papers (97, 102).
   Applications of TLC to pesticide analysis have been  reviewed (103, 104).


7K   PRACTICAL CONSIDERATIONS  IN TLC

   Spots are applied to the thin layer using simple disposable capillaries,
   GC syringes, or automatic multiple spotting devices.  All initial zones
   should be of small, uniform size, and only enough pesticide is spotted
   to allow for detection after the run.  Care should be taken that the
   spotting pipet does not penetrate the surface  layer.  Standard solutions
   must  be spotted on the same plate as the sample, preferably on both sides
   of the sample spot.

   Layers are hand-coated with a commercial adjustable  spreading  device
   (Figure 7-B) or purchased pre-coated on glass, plastic, or aluminum
   backing.  Analytical layers are usually 250 ym thick.  Pre-coated layers
   are of high purity and uniformity and are used almost exclusively in
   most  laboratories, especially for in situ quantitation by densitometry.
   Substitution of one brand of adsorbent for another or pre-coated for hand-
   coated plates often cannot  be directly made.  For example, silica gels  with
   differing polarities or surface hardness (binders) may require modified
   solvent systems or detection reagents if similar results are to be
   obtained.  Pre-coated silica gel plates, especially  those prepared with
   an organic binder, are generally used as received from the manufacturer
   without activation.

   Silica gel and alumina layers usually give the best  results, but polyamide,
   microcrystalline cellulose, kieselguhr, zinc carbonate (105),  and magnesium
                                      -265-

-------
                                                         Section 7K.

oxide, among other adsorbents, have also been used.  For reversed phase
TLC,  hydrophobic C,ft  chemically bonded silica gel plates are commercially
available.         *•*

      Figure 7-B   Desaga/Brinkmann Adjustable Applicator For Coating
                  Regular or Gradient Layer Plates, Brinkmann Instru-
                  ments, Inc.   -                   >>,,,.,,
Chromatography is carried out in a development chamber, most often a
rectangular, glass, paper-lined tank saturated with solvent vapors
(Figure 7-C). Low volume "sandwich" chambers are also used.  Both
saturated and unsaturated atmospheres have teen used to advantage and
should be tested for optimum results in any,'particular application.
Ascending development for a distance of 10-20 cm is typical.  It is
important to follow exactly all stated conditions when attempting to
reproduce a separation.  The temperature, development chamber design
and equilibration, and water content of the adsorbent are probably
the most frequent sources of variation among laboratories.

The technique termed "high performance thin layer chromatography"
(HPTLC) has become increasingly important for separations and in situ
quantitative analysis in the recent past.  HPTLC is carried out on
10 x 10 cm, 7.5 x 7.5 cm, or 5 x 5 cm pre-coated layers of silica gel
with a smaller particle size and a narrower particle size distribution
than in conventional TLC plates, thereby giving improved resolution and
sensitivity of detection.  Volumes no larger than 1 PI must be spotted
for these advantages to be fully realized.  For manual application,
spotting is usually done with a Pt-Ir tipped Nanopipet (or equivalent),
or this type of pipet is used with an automatic spotting device that
controls both the pressure of the pipet tip on the layer and the duration
of contact.  Solvent development is carried out in a miniature glass
                                   -266-

-------
                                                         Section 7K

rectangular chamber or in a commercial, automatic U-chamber device pro-
ducing radial.zones (106) (a special radial scanner is needed to quanti-
tate these separations).  High resolution is achieved rapidly with short
development distances (i.e., 5 cm or less).

In a typical residue analysis, it is virtually impossible to apply the
whole cleaned-up sample extract or an appropriate, accurate aliquot, as
a spot of 1 yl or less, so HPTLC has not yet been widely used for actual
samples.  New approaches are appearing that may solve this problem by
allowing a larger sample to be applied without sacrificing the benefits
of the HP layer.  One proposed solution utilizes a two-section plate
with a high performance analytical layer above a spotting region; initial
development concentrates the diffusely applied sample into a narrow
zone at the interface of these layers (107).  Another possibility is
the use of programmed multiple development (apparatus from Regis Chemical
Co.)» which causes large initial spots to be narrowed during migration
on the HP layer (108).  HPTLC plates are available from Merck and What-
man, and HPTLC equipment from Camag, Inc. and Fotodyne Corp.

Figure 7-C.  Desaga Rectangular TLC Tanks, Brinkmann Instruments, Inc.
The following solvent systems have proved tov be generally useful for
separation of a wide range of pesticides oh silica gel thin layers:
benzene mixed with varying amounts of ethanol for polar compounds or
with hexane  for those which are  less polar; and a mixture of hydro-
carbon plus  acetone plus  chloroform, with the addition of methanol for
more polar pesticides.  Examples include pentane-acetone-chloroform
(65:30:5 v/v) or pentane-acetone-methanol-chloroform (70:15:10:5 v/v).
The purpose  of the chloroform is to control the evaporation of acetone
in the atmosphere of an unsaturated tank.  Proportions of the com-
ponents are  changed to suit the  requirements of specific separations.

After development and air drying of the layer, spot detection may be
achieved in  a number of ways.  Few pesticides are naturally colored, but
                                    -267-

-------
                                                            Section 7L
   colored derivatives may be made prior to spotting, e.g., dyes, formed
   from aromatic amine moieties of urea herbicides by coupling with
   N-ethyl-1-naphthylamine (FDA PAM, Vol. II, Sec. 120.216).  Colorless
   spots can be detected by applying a chromogenic reagent, either by
   spraying or dipping.  A commercial aerosol spray device is shown in
   Figure 7-D.  Dipping is the preferred method of application, if feasible,
   because of the uniformity achieved and the hazards involved in careless
   spraying of corrosive, toxic, or carcinogenic reagents.  A Thomas-
   Mitchell dipping tank is recommended.  Sometimes the reagent can be
   incorporated in the layer prior to development or included in the
   developing solvent.  Naturally fluorescent spots can be detected under
   short (254 nm) or long (366 nm) wave DV light, or fluorescence may be
   induced by application of fluorogenic reagents after development or
   preparation of fluorescent derivatives (e.g., dansyl compounds) prior
   to spotting (109).  Spots that absorb UV light are detected as quenched
   (dark) spots on layers containing phosphor activated by UV light (usually
   254 nm).  Radioactive (labeled) pesticides are detected by autoradiography
   and some fungicides are detected by direct bipautography.


        Figure 7-D.  TLC Aerosol Sprayer, Brinkmann Instruments,  Inc.
TL   QUANTITATIVE TLC

   Quantitatlon of separated spots may be achieved by "eyeball" comparison
   between sample and standard spots run on the same plate or by some inde-
   pendent analytical method (e.g., spectrophotometry or GC)  after scraping
   the spot, collecting, and eluting the pesticide from the adsorbent.
   Manual elution is simply carried out by scraping the area containing the
   pesticide spot, collecting the scrapings in a vial or tube, adding
   solvent and agitating (vortexing), filtering the adsorbent, and concen-
   trating the filtrate containing the pesticide.  An automated elution
   system is available from Camag, Inc. (110).  Radioactive spots can be
   quantitated by scintillation counting after scraping or by automatic
   scanning of radioactivity on the layer.
                                     -268-

-------
                                                         Section 7L


Colored, fluorescent, or quenched spots may be scanned on the layer when
a spectrodensitometer is available.  Quantitation is achieved by scanning
sample and standard spots in the optimum instrumental mode and treating
the resultant peaks, representing the amount of light absorbed or
emitted, in the same manner as GC peaks for calculations.  A versatile
densitometer is capable of scanning in single or double beam and re-
flectance or transmission modes, and has monochrometers or filters
for selection of the best wavelengths of incident and emitted (for
fluorescence) energy.  Important considerations for densitometry are
adequate extract cleanup (111), precise and accurate spotting, uniform
layers, Rj? values between 0.3 and 0.7, uniform application of detection
reagents, and optimum' use of a good densitometer. •

Fluorescence densitometry .has proven to be the most advantageous mode
for pesticide analysis in terms of sensitivity and selectivity.  If
the compound is naturally fluorescent (e.g., benomyl, Maretin, quino-
methionate), the procedure is usually straightforward and measurements
can be made immediately after separation.  Sensitivity and reproduci-
bility are usually very high because no reagents are added or sprayed
on the chromatogram to produce background fluorescence..  For the majority
of pesticides that are not fluorescent, however, some kind of treatment
is required.  Possibilities for producing fluorescence include treatment
of the plate with heat, acid, base, inorganic salts, or a combination of
these; preparing a derivative in solution before spotting; or applying
a fluorogenic reagent to the layer after separation.  All of these
options are included in the papers cited in Table 7-2.

Manual spotting is best performed with 1 or 2 yl Microcap disposable
pipets, using repeated spotting with drying in between for larger volumes.
It has been shown that sample delivery errors below 1% are feasible with
Microcaps (112).  Larger volumes of sample extracts are conveniently and
reproducibly spotted with a device such as the Kontes automatic spotter
that applies milliliter volumes of one to six samples or standards in
small, uniform zones with little operator attention.  Solutions are
loaded into 5 ml capacity glass tubes and are delivered onto the layer
through Teflon coated needles, the rate of flow and spot size being
controlled by a stream of nitrogen or air focused onto the spotting
location.  The Kontes spotter is pictured in Figure 7-L and described
in reference (113).  Manual spotting of larger volumes onto commercial
plates with an inert -preadsorbent spotting area is quickly and re-
producibly done with a Drummond digital microdispenser.  Samples and  •-•
standards, including total extracts, so applied are narrowed to a common,
small initial zone size at the silica gel interface (114).

Pre-coated layers are recommended for quantitative TLC since it is very
difficult to hand-coat layers with adequate uniformity.  They are
generally purified before use by a predevelopment with chloroform-
methanol (1:1 v/v) followed by evaporation of the solvent in a dust
free atmosphere.  Uniform application of chromogenic or fluorogenic
reagents is better achieved by dipping than by spraying.  However,
dipping is not always possible; its use depends on the reagent solvent,
adsorbent, and type of compounds on the layer.
                                    -269-

-------
 Figure 7-E.
                                                          Section  7L
Fiber Optics Thin Layer Scanner and Automatic Spot Applicator,
Kontes Glass Company, Inc.
When a new densitometric method is developed, the spots of interest should
be scanned in all possible modes and directions and at a variety of wave-
lengths in order to obtain the best signal to noise ratios and selec-
tivities.  The optimum conditions are then used to obtain the calibra-
tion curve (linear range) and perform the analysis.  Samples and
bracketing standards should always be chromatographed on the same plate.

Thin layer densitometry is capable of precision of 1-2% on a routine
basis and can rival GC and HPLC for determination of certain pesticide
residues in the hands of an experienced operator.  A book covering the
principles and experimental details of thin layer densitometry, including
a chapter on pesticide analysis, has been published (115).  Table 7-2
contains some recent, selected applications of thin layer densitometry.
A fiber optics scanner specifically designed for pesticide analysis (116)
is available from Kontes (Figure 7-E).
                                   -270-

-------
                     Section 7L
TABLE 7-2
Compounds
Acidic herbicides

Bayrusil
Benomyl
Benomyl, carben-
dazin, and 2-AB
Captan, captafol
Carbaryl

Carbaryl

Chloranben

Chlorophenoxy
acid herbicides
Coumaphos
Coumaphos
Coumaphos and
0-analog
DDT
DDT
Fenitrothion
Fenitrothion,
breakdown
products, and
related compounds
Gibberellins
PESTICIDES QUANTITATED BY THIN
Sample matrix Scanning mode
.standards only

foods
cucumber
fiuita, vegetables

apple, potato
potato

apples, water,
lettuce
bean, tomato

water

water
water
eggs

water
water
water
standards



apple pulp
fluorescence

fluorescence
fluorescence
quench

fluorescence
visible

visible

visible

visible

fluorescence
fluorescence
fluorescence

visible
visible
fluorescence
fluorescence



fluorescence
LAYER DENSITCttffiTRY
Detection" method
4-bromoethyl-7-
methoxycoumarin
heating
—
fluorescent layers

NaC103
£-nitrobenzenedi-
azonlum fluoborate
p_-nltrobenzenediazo-
nium fluoborate
Bratton-Karshall
reagent
AgH03

heating
heating
heating

AgN03
AgN03
SnCl^fiuorescanlne
fluorescamine



H,SO.
Reference
'(125)

(134)
(144)
(118)

(117)
(135)

(142)

(133)

(122)

(127)
(141)
(131)

(135)
(137)
(121)
(145)



(140) -
-271-

-------
                                                                                 Section  7L
TABL6 7-2 (Continued)
   CoirpoundB	
 SampJe matrix
Scanning mode    Detection method
                      Reference
Glyphosate (via
H-nitroso
derivative)
Herbicide* contain-
ing KHj or OH groups
Karetln
HCPA and
Tcrbicil
OC1 pesticides
OF insecticides

OF pesticides

OF pesticides
Qulnocethlonate
Thiabcndazole
Thiourea
B-Trinzines
Triasines
shoots and
roots

water, soil

milk,, eggs
apples

human autopsy samples
standards only

vater

tissues
crops
fruits
citrus fruits
standards only
vater
fluorescence

fluorescence

fluorescence
UV absofbance

visible
visible

fluorescence

visible
fluorescence
fluorescence
UV absorbance
quench
visible
fluorescamine
dansyl chloride
AgN03
AgNOj or enzyme
inhibition
hydrolysis/dansyl
chloride
palladium chloride
fluorescent layers
iodine
(138)

(119)

(129)
(143)

(123)
(126)

(128)

(124)
(130)
(132)
(136)
(120)
(139)
 Earlier analyses are reviewed by J. D. MacKell and R. W. Frei in J. Chromatogr.  Sci.,  13,  279  (1975).
                                                      -272-

-------
                                                            Section 7H
7M   THIN LAYER SYSTEMS

   a.  Chlorinated Pesticides

       Extracts of fatty and nonfatty foods cleaned-up on a Florisil
   column are chromatographed on prewashed alumina layers developed with
   heptane (for the 6% diethyl ether-petroleum ether Florisil eluate) or '
   2% acetone in heptane (15% diethyl ether fraction).  Detection is pro-
   vided by spraying with AgN03-2-phenoxyethanol reagent in ethanbl or
   acetone and exposing to high intensity short-wave UV light to produce
   brown to purplish-^black spots.  The construction of a DV light
   apparatus containing four 15 watt lamps for rapid color development
   and allowing a variable distance between the TLC plate and the light
   source is described in the Canadian PAM, Section 14.10*  Thin layer
   media must be very low in chlorine content, and other precautions and
   care must be taken to prevent large areas of the plate from turning
   brown or gray, thereby reducing the contrast of the spots with back-
   ground.  A sensitivity in the 5-500 ng range is possible with AgNOs
   reagent, with a light steaming before spraying often aiding the detection.
   Conventional 20 cm x 20 cm glass plates, commercial pre-coated TLC sheets,
   or 3-1/4 inch by 4 inch microslides, may be employed.  Complete details of
   these methods plus Ry values for numerous compounds in the aforementioned
   two solvent systems, as well as for an alternative system consisting of
   immobile dimethylformamide on alumina and isooctane as the developing
   solvent, are given in Sections 410, 411, and 413 of the FDA PAM.  Silver
   nitrate has been incorporated into acid-washed alumina before the plates
   are coated so that only exposure to UV light is required for spot visualiza-
   tion  (FDA PAM, Section 412).  The AgN03 detection method has recently been
   studied in detail for the determination of chlorinated insecticides and
   herbicides (146).
               0                 •  •                  •
   Similar TLC procedures are described in detail in the EPA PAM, Section
   12,B  for the determination of chlorinated pesticide residues in. serum
   and adipose tissue.  An extract from 50 g of serum, cleaned-up on Florisil
   and concentrated to 100 yl before spotting, will produce a visible spot
   at 2  ppb, assuming that 10 ng of pesticide is detectable.  An adipose
   tissue extract from a 5 g sample, concentrated to 500 Ul,. will give a
   readable spot at 10 ppb.  The method involves TLC of the 6% and 15%
   Florisil column eluates as above, with additional prior cleanup of the
   15% fraction on an alumina micro-plate developed with acetonitrile.

   Silica gel layers developed with hexane, 1% acetone-hexane, 10-50%
   benzene-hexane, or 1% ethanol-hexane are recommended for screening
   chlorinated pesticides in foods at 0.1 ppm levels  (Canadian PAM, Pro-
   cedures 9.1 and 12.4).  Complete details of plate preparation, extract
   concentration, and visualization with silver nitrate reagent are given
   in  this source, along with figures of spot locations for eleven common
   pesticides  in  four mobile solvents.
                                       -273-

-------
                                                         Section 7M


For complex pesticide mixtures, tvo dimensional or multiple development
techniques may be helpful.  The former was used to identify organochlorine
pesticides in blood and tissues (147) and the latter  (148) for the separa-
tion of 13 common pesticides.

Extensive listings of additional solvent systems, corresponding Rp values,
and detection reagents for chlorinated pesticides will be found in
references (103, 123, 149, and 150).  In reference (123), 26 solvent
systems and 14 chromogenic reagents are evaluated for the determination
of 12 organochlorine pesticides in blood, urine, and tissue samples.


b.  Organophosphorus (OP) Pesticides

    Cleaned-up extracts may be developed with methylcyclohexane on DMF-
coated alumina layers and detection made by spraying with tetrabromo-
phenolphthalein ethyl ester, AgN03, and citric acid.  This reagent reacts
only with thiophosphoryl compounds to give blue or magenta spots (FDA
PAM, Section 431; EPA PAM, Section 12,B).  Thio and nonthio organophos-
phates are developed on silica gel layers with isooctane-acetone-chloro-
form (70:25:5 v/v) and .detected as blue or magenta spots by treatment
with 2^nitrobenzyl pyridine and tetraethylpentamine spray (FDA PAM,
Section 432).

A two dimensional procedure (IDA PAM, Section 614.11; 151) has the
significant advantage of specificity, obtained by bromine vapor oxidation
of the OP pesticides before development in the second direction.  Silica
gel layers with toluene, 25 percent heptane in ethyl acetate, or ethyl
acetate as developing solvents were used along with the Storherr charcoal
column cleanup procedure and enzymatic detection with commercial ho-rse
serum cholinesterase and indoxyl acetate to identify 18 pesticides in
crops at 0.01 ppm levels.  A sandwich type chamber is specified for
development to obtain the requisite resolution and sensitivity.  The same
procedure should be well suited to OP pesticides in human and environ-
mental samples after appropriate cleanup.

Enzyme inhibition techniques are important for the selective and sensitive
(pg-ng amounts) detection of enzyme inhibitors such as OP and carbamate
insecticides and metabolites.  These compounds inhibit esterases and
thereby prevent hydrolysis of a chromogenic substrate.  Procedures include
separation by TLC (on silica gel layers sometimes as thick as 450 Jim),
optional treatment with bromine vapor or UV light, and spraying of the
layer with enzyme and substrate solutions.  Areas corresponding to
inhibitors are visible as white spots on an Intensely colored background;
i.e., inhibited enzyme is surrounded by enzyme free to hydrolyze the sub-
strate and thus produce color.  While many OP pesticides are inhibitors
per se, bromine or UV treatment is required to convert others to active
inhibitors.  For carbamates, UV or bromine treatment may produce no change
or increased or decreased inhibition, depending on the compound.  Sample
extracts often require minimal cleanup prior to TLC analysis with enzymatic
detection; for example, hexane extracts of many foods can be directly
chromatographed.  Section 9.2 of the Canadian PAM provides procedural
                                   -274-

-------
                                                          Section 7M
 details,  tables of sensitivities and effects of bromine and UV treat-
 ment for.OF and carbamate pesticides, and diagrams of mobilities with
 hexane-acetone (8:2 v/v), a generally useful development solvent for
 TLC on 450 Vim silica gel layers.  The preparation of these layers is
 detailed  in Section 12.4 of the Canadian PAM.  Several different
 esterases have been compared for the detection of 65 OF and carbamate
 pesticides in vegetables and fruits (152).

 TLC enzyme inhibition methods and applications to pesticides have been
 reviewed  (153-155) as have the merits of TLC for analysis of residues
 (156). The separation and detection of 42 phosphate compounds using
 five ternary solvent systems on three adsorbents and three selective
 chromogenic sprays have been reported (157).  Twenty five solvent
 systems and several visualization reagents were evaluated for detection
 of 12 OF  insecticides in tissues (124).


 c.  Chlorophenoxy Acid Herbicides

     Extracts containing methylated chlorophenoxy acids are cleaned-up on
 a Florisil column and chromatographed on alumina layers using hexane
 saturated with acetonitrile as the developing solvent.  Cleaned-up
 extracts  containing free acids are developed for a distance of 3.5 cm
 on a pre-coated silica gel sheet with cyclohexane->acetic acid (10:1 v/v),
 then the  sheet is dried and developed for 15 cm in the.same direction
 with benzene-petroleum ether (3:1 v/v).  Spraying with silver nitrate
. chromogenic reagent produces black spots with a sensitivity of ca 50 ng
 for the esters and 100-500 ng for the free acids.  Details of both
 methods and % values are given in Sections 421 and 422 of the FDA PAM.
 Other detection reagents for these pesticides include Rhodamine B and
 Bromocresol green indicators (158).


 d.  Other Pesticide Classes

     The TLC of other classes of pesticides including carbafflates, ureas,
 phenols,  dithiocarbamates, triazines, and organomercurials was reviewed
 in references  (103) and  (104).  Applications, solvent systems, detection,
 and quantitation are covered in these references.  TLC is particularly
 applicable to herbicides, many of which are polar and not susceptible
 to gas chromatographic analysis without derivative formation.  Studies
 have been reported for the TLC of triazine herbicides on silica gel
 (120, 159) and polyamide (160); determination of 11 urea herbicides in
 water (161); detection of dithiocarbamate fungicides with Congo red (162);
 separation of carbamate and phenylurea pesticides on polyamide  (163);
 comparison of six reagents for detection of carbamate and phenylurea
 pesticides on EPTLC plates (164); and separation and identification of
 carbamate pesticides in post mortem material  (165).

 The TLC of five dithiocarbamate residues  in chloroform extracts of leaves
 is detailed in Section 9.3 of the Canadian PAM.  Silica gel layers developed
                                     -275-

-------
                                                             Section 7N


   with benzene for dimethyldithiocarbamates or acetic acld-methanol-benzene
   (1:2:12 v/v) for ethylenebisdithiocarbamates are used,  with detection as
   yellow, brown, or green spots after  a cupric chloride-hydroxylamine
   hydrochloride spray.


7N   REFERENCES

   (1)  Eisenbeiss, F., and Seiper, H.,  J.  Chromatogr..  83.  439 (1973).

   (2)  Larose, R. H., J. Assoc. Off. Anal.  Chen.,  75_,  1046  (1974).

   (3)  Summermann, W., Rohleder, H., and Korte,  F.,  Z.  Lebensm.  Unters.
        Forsch., 166. 137 (1978).

   (4)  Mittelstaedt, W., Still, G. G.,  Durbeck,  H.,  and Fuhr,  F.,  J.  Agr.
        Food Chem.. 25,, 908 (1977).

   (5)  Snyder, L. R., J. Chromatogr. Sci..  15,  441  (1977).

   (6)  Baumann, F., Pecsar, R. E., Wadsworth,  R., and  Thompson,  B., Amer.
        Lab., p. 97, September  (1976).

   (7)  Horvath, C., Melander, W., Molnar,  I.,  and Molnar, P.,  Anal. Chem.,
        14, 2295 (1975).

   (8)  Majors, R. E., J. Chromatogr. Sci..  15,  334  (1977)..

   (9)  Saunders,  D. L., J.  Chromatogr.  Sci.. 15, 372 (1977).

  (10)  Bradley, T., Spectra-Physics Chromatography Review.  4^(3) , 8  (1978).

  (11)  Horvath, C., and Melander, W., J. Chromatogr. Sci..  15, 393  (1977).

  (12)  Greenwood, J. V... Spectra-Physics Chroma tography. Review. J3  (2),  1  (1977),

  (13)  Snyder, L. R., and Saunders, D.  L.,  J.  Chromatogr. Sci..  7_t 195  (1969).

  (14)  Snyder, L. R., J.  Chromatogr.. 92. 223  (1974).

  (15)  Bakalyar,  S. R., Amer. Lab., p.  43,  June  (1978).

  (16)  Dlebold, G. J.,  and Zare, R. N., Science. 196. 1439  (1977).

  (17)  Rucki,  R.  J., Talanta. Z7, 147 (1980).

  (18)  Kissinger, P.  T.,  Anal. Chem.~. 49(4). 447A (1977).

  (19)  Lores,  E,  M., Bristol, D.  W.,  and Moseman, R. F., J. Chromatogr. Sci.,
        16, 358 (1978).                                    	''—	
                                      -276-

-------
                                                          Section  7N


(20)  Willmott, F. W., and Dolphin, R. J. ,  J. Chromatogr. Sci., 12,
      695 (1974).

(21)  Ettre, L. S.,' J. Chromatogr. Sci.. 16, 396  (1978).

(22)  Julian, B. G., Vandenbom, H. W., and Kirkland, J. J.,  J.  Chromatogr.,
      112, 443 (1975).

(23)  Krull, I. S., Goff, U., and Ashworth, R. B., Amer. Lab.,  p.  31,
      August (1978).

(24)  Chamberlain, W. J., Walters, D. B., and Chortyk, 0. T., Anal.  Chin
      Acta, 2i» 213 (1975).

(25)  Rabel, F. M.. Amer. Lab.. p. 81, January  (1980).

(26)  Kirkland, J. J. «• Chromatographia. £, 661  (1975).

(27)  Tseng, P. K., and Rogers, L. B., J. Chromatogr Sci., 16,  436 (1978).

(28)  Berg, J., Varian Instrument Applications,  11(2), 5  (1977).

(29)  Bakalyar, S. R., and Henry, R.  A., J. Chromatogr.,. 126.  327  (1976).

(30)  Lawrence,- J. F., and Turton, D., J. Chromatogr., 159.  207 (1978).

(31)  Pribyl,  J.,  and Herzel,  F., J.  Chromatogr., 166, 272  (1978).

(32)  Conlon,  R.  D., Anal. Chem.. 41(4). 107A (1969).

(33)  Munk, M. N., J. Chromatogr. Sci..  8. 491  (1970).

(34)  Veening, H., J. Chem.  Educ.. 41(9),A549 (1970).

(35)  Taggart, W.  P., Industrial Research, p.  76, February  (1974).

(36)  Karasak, F.  W., Research/Development,  24_(6), 52  (1973).

(37)  Brown, P.  R.,  High Pressure Liquid Chromatography,  Academic Press,
      N.Y.,  1973.

(38)  Parris,  N.  A.,  Instrumental Liquid Chromatography,  Elsevier
      Scientific Publishing  Co., N.Y.,  329 pp.  (1976).

(39)  Jackson, M. T., Jr., and Walker,  J.  Q., J. Chromatogr. Sci.. 15,
      490 (1977).

(40)  Creed,  C.  G.,  Research/Development.  27(9), 40 (1976).

(41)  Ivie,  K.,  Chapter 2 in Analytical Methods for Pesticides and Plant
      Growth Regulators,  Vol.  XI, Zweig, G.  and Sherma,  J.,  eds., Academic
      Press,  NY (1980).
                                      -277-

-------
                                                           Section 7N
 (42)

 (43)

 (44)


 (45)


 (46)

 (47)



 (48)

 (49)


 (50)

 (51)


 (52)

 (53)

 (54)


 (55)



 (56)


 (57)


 (58)


(59)


(60)
 Moye, H. A., J. Chromatogr. Sci.. 13, 268 (1975).

 Schooley, D. A., and Quistad, G. B., Prog. Drug Metab.r 3_, 1-113  (.19  9).

 Self, C., McKerrell, E. H., and Webber, T. J. N., Proc. Anal. Div.
 Chem. Soc.. 12, 288 (1975).                       ~~"~"—~~

 Hoodless, R. A., Sidwell, J. A., Skinner, J. C., and Treble, R. D.,
 J.  Chromatogr.. 166, 279 (1978).

 Werner,  W., GIT Fachz. Lab.. 22_, 785 (1978).

 McFadden, W. H., Bradford, D, C., Games, D.  E., and Gower, J. L.,
 Amer. Lab., p.  55, October (1977); McFadden, W. H., Schwartz, H. L.,
 and Evans, S.,  J. Chromatogr.. 122, 389 (1976).

 Arpino,  P. J.,  and Guiochon, G., Anal.  Chem.. 51(7). 682A (1979).

 Vestal,  M. L.,  NBS Spec. Publ. 519 (Trace Org.  Anal.; New Front.
 Anal. Chem.), 647 (1979).

 Oehler,  D. D.,  and Holman, G. M.,  J. Agr.  Food  Chem..  23,  590 (1975).

 Moye, H.  A., and Scherer, S. J., oral presentation at the 91st AOAC
 Meeting,  October 17-20, 1977, Washington,  DC, Abstract No. 79.

 Lawrence,  J. F.,  and Leduc,  R., J.  Agr.  Food Chem.,  25 » 1362 (1977).

 Aten,  C.  F., and Bourke, J.  F., J.  Agr.  Food Chem..  25. 1428 (1977).

 Farrow,  J.  E.,  Hoodless, R.  A., Sargent, M., and Sidwell,  J. A.,
 Analyst.  102, 752 (1977).

 Mtmdy, D.  E., and Machin, A.  F., J.  Chromatogr..  139.  321  (1977);
 Mundy, D.  E., Quick, M.  P.,  and Machin,  A. F.,  J.  Chromatogr..
 121,  335  (1976).

Paschal, D.  C.,  Bicknell, R.,  and Dresbach,  D., Anal.  Chem.. 49,
 1551  (1977).                                    	  —

 Bjornsson,  T. D.,  Blaschke,  T. F.,  and Meffin,  P.  J.,  J. Pharm.
 Sci.,  66,  142 (1977).                                 	

Kvalvag, J., Ott,  D. E.,  and Gunther, F. A.,  J. Assoc.  Off.  Anal.
 Chem.. .60,  911  (1977).                        ~"~
Farrington, D. S., Hopkins, R. G., and Ruzicka,  J. H.'A.,  Analyst.
102, 377 (1977).                                           	—

Selim, S., Cook, R. F., and Leppert, B. C., J, Agr. Food Chem.. 25.
567 (1977).                                 	a	~  —
                                   -278-

-------
                                                           Section 7N
'(61)  Byast, T. H.,  J.  Chromatogr.,  134.  216 (1977).

 (62)  Fasco, M. J.,  Piper,  L. J.,  and Kaminsky, L. S., J. Chroma togr.. 131,
      365  (1977).

 (63)  Lawrence, J. F..  J. Agr.  Food  Chen.,  25,  211 (1977).

 (64)  Brinkman, U. A.  Th.,  DeKok,  A., DeVries,  G., and Reymer, H. G. M.,
      J. Chromatogr.,  128,  101 (1976).

 (65)  Lawrence,  J. F.,  J.  Assoc. Off. Anal. Chem., 59, 1066 (1976).

 (66)  Maeda, M.,  and Tsuji, A., J. Chromatogr.. 120, 449  (1976).

 (67)  Pryde, A.,  and Darby, F. J., J. Chromatogr.. 115, 107 (1975).

 (68)  Pfeiffer,  C. D.,  J.  Chromatogr. Sci.. 14, 386  (1976).

 (69)  Byast, T.  H.,  and Cotterill, E. G., J. Chromatogr., 104, 211  (1975).

 (70)  Smith, A.  E.,  and Lord, K. A., J. Chromatogr., 107, 407  (1975).

 (71)  Whitacre,  D. M., Atallah, Y. H., Forrette,  J.  E., and Suzuki,  H. K.,
      in Analytical  Methods for Pesticides and Plant Growth Regulators
      Volume X,  Zweig, G.  and Sherma, J.,  eds., Academic  Press, OT,  p. 367
       (1981).

 (72)  Cabras,  P., Meloni,  M., and Pirisi,  F. M..  J.  Chromatogr..  176, 473
       (1979).
                                                                       i
 (73)  Ernst, G.  F.,  and Verveld-Roeder, S. Y., J. Chromatogr., 174,  269  (1979)

 (74)   Krause,  R. T., J. Chromatogr.  Sci.,  16_,  281 (1978).

 (75)   Mbring,  S. E., and McChesney,  J, D., J.  Assoc. Off. Anal.  Chem.,  62,
       774 (1979).
       Wong, L. T., Gnanaprakasam,  S., and  Thomas,  B.  H.,  J.  Chromatogr.,
       135, 149 (1977).
(76)


(77)


(78)


(79)

(80)  Pflugmacher, J., and Ebing, W., J. Chromatogr.,  151,  171  (1978).
       Thean, J. E., Fong, W. G., Lorenz,  D.  R.,  and Stephens,  T.  L.,
       J. Assoc. Off. Anal.  Chem.,  61,  115 (1978).

       DiPrima, S. J.,  Cannizzaro,  R. D.,  Roger,  J.-C.,  and Ferrell,  D.,
       J. Agr. Food Chem., 26_, 968  (1978).

       Lawrence, J. F.,  and  Leduc,  R.,  J.  Assoc.  Off. Anal. Chem., 61, 872
       (1978).
                                     -279-

-------
                                                            Section 7N


  (81)  Newsome,  W.  H.,  J.  Agr.  Pood Chen,. 26_, 1325 (1978).

  (82)  Bowman, M. C., Holder,  C.  L., and Bone, L. I., J. Assoc. Off. Anal.
       Chem.. ,61, 1445  (1978).                                   ~

  (83)  Lawrence, J. F.,  and Leduc,  R.,  J. Chromatogr.. 152, 507 (1978).

  (84)  Ott, D. E.,  J. Assoc. Off. Anal. Chem.. 62, 93 (1979).

  (85)  Cochrane, W. P.,  and Lanouette,  M., J. Assoc.  Off. Anal. Chem.. 62,
       100  (1979).                          	—	

  (86)  Brealey,  C.  J.,  and Lawrence, D. K.. J. Chromatogr.. 168,  461 (1979).

  (87)  Moye, H. A., and Hheaton,  T. A., J. Agr.  Food  Chem.. .27(2),  291 (1979).

  (88)  Crathorne, B., and  Watts,  C. D., J. Chromatogr..  169,  436  (1979).

  (89)  Stoeber,  I., Z. Anal. Chem.. 293. 370 (1978);   Stoeber,  I.,  and
       Reupert, R., Vom Wasser, 51, 273 (1978).

  (90)  Otsuki, A., and Takaku,  T.,  Anal. Chem..  J51(7), 833. (1979).

  (91)  Rus, V., Rev. Chim.  (Bucharest). 29(11),  1088  (1978).

  (92)  Ramsteiner, K. A.,  and Hoerman,  W.  P., J. Agr.  Food Chem., 27(5)
       934 (1979).                            		

  (93)  Lord, K. A., Cayley, G. R.,  Smart,  L.  E., and Manlove, R., Analyst.
       105, 257  (1980).                                           '	*—^

  (94)  Mills, P. A., J.  Assoc.  Off.  Agric. Chem..  42, 734  (1959); 44, 171 (1961).

  (95)  Mitchell, L.  C.,  J. Assoc. Off.  Agric.  Chem..  40,  999  (1957).

  (96)  Sherma, J., and Zweig., Paper Chromatography. Academic Press,  NY,
       Chapter 12 (Pesticides), pp.  355-396  (1971).

  (97)  Getz, M. E.,  Advances in Chemistry  Series 104.  ACS, Chapter  8,  (1971).

  (98)  Getz, M. E.,  Residue Reviews.  2.*  9  (1963).

  (99)  Coffin, D. E., J. Assoc. Off.  Anal.  Chem..  49,  1018 (1966).

(100)  Heatherington, R. M., and Parouchais,  C., J. Assoc.  Off. Anal. Chem..
       53, 146 (1970).                           ~~	"	

(101)  Stahl, E., Thin Layer Chromatography.  2nd ed.,  Springer-Verlag, NY  (1969)

(102)  Kovacs, M. F., Jr., J. Assoc.  Off.  Agric. Chem.. 46, 884 (1963): 47,
       1097 (1964).                          	~
                                      -280-

-------
                                                          Section 7N


(103)   Sherma, J., Chapter 1 In Analytical Methods for Pesticides and
       Plant Growth Regulators, Vol. XI, Sherma, J., and Zweig, G., eds.,
       Academic Press, NY (1980); also Chapter 1 in Volume VII of the
       same series (1973).

(104)   Sherma, J., Critical Reviews of Analytical Chemistry, CRC Press,
       pp. 333-338, August (1973).

(105)   Reichling, J., 2. Anal. Chem., 281, 139  (1976).

(106)   Vitek, R. K., and Kent, D. M.. Amer. Lab., p. 71, January  (1978).

(107)   Halpaap, H., and Krebs, K.-F., J. Chromatogr.. 142, 823  (1977).

(108)   Faber, D. B.. J. Chromatogr.. 142, 422  (1977).

(109)   Lawrence, J. F., and Frei, R. W., Chromatogr. Rev., 18,  253  (1974).
       (This  is  a  review  of theory, factors affecting emission, and appli-
       cations of  fluorometric derivatization  for pesticide  analysis  by
       HPLC and  TLC.)

(110)  Vitek, R. K.,  Seul, C.  J., Baier, M., and Lad, E., Amer. Lab..
       p.  109, February  (1974).

(Ill)  Sherma, J., Sample Preparation  for  Quantitative  TtC,  in Thin Layer
       Chromatography;   Quantitative Environmental  andjClinical Applica-
       tions, Touchstone, J.  C., and Rogers, D., eds.,  Wiley-Interscience,
       NY,  in press.

(112)  Emanuel,  C. F.,  Anal.  Chem.. 45, 1568 (1973).

(113)  Getz,  M.  E.,  J.  Assoc. Off.  Anal.  Chem.. 54, 982 (1971).

 (114)  Sherma,  J., Amer.  Lab., p. 105, October (1978).

 (115)  Touchstone, J. C., and Sherma,  J.,  Densitometry j.njlh.i'a Layer
       Chromatography. Wiley-Interscience, NY (1979).

 (116)  Beroza,  M., Hill, K.  R., and Norris, K. H.,  Anal. Chea., 40, 1608
        (1968).

 (117)   Francoeur, Y., and Mallet, V., J. Assoc. Off. Anal. Chem., _60_, 1328
        (1977).

 (118)   Polzhofer, K., Z. Lebensm. TJnters. Forsch., 163, xu9  (1977).

 (119)   Pribyl, J., and Herzel,  F., Z. Anal. Chem.. 286. 95  (1977).

 (120)   Jork, H., and Roth, B..  J. Chromatogr..  144, 39  (1977).

 (121)  Berkane, K., Cassie,  G.  E., and Mallet, V. N., J. Chromatogr..  139_,
        386 (1977).
                                      -281-

-------
                                                           Section 7N


 (122)  Sherma, J., and Koropchack,  J.,  Anal.  Chim.  Acta,  91, 259 (1977).

 (123)  Tewarii S. N., and  Sharma,  I.  C.,  J.  Chromatogr..  131, 275 (1977).

 (124)  Tewari, S. N., and  Harpalani,  S. P.,  J.  Chromatogr..  130. 229 (1977).

 (125)  Dunges, W., Chroma tographia. j),  624  (1976).

 (126)  Stefanac,  Z., Stengl, B., and  Vasilic, Z., J.  Chromatogr.. 124,
       127  (1976).                                                 	

 (127)  Volpe, Y., and Mallet, V. N.,  Anal. Chim. Acta.  81, 111 (1976).

 (128)  Lawrence,  J. F., Renault, C.,  and  Frei,  R. W., J.  Chromatogr.. 121,
       343  (1976).                                                    	

 (129)  Zakrevsky, J. G., and Mallet,  V. N., J.  Agr. Food  Chem..  23,  754
       (1975) .                              	 	—	  —•*.

 (130)  Francoeur, Y., and Mallet, V.  N.,  J. Assoc.  Off. Anal.  Chem.,  59,
       172  (1976).                                      ~~—

 (131)  Zakrevsky, J. G., and Mallet,  V. N., J.  Assoc. Off. Anal. Chem.. 58,
      ,554  (1975).	

 (132)  Ottender, H., and Hezel, U., J.  Chromatogr.. 109,  181 (1975).

 (133)  Sherma, J., and Touchstone, J. C., Chromatographia, _8,  261 (1975).

 (134)  Zakrevsky, J. G.,. and Mallet,  V. N., Bull. Environ. Contain. Toxicol.,
       13, 633 (1975).                          ~~~

 (135)  Sherma, J., Klopping, K., and  Getz, M. E., Amer. Lab..  p.  66,
       December (1977).

 (136)  Mandrou, B., Brun, S., and Kingkate, A., J. Asspc. Off. Anal. Chem.,
       60, 699 (1977).                                          =""

(137)  Sherma, J., and Bloomer, K., J. Chromatogr.. 135,  235  (1977).

(138)  Young, J.  C., Khan, S. U., and Marriage, P. B., J. Agr. Food Chem.,
       25, 918 (1977).	

(139)  Wislowska, E., and Kostowska,  B.,  Chem. Anal.(Warsaw).  22,  975 (1977).

(140)  Winkler, V. W., in Analytical Methods for Pesticides and  Plant Growth
       Regulators, Vol. X, Zweig, G., and Sherma, J., eds., Academic Press,
       NY, pp. 545-559 (1978).

(141)  Mallet, V. N., and Volpe, Y., J. Chromatogr.. 97., 415  (1978).

(142)  Sherma, J., Kovalchick, A. J., and Mack, R., J. Assoc. Off. Anal.
       Chem., 61, 616 (1978).
                                    -282-

-------
                                                          Section 7N

(143)   Polzhofer, K., Z. Lebensm. Uhters. Forsch., 167,- 162  (1978).
                                                       •
(144)   Masao, M., Yasuo, I., and Hiroaki, H., Bull. Aerie. Chem. Insp.
       Stn., 17, 43 (1977).

(145)   Zakrevsky, J. G., and Mallet, V. N.,  J. Chromatogr.,  132, 315  (1977).

(146)   Ludwick, A. G., Lau, A. N. K., and Ludwick, L. M., J.  Assoc. Off.
       Anal. Chem., 60, 1077 (1977).

(147)   Eliakis, C. E., Coutselinis,  A. S., and Eliakis, D. C.,  Analyst,  93.,
       368  (1968).

(148)   Szokolay, A., and Madaric, A., J. Chromatogr.. 42,  509 (1969).

(149)   Walker, K. C., and  Beroza,.M., J. Aasoc.  Off. Agric.  Chem.,  46,  250
       (1963).

(150)  Thielemann,  H.,  Z.  Chemie Lpz..  14,  292  (1974).

(151)  Gardner, A.  M..'J.  Assoc.  Off. Anal.  Chem..  54.,  517 (1971).

(152)  Ernst,  G.  F.,  Pieterse,  C.,  and  Marterns, L.  J.  H., J. Chromatogr..
       133. 245 (1977).

(153)  Mendoza,  C.  E.,  Residue Reviews.  43, 105 (1972)  and 50_, 43 (1974);
       J.  Chromatogr..  78. 29  (1973).

(154)  Garner, A. M.,  FDA By-Lines, 2(3),  173,  January'(1972)„

(155)  Villeneuve,  D.  C.,  Advances in Chemistry Series 104. ACS, Chapter
       3 (1971).

 (156)  Abbott, D. C., and Thompson, J., Residue Reviews. 11, 1 .(1965);
       Abbott, D. C., and Egan, H., Analyst. 92, 475 (1967).

 (157)  Getz, M. E., and Wheeler, H. G., J. Assoc. Off. Anal. Chem..  51, 1101  (1968)

 (158)   Thielemann, H., Z. Anal. Chem., 272, 286  (1974).

 (159)   Van den Heede, M., and Heyndricks, A., Meded. Fac. Landbouwvet.
        Rilksuniv. Gent. 41, 1457 (1976) (Analytical Abstracts, abstract
        5G20, November, 1977; Chem.  Abst. 86. 84568c, (1977).

 (160)  Reichling, J., and Fischer,  H., Z. Anal.  Chem.. 115_,  670  (1975).

 (161)  Deleu, R., Barthelemy, J.-P., and Copin, A., J. Chromatogr..  134.,
        483  (1977).

 (162)  Ragab, M. T. H., Anal. Lett., £, 295 (1976).
                                      -283-

-------
                                                           Section 7N
(163)  Reichling, H., and Fischer, H., Z. Anal.  Chem..  276,  301 (1975),



(164)  Davies, R. D., J. Chromatoer.. 170, 453  (1979).



(165)  Tewari, S. N., and Singh, R., J. Chromatogr.. 172,  528  (1979).
                                    -284-

-------
                                   Section 8
                      SAMP LING,  EXTRACTION,  AND
                         PROCEDURES IN PESTICIDE ANALYSIS
   This section treats a .number of miscellaneous topics important in residue
   analysis.   These include general considerations for collection and ex-
   traction of extracts.  Specific procedures for extraction and cleanup of
   pesticide and metabolite residues are discussed in Section 9.
                    SAMPLE COLLECTION, PREPARATION,  AND STORAGE
8A   GENERAL CONSIDERATIONS IN SAMPLING

   Special considerations must be given to the procurements storage, and trans-
   portation of samples to be analyzed for pesticide residues.  Procedures
   should ensure, as well as possible, that the pesticides originally present
   have not undergone degradation or concentration and that potentially inter-
   fering impurities have not been added.  Plastics must be rigidly avoided
   as containers for samples to be examined by electron capture GC because
   minute traces of materials such as polyethylece may produce spurious re-
   sponses.  Similarly, metal containers may contain trace impurities such as
   oil films, lacquers, or rosin from soldered joints that will cause inter-
   ference in GC analysis.  In general, glass jars or bottles with aluminum
   foil or Teflon-lined lids are the most suitable sample containers, although
   it is sometimes possible for pesticides in stored extracts to be adsorbed
   onto the glass surfaces.  Glass containers should be carefully precleaned
   as outlined in Subsection 3L in Section 3.  Aluminum foil can be cleaned
   by agitating it in analytical reagent grade acetone followed by several
   rinsings with pesticide grade ethyl acetate and hexane.  Plastic containers
   may be used, if necessary, only when non-interference with the subsequent
   analysis has been proved at its limit of detection.  Important variables
   in the sampling and storage processes include the size of the sample,
   source, stability, contamination,  intended use, behavior of the pesticides,
   and the temperature and time of storage.

   Readers interested in  a more exhaustive discussion of sampling and storage
   procedures than provided in the following sections of this chapter are
   referred  to the publication "Guidelines on Sampling and Statistical
   Methodologies for Ambient Pesticide Monitoring"  (Monitoring Panel, Federal
   Working Group on Pesticide Management, Washington, DC, 1974),  This 60-page
                                        -285-

-------
                                                               Section 8B, 80

    manual contains  chapters  on .statistics  and  study design, air, soil, the
    hydrologic environment, estuaries,  fresh water fish, wildlife, foods and
    feeds,  and human tissues.   The  20-page  booklet entitled "Guidelines for
    Sampling  and Transporting Samples for Pesticide Residue Analysis"  (Federal
    Interdepartmental Committee on  Pesticides Check Sample Program, London,
    Ontario,  Canada,  April, 1979) contains  detailed information on dry feeds,
    plants  and soils,  food products, wildlife,  tissues, forest substrates,
    water,  and fish.   The influence of  sampling methods on residue analytical
    results,  sampling criteria, and statistics  of sampling data have been de-
    scribed (1).


8B   REPRESENTATIVE  VS. BIASED SAMPLING

    Samples collected for the purpose -of assessing tolerance infringements> such
    as with agricultural and food products, should be random and representative.
    To the  contrary, most environmental samples are deliberately chosen to be
    biased  in nature.  For example, a sample of water to be analyzed for the
    highest possible pollution  in a stream  or lake would best be "taken as a grab
    sample from the point of maximum pollution introduction (such as an effluent
    pipe from a factory) rather than from the center of the river where it might
    be most representative.  If,, on the other hand, the objective is an average
    residue profile of the entire body of water, the final sample would preferably
    be a composite of a number of subsamples taken at various locations and water
    depths.  Analysis of a sick bird or fish in the middle of a metabolic cycle
   would usually be more useful for determining any pesticide contamination than
    a dead specimen that is likely  to contain only metabolites.  Similarly, human
    stomach washings  (lavage)  taken at an early stage are more likely to contain
   parent pesticides and to be useful for indication of pesticide poisoning.

   It is important that the analyst be aware of these considerations and that
   he be consulted when the sample to be collected is decided so that it is
   valid for the purpose of the analysis and valuable time is not wasted on a
   worthless sample.


8C   SAMPLE CONTAINERS

   Section 2 of the EPA Pesticide Analytical Manual specifies suitable sample
   containers for various sample types.  These include wide mouth glass  bottles
   with Teflon or aluminum foil lined screw caps for autopsy  tissue  samples of
   less than 25 g, glass vials of at least  7 ml capacity for  blood (avoid  rubber
   or cork caps),  empty pesticide-grade solvent bottles for water samples, and
   pint or quart capacity mason jars for larger environmental or  agricultural
   samples.  Sample collection glassware should be scrupulously cleaned  as out*-
   lined in Section 3 of this Manual.   Special precautions must be taken in
   preparing glass containers and caps  and  taking samples for PCP  analyses
   because of the ubiquity of the chemical.  These are  outlined in Section
   5,A,(A),Ca),IV of the EPA  PAM.   Specimens  intended for organoehlorine com-
   pound analysis  are never wrapped directly in paper,  cardboard, or plastic.
                                        -286-

-------
                                                           Section 8C

It is common practice in some laboratories to wrap tissue or other samples
in aluminum foil prior to analysis.  Figure 8-A, part a, shows a gas chro-
matogram of a pentane rinse of the shiny side of commercially available
aluminum foil.  The amount of rinse injected corresponded to 2 sq. cm of
foil.  The GC conditions included the use of a 10% OV-210 on Gas-Chrom Q
column and a 63Ni EC detector, typical of those used for analysis of pesti-
cide's and PCBs.  Part b shows the corresponding rinse of the dull side of
the same foil.  In general, the amount of interfering material was found
to vary with the brand and lot of foil.  However, the risk of contamination
from this source dictates that aluminum foil not be used for packaging
samples without a thorough acetone prerinse (2).


         Figure 8-A.  Gas chromatogram of pentane rinse of aluminum
                      foil on OV-210 column with 63Ni electron capture
                      detector.  Amount injected corresponds to 2 sq. cm
                      of foil,  a « shiny side of foil; b « dull side of
                      foil (2).
                                       -287-

-------
                                                                Section 8D
   It is good procedure to clearly label collected  samples with all pertinent
   information such as a code number, date and  time of collection, type of
   sample, place and method of collection, description of collection site, size
   of sample,  etc.   All samples that are perishable are shipped to the laboratory
   in styrofoam containers with dry ice.  A detailed description of systematic
   procedures used  for receiving, numbering, and  storing environmental samples
   at the National  Monitoring and Residue Analysis  Laboratory, Gulfport, MS,
   has been published (3).  A strategy for documenting the chain of custody
   of samples  that  has the potential for being  used as evidence in a legal
   proceeding or agency enforcement action is detailed in Section IV of the
   EPA National Enforcement Investigations Center Pesticide Product Laboratory
   Procedures  Manual (see Section 3E of this Manual).


8D   SAMPLE COMPOSITING

   After collection of a valid gross sample, compositing or reduction to an
   analytical  size  sample may be required, especially  for agricultural and food
   samples.  The general requirement is that the  small analytical sample must
   be fully representative of the gross sample  collected.   The exact steps in
   the compositing  procedure will depend on the particular sample involved.
   Figure 8-B  shows typical steps in reduction  of a  gross  sample of an agri-
   cultural product collected in the field, during processing,  or at the'market.
       Figure 8-B.  Typical steps in  reduction of a gross sample
                           Remove peal or husk (if necessory) and
                           reduce size of large units by cutting or chopping
1
Peel or husk
i
JMix and quarte
|
Subsample
sib



0


1
Fiesh or kernel!
|
Mix and quarter
i
Subsample
6lb
                                                    if necessary, reduce
                                                    size of large units by
                                                    cutting or chapping
                                                          I
                                                      | Mix and quarter I
                             Subdivide
                                          Subdivide
                                                        Subdivide
                                     -288-

-------
                                                              Section 8E
8E   STORAGE OF SAMPLES
   As a general rule, samples should be analyzed as soon as possible after
   their collection.  If storage is necessary, it should be under prescribed
   conditions that preserve the integrity of the original sample.  Samples other
   than water are ordinarily stored in a freezer below 0°C, but, even then,
   physical and chemical changes may occur in either the sample or in the
   residues sought.  Because many pesticides are photodegradable, it is
   advisable to protect samples and any solutions or extracts from needless
   exposure to light.

   Tissue samples that are to be extracted within 24 hours may be held at normal
   refrigerator temperature (+2 to +4°C).  If extraction is not to be carried
   out within this time, the. samples should be deep frozen at -12 to -18°C.  If
   tissues are stored in a "self-defrosting" freezer in unsealed containers,
   the weight can markedly decrease due to desiccation.  If the tissues were
   not weighed prior to freezing, or if they are to be subdivided at a later
   time, this desiccation may make it impossible to relate the amount of sub-
   stance determined analytically to its original concentration in the tissue
   (2).  A related problem occurs when samples experience  repeated freezing
   and thawing.  Adipose tissue in particular has a tendency to "leak" lipid
   when  the cell membranes are disrupted by a freeze-thaw cycle.  In a series
   of experiments in which such cycles were deliberately applied to a collection
   of samples of adipose tissue from a rat, the apparent lipid content of the
   tissue  (mg per gram of tissue) decreased by an average of 10% after three
   freezings.  This loss was only apparent, and was not observed if the tissue
   was extracted in the original storage container  (2).

   Blood samples that are to be separated  for subsequent analysis of the serum
   should be  centrifuged as soon as possible  after  drawing.  If the serum is
   to.be analyzed within a  3-day period,  storage at +2  to +4°C is suitable.  If
   storage is  to be for  longer periods,  it is preferable to  deep freeze at
   -12  to  -18°C.   Otherwise, DDT may  degrade  in contact with broken red blood
   cells (hemoglobin).

   Agricultural or environmental  samples that are  to  be analyzed for organo-
   phosphates should be placed in tight containers  and  stored  in deep  freeze as
    soon as possible after sampling unless sample preparation is  to be  conducted
   within a very few hours.  No  difference was  found in measured residue  levels
    for a series of OP pesticides  when food samples were extracted  immediately or
    after storage at -17°C for several months  (4).

    Water samples should be extracted at once, if all all possible,  or stored in
    the dark at 4°C to avoid rupture of the container as a result of  freezing.
    Pesticides can be adsorbed on the glass container during storage,  so  the
    container should be rinsed with solvent if the extraction is not  made  in the
    container itself.  For carbamates, the sample is acidified  immediately after
    collection with sulfuric acid and 10 g of sodium hydroxide  are  added for
    each liter of sample.  Maximum storage is 24 hours for all compounds except
    chlorinated hydrocarbons, which can be held for up to 30 days.
                                           -289-

-------
                                                               Section  8F

   Whole fish can be stored for  up  to  six months  if an  even  temperature  of at
   least -26°C is maintained with a good  glaze on the sample and rapid initial
   freezing.   Homogenized  samples require less storage  space, but these  samples
   should be  monitored for stability of the  compounds of interest if held
   longer than one month.

   If lengthy storage is required prior to analysis, a  good  alternative  to
   storage of sample is to extract  the sample at  once,  remove most or  all of
   the solvent,  and  store  the extract  at  a low temperature.  Decomposition in
   samples that must be stored can  be  evaluated by storing spiked controls
   along with the samples.   Organophosphorus pesticides field-extracted with
   chloroform from water were successfully preserved for three weeks upon re-
   frigeration.   Of  the 16 compounds tested, only EPN and malathion were not
   stable (5).

   If freezing is not possible, wildlife  and fish samples may be preserved in
   formalin or alcohol.  Because  analytical results are usually in terms of
   wet weight,  the wet  or  "fresh" weight  of the sample before it was preserved
   should be  recorded,  as well as the volume of preservative used in each jar.
   Specimens  preserved  in  formalin  or alcohol must be accompanied by a "control"
   jar.   This jar must  contain the  same mixture used in preserving the specimens,
   and must be prepared (i.e., rinsed and sealed) in the same manner as the jars
   containing  specimens.  This may not be equivalent to freezing for storage
   of samples, however.  For example, Abate was partially converted to Abate
   sulfoxide in  fresh samples stored in formalin or formalin plus 5% acetic
   acid, but not in frozen samples  (6).  Formaldehyde should be checked for the
   presence of PCB contamination prior to use as a sample preservative (7).

   Comments pertinent to collecting samples of 'different types will be pre-
   sented in the Subsections 8F to 8K.   Methods, for the analysis of the various
   sample types are surveyed in Section 9 of this Manual.


8F   SAMPLING OF AGRICULTURAL AND FOOD PRODUCTS

   Procedures for sampling, sample preparation,  sample  compositing,  and sample
   reporting, as required by Federal law,  for all commodity types are outlined
   in detail in Sections 140-143  of the FDA Pesticide Analytical Manual,  Volume
   I.  Section 3 of the Canadian  Department  of National Health and Welfare
   Analytical Manual for Pesticide Residues  in Foods  gives  guidelines for
   systematically obtaining representative samples of processed and  packaged
   foods, bulk foods, and field crops and  for handling,  shipping, and storing
   samples.  Recommended minimum  sizes  are tabulated  for different samples, with
   a  general sample requirement of n product units, where n equals the  square
   root of the total but need not exceed 10-15 separate  units.

   Section 4 of the same Canadian PAM covers  laboratory  preparation of  analytical
   samples from gross samples of  fresh, frozen, and canned  vegetables,  fruits,
   and juices; dry cereal grains, flakes,  dehydrated fruits and vegetables;
   animal tissues; eggs; butter and  margarine; milk and  cream;  cheese and nuts;
   fats  and oil; and fish and fish products.
                                        -290-

-------
                                                              Section 8G

   It is suggested that readers interested iii analysis of sample substrates
   of this type for legal compliance to tolerance levels should refer to these
   two excellent -sources of information.  Sampling methods for trace organic
   analysis of foods have also Been described by Horwitz and Howard (8) .  If
   the purpose of an analysis is to obtain information on maximum residue
   levels in a particular situation, biased sampling would be used, e.g., the
   lower perimeter of fruit would be sampled from certain trees most likely
   to have received a higher dose of pesticide spray.
8G   SAMPLING OF BIOLOGICAL MATERIALS

   Adipose tissue, blood, and urine samples from live and autopsy animal and
   human subjects are commonly analyzed for pesticide residues.  The amounts
   of sample required, the time of collection, and the compound to be detected
   are determined by the nature of the pesticide (s) of interest.  Pesticides
   that degrade or .are metabolized readily may be absent in a particular
   sample, but their original presence can be deduced by determination of
   metabolites such as alkyl phosphates from OP pesticides, phenols from
   chlorophenoxy acid herbicides or carbamate insecticides, or DBA from DDT.
   If body tissues or fluids are analyzed quickly in cases of high exposure,
   the chance of finding the parent pesticide is greatly enhanced.  If exposure
   is low or a long time has elapsed after exposure, the analyst must be
   familiar with pesticide metabolism in order to choose appropriate samples
   and metabolites to determine.  For example, the highest concentration of
   organophosphorus pesticide urinary metabolites will be found from four to
   eight hours after the donor's exposure (EPA PAM, Section 6, A, 2, (a) ,V).  When
   concentrations of pesticides or metabolites are expected to be small, samples
   must be larger, e.g., morning urine samples or 24-hour pooled specimens.

   The majority of human adipose tissue samples are taken during autopsy by an
   attending physician.  Samples should be placed in a clean glass container
   with a foil- lined  (never rubber- or cardboard -lined) screw cap.  The aluminum
   foil should be prerinsed with acetone.  Plastic bags or bottles must be
   avoided since they can contribute traces of impurities such as phthalates to
   the sample, causing spurious GC peaks when the final concentrate is examined
   by EC-GC or GC-MS  (EPA PAM, Section 5,A,(l),(a) ,V).  Up to 2% of radiolabeled
   DDT was found to be lost by "extraction" into the plastic when liver samples
   were stored in polyethylene bottles at 4°C overnight.  This radioactivity was
   not removed when the bottles were washed, so that the loss for one sample could
   constitute a contamination for the next sample stored in the same bottle (2) .

   Whole blood samples are. transferred to glass vials with Teflon or foil lined
   screw caps* and the required serum aliquot is removed after a period of
   settling in a refrigerator and subsequent  centrifugation.  Serum 'is stored
   in a refrigerator  at  2-5°C if the analysis is to be performed within 24 hours
   or in a deep freeze  (-15 to -25°C) for longer storage periods.  The analysis
   of chlorinated pesticides is not adversely affected by such storage for
   periods up to  six  months  (EPA PAM, Section 5 ,A, (2) , (a) »
                                          -291-

-------
                                                              Section 8H
8H   AIR SAMPLING
   The EPA has in the past operated a nationwide air monitoring program in
   order to gather information on the extent of human exposure to airborne
   pesticides.  This program utilized Greenburgh-Smith impingers containing
   ethylene glycol for trapping organophosphorus and halogenated hydrocarbon
   insecticides both in the vapor phase and as dusts.  The air was drawn
   through the impingers by means of a vacuum pump,  the amount sampled  (cu.  m)
   being.controlled by means of a flow meter and timer.  The ethylene glycol
   sampling procedure did not prove acceptable in terms of,convenience  or
   reliability, and the EPA national air sampling program was discontinued.
   However, the ethylene glycol impinger continues to be used by some labora-
   tories in local monitoring programs (3,  9,  10).

   Robert G.  Lewis of the U.S.  EPA Health Research Effects Laboratory has  .
   written an extensive review of sampling  methods for airborne pesticides"(11).
   Included are discussions of many types of accumulative samplers (e.g.,  im-
   pactors, bubblers,  liquids supported on  solid substrates,  polymer  foams,
   etc.),  reactive samplers,  continuous and sequential samplers,  and  grab
   samplers.   Recent reports  of pesticide recovery from air have included  the
   use of tubes containing XAD-2 resin for  trapping  2,4~D acid, and its  ester
   and amide derivative .with 86-96% efficiency (12);  hexylene glycol  contained
   in glass scrubbers  for recovery of dieldrin and heptachlor at  0.1  ng/cu. m
   (13);  and XAD-2 resin for organothiophosphates  (14).

   One of the most promising  approaches to  the sampling of air involves use  of
   polyurethane foam.   The updated review of air sampling methods  in  Section
   8,A of the EPA PAM contains  a discussion of this method in addition  to  other
   approaches and apparatus recommended by  the EPA for high volume ambient air
   sampling,  indoor air sampling,  crop re-entry monitoring, and workplace  and
   personnel  monitoring.    Polyurethane foam  vapor traps  following a particle
   filter have been evaluated (15)  for sampling of pesticides, PCBs, and poly-
   chlorinated naphthalenes.  Collection rates up  to  360  cu. m of air per 24
   hours  and  sensitivities  as low as  1 cu. m for some compounds can be achieved.
   The filters and plugs were Soxhlet extracted with  hexane-ethyl ether (95:5 v/v)
   at  4 cycles per hour for 16-24  hours,  and OC1 pesticides were determined by
   EC-GC after alumina column cleanup and OP pesticides by FPD-GC without clean-
   up.  Collection was  generally satisfactory  but was poor for the more-volatile
   OC1 compounds.   Recovery was  ca 75%  for  OP pesticides.  It was shown that a
   second  trap in  series with the  first  did  not  necessarily improve recovery
  values.  The collection  of dieldrin,  lindane, trifluralin, dacthal, chlordane,
   and heptachlor  on polyurethane  foam was studied and optimum plug size and
   shape for  any chosen sampling rate were given.  Trapping efficiency depended
   on  pesticide vapor pressure and the flow  rate of air.  The quantitation
  limit was  ca 0.1 ng/cu. m in a 5  cu. m air sample.  It was crucial that the
  plugs were carefully protected  from laboratory contamination and Soxhlet ex-
   tracted with pesticide-grade acetone and hexane prior to use if clean blanks
  and  highest sensitivity were to be achieved  (16).   Other collection efficiency
  data for pesticides  and PCBs are included in Section 8,B of the EPA PAM.

  Because of possible  pesticide degradation, air sampling apparatus should
  be shielded from light during sample  collection.
                                        -292-

-------
                                                              Section 81
81   WATER SAMPLING
   The design, of a comprehensive pesticide sampling program for environmental
   waters is a specialized topic that is covered in publications available
   from the Water Quality Control Division of the USEPA, National Environ-
   mental Research Center, Cincinnati, Ohio.  Important considerations include
   the objective of the study, frequency of sampling, location of sampling
   stations as related to hydrologic conditions, and the 'selection of sampling
   methods.  The following is a brief review of some important selected
   factors in a sampling program.

   a.  Grab Samples
       Water can be collected by taking one instantaneous ("grab") sample from
   a given location,  directly filling the sample container.   The usual technique
   is to submerge the container a few inches below the water surface during
   filling to avoid skimming off any floating film that would be least repre-
   sentative of the vertical water column.  Several collections should be taken
   at various depths and locations to provide a more representative sample.
   Care should be exercised to avoid disturbing bottom sediment.  Discrete
   samples from various depths can be obtained with standard samplers consisting
   of a metal outer container with a glass sample collection bottle inside
   (e.g*., Precision and Esmarch samplers, EPA PAM, Section 10,A,II).  Grab
   sampling is often sufficient for lakes, reservoirs, etc., that are not subject
   to rapid transitional changes.

   Grab samples less than 2 liters are collected in wide mouth glass bottles,
   and samples of one gallon or more in the glass bottles in which pesticide
   quality solvents are supplied.  All bottle caps should be Teflon lined.  The
   sample size is dictated primarily by the expected residue levels, the sensi-
   tivity of the analysis, and the need to run duplicate, spiked, and background
   analyses.  A 500-1000 ml sample may suffice from water where pesticide levels
   are expectedly high, while 2 liters or more may be needed for a surveillance
   program where no high levels are anticipated.  Rainwater is collected in
   clean glass containers rather than metal or plastic.  Samples should include
   information that will help the analyst choose a proper analytical method and
   interpret the results.  This includes the location of sampling, depth, sus-
   pected contaminants, type of sample (surface water, waste discharge, etc.),
   and agricultural activity or spills in the immediate area or upstream.

   Many pesticides are unstable in water, so samples should be analyzed as soon
   as possible after collection, ideally within a few hours.  If this is im-
   practical because of distance from the sampling site to the laboratory and/or
   the  laboratory work load, store the sample in a refrigerator or freezer.
   Samples being examined solely for organochlorine residues may be held up to
   a week under refrigeration at 2 to 4°C with no adverse effect.  Those samples
   to be analyzed for organophosphorus or carbamate pesticides should be frozen
   immediately after drawing the sample because of rapid degradation in aqueous
   media  (Table 1, Section  10,A of the EPA PAM shows data for the degradation
   rate of  29 pesticides  in water at ambient temperature in sealed containers
    (I?)). pH adjustment is  required  for  some samples immediately after collection
    (e.g., adjustment to pH  2 with sulfuric acid for phenoxy acid herbicides).
   Holding  time and  storage conditions must be reported along with the analytical
   results  and corrections  made if rates  of pesticide  degradation are known.
   Exposure of samples  to sunlight should be avoided.
                                         -293-

-------
                                                           Section 81
 Every effort should be made to perform the solvent extraction step at the
 earliest possible time after sampling, irrespective of the classes of
 pesticides suspected of being present.  Especially unstable pesticides can
 be extracted immediately in the field.  The resulting extracts can be safely
 stored for periods up to three or four weeks at -15 to -20°C before pro-
 ceeding with subsequent cleanup and determinative steps.   One disadvantage
 of glass sample bottles is possible breakage in shipment,  and care should
 be exercised in proper packaging to avoid this.  Another  disadvantage is
 the already-mentioned possibility of pesticide adsorption on glass surfaces.
 Reduced recovery (>90Z to 46-681) of DDT in water analysis upon storage has
 also been noted due to adsorption on suspended matter in  the sample (18).

 The assumption made Is that a grab sample is at least representative of the
 immediate water mass from which it was taken and somewhat  representative of
 the water that will pass the sampling point during some limited future time
 interval.   The grab sample is amenable to use in both random and nonrandom
 sampling programs.   The number,, frequency, and distribution of samples
 collected will depend on the study objectives and the variability  within the
 "population" being sampled.

 After sampling,  pesticides are extracted from water,  cleaned up and concen-
 trated as  necessary, and determined by GC or an alternative method.  Pesti-
 cides in clean water (e.g.,  drinking water)  can be detected at 5-500 ppt
 levels by electron capture GC without the need for extensive extract clean-
 up.   Impurities in "dirty" samples will require additional cleanup steps,
 and background problems will cause difficulty in analyzing these low levels
 accurately.

 b.   Continuous Samplers

     Continuous and automatic devices are often used for sampling flowing
 bodies of  water such as rivers and streams.   Activated carbon filters, have
 been widely  used for adsorption of pesticides and other kinds  of organlcs
 in natural waters  since they were developed  and Introduced by  the tijS, Public
 Health Service in  1951 (19).   The technique  involves passage of  a continuous,
 constantly controlled volume of water through a column of  activated carbon
 followed by  desorption by  means of elution or by Soxhlet extraction with a
 suitable solvent or combination of solvents.   The variable efficiency and
 consistency  of pesticide adsorption and  desorption  from the adsorbent prior
 to determination, ease of  contamination with extraneous organic substances,
 and bacterial  and oxidizing  attack on the  sorbed pesticides have caused
 problems with  carbon columns  (20,  21).

 filter materials which have been recommended as alternatives to carbon for
 collection of  pesticides (usually chlorinated insecticides) from natural
waters include reversed  liquid-liquid partition  systems (a hydrophobia phase
 coated on  a  support)  and other adsorbents.   Carbowax 4000   (5 g) and ii-undecane
 (22), silicones  chemically bonded to diatomaceous earth support (23),
 covalently bonded aromatic and alkyl chlorosilanes on Celite (24),  porous
polyurethane foam columns  (for pesticides and PCBs) (25, 26), polyethylene
 film  (20-25 pm thickness)  (27), and polyurethane foam coated with selective
adsorbents (28) have all been used with varying success.
                                      -294-

-------
                                                          Section 81


The XAD macroreticular adsorbent resins (XAD-1, -2, -4, and -7) have.been
used to collect organlcs from both potable (29, 30) and sea (31) water.
Optimum conditions for use with XAD-4 resin were found to be 2 g of adsor-
bent, a flow rate through the resin of 8 ml/minute, and 100 ml hexane-
diethyl ether (10:1 v/v) as elutlng solvent.  Among 10 chlorinated
insecticides studied, only aldrin and p,p'-DDE were not quantitatively re-
covered, and recovery of PCBs was 76% (32).  Details for use of XAD-2 and
-4 resins for many classes of trace organic water contaminants have been
published (33) and recoveries between 81 and 96% were reported for 20 ppt
levels of atrazine, lindane, dieldrin, DDT, and DDE (47% for aldrin).  An
EPA report (34) recommends XAD-2 resin for routine monitoring of sea water
for chlorinated insecticides and PCBs.  Average recovery for XAD-2 ex-
traction of fortified natural waters collected across Canada was 85% for
the 10-100 ng/liter levels of ten OC1 pesticid.es (recovery of mirex was
unacceptably low) and 82% for 250 ng/liter levels of PCBs; blanks from the
resin were a low 4 ng PCBs/liter (35).  Concentrations as low as 0.1 ppt of
PCBs and organochlorinated pesticides were detected by recovery from water
on small XAD-2 columns  (36), and ng levels of carbamates were recovered
(86-108% at 0.01-1 ppm  levels) with the same adsorbent (37).  Amberlite
XAD-4, porous polyurethane foam, and undecane plus Carbowax 4000 on Chromo-
sorb were comparable for extracting ten OC1 insecticides from environmental
water samples (22).

Continuous liquid-liquid extractors are an alternative to the filter-adsor-
bent .processes preferred by some analysts.  A multi-chamber extractor with
internal solvent renewal replenishing (38) allowed extraction of 135 liters
of water at rates of 0.5-1.0 liter/hour and recovered greater than 97% of
ppb levels of pesticides.  Subsequently, a similar modified apparatus per-
mitted use of both heavier- and lighter-than-water solvents (39).  A simple
and rugged field version of the Kahn and Wayman apparatus (38)" excluded
solvent recycling and was based on mixed settling  (40).  -This apparatus,
which consisted of an extraction unit, magnetic stirrer, and pump, provided
quantitative recovery of pesticides and PCBs at levels of 0.1-1.0 ng/liter
of river water.

More recently, a similar in situ apparatus designed to solvent-extract large
amounts of sea and river water continuously while situated at a desired depth
at the  sampling site has been described  (41).  A Teflon  helix, continuous
liquid-liquid extractor, plus a continuous evaporative concentrator re-
covered pg to ng per liter amounts of OP pesticides from river and sea water
or from secondary  sewage effluent with >80%  efficiency  (42).  A comparative
study of  recoveries  from river water by continuous extraction and activated
carbon  filters  showed  that  the recoveries were similar but the former was
less costly  (43).

The  theory for  extracting  chlorinated pesticides continuously from water
with a  stationary immiscible  solvent  is discussed  in reference  (40).

 See  Section  10,A,III-V of  the EPA PAM for  updated material on water sampling.
                                       -295-

-------
                                                              Sections 8J, 8K

 8J   SAMPLING OF HOUSE DUST, SOIL, AND STEEAM BOTTOM SEDIMENT

    House dust is collected with a vacuum cleaner, air dried, and sieved prior
    to analysis.  Soil is sampled By collecting cores or Borings of a known
    diameter cut to a depth of ca. 3-4 inches or more from the centers of plots
    1 sq. m In size.  Ten to twenty cores representing a surface area of at
    least 200 sq. m are recommended.  The first 2 inches of core, containing
    the grass or crop cover and roots, are separated from the underlying soil.
    Corings representing each layer of soil are combined, quartered,  and
    divided into 2 IB samples for analysis.  Soils are analyzed in an air-dry
    state after sieving to remove foreign material.   Another reported procedure
    for soil sampling (3) involves collection of cores 1-3 inches deep and
    3 inches in diameter with a hand-operated auger; on a 1/4 acre site
    C10S feet x 105 feet), sampling Begins 7.5 feet  from the Border of the site,
    and a core is collected every 15 feet until 7 cores are oBtained.   The
    process is repeated along parallel lines separated By 15 feet from the
    original sampling line,  until a total.of 49 cores are collected.   The cores
    are sieved through a hardware cloth screen into  a 3 gallon galvanized pail
    and thoroughly mixed.  The sample is transferred to two one-half  gallon cans
    with lids for shipment to the laBoratory.   There is no way to collect a
    truly representative soil sample,  and reproduclBility of results  on different
    samples taken from the same area is often expectedly poor.

    Sediment from the Bottom of a Body of water provides information  concerning
    the degree of pollution  resulting from pesticides,  particularly those that
    are not readily degradaBle.   This  information comBined with residue .data on
    the water and resident Biological life  gives an  overall pesticide  cohtaminar
    tion profile  of the Body of water.   Bottom sediment varies with respect to
    Both particle size composition (surface adsorptive  power)  and organic content.
    Therefore,  sample sites  should Be  selected at random in an effort  to  collect
    samples representing a rang'e of variation.   In some cases  consultation with
    an  oceanographer can indicate where one would Be likely to find the maximum
    amounts of pollution from considerations such as currents and industrial
    effluent discharges.

   ABout a quart of sediment is a typical  sample size.  Actual collection is
   accomplished "i±th one  of a variety  of core  samplers  or dredges.  A, diagram
   of  a dredge-type device  for collecting  sediment  samples has Been puBlished
    (3).   The dredge is  thrown into  the water at  least  10  times to collect
   samples, which are transferred each time to a. galvanized pail.  The total
   sample is mixed and  transferred  to  one-half gallon cans  (with a hole in each
   lid to  release  any gas Buildup from organic matter- in  the sample)  for ship-
   ment  to  the laBoratory.  A simple Bottom sediment collector composed of a
   steel  can attached to  the  end of an aluminum  pole has also Been described
    (44).   Samples may Be preserved with formalin or a variety of other steri-
   lants provided  they do not affect the analyses to Be run.  Samples are air
   dried and ground prior to analysis.  They are stored, if necessary, in a
   freezer if volatile compounds such as 2,4-D ester may Be present.


8K   MARINE BIOLOGICAL AND WILDLIFE SAMPLES

   A problem sometimes encountered when collecting plankton and Bottom
   organisms is oBtaining the minimum weight necessary for successful analysis.
                                         -296-

-------
                                                             Section 8L
  As a general mle, a minimum of about 10 g will be required.  Collected
  organisms can be frozen at once or preserved with 5-10% formalin or 70%
  ethanol, prepared with distilled water rather than the water from which the
  collection was made.  This eliminates the possibility of pesticides In the
  water concentrating in the organisms over a period of time.  Any added pre-
  servative must be extracted and analyzed to determine if exchange of pesti-
  cides from the organisms to the preservative has occurred.

  Sufficient masses of plankton are collected by use of a tow net behind a
  boat or by pumping water through a net.  Bottom fauna are collected with
  dredges or dip nets.  Samples are washed through a screen and organisms
  are hand picked from the remaining debris.

  Fish are collected utilizing seines, gill nets, traps, electrocution de-
  vices, otter trawls, or angling.  Wrapping the fish in aluminum foil and
  preservation by .quick freezing in dry ice is most desirable.  When this is
  not possible, liquid preservatives are used.  Larger fish should be injected
  with preservative from a syringe to prevent decomposition of internal organs.
  Fish stored in formalin plus 5% ^28303 showed no loss of Abate (temephos)
  residues  (>1 ppb) for up to three weeks (45).

  Fish can be analyzed whole to yield data on gross contamination, or the fish
  can be sectioned to obtain information on edible and non-edible parts.
  Analyses of individual organs and tissues yield information on distribution
  of pesticides in the"fish.  Analysis of blood from a dying fish may be
  valuable for determining probable cause of death where pesticide exposure is
  suspected.  The blood is obtained by cutting the tail at the caudal peduncle
  and collecting and freezing the blood in a small vial.

  Invertebrate samples are collected in pitfall traps, as described by Wojicfc
  et -JL!.   (46).  Bird samples are collected using Japanese mist nets placed
  near a water source or in a cove where the net Is not visible.  Traps baited
  with peanut butter or some other foodstuff are employed for sampling mammals.
  These traps, which are available in a variety of sizes, have a trap door that
  closes when the animal enters  to take the food.  Non-crop vegetation samples
  are obtained with shears, sickles, pocket knives, etc., usually from the
  same sampling area as soil samples.  All of these samples are sorted, wrapped
  in aluminum foil with the shiny  side out, tagged, and placed in a plastic
  bag for shipping  (3).

  Some of the material In  the  sections on sampling was adapted from an EPA
  training course manual  (47).   A review that Includes some of the above
  sampling procedures and  additional methods for  collection of environmental
  samples has been  published  (3).


8L   CONTROL OF PROCEDURES  FOR EXTRACTION OF RESIDUES

   Specific procedures  for the extraction and cleanup  of pesticide multi-
   residues In many sample types are surveyed in Section 9 of  this Manual.
   This  subsection discusses general considerations  of pesticide extraction
   from collected samples.
                                       -297-

-------
                                                           Section 8L
 Many solvents are employed for extracting residues,  depending on the
 polarity of the pesticide and the amount of co-extractives  expected from
 the particular substrate.  Solvents range from hexane or petroleum ether
 for nonpolar organochlorine and organophosphorus compounds  to methylene
 chloride (dichloromethane) for polar carbamates.   Chloroform, diethyl ether,
 ethyl acetate, benzene,  acetonitrile, methanol, acetone, and  various "two-
 and three-component mixtures of these have all been  widely  used.   Addition
 of acid to the organic solvents may aid extraction of acidic  pesticides  such
 as 2,4-D herbicide.   Acetonitrile is an excellent general purpose extraction
 solvent for low fat-content samples (acetonitrile plus ca  35% water for low
 moisture samples),  and hexane/acetonitrile systems are widely recommended
 for partition cleanup.

 Although it has been shown in some cases that  recovery of pesticides from
 tissues during extraction does not necessarily correlate with recovery of
 lipids,  it is usually desirable to use  an extraction solvent  that will
 quantitatively extract lipids with the  pollutants for reporting purposes
 (2).   A study (48)  has compared the recovery of lipid by nine solvent
 mixtures from human adipose tissue for  pesticide determination.   Extraction
 procedures should always be validated for each class of compounds in each
 type of sample matrix to which it is applied.   In addition  to the nature of
 the analyte,  the toughness,  water content,  and lipid content  of the sample
 matrix will Influence the effectiveness of a given extraction procedure  (2).

 Different techniques are employed for bringing, the extraction solvent and
 sample into contact.  The best extraction is obtained,  in general, by
 achieving the most  intimate contact between the two,  although the type of
 residue is an important  distinction.  When emulsions result from  vigorous
 shaking or mixing during extraction procedures,  centrifugation will usually
 be effective In separating solvent layers.   A  surface residue can usually
 be extracted by a simple washing procedure,  while the more  common internal
 residues can be extracted only after fine maceration of the sample.  Some
 soil samples  tenaciously bind pesticides and require long periods  (e.g.,
 8  hours)  of Soxhlet  extraction rather than shorter periods of blending as
 is common with plant materials.   Blending of the  sample plus  solvent in a
 Waring blender (Figure 8-C),  Omni-Mixer,  Wiley mill,  or Hobart food chopper
 is probably the most usual extraction procedure in use  today,  especially
 for biological, plant, and food samples.  Some additional sample subdivision,
 such as cutting, chopping, or grinding usually precedes the blending operation.
A 5 minute period of blending at a moderate speed is typical for many  samples.
A special device for aiding formation of a homogenous sample has been  des-
 cribed  (49).  The device, consisting of a handle and shaped aluminum sheet,
 fits inside a blender jar and serves to gently push bulky samples into  the
 cutting blades during the blending operation.  A liquid-nitrogen cooled
 freeze grinder for biological materials containing labile pesticides has
also been devised (50).

Blending with a solvent followed by filtering or centrifuging is particularly
efficient for most vegetable samples.  The water in the sample may give
                                      -298-

-------
                                                           Section 8L
  Figure 8-C.  Waring Aseptic Dispersail Model AS-1  (Shown on 702-CR Base)
rise to emulsions with nonpolar solvents, and this can often be avoided by
use of a drying agent such as anhydrous Na2S04 or 2-propanol together with,
or before, the solvent.  Meat samples containing too much connective tissue
for a blender to deal with effectively should be first comminuted by a
grinder.  Simple heating of minced sample in a beaker on a steam bath with
solvent can be effective, possibly after grinding the sample with Na2S04
and sharp sand to help break down some connective tissue.  More volatile
pesticides (e.g., lindane) might be lost in this way.

A comparative study of the efficiencies in the extraction of carbofuran
from radishes was made using three blenders, a Polytron ultrasonic homo-
genizer, a Lourdes blender, and a Waring blender.  The least efficient
blender extracted 90% as much as the most efficient, and all three were con-
sidered useful for accurate pesticide analysis (51).

In some cases, more exhaustive extraction of residues from difficult samples
can be obtained by Soxhlet extraction for periods up to 12 hours or longer
with a solvent such as methanol-chloroform (1:1 v/v) (52).  Soxhlet thimbles
may require exhaustive extraction prior to use so they do not contribute
                                    -299-

-------
                                                           Section 8L

interferences to the analysis (53).  Preliminary steps such as drying,
grinding, or chopping normally precede Soxhlet extraction, but care must
be exercised since some pesticides have been shown to be unstable in the
presence of homogenized samples (54).  Even Soxhlet extraction may not give
complete extraction in all cases, and only studies with samples to which
radioactive tracers have been applied can indicate the absolute extraction
efficiency in any particular case.  The usual evaluation procedure of
spiking a sample with pesticide and looking for quantitative extraction
is less reliable than the radiotracer method because the spiked chemical
will not be naturally incorporated in the same matrix as would the tracer.
Radiotracers are not always available or feasible to use, however.  The
most important factor in preparation of a valid spiked sample to accurately
indicate recovery of endogenous compound may be the solvent in which the
spike is dissolved.  In one study of the extraction of mirex from fish
muscle, recovery varied from 41 to 89% with a common extraction procedure
but different spiking solvents (2).

Water samples (100-500 ml) are generally extracted by shaking with an
appropriate solvent (3 x 100 ml) in a separatory funnel (55).  Soils are
extracted by a variety of methods such as shaking, soaking, blending,
Soxhlet or Goldfisch extraction, or refluxing. -Two 15 minute extractions
in an ultrasonic generator were found comparable to a 24 hour Soxhlet
extraction for removal of s-triazine herbicides from fortified soils (56),
and a 30 second extraction technique using a Brinkmann Polytron ultrasonic
generator gave better recoveries of several chlorinated insecticides from
soil than did 8 hours of Soxhlet extraction (57).

An apparatus that simultaneously Soxhlet extracts pesticides and concen-
trates the resulting extract has been designed (58).  Advantages of this
cyclic extraction-evaporation system are that distillation of solvents prior
to extraction can often be omitted, and excess solvent is re-utilized for
extraction.

The most efficient solvent and parameters for extraction of pesticides from
water can be determined using the ^-values originally suggested by Beroza
and co-workers for use in residue confirmation (Subsection 10F in Section
10).  The £-value is the fraction of total pesticide.that is distributed
into the nonpolar phase of an equivolume immiscible pair of solvents.  This
approach was used to study the extraction of OP pesticides from water (59),
and the best solvents were benzene, ethyl acetate, or diethyl ether for
diazinon and diazoxon at pH 7.4, ethyl acetate for malathion at pH 6, and
diethyl ether or ethyl acetate for fenthion (Baytex) at pH 3.4.  ^-Values
can also be used to theoretically select water-to-solvent ratios and the
optimum number of extractions for maximum recovery of a pesticide in water
(60).  As a practical example (61), diethyl ether or ethyl acetate was
found best for extraction of 2,4-D acid and esters and benzene for 2,4,5-T
                                     -300-

-------
                                                              Section 8M
   acid and esters.  A 99% recovery of 2,4-D from one liter of aqueous solution
   was obtained by a two stage serial extraction with 200 ml and 50 ml of ethyl
   acetate under conditions predicted by p_-values.
8M   CONTROL OF METHODOLOGY FOR CONCENTRATION OF SAMPLE SOLUTIONS AND
     FRACTIONAL COLUMN ELUATES

   The concentration of cleaned-up sample in the injection or spotting solution
   is one important factor that determines if sufficient residue is available
   for detection by GC, LC, or TLC.  The analyst must determine this and con-
   centrate final solutions according to the least sensitive pesticide in the
   method's scope.

   Purified extracts or eluate solutions containing even somewhat volatile com-
   pounds are concentrated with minimum losses to a volume of ca  5-10 ml using
   a Kuderna-Danish evaporative concentrator flask fitted on top with a 3-ball
   Snyder reflux column and a collection tube on the bottom (Figure 8-D).
                                              Figure 8-D.  Kuderna-Danish; J
                                              Evaporative Concentrator, Kontes
                                              Glass Co. No. K-570000.
   The tube is heated in steam water bath in a hood.  The apparatus should be
   mounted or held so the lower rounded flask surface is bathed in steam.  Flasks,
   which range in size from 125-1000 ml, should be initially charged with
   40-60% of their nominal volume, and the column should be pre-wet with ca  1 ml
   of solvent before beginning concentration to prevent possible initial small
   loss of pesticides.  Refluxing is continued until the final concentrate is
                                        -301-

-------
                                                           Section 8M


collected in the lower tube.  Boiling chips are required for smooth operation
of the K-D evaporator, and carborundum, checked for absence of contamination,
is recommended in preference to porcelain, vanadium, or glass chips,  A
Snyder column modified By putting a distillation trap below the Vigreaux
bubble condensing system increased the degree and consistency of recovery
of nanogram amounts of HCE isomers upon Kuderna-Danish evaporation (.62).

For concentration from 5 ml to smaller volumes (as low as 50-100 pi), the
concentrate is cooled, the collection tube is removed from the K-D flask,
and a fresh chip is added.  A micro-Snyder reflux column (Figure 8-E) is
fitted directly to top of the tube, and evaporation is begun by holding the
bottom of the tube in a steam or hot water bath.  Evaporation is continued,
with care to avoid bumping, to slightly below the desired volume.  The tube
is withdrawn from the water when boiling agitation becomes too vigorous;
immersion and withdrawal are alternated based on observation of boil agita-
tion.  The apparatus is cooled 3-5 minutes, and condensate is allowed to
drain down into the tube before the column is removed.  The sides of the
tube and column joint are rinsed with solvent to avoid hang-up of pesticides
on upper glass surfaces.  A 1-2 ml syringe is useful for performing this
rinse.  Finally, further fresh solvent is added to dilute up to the desired
volume, if necessary.

A special rack that simultaneously agitates and evaporates solutions in six
concentrator tubes fitted with micro-Snyder columns in a time equal to a
single tube is described in the EPA FAM, Section 5,A,3,a.
                                           Figure 8-E.  Semi-Micro Kuderna
                                           Danish Apparatus, Kontes Glass
                                           Co., No. K-569250.
                                     -302-

-------
                                                           Section 8M

Extracts containing fats, oils, or plant extractives, or purified extracts
to which "keeper solution" has been added, can be evaporated on a rotating
vacuum type evaporator with the water bath at, or just slightly above,
room temperature (Figure 8-F).  A double-reservoir rotoevaporation vessel
facilitating collection, concentration, and final volume calibration of
column eluates and eliminating a number of manual transfer steps has been
designed (63).
                                         Figure 8-F.  Rotary Evaporator,
                                         Kontes Glass Co., No. K-570160
Extracts contained in a beaker or a centrifuge tube immersed in a water
bath at 40°C can be evaporated under a stream of nitrogen adjusted to cause
gentle depression on the surface of the solution.  The nitrogen should be
passed through well maintained scrubber tubes to remove contaminants that
could cause pesticide degradation.  Warming a tube by holding it in the
hand is a useful, gentle evaporation aid during nitrogen blow-down.
Figure 8-6 shows the Organomation Associates, Inc., N-Evap apparatus
that is widely used for evaporation by nitrogen blow-down.

An evaporation assembly combining an evaporative concentrator tube, a
Ruderna-Danish flask, and a rotary vacuum evaporator (Figure 8-F) is shown
in Section 10,A of the EPA PAM, Figure 1.  The concentrator tube is not
immersed in a high temperature water bath as usual, but rather in a 35°C
water bath to minimize degradation of heat labile pesticides.  This apparatus
confines the concentrated extract to one container, thereby eliminating the
need for transfer.  One hundred ml of methylene chloride can be reduced to
5 ml in ca  20 minutes with a vacuum of 125 mm of mercury.
                                     -303-

-------
                                                           Section 8M
                                            Figure 8-G.  Model 111 12-position
                                            N-Evap* apparatus, Organomation
                                            Associates, Inc., Northborough, MA.
Another multitube apparatus for nitrogen evaporation is available from
Kontes Glass Co. (64).  Concentration is rapid until the solution reaches
0.5-1.0 ml, at which point evaporation slows markedly because this last
volume is below the heating zone of the evaporator block.  Thus, losses
of pesticides from inadvertent evaporation to dryness (65) are avoided,
and a minimum of analyst attention is required (Fig. 8-H).
                                             Figure 8-H.  Ebullator,
                                             Kontes Glass Co.
                                    -304-

-------
                                                          Section 8M

A distillation column is fixed on top of the tube holding the sample, and
small bore stainless steel needlestock tubing is fitted through the column
down into the tip of the tube to direct a stream of micro bubbles of nitro-
gen through the solution to Initiate and maintain ebullation.  Recoveries
of seven chlorinated pesticides after concentration for 2 hours in this
apparatus were greater than 94% with both hexane and benzene solvents.

It is important to avoid pesticide loss or decomposition during evaporation
steps.  Numerous reports have been made (e.g., 65, 66) of severe pesticide
loss during concentration steps, even in the presence of sample coextrac-
tives.  There was no correlation between the amount of coextractives and
evaporative losses, but apparently the nature of the coextractives may be
important.  In most situations, organochlorine and organophosphorus pesti-
cides can be concentrated to small volumes without loss ,by the K-D evapo-
rative procedures described at the beginning of this subsection.  Some
recoveries from 100- and 1000-fold concentrations carried out in K-D
assemblies are shown in the following table.  The recoveries are quite
acceptable when concentrating to 1 ml, but when concentrating to 100 yl
without a keeper, recoveries become marginal.  Using a keeper, such as a
paraffin oil, helps retain the compounds and greatly reduces losses.  How-
ever, the keeper may interfere with some analysis, especially by flame
ionization detection or mass spectrometry.

                               TABLE 8-1
 Losses of pesticides on evaporation in Kuderaa-Danish  concentrators  (67)
Pesticide
   Original
  amount in
100 ml hexane
    
                                      % Recovery-on concentration to *
                                10 ml '
1 ml
0.1 ml
0.1 ml with
 "keeper"
Diazinon
Aldrin
Malathion
Parathion
Dieldrin
_p_,jg;-DDT
40
1.0
40
10
2.5
5.0
102 (2.4)
103 (2.8)
85 (2.4)
93 (4.4)
103 (5.6)
96 (9.2)
85 (4.4)
85 (4.0)
91 (5.2)
84 (4.0)
92 (4.0)
91 (5.6)
71 (1.8)
69 (1.2)
77 (3.0)
70 (1.2)
78 (0.6)
78 (3.0)
83 (1.8)
81 (0.6)
88 (0.6)
82 (2.4)
90 (1.2)
90 (2.4)
  Averages  of  six determinations  for each pesticide.   Standard deviations
  given in  parentheses.   Gentle stream of nitrogen used  to  assist concentra-
  tion below 10 ml.
                                      -305-

-------
                                                            Section  8M

Evaporation  to  dryness  should never  occur.   If  the  complete removal of a
particular solvent  is required,  solvent  exchange  can be  carried out so that
the sample never gets to dryness.  For example, hexane can  be completely
removed by boiling-down to a low volume  and  adding  small volumes of acetone
as evaporation  continues until all hexane is eliminated.

The use of air  for  concentration of  an extract  should best  be avoided.
Satisfactory recoveries are obtainable when  the residue  levels are  rela-
tively high, but significant losses  have been documented of even the more
stable pesticides at low concentration levels (65).

A commercial tube heater that avoids evaporation  to dryness with micro K-D
apparatus was originally described by Beroza and  Bowman  (68) (Figure 8-1).
Six extended-tip K-D concentrator tubes  are  accommo.dated, and simultaneous
evaporation  to  less than 1 ml can be carried out  without attention.
                                         Figure 8-1.  Tube Heater,
                                         Kontes Glass Co., K-720000
Other reports of pesticide loss include dieldrin and DDT when an extract
was evaporated in the presence of light (69), mirex upon evaporation of
aqueous solutions (70), and carbamate pesticides when evaporated in a K-D
apparatus (71).  In the latter case, rotary vacuum evaporation (Figure 8-F)
at 50-55°C with addition of a keeper, solution was recommended.  Many
carbamates can be successfully evaporated under a nitrogen stream without
loss after adding a keeper.  A satisfactory general purpose keeper is 5
drops of 1% paraffin oil in hexane.  Solutions containing the herbicide
Balan (benefin) cannot be evaporated in a current of air without loss of
pesticide, whereas rotary evaporation at a temperature of 50°C or less is
successful.  All evaporation, and concentration steps should be checked with
spiked samples if any question of pesticide loss should arise*
                                      -306-

-------
                                                             Section  8N


   The Importance of clean glassware in all parts of a pesticide analysis has
   been stressed several times earlier in this Manual.  The special importance
   of clean,glassware to be used for concentration of solutions to small
   volumes cannot be overemphasized.

   The final solution to be used for the determinative step must be composed
   of a solvent appropriate for the particular analytical procedure.  Choice
   of a volatile solvent for partition and column cleanup procedures is
   advantageous because evaporation to an appropriate volume can be carried
   out quickly enough to be practical.  If a different solvent is required
   for the final sample solution, solvent exchange can be carried out by
   taking up the nearly dry residue in the new solvent after evaporation.
   Solvents for GC and LC are restricted by the selectivity of the detector,
   while for TLC almost any volatile solvent is useful for the solution to be
   spotted.  Chlorinated solvents cannot be present in the injected solution
   when an EC or the Cl modes of the MC or electrolytic detectors are to be
   used.  Acetonitrile has an adverse effect on the response of the EC
   detector, while aromatic and halogenated compounds and acetonitrile increase
   the response of the thermionic detector.  The most volatile solvent possible
   should be used to shorten the venting period and minimize loss of early
   eluting pesticides for those detectors that require solvent venting (e.g.,
   FPD and CCD).  A solvent free of UV absorption is required for the detection
   by the ultraviolet LC monitor.


8N   REFERENCES  .

   (1)  Ambrus, A., Adv. Pest.' Sci.. Plenary Lect. Symp. Pap. Int. Congr. Pestic.
        Chem.. 4th. 1978. j3» 62° (1979).

   (2)  Albro, P. W., Ann. NY Acad. Scj., 320, 19 (1979).

   (3)  Ford, J. H., McDaniel, C. A., White, F. C., Vest, R. E., and Roberts,
        RJ E., J. Chromatogr. Sci.. 13, 291 (1975).
        'H                     '         d
   (4)  Dick, G. L., Heenan, M. P., Love, J. L., Udy, P. B., and Davidson, F.,
        N. Z. J. Sci.. 21, 71 (1978).

   (5)  Bourne, S. . J. Environ. Sci. Heath B.. 13(2), 75 (1978).

   (6)  Mies, J. W., Dole, W. E.,  and Churchill, F. C., Arch. Environ. Contain.
        Toxicol., 5., 29  (1976).

   (7)  Kurtz, D. A., Pestic. Monit. J.. U.(4), 190 (1978).

   (8)  Horwitz, W., and Howard, J. W., NBS (US) Spec. Publ. 519:231  (1979).

   (9)  Arthur, R. D., Cain, J. D., and Barrentine, B. F., Bull. Environ.
        Contain. Tosicol., 15, 129 (1976).

   (10)  Kutz, F. W., Yobs, A. R., and Yang, H. S. C., Chapter 4 in Air
        Pollution from Pesticides and Agricultural Processes!, Lee, R. E., Jr.,
        ed., CRC Press, Boca Raton,  FL.
                                        -307-

-------
                                                             Section 8N


 (11)  Lewis, R. G., Sampling and Analyses of Airborne Pesticides,  in Air
       Pollution from Pesticides and Agricultural•Processes. Lee, R. E., Jr.,
       ed., CRC Press, Boca Raton, FL, pp. 52-94  (1976).

 (12)  Johnson, E. R., Yu, T. C., and Montgomery, M. L., Bull. Environ. Contam.
       Toxicol., 1J, 369 (1977).                            ~~"~	

 (13)  Taylor, A. W., Glotfelty, D. E., Turner, B. C., Silver, R. E.,
       Freeman, H. P., and Weiss, A., J. Agr. Food Chem.. 25, 542 (1977).

 (14)  Kaminsky, F., and Melcher, R. G., Am. Ind. Eve. Assoc. J.. 39(8),
       678 (1978).                       		  —    '

 (15)  Lewis, R. G., Brown, A.  R., and Jackson, M. D., Anal. Chem.. 49.
       1668  (1977).                                    	—

 (16)  Turner, B. C.,  and Glotfelty, D. E., Anal. Chem.. £9, 7 (1977).

 (17)  Eichelberger, J.  W., and Lichtenberg, J. J., Environ. Sci. and Technol.,
       5,  541 (1971).                                	:	'

 (18)  Wilson, A.  J.,  Bull. Environ. Contam. Toxicol.. 15, 515 (1976).

 (19)  Braus,  H.,  Middleton, F.  M., and Walton, G., Anal.  Chem.. 23, 1160
       (1951).

 (20)   Sproul, 0.  J.,  and Ryckman,  D.  W.,  J..Water" Pollut.  Contr.  Fed.,  33.
       1188  (1961).	 '  ~'

 (21)   Ruzicka,  J.  H. A., and Abbott,  D.  C.,  Talanta. .20,  1261 (1973).

 (22)  Musty,  P.  R., and Nickless,  G.,  J.  Chromatogr.. 120,  369  (1976).

 (23)  Aue, W. A.,  Kapila,  S., and  Hastings,  C.  R.,  J.  Chromatogr..  73,
       99  (1972).	*-   —'


 (24)  Ito, T., Water Resource Research Institute of the University  of North
      Carolina, Report No.  54,  August  (1971).

 (25)  Gesser, H. D., Chow,  A.,  Davis,  F.  C., Uthe, J. F., and Reinke, J.,
      Anal. Lett., ±, 883  (1971);  Gesser, H. D.,  Sparling,  A. B., ChowT A.,
      and Turner, C. W., J. Am. Wat. Wks. Ass.,  65.  220 (1973).

 (26)  Musty, P. R., and Nickless,  G.,  J.  Chromatogr... 100.  83 (1974).

 (27)  Weil, L., Quentin, K.-E., and Gitzowa, S., Gas -u. WassFach.  113,
      64 (1972).	'

(28)  Uthe, J. F., Reinke, J.,  and Gesser, H. D., Environ. Lett.. .3, 117
      (1972).
                                      -308-

-------
                                                            Section 8N


(29)   Burnham, A. K., Calder, G. V., Fritz, J. S., Junk, G. A., Svec, H. J.,
      and Willis. R.. Anal. Chem., 44. 139 (1972).

(30)   McNeil, E. E., Otson, R., Miles, W. F., and Rajabalee, F. J., J_.
      Chromatogr., 132, 277 (1977).

(31)   Riley, J. P., and Taylor, D., Anal. Ghim. Acta. 46,  307  (1969).

(32)   Musty, P. R., and Nickless,  G., J. Chromatogr., 89,  185  (1974).

(33)   Junk, G. A., Richard, J. J., Grieser, M. D., Witiak, D., Witiak,
      J. L., Arguello, M. D., Vick, R.,.Svec, H.  J., Fritz, J. S.,*and
      Calder, G. V., J. Chromatogr., 99, 745  (1974).

(34)   Harvey, G. R., U.S. EPA Report R2-73-177,  32 pp  (1973).

(35)   Coburn, J. A., Valdmanis, I. A., and Chau,  A.  S.  Y., J.  Assoc.  Off.
      Anal. Chem..  60_, 224  (1977).

(36)   Rees, G. A.  V., and Au, L.,  Bull. Environ.  Contam.  Toxicol., 22(4/5)
      561  (1979).

(37)  Sundaram,  K. M. S.,  Szeto, S. Y., and Hindle,  R., J. Chromatoer.. 177(1),
      29  (1979).

(38)  Kahn, L.,  and Wayman, C.  H.', Anal. Chem.,  36.  1340 (1964).

(39)  Goldberg,  M.  C., DeLong,  L., and Sinclair, M., Anal. Chem., 45, 89
      (1973).

(40)  Ahnoff,  M.,  and Josefsson,  B.,  Anal.  Chem.. 46. 658  (1974).

(41)  Ahnoff,  M.,  and Josefsson,  B.,  Anal.  Chem., 48, 1268 (1976).

(42)  Wu,  C.,  and Suffet,  I.  H.,  Anal. Chem., 4£, 231  (1977).

(43)  Brodtmann, N. V.,  J. Am.  Wat. Wks.  Ass., 67, 558 (1975).

(44)  Miles,  J.  R. W.,  Pestic.  Monit. J.,  10, 87 (1976).

(45)  Miles,  J.  W., Dale,  W.  E., and Churchill, F. C., Arch. Environ.
      Contam.  Toxicol..  5_, 29 (1976).

(46)  Wojcik,  D. P., Banks, W.  A., Hicks,  D.  M., and Plumley, J. K., Florida
      Entomologist. 55_,  115 (1972).

 (47)  Pesticide Residue Analysis in Water. U.S. EPA, Office of Water Programs,
      distributed by NTIS (No.  PB-238 072) September (1974).
                                       -309-

-------
                                                             Section 8N


 (48)  Mes,  J.,  and Campbell,  D.  S.,  Bull. Environ. Contam. Toxicol.. 16,
      53  (1976).                                                     ' —

 (49)  Analytical  Methods for Pesticide '•Residues in Foods, Canadian Depart-
      ment  of National Health and Welfare, Section 14.4.

 (50)  Howells,  K. J.,  Shaw,  T. C., Rogers, P. P., and Galbraith, K. A.,
      Lab.  Pract..  23, 248 (1974).

 (51)  Wheeler,  W. B.,  Thompson,  N. P., Edelstein, R,  L., and Krause, R. J.,
      Bull. Environ. Contam.  Toxicol.. 23(3), 387 (1979).

 (52)  U.S.  FDA  Pesticide Analytical  Manual, Vol.  I, Section 253.

 (53)  Telling,  G. M.,  Sissons, D. J.,  and Brinkman, H. W., J. Chromatogr..
      137,  405  (1977).                                      	*~

 (54)  Nutahara, M., and Yamamoto, M.,  J.  Pestic.  Sci.. 3/2), 101 (1978).

 (55)  Agemian,  H.,  and Chau,  A.  S. Y., J.  Assoc.  Off. Anal.'Chem.>  60,  1070
      (1977).                          	   —

 (56)  Hill, B.  D.,  and Stobbe, E. H.,  J.  Agr. Food Chem.. 22, 1143  (1974).

 (57)  Johnsen,  R. E.,  and Starr,  R.  I., J.  Agr. Food  Chem.. 20,  48  (1972).

 (58)  Voss, G., and Blass, G., Analyst, 98. 811 (1973).

 (59)  Suffet, I.  H., and Faust,"  S. D., J.  Agr.  Food Chem.. 20, 52  (1972).

 (60)  Suffet, I.  H., J.  Agr.  Food Chem..  21,  288  (1973).

 (61)  Suffet, I.  H., J.  Agr.  Food Chem..  21,  591  (1973).

 (62)  Malaiyandi, M., J. Assoc.  Off. Anal.  Chem..  61, 1459 (1978).

 (63)  May, T. W., and Stalling, D. L., Anal.  Chem., j>l(l), 169 (1979).

 (64)  Beroza, M., Bowman, M.  C.,  and Bierl, B. A., Anal.  Chem..  44,  2411
      (1972).                                                   —

 (65)  Burke, J.  A., Mills, P.  A., and  Bostwick, D. C., J.  Assoc. Off. Anal.
      Chem., 49, 999 (1966).                                       '.   ""	

(66)  Chiba, M., and Mbrley, H. V., J. Assoc. Off. Anal.  Chem., 51,  55  (1968).

(67)  Lewis, R.  G., Accuracy  in Trace  Organic Analysis, in Accuracy  in Trace
      Analysis:   Sampling, Sample Handling, and Analyses.  Vol. 1, La Fleur, D.,
      ed., NBS Special Publication 422, Washington, DC, pp.  9-33 (1976).
                                       -310-

-------
                                                             Section 8N


(68)  Beroza, M., and Bowman, M. C., Anal.  Chem.,  39_, 1200 (1967).

(69)  McKinley, W. P.^ and Savary,  G.,  J. Aar.  Food Chem.. 10,, 229  (1962)

(70)  Stein, V. B., and Pittroan, K. A.,  Bull.  Environ. Contain. Toad-col.,
      19(6), 755 (1978).

(71)  Storherr, R. W., J. Assoc. Off. Anal. Chem.. 55, 283 (1972).
                                         -311-

-------
                                   Section 9
            MULTIRESIDIE EXTRACTION AND ISOLATION  PROCEDURES FOR
              PESTICIDES AND METABOLITES AND  RELATED OTOUNDS
   This  section presents brief  descriptions 'and  quality  control aspects
   of widely used multiresidue  analytical procedures for different sample
   substrates.  A few methods for important individual residues are also
   included.  Many of the problem areas are treated in a general manner
   elsewhere in this Manual, but they are high-lighted again here in re-
   lation to the specific methods.  References are given in each case
   to sources of detailed methodology.  Control  of procedures for
   collection of samples is covered in Subsections 8A-8K in Section 8
   and for sample extraction and extract concentration in Subsections 8L
   and 8M in Section 8.


                           CHLORINATED PESTICIDES


9A   TISSUE,  FAT,  AND FOOD ANALYSIS BY THE MILLS, ONLEY,  GAITHER PROCEDURE

       a.  Analysis of Tissue and Fat

           The modified Mills,  Onley,  Gaither method described in Section
   5,A,(l),(a)  of  the EPA Pesticide Analytical Manual has been determined
   by a number of  interlaboratory collaborative studies to  yield very
   acceptable precision and accuracy for the analysis of  a  number of
   chlorinated pesticides and metabolites  in human or animal  fatty tissues.
   However, many polar OP and carbamate  pesticides are  not  recovered.
   (This method involves dry maceration  of a 5 gram sample  with sand  and
   anhydrous  Na2S04,  isolation  of fat  by repeated extraction with petroleum
,   ether, extraction  of  residues into  acetonitrile,  and then partitioning
   back into  petroleum ether after adding  2% NaCl,  drying by elution  through
   a  column of  Na2S04,  concentration of  the  eluate,  cleanup on  a  Florisil
   column, and  EC-GC  after  reconcentration of  column  eluates.)   If necessary,
   further cleanup  of the 15% ether-petroleum  ether Florisil eluate is carried
   out  on a MgO-Celite  column.   Pooled blood serum can  be analyzed by the MOG
   Florisil procedure after  extraction with  a  hexane-acetonitrile  solvent
   system [EPA  PAM, Section  5,A, (3),(a),VIII].
                                     -312-

-------
                                                   Section 9A
(1)   Some analysts, with hope of saving time, have combined 6% and 15%
     ethyl ether-petroleum ether Florisil column fractions and have then
     attempted gas'chromatography on the mixture.  With some luck this
     approach might prove successful, but there is a good chance that it
     could lead to erroneous conclusions.  For example, in one documented
     instance, an analyst reported the presence of aldrin in a human fat
     sample.  Other collaborators on the sample analysis found the same
     peak in the 15% eluate, making its identification as aldrin impossible
     since this compound elutes wholly in the 6% fraction.  By combining
     the fractions, the analyst inadvertently neglected the use of se-
     lective adsorption as a valuable identification tool.


 (2)  The polarity of  the  ethyl  ether-petroleum ether eluting  solutions
     exerts a profound  effect on  the elution pattern of  several pesti-
     cidal  compounds.   The  amount of ethanbl,  a relatively polar  solvent,
     in the ethyl ether is  a critical factor as illustrated in Figure
     4-A in Section 4.  As  indicated in this figure, with nd  ethanol,
     dieldrin would be  expected to yield only 87% recovery in Fraction
     II with the balance  being  retained on the column.   If twice  the
     proper amount of ethanol is  present, approximately  7% should elute
     in Fraction I, giving  a 93%  recovery in Fraction  II.   If 2%  ethanol
     is present and all the dieldrin still does not elute in  Fraction II,
     the presence of  moisture in the system may be the cause. An excess
     of moisture may  result in  all or most of the dieldrin eluting in the
     6% fraction.                                 .   "

 (3)  The "activity characteristics of Florisil may vary somewhat  from lot
     to lot.  Each lot, when received at a laboratory, should be  care-
     fully evaluated  to be  certain,the compound elution characteristics
     are satisfactory.
 (4)
Storage and holding temperature of Florisil are critical.  The oven
used for holding this (and other adsorbents) should be confined
exclusively to this usage and not used as an all-purpose drying
       Florisil will readily pick up air-borne contaminants that
      oven.
      may result in spurious chromatographic peaks.  If the oven tempera-
      ture varies more than +_ 1°C, considerable influence may be observed
      in the retention characteristics.  The recommended activation tempera-
      ture is 130°C.

 (5)  Anhydrous Na2S(>4 used to top the Florisil column, even AR grade,
      frequently contains sufficient impurities to result in spurious
      peaks in the blank eluates.  Because of the prevalence of this
     j situation, it is good practice to Soxhlet extract all lots of
     'this salt before use.

 (6)  The presence of peroxides in ethyl ether can result in extremely
      low recoveries of organophosphorus compounds and also poses a
      serious safety hazard.  Methods have been set forth for the re-
      moval of peroxides from ether but have not proven wholly satis-
      factory.  The purity of petroleum ether is also critical and may
      exert a profound effect on  the recovery of certain of the organo-
      phosphorus compounds.
                                   -313-

-------
                                                        Section 9A


 (7)   Glassware must be meticulously cleaned to remove electron capturing
      contaminants.   Reagent blanks must be run with each set of samples.

 (8)   Most chlorinated pesticides  should be recovered in the range of
      85-100%.   HCB  is an exception because of an unfavorable partition
      ratio in  the acetonitrile-petroleum ether solvent system.   An
      aldrin spike can be added to the minced fat at the start of the
      procedure if this pesticide  is known to be absent.   Recovery of
      this spike should not  be less than 70%.


 If improper Florisil fractionation occurs during an analysis,  the
 following points should be  considered:   Florisil that is  too retentive
 could result from (a)  improper activation temperature,  (b)  improper
 percentage of  ethyl ether in petroleum  ether,  and ethyl ether that does
 not contain the required 2% ethanol (read the  label on the  container
 carefully).  Florisil that  appears insufficiently retentive might re-
 sult  from (a)  or (b)  above,  or from residual amounts of a polar solvent
 in the sample  or standard being placed  on the  column.  Likely possi-
 bilities  are acetonitrile from the sample partition cleanup step (if
 drying steps are not performed properly)  or incomplete removal  of benzene
 (or other solvent more polar than hexane)  from a standard solution placed
 on the column.   Other sources of  Florisil problems  are. undoubtedly
 possible.   See also Subsection 9M for further  comments on pesticide
 elution from Florisil.


      b.   Analysis of Fatty  and Nonfatty Foods  Using Florisil Cleanup

          The Mills,  Onley,  Gaither column method for determining nonionic
 chlorinated  pesticides  in fatty foods is similar to that  outlined in Sub-
 section 7Aa  and is  described in detail  in Sections  211, 231, and 232 of
 the FDA PAM.   Eluants  are 6,  15,  and 50% ethyl ether in petroleum ether.
 The method for nonfatty foods  (FDA PAM,  Sections  212 and  232) involves
 extraction of  pesticides  with acetonitrile or  water-racetonitrile  and
 partition into  petroleum ether prior to Florisil column chromatography
 and EC-GC,   The FDA PAM lists  pesticides  recovered  through  these  pro-
 cedures [results  for some 200  pesticides  and other  chemicals are  given
 in Table  201-A and  over 300  compounds have been tested (1)], samples to
which they are  applicable,  and supplemental cleanup  procedures  for the
 Florisil  column fractions.   This AOAC multiresidue method is currently
 official  for 26  OC1  and OP pesticides and  PCBs  in various groupings of
 42 nonfatty  and 4 fatty foods  (1).   The problem areas are the same as
 those given  in  Subsections  9Aa and  9M.  The elution  pattern of more than
 150 pesticides  from the U.S. FDA Florisil  column  eluted with 6, 15, 20,
 30, 50, and  65% diethyl ether  in petroleum ether  is  tabulated in  Section
 7.2(b) of  the Canadian  PAM.  Free fatty acids  in high quantity are not
 sufficiently removed by this FDA/AOAC procedure to prevent interference
with pesticide  determination using electron capture, KC1 thermionic,  and
Hall electrolytic conductivity  detectors  (2).  A potassium permanganate/
 dilute sulfuric acid oxidation procedure was developed to supplement
Florisil  chromatography for cleanup  of  chlorinated pesticide residues in
vegetable  extracts.   Twelve chlorinated pesticides were completely re-
 covered,   and only aldrin was lost via decomposition  (3).
                                  -314-

-------
                                                         Section 9A


Thirteen chlorinated pesticides were determined in milk by GC after Florisil
column cleanup.  Of the several systems tested, extraction of milk with
20 ml hexane plus 5 ml acetonitrile plus 1 ml ethanol produced the highest
pesticide recoveries and lowest fat extraction (4).

In order to obtain more efficient cleanup of extracts of fatty foods and
recovery of additional pesticides of higher polarity (e.g., organo-
phosphates), a new elution system consisting of three different mixtures
of methylene chloride, hexane, and acetonitrile was devised as replace-
ment for the traditional diethyl ether-petroleum ether eluants.  These
eluant mixtures are methylene chloride-hexane  (20:80 v/v); methylene
chloride-acetonitrile-hexane (50:0.35:49.65 v/v); and methylene chloride-
acetonitrile-hexane  (50:1.5:48.5 v/v).  At least 50 pesticides and re-
lated chemicals have been recovered, in groupings different from the
mixed ether'systems, with these new solvents  (5).  Table 201-A of the
.FDA PAM also includes data on the elution characteristics of compounds
using the methylene chloride/hexane/acetonitrile system  (FDA PAM, Section
252).  A silver nitrate-coated Florisil column has provided cleanup of
fatty and vegetable sample extracts and fractionation of chlorinated pesti-
cides and phthalate esters prior to their simultaneous analysis by gas
chromatography  (6).

Malathion and  some other organophosphorus pesticides require 50% diethyl
ether-petroleum ether for elution from Florisil.  This elution, which
must be preceded  by  elution with the 6% and  15% eluants, has been found
occasionally to be inconsistent.  OP pesticides can be lost through de-
gradation on the  Florisil column and during  subsequent evaporations, or
when water  dilution  of  the acetonitrile extract for residue transfer to
petroleum ether is carried out.  Recoveries  are tested by  carrying known
amounts of  pesticides through  the procedure  in the absence of  crop sub-
strate.  Only  23  of  70  OP pesticides and metabolites  tested through the
MOG procedure  were recovered,  and not  all recoveries were  complete.  The
AOAC has validated the  procedure only  for carbophenothion, diazinon,
ethion, malathion, methyl and  ethyl parathion, and ronnel  in 18  fruit and
vegetable crops (7-9).                                      ,

Beckman and Garber  (10) recommended the  solvent series benzene,  diethyl
ether-benzene  (1:2 v/v),  acetone,  and  methanol for elution of  Florisil
columns.  The  elution pattern  and  recovery  of 65  OP pesticides were studied,
but sample  extracts  were not tested.   This  system was later found  to be
applicable  to  the determination of methyl and ethyl parathion, malathion,
malaoxon, and paraoxon residues in apples  and lettuce, although  "all-Florisil"
 columns were not generally recommended as  the best choice  for  cleanup of OP
pesticides  (11).

A novel use of Florisil was the development of a  partition column  consisting
 of acetonitrile on Florisil for the separation of some pesticides  from fish,
 beef,  and butter fat (12).   The technique was useful for cleanup of pesti-
 cides having favorable ^-values (Section 10F) in a hexane-acetonitrile
 system, which included dimethoate, temephos, methyl parathion,, fenitrothion,
  •.rufomate,  malathion, and parathion.
                                      -315-

-------
                                                             Sections 9B, 9C, 9D


    A multiresidue method for organochlorine,. organophosphorus, dinitrophenyl,
    and carbamate pesticides in applies and other high-water crops was devised
    using Florisil for cleanup (13).  Carbamates were eluted from one column
    with toluene-acetone (19:1 v/v) and acetylated with trifluoroacetamide for
    EC-GC determination.  Organochlorine and organophosphorus compounds were
    eluted from a separate column with toluene-acetone (49:1 v/v) and determined
    by GC.   Dinitrophenyl compounds were then eluted from this column with 95%
    ethanol, cleaned-up by solvent partitioning, methylated, and determined by
    EC-GC.   Most recoveries were greater than 75%,  even for polar compounds.

 9B   HCB AND MIREX IN ADIPOSE TISSUE

    Section 5,A,l,(b)  of the EPA PAM describes the  determination of hexachloro-
    benzene (HCB)  and  mirex in fatty tissue with confirmation of HCB by formation
    of Msj-isopropoxytetrachlorobenzene.   The sample is dissolved- in hexane
    and applied directly to a Florisil column.   The HCB and mirex residues  are
    eluted with hexane and determined by direct EC-GC of the concentrated eluate.
    HCB is  then reacted with 2-propanol,  and the BITS derivative is chromatographed
    to provide confirmation of HCB.   Mirex does not survive this reaction.  Other
    common  pesticides,  some of which are  altered by the reaction,  are all sepa-
    rated from the HCB derivative on the  OV-17/OV-210 column used (14).


 9C   HUMAN OR ANIMAL  TISSUE AND  HUMAN MILK ANALYSIS BY THE FLORISIL MICROMETHOD

    If the  size of the available  tissue sample  is so small  as to make the macro
   MOG method unsuitable,  a micromethod  is described in Section 5,A,(2) of the
   EPA PAM requiring  as  little as 200-500  mg of  sample.  The sample is extracted
   with  acetonitrile,  pesticides are  partitioned into hexane, fractionated on
   a  1.6 gram Florisil column  (eluate I: 12  ml of hexane plus 12 ml..of 1%
   methanol in hexane; eluate  II: 12  ml  of 1% methanol  in hexane),  and concen-
   trated  fractions determined by EC-GC.   Several pesticides, including O-BHC,
   lindane, diazinon,  DDD,  and toxaphene,  split between fractions.  Florisil
   columns must be conditioned at 130°C  at least overnight before xteing.  Pre-
   cautions concerning use  of Florisil are similar  to those outlined in Sub-
   section  9Aa.  Virtues of  the micromethod  include a low background level and
   savings  in  the volume of  solvent required.

   Miniaturization of Florisil column cleanup has been reported in several papers
   (15, 16).   One procedure has been successfully studied by several labora-
   tories  (17).


9D   HUMAN BLOOD OR SERUM

   A 2 ml aliquot of serum is extracted with 6 ml hexane for 2 hours on a slow
   speed rotary mixer.  After concentration, the hexane layer is analyzed by
   EC-GC [EPA PAM, Section 5,A, (3),(a)].   The procedure involves no cleanup,
   but, if carefully handled, it is capable of yielding recoveries of chlorinated
   pesticides comparable to that obtained from a full MOG cleanup technique
   (see Tables 2-4 to  2-9 in Section 2).   Since all pesticides  will be present
   in one extract, a GC column must be chosen that  will separate the expected
                                       -316-

-------
 ,                                                           Section 9E

  pesticides.  Certain  serum samples will yield a very  late eluting extraneous
  peak (probably  a phthalate)  that  is  sometimes large enough to  distort  a
  following  chromatogram if  time  is not  allowed for its elution  from the
  column.  Blood  samples should never  be stored in  containers with polyethylene
  or  rubber  caps.  Hexane was proven superior to hexane-formic acid for
  extraction of dieldrin, lindane,  and DDT  from serum  (18).  Microcoulometric
  GC  determination after sulfuric acid extraction was successfully applied to
  24  organochlorine  pesticides in blood  at  1 ppb levels with no  cleanup  (19).

  Blood samples are  not always analyzed  without cleanup steps.  Monitoring
  of  fur seal blood  for OC1  pesticides and  PCBs required chromatography  of
  the hexane extract on a 2.3 gram Florisil column prior to EC-GC (20),
  Another monitoring study  of pesticides in human  blood was carried out  by
  hexane extraction  of acidified  samples followed  by cleanup on a 1 gram
  Florisil column and EC-GC (21).

  Hexane-acetone  (9:1 v/v)  was a  better  extractant for DDT and BHC isomers
   in human blood  than was pure hexane.  Maximum recoveries occurred when
   serum was treated  with formic acid before extraction.  Ease of extract-
   ability decreased in the  order:  f-BHC > a-BHC > g-BHC > £,£.'-DDE > p_,p_'-DDT >
   o^'-DDT (22).

9E   PENTACHLOROPHENOL  (PCP)  IN BLOOD AND URINE

   Acidified blood is extracted with benzene on a Roto-Rack for 2 hours
   followed by methylation of PCP and determination by EC-GC  [EPA PAM,
   Section 5,A,(3),(b)].  Urine is  acidified, and hydrolysis carried out for
   one hour to free conjugated PCP.  PCP and phenol metabolites of PCP and
   HCB are extracted with benzene, methylated with diazomethahe, the methylated
   phenols are cleaned  up and fractionated on an acid alumina- column, and
   determination is carried  out by  EC-GC [EPA PAM, Section  5,A, (4),(a)].  The
   following comments pertain to  these methods:

   a.  The alkylating reagent diazomethane is a hazardous chemical and must
       be handled with  extreme  caution.*
   —  Diazomethane and related alkylating reagents (e.g., diazoethane, diazo-
      pentane) have been widely used in pesticide residue analysis and are
      cited in several procedures in this Manual and the EPA PAM.  These com-
      pounds and their precursors are toxic and carcinogenic and are irritating
      to the skin.  Solutions have been known to explode inexplicably.  It is
      recommended that safer substitutes be found for these reagents whenever
      possible, for example BF3-methanol for methylation of acid herbicides
      and acetic anhydride for acetylation of pentachlorophenol (Section 9Ac).
      Substitution of one reagent for another, however, can require a large
      amount of effort to check the validity of the procedure with the new
      reagent.  If diazolkane reagents must be used to reproduce established
      analytical procedures, take care to keep from direct contact with the
   -  skin.  Wear disposable vinyl gloves and safety goggles, and avoid
      breathing of vapors.  Work behind a safety shield in an efficient hood
      or inside a radiological glove box.  Do not prepare or store reagents
      in ground glass stoppered or etched glassware.  Avoid strong light.
                                        -317-

-------
 b.
 d.
                                                          Section 9E


     A 1.5% OV-17/1.95% QF-1 column is not recommended since the relative
     retention values for 2,4-D methyl ester and PCP methyl esters are
     identical and these pesticides would not be differentiated.

     All reagents including distilled water must be pre-extracted with
     hexane to remove interfering materials.   Reagent blanks should be
     carried through the entire procedures with each set  of samples and
     standards.

     Glassware should be washed with dilute NaOH followed by deionized
     water and acetone.

     Contact between wooden or  paper materials  and  glassware should not
     be  permitted as some of these materials  have been found to contain
     significant  levels  of PCP.

     Other ether  derivatives (e.g.,  ethyl,  propyl,  amyl,  etc.) can be pre-
     pared and characterized for confirmation of  PCP  identity.

     Fortified samples should be analyzed along with  each series of actual
     samples  to verify adequate  recovery of PCP and the other phenols of
     interest.  Because  of  the ubiquity of  PCP, the "blank" used for forti-
   .fication must be analyzed,  and  a correction must be made for the
     amount of PCP found.

h. "A reagent blank consisting  of 5 ml of pre-extracted distilled water
    should also be carried  through  the entire procedures along with
    samples.
 f.
 g-
 i.
    Confirmation of PCP is based on chemical ionization mass spectrometry
    or extraction j^-values.
In the methods described above  (23, 24), phenols were chromatographed on
conventional GC columns after derivatization to a more easily chromato-
graphed compound.  The derivatization step exposes the analyst to a toxic
derlvatizing reagent and increases the possibilities of error.  It has
been demonstrated that support  coated polyester columns are suitable for
determining free chlorinated phenols in urine at subnanogram levels with-
out the need for derivatization (25).

A method for monitoring PCP in  fish and other environmental samples with a
± 2% accuracy and precision has been described (26).  PCP was extracted
from fish tissue and converted  to pentachloroanisole (PCA) by means of
alkylation in the presence of potassium carbonate as a condensing agent.
After adding pentachlorophenetole as internal standard,  determination was
carried out by electron impact mass fragmentography monitoring 280 m/e for
PCA and 294 m/e for the internal standard.
                                   -318-

-------
                                                            Sections 9F, 9G


97   BIS(2-CHLOROPHENYL)ACETIC ACID (j5,2.!-DBA) IN HUMAN URINE

   The excretion level of this metabolite is a sensitive indicator of exposure
   to £,£f-DDT.  Urine is extracted three times with an equal volume of 2%
   acetic acid in hexane, the combined extracts are evaporated to remove
   residual water or acetic acid, DDA is converted to its methyl ester by
   reaction with BF3-methanol reagent, and the ester is extracted with hexane
   and determined by GC with microcoulometric or EC detection [EPA PAM,
   Section 5,A, (4),(b)].

   Microcolumn Florisil cleanup  (Subsection 9C) is required when the poorly
   selective EC detector is used.  DDA should elute completely in Fraction
   II.  Concentration and injection volumes depend upon the sensitivity of
   the detector employed.  A column of 5% OV-210 at 175-180°C will separate
   DDA from p,£*-DDE (which usually is also present in high exposure donors),
   whereas 4% SE-30/6% QF-1 or 1.5% OV-17/1.95% QF-1 columns at 200°C will not
   resolve these compounds.

   A very similar procedure involving diazomethane methylation, no cleanup,
   propyl ester internal standard, a 1%  QF-1 column at 190°C, and a 63Ni
   pulsed EC detector has been reported.  The calibration curve was linear
   up to 1.5 yg DDA per liter, the coefficient of variation was ca 8%, the
   absolute detection limit was  0.05 ng, and 20-30 samples could be run per
   day  (27).

96   2,4-D AND 2,4,5-T IN URINE

   A method is described in the  EPA'PAM, Section 5,A,(4),(c), for determining
   these herbicides and their degradation products 2,4-dichlorophenol and
   2,4,5-trichlorophenol in human and animal urine.  Phenolic conjugates are
   hydrolyzed  in acid, and free  phenols  and acids are extracted with benzene
   and  ethylated with diazoethane.  Cleanup and fractionation of derivatives
   is carried  out on a silica gel column (1 gram, containing 1.5% water), and
   determination of concentrated eluates by EC-GC on a 4% SE-30/6% OV-210
   column.

   Deactivated silica gel  (Subsection 4Ad in  Section 4) columns should be
   prepared just prior  to use.   Because  of  the differences in temperature and
   humidity from one laboratory  to  another, silica gel elution parameters
   should be established by each analyst under local conditions.  The per-
   centage water added  for deactivation  should be increased if the  compounds
   of interest elute in a  later  fraction than that indicated in the detailed
   procedure,  or,  the percentage of benzene in the benzene-hexane eluant can
   be increased.   Early elution  would be remedied by less deactivation  or less
   polar solvents.   Spiked control  urine, rather than standard compounds,
   should be used  to determine the  elution  pattern.  See the footnote on
   page 6 concerning the hazards associated with the ethylating reagent.
   Alkylated  standards  are stable for one month  if stored in a freezer
    (-18°C) when not  in use.
                                        -319-

-------
                                                            Section 9H

   A multiresidue scheme for phenol metabolites and including 2,4-D, 2,4,5-T,
   and silvex is discussed in Subsection 9R.  A method for monitoring 2,4-D
   in the urine of pesticide spray operators at 0.1 ppm involved cleanup on
   XAD-2 resin, quantitation by GC of the methyl ester, and confirmation by
   trans-butylation to the ii-butyl ester.  Recovery was 94 ± 6% for five
   fortified samples (28).


9H   KEPONE IN HUMAN BLOOD FOR ENVIRONMENTAL SAMPLES

   The determination of Kepone in human blood, air, river water, bottom
   sediments, and fish is described in the EPA PAM, Section 5,A,(5),(a).
   This is based on the research of Moseman gt al. (29), Hodgson et al. (30),
   and Earless ^£ al. (31).  Samples are extracted, and the extracts are
   cleaned up by chromatography on a micro Florisil column, base partitioning,
   or gel permeation chromatography.  Kepone is determined by EC-GC with
   multiple columns.  Confirmation is by chemical conversion to mirex
   followed by further cleanup prior to EC-GC (32); detection with a Hall
   conductivity detector in the Cl-mode; or chemical ionization mass
   spectrometry (31).

   It is mandatory to use 1-2% methanol in benzene for all sample and standard
   solutions injected for EC-GC to obtain the maximum reproducible response.
   Sufficient control and spiked reference materials should be utilized to
   ensure the validity of analytical results for all sample types.  Elution
   patterns for the Florisil columns should be carefully established by each
   analyst by eluting standard Kepone under local laboratory conditions.
   Analytical standards should be validated by cross-reference analysis of
   difrerent preparations of analytical grade Kepone with agreement within
   i 5% of the established purity.

   The analysis of field-collected avian tissues and eggs for Kepone residues
   has been reported (33).   Samples were extracted with benzene-isopropanol
   (2:1 v/v) and extracts cleaned up with fuming H2S04-concentrated H2S04
   (1:1 v/v).  Separation of Kepone from OC1 pesticides and PCBs was obtained
   on a 10 gram 130°C-activated Florisil column eluted with 100 ml of benzene-
   acetone (95:5 v/v) followed by 200 ml of benzene-methanol (90:10 v/v); the
   second eluate contained the Kepone.   Determination was by EC-GC on a 4%
   SE-30/6% QF-1 column and confirmation by GC-MS.  Recoveries averaged 86%
   at 1 ppm.  Procedures for determination of Kepone in serum, plasma,  urine
   and fat have been reported.   After addition of #2804,  samples were extracted
   with hexane-acetone (17:3 v/v), extracts were evaporated,  and the residue
   dissolved in benzene-methanol (99:1 v/v).   The extraction was modified for
   feces and bile.   Programmed temperature GC with pulsed EC detection on a
   4% SE-30/6% QF-1 column provided linear calibration curves for 10 pg-100 ng
   of Kepone (5 ppb-50 ppm/gram sample)  (34).   Determination of Kepone  in eels
   (35);  blue fish or shrimp (36); finfish, shellfish, and crustaceans  (37);
   water and sediment (38);  and soil and mullet (39)  using gas chromatography
   have also been reported in the literature.
                                      -320-

-------
                                                            Section 91
91   GEL PERMEATION CHROMATOGRAPHY
     a.  Gel Permeation Chromatographic Cleanup of Adipose Tissue

         (1)  Theory

              Gel permeation chromatography (GPC) is a form of liquid chromatog-
   raphy by which compounds are separated according to molecular size.  It is
   particularly useful in separating very large molecules such as lipids and
   cholesterol found in adipose tissue samples from the smaller molecules of
   pesticides, PCBs, etc.  The method is as effective as the MOG procedure
   for cleanup in pesticide residue analyses (4QJ and has the added advantages
   that removal of fat is more complete and recoveries of pesticides are nearly
   quantitative.  Hence, it is the ideal choice for GC-MS analyses, where
   maximum detectability of pesticides is needed and minute quantities of lipid
   materials can cause serious interferences.

              Porous polymer beads (e.g., BioBeads SX-3) are used as gel
   particles and organic solvents (e.g., toluene, ethyl acetate, or cyclo-
   hexane) are used for the mobile phase.  The elution process is very simple
   (isocratic only); the same solvent system is used for column preparation,
   elution, and washing.  Macromolecules cannot permeate the porous gel and
   are rapidly eluted or "dumped" from the column.  Molecules that can enter
   the pores of the beads are temporarily retained to greater or lesser
   extents depending on their molecular volumes.  Hence, large-volume pesti-
   cides such as mirex elute first (in this case, following shortly after
   cholesterol), while small-volume pesticides such as HCB elute last.  Since
   molecular volume rather than molecular weight dictates the order of elution,
   all equatorial B-BHC elutes after the other BHC isomers.

          (2)  Equipment
              The gel permeation chromatograph is an AutoPrep Model 1001
   (Analytical  Biochemistry Laboratories, Inc., Columbia, MO), equipped with
   a  2.5  cm id  x 60 cm glass column  (Chromaflex3* R-422350/6025, Kontes, Vine-
   land,  NJ, or equivalent) packed with 200- to 400-mesh BioBeads SX-3  (BioRad
   Laboratories, Richmond, CA).

          (3)  Column Preparation and Operation

               (a)  Prepare a slurry  of  ca 60 g of BioBeads SX-3 in pesticide
   quality  (or  equivalent) toluene-ethyl acetate  (1:3 v/v).  This will be
   sufficient to pack a  column about 25 cm long.

               (b)  Add small volumes of resin and solvent  to the column.  "Each
   addition of  resin must be in  contact with enough solvent to swell  the resin
   before the next addition.

               (c)  After the resin  is  transferred to  the column, compress  the
   gel to approximately  25 cm, allowing solvent  to flow out of the column  exit.
                                        -321-

-------
                                                         Section 91
            (d)  Add only ethyl acetate-toluene  (3:1 v/v) to the solvent
reservoir.  Addition of other solvents to the system via sample introduction
will change the gel swelling ratio and must be  kept to a minimum  (i.e.,
< 5% v/v of aliquot injected).

            (e)  Install the column and start the pump.  The pump  operating
pressure should be 5-7 psi (not to exceed 10 psi).

            (f)  Adjust the pumping rate to approximately 5 ml/minute with
the pump vernier control valve.

            (g)  Set the timer to collect for 20 minutes and check the actual
pumping rate.

            (h)  The GPC elution pattern of the  pesticides of interest should
be established for standards before introduction of biological samples
into the gel permeation chromatograph,


       (4)  Procedure for GPC Cleanup

            (a)  Start up the GPC instrument and elute the column with ethyl
acetate-toluene (3:1 v/v) until it is purged of entrained air.

            (b)  Introduce 
-------
                                                         Section 91


mini-alumina column for improved purification of pesticide extracts from
fat samples.  In many cases, GPC fractions require no further cleanup prior
to determination of residues by GC.


d.  Application of GPC

    The original GPC system consisting of BioBeads SX-2 crosslinfced poly-
styrene gel and cyclohexane was designed by Stalling et al.  (41) for re-
moval of lipids from extracts of samples such as fish before EC-GC determina-
tion of commonly occurring pesticide and PCB residues.  This excellent
method was later improved considerably by the use of BioBeads' SX-3 and
ethyl acetate-toluene.  A broad range of OC1 and OP pesticides can be
recovered in good yields from fats and oils (42, Subsection b above).  An
evaluation of the GPC system (43) with different sample types indicated
that ca 98% of the fat or oil content of the extract is generally eluted
prior to the pesticide fraction and that this cleanup may be superior to
that achieved by acetonitrile partition and Florisil adsorption.  However,
although recoveries were higher by GPC than by Florisil adsorption, pre-
cision was poorer with the former method.  Analyses can be automated since
an important feature of GPC is that the same column can be, used repeatedly
over long periods without significant change in elution volumes or recoveries.

With the GPC procedure described in Subsection a above, organochlorine
pesticides have been determined and confirmed in human tissue and milk
(EPA PAM Section 12,A).  Samples, are extracted and cleaned up by a modified
Mills, Onley, Gaither procedure.  After further cleanup of Florisil fractions
by GPC, determination is carried out by GC on a Carbowax 20M column with a
Hall electrolytic conductivity detector.

Recoveries ranging from 88-106% were reported for disulfoton, diazinon,
methyl parathion, malathion, parathion, dichlorvos, and fensulfothion in
an evaluation study of the automated gel permeation chromatographic cleanup
techniques using BioBeads SX-3 gel and an ethyl acetate-toluene  (3:1 v/v)
elution solvent  (42).  A solvent composed of cyclohexane-methylene chloride
 (85:15 v/v) with BioBeads SX-3 provided adequate cleanup for EC-GC (no
liquid partitioning) of 9 OP pesticides and 2 metabolites  and 16 nonionic
OC1 pesticides in vegetable oils at 0.05-1.0 ppm.  Vegetables, fruits, and
crops were  analyzed for 26 organophosphorus pesticides and metabolites at
0.05-0.10 ppm levels using automated GPC for cleanup followed by FPD-GC.
Recoveries  of 7  compounds from 12  sample types were in the range 83-103%,
and 8  compounds  could be determined simultaneously  (44).   Carbamate and
organophosphorus compounds  in  several plant crops were recovered at levels
between 82  and 104% by automated GPC  (45).

Gel chromatography on  Sephadex LH-20 has also been reported  (46, 47) for
cleanup of  organochlorines  and organophosphates prior to GC, but this
approach  is now  of minimal  importance  in residue analysis.
                                     -323-

-------
                                                             Sections 9J,  9K
 9J   DETERMINATION OF CHLOROPHENOXY HERBICIDES IN 
-------
                                                            Sections 91, 9M
9L   CLEANUP ON SILICA GEL
   Silicar CC-4 silica gel (50) has been widely used for cleanup and fraction-
   ation of OC1 insecticides in various monitoring programs (51).  For
   example, in a study of duck wing contamination (52), a 15 g column was
   eluted with 60 ml of petroleum ether (HCB, mirex recovered), 350 ml of
   petroleum ether (PCBs, some DDE), and 150 ml of methylene chloride-
   hexane-acetonltrile (80:19:1 v/v) (remainder of DDE, IDE, DDT, and other
   OC1 compounds).  The same elution sequence was used to determine OC1
   residues in herons (53).  A modified sequence with four eluants, used
   to assess contamination of Bald Eagles, allowed collection of dieldrln
   and endrin in a discrete fraction:  80 ml petroleum ether (HCB and mirex);
   320 ml petroleum ether (PCBs, PBBs, DDE); 275 ml hexane-methylene chloride
   (85:15 v/v) (OC1 compounds, except endrin and dieldrin); 200 ml methylene
   chloride-hexane-acetonitrile  (80:19:1 v/v)  (endrin and dieldrin) (54).


9M   CLEANUP ON DEACTIVATED FLORISIL AND SILICA GEL  (see also Section 9G)

   The method of Osadchuk et_ al. is described  in the Canadian PAM, Section
   7.2,  Deactivated Florisil is prepared as outlined in Subsection 4Ac in
   Section 4 of this Manual.  The elution behavior of over 50 pesticides on
   Florisil deactivated, with 2% water has been determined for use after
   extraction and partition cleanup of residues.  A  30 cm'x 2.5 cm id column
   containing 15 cm of adsorbent is eluted with 300 ml portions of the
   appropriate eluting mixture(s) ranging from pure hexane to 5-30% methylene
   chloride in hexane to 5-30% ethyl acetate in hexane  (Table 9-1).  If the
   analyst wishes to screen a sample extract for a larger number of pesti-
   cides in one or two GC injections, the less polar eluants may be by-passed
   and only the more polar used.  However, some sample types may be inadequately
   cleaned-up by this procedure  or mutually interfering residues may occur in
   the same fraction.

   The following factors affect  the success of this  Florisil procedure:

        a.  Pesticides  containing a mercaptan  function are oxidized on the
   Florisil column.  For example, phorate, captan, carbophenothion,, chloro-
   benside, disulfoton, and demeton have losses ranging from 20-100%.  The
   oxidation  proceeds to the  sulfoxide and then to the sulfone.  Therefore,
   non-detection of  such pesticides does not guarantee they were not
   originally present in the  sample.  The degree of  oxidation by Florisil
   increases  with a  lower  extent of water deactivation  (greater adsorbent
   activity)  or  a greater  time of  contact with the column  and may also be
   affected by  the pH of the  particular Florisil used.

         b."  Oxygen analogs of organophosphorus pesticides  are  strongly adsorbed
    on Florisil  and cannot  be  completely  eluted even  with very  polar  solvents.

         c.   The 2% deactivated Florisil  column can  tolerate up to one  gram
    of fat or oil (30% methylene chloride in hexane  or  less polar eluants)
   without extraneous EC-GC response.

         d.   Up to two grams of fat or oil can be  applied  directly to  the column
    and eluted with 10% methylene chloride in hexane to recover BHC  isomers,  the
    DDT group, PCBs,  and HCB.
                                         -325-

-------
    Table 9-1
                                                           Section  9M
                         ORDER OP  EU'TION  OF  PESTICIDES  FHOM  FLORISIL  PARTIALLY

                        DEACTIVATED WITH & WATER USINO  300 mi  VOLUME  OF ELVENTS

                                        (?rom tht Canadian 7AM)
         PESTICIDES
                          Hexoho
                                                  in
                                                                              EtOAo InHexnne
                                                                           Parc»nt
                                                                          "Reeoverioo
 Aroclor 125^      PCB
 Chlordani
 Toxaphene
 Strobar.e

 Chlordane
 Aldrin
 Hexachlorobenzfin*
 Heptachlor
 p.p'-DDS
 o,p«-DDT
 Kirex
 Iiobenzan
 p.p'-RDT
 a-BHC
 Porthane
 p.p'-DDD
 Chlorbonside       M

 PCNB
 TCKB
 P-BKC
 f-BHO
 Dlcofol
 Konnol             OP
 Kcpachlor  epoxide
 Dicltlofonthion     OP
 Phorat*            6^M

 Carbophcnothicn    O^M
 Endosulfan I
 Dlcldrir.
 Chlorpyrifoo       OP
 Endrin
 Hethoxychlor
 Parathion          OP
 Ethion2            OP
 2.4-D eiethyl esttr
 2A5-T ncthyl ester
Anilazina
Ovex
Fenitrothion       OP
 Tetradifon
Diazinon           OP
Chlorothalonil  .
Methyl Parathion^  OP
Sulphenone
Dioxathion        OP
Malathion         OP
Atrazine*
K
OP

OP
  Endosulfan II
  Captan
  Fhoamet
  DCPA
  Arinphonnothyl.
                                               S(60J«)
                                                                    S(50JC)
                                                         +

                                                       s(75J«)
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >95
                                                                               >9S
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                              >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                               >95
                                                                              >95
                                                                               >95
                                                                              >95
                                                                              >95
                                                                              >95
                                                                              >95
                                                                                                   >95
                                                                                                   >95
                                                                                                   >90
                                                                                                >95
                                                                                               /v80
                                                                                                >95
                                                                                                >90
                                                                                                >95
  Not«  A 30^ CH?Cl2 fraction v/an eluted prior to all ethyl acetate fractions. All others were «incl«
                                                                                                  elution*.


        + - mostly *lutcs in  first  250 ml
        • - larce aisount in  250-300 ml  fraction
        S - some (ao percent)




Footnoea*: 1.  OP « organophosphorus; H - mercaptan; PCB - polychlorlnated biph*nyl
          2.  Higher recoveries are obtained by elutloa with more polar eluents
          3.  Beoainlng methyl parachion eluCes In another 50 ml of 51 EtOAc
          4.  Detected by alkali flame detector
                                                  -326-

-------
                                                            Sections 9N, 90
   The approach in d above has been used to determine HCB and mirex in fish
   and butterfat by elution with acetonitrile from a column composed of the
   fat or oil distributed on unactivated Florisil.  This procedure has been
   collaboratively studied (55) and adopted by the AOAC as official final
   action (56).  Care must be taken in analyses for HCB not to use plastic
   wash bottles, since this compound was found as a contaminant in 30 of 34
   such bottles tested (57).

   A one-step Florisil column cleanup described by Langlols et al. (58) has
   been widely used to isolate organochlorlnes and PCBs.  It is similar to
   the method just described but employs deactivated Florisil.  Activated
   Florisil is"equilibrated with 5% water, and 1 g of fat from fish or other
   extracts is thoroughly mixed with 25 g of this Florisil.  The adsorbent
   is placed on top of a second 25 g portion of conditioned Florisil in a
   25 mm id column.  The column is eluted with 300 ml of hexane-methylene
   chloride (4:1 v/v) (59).

   Silica gel deactivated with 30% water has been used to isolate organo-
   chlorines from lipids (60).  A micro column of this silica gel eluted
   with petroleum ether has been shown (61) to yield especially pure eluates.
   Small columns of precisely deactivated silicic acid (3 g, 3.3% water) were
   found to separate £,£f-DDT, cis- and trans-chlordane, 2.>£.'~DDE» *a& PCBs
   from the majority of toxaphene components.  This fractionation greatly
   simplified the analysis of the pesticides (62).


9N   LOW TEMPERATURE PRECIPITATION

   This procedure (Canadian PAM, Section 7.4) is used to separate fats, oils,
   and water from acetone-benzene-acid extracts of biological samples by pre-
   cipitation at -78°C.  The special low temperature cleanup apparatus is
   described in detail (Canadian PAM, Section 14.5).  Many apolar and polar
   residues and metabolites (e.g., DDT, 2,4-D acid and ester, parathion, and
   paraoxon) are retained in the acetone supernate and can be determined by
   EC-GC.  Forty pesticides have been quantitatively (80+ percent) recovered
   from a variety of plant and animal products at levels greater than 0.05 ppm.
   Freeze-out has been recently employed for the removal of lipids prior to
   Florisil chromatography and EC-GC in the determination of methoxychlor
   residues in microsamples of animal tissues and water at 10 ppb and 1 ppb
   levels, respectively (15), and for cleanup of human milk samples for ECB
   and other chlorinated pesticides by EC-GC (63).

90   CLEANUP ON ALUMINA

   Hexane extracts of animal tissues are cleaned-up and prefraetionated on
   narrow bore columns dry-packed with partially deactivated alumina and
   silica gel by the method of Holden and Marsden (64).  The initial alumina
   column eluted with hexane provides removal of lipids, while the second
   column affords pre-GC separation of residues plus further cleanup.  Table
   9-2 shows the elution order of chlorinated insecticides with hexane and
   10% diethyl ether-hexane eluants.  Alumina is activated at SOO°C and silica
                                      -327-

-------
                                                          Section 90

 gel at  150°C before deactivation with 5% (w/w) water.   Interferences con-
 tributed by columns in the Holden-Marsden method have been removed by
 methylene chloride treatment of the columns.   Basic alumina was recommended
 for easier control of activity and faster pesticide elution (65).

 Another alumina-silica column scheme (66) was devised for separation of
 17  OC1  residues in 4 eluates, each containing pesticides separable on a
 4%  SE-30/6% OV-210 GC column.  Microcolumns deactivated with 3-4% water
 were used.
Table  9-2


     ORDER OF ELUTION OF ORGANOCELORifoES FROM DEACTIVATED SILICA GEL

            ACCORDING TO THE METHOD OF HOLDEN AND MARSDEN (64)
Eluted in order
  by hexane
Eluted in order by 10%
diethyl ether in hexane
Hexachlorobenzene

Aldrin

PCBs

j3,_p_f-DDE

Heptachlor

p_,£'-MDE  (DDMU)  .
     Endrin
     Chlordane

     p_,p_'-DCBP
     Toxaphene

     p_,p_f-TDE

     Telodrin

     Heptachlor epoxide
     a-BHC

     Perthane

     B-BHC
     Kelthane

     Y-BHC
     Dieldrin
     Methoxychlor
Organochlorine insecticide residues in  fatty  foodstuffs were  determined  (67)
by using a cleanup technique based on a single  22  g  column of activity-4
basic or neutral alumina.  Concentrated hexane  extracts of samples, con-
taining 0.4-0.5 grams of fat, were transferred  to  the column, and pesti-
cides were eluted with 150 ml of hexane prior to determination by EC-GC.
                                     -328-

-------
                                                            Section  9P

  Recoveries of 15  Insecticides  from vegetable  oil  samples  spiked at levels
  of 5-250 yg/kg were between  70-124%.   Routine determinations were carried
  out  for cyclodienes,  EEC isomers,  and  HCB  at  the  5-10 yg/kg level and
  DDT-type compounds at the 20-30 yg/kg  level.   Results of  collaborative
  studies were reported.   If PCBs were present, the column  was eluted with
  10 ml  and then 150 ml of hexane.   The  first fraction contained all the
  PCBs and all or most  of any  residues of  aldrin, heptachlor, HCB,  £,£f-DDE
  and  j>,£f-DDT.  The second fraction contained  all  the BHC  isomers, heptachlor
  epoxide, dieldrin, endrin, £,£*-DDD, methoxychlor, and  Endosulfan A.   Com-
  pounds splitting  between fractions included methoxychlor, toxaphene, per-
  thane, chlordane, and strobane.  Further collaborative  study  (68) of the
  method found it satisfactory for determining  residues of  hexachlorobenzene
  and  g-HCH in butterfat and mutton  fat; a-HCH, Y-HCH, £,£»-DDT, and £,£f-DDE
  in chicken fat; g-HCH, dieldrin, hexachlorobenzene, and TDE In pork fat;
  DDT  isomers in eggs;  and other OC1 insecticides in these  and  other .samples
  of animal origin.

  A microcolumn of  2.0  g of Woelm basic  alumina deactivated with 11% water has
  been used for cleanup of water extracts  and fractionation of  residues.
  Petroleum ether  (5 ml) eluted HCB, a-  and  Y-BHC,  heptachlor epoxide  (10%),
  £,£f-DDE, £,£'-DDT, TDE, £,_p_'-DDT, telodrin,  isodrin,  aldrin,  and heptachlor.
  Subsequent elution with 10 ml of petroleum ether-ethyl ether  (80:20 v/v)
  recovered  &-BHC, heptachlor epoxide (90%), dieldrin,  and endrin (69).

  In a comparative  study (70), basic alumina was found to retain lipids
  better than Florisil, which  in turn held more than silicic acid.   It was
  also found  that  deactivation and elution with less polar solvents gave a
  superior separation of organpchlorine pesticides  from lipids  than activated
  adsorbents  and more polar eluants.  Saponification with ethanolic NaOH
  followed by alumina column chromatography provided efficient  removal  of
  lipids prior to  GC determination of several OC1 insecticides  (DDT was  con-
  verted to DDE)  (71).   A procedure for evaluation of the fat capacity  of an
  aluminum column has been described (68).

9P  MISCELLANEOUS  MULTIRESIDUE CLEANUP PROCEDURES

   Other multiresidue procedures include the following:  The method of  de
   Faubert Maunder  (72)  employs partition with dimethylformamlde (DMF)  to
   diminish the amount of fat carried over with the pesticides from fatty
   samples.  A hexane extract of the sample is extracted three times with
   hexane-saturated DMF; ,the combined DMF phases are washed with a DMF-
   saturated hexane and  then shaken with a large volume of 2% Na2SO^ solution.
   On  standing, a hexane layer containing chlorinated pesticides forms  on top
   of  the solution, this layer is separated, and residues are cleaned-up on
   an  alumina column and determined by EC-GC.  Like  the AOAC-MOG procedure
    (Section 9A), this method does not give good recoveries of hexachlorobenzene
   from  fatty samples.

   Wood  (73) proposed a rapid method for small  samples using  dimethyl sulfoxide
    (DMSO).  This is a good solvent for chlorinated pesticides that  dissolves
   only  low amounts of  oil  or  fat.  The  fatty sample is mixed with  Celite
                                        -329-

-------
                                                          Section 9P

 (1:15 w/w) and packed into a small column, and the pesticides are eluted
 with DMSO.  The eluate is adsorbed directly onto the top of a larger Florisil
 column and the residues then eluted. with hexane from the Florisil.   The
 method does not seem to he widely used.

 The de Faubert Maunder and Wood methods  have been compared with the
 standard FDA-AOAC Florisil procedure for analysis of chlorinated pesticides
 in a variety of foodstuffs (74).   No gross general differences were found
 in results, but one method might be advantageous for a  particular sample
 type.

 A rapid DMSO-petroleum ether partitioning cleanup method employing  test
 tubes and syringes in place of separatory funnels was found to recover
 60 OC1,  OP, and carbamate pesticides at  levels  > 50%.   Losses  were  found
 to be consistent,  so the use of correction factors was  proposed.  Crops
 containing 0.1-10  ppm levels were tested for analysis by GC (EC and FPD
 detectors), TLC, and HPLC (75).

 A reuseable,  macroporous silica gel column provided fractionation and
 88-105% recoveries of 0.1-1 ppm levels of different classes  of pesticides
 when eluted with a series of solvents of increasing polarity (76).

 Thin layer chromatography (TLC)  on 1-5 mm layers can provide cleanup if
 a minimal amount of fatty material is present in the extract.   Sample is
 applied as a  streak and  developed along  with standard marker compounds
 on the same plate  to allow location of the pesticide zones.  These bands
 are removed by scraping  and are extracted to recover the separated pesti-
 cides.  Modified layers  have been devised with  capability for  increased
 sample loading, e.g., multiband or wedge-layer  chromatoplates  (77).  With
 the latter, cleanup  and  determination can be combined on the same layer
without  intervening  elution.

 The use of ion exchange  resins  for cleanup  of ionic pesticides has" been
 reviewed  (78).  For  example,  acidic residues  such  as chlorophenols and
phenoxy acids  in extracts  of organic tissues, soil, and water will bind
under alkaline conditions  to a strong base  anion exchange resin.  After
washing out impurities,  the  residues  can be eluted from  the resin column
by  an acidic eluant  and  determined  by EC-GC after appropriate derivatization
reactions  (79).

The results of international  cooperative studies of OC1 pesticide, PCS,  and
Hg  residues in wildlife have been reported  (80).  The analytical methods
were based  on  extraction,  cleanup,  and GC determination, but no two labora-
tories used exactly  the  same procedure.  Nonetheless, there was reasonable
agreement among laboratories  in analysis of test samples, the coefficient
of variation for different chlorinated compounds ranging from 10-17%.
Collaborative  testing of a multiresidue method for chlorinated hydrocarbon
and other fumigant residues among 8 foreign laboratories was successfully
completed,  and results were reported  (81).
                                    -330-

-------
                                                            Section 9Q
                          ORGANOPHOSPHORUS PESTICIDES
9Q   DETERMINATION OF METABOLITES OR HYDROLYSIS PRODUCTS IN HUMAN URINE,
     BLOOD, AND OTHER TISSUES.

   The determination of intact organophosphorus pesticides in tissue or
   blood from suspected poisoning victims is described in Section 6,A,(1)
   of the EPA PAM (82).  However, in cases of low exposure or in high
   exposure cases after several hours, the probability of detecting parent
   compounds is greatly reduced because of rapid metabolism (83).  In most
   instances, the determination of alkyl phosphate metabolites in urine
   provides a measure of the extent of human exposure to the parent OP
   pesticide.  Section 6,A, (2),(a) of the EPA PAM and reference (84) contain
   a sensitive and selective analytical procedure for alkyl phosphate and
   phosphonate metabolites (hydrolysis products) of important pesticides.

   OP metabolites in urine are extracted quantitatively with an anion-exchange
   resin after addition of acetone in a 10:1 ratio to precipitate some inter<~
   fering compounds.  The compounds are eluted from the resin, derivatized
   with diazopentane (see footnote on page 6 for precautions when using this
   reagent), and the derivatives determined by FPD (P-mode)-GC.  If very low
   levels of alkyl phosphate metabolites are present, further cleanup on a
   2.4 gram silica gel column deactivated with 20% water is•carried out.
   Confirmation is by FPD-GC using both the P and S detector modes (recall
   that the S-mode is 5 to 10 times less sensitive).  Analysis can be made
   at the 0.1 ppm level, so that the excretion of alkyl phosphates in urine
   can be detected at pesticide levels much lower than those that result in
   cholinesterase inhibition.  The general class of organophosphate pesticide
   (but not the exact compound) involved in the exposure may be deduced by
   characterizing the metabolite(s) excreted.  These analytical methods have
   been applied to the analysis of the urine of rats exposed to a group of
   aromatic and aliphatic Of and phosphonate pesticides (85).       A

   Because of the complexity of this method, routine analyses should be
   validated by simultaneous analysis of spiked SPRM's.  As outlined in
   Section 3, one SPRM is analyzed along with each unknown if only occasional
   analyses are performed, or the ratio of SPRM to routine analyses is at
   least 10% when larger numbers are involved.  Because of the possible in-
   stability of urine samples spiked with alkyl phosphates, large samples of
   SPRM should not be prepared ahead of time for periodic analyses.  A
   method for preparation of individual SPRM as needed is detailed in
   Section 6,A,(2),(a),XI of the EPA PAM.  However, it has been shown (86)
   that dialkyl phosphate metabolites do not break down or disappear in urine
   samples frozen for up 20 weeks prior to analysis.

   Underivatized compounds may accumulate on the GC column after periods of
   extended use.  Injection of 1 Ul of diazopentane .solution should be made
   every two weeks to react with these compounds.  If peaks appear following
   this injection, the column should be reconditioned (Subsection 41 in
   Section 4).  Further confirmation of any particular metabolite can be
   accomplished by preparing its hexyl derivative.
                                       -331-

-------
                                                            Section 9R


   Reproducibility of this method is not as good as is desirable for a re-
   liable, routine analytical method.  This can be seen in Table 3 in EPA
   PAM Section 6,A,2,(a), where recovery variabilities of 15% or greater are
   reported for six analyses at the two highest spiking levels.   In the study
   of freezer storage of alkyl phosphate metabolites described above (86),
   the method was found to give both low and highly variable recoveries.
   Because of the unreliable quantitation obtained, the method currently
   described in Section 6,A,2,(a) of the EPA PAM should be considered only
   semi-quantitative.

   In addition to alkyl phosphates, significant amounts of the corresponding
   mono- and dicarboxylic acids are found in the urine of humans exposed to.
   malathion.  A silica gel cleanup FPD-GC method for determining these
   acids as a measure of exposure to malathion has been devised (87).  Urine
   is extracted$ the extract is alkylated, and derivatized carboxylic acids  .
   are cleaned up according to a previously published (88) alkyl phosphate
   method.  Additional cleanup by solvent partitioning with ether and sili'ca
   gel chromatography [elution with benzene followed by ethyl acetate-benzene
   (10:90 v/v), collected as one fraction]-is also employed,  Derivatized MCA
   and DCA are determined on a 4% SE-30/6% QF-1 column at 200°C.

   A reportedly simple and rapid method for quantitation of the metabolites of
   malathion and other OP pesticides has been published (89).  The omission of
   an extraction at low pH and the mild condition of anion-exchange chromatog-
   raphy on QAE-Sephadex prevented degradation of a malathion metabolite that
   takes place under strongly acid conditions.  Disadvantages of the commonly
   used partition fractionation of malathion and malaoxon metabolites were
   discussed.

   Another new method also employing an ion exchange resin for determination
   of mono- and diportic alkyl and aryl phosphates, phosphonates, and thio
   analogs in human urine has been reported to have a detection limit of less
   than 2 pmole for each of these classes of compounds.  The acids were
   protonated by passing through a hydrogen-form cation exchange resin.
   Benzyl esters were formed by refluxing the column effluent with 3-benzyl-
   l-£-tolyltriazene in acetone, partitioned into cyclohexane, and determined
   by GC (5% OV-210 column) with a P-mode FPD.  Inorganic (^-phosphate did not
   interfere, but could be removed by calcium hydroxide precipitation if
   desired (90).

   Urinary dialkyl phosphate metabolites have also been determined using
   l-(4-nitrobenzyl)-3-(4-tolyl)triazene as derivatizing reagent.  Urine was
   lyophilized, dialkyl phosphates were derivatized, and cleanup was carried
   out by anhydrous nickel sulfate adsorption and silica gel chromatography.
   GC analysis determined the metabolites at levels as low as 0.01 ppm (91).

9R   DETERMINATION OF £-NITROPHENOL  (PNP) AND OTHER PHENOLS IN URINE

   Urinary PNP, the phenolic metabolite of ethyl and methyl parathion, EPN
   (0-ethyl 0-jj-nitrophenyl (phenylphosphonothioate) nitrofen, etc., can be
                                       -332-

-------
                                                            Section 9S
   measured as an indicator of exposure to these organophosphorus pesticides.
   A small volume of urine is hydrolyzed with HC1 to form free PNP, then made
   alkaline and cleaned-up by extraction with .benzene-ether, and finally re-
   acidified and extracted with benzene-ether to remove PNP.  An aliquot of
   dried extract is analyzed by EC-GC with on-column conversion of PNP to the
   volatile trimethylsilyl derivative [EPA PAM, Section 6,A,(2),(b)].

   A multiresidue analytical procedure for halo- and nitrophenols from a range
   of biodegradable pesticides (organophosphates, phenoxy acids, organohalides)
   is also useful for determining exposure to these pesticides (92, 93).  A
   one to five ml sample is treated with a 1/5 volume of concentrated hydro-
   chloric acid and the mixture refluxed at 100°C for one hour.  The phenols
   are extracted with diethyl ether, ethylated by reaction with dlazoethane,
   and the ethyl-ethers chromatographed on a silica gel column (2 grams, 2%
   water deactivation).  (See the footnote on page 6 concerning precautions
   when using diazoalkanes)*  Elution with various concentrations of benzene
   in hexane purifies and fractionates the phenolic ethers, which are finally
   determined by EC-GC.

   Ten phenols, including the pesticides pentachlorophenol- and DNOC
   (4,6-dinitro-o_-cresol), plus the herbicides 2,4-D, 2,4-,5-T, and silvex can
   be determined by this scheme on one sample.   All halogenated phenols are
   eluted with 20% benzene-hexane, while nitrophenols and phenoxy acids elute
   in the 60 and 80% fractions.  The phenoxy acids are detected intact along
   with 2,4-dichlorophenol and 2,4,5-trichlorophenol, their potential mammalian
   metabolites.

   A method for the determination of residues of the herbicide DNBP
   (2-sec-butyl-4,6-dinitrophenol) in feed, blood, urine, feces, and tissues
   by EC-GC has been devised in the EPA Health Effects Research Laboratory (94).
   After extraction, the sample is reacted with diazomethane (see footnote on
   page 6 concerning precautions when using diazoalkane) to produce the methyl
   ether of DNBP."'  Cleanup and recovery of the derivative is obtained on acid
   alumina column eluted with hexane-benzene (40:60 v/v).  Average recoveries
   of greater than 85% were obtained from samples fortified at 0.1-30 ppm
   levels.


9S   SWEEP CO-DISTILLATION

   Sweep co-distillation has proven to be a simple time saving cleanup technique
   that eliminates the need for specialized adsorbents and large volumes of
   purified solvents (8, 95-100).  The technique can be used for OC1 and OP
   residues in fruits and vegetables, or fats and oils.  The procedural details
   are different for the two sample types; however, the cleanup principle is
   essentially the same.  The concentrated sample, in an organic solvent, is
   injected into a heated tube swept with 600 ml N2/min.  Sample extractives
   remain in the tube while volatilized components are swept into a simple
   condensing train.  After a 30 minute sweep time, the transfer lines are
   disconnected and condensed pesticides are rinsed with organic solvent into
   the sample tube.  After volume adjustments, the sample may generally be
                                       -333-

-------
                                                          Section 9S

 analyzed by GC without further cleanup.   If sensitivity levels  in the  low
 part  per billion range are desired,  an auxiliary cleanup is•recommended.
 The combination  of sweep  co-distillation and the micro  Florisil column
 [EPA  PAM Section 5,A,(2)]  has  proven to  be a thorough cleanup for fat
 samples.  A sulfuric  acid/Celite column  can be  adapted  as an optional
 automatic cleanup step (101).

 Figure  9-A  is  a  schematic  diagram of the apparatus  as originally used  for
 cleanup of  fruits and vegetables for determination  of OP residues.  The
 glass wool-packed tube was placed inside a heated copper tube.   A nitrogen
 sweep of 600 ml/min was used.   Two gram  aliquots of sample were injected
 followed by ethyl acetate  injections every three minutes. See  FDA PAM
 Section 232.2  for method details.

 OC1 and' OP  residues in a variety of  edible fats and oils have been
 determined  by  a  modified version of  the  sweep co-distillation cleanup
 system  (102,- 103).  A tube packed with glass wool,  sand, and glass beads
 is operated in a vertical  position with  the injection port on bottom.
 The cleanup is effected by the 250°  heat and the nitrogen carrier gas
 distributing the oil  upward through  one-half to three-fourths of the glass
 bead  packed column with a  percolation type action.   Pesticides  are volatilized
 and swept into the collector trap.   Recent study of sweep co-distillation of
 fats  has  shown that follow-up  injections of solvent are not  necessary.  After
 initial injection of  the sample, the equipment  may  be left unattended  for the
 30 min  sweep operation.

Figure 9-B shows the appearance  of a commercial version of the "Sweep
Co-distiller"  (Kontes Glass Co., Vineland, NJ).   This apparatus  permits
Simultaneous cleanup of four samples with a 30 min sweep time.   The 30 cm
tube allows  efficient cleanup for OC1 or OP residues in samples  of fats,
oils,  milk,  and crops (operated  in vertical position at 250°C with 600 ml
N2/min)  (104).   The tube for fat cleanup may be purchased prepacked, but
packing in the laboratory is preferable for consistent tube uniformity.
The empty tube may also be prepared for fruit and vegetable cleanup by
packing withC'15 cm glass wool in the injection end with remaining space
filled with glass beads.  The oven would be swiveled to a horizontal position
for the fruit and vegetable cleanup.   Operational parameters for the latter
application may be found in the FDA PAM,  Section 232.2.

In a preliminary evaluation of the Kontes apparatus by Watts, common
organochlorine pesticides were quantitatively recovered from chicken fat,
and the fat  residue was reduced  to less than 1% of the original  sample.
Similar results have been obtained by Luke with both the Kontes  and labora-
tory-assembled sweep-co-distillation units.  An oven temperature of
227-230°C was used, and no solvent injections were made after the sample
was applied.  Reproducible, quantitative recoveries were obtained for
organophosphorus and organochlorine pesticides from beef fat and butterfat
 (105).

The operating principle of sweep co-distillation has been presented diagram-
matically, and recoveries of 36  OP pesticides in 20 substrates (0.03-0.5  ppm)
and 30 OC1 pesticides in 14 substrates (0.003-0.05 ppm)  are tabulated (106).
                                    -334-

-------
                                                                 Section 9S
Figure  9-A.   Sweep  co-distillation apparatus,  schematic diagram.
                                             ^Insulation
                                             .-Asbestos
                                             -Heating tape

                        To pyrometer
                     Septum	.
                     Scrubber
                     tube
                                  Water
                                  and ice
                                  bath
                    Storherr tube
                    containing 5-6 inches
                    silanized glass wool
                   Nitrogen
            -Coding  in|e*
             coil
_ Glass wool
 (silanized)

—4cm Anakrom

 -Adapter
 -Glass wool
  (silanized)
                                  -J 19/22
                                  "'Concentration
                                  tube

                                  -Beaker
                                  of water
  Figure 9-B.   Sweep co-distillation apparatus,  (oven positioned  for
                  fruit and vegetable  cleanup),  Kontes Glass  Co.,
                  K-500750.
                                     -335-

-------
                                                                      Sections 9T,  9U
             Heath and Black (107)  recommended the following modifications  for faster
             and more convenient cleanup of organochlorine residues  in animal  fat:   no
             solvent introduction;  230°C distillation temperature; 600 ml/minute nitrogen
             flow; 6.7 mm distillation tubes with simplified packing;  and incorporation
             of a U-tube condenser  that allows direct introduction, onto a Florisil  column
             for secondary cleanup.

             A literature review of the applications  of sweep co-distillation  including
             a comparison to other  cleanup methods has been published  (108).


          9T   CHARCOAL CLEANUP OF  NONFATTY FOOD EXTRACTS

             A general determinative method for organophosphorus  pesticide  residues in
             nonfatty foods is  based on the FDA acetonitrile (or  water/acetonitrile)
             extraction procedure followed-by dilution with methylene  chloride to
             separate water,  cleanup on a short charcoal column,  and analysis  by GC with
             a P-selective detector.   The chromatographic tube (300  mm x 22 mm id)  is
             packed dry with a  one  gram layer of Celite 545 followed by 6 grams of
             adsorbent mixture  (acid-treated Norit SG-X or Nuchar C-190 charcoal-hydrated
             magnesium oxide-Celite 545,  1:2:4 w/w) and finally glass  wool  topping,  and
             the column is eluted with acetonitrile-benzene (1:1  v/v);   The satisfactory
             recovery of 41 pesticides and alteration products from  kale and 9 typical
             pesticides from other  low and high sugar content crops  was demonstrated
             (109).   A collaborative study (110)  of this method for  residues of six OP
             compounds in apples and green beans verified recoveries between 86 and
             125% when either a thermionic or FPD detector was employed.  The  method
             is described in the FDA PAM,  Section 232.3., and recoveries of 51 pesticides
             and related chemicals  are listed in Table 201-H of the  FDA manual.  Sections
             4Ae and f of this  Manual  describe procedures for purification  of  Celite
             and carbon adsorbents, 'respectively.


          9U   ACETONE EXTRACTION .,"

             The FDA PAM contains details of a procedure for determination  of  polar
             organophosphate and organonitrogen pesticides in nonfatty samples (FDA PAM
             Sections 232.4 and 242.1).   Samples are  blended with acetone and  filtered,
             pesticides are extracted  from the aqueous filtrates  into  petroleum ether-
             methylene chloride,  and an aliquot of concentrated extract is  determined
             by GC with a P- or N-selective detector.   Lack of a  column cleanup  step
             allows determination of many polar compounds that would not be recovered
             from adsorbents  such as Florisil or charcoal,  but a  specific detector,
             rather than electron capture,  must be used.   Repeated injection of impure
             extracts can shorten column life,  so  that packing material at  the head of
             column will need to be  replaced often.   A short (0.6 or 0.9 m) column of
             & polar phase, such as  DEGS or Carbowax  20M,  will probably be  advantageous
             for the chromatography  of polar compounds.   If it is desired to examine
             some pesticides  with the  electron capture detector,  cleanup of acetone
             extracts of nonfatty foods is  carried out on a Florisil column (FDA PAM
             Section 212.2).  A list of pesticides recovered through these  procedures,
             with and without Florisil cleanup,  is given in the FDA  PAM,  Table 201-1.
_
                                                -336-

-------
                                                            Section 9V
9V   MISCELLANEOUS MULTtRESIDUE CLEANUP PROCEDURES
   Nine extraction procedures were compared for efficiency of removal of six
   OP pesticides and metabolites from field treated crops.  Soxhlet extraction
   of the finely chopped crops with chloroform-methanol (90:10 v/v) proved
   most reliable and efficient (111).

   Alumina has not proven totally satisfactory;for cleanup of OP compounds
   since recovery of the more polar compounds is not complete (112).  Using
   alumina (activity II to III) and petroleum ether and petroleum ether-
   acetone (97:3 v/v) as eluants, Renvall and Akerblom (113) eluted only 13
   of the 31 OP compounds they tested.  However, many residue analyses are
   based on alumina column cleanup, e.g., the determination of carbophenothion
   in goose tissues (114) and monocrotophos in tobacco (115) by FPD-GC.

   The Abbott et al. method (116), involving cleanup by solvent partition with-
   out column chromatography, has proven adequate for analyses of seven types
   of foods for 39 pesticides and metabolites when detection was made with a
   thermionic detector.  Finely chopped sample is mixed with anhydrous sodium
   sulfate and extracted with acetonitrile.  The extract is diluted with a
   large volume of aqueous sodium sulfate, and the pesticides are extracted
   into chloroform.  The chloroform solution is dried and concentrated for
   GC.  Other determinations without column cleanup have been reported.
   Methyl parathion, diazinon, malathion, and phorate were,determined in plant,
   animal, water, and soil samples by EC-GC following only hexane extraction
   and partition with aqueous acetonitrile (117).  Azinphosemethyl and
   dimethoate residues in apple leaves were determined by FPD-GC following
   ethyl acetate extraction and cleanup by methylene chloride-water and hexane-
   kcetonitrile partitionings  (118).  A multiresidue analysis of 14 pesticides
   '?,n natural waters at ppb levels involving extraction and concentration
   before FPD-GC has been reported (119).

   The elution pattern of a series of representative OP pesticides from a
   column  (Kontes, Size 22) containing one gram of Wbelm silica gel deactivated
   with 1.5% water and prewashed with 8 ml hexane before applying the sample
   mixture is as follows:
                Eluant

         7 ml hexane
         8 ml 60% benzene-hexane
         8 ml benzene
         8 ml 8% ethyl acetate-benzene
         8 ml 50% ethyl  acetate-benzene
Pesticides Eluted


carbophenothion
ethyl parathion

malathion; diazinon
paraoxon
                                        -337-

-------
                                                          Section 9V

 Silica gel or silicic acid columns have been used for cleanup of animal,
 plant, soil, and water extracts prior to GC determination of OP pesticides
 (120-122)  and to separate OP pesticides and metabolites into groups to
 facilitate their identification by GC (123).  A tandem column of silica
 gel and alumina was used to separate leptophos and its oxygen and 2,5-
 dichlorophenol analogs prior to determination by FPD-GC (124).

 A rapid, simple approach has been developed for approximately the total
 residues of pesticides such as fenthion, disulfoton,  and phorate, which
 may consist of the parent pesticide and up to five metabolites formed by
 oxidation of thionophosphate and sulfide groups in each molecule.  The
 insecticides and any metabolites are oxidized to t;he  oxygen analog sulfone
 with m-chloroperbenzoic acid, followed by removal of  the acid on an alumina
 column and determination of the sulfone by ITD-GC. Quantitative re-
 coveries of parent pesticides and metabolites from corn, milk, grass,
 and feces  have been demonstrated (125).   Metasystpx-R and its sulfone
 were determined in plant and animal tissues and water at 10 ppb levels as
 the sulfone after oxidation by KMnO^(126).  '

 A method for 40 organophosphorus pesticide residues in plant material
 involved extraction with acetonitrile, partition into methylene chloride,
 and GC with a P-selective thermionic detector (127).   The same authors
 used this  extraction and cleanup procedure for plant  material subsequent
 to oxidation with potassium permanganate to convert organophosphorus
 pesticides containing thioether groups (e.g., demeton, disulfoton, phorate)
 to sulfones (128).

A  collaborative study by 12  laboratories of  the methods  of Abbott et al.
 (116;  see above,  this Subsection), Watts ££ al., and Sissions and Telling
was  conducted for OP pesticides  in fruits  and vegetables  (129).  The method
of Abbott _et al.  was found satisfactory for  determination of malathion,
dichlorvos, dimethoate, omethoate, and parathion in 6 fruit and vegetable
crops  (>90% average for all  pesticides and crops at 0.5-2 mg/kg) and was
judged widely applicable to  the  determination of many other nonpolar and
medium-polarity OP pesticides and to a wider range of samples.  The method
of Watts et al. (130), involving ethyl acetate extraction and cleanup on
a column of activated charcoal-magnesium oxide-Celite eluted with ethyl
acetate-acetone-toluene (an  early version  of the procedure described in
Section 9T), was  found satisfactory for the same pesticides plus azinphos-
methyl in 6 crops  (>90% average  recovery at 0.5-2 mg/kg) and was also Judged
to be much more widely applicable.  The method of Sissions and Telling (131),
employing cleanup by batch addition of charcoal followed by hexane and
hexane-acetone (98:2 v/v) elution through  an activity -5 alumina column
was not successful for the more  polar pesticides studied.  Details and
modifications of  these methods are discussed in the report of the collabora-
tive study.
                                     -338-

-------
                                                           Section 9W
  The methods of Abbott et al. (.116) and Sissons and Telling (112) and the
  sweep co-distillation method (Section 9S) were compared for determination
  of different OP pesticide residues in various vegetable crops.  There was
  no significant difference for most pesticide-crop combinations, except
  that sweep co-distillation tended to give lower results for polar compounds
  such as omethoate (132).

  A tabulation has been made  (133)  of the validated applicability of  five
  multiresidue analytical methods  to the .determination of some  50 OP
  insecticides, acaricides, and nematocides.   These procedures  were the
  AOAC  (llth ed.) 29.001-29.027 general Florisil cleanup method for OC1
  and OP pesticides;  the AOAC (llth ed.) 29.028-29.033 multiple residue
  carbon column cleanup method for OP pesticides; the  AOAC  (llth ed.)
  29.034-29.038 single sweep  oscillographie polarographic confirmatory
  method;  the Abbott\et al. method for total  diet studies  (116); and  an
  undescribed German  procedure (134).  In  addition, individual  determina-
  tions of some of  the compounds by other  special methods were  reviewed.
  It was stated that, in  general,  the multiresidue methods  were not usually
  suitable for metabolites, requiring separate analysis for the parent
  and metabolite; each method should be compound validated  in the worker's
  own  laboratory; and that  differences in  results were more likely  to arise
  from sampling problems  than from the analytical methods' themselves.

  The  use  of  a  selective  detector  sometimes  allows determination of OP
  pesticides with no  cleanup.  Por example,  a collaborative study of  the
  analysis of wheat for  chlorpyrifos methyl,  fenitrothion,  malathion,
  methacrifos,  and  pirimiphos methyl  involved only methanol extraction for
  40 hours followed by GC of  an aliquot  using a FPD or alkali flame ioniza-
   tion detector (135).

      CARBAMATE PESTICIDES AND METABOLITES AND MISCELLANEOUS. HERBICIDES


9W   1-NAPHTHOL IN URINE

   Humans exposed to the N-methyl carbamate insecticide carbaryl excrete in
   urine relatively large quantities of the metabolite 1-naphthol conjugated
   as either the sulfate or glucuronide.   Determination of 1-naphthol is
   made by subjecting 5 ml of urine to acid hydrolysis under reflux to break
   conjugates, extracting the 1-naphthol with benzene, and derivatization
   with chloroacetic anhydride solution.   After cleanup on a small silica
   gel column (1 gram, 1.5% water), the derivative is quantitated by EC-GC
   against a peak from standard 1-naphthol similarly derivatized.  Details
   are found in Section 7,A of the EPA PAM.

   Elution patterns from the silica gel column must be established at the
   temperature and humidity conditions prevalent in each laboratory.  Spiked
   control urine treated in the same manner as routine samples  is used for
   this purpose.  Traces of water can affect  the derivatization reaction and
   must be avoided.   Derivatized standards are stable  for about 6 months
   if stored in a refrigerator.
                                        -339-

-------
                                                            Sections 9X; 9Y
9X   ANALYSIS OF AMINE METABOLITES IN URINE
   A method for determination, of amine metabolites from anilide, urea, and
   carbamate pesticides was developed in the EPA Research Triangle Park
   Laboratories (136).  Pentafluoropropionic anhydride was the preferred
   derivatization reagent for the aniline compounds, with cleanup on 1 gram
   deactivated (3) percent water) silica gel columns.  Determination was by
   EC-GC on a 3% OV-1 column.  Recoveries ranged from 85-90% at 1.0 and
   0.1 ppm.


9Y   OTHER INDIRECT (DERIVATIZATION) METHODS OF ANALYSIS

   Numerous derivatization methods have been used for the indirect measure-
   ment of residue levels of parent carbamate insecticides in a variety of
   agricultural crops and other substrates.  These have involved derivati-
   zation of the amine or phenol moieties of the pesticides after hydrolysis,
   or, less often, the intact insecticide.  These derivative methods include
   reaction of intact insecticides with bromine, silylating reagents, acetic
   anhydride, and trifluoroacetic anhydride.  Phenols resulting from alkaline
   hydrolysis of the parent insecticides have been reacted with bromine
   (with or without simultaneous esterification), silylating reagents,
   mono- and trichloroacetyl chloride, pentafluorobenzyl bromide, and
   l-fluoro-2,4-dinitrobenzene.  The latter reagent is used for derivatization
   of carbamate insecticides in the method for water analysis (Section 9A,C)
   discussed in this Manual and described in detail in the EPA PAM, Section
   10, A.

   Amine hydrolysis products of carbamate insecticides have been reacted with
   l-fluoro-2,4-dinitrobenzene and 4-bromobenzoyl chloride.  These and other
   reactions have been surveyed in a review article (137) in which pertinent
   references are given.

   GC methods for phenyl substituted urea and carbamate herbicides are usually
   based on hydrolysis followed by determination of the corresponding aniline.
   Anilines have been derivatized with halogen, 4-chloro-a,a,a-trifluoro-3,5-
   dinitrotoluene, l-fluoro-2,4-dinitrobenzene, and pentafluoropropionic
   anhydride.  These reactions are also reviewed in reference (137).

   A fluorogenic labeling derivatization reaction with dansyl chloride has"
   been combined with EPLC for the determination of N-methylcarbamate insecti-
   cides in soil and water.   No preliminary cleanup was required, and de-
   tection limits were 1-10 ng/4 pi injection (138).

   Ten triazine herbicides were determined in vegetables at levels of
   0.13-0.86 ppm by preparation of heptafluorobutyryl derivatives.  Compared
   to the parent compounds,  the products were at least 300 times more sensi-
   tive to electron capture detection and 5-10 fold more sensitive to Cl-mode
   electrolytic conductivity detection (139).
                                       -340-

-------
                                                            Sections 9Z, 9A,A
9Z   DIRECT METHODS OF ANALYSIS
   Determinations of intact, underivatized N-methylcarbamate insecticides
   are hampered by their decomposition on GC columns under ordinary operating
   conditions (140).  Losses can be minimized by the use of specially prepared,
   conditioned, and maintained .columns.  Presilanized supports do not provide
   sufficient deactivation to prevent degradation of carbamates, so it is
   necessary to employ in situ silanization both during initial conditioning
   and thereafter to restore column performance.  Examples of direct analyses
   of crop extracts include a multiresidue method (141) on 5-6% DC-200 after
   acetonitrile partition and charcoal cleanup as for OP pesticides (109),
   determination of carbofuran and other carbamates on 20% SE-30 (142), and
   determination of 0.2-15 ng of carbaryl and 1-naphthol on a short column
   of 3% SE-30 (143).  Highly deactivated GC columns prepared from acid
   washed Chromosorb W support that is surface modified with Carbowax 20M
   have also been successfully used for chromatography of intact N-methyl-
   carbamates without degradation on the column.  Such columns, which are
   extremely promising for performing analyses without required derivatization,
   are described in Sections 4J and 5Lb of this Manual.

   Urea and N-arylcarbamate herbicides are, in general, more thermally stable
   than carbamate insecticides, and' are, therefore, more amenable to direct
   determination by GC,  For example, columns of 5% E-301 methyl silicone
   at 150°C  (144), 10% DC-200/15% QF-1 (1:1) at 160°C  (145), and'5 and 10%
   DC-200  (146) have been successfully used, the former for multiresidues of
   urea herbicides and the latter two for carbamate herbicides in foods. '
   However, decomposition of compounds on these column types has been noted
   under certain conditions, and determinations are therefore often made via
   thermally stable derivatives of hydrolysis products or directly on Carbowax
   20M-treated columns.  As an example of the latter, carbamate insecticides
   and herbicides have been directly chromatographed on Carbowax 20M modified
   (Ultra-Bond) supports containing 1-3% of a liquid phase such as OV-17,
   OV-101, or OV-210.  The Hall electrolytic conductivity detector was used,
   and determinations in soil were demonstrated  (147),

   s-Triazine herbicide residues were  determined in urine by hexane extraction
   from a  sample at pH 12,  drying of the extract by passage through a sodium
   sulfate column,  concentration of the extract, and GC using a N-mode Hall
   conductivity detector  (148).  Similar N-specific GC methods involving
   cleanup were used  to monitor triazines in European  streams  (149).
 9A.A
ANALYSIS OF PLANT AND FOOD MATERIALS
    Extraction of urea and carbamate pesticides from plant materials usually
    involves blending with methylene chloride, acetone,  chloroform,  acetoni-
    trile,  or an alcohol (or these solvents plus anhydrous Na2S04).   If the
    presence of conjugates of hydroxy metabolites is suspected,  hydrolysis
    with an acid during extraction may be included (Section 9A,L).

    Cleanup steps include solvent partition and/or liquid column chromatog-
    raphy,  the exact nature of which are pesticide- and sample-dependent.
                                       -341-

-------
                                                           Section 9,A,B

   For  example,  a column  of 4:1 MgO-cellulose was used  for cleanup of
   carbamate herbicide  residues from a variety  of foods (145), while Florisil
   was  employed  after acetonitrile-petroleum ether partition  for  the multi-
   residue,  multiclass  determination of carbamate, urea, and  amide residues
   (150).  Methods for  extraction, cleanup,  and GC of carbamates,  ureas,
   and  other classes of herbicides  (triazines,  uracils, phenols) have been
   reviewed  (151-153).  A multiresidue method for twelve triazine  herbicides
   in crops, water, and soils involving methanol extraction,  alumina column
   cleanup,  and  gas chromatography with a Carbowax column  and thermionic,
   microcoulometric, FPD, and electrolytic conductivity detectors  has been
   reported  (154).  Residues of 15 organonitrogen herbicides and fungicides
   were screened in foods by acetone extraction, partition  and Florisil
   (2Z  water) column cleanup, and CCD-GC determination  (155).  Herbicides
   of different  types were  determined in crops  at tolerance levels with no
   column cleanup  prior to  GC with N- and Cl-mode conductivity detection
   (156).  Total .residues of Mesurol and dts sulfoxide  and sulfone metabolites
   in plant and animal  tissues were  determined by oxidation of the extract
   with KMn04 to convert all residues to the corresponding sulfone, which
   was  detected at a limit  of 0.03 ppm by a S-mode FPD  (157).

9A,B   AIR ANALYSIS

   Section 8,B of the EPA PAM contains details of analytical methods for
   chlorinated, organophosphorus,  and N-methylcarbamate insecticides
   collected by one of the procedures described in Section 8,A of the
   EPA PAM or Section 8H of this Manual.   The sampling medium is extracted
   with hexane-diethyl ether (95:5 v/v).   Chlorinated pesticides and PCBs
   are measured by EC-GC after column chromatographic cleanup on alumina.
   PCBs are separated from technical chlordane and other pesticides by
   column chromatography on silicic acid deactivated with distilled water.
   Organophosphorus pesticides are determined by direct injection of an
   aliquot of extract into a chromatograph equipped with a flame photometric
   detector.   Carbamate pesticides are determined directly by GC using a
   N-mode Hall detector and a 3% OV-101/Ultra Bond 20M column.  As an
   alternative for carbamates,  derivatization is carried out with  o-bromo-
   2,3,4,5,6-pentafluorotoluene.   The derivatives  are cleaned-up and
   fractionated on a column containing-1  g of deactivated silica gel  and
   determined by EC-GC.   Collection efficiencies of  OC1  and OP pesticides
   and  PCBs using different samplers  and  collection  media as determined
   with these analytical procedures are tabulated in Section 8,B of the
   EPA PAM.

   The  air analysis method reported earlier in the EPA PAM was a multiclass,
   multiresidue procedure  (158) for residues  collected in  ethylene  glycol,
   in which prefractionation was carried out  on  a 1  g column of  silica  gel
   deactivated with 20%  water.  A  still earlier  method for  air analysis
   (159) included Florisil column  chromatography and  was the basis  of the
   former EPA National Pesticide Monitoring Program  (160).   Details of
   these can  be found in earlier editions of  the EPA  PAM, but  use of the
   current, more  widely  applicable and tested procedures, described above,
   is generally recommended.
                                     -342-

-------
                                                          Section 9A,C
9A,C   WATER ANALYSIS

   A broadly applicable multiresidue, multiclass method for the monitoring
   of water samples for pesticides is presented in Section 10,A of the
   EPA PAM (161).  Recovery studies were conducted on 42 halogenated com-
   pounds, 38 OP compounds, and 7 carbamates.  Recoveries of >80% were
   achieved for 58 of the 87 compounds, 60-80% recovery for 13 compounds,
   and <60% for the remaining 16 compounds  (concentration levels 0.09-400 ppb).
   Pesticides are extracted from water with methylene chloride, and the con-
   centrated extract is chromatographed on a 1 gram deactivated (20% water)
   silica gel column with four different solvents of increasing polarity to
   separate the pesticides into groups.  OC1 compounds are determined by
   EC-GC, OP compounds by FPD-GC, and carbamates by EC-GC after conversion
   to 2,4-dinitrophenyl ether derivatives.  Low recoveries were in most
   cases traced to losses during the silica gel chroma tography step.
   Evaporation of solutions by air blowdown should not be used because losses
   of all three classes of pesticides may occur.  Concentrations are carried
   out under a gentle stream of nitrogen.   It is important to apply the con-
   centrated extract to the silica gel column at the exact moment the last
   of the hexane prewash reaches the top surface of the column.  The total
   0.5 ml extract plus the 1.0 ml hexane rinse must be transferred to the
   column without loss to minimize the recovery error.  Solvents contained
   in several eluate fractions from  the silica gel column may interfere in
   the GC and carbamate derivatization steps.  It is critical to follow the
   directions for solvent removal and exchange outlined"in Section 10,A of
   the EPA PAM.  Sufficient silica gel should be activated (at 175°C) to
   provide only  a one-week supply, and deactivation should be carried out
   only  on the amount required for a 2 or  3 day. period.  Longer storage
   periods may result in a shift of  the pesticide elution pattern of the
   final deactivated columns.  Each  lot of  silica gel  should be tested for
   the proper elution pattern with representative pesticide  standards
   eluting in each  fraction.  A number of  the OP compounds require con-
   siderable column pre-conditioning by repetitive  injection of high-
   concentration standards  in order  to obtain  linearity of response and
   accurate  quantitation.   Confirmation of pesticide  identity should be
   made  by several  techniques outlined  in Section  10.

    Section 10,B  of  the  EPA PAM describes  the determination of some free
    acid herbicides  (e.g.,  MCPA,  2,4-D,  2,4,5-T)  in water.  The water  is
    adjusted to  pH 3 and extracted with methylene chloride.   The  extract  is
    taken to  dryness,  pesticides are esterified with 10% BCls in  2-chloroethanol,
    and the resulting esters are extracted with hexane,  concentrated,  and
    determined by EC-GC.   If cleanup is required,  chromatography  on  silica
    gel deactivated with 20% water is employed.   This  procedure  is a  further
    extension of the multiclass,  multiresidue procedures described directly
    above.  When preparing the BC^-chloroethanol esterification  reagent,
    work in an efficient exhaust hood and wear disposable  vinyl  gloves  because
    2-chloroethanol is toxib by dermal contact or when inhaled.

    The reagent is stable for at least thirty days if kept stoppered and
    refrigerated.  As usual,  spiked reference material containing the same
    pesticides at comparable concentrations as in the sample (if  these are
                                      -343-

-------
                                                        Section 9A,C

 known) should be analyzed in parallel.   Other aspects of quality control
 are as discussed in the preceding paragraph.   BCl3-methanol was also
 chosen in another study (162) as the best derivatization reagent for
 the determination of 8 phenylalkanoic acid herbicides in water
 (0.01-2.5 yg/L); solvent partition and silica gel (5% water deactivated)
 minicolumn cleanup and EC-GC with an OV-17/QF-1 column were employed.

 The 1979 Analytical Methods Manual of the Inland Waters Directorate,
 Water Quality Branch,  Environment Canada, Ottawa, contains  detailed
 methods for the analysis of organochlorinated pesticides and PCBs,
 organophosphorus pesticides (two procedures),  phenoxy acid  herbicides
 (two procedures), pentachlorophenol, and N-methyl carbamates in waters.
 The method for organochlorines,  employing benzene extraction, Florisil
 column cleanup,  and EC-GC,  has detection limits ranging from 0.001-0.01  ppb.
 The first OP procedure determines dimethoate,  fenitrothion,  and phosphamidon
 and the second determines 14 other OP pesticides, all at 0.005-0.1  ppb
 levels by FPD-GC without cleanup.   Phenoxy acid herbicides  (2,4-D;  2,4,5-T;
 Silvex) are extracted  with  chloroform from acidified  water  and converted
 to  their methyl  esters utilizing BF3-methanol  prior to cleanup on a
 Florisil column  and EC-GC determination at 0.01 ppb levels.   A second
 procedure determines 8 phenoxy acid herbicides at 0.01-2.5  yg/L levels
 by  extraction of acidified  water with ethyl acetate,  back extraction of
 the polar herbicides into KHCOg,  further concentration of acids by  methy-
 lene chloride extraction to a final volume of  1 ml, esterification  with
 BCl3/2-chloroethanol reagent,  and EC-GC of the resultant 2-chloroethyl
 esters.   A separate procedure for MCPA  (4-chloro-2-methylphenoxyacetic
 acid)  and MCPB [4-(4-chloro-2-methylphenoxy) butyric  acid]  in natural
 water at 0.1-0.2 yg/L  levels is  based on extraction from an  acidified
 sample with methylene  chloride,  derivatization to pentafluorobenzyl esters,
 cleanup and fractionation on a silica gel column,  and EC-GC  determination.
 PCP is detected  at 0.01 yg/L by  benzene extraction from acidified water,
 partition into potassium carbonate solution, acetylation with, acetic
 anhydride,  partition into hexane,  and EC-GC.   Five N-methyl  carbamates are
 determined at 0.10-1.0 yg/L. levels by extraction from acidified water with
 methylene chloride,  partition  with base to remove phenols and acids present
 in  the extract,  hydrolysis  with methanolic KOH to  the respective  phenols,
 extraction of the phenols with methylene chloride  and derivatization with
 penafluorobenzyl bromide, cleanup  and fractionation of the ether  derivatives
 on  a silica gel  microcolumn, and EC-GC  of the  column  eluates.

 Cleanup is  often not required  for  EC-GC analysis of surface water samples
 (163)  and is  usually not required  for any type of water  if a  selective GC
 detector is employed.  For  example,  the multiresidue analysis of 14 OP
 pesticides  in natural waters has been carried  out at ppb levels by
 extraction,  concentration,  and direct GC with  a FPD detector  in the P- and
 S-modes  (119).   Results  of  an  interlaboratory  study of the analysis of 15
water  samples  for 10 OC1 pesticides without any column cleanup have been
 reported (164).  Where needed, cleanup  and separation of common chlorinated
 and OP  insecticides extracted  from water have been successfully carried
out in silica  gel microcolumns (165,  166) and  columns of deactivated
 (5-20% H20) silica gel  (above) and  alumina  (167).
                                    -344-

-------
                                                         Section 9A,D

  Extracts of water, sediment, sludge, sewage, and soil often contain
  large amounts of elemental sulfur, which interfere in the GC analysis
  of early eluting pesticides with the EC or FPD detectors.  Chemical
  desulf urization with Raney copper powder (168) or copper ribbon (169) ,
  precipitation with metallic mercury (170) , reaction with CN  (171),
  and treatment with tetrabutylammonium sulf ite to produce an ion pair
  with sulfur as 8203" (172) have been used to remove such interference.
  See also Subsection 9A,D) .

  Polar phosphorus, urea, and carbamate pesticides are extracted from
  water with more polar solvents such as chloroform or methylene chloride.
  Extraction of acidic or basic compounds is aided by adjusting the water
  sample to a controlled pH value.  An XAD macroreticular resin can also
  be used for residue isolation and collection.  Determination by GC is
  carried out using an appropriate selective detector after extract con-
  centration and any required cleanup and /or derivatization steps.  As
  an example, carbaryl and 1-naphthol have been determined in natural
  water at 2.5-10 ppb levels  (82-102% recovery at 5 ppb) by -EC-GC after
  methylene chloride extraction, cleanup on an XAD-8 column, and derivatiza-
  tion with heptaf luorobutyric anhydride reagent  (173) .  Sixteen organo-
  phosphorus pesticides were  determined. in drinking water at ng/liter levels
  by extraction with Amberlite XAD-2  resin, elution from the resin with
  hexane-acetone  (85:15 v/v), and GC  of the concentrated effluent using a
  nitrogen-phosphorus selective detector  (174) .

  Chlorophenoxy herbicides and their  esters have been determined by adjusting
  the water  sample  to pH  2, extracting with benzene or diethyl ether,
  methylating  the acids with  diazome thane  or  BF3-methanol,  followed by
  gas  chromatography with an  electron capture or microcoulometric detector
   (175)  (see the  footnote on  page  6 concerning the hazards  of diazome thane) .
  PCP  has been determined in  marine biota  and sea- water  by  EC-GC of the
  amyl diazohydrocarbon derivative after Florisil cleanup  (0.002 ppb) and
  by HPLC of the  free phenol  without  cleanup  (2 ppb)  (176).

  TLC  determinations of  carbamate,  urea,  triazine, and uracil herbicide
  residues in water have  been reviewed (137,  177), as have  the  extraction,
  cleanup,  GC determination,  and confirmation of  chlorinated insecticides
   in water and soils  (178).
9A,D   SOIL, HOUSE DUST, AND BOTTOM SEDIMENT

   The analysis of soil and house dust for organochlorine pesticides is
   described in Section 11, A of the EPA PAM.  Homogenized samples are
   Soxhlet-extracted with acetone-hexane, extract is concentrated in a
   K-D evaporator, and cleanup carried out on successive aluminum oxide
   and Florisil columns.  Eluates are concentrated as required and deter-
   mined by EC-GC.  A similar AOAC method has been declared official final
   action for residues of aldrin, £,.p_'-DDE, jgi^'-DDT, o.,£*-DDT, £,£f-TDE,
   dieldrin, endrin, heptachlor, heptachlor epoxide, and lindane (179).
   Section 11, C of the EPA PAM references a procedure (180) for direct
   GC determination of carbamate pesticides in soils using Carbowax 20M-
   modif ied supports and the Hall electrolytic conductivity detector.
   This method is now being investigated by the EPA for possible future
   inclusion in the EPA PAM.
                                      -345-

-------
                                                       Section 9A,D

Sediment samples are partially air  dried, mixed with sodium sulfate,
and packed into a chromatographic column.  The pesticides are extracted
from the column by elution with hexane-acetone (1:1 v/v).  The extract
is washed with water to remove acetone, and the pesticides extracted
from water with 15% methylene chloride in hexane.  The extract is dried
with sodium-sulfate, concentrated to a suitable volume, and cleaned-up
on a Florisil column.  After desulfurization with copper, determination
of organochlorine pesticides is by  EC-GC.  Details of the entire pro-
cedure are presented in Section 11, B of the EPA PAM.  Air drying of the
sample required 1-3 days, depending on the soil type.  Such samples will
contain at least 50% water.  Pesticide concentrations are expressed on
a "dry" basis, requiring determination of the dry weight of sediment by
weighing a separate, air-dried sample before and after heating overnight
at 100-110°C.  Storage of soils in  light can cause formation of artifacts
of OC1 pesticides (181).  Moistening of dried soil with water (e.g.,
80ml/30Qml) may increase extraction of pesticides by solvents such as
hexane-isopropanol (3:1 v/v) (182).

Sediment samples may contain elemental sulfur that will be recovered
through the normal extraction and cleanup procedures for organochlorine
and organophosphate pesticides and  detected by the EC,-FPD (P- or S-modes>,
and conductivity detectors.  With the recommended GC columns and operating
parameters, sulfur can completely mask the chromatogram from the solvent
peak through the aldrin peak.  The  technique described in Section 11,B,VX
of the EPA PAM for desulfurization employs vigorous agitation for one
minute with bright metallic copper.  Some pesticides may be degraded by
this treatment (e.g..^ OPs, heptachlor), but these are not likely to be
found in routine sediment samples because of breakdown in the aquatic
environment.  The procedure should be carried out if the presence of sulfur
is indicated by an exploratory injection from the final extract concentrate
or if sulfur crystallizes out when the 6 and 15% ethyl ether eluates from
the Florisil column are concentrated.  During determination of atrazine
residues in soil containing-high levels of ammonium nitrate fertilizers,
the response produced by the N-thermionic detector was not constant for
standards and samples due to the presence of the fertilizer in the sample
extracts (183.).

Part 5 of the 1979 Environment Canada Analytical Methods Manual,  Inland
Water Directorate, Ottawa,. Canada, contains a method for organochlorine
pesticides and PCBs in sediment and fish.  Nineteen compounds are determined
at 0.001-0.05 mg/kg levels by extraction of previously frozen samples with
acetonitrile, partition: with petroleum ether after appropriate dilution
with water, and cleanup and separation into four fractions on a Florisil
column.  Each fraction is determined by EC-GC.   Sulfur is removed by
precipitation with copper powder or mercury.

Nine chlorinated insecticides were determined by  a modified GC procedure
(184) with recoveries of 75-99% from suspended sediment and bottom
material.  Extraction was with acetone and hexane added separately,
coextractlves (including PCBs) were Isolated by alumina and silica gel
column chromatography,  and EC-GC was used to analyze the various  column
eluates.   Some soil analyses have been carried out by EC-GC with  tut
required column cleanup (185),  but this is not common.

                                  -346-

-------
                                                        Section 9A,D


 Shake (blending)  Soxhlet,  and column extraction methods were compared
 for efficiency in removing some twenty chlorinated insecticides from a
 sandy loam soil.   There was no statistical difference aiaong the three
 methods for the majority of pesticides, but shake extraction was sig-
 nificantly more efficient for BHC isomers (186).  The shake extraction
 method with hexane-acetone after moistening the soil with 0.2 M
 was studied collaboratively using standard AOAC analytical methods
 (Jlorisil cleanup and EC-GC) (187) and found to give excellent recoveries
 for six insecticides in three different soils (188),

 Soil residues of  chlorfenvinphos, chlormephos, disulfoton, phorate, and
 pirimiphos-ethyl  were determined by GC with thermionic detection.  Ex-
 tracted compounds were cleaned-up on a carbon-cellulose column.  Re-
 coveries ranged from 95-101% (189).  Another group of OP pesticides was
 determined in soil by GC with the thermionic detector, following ex-
 traction with acetone-hexane-benzene (1:1:1 v/v).  Florisil was used to
 clean up and fractionate the residues,  Dichlofenthion, chlorpyrifos,
 ethion, fonofos,  and leptophos were eluted with benzene-hexane (9:1 v/v)
 and parathion, diazinon, chlorfenvinphos, malathion, phosmet, azinphos
-methyl, diazoson, and paraoxon with hexane-acetone  (95:5 v/v) (190).

 A multiresidue GC procedure for the herbicides dichlobenil, dinitramine,
 triallate, and trifluralin in soils was  described by Smith  (191).  Ex-
 traction was carried out with acetonitrile-water  (9:1 v/v) in a Sonic
 Dismembrator, herbicides were partitioned into hexane, and aliquots
 injected directly into an EC chromatograph.  Recoveries were 92-107%
 from three soils at 0.05-0.5 ppm  levels.  Acetonitrile-water mixtures
 have proven to be especially efficient solvents for residues of herbicides
 of different chemical classes  (192).   Anilide herbicides- were determined
 by GC after extraction  from soil  by blending with acetone  (193).  Urea
 and  carbamate herbicides were recovered  from soils  by shaking with
 methanol  (194) or acetone  (195) and by alkaline hydrolysis  and steam
 distillation  (196).  lodinated  (196) and 2,4-dinitrophenyl  (195) deriva-
 tives were used  for EC-GC  determination  of  the  herbicides.  Triazines
 were extracted with diethyl ether from soil treated with ammonia  (197)
 and  uracils with 1.5 N HaOH (198).  Nineteen acidic, neutral, and basic
 herbicides have  been determined in soils by two dimensional TLC  (199).
 Carbofuran residues in soil were determined at  the  0.1 mg/kg  level with-
 out  cleanup by EC-GC after ammonium acetate extraction and formation of
 the  dinitrophenyl-ether derivative (200).   Uracils  have  been  recovered
 by elution with  water  from a column prepared by mixing soil with Celite
 and  Ca(OH)2J  the eluate was acidified and extracted with CHCl^,  and uracil
 determination was by RbCl thermionic-GC (201).

 The  electrolytic conductivity detector has been used to  determine nitrogen-
 containing residue* in crude soil extracts. A detector maintenance
 program for decontamination of the transfer lines and vent valve pro-
 vided reliable operation with little "down time"" even though lengthy
 extract cleanup  was not carried out (202).

 The drying and storage of soils can have an effect  on residue analysis.
 For example,  the extractable atrazine content of soil samples was  reduced
                                  -347-

-------
                                                           Sections 9A,E, 9AF

    by drying at 45°C for 24 hours.  Dried samples originally containing
    1 ppm of atrazine showed no further significant loss when stored up to
    180 days at room temperature, but there was significant loss between
    180 and 360 days.  Dried samples originally containing 10 ppm of atrazine
    showed significant loss after 90 days of storage (203).

    The analysis of pesticides of many classes in soils and plants has been
    reviewed (204).   Results of the U.S.  EPA National Soils Monitoring
    program employing Florisil cleanup of extracts prior to EC- or FPD-GC
    for OC1 and OP pesticides, and partition cleanup of extracts prior to
    GC determination of atrazine with an  N-selective thermionic detector
    have been published (205).

              POLYCHLORINATED BIPHENYLS (PCBs), OTHER COMPOUNDS

9A,E   PESTICIDE-PCB MIXTURES

    PCBs are among the most ubiquitous and persistent chlorinated pollutants
    found today in the environment.   The  residue  analyst  is concerned  not
    only with the  detection and quantitative estimation of PCBs  but with
    their effect on  the reliable determination of pesticide.residues.  PCB
    interference may occur  with most  common chlorinated pesticides in  residue
    analysis,  and  the- residue  chemist must be aware  of the nature  of this
    interference with respect  to the  GC columns being used and  their opera-
    ting parameters.   Interference  in routine analysis  is  possible with
   ja,£f-DDT, jo,jp_'-DDT, £,2>DDD, and £,.p_f-DDE, as well as with  early  eluting
    pesticides  such  as BEC  isomers, aldrin,  heptachlor, and heptachlor
    epoxide,  since prominent PCB peaks  have retention times similar to these
    pesticides  on  the recommended GC  columns.

    PCBs are  frequently detected in human  adipose  tissues,  often at concentra-
    tions similar  to  those  of  chlorinated  pesticides, and  interference with
    pesticide analysis  can  be  significant,  depending upon  the columns and
    operating parameters used.   These interferences demonstrate the non-
    specificity of the  electron  capture GC detector and the need for careful
    confirmation by use of  at  least two GC  columns, TLC, chemical reactions,
    etc.  (Section 10).


9A,F   APPEARANCE OF PCB CHROMATOGRAMS

   Whenever an analyst observes a conglomerate of chromatographic peaks upon
   injection of a biological substrate into an EC detection system, the
   possibility of the presence of PCBs should be considered.  For example,
   Figure 9-C shows a chromatogram resulting from the injection of 10 ng
   Aroclor 1254 on a 4% SE-30/6% QF-1 column operated at 200°C with a
   carrier flow of 70 ml/minute.  The first isomer peak of consequence has
   an absolute retention of about 6 minutes and the final peak about 38
   minutes.  Figure 9-D represents the chromatogram of 6 ng Aroclor 1260
   under the same conditions, and major peaks ranging from 8 minutes to
   nearly one hour are seen.  Aroclors 1254 and 1260 have shown up most
   widely in a variety of environmental and tissue samples.
                                     -348-

-------
                                                 Section 9A,F
Figure 9-C.  Aroclor 1254.  Column 4% SE-30/6% QF-1,
             200°C,  carrier flow 70 ml/min.
   1
8    12    16    20   24    28    32'   36    40
   Figure 9-D.   Aroclor 1260.  Column 4%  SE-20/6% QF-1,
                 200°C, carrier flow 70 ml/min.
               16   '24      32     40

                  Retention, minutes
                              48
                                                   56
                           -349-

-------
                                                         Section 9A,F
 The  type of confusion evident when pesticides and PCBs are present in
 the  same substrate is illustrated in Figure 9-E, showing Aroclor 1254
 co-chromatographed with a mixture of eight chlorinated insecticides.
 Aldrin  (peak 1), £,,p_'-DDE (3), £,£»-DDD (5), £,_p_'-DDT (6), Dilan I (7),
 and  methoxychlor  (7)  are seen to overlap PCB peaks so closely that
 differentiation would be impossible.  Heptachlor epoxide (2) and dieldrin
 (4)  (in large quantities)  are partially separated, while Dilan II is
 fairly  well separated.   A co-chromatogram of Aroclor 1260 with the same
 pesticide mixture  would show good separation of aldrin and Dilan II,
 partial separation of heptachlor epoxide and Dilan I or methoxychlor,
 appearance  of disproportionately large Aroclor peaks at the retention
 locations of chlorinated pesticides should alert the analyst to the
 possible presence  of  these OC1 pesticides in the PCB sample.
     Figure 9-E.
Aroclor 1254  (solid line) and pesticide mixture
(dotted line).  Column 4% SE-30/6% QF-1,  200°C,
carrier flow  70 ml/min.
                                                   1 AldVjn
                                                   2 H«p».
                                                   3 p.p'-DDE
                                                   4 Di«Wrin
                                                   5 p.p'- DOO
                                                   6 p. p" DDT
                                                   7 Dilonl 4.M»lhorychlor
                     8    12    16    20
                       Retention, minutes
                      24
28
                                 32
Confusing chromatograms also result when PCBs are mixed with the multi-
peak pesticides  chlordane or toxaphene.   Figure 9-F shows the co-chromato-
gram of chlordane  and  Aroclor 1254.  The only clean separation is the
first peak of  the  earliest major pair of chlordane peaks, while partial
separation is  obtained for the second peak of the third pair. -The early
minor chlordane  peaks  are well separated but are of little value for
quantitation of  chlordane.   Aroclor 1260 does not interfere as seriously
with chlordane under these same chromatographic parameters since the
first PCB peak does not elute until after first two major chlordane
peaks.
                                  -350-

-------
                                                       Section 9A,F


 Figure 9-F.   Aroclor 1254 (solid line) and chlordane  (dotted Line)
              Column 4% SE-30/6% QF-1, 200°C, carrier  flow 70 ml/min.
                4Retenlion , minutes
Figure 9-G shows a mixture of Aroclor  1254 with toxaphene.   Analyses of
toxaphene, chlordane, and PCBs are  further confused because the chromato-
grams of environmental samples never exactly resemble those of standards,
Chlordane is not very widespread in environmental samples,  so its mutual
analysis with PCBs is less likely to be a problem.
  Figure 9-G.
Aroclor 1254  (solid line) and toxaphene  (dotted'line)
Column 4% SE-30/6% QF-1, 200°C, carrier  flow 70 ml/min.
                            12         *>
                       Retention, minutes
                                28
                                                          34)   40
                                   -351-

-------
                                                          Section 9A,G


   The actual effect of PCBs on quantisation of chlorinated pesticides is
   highly dependent on the levels involved, the pesticide of interest, and
   the attenuation Tfeing used.  For example, if the ratio of PCBs to pesti-
   cides is 10 ppm to 3 ppm, an attenuation can be used that will give an
   adequate peak for DDE while DDT (for example) and PCBs will hardly be
   seen.  If quantitation of DDT is required, however, a lower attenuation
   will be required (because of its lower response) to give an adequate
   peak size, the DDE peak will be off-scale, and PCB peaks will be more
   noticeable.  At a ratio of 25 ppm PCB to 3 ppm pesticide, quantitation
   of DDT will definitely be affected, and with 100 ppm PCB to 3 ppm pesti-
   cide and attenuation to keep DDE on scale, determination of the latter
   would be affected.


9A,G   METHODS FOR SEPARATION AND ANALYSIS OF PESTICIDES AND PCBs

       a.  Published Procedures and Data

           The EPA PAM contains macro and micro methods for determining
   PCBs in human milk in Sections 9,B,(1) and 9,B,(2), respectively.  In
   the macro method, the milk sample (4-24 grams) is extracted with acetone
   and hexane, PCBs are transferred to the hexane layer by adding sodium
   sulfate solution, and the hexane is dried by passage through a sodium
   sulfate column.  Part of the sample is used for a lipid determination,
   and the rest is partitioned with acetonitrile and then fractionated on
   an activated Florisil column 10 cm in height.  Identification and
   quantitation of PCBs is carried out by EC-GC and confirmation by use
   of different GC columns, and the electrolytic conductivity detector
   (Cl-mode), chemical derivatization by perchlorination, and GC-MS of
   pooled samples.

   In the micro method, a 0.5 gram sample of milk is extracted with ace- '
   tonitrile, residues are partitioned into hexane, the hexane is concentra-
   ted, and the PCBs are eluted through a 1 gram deactivated (3% water)
   Florisil column.  The eluted PCBs are further separated from chlorinated
   pesticides on a micro silicic acid column.  Chemical derivatization by
   perchlorination to yield decachlorobiphenyl (DCB) followed by EC-GC
   is used to confirm PCBs.  Neither the macro nor micro methods are
   capable of accurately identifying or quantitating absolute levels of
   PCBs, but they provide semi-quantitative results.

   Filter paper, glass wool, and sodium sulfate are likely sources of PCB
   contamination in the macro method, and these materials must be thoroughly
   precleaned with pesticide grade solvents as described in Section 3K.
   Each sample analyzed requires a total volume of ca 2000 ml of solvent,
   and care must be taken in concentration of this large volume to the
   final 1-5 ml for analysis.  One blank and one fortified goat's milk sample
   should be run with every set of 10 human milk samples for both the macro
   and micro methods.  Details for preparing these samples are described
   in Section 9,B,(1), XIV and 9,B,(2), X of the EPA PAM.  The amount of
   Florisil needed for a proper elution pattern should be determined for
                                     -352-

-------
                                                       Section 9A,G
each different lot by elution of analytical standards.  Proper separation
of PCBs and .pesticides on the silicic acid column should be checked by
chromatographing standard compounds and analyzing-both eluate fractions.
The Aroclor standard providing a chromatogram most closely resembling
that of the sample should be used for quantitation of that sample.

Analysts inexperienced with the method should be guided through the pro-
cedures at least four times by a person experienced with the procedure,
using duplicate samples already analyzed by the experienced worker.
Then the analyst should be required to demonstrate proficiency on an
additional set of four spiked samples without aid before handling actual
samples.   ,

The EPA Manual also describes the separation .of PCBs from DDT and its
analogs by the method of Armour and Burke  (206) (Section 9,C), and a
thin layer method for semiquantitative estimation of PCBs in adipose
tissue  (Section 9,D).  Section 9,E illustrates chromatograms of different
Aroclors on 4% SE-30/6%- OV-210 or QF-1 and 1.5% OV-17/1.95% QF-1 GC
columns, and  Section 9,F tabulates relative retention values and re-
sponse values of six Aroclors on OV-17/QF-1, SE-30/QF-1, and OV-210
columns.  Retention indices have been calculated  for-all 210 possible
individual PCBs on 13 GC phases, and recommendations were made for the
best phase  combinations for separations  (OV-210,  Apiezon L, and OV-225
were among  the best single columns; OV-3 + CHDMS  and OV-3 or OV-25 + poly
MPE were ,the  most discriminating pair)  (207).  HPLC and capillary  column
GC have also  been used  to  separate PCB mixtures  (208,  209).

Crist  and Moseman  (210) of the  EPA reported a  simplified micro perchlorina-
tion method for  determination of PCBs in biological samples.  A sample
was  cleaned-up by  the modified  MOG procedure  (Section 7Aa), and the
PCBs were perchlorinated with SbCl5  to  decachlorobiphenyl  (DCB), which
was  cleaned-up by hexane  partitioning and chromatography on a  1.6  gram
 column of activated 4j.orisil.   Details  are in  Sectio^ 9,B, (2),IX  of  the
 EPA PAM.   The presence of impurities in SbCls  reagent that  can cause
 erratic recoveries of PCBs was  noted by Trotter and Young  (211),  and
 DCB impurity was detected in various brands of the reagents used  in the
 Crist and Moseman procedure (210).

         b.  PCB Cleanup and Separation Systems

             Depending upon the particular pesticides and PCBs present,
 the amounts of each, and the purpose of the analysis, it may or may not
 be necessary to separate PCBs and pesticides present in the same extract
 before the determinative step.  Some combinations may permit quantita-
 tion of each without prior separation, others will require a separation
 before determination, and still others may require a separation pro-
 cedure that  destroys or converts some, of  the compounds to permit quantita-
 tion of those remaining unchanged.

 PCBs are eluted with 6% ethyl ether-petroleum ether in the modified MOG
 procedure described in the EPA PAM, Section 5,A,(1),(a) and in the FDA
                                   -353-

-------
                                                        Section 9A,G

 PAM multiresidue procedures, Sections 211 and 212.  They elute with
 eluant 1 of the alternative methylene chloride elution system (Section
 9A,B of this Manual and Section 252 of the FDA PAM).  A study by Lieb
 and Bills (212) found that the storage temperature of Florisil aftdr
 initial activation (overnight, 130°C) influenced the GC pattern obtained
 for Aroclor 1254 separated from lipids on a column of the Florisil.  To
 avoid selective adsorption of some PCB components and erroneous PCB
 analyses, storage of activated Florisil at room temperature was
 recommended.  This, however, is in opposition to the procedure
 recommended for routine pesticide work (continuous storage at 130°C
 until use) and should be studied further.  Hydroxy PCB metabolites
 extracted from cow's milk were cleaned up by extraction with aqueous
 alkali and re-extraction of the acidified aqueous solution with organic
 solvent prior to further TLC cleanup and GC-MS determination (213).

 The method of Armour and Burke (206) has been mo$t used for pesticide-
 PCB separation.   The 6% ethyl ether-petroleum ether Florisil column
 eluate or eluate 1 of the alternative procedure (Section 9A,b of this
 Manual) is concentrated to an appropriate volume and a 5 ml or smaller
 aliquot applied to a column of partially deactivated silicic acid and
 Celite, standardized before use to effect the best possible separation
 between _p_,_p_f-DDE and Aroclor 1254.  Petroleum ether followed by ace-
 tonitrile-hexane-methylene chloride (1:19:80 v/v)  are used to elute the
 column, both fractions being collected in a K-D evaporation flask.   The
 eluates -are  concentrated and subjected to EC-GC.   Mixed results  have been
 reported  with this silicic acid separation system.  PCBs and polychlorinated
 terphenyls split between the two fractions (EPA PAM,  Section 9,C,  Table 1)
 as  do  the pesticides  aldrin .and 2.,2.'-DDE (Canadian PAM,  Section 7.5).
 Polychlorinated  naphthalenes (214) and dioxlns elute in the first  fraction
 and most  other chlorinated pesticides (e.g.,  chlordane,  toxaphene,  DDT,
 heptachlor,  lindane)  in the second.   Because of the division of  some
 compounds between the two  eluates, GC columns must be carefully  chosen
 to  separate  the  components present in each fraction.   The  tables of
 relative  peak heights and  peak  retentions in the EPA PAM can help  in this
 selection.   The  chemist running this procedure for the first time should
 perform a sufficient  number of  recovery  trials with spiked  samples  to
 gain confidence  in its  reliability.   Impurities present  in  silicic  acid
 adsorbent batches,  their effect on separations, and means for their
 removal have been described (215).   Pesticide-PCB  separations were
 found reproducible only for individual batches of  adsorbent.  Porter
 and Burke have reported the separations of TCDD from PCBs on acidic,
 basic,  or neutral aluminum oxide;  PCBs were eluted with hexane-methylene
 chloride  (99:1 v/v) and di,  tri-,  and tetrachlorodibenzo-p-dioxins with
 hexane-methylene  chloride  (80:20 v/v)  (216).

A slightly modified version of  the Armour-Burke method is detailed  in
 the Canadian PAM Section 7.5, and  the method has been miniaturized  for
 determining  chlorinated  pesticide  and PCB  residues  in fish.  In the latter
method, the  sample  is dried with Na2SO^ and packed  into a column, which
 is eluted with petroleum ether.  Cleanup and separation of the extract
 is on 1 cm (id) Florisil and silica  gel columns (217).
                                  -354-

-------
                                                       Section 9A,G
Other colunsns used in various multiresidue cleanup procedures provide
at least partial separations of organochlorine pesticides from PCBs.
These include columns of activated alumina (67; Section 90); silica
plus alumina (64; Section 90); 60 A° silica gel eluted with pentane
(elutes PCBs and mirex) and benzene (recovers DDE, DDT, TDE, BHC,
dieldrin) (218); and charcoal (219, 220).  Elution of a charcoal column
with diethyl ether-acetone (3:1 v/v) removes OC1 pesticides while PCBs
are then eluted with benzene  (59, 221).  A Norit C-170 charcoal-poly-
urethane foam (40:60 w/w) mixture is especially useful for separation
of mirex and photomirex from PCBs (222).

Section 251.2 of the FDA PAM describes derivatization.and micro-column
chromatography procedures -for removal of DDT and related compounds
from extracts containing PCBs.  Cleaned-up extracts are treated with
alkali to convert DDT to DDE and TDE to its olefin; PCBs are unchanged.
Subsequent oxidation of the solution with chromium trioxide in acetic
acid converts DDE and TDE olefin to dichlorobenzophenone, but again
PCBs remain intact.  PCBs are then separated from polar dichloro-
benzophenone by elution with petroleum ether from a micro activated
Florisil column.  Dichlorobenzophenone is eluted, if required, with ethyl
ether-petroleum ether  (1:1 v/v).  Recoveries of Aroclors 1242, 1254, and
1260 ranged from 77-100%  (0.4-56 ug amounts), while DDT, DDE, and TDE
were recovered  (as dichlorobenzophenone) between 5-86%  (2-33 ug).  The
same reactions used in this GC determinative procedure have been applied
to TLC estimation  (Subsection d.) and confirmation  (Subsection e.) -of
PCBs.

Other techniques for separating PCBs from DDT and its analogs by chemical
derivatization and column chromatography include:  dehydrochlorination
with l,5-diazobicyclo(5.4.0)undec-5-ene reagent  (223);  sodium dichromate
in acetic acid plus suifuric  acid for conversion of DDE to  dichloro-
benzophenone without affecting DDT, TDE, or PCBs  (224) ; oxidation by
chromic  acid glacial acetic acid reagent followed by silica gel column
chromatography  (225);  a silica gel  tube with MgO catalyst for conversion
of DDT to DDE and TDE  to DMU  without effect on PCBs  (226) ;  heating
cleaned-up  fish or serum extracts with KOH/ethanol  to convert OC1 pesti-
cides to alkenes,  oxidation with Cr203 to more polar compounds, and
separation  from PCBs on a Florisil  column  (227) ; and reaction of fish
tissue extract with  fuming nitric acid  followed by  separation of nitrated
PCBs  from mirex analogs by chromatography on a micro Florisil column  (228) .

Aroclor  1254 residues  in blood have been determined by  extraction with
hexane-saturated acetonitrile and  cleanup on an alumina column.  Eluates
were  analyzed by EC-GC on an OV-1  column (229).  PCT, PCB,  and DDT
 residues in blood (5-11 ppb)  were  determined by heating with ethanol
 and KOH to  dehydro chlorinate DDT,  extracting with hexane, washing with
H2S04,  and chromatographing on  a mixed silica + Florisil + Na2S04 column.
 The hexane eluate was  concentrated and analyzed by  EC-GC  on a 2% OV-1
 column,  and confirmation was by MS (230).
                                   -355-

-------
                                                       Section 9A,G
        c.  GC Quantisation of PCBs

            One of the most difficult aspects of PCB quantitation is to
obtain a match between the sample and a standard.-  Because individual
congeners in the original source are likely to be distributed differ-
ently due to varying volatilities, solubilities, and reactivities,
biological and environmental samples seldom have a GC peak pattern
that will exactly match any Aroclor standard, and even commercial
preparations of PCBs vary in abundance of minor components from batch
to batch.  In addition, detector response of different PCB isomers can
vary by as much as 10,000 fold, so that any similarity between samples
and a known commercial PCB mixture is likely to be purely fortuitous.
For example, the upper chromatogram in Figure 9-H is that of a standard
PCB mixture, Aroclor 1242.  Below it is the same mixture added to an
ambient air sample at a level equivalent to 100 ng/m^.  The PCB mixture
   Figure 9-H.
Gas chromatograms of Aroclor 1242.  (A) standard
fortification solution diluted 10 times to .200 pg/yl;
(B) residue in upper foam trap after 24,hours at
225 L/minute.  Numbers indicate peaks used for quantita-
tion.
                                     -356-

-------
                                                       Section 9A,G
was added to the air stream as vapor while the sample was being
collected, resulting in alteration of the relative peak ratios because
the various components were trapped with efficiencies ranging from
40-95% and also because of contributions from background materials.
It would be difficult to identify the added PCBs in this sample from
GC data alone.  Figure 9-1 shows an even more difficult, but no less
typical, case.  Chromatogram A is Aroclor 1016, while B is what was
isolated from the brain of a 'rat that had been fed this PCB mixture
for one year (231).  The problem of accuracy in PCB analysis is 'not
easily solvable.  Because of the complexity'of commercial mixtures,
identification of individual components is not practical.  Complete
separation by GC is impossible, even with capillary columns.  The mass
spectrometer cannot usually distinguish between all PCB isomers.

The most widely used practical approach for PCB quantitation is to
compare the total area or height of detector response for the residue
peaks to the total area or height of response obtained under the same
conditions for a known weight of the commercial Aroclor standard with
the most similar pattern.  Only those peaks from the sample that can
be attributed to chlorobiphenyls are used, and these peaks must also
be present in the chromatogram of reference materials.  If the presence
of more than one Aroclor is clearly indicated, the residue may be
quantitated using mixtures of Aroclor standards judged appropriate
for different portions of the sample chromatogram.  In one interlaboratory
study of PCB analysis using Aroclor -1254 as reference standard, quantita-
tion via the three specific peaks with retention times, relative to DDE,
of 127, 147, and 177 produced the best interlaboratory agreement (232).
        Figure 9-1.

 Electron  capture  gas  chromato-
 grams  of  (a) Aroclor  1016
 standard  and  (B)  PCB  residue
 extracted from brain  of  rat
 fed on diet containing
 Aroclor 1016  for  one  year.

                                         2     4     «     S

                                           TIME, minutes
                                    -357-

-------
                                                        Section 9A,G

 The procedure of Webb and McCall (233)  has an advantage in that residues
 can be quantitatively measured on a GC  peak-to-peak basis  against  a
 series of reference Aroclors  with known weight percentage  compositions
 for individual peaks.   Reference Aroclors  1016, 1242,  1248,  1254,  and
 1260 have been characterized  using a Hall  electrolytic conductivity
 detector for Cl measurement and chemical ionization MS with single
 ion monitoring for molecular  weights (234).   A ten laboratory collabora-
 tive study of PCB quantitation in synthetic standard mixtures,  milk, and
 chicken fat samples indicated that greater overall precision was possible
 using  the individual peak method compared  to total area or height  (235).
 The Individual peak method has been made official  first action by  AOAC
 as  an  alternative to the  total area quantitation procedure (56).

 All PCB components may be converted by  perchlorination with SbCl5  to
 a single derivative (decachlorobiphenyl),  and the  total PCB content may.
 then be'measured as this  compound (210,  236; Subsection 9A,Ea).  Quantita-
 tive data are not truly valid with this  approach unless the average
 chlorine content of the original PCBs in the sample before chlorination
 is  known.   A related approach to quantitation is dehydrodechlorination
 of  PCBs to biphenyl by lithium aluminum hydride in dodecane,  followed
 by  HPLC with a UV detector at 248 nm.   The absolute detection limit was
 100 ng,  and DDT isomers,  chloronaphthalenes, and PCTs  were determined
 simultaneously (237).

 Other  quantitation approaches that have  been attempted include  estimation
 of  the weight of PCB injected by dividing  the retention time  x peak
 height for all PCB peaks  by the product  of peak height and retention time
 for 1  ng jp^jfr'-DDE on the  same GC column  (238),  and peak resolution and
 matching by a computer (239).   GC-MS with  individual mass  monitoring using
 a minicomputer-controlled spectrometer has been reported (240).    This
 method provided sensitive qualitative and  quantitative analysis of sediment
 extract without the need  for  elaborate column adsorption separations prior
 to  GC.   Beroza and Bowman's ;p_-values have  been applied to  quantitation of
£,p_f-DDT in the presence  of non-resolved PCB peaks,  and results within
 11% of actual were reported (241).   The  USFDA approach to  chemical pro-
 filing of  PCB content  in  a sample to select  the most suitable quantitation
 standard has been discussed (242).

           d.  PCB Estimation by TLC

              The semiquantitative TLC procedure (243)  for  determination of
 PCBs in adipose tissue utilizes the 6% eluate of the Florisil cleanup.
 The concentrate is treated with KOH to dehydrochlorinate DDT and DDD
 to  their olefins,  thereby eliminating the  problem  of separating the pesti-
 cides  from the PCBs.   Any interfering DDE  is  then  oxidized to jj,j3'-dichloro-
 benzophenone,  which has an % value different  from the PCBs on a silver
 nitrate-Impregnated alumina layer developed with 5% benzene In hexane.
 The PCBs give one spot for all  formulations,  and this is quantitated
 against  a  standard Aroclor 1254 or 1260  spot run on  the same plate.  The
 best standard can be chosen after examining a preliminary GC trace.  The
 final values obtained  by  this method should be  considered as approximations,
                                 -358-

-------
                                                          Section  9A,H


  with  a precision of  roughly  ±50% indicated by recovery studies.   The
  minimum level of detection is  ca 1 ppm by exposure  of the layer to
  UV light.  An EPA human monitoring program for PCB  levels in adipose
  tissue has utilized  this  TLC procedure with  confirmatidn by combined
  GC-MS (244).

                e.   Confirmation  of PCBs
                    Confirmation  of PCBs has been obtained by perchlorina-
  tion  (210, 236), alkali treatment (245),  and reaction with chromic  acid
   (chromium trioxide).   The stability of PCBs  to alkali is useful for
  confirming the identity of PCB residues,  and at the same time,  con-
  version of DDT to DDE by  the alkali removes  some interference to
  quantitation of PCBs.  Treatment with alkali also provides additional
  cleanup for  many sample types.  Resistance to oxidation with chromic
  acid-acetic  acid reagent  is  also useful evidence for identifying PCBs
  in the presence of reactive  pesticides such  as DDE  and DDT (243) and
  chlorinated  naphthalenes  (246).  However, it has been reported (247)
  that  alteration of PCB chromatographic patterns can occur upon chromium
  trioxide  acid digestion of animal tissue extracts,  including changes
  in peak areas and disappearance of several PCB homologs.

  Two-dimensional (248) or  multi-development reversed phase (249) TLC
  systems,  which separate PCBs from DDE, DDT,  and other pesticides, can
  aid identification.   PCBs are  destroyed by DV irradiation, but many
  pesticides may be altered as well (250).   Toxaphene survives UV treat-
  ment  that destroys PCBs and  can be confirmed in- mixtures in this way.
  Mirex, a  late eluting pesticide that usually is not interfered with by
  PCBs, also withstands TJV  irradiation and can thus be confirmed.  Irradia-
   tion with controlled UV wavelengths has provided identification and
   determination of aldrin,  dieldrin, heptachlor, and  heptachlor epoxide
   in mixture with PCBs.  Photoisomerization reactions of the pesticides,
  producing products with longer retention times, were induced with wave-
   lengths  >290 nm; subsequent  irradiation with wavelengths >230 nm yielded
  photodechlorinated products  of PCBs with shorter retention times (251).
  Most organochlorine  pesticides are destroyed by reaction with HN03-H2S04
  whereas  PCBs and toxaphene are unaffected.  Chlorinated pesticides were
   selectively  detected in the presence of PCBs by use of a modified
   Coulson conductivity detector at 600°C with a hydrogen flow of 1-2 ml/min.
   (252).  Mirex and PCBs have been differentiated based on the low sensi-
   tivity of the Hall detector for the latter (253).  A collection of spectra
   helpful in confirming isolated residues of PCBs has been published (254).

9A,H   DETEBMINATION OF POLYBROMINATED BIPHENTfLS

   Polybrominated biphenyls   (PBBs) were manufactured for use as a fire
   retardant from 1970 to 1973.  Since the summer of 1973, when PBBs were
   accidentally mixed with dairy feed resulting in the contamination of
   livestock and food products, the sensitive determination of low levels
   of PBB residues has  been of interest to the PDA and the EPA.  One
                                     -359-

-------
                                                                     Section 9A,I

              commercial PBB fire retardant (Firemaster BP-6) has been chemically and
              toxicologically evaluated; 13 different polybromobiphenyls were found
              plus a brominated naphthalene contaminant, and biological effects were
              ascribed to PBBs (255).

              A rapid method has been developed for analysis of plasma, feces, milk,
              and bile using all disposable glassware to reduce the amount of labora-
              tory background and cross-contamination of samples (256).  The authors
              found that this type of equipment was necessary because methods that had
              proven to be effective for decontamination of PCBs were not effective
              for PBBs.  The methodology involved multiple extraction of ethanol-
              denatured sample (except for' feces) with petroleum ether-diethyl ether
              (1:1 v/v) in a disposable test tube, followed by cleanup on a miniature
              adsorption column packed in a 23 -cm disposable Pasteur pipet.   The
              column contained Florisil, silica gel, and sodium sulfate in different
              proportions, depending on the sample.   The column was eluted with 5 or
              10 ml of petroleum ether-benzene (98:2 v/v)  into a disposable screw
              top vial.  Determination of PBBs in the concentrated eluate was made by
              EC-GC on a silanized 5Z OV-17 column.   Mean recoveries for the six major
              components of a commercial PBB mixture were approximately 96% for plasma,
              59Z for feces,  98Z for milk,  and 89% for bile at 0.05-50.0 ppm levels.
              The maximum background level was 0.0007 ppm for the major hexabromobiphenyl
              peak,  corresponding to a minimum detectable limit of ca 0.001  ppm.

              The separation and characterization of PBBs  by chromatography  and
              spectroscopy have been studied (257).   Columns containing 1% SE-30  or
              2% OV-17 were used for GC-FID-MS,  5 pm silica gel 60 (Merck) columns
              for HPLC (UV detection),  and  paraffin-coated kieselguhr for reversed
              phase  TLC.   In addition to quadrupole  MS, NMR and UV spectroscopy were
              evaluated.   Capillary GC has  been used to separate PBBs,  as well  as
              chlorinated dibenzofurans and anisoles (258).
                
-------
                                                          Section 9A,J
   toluene (3:1 v/v).   GC-MS analyses to confirm the polychlorinated ter-
   phenyl residues were made using a 152 cm x 2 mm id glass column packed
   with 32 OV-1 on Gas Chrom Q.  The column oven was programmed from
   150°C (1 minute) to 300°C at 4°C/minute.  Mass spectra were acquired
   over the range 420 to 720 amu at 6 seconds/scan for PCT confirmation
   (261).

9A,J   SEPARATION AND DETERMINATION OF DIOXINS

   Section 9,6 of the EPA PAM contains sample preparation and capillary
   column GC-MS techniques developed and currently applied by EPA labora-
   tories for isolation, detection, quantitation, and confirmation of
   2,3,7,8-tetrachlorodibenzo-£--dioxin (TCDD) residues "(262).  Tissues,
   milk, water, soil, and sediment samples are subjected to an "acid-base"
   sample preparation Involving saponificatiori with hot caustic solution,
   followed by extraction with hexane, washing with concentrated sulfuric
   acid, cleanup by alumina column chromatography, and capillary column
   GC-high resolution mass spectrometric multiple ion selection analysis
   for TCDD residues.  Fish tissue is subjected to a "neutral" cleanup
   procedure in which extraction is carried out with acetonitrile, and
   cleanup by solvent partitioning and Florisil column chromatography
   precedes alumina column chromatography.  ^'Ci-TCSD is added to all
   samples as an internal standard or marker to monitor and determine the
   cleanup efficiency.  Sensitivity of the procedure is in the 0.02-100 ppt
   concentration range.  Extreme care and very clean laboratory practices
   are mandatory for low ppt analyses.

   The efficiency, accuracy, precision, and validity of ppt TCDD analyses
   depend on an incorporated quality assurance program that is described
   as part of the  procedure in Section 9,G of the EPA PAM.  It is important
   that  TCDD analyses be conducted only by trained personnel with strict
   safety procedures in effect.  The hazards and analysis of TCDD have
   been  reviewed  (263).

   Reports from laboratories that have conducted environmental monitoring
   projects for TCDD and have  developed and applied analytical cleanup
   systems and mass spectrometric methods  of analysis for ppt levels of
   TCDD  residues  in environmental, biological, human, and agricultural
   samples and  chemical formulations are contained in references  (264-275).
   It has been  shown  (276)  that  analysis of environmental samples by low
   resolution GC-low resolution  MS  alone is acceptable if suitable  control
   samples are  available  to show the absence of  interferences.  When
   suitable controls  are  not available  and when  cleanup  is nonspecific,
   positive results  for TCDD must be confirmed by high resolution MS,
   preferably  using mass  fragmentography with  single or  multiple ion
   detection and/or  chemical ionization (277).   A recent HPLC method  (278)
   shows promise  of  being specific  for  TCDD and  eliminating  the need for
   high resolutions MS confirmation.
                                      -361-

-------
                                                          Sections 9A,K, 9AL
   Chlorinated benzyl phenyl ethers have been identified as a possible
   serious interference in the GC-MS determination of chlorinated dibenzo-p_-
   dloxins.  These compounds, which have been extracted from 2,4,5-trichlo-
   phenol, have retention times and MS responses similar to TCDDs (279).

9A,K   DETERMINATION OF ETHYLENETHIOUREA (ETU)

   Because of its toxicological significance, constant occurrence as a
   terminal residue following crop treatments with ethylenebisdithiocarba-
   mate fungicides, and its actual presence in technical ethylenebis-
   dithiocarbamates, analytical methods for determination of ETU have
   become extremely important.  A method for ETU "in apples (280) was based
   on reaction with benzyl chloride to give 2-benzylmercaptoimidazoline,
   which is subsequently treated with trifluoroacetic•anhydride to yield
   2-benzylmercapto-J$r-trifluoroacetylimidazoline.  This derivative is
   measured by EC-GO.

   ETU residues were measured in various crops by methanol extraction,
   alumina column cleanup, and derlvatization with 1-bromobutane in the
   presence of DMF and sodium borohydride.  The resulting 2-butyl-
   mercaptoimidazoline was measured down to 0.01 mg/kg with an FPD detector
   (281).  A similar method that determines ETU in milk, fruits, and
   vegetables as the same derivative has been collaboratively studied (282)
   and recommended as an AOAC official first action method (283).

   EC-GO as well as S-mode FPD-GC have been used to determine ETU residues
   from crops after derivatization with nt-trifluoromethylbenzyl chloride
   (284).  The trifluoroacetylated S-benzyl derivative has also been used
   to determine ETU residues on tomatoes (285) .>  ETU residues on fruit and -
   vegetable crops were determined at 0.01-0.1 ppm levels without derivatiza-
   tion (286).  After methanol extraction and cleanup by hexane/aqueous
   NH4C1 partition and alumina column chromatography, GC was performed on a
   3Z Versamid 900 column with S-mode FPD detection.  Recoveries ranged from
   62-95%.

   The occurrence, chemistry, and metabolism of ETU and analytical methods
   for its determination have been reviewed (287).

9A,L   DETERMINATION OF CONJUGATED PESTICIDE RESIDUES

   Pesticides and pesticide metabolites are known to form carbohydrate
   (glycoside, glucuronide), amino acid, sulfate, alkyl, and acyl conjugates
   in various plant, animal, and soil systems.  Because of the potential
   biological activity of many of these conjugates, their identification
   and determination has become an important task for the pesticide analyst.

   Because conjugates are, in general, more polar and less lipophilic than
   the parent pesticides, analytical methods are designed to take into
   account these differences.  In addition, the lability of certain conjugates
                                     -362-

-------
                                                       Section 9A,L


may dictate the analytical approach taken when isolating and identifying
the intact compound or a derivative, e.g., the need for protection of
labile moieties from hydrolysis during extraction or the choice of column
LC rather than GC for separation of thermally unstable or nonvolatile
conjugates.  Analysis of enzymatic or chemical hydrolysis products is
useful confirmatory information for conjugates that, have been identified
intact or may serve for the quantitation of a conjugate residue.

Different types of mass spectrometry (Section 10), including electron
impact, chemical ionization, field desorption, and laser ionization,
are probably the most powerful and widely used tools for structural
analysis of conjugates.  The field desorption method is especially
useful due to its applicability to polar materials.  NMR, particularly
using proton nuclei with the sensitive Fourier transform technique,
is another important aid for structure elucidation (Section 10K), as
are traditional 1R and UV absorption spectrometry and micro-XR (Section
10J).

Specific isolation methods depend on the exact nature of the conjugate
of interest and the sample matrix.  Most conjugates are extractable with
water, alcohol, and water-alcohol mixtures from insects, plants, or
tissues.  Samples may be freeze-drled and pre-extracted with an organic
solvent to remove lipophilic materials.  Purification, separation, and
concentration of conjugates have been carried out using simple solvent
partitioning, counter-current liquid-liquid distribution, extraction
with liquid anlon-exchangers, Amberlite XAD-2 polymer columns, silicic
acid columns, Porapak Q resin columns, Sephadex LH-20 gel columns,
DEAE-cellulose and DEAE-Sephadex anion-exchange columns, Sephadex 6
gel columns, Biogel F columns, cation-exchange resin amino-acid analyzer
columns, liquid anion-exchange paper chromatography, TLC, and GC of
conjugates either directly or after forming a volatile derivative.

Most analytical work on pesticide conjugates to date has been conducted
for structural identifications or metabolism studies.  The usual radio-
tracer detection techniques are widely used in metabolism research.  A
review of analytical methods for different conjugate types, including
many literature references, and examples of applications to different
research problems will be found in reference (288).  This volume also
contains Information on the nature and analysis of "bound" or unextract-
able pesticide residues.  One approach to the analysis of bound residues
was reported for chloroaniline bound to lignin fractions of plants based
on release by pyrolysis (289); pyrolysates containing intact chloro-
anilines were collected and derivatized as trifluoroacetanllides, which
were purified and determined by EC-GC.

Relatively little attention has been given to the recovery of pesticide
conjugates by analytical procedures designed to determine the parent
residues.  When the problem is addressed, the usual approach is that
which was taken to determine PCP residues in urine (24).  The analytical
procedure for intact PCP residues was modified to include an acid
hydrolysis, the purpose of which was to free the conjugated forms of
the pesticide and allow its derivatization along with the unchanged
parent compound (see Section 9E).
                                  -363-

-------
                                                          Sections 9A,M, 9A,N


   A few similar procedures for other pesticides have been published, the
   hydrolysis step in some cases serving both to break the conjugate and
   to hydrolyze the parent pesticide to a new form prior to determination
   (e.g., hydrolysis of carbamate insecticides to the corresponding phenol,
   which is derivatized, cleaned-up, and determined by GC).  3-Hydroxy-
   carbofuran, the major carbofuran metabolite produced in animals, is
   present as the water soluble glucuronide conjugate.  A mild acid
   hydrolysis was used to free the conjugated form of the metabolite and
   allow its extraction with organic solvent along with the parent com-
   pound (290).

   Conjugates of 2,4,5-T in biological samples have been broken and the
   free acids released by a basic hydrolysis step (291).  Residues of
   conjugated iodofenphos phenol metabolites were recovered from liver and
   kidney tissue by. extraction with ethanol-water-lN sodium hydroxide
   (90:10:1 v/v), hydrolysis with IN sulfuric acid, and hexane + ethyl
   ether extraction (292).


9A,M   REVIEWS OF ANALYTICAL METHODS FOR PESTICIDES, PCBs, AND OTHER
       NON-PESTICIDE POLLUTANTS

   See Subsection 1G in Chapter 1 for a general bibliography of important
   books and reviews on the analysis of pesticides and related pollutants.,


9A,N   REFERENCES

   (1)  McMahon, B. M., and Burke, J. A., J. Assoc. Off. Anal. Cheat., 61,
        640 (1978).                       ~~                   ~~~"  "~

   (2)  Sonobe, H., Carver, R. A., and Kamps, L. R., J. Agr. Food Chem..
        28(2), 265 (1980).

   (3)  Suzuki, T., Ishikawa, K., Sato, N., and Sakai, K. I., J. Assoc. Off.
        Anal. Chem., 62(3), 685 (1979).

   (4)  Suzuki, T., Ishikawa, K., Sato, N., and Sakai, K. I., J. Assoc. Off.
        Anal. Chem.. 62(3). 681 (1979).

   (5)  Mills, P.  A., Bong, B. A., Kamps, L. R., and Burke, J. A., J. Assoc.
        Off. Anal. Chem.. 55, 39 (1972); Mitchell, L, E., J. Assoc. Off.
        Anal. Chem.. 59, 209 (1976).                                   *~

   (6)  Suzuki, T., Ishikawa, K., Sato, N., and Sakai, K. I., J. Assoc. Off.
        Anal. Chem.. .62(3), 689 (1979).

   (7)  Vessel, J. R., J. Assoc. Off. Anal. Chem., 50, 430 (1967).

   (8)  McCully, K. A., J. Assoc. Off. Anal. Chem.. 55_, 291 (1972).

   (9)  Pardue, J. R., J. Assoc. Off.Anal. Chem., 54. 359 (1971),
                                    -364-

-------
                                                        Section 9A,N


(10)   Beckman, H., and Garber, D., J. Assoc. Off. Anal.  Chem., 52,  286
      (1969).                                           .        ~~  '   '

(11)   Versino, B,, van der Venne, M. Th., and Vissers,  H.,  J.  Assoc.
      Off.  Anal. Chem.. 54, 147  (1971).

(12)   Dale, W. E., and Miles, J. W., J. Assoc. Off. Anal. Chem.,  59,
      165 (1976).                                            "~~

(13)   Johansson, C. E., Pestic.  Sci.. £, 313 (1978).

(14)   Crist, H. L., Moseman, R.  F., and Noneman, J. W.,  Bull.  Environ.
      Contam. Toxicol.. 14. 273  (1975).

(15)   Solomon; J., and Lockhart, W. L., J. Assoc. Off.  Anal.  Chem.. 60,
      690 (1977).

(16)   Crist, H. L., and Moseman, R. F., J. Assoc. Off.  Anal.  Chem.»
      60, 1277 (1977).

(17)   Stijve, T., and Brane, E., Dt. Lebensmitt. Rdsch.. 73,  41 (1977).

(18)   Franken, J. J., and Luyten, B. J. M., J, Assoc.  Off.  Anal.  Chem.,
      59. 1279 (1976).

(19)   Griffith, F. D., and Blanke, R. V., J, Assoc. Off. Anal. Chem.,
      57, 595 (1974).

(20)   Kurtz, D. A., and Kim, K.  C., Pestic. Monit.  J..  10,  79 (1976).

(21)   Reiner, E., Krauthacker,  B., Stipcevic, M., and Stefanic, Z.,
      Pest. Monit. J.. 11, 54  (1977),

(22)   Gupta, R. C., Karnik, A.  B., Nigam, S. K., and  Kashyap, S.  K.,
      Analyst. 103, 723 (1978).

(23)   Edgerton, t. R., Moseman,  R. F., Linder, R. E.,  and Wright, L.  H.,
      J. Chromatogr., 170(2),  331 (1979).

(24)   Edgerton, T. R., and Moseman, R. F., J. Agr.  Food Chem.« 27(1),
      197 (1979).

(25)   Edgerton, T. R., and Moseman, R. F., J. Chromatogr. Sci.« 18(1),
      25 (1980).

(26)   Wu, A., Lech, J. J., Glickman, A.,  and Pearson,  M. L.,  J. Assoc.
      Off. Anal. Chem.. .61(5),  1303  (1978).

(27)   Delia Fiorentina, H., De Graeve, J., Grogna,  M., and  De Viest,  F.,
      J. Chromatogr.. 157, 421 (1978).
                                   -365-

-------
                                                        Section 9A,N


(28)  Smith, A. E., and Hayden, B. J., J.  Chromatogr..  171,  482 (1979).

(29)  Moseman, R. F., Cristj H. L., Edgerton,  T.  R., and Ward, M.  K.,
      Arch. Environ. Contain. Toxicol., £,  221  (1977).

(30)  Hodgson, D. W., Kantor, E. J.,  and Mann, J.  B., Arch.  Environ.
      Contain. Toxjcol.. 7_t 99 (1978).

(31)  Earless, R. L., Harris, D. E.,  Sovocool, G.  W., Zehr,  R. D., Wilson,
      N. K., and Oswald, E. 0., Biomed. Mass.  Spectrom.. 5/3), 232 (1978).

(32)  Moseman, R. F., Ward, M. K., Crist,  H. L.,  and Zehr, R.  D.,  J. Agr.
      Food Chem.. 26, 965  (1978).-

(33)  Stafford, C. J., Reichel, W. L., Swineford,  D. M., Prouty, R. M.,
      and Gay, M. L., J. Assoc. Off.  Anal.  Chem.,  61, 8 (1978).

(34)  Blanfce, R. V., Farris, M. W., Griffith,  F.  P., and Guzelian, P.,
      J. Anal. Toxicoli. JL, 57 (1977).

(35)  Meijs, A. W. H. M.,  and Ernst,  G. F., J. Chromatogr..  171. 486 (1979).

(36)  Mady, N., Smith, D., Smith,  J., and  Wezwick, C.,  NBS (US) Spec. Publ.
      519, 341 (1979).

(37)  Carver, R. A., Borsetti, A.  P., and  Kamps,  L.  R., J. Assoc.  Off.
      Anal. Chem.. 61,, 877 (1978).

(38)  Salek, F. Y., and Lee, G. F., Environ. Sci.  Technol..  12, 297 (1978).

(39)  Borsetti, A. P., and Roach,  J.  A. G., Bull,  Environ. Contain. Toacicol.,
      20, 241 (1978).

(40)  Johnson, L. D., Waltz, R. H., Ussary, J. P., and  Kaiser, F.  E;.,
      J. Assoc. Off. Anal. Chem..  59, 174  (1976).

(41)  Stalling, D. C., Tindle, R.  C., and  Johnson, J. L., J. Assoc. Off.
      Anal. Chem.. 55, 32  (1972).

(42)  Johnson, L. D., Waltz, R. H., Ussary, J. P., and  Kaiser, F.  E.,
      J. Assoc. Off. Anal. Chem..  5£, 174  (1976).

(43)  Griffitt, K. R., and Craun,  J.  C., J. Assoc. Off. Anal.  Chem.. 57.
      168  (1974).

(44)  Ault, J. A., Schofield, C. M.,  Johnson,  L.  D., and Waltz, R. H.,
      J. Agr. Food Chem..  27(4), 825  (1979).

(45)  Pflugmacher, J., and Ebing,  W., J. Chromatogr., 160(1),  213 (1978).

(46)  Pflugmacher, J., and Ebing,  W., J. Chromatogr., 93, 457 (1974);
      J. Chromatogr.. 151, 171  (1978).
                                    -366-

-------
                                                         Section 9A,N


 (47)  Wolff,  G.,  and Ebing, W., J. Chromatogr.. 147,  213  (1978).

 (48)  Bovey,  R. W.,  and Diaz-Colon, J. P., Tex. Agric.  Exp.  Stn.  Misc.
      Publ..  1381.  33 pp (1978).

 (49)  Brown,  M. J.,  J.  Agr. Food Chem.. 2!3, 334 (1975).

 (50)  Cromartie,  E.,  Reichel, W. L., Locke, L. N., Basile, A.  A.,
      Kaiser,  T.  E.,  Lament, T.  G,, Mulhern, B. M., Prouty,  R.  M.,  and
      Swineford,  D. M., Pestic.  Monit. J.. 9/1), 11 (1975).

 (51)  Blus, L. J., and  Lament, T. G., Pestic. Monit. J.. JL3_(2), 56  (1979).

 (52)  White,  D. H., Pestic. Monit. J.. 13(1), 12 (1979).

 (53)  Ohlendorf,  H. M., Elder, J. B., Stendell, R. C., Hensler, G.  L.,
      and Johnson, R. W,,  Pestic. Monit.  J.. 13_(3), 115 (1979),

 (54)  Kaiser,  T.  E.,  Reichel, W. L., and Locke, L. N., Pestic. Monit. J..
      13(4),  145  (1980).                                ""•	

 (55)  Bong, R. L,, J. Assoc.  Off. Anal.  Chem.. 60. 229  (1977).

 (56)  Burke, J. A., J.  Assoc. Off. Anal.  Chem.. 61(2), 359 (1978).

 (57)  Rourke,  D.  R., Mueller, W. F.,  and Yang, R.  S.  H., J. Assoc.  Off.
      Anal. Chem.. 60.  233  (1977).                                ~"~

 (58)  Langlois, E. B.,  Stemp, A. R.,  and Liska, J.,  J. Milk Food Technol..
      27(7), 202  (1964).                            ~~"	'	

 (59)  Frank, R.,  Braun, H.  E., Holdrinet,  M.,  Dodge,. D. P., and Nepszy,
      S. J., Pestic. Monit.  J..  12_(2),  69 (1978).

 (60) .Steinwandter, H., and Schlueter,  H.,  Z.  Anal.  Chem.. 286. 90  (1977).

 (61)  Steinwandter, H., and Schlueter,  H.,  Dt.  Lebensmitt. Rdsch..  74,
      139 (1978),                           ~~	~	•   .   ".."•.	

 (62)  Bidleman, T. F., Mathews,  J.  R., Olney,  C. E,,  and Rice, C.  P.,
      J; Assoc. Off. Anal.  Chem..  61,  820  (1978).           .   ,

(63)  Brevik, E. M., Bull.  Environ.  Contam.  Toxicol..  19,'281 (1978).

(64)  Holden, A. V., and Marsden,  K., J.  Chromatogr..  44,  481 (1969),

(65)  Duinker, J. C., and Hillebrandi M.. Th. J.. J. Chromatogr.. 150.
      195 (1978).                                       '	~~". 	

(66)  Wells, D. E., and Johnstone,  S. J., J. Chromatogr..  140. 17  (1977).
                                   -367-

-------
                                                        Section,9A,N


(67)  Telling, G. M,, Sissons,  D,  J,,  and Brinknan,  H.  W.,  J.  Chromatogr.,
      137. 405 (1977).

(68)  Panel on Determination of Organochlorine Pesticides  in Foodstuffs
      of Animal Origin, Ministry of Agriculture,  Fisheries  and Food
      (Great Britain), Analyst.  104. 425  (1979).

(69)  Wegman, R. C.  C., and Greve, P.  A., Pestic.  Monit. J.. 12(3), 149
      (1978).                                           ~™~ ~*~

(70)  Claeys, R. R., and Inman,  R. D.,  J. Assoc.  Off. Anal. Chem.. 57.
      399 (1974).

(71)  Lea, R., J. Assoc. Off. Anal. Chem.,  59.  224 (1976).

(72)  de Faubert Maunder, M. J., Egan,  H.,  Godley, E. W., Hammond, E. W.,
      Roburn, J., and Thomson,  J., Analyst,  89, 168  (1964).

(73)  Wood, N. F., Analyst. 94,  399  (1969).

(74)  Smart, N. A., Hill, A. R.  C., and Roughan,  P.  A., J.  Assoc. Off.
      Anal. Chem.. ;57, 153 (1974).

(75)  Friestad, H.-O., J. Assoc. Off.  Anal..  Chem.,. 60.  268  (1977).

(76)  Getz, M. E., Talanta. 22,  935  (1975).

(77)  Abbott, D_ C., and Thomson, J.,  Chem.  Ind.  (London),  p.  481 (1964);
      Analyst. 89_, 613 (1964).

(78)  Calderbank, A., Residue Reviews.  12,  14  (1966).

(79)  Renberg, L., Anal. Chem..  46_, 459 (1974).

(80)  Holden, A. V., Pestic. Monit. J.. T.*  37  (1973); Environ.  Qual.  Saf. .
      Suppl.. ^3, 40  (1975).

(81)  Report by the Panel on Fumigant  Residues  on Grain, Analyst. 99.
      570 (1974>.

(82)  Lores, E. M. > Bradway, D.  E., and Moseman,  R.  F., Arch.  Environ.
      Health. 33(5), 270 (1978).

(83)  Fournier,. E., Sonnier, M., and Dally,  S.., Clin. TorLcol..  12(4),
      457 (1978).                               "      i           ~~

(84)  Lores,. E. M., and Bradway, D. E., J. Agr. Food Chem.. 25,  75 (1977).
                                  -368-

-------
                                                        Section 9A,N


 (85)   Bradway,  D.  E.,  Shafik, T. M., and Lores, E. M., J. Agr. Food
       Chem..  25,  1353  (1977).

 (86)   Ito,  G.,  Kilgore, W.  W., and Seabury, J. J., Bull. Environ. Contarn.
       ToKicol.. 22,  530 (1979).

 (87)   Bradway,  D.  E.,  and Shafik, T. M., J. Agr. Food Chem.. 25, 1342
       (1977).                                                 ~~

 (88)   Shafik,  T.  M., and Enos, H. F., J. Agr. Food Chem., 17, 1186 (1969);
       Shafik,  T.  M., Bradway, D. E., and Enos, H. F., J. Agr. Food Chem..
       19, 885  (1971).

 (89)   Houx, N.  W.  H*,  Jongen, W. M. F., de Vries, A. W., Welling, 'W., and
       Dekker,  A.,  Pestic. Sci.. 10(3), 185 (1979).

 (90)   Daughton, C. 6., Cook, A. M., and Alexander, M., Anal. Chem..  51(12),
       1949  (1979).

 (91)   Takade,  D.  Y., Reynolds, J. M., and Nelson, J. H., J. Agr. Food Chem.,
       27(4),  746 (1979).

 (92)   Shafik,  T.  M.,  Sullivan, H. C., and Enos, H. F., J. Agr. Food  Chem..
       21, 295 (1973).

 (93)   Kirby,  K. W.,  Keiser, J. E., Groene, J., and Slach, E. F.,
       J. Agr.  Food Chem.. 27(4), 757  (1979).

 (94)   Edgerton, T. R., and Moseman, R. F., J. Agr. Food Chem., J26, 425
       (1978).

 (95)   Hladka,  A., Kovac, J., and Krampl, V.,  Z. Anal. Chem.. 274, 371  (1975),

 (96)   Kovac,  J., Batova, V., Hankova, A., and Szokolay, A., Bull. Environ.
       Contam.  Toxicol., 13, 692  (1975).

 (97)   Storherr, R. W., and Watts, R. R., J. Assoc. Off. Anal. Chem.. 48.
       1154  (1965).

 (98)   Watts,  R. R.,  and Storherr,r R. W., J. Assoc. Off. Anal. Chem.. 48.
       1158  (1965).

 (99)   Watts,  R. R.,  and Storherr, R. W., J. Assoc. Off. Anal. Chem.. 50,
       581  (1967).

(100)   Storherr, R. W., and Watts, R. R., J. Assoc. Off. Anal. Chem.. 51,
       662  (1968).

(101)   Eichner, M.. Chem. Labor. Betrieb. 2!8,  299  (1977).
                                   -369-

-------
                                                        Section 9A,N
(102)  Storherr, R. W., Murray,  E.  J.,  Klein, I., and Rosenberg, L, A.,
       J. Assoc. Off. Anal.  Chen..  50 605 (1967).

(103)  Malone, B., and Burke, J. A.,  J.  Assoc.. Off. Anal. Chem.. 52, 790
       (1969).                                             ~~~

(104)  Gay, L., and Cerney,  M.,  Mitt. Geb.  Lebensmittelunters. u. Hyp..
       66, 225 (1975).
(105)  Luke, B., results of  research  reported by McCully, K. A. in
       J. Assoc. Off. Anal.  Chem..  .62(2), 385 (1979).

(106)  Eichner, M., Z. Lebensm.  Untera.  Forsch.. 167(4), 245 (1978).

(107)  Heath, A. B., and Black,  R.  R.,  J. Asaoc. Off. Anal. Chem.. .62_(4),
       757 (1979).

(108)  Pflugmacher, J., and  Ebing,  W. ,  Landwirtsch. Porsch.» 32(1-2),
       82 (1979).

(109)  Storherr, R. W., Ott, P., and  Watts, R, R., J. Assoc. Off. Anal.
       Chem., 54_, 513  (1971).                       .

(110)  Laski, R. R., J. Assoc. Off. Anal. Chem.. 57, 930 (1974).

(Ill)  Bowman", M. C., Beroza, M., and Leuck, D. B., J. Agr. Food Chem..
       16, 79.6 (1968).                                    .

(112)  Sissons, D. J., and Telling, G.  M.,  J. Chromatoer..  47, 328 (1970).

(113)  Renvall, S.,. and Akerblom, M., Residue Reviews, 34,  1 (1971).

(114)  Jennings, D. M., Bunyan,  P.  J.,  Brown, P. M., Stanley, P. I., and
       Jones, ]?. J. S.» Pestie.  Sci., J5,  245 (1975).

(115)  Domanski, J. J., Haire, P. L., and Sheets, T. J., Arch. • Jfoviron.
       Contam. Toactcol>, £,  377  (1976).

(116)  Abbott, D. C., Crisp, S., Tarrant, K. R., and Tatton, J. 0*6.,
       Pestle. Scir.. 1., 10 (1970.

(117)i  Kadoum, A* MT, Bull.  Environ.  Contam. Toxicol.. 3^, 247  (1968).

(118)  Free, I>. J., Butler,  E. P»,  Kimball, E. R., and Stewart, D. K. R.,
       J> Econ. Entomol.. 69, 473  (1976).

(119)  Ripley, B. F., Wilkinson, R. J. * and Chau, A. S. Y.» J. Assoc. Off.
       Anal. Chem... 57. 1033 (1974).

(120)  Kadoum, A, M., Bull.  Environ.  Gontanu Toxicol.. 3^, 65 (1968) and
       2, 264 (1967).
                                   -370-

-------
                                                       Section  9A,N

(121)  Ivey, M. C., and Mann, H. D.><, J. Agr. Food Chem..  23,  319  (1975).

(122)  Ivey, M. C., J. Assoc. Off. Anal, Cheta.. .59, 261  (1976).

(123)  Bowman, M. C., and Hill, K. R., J. Agr. Food Chen..  19,  342
       (1971).

(124)  Aharorison, N., and Ben-Aziz, A., J. Agr. Food Chem.. 22. 704 (1974).

(125)  Bowman, M. C., and Beroza, M., J. Assoc. Off. Anal.  Chem.. 52.
       1231 (1969).

(126)  Thornton, J. S., Olson, T. J., and Wagner, K., J.  Agr. Food  Chem..
       25, 573 (1977).

(127)  Hild, J., and Thier, H. P., Deut. Lebensm. Rundsch.. 73. 330 (1977).

(128)  Hild, J., and Thier, H. P., Z. Lebensa. Unters. Forseh.. 166, 9 (1978),

(129)  Report of the Ministry of Agriculture,  Fisheries,  and  Foods, Analyst,
       102, 858  (1977).

(130)  Watts, R. R. Storherr, R. W., Pardue, J. R., and  Osgood, T.,
       J. Assoc. Off. Anal. Chem., 52, 522  (1969).

(131)  Sissons, D. J., and Telling, G. M., J.  Chromatogr..  47,  328  (1970);
       48, 468 (1970).  .

(132)  Smart, N. A., Hill, A. R. C., and Roughan, P. A.,  Analyst» 103  (1228),
       770  (1978).

(133)  Frehse, H., Anal. Chem.. 59., 896  (1976).

(134)  Becker, G., Method 58 in Deutsche Forschungsgemeinschaft (Lieferung
       1974) Ruckstands Analytik von Pflanzenschutzmitteln. No. 3., Verlag
       Chemie, Weinheim/Bergstr., W. Germany.

(135)  Desmarchelier, J., Pestic. Sci.. 8,  473 (1977).

(136)  Bradway, D, E., and Shafik, T» M,, J. Chromatogr.  Sci.,  15,  322 (1977),

(137)  Sherma, J., Critical Reviews in Analytical  Chemistry,  (CRC Press)
       1 (3), PP, 299-354  (1973).

(138)  Frei, R. W., Lawrence, J. F,, Hope,  J., and Cassidy, R.  M.,
       J. Chromatogr. Sci., 12> 40  (1974).

(139>  Bailey, R., Le Bel, G.,  and Lawrence, J. F., J. Chromatogr.. 161(1).
       251  (1978).
                                  -371-

-------
                                                        Section 9A,N


(140)  Strother, A.,  J.  Gas Chromatogr.. £, 110 (1968).

(141)  Storherr, R. W.,  J.  Assoc.  Off. Anal. Chem.. 55, 283 (1972).

(142)  'Cook, R. F., Stanovick,  R.  P.,  and Caasil, ;C. C., J. Agr. Food
       Chem.. 17,  277 (1969).

(143)  Lewis, D. L.,  and Paris, D.  F., J. Agr.  Food Chem.. 22, 148 (1974).

(144)  McKone, C.  E., and Hance, R. J.,  J. Chromatogr.. 36. 234 (1968).

(145)  Onley, J. H.,  and Yip, G.,  J. Assoc. Off. Anal.  Chem.. 54, 1366
       (1971).                                        '

(146)  LasfcL, R. R.,  and Watts, R.  R., J. Assoc. Off.  Anal. Chem.. 56,
       328 (1973).                     ~

(147)  Hall, R. C., and  Harris, D.  E., J. Chromatogr..  169. 245 (1979).

(148)  Erickson, M. D.,  Fank, C. W., and Morgan, D. P., J. Agr. Food Chem..
       27(4), 740  (1979).                                       —^—

(149)  Hoermann, W. D.,  Tournayre,  J.  C., and Egli, H., Pestle... Mbnit. J..
       13_(3), 128  (1979).                                        -^__

(150)  Thier, H. P., Angew.  Chem.  Int. Ed.. 10, 860 (1971).

(151)  Williams, I. H.,  Residue Reviews.  J38, 1  (1971).

(152)  Cochrane, W. P.,  and Purkayastha,  R., lexicological and Environ-
       mental Chemistry  Reviews. JL, 137  (1973).

(153)  Byast, T. H., Cotterill,  E.  G., and Hance, R. J.,  Methods of
       Analysis for Herbicide Residues.  Technical Report No.  15, 2nd
       ed., Agricultural Research Council,  Weed Research Organization,
       Yarnton, Oxford,  England (1977),  65 pp.

(154)  Ramsteiner, R., Hermann, W.  D., and Eberle,  D. 0., J.  Assoc.  Off.
       Anal.  Chem.. 57,  192  (1974).

(155)  Lawrence, J. F., and McLeod, H. A.,  J. Assoc. Off. Anal.  Chem..
       60, 979 (1977).

(156)  Lawrence, J. F., J. Chromatogr..  121, 85 (1976).

(157)  Thornton, J. S., and Draeger, G.,  Int. J. Environ.  Anal.  Chem..
       1, 229 (1973).

(158)  Sherma, J., and Shafik,  T. M., Arch.  Environ. Contain.  Toxicol.. 3,
       55 (1975).                             "	
                                  -372-

-------
                                                        Section 9A,N
(159)  Arthur, R. D., Cain, J, D., and Barrentine,  B.  F., Bull. Environ.
       Contain. Toxicol.. 15, 129  (1976); Pestic.  Mbnit.  J..  10, 168 (1977).

(160)  Kutz, F. W., Yobs, A. R.,  and Yang, H.  S.  C.,  Chapter 4 in Air
       Pollution from Pesticides  and Agricultural Processes„ Lee, R. E.,
       ed., CRC Press, West Palm  Beach, FL,  1976.

(161)  Thompson, J. F., and Reid, S. J., Arch.  Environ.  Contam. Toxicol.,
       .6, 143  (1977).

(162)  Agemian, H., and Chau, A.  S. Y., J. Assoc. Off. Anal, Chem., 60,
       1070 (1977).

(163)  Engst, R., and Knoll, R.,  Nahrung.  17,  837 (1973).

(164)  Aspila, K. I., Carron, J.*  M., and Chau,  A. S.  Y.,  J.  Assoc. Off.
       Anal. Chem., 60, 1097  (1977).

(165)  Johnson, L, G., Bull. Environ. Gontam.  Toxieol..  5_, 542 (1970).

(166)  Leoni, V., J. Chromatogr.. JS2, 63  (1971).

(167)  Law, L. M., and Goerlitz,  D. F., J. Assoc. Off. Anal» Chem.. 53,
       1276 (1970).

(168)  Schutzmann, R. L., Woodham, D. W.,  and  Collier, C. W., J. Assoc.
       Off. Anal. Chem., 54, 1117 (1971).

(169)  Johnson, R. E.,. and Munsell, L. Y., Bull.  Environ. jCpntani. ToadLcol.,
       17, 573  (1977).

(170)  Goerlitz, D. F., and Law,  L. M.. Bull.  Environ, Contam. TcTtJcol.,
       £, 9 (1971).

(171)  Mattson, P. E., and Nygren,.  S., J>  Chromatogr.. 124,  265 (1976).

(172)  Jensen,  S., Renberg, L., and Reutergardh,  L,,  Anal. Chem., 49,
       316  (1977).

(173)  Nagawawa, K.r Uchiyama, H.,  Ogamo,  A.,  and Shinozuka, T., J.
       Chromatogr., 144v 77  (1977).

(174)  La Bel,  G. C., Williams, D.  t.,  Griffith,  G.,  and Benoit, F. M.,
       J. Assoc.  Off. Anal.  Chem.,  62_(2),  241  (1979).

(175)  Devine,  J» M,, and Zweig,  G., J. Assoc. Off. Anal. Chem., 52, 187
       (1969).

(176)  Faas, L. F., and Moore, J. C.,  J. Agr.  Food Chem.> 27(3), 554 (1979),
                                   -373-

-------
                                                        Section  9A,N


 (177)   Sherma, J., "Thin Layer Chromatography:  Recent Advances", Chapter
        1 in Analytical Methods for Pesticides and. Plant Growth  Regula-
        tors. Volume VII, Sherma, J. and Zweig, G., editors, Academic
        Press, N.Y., 1973, pp. 1-88.

 (178)   Pionke, H. B., and Chesters, G., Soil Sci. Soc. Am. Proc., 32.
        749 (1968).                      	—'

 (179)   Burke, J.  A., J. Assoc. Off. Anal. Chem.. 59, 339  (1976).

 (180)   Hall, R.  C., and Harris, D. E., J. Chromatogr.. 169. 245  (1979).

 (181)   Maini, P,", and Collins, A., Bull. Environ. Contain. Toxicol.. 14.
        593 (1975).                      ~       !	  —

 (182)   Gary, A,  E., Douglas,  P., Tai, H., Mitchell, W. G., and Wiersma,
        G.  B., Pestic. Monit.  J.. 13(1), 17 (1979).

 (183)   Cotterill, E.  G., Analyst. 104(1242),  878 (1979)'.

 (184)   Goerlitz,  D. F., and Law., L.  M., J. Assoc. Off. Anal.  Chem.. 57.
        176 (1974).	  -~

 (185)   Suzuki, M.,  Yamato,  Y., and Watanabe,  T.,  Pestic.  Mbnit.  J..  11,
        54  (1977).	  —-

 (186)   Nash,  R. G., Harris, P. D., Ensor, P,  D.,  and Woolson,  E. A.,
        J.  Assoc.  Off.  Anal. Chem.. 56.  728 (1973).

 (187)   Official Methods of Analysis,  (1980),  13th Edition, AOAC,
       Washington,  B.C.,  Sections 29.015 and  29.018.

 (188)  Woolson, E.  A.,  J. Assoc.  Off. Anal. Chem.. 57. 604 (1974).

 (189)   Suett, D. L.,  Pestic.  Sci.. 6., 385 (1975).

 (190)  Miles, J. R. W., Harris,  C.  R., and Moy, P., J. Econ. Entomol..
       1, 97  (1978).                                	;	

 (191)  Smith, A. E., J. Chromatogr..  97.  103  (1974).

 (192)  Smith, A. E., J. Chromatogr..  129.  309 (1976).

 (193)  Bartha, R., J. Agr. Food  Chem.. 16. 602  (1968).

(194)  McKone, C. E., J. Chromatogr.. JJ4,  60  (1969).

(195)  Cohen, I. C., and Wheals,  B. B...J. Chromatogr.. 43. 233  (1969).
                                   -374-

-------
                                                       Section  9A,N

(196)   Baunok, I., and Gelssbuhler, H.. Bull. Environ. Contain.  Toxicol..
       J5,  7 (1968).

(197)   Abbott, D. C., Bunting, Ji A., and Thomson, J., Analyst, 90, 357
       (1965).

(198)   Jolliffe, V. A., Day, B. E., Jordan, L.  S., and Mann,  J. B,,
       J.  Agr. Food Chen.. 15, 174  (1967).

(199)   Smith, A. E., and Fitzpatrick, A., J.  Chromatogr..  57, 303 (1971).

(200)   Caro, J. H., Glotfelty, D. E., Freeman,  H. P., and  Taylor, A. W.,
       J.  Assoc. Off. Anal. Chea.,  56_,  1319  (1973).

(201)   Caverly, D. J,, and Denney,  R. C., Analyst. 10£,  576 (1977).

(202)   Sarna, L. P., and Webster, G. R.  B., J.  Assoc. Off. Anal. Chem.,
       57, 1279  (1974).

(203)   Swain, D. J., J. Aer. Food Chem..  27(4), 915  (1979).

(204)   Thier, H. P.> Angew. Chem. Int.  Ed.,  13, 217  (1974)„

(205)  Carey, A. E., Gowen, J. A.,  Tai,  H., Mitchell, W. G., and Wiersma,
       G. B., Pestic. Monit. J.,  12(4),  209  (1978).

(206)  Armour, J.  A., and Burke,  J. A.,  J. Assoc.  Off.  Anal. Chem., 53,
       761  (1970);  FDA PAM  (Section 251.1.

(207)  Albro, P. W., Baseman,  J".  K., Clemmer, T. A., and Corbett, B. J.,
       J.  ChromatQgr^, 136,  147  (1977).

(208)  Krupcik,  J., Kriz, J.,  Prusova,  D.,  Suchanek, P., and Cervenka, Z»,
       J.  Chromatogr., 142.  797  (1977).

(209)  Brinkman, U. A. Th.,  Seetz,  F.  L., and Reymer, H. G. M., J.  Chromatogr.,
       116.  353  (1976).

(210)  Crist, H. L., and Moseman, R. F., J.  Assoc. Off. Anal.  Chem.,  60_,
       1277 (1977).

(211)  Trotter,  W. J., and Young, S. J. V.,  J. Asaoc. Off. Anal. Chem..
       58,  466 (1975).

 (212)  Lieb, A.  J., and Bills, D. D., Bull.  Environ. Contam. Toxicol..
       12, 328 (1974).

 (213)  Gardner,  A. M.,  Righter,  H. F., and Roach, J. A. G.,  J.  Assoc.  Off.
       Anal. Chem.. 59,  273 (1976).
                                   -375-

-------
                                                                      Section  9A,N
               (214)  Armour,  J.  A. ,  and Burke, J. A. , J, Assoc. Off. Anal. Chem. . 54,
                                                                                   r "J
 (215)



 (216)



 (217)

 (218)



 (219)



 (220)

 (221)



 (222)



 (223)

 (224)

 (225)

 (226)



 (227)

 (228)



 (229)



(230)



(231)



(232)
                     Hucklns, J.  N. ,  Stalling, D. L. , and Johnson, J. L. , J. Assoc.
                     Off. Anal. Chem. . £9,  975 (1976).                    - ~~
                     Porter, M. L. ,  and Burke,  J.  A,, J. Assoc. Off. Anal. Chen., 54.
                     1426  (1971).                                    — - - - —  —


                     Erney, D. R. , J. Assoc.  Off.  Anal.  Chen.. _57., 576 (1974).

                     Picer, M. , Picer,  N. ,  and  Ahel,  M. , Pestic.  Monit. J.. 12(3) , 102
                     (1978) ; Picer, M. ,  and Ahel,  M. , J. Chromatopr. . 150, 119 (1978) .

                     Berg, 0. W. , Diosady,  P. L. ,  and Rees,  G. A. U. , J.  Assoc. Off.
                     Anal. Chem.. 53, 761  (1970).                      - ""•
                     Holdrinet, M, , J. Assoc. Off. Anal.  Chem..  5_7(3) ,  580 (1974).

                     Tiechman, J. , Bevenue, A., and Hylin,  J.  W. ,  J. Chromatoer. . 151.
                     155 (1978).                                   ~ - &—   -


                     Babjafc, L. J. , and Chau, A. S. Y. , J.  Assoc.  Off.  Anal.  Chem.. 62(5),
                     1174 (1979).                                  --- —  —


                     Miles,  J. R. W. , J. Assoc. Off. Anal.  Chem..  55. 1039 (1972).

                     Jensen, S., and Sundstrom, G. , Ambio.. 3,, 70  (1974).

                     Underwood, J.  C. , Bull. Environ. Contam. Toacicol. . 2.1(6),  787 (1979),

                     Luckas, B. , Pscheidl, H. , and Haberland, D. .  Nahrung.  20,  KL-K2
                     (1976).                                               —


                     Trotter,  W. J., J.  Assoc. Off.  Anal. Chem.. 58, 461 (1975).

                     Insalaco, S. E. , Makarewicz,  J.  C.", and Vestling, M. W. , J. Acr.
                     Food Chem.. 28(2),  261 (1980).                            ~ - 6—


                     Welborn,  M.  E. , Borchard, R.  E. ,  Hansen, L.  G. , and Metcalf,  R. L.
                     J. Assoc.  Off.  Anal.  Chem.. SJ,  1248 (1974),

                     Doguchi,  M. , and Fukano,  S, ,  Bull. Environ.  Contam. Toxicol.. 13,
                     57 (1975).                                   — - "" -  —


                     Bruse,  V.  W. , Kimbrough,  R. D., Villanueva,  E. C. ,  Jennings,  R. ,
                    Linder, E.,  and Sovocool, G. W. ,  Arch.  Environ.  Health. 29, 301 (1974)

                    Cochrane,  W. P.,  and  Whitney, W. ,  Advances in Pesticide Science.
                    Geissbuehler, H, , ed.,  (3), 664  (1979)";       ~~ - '
.
                                               -376-

-------
                                     ,                   Section 9A,N

(233)  Webb, R. W., and McCall, A. C., J. Chromatogr.  Sci.. JU, 366 (1973).

(234)  Sawyer, L. D., J. Assoc. Off. Anal. Chen..  61,  272 (1978).

(235)  Sawyer, L. D.. J. Assoc. Off. Anal. Chen..  61,  282 (1978).

(236)  Armour, J. A., J. Assoc. Off. Anal. Cheat..  56_,  302 (1973).

(237)  Seidl, G., and Ballschmiter, K.,  Z. Anal. Chem.. 296_(4), 281 (1979).

(238)  Collins, G. B., Holmes, D.C,,  and Jackson, F.  J,, J.  Chromatogr..
       71, 443  (1972).

(239)  Zobel, M. G. R., J. Assoc.  Off. Anal.  Chem., 57, 791 (1974).

(240)  EicheJberger, J. W., Harris, L. E., and Budde,  W. L.,  Anal. Chem..
       46, 227  (1974).

(241)  Zelenski, S. G., Tashiro, J., Worthen, L. R., and Oln.ey, C. E.,
       J. Chromatogr..  84, 67  (1973).

(242)  Cairns,  '£., and  Jacobson, R. A.,  J. Chem.  Inf.  Comput._Sci., 17,
       105  (1977).

(243)  Mulhern,  B. M.,  Cromartie,  E.,  Reichel, W.  L.,  and Belisle, A. A.,
       J. Assoc. Off. Anal. Chem.. 54, 548 (1970).

(244)  Yobs, A.  R., Abstracts  of Papers  WATR 18;  Biros, F. J., Enos, H. F,,
       Walker,  A.  C., and Thompson, J. F,., Abstracts of Papers WATR 19;
       164th National ACS Meeting, New IJork  City,  August, 1972.

(245)  Young,  S.  J. V., and Burke, J.  A,, Bull. Environ. Contain. Toxicol..
       2, 160  (1972).

(246)  Holmes,  D.  C., and Wallen,  M.,  J. Chromatogr.. 71» 562  (1972).

(247)  Szelewski,  M.  J.,  Hill, D.  R,,  Spiegel, S.  J., and Tifft, E. C.,
       Jr.. Anal.  Chem.,  51,  2405  (1979).

(248)  Fehringer,  N.  V.,  and  Westall,  J^ E., J. Chromatogr..  57, 397  (1971)

(249)  de Vos,  R., and  Peet,  E. W., Bull. Environ. Contam. Toxicol., ,6_,
       164  (1971).
        Lewis,  R.  G., Hanisch, R. G., MacLeod, K., and Sovocool, G. W.,
        J.  Agr. Food Chem.. 24, 1030 (1976). .
(250)


(251)  Mansour, M., and Parlar, H., J. Agr. Food Chem..  26,  483  (1978).
                                   -377-

-------
 (252)
                                                 Section 9A,N


Dolan, J. W., Hall, R. C,,  and Todd,  T.  M.,  J.  Assoc. Off. Anal.
Chem.. 55, 537  (1972).	 -   -                      "~~	
 (253)  Ballet,  D.  J.,  Norstrom, R. J., Onuska, F. I., Comba, M. E.,  and
       Sampson, R.,  J. Agr. Food Chem.. _24, 1189 (1976).

 (254)  MacNeil, J. D., Bull. Environ. Contain. Toxicol.. 15, 66  (1976).

 (255)  Hass, J. R., McConnell, E, E., and Harvan, D. J., J. Aer. Food
       Chem.. 26,  94  (1978).                         "	

 (256)  Willett, L. B.% Brumm, C..J., and Williams, C. L,, J. Agr. Food
       Chem., 26,  122  (1978). '                         •       	

 (257)  DeKok, J. J., J.  Chromatogr.. 142_, 367 (1977).

 (258)  Farrell,  T. J., J.  Chromatogr.  Sci.. 18(1), 10 (1980).

 (259)  Fehringer, N. V., J.  Assoc. Off.  Anal. Chem.. 58, 978 (1975).

 (260)  Chittam,  B., Safe,  S., Ruzo, L.  0., Hutzinger, 0., and Zitko, V.,
       J. Agr.  Food Chem..  25_, 323 (1977).

 (261)  Wright,  L. H.,  Lewis, R.  G., Crist, H, L., Sovocool, G.  W., and
       Simpson,  J. M., J. Anal.  ToxLcpl.. .2(3),  76 (1978).

 (262)  Harless,  R. L., Oswald, E.  0.,  Wilkinson,  M.  K.,  Dupuy,  A. E., Jr.,
       McDaniel, D. D.,  and Tai,  H., Anal. Chem.. 52_(8), 1239-1244 (1980).

 (263)  Firestone, D.,  Ecol.  Bull.. 27.,  39 (1978).

 (264)  O'Keefe,  P. W., Meselson,  M. S.,  and Baughman, R. W., J.  Assoc.
       Off. Anal. Chem.. 61,  621  (1978).                         '•        .

 (265)  Shadoff, L. A., and  Hummel, R.  A., Biomed. Mass.  Spectrom. 5, 7
       (1978).                             	•*	

 (266)  Shadoff, L. A., Hummel, R.  A.,  Lamparski,  L.,  and Davidson,  J. H.,
       Bull. Environ.  Contam.  ToadLcol. 18. 478 (1977).

 (267)  di Dimenico, A., Anal.  Chem.. 61(6),  735  (1979).

 (268)  Buser, H. R., in Dioxin;   lexicological and Chemical Aspects.
       Cattabani, E., Cavallaro, A., and  Galli, G., eds., Spectrum  Publ.
       Inc., NY, Chapter 4, pp. 27-41  (1978).

(269)  Freudenthal, J., in Dioatin;   ToxJcological and Chemical Aspects.
       Cattabani, E., Cavallaro, A., and  Galli, G., eds., Spectrum  Publ.
       Inc., NY, Chapter 5, pp. 43-50  (1978).
                                  -378-

-------
                                                 Section 9A,N


Harless, R. L. , and Oswald, E.  0.,  in Dioxin;   Toxicologies! and
Chemical Aspects, Cattabani,  E.,  Cavallaro,  A., and Galli,j3. ,
   i ,~ Spectrum Publ. Inc., NY,  Chapter 6,  pp.  51-57 (1978)"
(270)
(271)  Cruiranett, W.'B.,  Nature (London), 283/5745), 330 (1980).

(272)  Hummel, R, A.,  J.  Agr.  Food Ghent. , 25_, 1049 (1977).

(273)  Buser, H. R. , Anal.  Chem.  49, 918 (1977).

(274)  Buser, E. R.' . J.  Chromatogr . 129. 303 (1976).

(275)  Rappe, C. , and  Marklund,  S., Cheroosphere 3, 281 (1978).

(276)  Hummel, R. A.,  and Shadoff, L. A., Anal. Chem.. 52(1), 191  (1980).

(277)  Hass, J.  R. , and  Friesen,  M. D. , N.Y. Aead. Sci. , Sci. Week,
       June 21-30  (1978),

(278)  Lamparsfci, L. L. , Nes trick, T. J. , and Stehl, R, H. , Anal.  Chem. ,
       51(9), 1453  (1979).

(279)  Shadoff,  L.  A., Anal. Chem., ^O(ll), 1586  (1978).

(280)  Newsome,  W.  H, , J. Agr. Food Chem.. 20, 967 (1972).

(281)  Haines, L.  D. ,  and Adler,  I. L, , J. Assoc. Off. Anal.  Chem., 56,
       333 (1973).

(282)  Yip, G. ,  J.  Assoc. Off. Anal. Chem., 60, 370, 1105 and 1111 (1977).

;;(283)  Newsome,  W.  H. , J. Assocl  Off. Anal. Chem., 61(2), 361 (1978).
•o •           '                  l>'  • •
(284)  King, R.  R. , J. Agr. Food Chem.. 25, 73  (1977).

(285)  Newsome,  W.  H. , J. Agr. Food Chem.. 24, 999 (1976).

 (286)  Otto,  S., Keller, W. , and Drescher, N. , J. Environ. Sci. Health.
       Part B,  12,  179 (1977).

 (287)  Committee on Terminal Pesticide Residues Applied Chemistry
       Division, ITIPAC,  Pure Appl. Chem.. 49. 675  (1977).

 (288)  Kaufman,  D.  D. , Still, G. G. , Paulson, G.  P., and Bandal, S. K. ,
       Bound  and Conjugated Pesticide Residues, ACS Symposium,  Series
       No. 29,  1976.

 (289)  Balba,  H. M. ,  Still, G. G. , and Mansager,  E, R. , presented  at the
       175th  ACS National Meeting, Anaheim, California, March 13-17, 1978,
       Pesticide Chemistry  Section Paper No. 6.
                                    -379-

-------
                                                        Section 9A,N

(290)  Cook, R. F., Carbofuran,  in Analytical Methods for Pesticides
       and Plant Growth Regulators. Vol.  VII. Shgt-™., .T.  *~A 7~-*c
       G., editors, 1973, p. 187;  J.  Agr.  Food Chem.. 25, 1013 (1977).

(291)  Nony, C. R., Bowman, M. C.,  Holder,  C. L.,  Young,  j.  F., and
       Oiler, W. L., J. Pharm. Sci..  65.  1810 (1977).

(292)  Ivey, M. C., and Oehler,  D.  D., J. Agr.  Food Chem.. 24, 1049 (1976),
                                  -380-

-------
                                  Section 10
              CONFIRM/VTORY AND OTHER DETERMINATIVE PROCEDURES
10A   REQUIREMENTS FOR POSITIVE CONFIRMATION OF PESTICIDE IDENTITY

    Obtaining convincing identification of a trace residue is-a major task
    of.the pesticide analyst.  The" identity of pesticide residues should
    always be confirmed by a method different from that used  in the initial
    determination since interpretation of results (e.g., decisions of a
    legal or health nature) as well as reliable quantitation .(selection
    of standards) depend on correct identification.  Multiresidue GC
    analytical methods do not provide irrefutable identification since
    interfering materials and artifacts are often observed, and metabolic
    and decomposition products may be encountered.

    A specific example of a serious identification problem is the determina-
    tion of the PCBs, which are easily mistaken for  pesticide residues such
    as £,£T-DDE and j3,£r-DDT.  Another important example concerns overlapping
    peaks when foods are screened for tolerance levels:  a 4% SE-30/6% QF-1
    column may give peaks at essentially  identical retention  times for endrin
    and .0,2.'-DDT, for Endosulfan  I and £,_p_'-DDE, and for B-BHC and lindane.
    Both DDT and DDE are very common pesticides with rather high tolerance
    levels.  Thus, if the analyst is unaware  that endrin and  endosulfan may
    produce corresponding GC responses, he may conclude that  observed peaks
    indicate only insignificant quantities of DDT and DDE relative to
    tolerance levels and that no  further  work is necessary.   Unfortunately,
    what appears  to be  insignificant response for DDT and DDE is very sub-
    stantial response for endrin  and Endosulfan I because of  lower GC
    sensitivity  to these compounds  and  lower  tolerance levels;  therefore,
    confirmation of  identity  is mandatory (1).

    Confirmatory evidence  is  especially important with the  relatively non-
     specific EC  detector.   One  difficulty is  that  determinations of  very-
     low pesticide concentrations  are  usually  required, and  many potentially
     useful confirmatory methods  (e.g.,  infrared  spectroscopy) require a
     greater quantity and/or purity of pesticide  than might  be available.
     The techniques  chosen for confirming various  residues will depend  on
     the nature of the pesticide,  the level found,  the type  and amount  of
     sample,  and the presence of other residues.   The lower  the concentration
     of pesticide present,  the fewer or less certain are the available methods
     for making positive identification.   If larger amounts of residue  are
     found and can be isolated in a reasonably pure state,  infrared (IR)
     spectroscopy and mass spectrometry can provide unequivocal identification.
                                      -381-

-------
                                                          Section 10A

 Considerations of set  theory  (2)  indicate that three independent "equivocal"
 results are required in  order to  be confident of the identity of a  pesticide
 residue.  These might  be elution  in a certain fraction from a liquid chroma-
 tography cleanup column,  a GC retention time, and a positive response of a
 selective GC detector.   Another possible combination that would be  a basis
 for confidence is the  GC retention times from a polar column and a  nonpolar
 column plus an Rp value  from  paper chromatography (PC) or thin  layer
 chromatography (TLC) or  an extraction £-value.  Still another would be a
 GC retention time, a PC  or TLC RF value, and the GC retention time  of a
 derivative formed by a chemical or photochemical reaction.

 The dependence or independence of measured values was studied by Elgar (3)
 who reported that many widely used confirmatory methods may not  give
 truly independent evidence of identity since they are measuring  the same
 chemical or physical properties.   Thus,  care must be exercised when
 deciding which methods to use in  combination to avoid doing a great deal
 of work without gaining  additional useful information.  Examples  of highly
 correlated (not independent)  values include GC retention times on certain
 stationary phases (Figures 5-A,A  in Section 5);  PC or TLC Rp values from
 certain adsorbent/solvent systems;  ^-values in different solvent pairs;
 and PC,  TLC, and ^-values.  These combinations will not provide  independent
 information for confirming residue identity.

 In Figure 10-A,  the correspondence between extraction ^-values in hexane/
 acetonitrile and isooctane/DMF  solvent pairs  (A),  and ^-values in hexane/
 acetonitrile and TLC Rp values with the  system silica gel/hexane  (B)  is
 shown by the generally straight line  along which the plotted data points
 lie.  The independence of TLC and  PC  data [Figure  10-A,  (C)]  and GC and
TLC  data (D)  is  illustrated by  the  scatter of the  points.   Clearly,  many
 combinations of  alternative columns,  selective detectors,  £-values or PC
or TLC,  and chemical derivatization can  be applied for purposes  of confirma-
 tion.,,
     o

Figure 10-A.  Degree of correspondence between different  types of data for
              residue confirmation.  A =  extraction £-values, B  = TLC vs
              £-values, C = PC vs TLC, D  =  TLC vs  GC
J.1.0

£0,8
,5
1 0.6
|o.Z


O
O
o
" 2?°
!/° . .
    0,2 0.4  0.6 0.3 1.0
      Ijoocuno/DMF
                     B   Extraction p-valuc v
                     •"• thin-layer chromatography
l.u
50.8
,2
{0.6

?0.4
e
g
£0.2



O
00
o


^D *
0 0°
o
                      0  0.2 0~.
-------
                                                           Section 10B
  When  the  analyst  is  making pesticide identifications,  common sense is
  necessary.  An  example  of  misapplied common sense would be reporting
  methyl  or ethyl parathion  in human fat;  metabolically  it is virtually
  impossible for  parathion to persist per  se and to appear in a tissue or
  body  fluid (except gastrointestinal).  The.persistence of heptachlor
  would also be very unlikely because body metabolism normally converts
  it  to heptachlor  epoxide.   Chromatography with EC detection of human
  adipose tissue  from  the general population often produces peaks with
  retention characteristics  very close or  identical to the RRT^ values for
  ct-BHC and/or  £,j>/-DDE.  However, the presence of these compounds has
  rarely, if ever,  been confirmed.  In these instances,  the peaks in question
  represent artifacts  that happen to have  the same retention times as these
  pesticides, and careful confirmation by  ancillary techniques would provide
  the proper identification.

  In  summary, since all methods and tests  regularly used in residue analysis
  are presumptive in nature  (the behavior  of an unknown  is compared to that
  of  a  known, standard material), it is most desirable to use a number of
  tests that measure different chemical or physical properties.  The initial
  GC  method should have been proved to recover and detect the pesticide
  residues  of interest, and  it is desirable that data are available on the
  behavior  of many pesticides and their metabolites and  degradation products
  in  the  various  operations  that comprise  the method. The analyst should
  be  familiar with and capable of fully using and interpreting these data
  and all other available information, including pesticide usage, the
  chemistry and metabolism of residues, common artifacts from sample sub-
  strates and reagents, and  the possibility of interfering residues, such
  as  PCBs and phthalate esters.  Analytical conclusions  must be reached  with
  an  open mind, common sense, and reasonable judgment.  The extent of
  confirmatory  effort  and the exact procedure chosen will depend on factors
  such  as the history  and significance of  the sample; nature and level of
  the residues; sairple type; purpose of the analysis; and practical con-
  siderations such as  time,  cost, number of samples, and available instrumen-
  tation.  Alternatives to  confirmation of residues in all samples are
  discussed in  Section IE.   "Unusual" residues should be verified in all
  analyses, even  at low levels, to support a decision to devote further
  effort  to tracing their origins.  Confirmatory methods should yield
   identical results with both the suspected sample residue and standard
   reference material  subjected concurrently to the same  tests.  Similar
   concentrations  of the sample and standard should be used in the comparative
   testing to demonstrate quantitative as well as qualitative confirmatory
   evidence  (FDA PAM,  Section 601).  The following subsections discuss the
  more  widely.used confirmatory procedures, some of which are also useful
   for residue quantitation.


10B    GC RELATIVE RETENTION  TIMES

   In most laboratories, the initial, tentative identification of a pesticide
   residue results from a multiresidue procedure involving extraction, cleanup,
                                      -383-

-------
                                                         Section 10B


 and  gas  chromatography.   Tables  of GC  retention times  for particular column-
 detector combinations  are normally used for the tentative identification.
 The  recovery of  a residue through the  preliminary cleanup steps should not
 be overlooked as valuable,  supplemental confirmatory evidence.   This is
 particularly true when such characteristic  properties  as the  ability to
 withstand acid or alkali  treatment or  elution  in a particular fraction
 from an  adsorbent column  is involved.

 The  following guidelines  are useful for the proper utilization  of
 retention times  in making compound identifications.

      a.   The use of relative retention or Kovats1  retention indicies (4)
 rather than  absolute retention is more reliable (Subsection 5N  in Section 5).

      b.   Be  highly suspicious cjf any peak with, a calculated relative re-
 tention  value (RRT) that  does not precisely match that of the standard
 or that  of the tables  [EPA PAM,  Section 4,A, (6)].   A  simple  aid is  to
 co-chromatograph some  pure  standard of the  suspect compound along with
 the  sample extract and observe the peak configuration  compared  to that
 of the sample alone.   If  some distortion is evident  in the configuration
 of a given suspect peak,  the identification can be safely negated.

      c.   If  cleanup is used on the sample,  always  run  the elution fractions
 separately.   Do  not pool  the elution cuts.   Selective  adsorption combined
 with GC  retention characteristics provides  a valuable  identification tool
 for  pesticide analysis.

      d.  NEVER rely on one  GC column for positive  identification.  Use an
 alternative  column providing a completely different  peak elution pattern.


As illustrated in Figure  5-A,A in Section 5, the combination  of  a nonpolar
DC-200 column with a slightly polar  DC-200/QF-1 column (plot A)  is not
very useful  for  confirmation.  Another highly  correlated  pair of phases
 is slightly  polar 4% SE-30/6% QF-1 with slightly polar 1.5% OV-17/1.95%
 QF-1.  To  the contrary, a combination  of DC-200 with highly polar DECS
 (plot B) or  highly polar  OV-210 with OV-17/QF-1 (plot C) would be a  good
 choice.  Other complimentary pairs are SE-30/QF-1 with either DECS or
 OV-210.

Specific examples  (5)  of  the utility of at  least two different GC columns
 for  sample diagnosis include the  following.  Identity of  certain early
eluting BHC  isomers, particularly the  alpha isomer, may be hindered by
 the presence  of hexachlorobenzene.   The  latter  is co-eluted with a-BHC
on silicone  columns and with B-BHC on Apiezon,   but all three compounds
are resolved  on a polar cyano-silicone  column.   Dieldrin and £,£f-DDE
are difficult to resolve  on  a number of  single  phase silicone columns
but are separated on Apiezon, cyano-silicone, and trifluoropropyl silicone
 (QF-1, OV-210, SP-2401).  On Apiezon,  dieldrin  elutes before DDE while
the order is reversed on  the cyano-silicone column.  On the QF-1 or OV-210
                                  -384-

-------
                                                        Section 10B
column, dieldrin elutes far later than £,j>/-DDE, to the extent of about
l.'Ax at 180°C column temperature.  Figure 10-B illustrates the confirma-
tion of organochlorine pesticides.by comparison of relative  retention
times on two columns of different polarities.


Figure 10-B.  Dual column confirmation of pesticides by electron capture
              detection*  Pesticides (from left):  lindane,  aldrin,
              dieldrin, o,£*-DDT, p_,_p_'-DDT, and unknown.  Top chromatogram:
              4% SE-30/6TOV-210 on Gas Chrom Q.  Bottom  chromatogram:
              1.5% OV-17/1.95% OV-210 on Gas Chrom Q.  Both  columns:
              6.3 mm x 183 cm glass.  Carrier gas:  nitrogen, 65 ml/min.
              Oven temperature:  200°C.  Chart speed:  1.^27  cm/min.
                                     -385-

-------
                                                        Section IOC
IOC   SELECTIVE GC DETECTORS

   The EC detector, being rather non-specific, responds to any electron
   capturing compounds injected in addition to pesticides.  For this
   reason, interpretation of results from EC-GC is facilitated if
   additional chromatograms are run using one or more of the highly
   selective detectors.  The thermionic, flame photometric, or conductivity
   detectors, described in Subsections 5E, 5F, and 5G, are especially useful
   for confirmation.  Because interference peaks may occur with even the
   most selective detectors available, the absence of a peak is really
   more conclusive than a positive response.  For example, if a peak on
   an electron capture chromatogram suspected of being a chlorinated
   pesticide does not appear when the sample is injected into a
   chromatograph with the same column and a Hall conductivity detector
   in the Cl mode, this is convincing evidence that the original peak
   was definitely not due to a chlorinated pesticide but most likely
   an artifact with a coincident retention time.  Appearance of the peak
   in the conductivity chromatogram indicates that the peak was due to a
   halogenated compound, but further confirmation is still required to
   prove that the peak truly represents the pesticide of interest and not
   an artifact.

   Because of the- selectivity of its filters, the flame photometric
   detector (FPD) may simplify confirmation of sulfur- and/or phosphorus-
   containing residues.  Identification of a thiophosphate is usually
   unequivocal if (a) its retention ratios on at least two different .GC
   columns of different polarity match with those of a standard, (b) the
   compound elutes in the correct fraction from a cleanup column, (c) it
   is detected by the FPD detector, and (d) the sulfur (394 nm) to phosphorus
   (526 nm) response ratio of the FPD matches the standard (Subsection 5F).
   Figure 10-C illustrates simultaneous chromatograms of parathion,
   malathion, and diazinon generated by monitoring both phosphorus and
   sulfur emissions with a dual flame photometric detector.
                                      -386-

-------
                                                           Figure  IOC
Figure 10-G.
          Gas chromatograms of phosphorothioate pesticides obtained

          simultaneously with a dual flame photometric  detector.
                                            I.;.;. ._^—:>-  •  .—-a)
                                                                  UH
                                                                  "  i-!:
M..rp:.-u...;.^.:;::io•;,-'— 3  —»::•  s

;r:.3 -_-:j::
«— Tla*/Mtn.
".: •••:?*—rj:.~i
 •r.trirr-o-nu

• • F  r ~*T



I'prl
Ul|-	,U
                                                                     . I .
                                    -387-

-------
                                                              Section  10D

     .Relying on  only  one GC  column may  lead  to  incqrrect  identification, even
     with'the added information  from the Florisil column  elution and  the FPD.
     For example, phosalone  and  azinphosmethyl  have  the same  relative retention
     time on an  OV-1  GC column;  both elute in the third (hexane-acetone,
     85:15 v/v)  fraction* from the 2% deactivated Florisil column recommended
     by the Canadian  PAM (Section 9M, this Manual),  and both  respond  equally
     to the FPD  since each compound  has one  P and two S atoms per molecule.
     A second example is phosalone and phosmet  (Imidan),  both of which have
     the same retention on OV-17, elute in the  third Florisil column  fraction*,
     and have one P and two  S atoms  per molecule (6).

     The selectivity  of an EC detector can be improved if the products formed
     in the detector  are allowed to  pass to  a second column with another EC
     detector; the resultant distinctive peak pattern can provide identifica-
     tion of OC1 pesticides  and  PCBs  (7).

     When it is possible to  use  two  gas chromatographic detectors, further
     confidence in qualitative accuracy can  be  achieved.  For example,
     simultaneous analysis by electron capture  and flame  photometric  gas
     chromatography is very useful for confirmation  of organophosphorus
     pesticides  (Figure 10-D).   The Hall microelectrolytic conductivity and
     nitrogen-phosphorus detectors are likewise very useful for dual-detector
     confirmation.

10D    THIN LAYER CHROMATOGRAPHY (TLC) RF VALUES

     Experimental aspects of TLC and  its use for screening and quantitation
     of residues were covered earlier in Subsections J through M in Section 7.

     TLC is perhaps the simplest confirmation technique for GC when levels''of
     residues present are high enough.  An aliquot of cleaned-up extract is
     evaporated to near dryness, a suitable  solvent is added,  and a detectable
     quantity of the sample is spotted on a  thin layer plate together with.
     appropriate standards.   An agreement of about + 2 mm in migration distance
     of the sample and standard spots is considered adequate since the movement
     of the sample is likely to be affected by co-extractives despite cleanup
     steps.   If the sample contains several pesticides,  different solvents
     and/or adsorbents may be required before all are separated and matched
     with standards.   Mixing together the sample and a standard and observing
     whether separation occurs (co-chromatography)  is another procedure for
     making comparisons.

     It is best not to rely on published or previously determined Rj- values
     for confirmations because differences in development  conditions from run
     to run cause these values to be non-reproducible.   Standards and samples
        A recently devised elution system for 2% deactivated Florisil columns,
        modified from that described in Section 9M, includes four eluents:
        hexane-methylene chloride•(95:5 v/v), hexane-methylene chloride
        (70:30 v/v), hexane-acetone (85:15 v/v), and hexane-acetone (1:1 v/v
        (6).
                                        -388-

-------
                                                        Section 10D
Figure 10-D.
Simultaneous gas chromatograms of organochlorine and
organophosphorus pesticides using electron .capture and
flame photometric detectors.  Columns:  6.3 mm x 183 cm,
4% SE-30/6% OV-210 on Gas Chrom Q.  .Carrier gas:  nitrogen.
Oven temperature:  200°C.  Detectors:  Pulsed 63Ni, 270°C;
flame photometric, 526 nm filter, 200°C.  Chart speed:
1.27 cm/min.
should always be run on adjacent areas of the same plate if possible.
If Rp values must be used, the value relative to the Rp of a standard
compound X run on the same plate (Rg value) will be more reliable than
the absolute Rp value for many of the same reasons that relative GC
retention times are more reliable than absolute retentions.  Chlorinated
pesticides are often referred to DDD, and phosphates to parathion, in
calculating R^ values.

Although TLC is very widely applied for pesticide confirmation, results
may not always be conclusive.  TLC confirmation of many pesticides, such
as toxaphene and chlordane, is greatly influenced by the degree of clean-
up on the sample extract and the level of detection.  Oils and waxes will
                                    -389-

-------
                                                        Section 10D


particularly.interfere with TLC, causing streaked zones and/or distorted
RF values that may completely negate its value for confirmation.  The
15% ethyl ether-petroleum ether Florisil column extract normally requires
further cleanup prior to TLC (FDA Pesticide Analytical Manual, Section
411.5).

Detection reagents yielding spots of different colors with different
pesticides are especially valuable for confirmation.  Diphenylamine-
zinc chloride reagent provides such differentiation for chlorinated
pesticides; various shades of purple, grey, green, and reddish-orange
colors are produced on the layer after spraying and oven heating (FDA
PAM, Section 612).  Identification of naturally fluorescent pesticides
is aided by heating the chromatogram, causing specific alterations in
recorded spectra (8).  Th.is heating procedure may, however, increase
background fluorescence from co-extracted compounds also present in
the sample.  TLC after fluorogenic labeling (9) of pesticide residues
is a. combination of chromatography with chemical derivatization (Sub-
section 10G) that can provide very specific detection of certain residues.
If sufficient pesticide is present in the thin layer spot, scraping,
collecting the adsorbent, and eluting the compound followed by mass
spectrometry (Subsection 10L) can provide unequivocal identification.
        »
It was mentioned earlier in this section that if additional independent
information is to be gained by running PC plus TLC or TLC in more than
one system, the systems must be very carefully chosen to be truly
"different".  The use of multiple Rp values for identification purposes
was studied by. Connors (10), who found that useful, uncorrelated data
can be obtained in several ways, such as by pairing aqueous with non-
aqueous systems, acidic with basic solvents or supports, aprotic with
protic solvents, polar with nonpolar solvents, hydrogen-bond donors with
hydrogen-bond acceptors, or reversed phase with normal phase systems.
The specific approach that might be successful depends on the chemical
nature of the pesticides to be confirmed.  The important point is that
different thin layer and/or PC systems chosen at random will not
necessarily provide the analyst with any additional, independent evidence
of identity.  Similar correlation studies were reported by Dale and
Court (11).

Permanent records of TLC plates for documentation should be made by one
or more of the following methods:  Xeroxing the original plate, spraying
with plastic to preserve the plate, hand tracing or charting, densitometry,
or color photography (12).  Where available, the latter appears to be the
preferred procedure.

Section 614.11 of the FDA PAM describes a method for confirmation of
organophosphorus pesticide residues by two-dimensional TLC.  It is
applicable at levels as low as 0.01 ppm in nonfatty food extracts cleaned
up by carbon column chromatography.  The pesticides are oxidized by
bromine vapor after the first development, and detection is made with
horse serum cholinesterase and indoxyl acetate substrate after the second
development.  The system provides good specificity because it involves
chromatography of both the parent pesticides and their derivatives.
                               -390-

-------
                                                             Sections IDE, 10F
10E    HIGH PERFORMANCE LIQUID CHROMATOGRAPHY (HPLC)
     See Subsections 7A through 71 in Section 7 for a discussion of this topic.
     HPLC has been used mainly for quantitation of residues in situations where
     GC is either not applicable or not convenient to use.  An HPLC retention
     time can serve as evidence to confirm GC in the same way as a PC or TLC
     Rp value.  The liquid chromatographic system should be carefully chosen
     to be "different" from the GC system (i.e., adsorption rather than
     partition), and the independence of the data must be clearly established
     if it is desired to use both PC or TLC and HPLC data for confirmation.
     The variable wavelength UV detector allows determination of the wave-
     length of maximum absorption for each pesticide.  Detection of HPLC
     effluents with a Cl-selective electrolytic conductivity detector (13)
     can also provide useful confirmatory evidence.
                             f  •
10F    EXTRACTION £-VALUES

     Extraction £-values (14-18) are a tool for identifying, pesticides at the
     low ng level.  The £-value is determined by equilibration of a solute
     between volumes of two immiscible liquid phases, followed by the analysis
     of one of the phases for the solute.  The £-value, defined as the
     fraction of total solute partitioning into the upper phase, can be
     derived from a single distribution between the solvents or from a
     multiple distribution, as in counter-current distribution.  ^-Values
     for most pesticides are appreciably different from those of normal
     co-extracted contaminants.  The determination of these values is
     simplified since only relative, rather than absolute, data are required,
     and sensitivity is at or only slightly above the level of EC-GC.

     Details including experimental procedures, formulas for calculating
     ^-values and the fractional amount extracted after repeated extractions,
     graphs for determining specificity in a given system, and ^-values for
     131 pesticides in six binary solvent systems (hexane-90% DMSO, heptane-
     90% ethanol, isooctane-80% acetone, hexane-acetonitrile, isooctane-DMF,
     and isooctane-85% DMF) are given in Section 621 of the FDA PAM, reference
     (15), and Section 12,C of the EPA PAM (data for 88 pesticides in the
     latter).  A device and method for determining ^-values with unequilibrated
     solvents or unequal phase volumes are given in the FDA PAM, Section 622.1
     and reference  (18).

     As mentioned earlier, the general technique of determining £-values has
     much in  common with the use of several GC columns, PC, and TLC in identi-
     fication studies since all systems may share the same partition mechanisms.
     Unless the analyst assures himself that the data are not correlate4, it is
     best to use either a PC c>r_ TLC Rp value or an extraction p_-value as one
     independent criterion of identity.  The great advantage of ^-values over
     PC or TLC is that the method is useful at levels amenable to quantitative
     analysis by EC-GC where sufficient residue might not be available for
     either of the  former techniques.
                                         -391-

-------
                                                             Section 10G
10G    DERIVATIZATION (CHEMICAL REACTION) TECHNIQUES
     Derivatives of pesticides are prepared for various reasons, such as to
     decrease volatility or increase detectability for HPLC or TLC; to
     increase volatility, stability, and/or detectability and avoid tailing
     peaks for gas chromatography; for removal of interferences in residue
     analysis (19); and to alter the structure to aid characterization.
     It is this latter topic that will be discussed in this subsection.

     Comparison of retention times on a given GC stationary phase before and
     after chemical derivatization is a relatively r.ecent innovation that
     is becoming increasingly important for corroboration of residue identity.
     Desirable characteristics of any chemical derivatization technique  include:

          a.   A specific product should be formed with at least as much  or
     more response to electron capture, or to other detection, compared  to
     the parent pesticide.

          b.   The product should have a different retention time than the
     parent,  preferably greater to differentiate it clearly from the background.

          c.   Reactions should be essentially quantitative, they should  use
     highly pure reagents and solvents, and they should be facile and rapid.
     Reagents 'and equipment should be inexpensive, if possible.

          d.   A cleanup method should be available to remove any background
     interferences introduced by the reaction.

          e.   If product structures and reaction mechanisms and limitations
     are known,  misidentifications can be avoided because the analyst can
     elucidate the extent and probable sources of error in the procedure.

          f.   Sensitivity should be at least in the 0.01 to 0.1 ppm range in
     terms of the parent pesticide, which is lower than the established
     tolerance values for most pesticides.

          g.   The same reaction should occur, and to the same degree, in
     both the sample extract and in a solution of the reference material of
     the suspected compound at the same concentration.   Matrix effects can
     play an  important role in the applicability of chemical derivatization
     for quantitation purposes.   A reaction might work very well for pure
     standards but may fail when applied to samples due to the effects of
     sample components.

          h.   The reaction should be safe to perform.


     Derivatization reactions for gas chromatography are usually carried out
     in solution, on the surface of a solid matrix,  or in a GC precolumn.
     Reactions in solution on a microscale are most common for residue level
                                        -392-

-------
                                                        Section 10G


work.  The reaction usually involves heating of the reactants in a small
sealed tube, after which the derivative is dissolved in a suitable solvent.
If direct injection into a gas chromatograph is not possible, cleanup
by solvent partitioning and/or column chromatography and concentration
steps may be applied.  Solid matrix reactions are generally carried
out by introduction of dissolved compound onto a microcolumn composed
of solid support  (e.g., alumina) mixed with reagent(s).  After a
specified reaction time, solvent is added to elute the derivative for
GC determination.  The advantages of this approach are simplicity,
reduced glassware needs, and ability to react many samples simultaneously.
However, the same derivative as formed in a solution reaction is nbt
always produced and/or eluted from the column in a solid matrix reaction
with the same  active  reagent.  GC precolumns are usually composed of a
reagent-solid  support mixture located in a heated area ahead of the
'analytical.column.  The sample is injected into the precolumn, and the
derivative is  formed  and swept by the carrier gas onto the analytical
column for determination.   Speed of operation is the greatest advantage
for  those reactions that are rapid enough to be feasible by the precolumn
technique.  The chromatograph is best fitted with a special injection
apparatus so injection can be made into the precolumn  for derivatization
or directly into  the  analytical column for normal operation.  In addition
to these three types, some derivatizations may be carried out in a hot
injection port or on  the analytical column itself.

Derivatization aimed  at  increasing detectability in HPLC  is usually  carried
out  "post column",  or after separation of the parent molecules rather  than
the  derivatives.   This  is  accomplished by inserting a  mixing  chamber at
the  end  of  the column and  pumping  in  reagent  to mix with  the  column
effluent.   The derivative  is formed  in a reaction coil, and  is measured
 subsequently in a suitable detector.  This  type of derivatization  can  be
easily automated  for routine analysis.  Derivatization for  the purpose
of providing detection  in TLC is  generally  carried out by spraying  chemical
 reagents,  also "post chromatography".  For  both of  these  liquid  chromatog-
 raphy techniques, derivatization for confirmation  of  identity is usually
 done in solution or "on column" (or "on-layer",  for TLC),  prior  to
 chromatography, as is the case for GC.

 The following subsections review some procedures for  confirming residue
 identity by chemical derivatization.   Table 651-A of  the FDA PAM contains
 an extensive further listing of derivatization methods for more than
 100 pesticides and related compounds of many chemical types,  including
 comments on the  level of applicability,  yields, and 60 references to the
 original papers.  A  review paper on chemical derivatization for GC and
 HPLC has been published (20).

      a.  Organochlorine Pesticides

          Most of the effort to date in the development of confirmatory
 derivatization tests has been confined to the organochlbrine insecticides.
 For these compounds, addition, oxidation, epoxidation, rearrangement,
                              -393-

-------
                                                        Section 10G
dechlorination, hydrolysis, reduction, and dehydrochlorination'are the
most commonly used reactions.  Examples of specific tests are shown in
Tables 10-1, 10-2, and 10-3, as reviewed through 1978 by Cochfane
(20-22).  Table 10-4 lists selected references for OC1 pesticide
derivatization methods published since 1978.  Section 9A,G,e discusses
confirmation reactions suitable for PCB-pesticide mixtures.  It must
be realized that these reactions destroy some pesticides (and Artifacts)
in addition to forming pesticide derivatives.
  TABLE 10-1
                   CONFIRMATORY DERIVATIZATION TESTS FOR
                  PESTICIDE AND METABOLITE RESIDUES (21)
           Pesticide
       DDT

       DDE
       DDD
       Methoxychlor
        Reaction
(a)  Dehydrochlorination
(b)  Dechlorination (of
Oxidation
Dehydrochlorination
Dehydrochlorination
                          ' -isomer)
       Aldrin
       Dieldrin

       Endrin

       Endosulfan

       Heptachlor
       Heptachlor epoxide
       cis- and trans-Chlordane
       Nonachlor
 (a)  Addition^-— Br2
                ^ tert-BuOCl
 (b)  Epoxidation
             cleavage
 Epoxide ^^— rearrangement
             acetylation
 (a)  Epoxide rearrangement
 (b)  Dechlorination
 Sulf ite reduction
                  acetylation
 (a)
     Ally lie ^- — hydroxylation
                 dechlorination
(b)  Addition
(c)  Epoxidation
Epoxide rearrangement
Dehydrochlorination
(a)  Dechlorination
(b)  Dehydrochlorination
                                  -394-

-------
TABLE 10-1 (Continued)
                                                            Section 10G
      Parent Pesticide
      Heptachlor
      trans-Chlordane
      cia-Chlordane
     cisj- and trans-
        Chlordane
      Endrin
      Endosulfan
  Metabolite
(a)  Chlordene
                        (b)  1-Hydroxy-
                             chlordene
   Group Reacted

(1) Allylic hydrogen
(2) Double bond
(1) Allylic hydrogen
(2) Double bond
(c)  l-Hydroxy-2,3-  (1) Hydroxyl
     epoxychlordene
                     (2) Epoxide

2-Chlorochlordene   iDouble bond or
3-Chlorochlordene   (. gem-dichloro group
1,2-Dichloro-        Chloro epoxide or
chlordene epoxide    gem-dichloro group
Photo-endrin         gem-Pichloro
Endosulfan diol      Hydroxyl
  Derivative-

 1-Bromochlordene
 Chlordene epoxide
 Silyl ether
 Chloroacetate
 Epoxide
 Silyl ether
 Trihydroxy
 chlordane
/Epoxide or
Ihexachloro
 Chloroacetate
 or heptachloro
 Pentachloro
 Acetate or silyl
   ether
                                  -395-

-------
                                                           Section 10G
      TABLE  10-2
        CONFIRMATORY TESTS FOR ORGANOCHLORINE PESTICIDES C22)

    Pesticide Class             Reagent or Reaction Type
 General


 Hexachlorobenzene (HCB)


 BHC  isomers

 Cyclodiene insecticides
Mirex

Kepone


PCBs

Chlorobiphenyls and PGP

DDT
CrCl2 reduction (26)1
KOH denydrochlorination (46)1

Base/alcohol
KOH hydrolysis/diazomethane

NaOMe/MeOH or GC alkaline precolumn

Comparison of 3 methods (D)
10 various reactions (D)
BCl3/2~chloroethanol (D/E)
UV irradiation (D/E/H)
H2S04 or 60% KOH (E/M)
_t-BuOK/£-BuOH or CrCl2 (E/M) '
Acid or "base-Al203 microcolumn (C/E/H/T/M)
Base-catalyzed intramolecular cyclization
Silylation/acetylation (T/M)

UV dechlorination

KOH/esterification
LiAlH4/PCl5

SbCl5 perchlorination

Acetylation and butylation

Reduction and/or oxidation
•^•Figures in parenthesis indicate the number of pesticides studied,
 letters indicate the particular pesticide(s) confirmed
C * chlordanes, D = dieldrin, E = endrin, H = heptachlor,
T » Thiodane (endosulfan) and M = and metabolites
                                  -396-

-------
TABLE  10-3
                             Section  10G
          CHEMICAL  DERIVATIZATION OF ORGANOCHLORINE,  PCB,  PBB,
                        AND  RELATED  COMPOUNDS3   (20)
                            Pesticide/
                             Compound
Derivatisation
Procedure
    Substrate
                            PCB/OC        EihanoKc KOH
                                          EthanottoKOH/
                                           CrO3
                                          Photoisomerisa-,
                                           lion (OCs)
                                          Photo-dechlprin-
                                           ation(PCBs)
                                          "MgO micro-
                                           reaction"

                            PCB/OC/chlorin-
                             ated paraffins  Photolysis
                           PCBs
                           Hydro.xylated
                             PCBs

                           PCB/Mirex
                           Mirex



                           Kepone

                           Kelvan


                           PBB

                           HCB


                           Endosulfan
                           Heptachlor and
                             Epoxide
                           Lindane

                           Toxaphene

                           Polychloro
                             naphthalenes

                           TCDD
TiO2 photode-
  chlorination
Perchlori nation
Silylation

Photolysis


Photolysis
Hematin
  dechlorination

Chlorination

Photolysis or
  oxidation

Photolysis

2-propanol/KOH
              Fish and fish prod.

              Fish, serum

              Environmental
                samples


              Fish


              Fish
                                                        Aqueous media
                                                        Fish
Photolysis
Acetyiation

Photolysis
NiCl2/NaBH4
Acetyiation
Human tissue
  and milk

Urine, feces.eggs

Model system

Blood, oyster


Potatoes

Feeds, dairy prod.

Adipox tissue,
  human milk


Soils

Foods

Soils
MethanolicKOH  —
Photolysis

Photolysis
Silica, soil
     PCB  = polychlorinated biphenyl;  PBB »  polybrominated biphenyl;

     HCB  = hexachlorobenzene;  TCDD  = tetrachlorodibenzodioxin;

     OC = organochlorines.

                                      -397-

-------
                                                          Section 10G
   TABLE 10-4
               CONFIRMATORY DERIVATIZATION REACTIONS FOR OC1

                       PESTICIDES PUBLISHED SINCE 1978
Compounds Studied
     Reagent (sensitivity)
Reference
Chlordane and mirex
Mirex and PCBs

 (in fish)

OC1 pesticides and PCBs

 (Harp seal tissue)

Chlorophenoxy acid

    herbicides
Cone. ^SO^-fuming HNOg .(1:1 v/v)

to remove PCBs and other

interferences

Reduction of mirex with

chromous chloride

Dechlorination using sodium

ethoxide

Pentafluorobenzyl bromide

  (10 ppb in urine)
    23
    24
    25
    26
  A confirmatory technique related to chemical derivatization is ultra-
  violet degradation or photolysis (27, 28; Table 652-A of the FDA PAM).
  Degradation products arising from UV treatment of chlorinated insecti-
  cides and detected by EC-GC can provide identification of these pesti-
  cides (28) at 75-100 pg levels.  Depending on the length of irradation
  (often ca 10 minutes), all of the parent pesticide may not be degraded.
  Solvent and sample blanks should be run to prove if background is reacted
  as well.  Isooctane is a good solvent because it is little affected by
  UV light.

  Section 12,D,(1) of the EPA PAM gives details of a microscale alkali
  dehydrochlorination method for use in multiresidue analysis.  This
  procedure produces derivatives for identity confirmation and provides
  supplemental cleanup for some troublesome extracts after Florisil
  chromatography.  Section 651.12 of the FDA PAM describes the micro-
  scale alkali treatment method that is"part of the AOAC official method
  for perthane.  Table 651.1 lists the behavior of about 40 compounds
                                     -398-

-------
                                                        Section 10G
under these reaction conditions.  Alkali reactions carried out on a GC
precolumn rather than in solution have proved advantageous in some
instances (29).  Section 5,A,(l),Cb) of the EPA PAM describes the
confirmation of HCB in fatty tissue by formation of the disubstituted
ether derivative bis-isopropoxytetrachlorobenzene (30).

Section 11 of the Canadian PAM gives complete details  for the following
tests:
            Pesticide(s)

     £,£f-DDT, £,£!-DDT, £,£f-TDE,
     methoxychlor

     £,£'-DDT, endrin

     Dieldrin, endrin

     Chlordane, heptachlor
     epoxide

     Aldrin, heptachlor,
     £»£f -DDE

     Aldrin

     Endosulfan

     Chlorophenoxy acid
     herbicides

     Captan
     Reagent

Sodium methylate


Chromous chloride

BC'l- in 2-chloroethanol

K-tert butoxide/tert-butanol,
silylation

Chromic acid .


m-Chloroperbenzoic acid

Alcoholic KOH
jtt-Propanol


Resorcinol
A special two stage, mixed phase 180 cm column consisting of 165 cm of
4% OV-1/6% QF-1 and 15 cm of 3% OV-1/6% OV-225 at the injector end is
recommended in the Canadian Manual for resolving HCB, BHC isomers, sulfur,
and aldrin for confirmatory purposes, because they are not resolved on
the 4% SE-30/6% QF-1 working column.

One method of differentiating PCBs from organochlorine pesticides is
by treating the residues with a non-fuming HN03-H2S04 mixture.  Organo-
chlorine pesticides are destroyed, whereas"PCBs (and toxaphene) are
unaffected.  Various confirmation methods for PCBs are covered in Sub-
section 9A,G,e in Section 9 of this Manual, and perchlorination is
described in Section 12,D,(2) of the EPA PAM.


     b.  Other Pesticide Classes

         Residues of organophosphorus pesticides may be confirmed by
alkaline hydrolysis followed by esterification of the resulting dialkyl
phosphates to trialkyl phosphates (31).  This procedure does not distinguish
pesticides that produce the same hydrolysis product.  According to McCully (32),
                                   -399-

-------
                                                                 Section 10G


  the three most  practical methods  for confirmation of OP  pesticides  are
  oxidation to oxygen analogs (33),  pentafluorobenzyl bromide derivatiza-
  tion of hydrolyzed phenols or thiophenols (34),  and chromous chloride
  reduction (35).   The sodium hypochlorite  oxidation method has the widest
  applicability,  but it  suffers from low  sensitivity, difficulty in
  analyzing the analog products,  and inability to  distinguish analogs
  originally present in  samples.  The CrCl2 method is simple but applicable
  only to OP pesticides  containing  a nitro  group.   The pentafluorobenzyl
  bromide procedure is intermediate in scope.   These and other reactions
  used to identify organophosphorus pesticides are listed  in Table 10-5,
  along with information on triazine, carbamate, and urea  pesticides.
  This table is from a review article (22)  that gives the  original
  references for  these reactions.   Table  10-6  contains a selection of
  more recent references.   The pentafluorobenzyl bromide derivatization
  procedure for OP pesticides is  being collaboratively evaluated on
  different substrates (36).

  Triazine herbicides have been confirmed by silylation, methoxylation
  (in sodium methoxide-methanol), methylation  (CH3l-NaH),  and hydrolysis-
  DNFB reactions  (37, 38),  and linuron has  also been confirmed by alkylation
  (with alkyl halide - NaH) (37).
 TABLE 10-5
       CONFIRMATORY TESTS FOR ORGANOPHOSPHORUS, TRIAZINE, UREA, AND CARBAMATE COMPOUNDS (22)
 Pentlcid. Clasa
 Orgftnophospluccs
 Triatinea
C«rban»te» and
ureas
Chlorophonoxy
•elds
                     Compound Type
                                                         Confii
!    General
    Phenol-generating compounds
    Aryl-H0_ and Aryl-CN
    containing compounds
 d)  P™ S compounds
 e)  -Nit *nd -Nllg containing oompounda

 f)  OH compounds (diatimm metabolites)
 g)  Crufornate

 a)  Chloro-a_-trlaEines
 b) Hydroxy-s-triauineB


 c) Cyanazlne and Metabolites

 a) Intact compound


 b) Phenol-generating compounds
                    o)  Amlne-gonerating compounds
a) Esters
                                                         Hydrolysla/mothylatlon
                                                         HyUrolyalB/PFD ether formation
                                                         Reduction (CrClg.PdCl.,,  Zn/HCl)

                                                         Oxidation (to P — 0)
                                                          i)Alkyation (Hall/Mel/DMSO)
                                                         ii)Ueainlnat Ion/methyl a tion
                                                         Silylation or alkylation
                                                         UV dechlorination
                                                          1) Alkylation
                                                          '  Silylation
                                                             Methoxylation
                                                             HydrolyslB/DNP formation
                                                             Sllylation
                                                             Alkylation
                                                             Chlurination  '
 li
 ill
 iv
  1
 11
 ill
                                                        Acid catalyzed cyollcation
                                                         11
                                                        ill
     Acetylation
     Silylation
     Alkylation
 Iv Perfluorlnatlon»
  1 Bromination
 11) Chloroacetylatlon
ill) Thlophosphorylation
 ivi Silylation
  v) Dichlorobenzene eulfonylation
 •'  DNT/DNP
    Pentafluorobaneylation
    Tn^l «n kl _..
                                         ___________
                                         lodinatioii
                                         Dromination
                                         -Bromobenzoylation
                                                            p-Br
                                                            2^^-
Vil
  1
 11
ill
 iv  _.
  v) DNT
 vi) PentafluorobenKylatlon
    (Amines in general)

  1) Transesterlfloation
 ii) -   '   '
                                                         -., Dromlnation
Mncludea trifluoroaeetylatlon,  pentafluoropropylation. and  heptafluorobutylatlon
                                  -400-

-------
       TABLE 10-6
                                                                   Section 10G
                    CONFIRMATORY DERIVATIZATION REACTIONS FOR PESTICIDES
                          OF VARIOUS CLASSES  PUBLISHED SINCE 1975
     Compounds Studied
    Reagent or Derivative (Sensitivity)
Reference
 -N0_-containing herbicides
   and fungicides

 Organonitrogen fungicides
   and herbicides

 Carbofuran and metabolites
 OP pesticides

 Sulfoxide-containing
   pesticides

 Carbaryl
 Carbamate insecticides

 Dimilin (TH 6040)

 Thiabendazole

 II-Aryl carbamates


 S-Containing carbamates
Carbamate  and urea
  herbicides

Dimilin  (TH 6040)
Azodrin

OP pesticides


s-Triazines
NT-Methyl carbamate
  insecticides and
  metabolites
 CrCl3 reduction to -NH2 followed by CCD-GC      (39)
   (0.5-1.0 ppm)

 Alkylation, methoxylation, trifluoroalkyla-     (40)
   tion (0.1 ppm)

 Heptafluorobutyric anhydride plus tri-          (41)
   methylamine catalyst (10 pg)

 In-block methylation with THAM (yg levels)      (42)

 Trifluoroacetic  anhydride (1 ppm)               (43)
 N-mono- and trichloroacetyl,  and IJ-nitroso      (44)
   derivatives

 Heptafluorobutyryl derivatives  (0.1  ppm)         (45)

 Trifluoroacetyl derivative  (0.02 ppm)            (46)

 Pentafluorobenzyl  chloride  (0.01 ppra)            (47)

 Flash heater reaction with  trimethylanilinium   (48)
   hydroxide (ng levels)

 Trimethylphenylammonium hydroxide injected      (49)
   with compound into gas chromatograph
   (20 ng)

 Alkylation  by NaH/CH3I (0.1 ppm)                 (50)


 Conversion  to N.N'-dimethyl analog with          (51)
   NaH/CH3I  (0.25 ng)

 Trifluoroacetylation (2 ppb)                     (52)

 Oxidation with neutralized NaOCl                 (53)
 (0.25-0.5 ppm in fruits and vegetables)

 Trifluoroacetic acid plus a silylation           (54)
   reagent (20 pmol)

Post-column HPLC fluorometric labeling           (55)
  with o-phthalaldehyde after basic
  hydrolysis
                                       -401-

-------
                                                              Sections  10H,  101
10H    SPECTROMETRY  (SPECTROPHOTOMETRY)
     Spectrophotometric methods  for residue  determination  (quantitation)
     usually are not as sensitive  or selective  as GC or TLC, and  for this
     reason they are.not as widely used  as in the early days of pesticide
     analysis before chromatographic methods were developed.  The appli-
     cability of spectrometry  is especially  limited for multiresidue determina-
     tions or analyses of a parent compound, metabolites,  and hydrolysis
     products.

     Spectrometry can be very  valuable,  however, in conjunction with chroma-
     tography as a confirmatory  tool, and it is this aspect that  will be
     stressed in the following subsections.


101    VISIBLE, UV, FLUORESCENCE,  AND PHOSPHORESCENCE

     Very few pesticides are naturally colored, so a chromophoric group must
     be formed by a reaction.or  added through derivatization before most
     pesticides can be measured  in the visible 'spectral region.   The colori-
     metric method then becomes  specific to the color forming 'group involved.
     The inferior sensitivity  of direct and indirect visible spectrophoto-
     metric methods limits their usefulness for confirmation in human and
     environmental monitoring where residues are generally present at low
     concentrations.

     The correlation between UV spectra and pesticide structure and the
     usefulness of'UV spectrophotometry in confirming identification have
     been reviewed (56).   Spectra-structure correlations can be of value to
     the analyst in identifying chromophores and therefore making confirma-
     tions,  especially in conjunction with spectral information obtained by
     other methods,  such as IR, NMR, and MS.   In some cases, extinction
     coefficients (absorptivities)  are sufficiently large to permit identifica-
     tions at  submicrogram levels.   If a suitable absorption wavelength of a
     pesticide can be chosen that  is free of interference from contaminants
     or solvents,  UV spectrophotometry can be performed directly without
     sample purification and at a greater saving of time.   However,  because
     absorption of UV energy is quite common for most organic compounds,
     rigorous  cleanup may be required to remove any interferences that  can
     absorb  in the spectral region where the pesticide will be measured.  The
     transparency of  many functional groups (and often large segments of
     complex molecules)  in the near UV spectra'l range imposes a  limitation
     on interpretations  of absorption bands in this  region.  Solvents must be
     carefully chosen to  be transparent at the wavelengths  absorbed  by  the
     pesticide.  UV absorbing groups can be added by chemical derivatization
     methods,  and  this procedure has been used to detect pesticides  by  HPLC
     UV detectors  and by  TLC.   UV spectra of  76  reference pesticides have
     been published  (57).   Visible and UV Spectrophotometric methods for
     pesticides  have  been reviewed (58),  including development of  color by
     azo coupling  and IT complexing and recent instrumental  developments.
                                        -402-

-------
                                                           Section 10J


   If a pesticide is naturally fluorescent or can be made fluorescent by
   derivatization, fluorescence spectrophotometry is likely to be more
   selective and sensitive than either visible or UV absorption methods.
   Concurrence of fluorescence excitation and emission spectra between
   samples and standards, recorded either in solution or directly on
   thin layer chromatograms, has served as' a valuable confirmatory aid
   for certain pesticides.  Fluorescence characteristics are dependent
   on a number of experimental conditions which must be closely controlled,
   e.g., solvent and pH effects (59).  Removal of naturally occurring
   fluorescent interferences from biological samples can-pose serious
   cleanup problems.  Fluorescence and phosphorescence methods for pesti-
   cides have been reviewed by Argauer (60), and the phosphorimetry of
   pesticides has been reviewed by Baeyens (61).


10J   INFRARED (IR)

   IR spectroscopy with micro sampling techniques is generally sensitive
   at the 1 yg level but has been used as low as the -0.1 yg level in some
   applications.  It is thus considerably less sensitive than GC or TLC
   and cannot be used unless enough sample is available to provide a
   sufficient concentration of pesticides for IR observation.  Sample
   extracts require a stringent cleanup procedure (e.g., partition plus-
   column adsorption chromatography) plus additional purification either
   by GC or TLC.  Thin layer spots are scraped and collected, and the
   pesticide is eluted from the adsorbent with an appropriate solution.
   Fractions can be collected from a gas chromatograph equipped with a
   stream splitter:  a small percentage of the effluent stream goes to
   the detector for monitoring purposes while the remainder goes to a
   collecting device.

   Potassium bromide (KBr) micro-pellet techniques using a pellet of 1-2 mm
   diameter are described in Section 12,E of the EPA PAM.  These methods
   were developed by R. C. Blinn of the American Cyanamid Co.  The key to
   their sensitivity is the ability to transfer the maximum amount of pesti-
   cide to a very small amount of KBr to be pressed into the micro-pellet.
   The equipment commonly used by Blinn for preparing micro-pellets by the
   syringe method is shown in Figure 10-E, and the technique for transfer
   of the sample-KBr mixture adhering to the syringe needle is pictured in
   Figure 10-F.  The method using a commercially available "wick stick"
   may be the most reliable and foolproof for preparing micro KBr pellets
   (Figure 10-G).  The sample is applied to the wedge of potassium bromide,
   which is then dipped at its base into a volatile solvent.  The solution
   migrates up the wedge to the tip where the solvent evaporates, and the
   compound becomes concentrated at the tip.  The tip is then cut or broken
   from the wedge and pressed into a micro-pellet.  A procedure using micro
   pellets of 0.5 mm diameter to measure 1.4 ug of DDT has been reported (62).
                                  -403-

-------
                                                        Section 10J


The Blinn microtechniques are sensitive and reliable, but considerable
experience is required to prepare pellets with a minimum of contamination.
They require the availability of a modern IR spectrophotometer including
a beam condenser and microcells.  Contamination from such sources as
the sample, solvent, reagents, atmosphere, and handling is their major
source of error.  The same "amounts of interferences that would be
inconsequential for macro-sampling techniques become a significant
percentage of a micro sample and contribute to the spectrum.  Clean
gloves should always be worn when preparing micro-pellets, and only
purified solvents and reagents and carefully cleaned equipment should
be used.  Inevitable losses due to handling and processing require that
the isolation procedure be started with sufficient sample to finally
achieve a useable spectrum.

Another method that is in effect in a micro-sampling technique involves
scale expansion, or electronic amplification of the signal from the
spectrometer.  This method increases pen response without an increase
in the sample concentration, but the response to all interferences and
electronic noise is increased as well.  All sources of interference
must, therefore, again be minimized.

IR microtechniques have been reviewed by Blinn (63), including discussion
of micro multiple internal reflectance.  Advantages of internal re*-
flectance include ease of applying (by dotting or streaking) sample to
the surface of the reflectance plate (crystal), minimizing of inter-
ferences from handling and reagents, and ease of recovery after IR
evaluation (samples made into pellets are essentially lost for further
scrutiny).  A disadvantage is lowered sensitivity compared to the KBr
micro-pellet method.  Sensitivity is increased by spreading a very thin
film of sample over the effective sample area of a very thin reflectance
plate.  The multiple reflectance method has been applied to the identifi-
cation of Thiram residues at 0.1 ppm on lettuce after extraction, Florisil
chromatography, and TLC (64).  An alternative micro-KBr technique with
sensitivity levels similar to the method in the EPA PAM is detailed in
the FDA PAM, Section 631.
                                    -404-

-------
                                                    Section 10J
Figure 10-E.   Equipment for preparation  of micro KBr pellets
               (photo courtesy of R. C. Blinn)
Figure  10-F.,  Illustration of technique of Curry  et al.
             '  (Fhoto courtesy of R.  C.  Blinn)
                    A,S Curry,, J F Read, C Srtwff, on4 *W Jftc*io$.
                                  -405-

-------
                                                         Section  10J
   Figure 10-G.   Illustration of wick-stick method
            STAINLESS STEEL
               HOLDER
                                          STAINLESS
                                          STEEL CAP
                                        	SAMPLE CONCENTRATED
                                                 AT TIP

                                           GLASS VIAL
                                           «8r "WICK-STICK"
                                           VOLATILE'SOLVENT
The FDA Manual  (Section  632),also  gives details of a qualitative micro
procedure for collection of  GC  fractions directly on powdered KBr fov
IR confirmation.  Interpretation of  IR spectra from collected fractions
must take into  consideration the stability of the pesticide of interest
to GC conditions.  The analyst  should  be sure he is measuring the spectrum
of unchanged pesticide rather than of  a degradation product.   In addition,
the specificity of the GC detector will often obscure elution of inter-
fering materials from the GC column, so that  a fraction presumably con-
taining isolated pesticide could be  totally unsuitable for IR characteriza-
tion.  These interfering materials might be from the sample substrate or
bleed or breakdown products  from the stationary phase of the column
packing.  The column exit line  should  be heated at least to column
temperature to  the point of  trapping,  otherwise condensates resulting
from previous samples may contaminate  the trapped compound.   Use of
splitters at the column  exit  is usually necessary'because of the high
sensitivity (detectors would  be overloaded by the pg quantities  for
IR)  and the destructive  nature  of  pesticide detectors.
                                    -406-

-------
                                                        Section 10J


Several different types of IR detectors directly coupled with gas chroma-
tographs (65) have become available commercially but have not proven
especially useful for pesticide residue work because of various dis-
advantages.  Trapping procedures have been used almost exclusively,
including the following methods:

     a.  Passing column effluent through solvent (66, 67).

     b.  Condensing effluent on a micro sodium chloride plate.

     c.  Condensing effluent on a thermo-electrically cooled capillary
plate for internal multiple reflectance IR.

     d.  Trapping fractions on column packing (68-70).  This procedure
is very efficient, and fractions are easily collected for subsequent
IR evaluation; reagent interferences are possible.

     e.  Collecting on a TLC plate for further cleanup prior to IR (71).

     f.  Trapping on Millipore or siliconized filter material (72, 73).

     g.  Using vatious types of liquid nitrogen or dry ice cold surface
traps (74).

     h.  Using a cool or cold small internal diameter tubing at the
GC vent (75).

     i.  Trapping directly on KBr powder supported by pipe cleaner inside
capillary tubing (FDA PAM, Section 632).  This procedure is probably the
most sensitive of any, tubes can be changed for each peak, and the
technique is free of sources of interferences.-


The choice of trapping procedure will depend on the amount of compound
available, IR technique to be used, purity of the compound eluted from
the GC column, and equipment available to the chemist.

IR spectra of over 400 reference pesticides have been published (76)  to
aid the analyst in matching spectra of unknown pesticides.  The ASTM
FIRST-1 computer search program (65) and similar computer retrieval
systems aid in matching sample and reference spectra when standards
cannot be easily chosen for a manual point-by-point comparison.
Reference Raman spectra of OC1,OP, and carbamate pesticides were
published (77).  Giang has compiled a bibliography with 855 references
published to 1976 on the use of IR spectrophotometry in pesticide analysis
(78).

An important recent development in IR analysis that is capable of sensi-
tivity at subnanogram (79) residue levels is the Fourier transform (FT)
or interferometric method.  In FT-IR, a Michelson interferometer is
used instead of the prism or grating and slits in a conventional spectro-
meter.  The slitless spectrometer has an advantage in energy throughput,
                              -407-

-------
                                                            Section 10K


   in addition to the so-called Fellget's or multiplex advantage that allows
   all wavelengths to be detected simultaneously throughout the spectral
   range.  The signal to noise ratio increases with consecutive accumulation
   of scans and is proportional to the square root of the number of scans.
   Since each scan requires only a few seconds and instrument stability is
   high, many cumulative scans can be made on each sample.  The fast scan
   capability is ideal for on-the-fly IR detection of GC effluents.  FT-IR
   spectroscopy has at least an order of magnitude greater resolving power,
   greater wavelength accuracy, and a greater scan range than does con-
   ventional dispersion IR spectroscopy.  There is also a much smaller image
   in the sample compartment without any special measures, making FT-IR
   ideal for microsamples.  A dedicated minicomputer, in addition to the
   basic FT-IR optical equipment and detector, is required to collect, process,
   and store the data.  FT-IR methodology and equipment have been reviewed
   (65).  There is no doubt that much use will be madg of FT-IR spectroscopy
   for pesticide determination and confirmation as the principles, techniques,
   and instrumentation become more familiar.
10K   NUCLEAR MAGNETIC RESONANCE (NMR)

   NMR spectroscopy has had only limited application in residue analysis
   "because of its low sensitivity relative to other analytical methods,
   e.g., GC-MS, IR, and UV.  Despite this drawback, it is one of the most
   valuable tools available for structural analysis and identity confirma-
   tion.  Current pulsed Fourier transform NMR spectrometers (80) allow
   routine acquisition of useful data on as little as 10 yg of a proton NMR.
   sample in a few minutes of experimental time.  The NMR sensitivity of
   13C is lower; with current commercial instrumentation, a practical sample
   size is greater than 20 mg, although 13C spectra of as little as 300 ug
   have been obtained on modified instruments (81) .  Useful information is
   provided by NMR in many areas relevant to the analysis of pesticides,
   their metabolites, and degradation products,  such as identification and
   structural characterization, molecular geometries, conformations and
   stereochemistry, chemical kinetics and equilibria, complex formation
   and binding, and electronic charge distributions.

   Residues of £,£f-DDT and p^ja'-DDE isolated from adipose and liver tissue
   samples have been analyzed by NMR (82) ,  with  semi-quantitative determina-
   tion of the relative concentrations of the pesticides.   Included in NMR
   studies of the metabolism, binding, and  degradation of pesticides are
   % spectra useful for identification of  £,£'-DDT (83,  84),  £,_p_'-DDA (85),
   aldrin and dieldrin derivatives  and other chlorinated pesticides (86) ,
   rotenoids (87), and dithiocarbamates (88).  Other IH reference spectra
   of organophosphorus (89), diphenylme thane (DDT type)  (90),  and carbamate
   (91)  pesticides have been published and  are useful for identity confirma-
   tion.   The application of NMR to pesticide analysis has been reviewed
   (80,  92, 93).

   Carbon-13 NMR spectra have been  published for a-BHC (94) ,  for several
   chlorinated biphenyls (95, 96),  and for  30 chlorinated polycyclodiene
   pesticides (97).   Studies of technical chlordane components (98),  mirex
   (99,  100),  and Kepone and its photo-products  (101)  have provided
                               -408-

-------
                                                      Section 10L
NMR data useful for confirmation and structural characterization.  Chlorine
nuclear quadrupole resonance spectrometry has been used to study the
structures of several chlorinated pesticides including BHC, aldrin,
endrin, endosulfan, and dieldrin (102-104). 31P- NMR chemical shifts
have been correlated with structures of some organophosphorus pesticides
(105), and 31P Fourier transform NMR has been used for the determination
of malathion at ppm levels (106).
10L  MASS SPECTROMETRY (MS)

The mass spectrometer is a very sensitive spectroscopic tool for pesticide
residue analysis, providing useful data on ng or pg residue levels.  Ions
are produced from neutral sample molecules and are then sorted according
to their mass-to-charge ratio (m/z).  The mass spectrum is a record of
these different ions and their relative abundance.  A mass spectrum is
usually quite characteristic of an individual pesticide, sometimes even
providing data that will differentiate geometric isomers.  Pesticide
identifications can be made by matching the mass spectrum of an unknown
sample with the mass spectrum of a known material.  This comparison is
especially valuable because it is based on many different peaks character-
istic of the unknown compound.  The composition of. an unknown compound can
be obtained without comparison to a reference material by making exact
mass measurements of the molecular ion and, other key fragment ions in
the spectrum.  Because the exact mass of every element has a unique
fractional value on a scale compared to 12C = 12.0000, any combination
of these elements into a chemical formula will have a unique fractional
mass, specific for that combination of elements.  The exact masses used
to determine the chemical composition of a compound can be obtained on
either a low or high resolution spectrometer.  The advantage of a high
resolution instrument is the ability to separate ions with different
compositions that are at the same nominal mass (e.g., m/z 28 for CO, N2,
C2H4) and to obtain accurate mass values for these ions.  The molecular
ion is the species .resulting from the removal of a single electron from
a molecule.  After the recommendation of Benyon (107), the symbol M   is used
to represent the odd-electron molecular ion formed from an even electron
molecule.
     a. MS Instrumentation and Operation

       (1)  Introduction

            Five components are common to most mass spectrometers: the
inlet system, the ion source, the mass analyzer, the detector, and the
readout system.'  In ad'dition, a vacuum must be maintained throughout the
spectrometer from inlet to detector so that ions formed in the source will
not be lost from collisions with atomspheric gas molecules.  A second
reason for maintenance of vacuum is to prevent oxidation of the filament
in the ion source and various other inside parts of the spectrometer and
                                -409-

-------
                                                         Section 10L

 electron multiplier.   A sample is introduced via the inlet system into the
 ion source,  where it  is ionized.   The function of the inlet system is to
 transfer the sample from a high pressure (i.e., 1 atm)  region into the
 vacuum of the spectrometer without seriously unbalancing the spectrometer
 operation.   The generated beam of ions is focused and separated in the
 mass analyzer according to the m/z ratios.   The detection system senses
 the mass-separated ion beams,  and the readout device translates the
 signal provided by the detection system into an output  that can be
 interpreted  by the analyst.   Several reviews of pesticide residue
 analyses by  MS have been written by Safe (108), Skinner and Greenhalgh
 (109),  and Ryan (110).   In addition, detailed reviews of mass spectrometry
 have been made by Burlingame et al.   (Ill)  and Alford (112).

          (2)   Inlet Systems:   Direct Insertion Probe

               Samples  may be introduced directly into the ion source  with
 a  direct insertion probe assembly.   For example,  the sample is loaded into
 a  short length of melting point capillary,  placed in'the heater well  at the
 end of  a probe,  and inserted to within a few millimeters of the ion source
 through a vacuum lock  that maintains a vacuum-tight  arrangement.   The
 temperature  is then increased  until  the sample  vaporizes and  a spectrum
 is obtained.   The introduction of trapped GC fractions  into  an inde-
 pendent mass  spectrometer by these techniques was used  for residue
 analysis prior to the  development of combined gas chromatography/mass
 spectrometry  (e.g., 113),  but  the latter procedure is now employed
 almost  exclusively.  The direct insertion probe is reserved mainly
 for samples that  cannot  be chromatographed,  because  of  thermal instability,
 low vapor pressure, and/or high polarity.  However,  combined with  specific
 ionization techniques, such  as  negative chemical  ionization, direct
 insertion can  provide  a  sensitive, rapid screening method  (113A).


                            Combined GC/MS

For Impure samples  such  as biological  extracts, the  gas  chromatograph of
a  coupled GC/MS instrument serves  as an efficient  inlet  system for intro-
duction of samples  into  the  spectrometer.  The  resolution provided by gas
chromatography offers extra sample cleanup in addition to any partition
and liquid chromatography  steps.  Temperature and  sometimes flow rate
programming have  proven  useful  for achieving high  chromatography resolution
with the  combined instrument.   Column  bleed can be a  serious proble in
GC/MS, since bleeding liquid phase is  also detected by the mass spectro-
meter and contributes spurious  ions  to  the analytical spectra.  Carefully
conditioned,  low  bleed columns  that  are  stable at high temperatures should
be  used whenever  possible.  Other approaches that alleviate problems from
column bleed include use of a short, bleed-absorbing column placed between
the analytical column and the GC/MS  interface, programming the flow rate
of  the carrier gas, and  computer subtraction of background resulting from
the bleed (114).
                                 -410-

-------
                                                        Section 10L


Compatability of the gas chromatograph and mass spectrometer is a problem
because of the large volume of carrier gas eluting from the chromatograph
and the need to operate the spectrometer at high vacuum (10~5 - l.0~6 mm Hg).
In the simplest approach, the two instruments are connected directly, and
a large pumping system is used to maintain the required vacuum in the mass
spectrometer.  This approach has been used successfully with GC columns
having flow rates up to ca 5.0 ml/minute.  Introduction of samples from
packed columns into the mass spectrometer requires removal of most of
the carrier gas by means of an interface between the two instruments.
At the same time, as much sample as possible should be retained so that
the gas flowing into the spectrometer is enriched in sample.  Three basic
types of sample enriching devices or separators have widespread use in
modern GC/MS systems, namely effusion (Watson-Biemann; Brunnee), jet
(Ryhage), and membrane (Llewellyn~Littlejohn).  Each has its own advantages
and limitations, and there appears to be no strong preference for one over
the other.  In all cases, some carrier gas enters the source along with
the sample molecules, and broadening of GC peaks by the interface may
occur.  In practice, most separators convey only 20-40% of the sample
in the GC effluent to the mass spectrometer.  The theory and operation of
separators have been described in detail by McFadden (115, 116).  The
mass spectrometer in a combined instrument must be able to scan through
an appropriate mass range, e.g., from mass 10 to mass 800, in a small
fraction of the time that it takes to elute the peaks from the gas
chromatograph.  •

                            Combined LC/MS

GC/MS is sometimes limited by the volatility or heat sensitivity of the
compounds under study.  To circumvent these difficulties, various methods
of interfacing a high pressure liquid chromatograph with a mass spectro-
meter have been explored.  The demands on a LC/MS interface are much more
extreme than for GC/MS, because of the greater enrichment required
(usually 10^) and the possible adverse effects (e.g., background inter-
ference, chemical ionization effects, filament damage, etc.) of excess
solvent entering the ion chamber.  Six methods have been used for LC/MS
interfacing.  Three methods, namely the high capacity atmospheric pressure
ionization source, the semipermeable dimethyl silicone membrane, and
modification of sample at the interface by reduction to hydrocarbon, have
not been widely, accepted.  Methods involving direct introduction with no
enrichment, direct introduction with jet enrichment, and mechanical
transfer using  a moving wire with belt are most promising and are under
active development.  All six methods have been described and compared
by McFadden  (116), with appropriate original literature references.

The most common commercial liquid chromatograph/mass spectrometer inter-
face  (117) (Figure 10-H) consists of a continuous belt that introduces
the LC effluent into a chamber at atmospheric pressure and then sequentially
passes it beneath an infrared heater and through two vacuum locks into a
vaporization chamber.  Under optimum conditions the LC solvent is evaporated
from  the belt by the heater and vacuum locks, leaving only a deposit of
the sample on the belt.  The vacuum locks also accomplish the transition
                                -411

-------
                                                         Section 10L

 from atmospheric pressure to the vacuum system of the mass spectrometer.
 In the vaporization chamber, a second heater volatilizes the sample in
 front of a nipple leading into the ion source.  A third heater cleans
 the helt before its return to the atmospheric chamber via the vacuum
 locks.
      Figure 10-H.
HPLC/MS interface developed under contract by
Finnigan Corporation* for EPA.
                                                  LC EFFLUENT
             FLASH
            VAPORIZER
                                          INFRARED
                                          REFLECTOR
  ION
 SOURCE
           CLEAN-UP
            HEATER
                                                    DRIVE
                                                    WHEELS
                                           SPRING
                                           LOADED
                                           IDLER
                                           WHEEL
This interface  is able  to  accommodate most  commonly used organic LC
solvents at optimum flow rates varying  from about  0.2  to 1.5 ml/minute.
The use of water as an  LC  solvent  generally requires ail LC effluent
splitter if reasonable  LC  flow rates are  to be used, since the maximum
capacity of the interface  for water appears to be  about 0.1 ml/minute.
The LC/MS system has been  successfully  applied to  the  analysis of a
large number of carbamate  pesticides (117).

The field of LC/MS is still under  development.  A  second LC/MS interface
has been introduced by  Hewlett-Packard, and other  commercial interfaces
are anticipated in the  near future.


         (3)  lonization Processes:  Electron Impact (El)

              The most widely used ionization source is the electron impact
type wherein gaseous molecules are ionized  by electrons emitted from a
* Mention of commercial products does not constitute endorsement by the U.S.
  EPA.
                                  -412-

-------
                                                        Section 10L


glowing filament.  These positive ions are accelerated into the analyzer
section.  The El source is relatively stable and easy to operate, and
has high ionization efficiency.  The overall quantity of positive ions
and the nature of the fragmentation process depend on the energy of the
ionizing electron beam.

Upon ionization of many compounds at low electron energy levels (0-20
electron volts or eV), a large fraction of the ion current tends to be
carried by unfragmented molecular ions.  However, the absolute intensity
is relatively -low.  At higher energy levels, fragmentation and rearrange-
ment are more prevalent, and the ion current is much higher.  Molecules
are often cleaved to such an extent that the molecular ion is absent from
the.mass spectrum or is of very low intensity.  (A great many other com-
pounds  form only fragments even at low eV values.)  Because mass spectra
are more reproducible when compounds are ionized by 60-80 eV electrons,
most, mass spectrometers are operated in this energy range.  It is note-
worthy  that the El source produces mass spectra that are quite repeatable
among instruments and distinctively characteristic of the compounds being
ionized.  This has led to the  collection of large libraries of mass
spectral data with which unknown spectra can be compared.  Such comparisons
often permit rapid identification of the unknown pesticide.


                       Chemical Ionization  (CI)

Chemical ionization  spectra are obtained by adding methane, helium, or
other reagent gas  (at  relatively high pressures of about 1 mm or 130 Pa Hg)
to  the  sample either as the GC carrier gas or  after removal of the GC
carrier gas by  the separator.  In the latter case, the CI reagent gas
is  introduced into the mass spectrometer just  ahead of the point at which
the effluent enters  the ion source, or into the source itself.  Electrons
produce reagent gas  ions  that  subsequently  ionize  sample molecules by
chemical reactions,  e.g., proton transfer, hydride abstraction, .ion
attachment, and resonance transfer.  The mass  spectra obtained with CI
are quite  different  from  those formed on electron  impact and are,  in
general, simple and  complementary  to electron  impact spectra for pesticide
confirmation.   Although CI usually  provides molecular ion  (MT-)  or
 (M + H)+ or (M  - H)+ peaks of  high  intensity,  a  study  (118) of a series
of chlorinated  and organophosphorus pesticides found no molecular  ion
or ions in the  molecular  ion  region produced  from electron  impact  or
chemical  ionization for  a number  of specific  compounds.  CIMS has  sensi-
 tivity at  least as good  as  that  of El  (119) and  offers  the  advantage  of
 allowing  characterization of  a sample's  chemical reactivity through the
 choice of  the reagent gases.   In addition to  methane  and helium,  other
 gases including isobutane,  hydrogen,  argon-water,  ammonia,  and nitric
 acid have been used successfully to produce CI spectra  (120).   Positive
 CI data for 29  OP insecticides and metabolites have been reported (121).
                              -413-

-------
                                                         Section 10L
                         Field lonization (FI)
 Field ionization involves passing a gaseous compound between an anode
 (usually a thin wire or sharp blade) and a cathode.  An extraordinarily
 high electric field, approximately 108 V/cm, is impressed on the anode,
 permitting (as most commonly explained) valence electrons of the sample
 to "tunnel" to the metal of the wire or blade.   Charge separation can
 take the form of electrons tunneling out of molecules, proton transfers
 (facilitated by tuneling) between molecules, separation of the oppositely
 charged ions in electrolytes, and so on (111).   A positive ion results,
 which can be separated according to mass-to-charge ratio and detected.
 This is a relatively low energy ionization method that often produces
 enhanced molecular ion intensities and a cleaner, spectrum for compounds
 with poor thermal stability.

 Fragmentations are less prevalent and different from those observed in
 normal El spectra.   The FIMS  of a number of pesticides has been studied
 (122), and FIMS has been combined sequentially with HPLC for the determina-
 tion of trifluralin (123).


                          Field Desorption (FD)

 Field  desorption MS is a modification of FI in which the sample is applied
 directly to a carbon or metallic filament anode.   As with FIMS, field
 desorption depends on application of very high electric fields (5000-10,000 V)
 to this anode.   Sample molecules in contact with the anode desorb as ions
 into the source,  where they are separated and mass analyzed.   Like FI,  field
 desorption produces ions of low internal energy,  and usually results in
 minimal sample fragmentation.   Unlike FI,  field desorption has no require-
 ment that the compound be volatile prior to ionization.   Mass- spectra can
 be obtained for samples that  are thermally unstable or have no appreciable
 vapor  pressure, as for example, salts.   The field desorption mass spectrum
 of endrin and its El spectrum, which has a low abundance of the molecular
 ion, are shown in Figure 10-1 (124).   Strong molecular ion peaks are pro-
 duced  for most pesticides (118), including highly polar pesticides and
 metabolites such as carbamates and ureas (124-126).   Impurities may also
 give only molecular ions, so  interpretation of  mass spectra is sometimes
 simplified and the necessity  of sample cleanup  reduced.   However,  assign-
 ment of molecular ions and  interpretation of spectra in biological samples
 can be complicated by the presence of (M + H)+,  (M + Na)+, or other ion
 adducts and cluster ions.   Other disadvantages  are that quantitative data
 are difficult to obtain by  FDMS, and valuable structural information pro-
.vided  by fragmentation is lost.


                  Atmospheric  Pressure lonization  (API)
                                         —12    —15
 A  novel method with high sensitivity (10    - 10     g)  involves generation
 of ions with an atmospheric pressure ionization source.   The  API instrument
                               -414-

-------
                                                         Section 10L
is essentially an electron capture  detector with suitable interfacing to
a mass spectrometer so that  its  ions  can be mass-identified.   The source
uses °%i on gold foil to produce electrons that can interact with
nitrogen and water passing through  the  ionization chamber at  atmospheric
pressure.  Preheated carrier gas enters the API source just behind the
sample injection port.  Both gas and  sample pass through the  *>%i source
block where ionization reactions take place.   Just beyond the chamber
is a small aperture "through  which the ions  pass on their way  to being
mass analyzed and detected.  With certain samples, this source generates
more ions for a given quantity of sample molecules than any other ion
source; this is reflected in the reference  to  the API mass spectrometer
as the "femtogram machine" (127).   The  ^%i source has been replaced by
a corona discharge (128) , producing identical  API mass spectra and limits
of detection but a greater dynamic  response range.  Qualitative and
quantitative applications of negative ion formation from pesticides in
the API mass spectrometer have been studied (129) .
     Figure 10-1.  Electron impact  (El)  and  field desorption (FD)
                   mass spectra of  endrin  (124).  •
                                                 "
             210   230  250   270  290   310  330   350  370   390
i«h
            40-
                 FD
             300  310   330   330  3IO
                                         370  360   390  400
                              -415-

-------
                                                        Section 10L
                     Negative Chemical lonization

An important new development in chemical ionization methodology is
simultaneously pulsed positive and negative CI mass spectrometry,
developed by Hunt et al.  (130).  In this method, both the positive
and negative ions produced in a CI source are alternately pulsed from
the source, with appropriate potentials, through a quadrupole analyzer
to two electron multipliers, one for positive and one for negative ions.
Positive and negative mass spectra are, thereby, measured "simultaneously".
Under favorable circumstances, negative CI can afford sensitivity two or
three orders of magnitude greater than that obtainable with positive CI
(131), making it a very relevant technique for residue analysis of
pesticides and dioxins.

The positive and negative methane (132) and isobutane (133) CI mass
spectra of selected polycyclic and aromatic chlorinated insecticides
of several types have been determined and published.  The negative
CI spectra with isobutane as enhancement gas were exceptionally simple,
with the most abundant ion for almost all compounds studied being
(M + Cl)~ (133).

Negative CIMS with methylene chloride reagent gas was the basis of a
multiresidue screening procedure for OC1 residues in environmental
substrates at 1 ng levels (134).  The four principle negative ion-forming'
reactions with methylene chloride as reagent gas were (1) resonance
capture of an electron to give M~ ; (2) chloride attachment to hydrogen-
bonding or carbon-bonding substrates to give (M + Cl)~; (3) deprotonization
or dissociative capture of an electron for relatively strong gas phase
acids to give (M - H)~ ; and (4) oxygen-chloride exchange to give
(M - CI +'0)~ (113A, 134).

Polychlorinated dibenzo-p_-dioxins were determined in biological samples
by methane negative CIMS, which was found to be as much as 1000-fold more
sensitive than methane positive CI and electron impact MS.  The use of
oxygen with or without methane resulted in decreased sensitivity but
increased selectivity for the dioxins.  Detection limits ranged from
100 to 500 pg for 2,3,7,8-TCDD down to ca 1-10 pg for 1,2,3,4,6,7,8-HpCDD
by selective ion monitoring (131).  The highly toxic 2,3,7,8~TCDD can be
distinguished from other isomers by negative chemical ionization and
reaction with oxygen to form dichloroquinoxide ions (134A, 134B).

A discussion of 13 methods for ionization of organic compounds in MS
has been published, including detailed consideration of chemical ioniza-
tion and field ionization and pesticide spectra (135).  Design con-
siderations of El, CI, FI, FD, and API sources have been described (136).
Five ionization methods were compared for producing positive and negative
ion mass spectra of typical organophosphorus pesticides.  The negative
ionization techniques were much more sensitive for the 16 compounds
tested (137).
                                 -416-

-------
                                                        Section 10L


         (4)  Mass Analyzer Systems

              Low resolution magnetic analyzer systems depend on bending of
the ion beam in a magnetic .field.  The magnetic field segregates the ions
into beams, each of a different m/z.  To obtain the mass spectrum, the
magnetic field is varied, and each m/z ion from light to heavy is
successively brought to focus on the exit slit.  Such analyzers are
referred to as single- or direction-focusing analyzers.  High resolution
instruments have an analyzer region with an electrostatic sector for
velocity or kinetic energy focusing plus a magnetic sector for separation
of fragments according to m/z ratio.

Quadrupole analyzers are based on mass separation in a radio frequency
(RFX•electric field.  This field is established on a set of four
precision parallel, usually circular, rods, with both a DC voltage
and an RF alternating voltage being applied to these rods.  Ions are
accelerated gently  (5-30 V) into the analyzer or filter region, and
begin to oscillate  between the rods.  At a given DC and RF level, ions
of a specified m/z  value undergo stable oscillations and pass  through
the length  of the analyzer tube  to  the detector.  Ions of lower or
higher mass will undergo increasingly erratic oscillations that eventually
result in  their striking the rods or walls.  The spectrum is obtained by
sweeping ^the applied RF voltage  and DC ramp voltage and measuring the
detector, current as a  function of time.

        , (5)  Resolution ,

     .         Resolution describes  the performance  of  the mass analyzer
in terms of its ability  to separate ions  of  different  masses  from one
another.   Resolution  is  expressed  in numerical form by the  equation
M/AM where M and M +  AM  are mass numbers  of  two  neighboring peaks of
equal  intensity  in the mass spectrum.  The criterion  for resolution is
a relative height  of  the valley between  peaks  of 10%,  with.each peak
contributing 5%  to the valley. .  For example,  an instrument  would, have
 a resolution of  100 if two peaks with a mass difference of 1 .part in
 100 (e.g.,- m/e 100 and 101)  were resolved to the 10%  level.   Low resolu-
 tion mass spectrometers  typically show maximum resolution values  between
 300 and 1000,  while high resolution instruments are capable of attaining
 resolutions well in excess of 104.   The advantage of a high resolution
 spectrometer is the capability of resolving ions with very little
 differences in mass and obtaining the masses of these ions accurately
 to 0.001 mass units or better.   Exact masses are determined using a
 computer coupled to the mass spectrometer or by peak matching known
 marker peaks and unknown peaks on an oscilloscope (138).  Once the
 exact mass of a key ion (often the molecular ion) is known, the elemental
 composition or formula of the molecular or fragment ion is obtained,
 again by using a computer or by consulting tabulations of the mass.es
 of different combinations of atoms.  Elements .indicated to be present
 by the mass spectral pattern or prior information about, the unknown sample
 are often needed to correctly evaluate the data.
                                  -417-

-------
                                                        Section 10L


Resolutions of the order of 1000 are attainable with low resolution
magnetic and quadrupole analyzer designs, although single-focusing
magnetic analyzers can attain higher resolutions with an extreme decrease
in sensitivity due to the narrow slits that must be used.  Resolution in
excess of 8000 is considered high, since this is the amount usually
necessary to resolve most mass doublets.  The extra focusing added in
& high resolution mass spectrometer reduces the overall number of ions
traversing the instrument, thus reducing the overall sensitivity.  To
overcome such a reduction, the mass range is usually scanned at a slow
rate.  To minimize the effects from slow scanning and decreased sensi-
tivity, only as much resolution as is necessary to perform the requited
analysis should be used, since the accuracy of an exact mass measurement
is independent of resolution as long as any mass doublets are separated.

References (108, 110, 139-141) review methods and applications of MS
and combined GO/MS to pesticide residue analysis, and references (111,
112) give a more general survey of GC/MS instrumentation, principles,
and techniques.


     b.  Examples of GC/MS Confirmation

         Figure 10-J shows the electron capture gas chromatogram obtained
by injection of an aliquot of the 6% ethyl ether Florisil column eluate
from cleanup of a human adipose tissue extract (142).  Figure 10-K shows
the total ion current chromatogram of the same eluate from GC-MS.  Although
the curves are drawn to different scales and are not directly comparable,
it is evident that many more compounds are identifiable in the latter
because of the general response of the mass spectrometer.  In general,
chromatograms traced by the total ion monitor are similar, but not
necessarily identical, in response and sensitivity to those traced
by a flame ionization detector.  Differences exist in sensitivities
to some compounds, and broadening occurs in some peaks in the interface
to the mass spectrometer.  Figure 10-L shows the mass spectrum of
standard j3,jj'-DDE, the major GC peak evident in both chromatograms in
Figures 10-J and 10-K.

The identification of pesticides from their mass spectra is often
complicated by the obscuring of low mass ions by impurity fragments,
especially in biological extracts.  For this reason, extra cleanup
of extracts may be needed for GC-MS as compared to GC alone.  For
example, alkaline hydrolysis has been used for the 15% ethyl ether
Florisil column eluate, while additional column adsorption cleanup
(e.g., alumina plus Florisil columns) or use of silica gel rather than
Florisil initially has been successful for the 6% ethyl ether eluate.
Gel permeation chromatography has also been successfully applied to the
6 and 15% fractions (143).
                                   -418-

-------
                                                      Section 10L
Figure 10-J.
Electron capture chromatogram of human adipose tissue
extract, 6% ether Florisil  column eluate
                       !

                       I
                       I?
             INJECTS?!
Figure 10-K,  Total ion  current  profile of the same human adipose
              tissue extract
                                   -419-

-------
                                                        Section 10L
     c.  The Mass Spectrometer as a GC Detector
         There are a number of ways to use the mass spectrometer as a
sensitive and selective GC dete'ctor.  These procedures require that the
analyst know what compound or compounds he/she is looking for and are
not applicable to totally unknown samples.

Selected ion monitoring (SIM), also called multiple ion detection (MID)
or multiple ion selection (MIS), involves automatic, continuous
monitoring of a few ions of different masses.  Tracings of the selected
masses are recorded simultaneously as rapid switching is accomplished
in the spectrometer to bring each ion into the detector in turn for a
short period of time.  Simultaneous recording of one or several compounds
can be achieved, with characterization of each being based on the
formation of one or more selected ions (144).  To use SIM effectively,
one should know the kind of compound sought and its MS characteristics.
Sensitivity of detection for SIM can sometimes be extended to the sub-
picogram range, which is considerably more sensitive than conventional
scanning because of the longer sampling time at each selected mass.
Sensitivity for a particular compound is influenced by the extent of
fragmentation and the fraction of the total ion current carried by the
selected ions.  Identification and quantitation of compounds can be
improved by exact mass measurement  (e.g., to 0.001 amu) of the specified
ion, but only at the expense of sensitivity (144A).  Figure 10-M shows
the m/z 405, 407, 409, and 411 ions of trans-nonachlor and isomers
monitored simultaneously in a human adipose tissue.  Total ion current
profiles (TICP) cannot be generated by the SIM technique because data
from only certain masses are collected.

Compounds not resolved by gas chromatography can still be detected with
certainty if their molecular (or other characteristic) ions can be re-
solved by SIM.  Recording the masses and relative intensities of several
ions formed from a single pesticide can increase the certainty of compound
identification.  SIM has been applied to the detection of organophosphorus
insecticides (145) and to carbofuran and metabolites in crops (146).

Repetitive scanning through a narrow mass range generates quantifiable
spectral envelopes from several ions at once.  This procedure, generally
sensitive at low ng levels, has been applied to pesticide analysis (147).

Reagent ion monitoring is an interesting variation of single ion monitoring,
wherein the intensity of reagent ions used in a chemical ionization source
is monitored as a function of time.  The intensities of reagent ions de-
crease when they react with material eluted from the GC column, providing
a chromatogram that is distinctive  from those produced by other detectors
(148).
                                   -420-

-------
                                                   Section 10L
Figure  10-L.   Total  mass spectrum of £,pf-DDE

I00-|

80-


>,60-
(/)
§
J=j 40-

4>
_>
•§ 20-
o:


A^'-DDE
/T^\ /7
C'~\ /~?~\
\ 	 f c X.
Cl Cl
1

. ISO
1
,= 105
75
I lib
ill fi i
li "i ' il I 1 II 1 i
.... . i . jllii Ji.ii.iilA.ajlLiylL Jli. J LuLt j. s J.L JjJi
1 1 ' 1 ' I ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 • 1 ' 1 ' IT")"!"")1 I p fT
50 100 150



~^\ ,
\>— i
^/

'6







lUt _ s
_j
6 2.
Q

gf
3_
K"
_ Q
3 -•
O
3


200 250 300
Figure 10-M.  Selected ion monitoring applied  to  a  human  adipose tissue
              extract.  The four masses  shown  have  been monitored as
              the extract elutes from an-OV-17/OV-210  GC  column at 180°C.
              The largest peak is trans-nonachlor,  the last  eluting
              peak is cis-nonachlor, and the peak preceding  trans-
              nonachlor is an isomeric nonachlor  also  observed  to be
             _ present in technical chlordane.
          411.0
         409.0
         407.0
         405.0
                                  -421-

-------
                                                        Section 10L
     d.  Computerization of GC-MS

         Combination of a computer with a GC-MS system can serve several
very useful functions.

         (1)  The computerized GC-MS data acquisition system permits rapid
processing of information from complex sample mixtures.  The mass spectra
of specific compounds in the mixture can be experimentally obtained and
automatically matched with a library file of standard mass spectra.  Com-
puter control of data acquisition may enable the operator to devise
relatively complex scanning procedures.  For example, different mass
ranges may be sampled for different time periods, masses may be sampled
for times related to the intensities being measured, or several dis-
continuous mass ranges may be sampled.

         (2)  Column bleed and other background can be conveniently sub-
tracted by the computer.

         (3)  Continuous repetitive scans can be made during the entire
chromatographic separation; for example, a spectrum can be scanned every
2-4 seconds.  In a typical GC/MS run, several hundred to more than a
thousand mass spectra may be acquired in this way, each one being a complete
spectrum over the mass range selected.


All spectra are stored, and chromatograms may later be reconstructed by
the computer by summing and plotting the total ion current detected in
each scan but excluding carrier gas ions or other interfering ions.
Reconstructed total ion current profile chromatograms (TICP) obtained
resemble those traced in real time by a conventional total ion monitor
of a magnetic deflection spectrometer.  A typical reconstructed GC/MS
total ion current profile of an extract of human fat is shown in Figure
10-N, with some of the components identified (142).

         (4)  The computer can trace the intensities of selected character-
istic masses from among the great quantity of data acquired by continuous
repetitive scanning.  The resulting mass chromatograms or extracted ion
current profiles (EICP) (149) resemble the single or selected ion profiles
described earlier and permit compounds and spectra of interest to be
located and the appropriate spectrum to be retrieved and plotted.  EICPs
can be individual selected ion current profiles (SICP) or summed sets of
several masses, all extracted from scanned data.  EICPs have an advantage
over SIM in that large numbers of ion profiles and complete spectra can
be examined rapidly after only one chromatographic separation, but this
computerized acquisition of .repetitively scanned spectra is of considerably
lower sensitivity (as much as 10 ) than SIM because of the longer integra-
tion time characteristic of the latter method (149).  Reference (140)
illustrates computer-generated selected ion current profiles.
                                   -422-

-------
                                                         Section 10L
Figure 10-N..
Computer reconstructed total ion chromatogram and mass
chromatograms of M = 237  (o,j>'-DDT and j),j>'-DDT)  and
M s 405 (trans-nonachlor) from a composite  human  adipose
tissue extract.  Column:  45.7 m SCOT column coated with
SE-30, programmed from 170-240°C at  2°C/minute (N.C.  * not
chlorinated).
                                                   M-OOT
          M« 237.0
                                   100         1W
                                      SCANNVMCR
           M-408.0
                     1
The limited mass range chromatogram  (148)  is  a variation of mass chroma-
tography that has proved especially  valuable  in  the  determination of
polychlorinated hydrocarbons.  In this  technique,  the  computer sums
ion intensities (collected from  repetitive scanning) through a limited
mass range as a function of scan number or time.   (The procedure has
also been termed selected ion  summation analysis or  SIS.)   For example,
the molecular ion cluster of mirex,  due to the contributions of 37ci
from each of the 12 chlorine atoms,  is  spread over a range of more than
27 V.    Instead of treating a single ion  (e.g., CiQ^di2+' > nominal
m/z 540), the entire cluster can be  summed to provide  increased sensi-
tivity with some sacrifice in  specificity.  The  method has been used to
identify dieldrin and HCB residues in lake trout (150).
                                -423-

-------
                                                         Section 10L

          (5)   Quantitation of peak areas in the selected ion profiles and
 ratios of these peaks can be provided.

               Computer coupled GC/MS equipment is extremely expensive,  and
 highly qualified personnel are needed for operation,  maintenance,  and
 interpretation of data.   A significant amount of "down-time" is to be
 anticipated because of the complex nature of the instrumentation.   Computer-
 ized data acquisition and processing for magnetic instruments,  quadrupole
 instruments,  and selected ion monitoring have been described (110),  as
 have techniques available for computer identification of unknown mass
 spectra using various retrieval systems (151).


      e.   Applications of  GC/MS to  Pesticide Analysis

          Reference spectra and fragmentation data for pesticides of  several
 types and for related chemicals have been published (105,  108,  152-156).
 Applications  of GC/MS include confirmation of  the 1-naphthyl chloroacetate
 derivative of 1-naphthol  (a carbaryl metabolite)  extracted from urine
 (157);  2,4-D,  2,4,5-T, and 2,4,5-TCP in urine  (158, 159);  organophosphorus
 pesticides in blood and urine (160,  161)  and food (162); multiple
 chlorinated insecticides  in human  adipose and  liver tissue (142, 143,
 163),  foods (164),  and soils (165);  toxaphene  in human and biological
 samples  (166);  Kepone in  human and environmental samples  (167);  chlordane-
 related  residues in human samples  (142,  168);  thiabendazole and
 5-hydroxythiabendazole in animal tissue (on-column  methylation plus  SIM)
 (169); dimethoate residues in "wheat  by  SIM at m/z 87  (170);  and mirex in
 fish  (171).

An important  application  of GC/MS  has been mutual determination and
 identification of PCBs in the presence  of chlorinated  pesticides (172).
 Insecticides mixed  with PCBs have  been  identified at levels  below 10  ng
without  complete separation on a GC  column by peak monitoring' MS as
described  earlier (173).   GC/MS  has  been  successfully.applied to the
detailed analysis of  complex pesticide mixtures,  such  as technical
chlordane  (168).  Pesticides and PCBs have also  been identified by
GC/MS using chlorine  isotope ratios  to  reconstruct  chromatograms that
are characteristic  for the number  of chlorine atoms found  in repetitive-
scan spectra  (174).

Special MS and GC/MS  techniques  that have  been applied to  the analysis
of simple and  complex pesticides in a variety of  sample substrates include
selected ion monitoring (175,  176), field  ionization (177), and field
desorption MS  (178).  Methods have also been developed for the determina-
tion of carbamates and ureas by  combined  liquid chromatography/mass
spectrometry  (117).  References  (108, 110-112, 179) contain reviews of
applications to residue analysis.  Symbolism and nomenclature of mass
spectrometry have been reviewed  (107).
                                -424-

-------
                                                           Section 10M
        f.  Mass Spectrometry/Mass Spectrometry (MS/MS)
            Mass spectrometrists have, within the last few years, investigated
   the possible elimination of any preseparation method, such as GC or LC,
   for the analysis of complex mixtures.  Instead, the mass spectrometer
   itself is used as the separation device, followed by a second mass
   spectrometric analysis of the sample.  This technique is called mass
   Spectrometry/mass Spectrometry.  Techniques are available to perform
   this method using quadrupole or magnetic sector instruments with positive
   or negative ions.  Operation commonly involves the separation of the ions
   of a particular m/z value, characteristic of a given compound present in
   a complex mixture, by a first mass spectrometer.  This ion current then
   encounters collisions with gas molecules, which impart considerable energy
   to them through the process of collisional activation.  The resulting
   energetic ions may-then decompose into characteristic fragments, which
   are then analyzed in a second or third mass analyzer region, as the case
   may be.  This method,holds promise as a rapid method of mixture analysis.
   Hunt et al. (179A) have used MS/MS to analyze nitrophenols in sewage
   sludge.  However, recent studies show that artifacts can be created in
   the analysis (111).   Other references involving MS/MS' include (179B, 179C).


10M   QUALITY ASSURANCE OF GC-LOW RESOLUTION MASS SPECTROMETRY

   This section reviews procedures to be followed for quality assurance of
   data derived from the mass spectrometer in the identification, confirma-
   tion, and quantitative determination of chlorinated insecticides, PCBs,
   hexachlorophene, and PBBs in human tissues and fluids.  These methods
   were developed at the Health Effects Research Laboratory, U.S. EPA
   Research Triangle Park, NC (180) for use in the EPA National Human
   Monitoring Program for adipose tissue and serum samples;   The procedures
   assure interpretable mass spectra of the highest experimentally obtainable
   quality for compound identification as well as quantitative accuracy when
   monitoring ion intensities (as by selected ion monitoring).  The specific
   pesticides of current interest are the following:
                      '-DDT
£»£.'
p_,p_'-DDT
£»£.' -DDE
£.»£* -DDE
£,£/ -DDD
p_,p_'-DDD
cc-BHC
g-BHC
Lindane (y-BHC)
5-BHC
Aldrin
Dieldrin
Heptachlor
Heptachlor epoxide
Endrin
Mirex                 '    '
Oxychlordane
trans-Nonachlor
Polychlorinated biphenyls
Hexachlorobenzene
Polybrominated biphenyls
Polychlorinated terphenyls
                                     -425-

-------
                                                        Section 10M
     a.  Introduction to Quality Assurance Procedures

         Correct identification of organic pollutants from gas chromatography-
mass spectrometry (GC/MS) data requires valid mass spectra of the compounds
detected.  This is independent of the actual method of interpretation of
the spectra, i.e., an empirical search for a match within a collection of
authentic spectra or an analysis from the principles of organic ion frag-
mentation.  A properly operating and well tuned GC/MS instrument is re-
quired to obtain valid mass spectra.

The purpose of the following procedure is to permit a check of the per-
formance of the total operating computerized GC/MS system.  Thus, with
a minimum expenditure of time, an operator can be reasonably sure that
the GO column, the enrichment device, the ion source, the ion separating
device, the ion detection device, the signal amplifying circuits, the
analog to digital converter, the data reduction system, and the data
output system are all functioning properly.

An unsuccessful test requires the examination of the individual sub-
systems and correction of the faulty component(s).  Environmental data
acquired after a successful system check are, in a real sense, validated
and of far more value than unvalidated data.  Environmental data acquired
after an unsuccessful test may be worthless and may cause erroneous
identifications.  It is recommended that the tests be applied often on
a working system, especially when there is a suspicion of a malfunction.

The procedure is written for a low resolution mass spectrometer such as
the Finnigan 3200 or the Hewlett Packard 5930A quadrupole-type mass
spectrometer, equipped with an automated data system such as the Finnigan
6000 or Hewlett-Packard 5933A system.  However, the test is clearly and
readily adapted to any GC/MS system by suitable modification of the
detailed procedure.       .';

There is a special need to closely monitor the performance of the quadrupole
mass spectrometer.  Unlike the magnetic deflection spectrometer, the active
ion separating element of a quadrupole spectrometer (the rods) is directly
contaminated during operation, and after prolonged operation is subject to
severely degraded performance.  Since degraded performance usually affects
the high mass region first, the test includes high mass end criteria.
High quality, high mass data are important since many environmentally
significant compounds have molecular and fragment ions in the 300-500 y
range.

A quadrupole mass spectrometer, which has been tuned to give a reference
compound spectrum that meets the criteria of this test, will, in general,
generate mass spectra of organic compounds that are very similar, if not
identical, to spectra generated by other types of mass spectrometers.  Thus,
quadrupole mass spectra will be directly comparable to spectra of authentic
samples in collections that have developed over the years, mainly from
magnetic sector mass spectrometers.
                                    -426-

-------
                                                        Section 10M


Assurance of mass spectral data is obtained through a set of two levels
of functionality tests.  The first test requires establishment of pro-
duction, dispersion, and detection of ions from a reference compound,
perfluorotri-n-butylamine (PFTBA).  Relative peak heights are adjusted
to conform to the known electron  impact spectrum, with a slight biasing
toward increased transmission of  ions higher than m/z = 200, which are
not commonly interfered with by tissue component fragments.

The second test of the GC-MS combination requires injection of a known
low-level standard sample while the operation is under computer control.
This is followed by periodic verification of the quality of spectra
compared to spectra of known ideal quality.  Chemical compounds used
may be bis(perfluorophenyl)phenylphosphine(or decafluorotriphenyl phosphine,
DFTPP).  Another set of compounds commonly used are aldrin and heptachlar
epoxide.  Heptachlor epoxide is useful as a representative member of the
important chlordane series of pesticides and, more generally, because the
M-C1 ion, six-chlorine isotope cluster beginning at 351 m/z allows a test
of sensitivity and resolution at a very useful mass, not provided for by
PFTBA or many other mass calibration compounds, but quite relevant to
pesticide work.  The appearance of the 351 m/z cluster may be examined
at 100, 10, and 1 ng levels, as instrument sensitivity requires, with
respect to appearance of the six-chlorine cluster versus statistical
appearance.  Resolution of 13C isotope peaks and relative abundance
versus the 81 m/e peak may also be determined.  Aldrin, injected as a
GC retention time test, also has its mass spectrum routinely compared
against the literature spectrum with respect to correctness of chloro
cluster statistics, sensitivity, and relative appearance of high and low
mass fragment ions.  The retention time of heptachlor epoxide relative
to aldrin (1.59 + 0.02) on a 1.5% OV-17/1.95% OV-210 column at 185°C may
also be determined, along with GC column resolution.  This test has the
advantage of providing a full functionality evaluation of the GC/MS system,
including sensitivity, data system acquisition, and recall of spectra.

     b.  Quality Assurance Procedures

         (1)  Using PFTBA (3M trade designation:  FC-43)  standard:

              (a)  Check on oscilloscope and/or light beam oscillograph
that 69, 131, 219, 264, 414, 502, and 614 m/z ions are present and in
reasonable relative abundance according to the following tabulation:
                              -427-

-------
                                                        Section 10M
       General Desired Appearance of the Mass Spectrum of PFTBA
                 Mass
                 (m/z)

                   69
                  100
                  114
                  119
                  131
                  219
                  264
                  414
                  426
                  502
                  614
  Relative
Abundance %

   100.0
    24.
     9,
    20.
    70.
    68.
    16.
     5,
     2,
     2.
     0.3
               Isotope Abundance Checks, Percentage Ratio of
                         Ion Signal Abundances
                           (70)/(69)
                         (220)/(219)
                         (503)7(502)
    1.1%
    4.4%

   10.3%
                   Tune mass spectrometer as required, with respect to
resolution, optimum peak shape, sensitivity, and minimum mass falloff
(refer to appropriate instrument manual for instructions).

              (c)  Calibrate data system and verify the calibration by
examining a PFTBA spectrum acquired under data system control (refer to
appropriate data system manual for programs).

         (2)  Run aldrin and/or heptachlor epoxide and examine the re-
constructed total ion chromatogram and mass spectra.

         (3)  Perform DFTPP test (optional).

         (4)  Go on to sample runs.

     c.  Preparation of Aldrin and/or Heptachlor Epoxide Standards

         Primary standards of aldrin and heptachlor epoxide can be obtained
from the Pesticide Repository, Health Effects Research Laboratory, EPA,
Research Triangle Park, NC.
                            -428-

-------
                                                         Section 10M


         Carefully weigh out  20 mg of the pesticide and dissolve in 100 ml
of n-hexane  (pesticide  quality, or  equivalent)  in a volumetric flask.   Keep
this  stock solution  under refrigeration.   Replace every 6 months.

         Prepare  a working standard of 20 ng/vl  concentration by diluting
1 ml  of  the  stock solution to 10 ml in a  volumetric flask.   These working
solutions  should  be  replaced  at least monthly.

      d.  Preparation of Decafluorotriphenylphosphine (DFTPP)  Standards

         Prepare  a stock solution  of DFTPP at  1  mg/ml concentration in
acetone  (pesticide quality, or equivalent).  This stock solution has been
shown to be  97+ percent stable after 6 months, and indications are  that
it will remain useable  for several years.-   Dilute an aliquot  of the stock
solution to  10 yg/ml (10 nl/yl)  concentration  in acetone.  The very small
quantity of  material present  in very dilute solutions is subject to
depreciation due  to  adsorption on  the walls of the glass container,
reaction with trace  impurities in  acetone,  etc.   Therefore,  this solution
may be useable only  in  the short term,  perhaps 1-3 weeks.

      e.  Quality  Assurance Test

         (1)  Adjust the GC column flow to  normal operational level
(e.g., 30  to 45 ml/min)  and set the desired oven temperature  (e.g.,
185°C).  The parameters should be  adjusted  to provide at least four
spectral scans during the elution  of the aldrin,  heptachlor epoxide,
or DFTPP standard.

         (2)  Set mass  spectrometer at  normal or high sensitivity as  desired.

         (3)  Calibrate the instrument.

         (4)  Inject 40 ng of  aldrin and/or  heptachlor epoxide (or  20 ng
of DFTPP)  on the  GC  column and note the time (or start stopwatch).

         (5)  After  the solvent  passes  through the analyzer and  the Vacuum
has recovered, turn  on  the ionizer  and  start scanning.

         (6)  Note the  exact retention  time  of the standard as it elutes
from  the column.  This  retention time can be used as  a daily  check of the
condition  of the GC  column and  separator by  comparing the values.  The
retention  times should  not vary significantly from day to day under identical
operating  conditions.

         (7)  Terminate the run, turn off the ion source and  electron
multiplier,  and reconstruct the  gas  chromatogram.

         (8)  Select a  spectrum number  on the front side of the GC peak as
near  the apex as possible  and  select  a  background  spectrum number immediately
preceding  the peak.
                                  -429-

-------
                                                        Section 10M
         (9)  Plot or display the mass spectrum and compare against a
reference spectrum-  The spectrum obtained on the test system should
contain ion abundances within limits given for the key ions in the
following tables.  Sensitivity is considered adequate if 40 ng or less
of either aldrin or heptachlor epoxide and 20 ng. or less of DFTPP pro-
vide good mass spectra.
  Reference Aldrin Mass Spectrum
    (5-chlorine cluster check)
                    Reference Heptachlor Epoxide Mass Spectrum
                            (6-chlorine cluster check)
m/z
261
262
263
264
265
267
269
271
Abund. (%)
61.5
4.7
100.0
7.8
65.0
21.1
3.4
0.2
(intensity of
 fragment)
(intensity of
 base peak)
(261)

 (66)
45%
m/z
351
352
353
354
355
357
359
361
363
(intensity of
fragment)
(intensity of
base peak)
Abund. (%)
51.2
5.6
100.0
11.2
81.2
35.2
8.5
1.1
0.06
(351) m
A7ff/
(81)
                   Reference Mass  Spectrum of DFTPP
        51
        68
        70
        127
        197
        198
        199
        275
        365
        441
        442
        443 (M+l)
        444 (M+2)
                        s Abundance Criteria

           30-60% of mass 198
           Less than 2% of mass 69
           Less than 2% of mass 69 (1.1% theoretical)
           40-60% of mass 198
           Less than 1% of mass 198
           Base peak, 100% relative abundance
           5-9% of .mass 198 (6.6% theoretical)
           10-30% of mass 198
           1% of mass 198
           Less than mass 443
           40-60% of mass 198 — this ion is very sensitive
            to spectrum number chosen and condition of
            equipment.  If greater than 60%, equipment is
            OK if all other criteria are met.
           17-23% of mass 442 (19.8% theoretical)
           1.86% (theoretical)
                                   -430-

-------
                                                         Section  10M


     f.  Protocol for Analysis of  Samples

         (1)  Sample Collection

              Samples of human adipose  tissue are obtained  through
cooperating medical pathologists and medical examiners  at hospitals in
cities selected according "to a proportionate, stratified-random  design.
The conterminous 48 states were divided into 9 census divisions, according
to the 1970 census of the United States.  A city within each census
division was selected from those already participating  in the National
Human Monitoring Program as the collection site for special projects.

              Blood sera samples are collected throughout the U.S.  by
means of a .cooperative arrangement between EPA and the  U.S. Public  Health
Service.  The PHS'program, called  the Health and Nutritional Examination
Survey II (HANES II), provides blood specimens from a probability sample
of persons 12 to 74 years old, along with various medical and nutritional
parameters and some information regarding pesticide use by the indi-
viduals sampled.  The blood is drawn into evacuated ampoules, allowed to
clot, and centrifuged, and the serum is decanted into a clean vial.

         (2)  Cleanup

              Tissues are normally extracted and cleaned, up according to
a modified Mills-Onley-Gaither procedure (Subsection 9A) by laboratories
under contract to the National Human Monitoring Program.  Concentrated
extracts, corresponding to the 6% and 15% ethyl ether in petroleum  ether
fraction from the Florisil cleanup column, are then sent to the ACB/HERL-
RTP for GC/MS analysis.  Composite samples, comprising  100-500 individual
samples, require additional cleanup before GC/MS analysis.  The usual
method of choice is gel permeation chromatography (GPC)  as described in
Section 91.  Bloodjsera samples (Section 9D) may or may not need GPC
cleanup.

         (3)  Analysis

              After cleanup, samples are concentrated by removal of solvent
at room temperature under a gentle stream of nitrogen.   The final volume is
usually 100 yl, but it may be smaller if levels of compounds sought are
particularly low.  Quantitative analysis is performed in the electron
impact (El) mode.  Aliquots of 5 to 50  ul are co-injected with aldrin
(e.g., 250 ng in hexane) as an internal standard into the GC/MS system.
A total ion current profile is generated, and retention  times relative
to aldrin are determined for each component of interest.  Mass spectral
data are recalled from the computer for each component of interest and
analyzed against reference mass spectra obtained from various literature
references (e.g., 142, 143) or from a reference library  such as the NIH-
EPA Chemical Information System (181),  or, most preferably,  generated
from authentic laboratory standards.  Relative retention times are also
compared to those of the reference material for further  confirmation.
                              -431-

-------
                                                         Sections ION, 100


   After identification, quantitative analyses are usually performed by se-
   lected ion monitoring (SIM).  An authentic reference sample is used for
   direct comparison.  Identification may be further confirmed by chemical
   ionization GC/MS, where available.

        g.  GC/MS Systems

            Manufacturer's operating manuals should be consulted for descrip-
   tions and detailed operating instructions for specific GC/MS systems.  The.
   previous edition of this Manual contained information on two GC-MS systems:
   the Hewlett Packard 5930A quadrupole MS, 5700A gas chromatograph, and
   5933A data system; and the Finnigan 3200 quadrupole MS and 9500 gas
   chromatograph.

            Another EPA Manual (182) contains specific information on the
   Finnigan 1015 and 3000 quadrupole GC/MS systems coupled with a PDP-8
   data system.  This Manual includes 10 chapters covering the following
   material:  (1)  introduction to broad spectrum organic analysis, routine
   monitoring of large numbers of target compounds, and real time selected
   ion monitoring; (2) detailed start-up and calibration procedures; (3)
   preparation methods for water samples; (4) information on OUTPUT programs
   for data analysis; (5) compound identification using PDP-8 software; (6)
   specialized techniques such as single ion monitoring, open tubular
   columns, chemical ionization, accurate mass measurement, standard
   additions, and sample spikes; (7) miscellaneous auxiliary software programs
   and housekeeping routines; (8) preventive maintenance; (9) trouble
   shooting; and (10) selected bibliography up to 1978, mostly to information
   from EPA laboratories.


ION   BIOLOGICAL METHODS

   Bioassay techniques, which include insecticidal activity, enzymatic, and
   immunological methods, have been described as providing an independent
   criterion of identity when combined with GC, chemical reactions, etc.
   (2).  These methods, which depend on the measurement of a physiological
   response of a test organism induced by exposure to the pesticide, have
   advantages of simplicity and sensitivity but are relatively non-specific
   so that their utility for confirmation is rather poor.  The insect
   bioassay technique has been reviewed (183).

   Specificity of enzyme inhibition is greatly enhanced by combination with
   TLC for detection and confirmation of organophosphate and certain carba-
   mate pesticides.  The Rj? value plus biological response provide important
   identity information at levels typically in the range of 500 pg to 10 ng
   for these compounds.


100   POLAROGRAPHY (VOLTAMMETRY)

   Voltammetry is the generic name for a group of electroanalytical methods
   in which current-vs-voltage curves are recorded when a gradually changing
                                 -432-

-------
                                                        Section 100


voltage is applied to a cell containing the solution to be analyzed, a
stable reference electrode, and a small-area working or indicator
electrode.  In the special case where.the indicator electrode is a
dropping mercury electrode, the technique is called polarography.  In
addition to classical DC polarography, in which the current is measured
for each drop as voltage is increased linearly with time, modern
variations include DC current sampled polarography, pulse polarography,
differential pulse polarography, linear sweep (rapid scan) polarography •,
and AC polarography.  These newer methods differ in the type of voltage
signal applied and/or the manner in which the current is measured, and
they are generally more sensitive and/or selective than traditional DC
polarography.

The use of polarography as a confirmatory test is described in Section
12,F of the EPA PAM and Sections 640 and 641 of the FDA PAM:  Procedures
and applications of polarography for both identification and determina-
tion of pesticide residues have been reviewed (184-186).

Polarographic identification of a pesticide residue is based on the
determination of the peak potential of the unknown in a cleaned-up
extract, and comparison with the potential of about, the same amount
of a reference standard under identical conditions.  As a check,
addition of the standard compound to the unknown should result in an
increase in the wave height but not appearance of another wave.  Mixtures
can be identified if the peak potentials of the components are sufficiently
separated.  Trapped GC fractions may be subjected to polarography to
confirm identifications based on retention times.  Instrumentation for
such modern voltammetric techniques as fast sweep oscillography provides
sensitivity comparable to colorimetry.  Pesticides not containing an
oxidizable or reducible functional group can be made amenable to polarography
by formation of a suitable derivative (e.g., nitro, halogen, carbonyl, etc).

Most polarographic studies have been applied to phosphorus-containing
insecticides such as parathion, diazinon, malathion, and carbophenothion
(187).  A collaborative study confirmed the usefulness of single sweep
oscillographic polarography for identifying such residues in non-fatty
foods (188).  Nitrophenol metabolites of OP pesticides were determined
in urine by polarography (189).  Thirty-eight herbicides have been
studied by single sweep derivative polarography (190), methylcarbamate
insecticides by AC polarography and cyclic voltammetry (191), and urea
herbicides by anodic polarography (192).  Published voltammetric reduction
potentials for about 100 organochlorine insecticides, PCBs, and naphthalenes
(3 electrode potentiostat, DMSO solvent) are a useful aid in identification
of residues  (193).  Parathion and related insecticides and metabolites
were polarographically determined in blood without extraction (194),.  The
voltammetry of 1,3,5-triazines  (195), propachlor herbicide (in soil) (196),
dithiocarbamates (197, 198), dinitroaniline herbicides (199), thiourea-
containing pesticides (200), trifluralin (in soils) (201), azomethine-
containing pesticides (e.g., Cytrolane, Cyolarie, chlordimeform) (202),
PCP (203), and phosmet (in apples) (204) has been reported.  Paraquat
can be directly determined in urine and serum by differential pulse
polarography at ca 0.04 yg/ml levels (205).
                                -433-

-------
                                                       Sections 10P, 10Q
10P   MISCELLANEOUS CONFIRMATORY METHODS

      a.  Carbon Skeleton Chromatography               '      .  ,

          Carbon skeleton chromatography (CSC) is useful in characterizing
   insecticide residues in amounts down to 5-100 ng.  Apparatus for CSC
   consists of a precolumn containing a hot (ca 300°C) catalyst attached
   to a gas chromatograph equipped with a flame ionization detector
   (available from National Instruments Laboratory, Rockville, MD).  the
   compound to be identified is injected directly on the catalyst .bed
   (e.g., 1% Pd on 60-80 mesh Gas-Chrom P) and is swept over the bed by
   hydrogen.carrier gas.- Nitrogen is introduced through the normal instru-
   ment inlet so that the detector yields optimum response.  While in the
   precolumn, all functional groups are stripped from the compound, and
   any multiple bonds are saturated.  The resulting hydrocarbons are
   carried into the chromatographic column where they are separated and
   identified by their retention characteristics relative to standards.
   This identification method, which is in effect a deriyatization pro-  .
   cedure, has been applied .to heptachlor, heptachlor epoxide, chlordane,
   aldrin, endrin, DDT and its analogs, and carbaryl.  Sufficient residue
   must be available for the method to be of value.  Techniques, applica-
   tions to many pesticide classes, and characterization of products of
   CSC (as well as some other precolumn reaction confirmatory methods)
   have been reported by Beroza and co-workers (206-209) and Asai et al.
   (210, 211).  Identification of 5-10 ng amounts of polychlorinated
   biphenyls, terphenyls, naphthalenes, dioxins, and dibenzofurans in
   biological samples has been demonstrated (212), and mixtures of
   polychlorinated naphthalenes, PCBs, PCTs, and OC1 pesticides have been
   analyzed (213).

      b.  Fragmentation Procedures

          GC fragmentation procedures are similar to CSC except that the
   reaction in the precolumn decomposes the pesticides, yielding character-
   istic fragment peak patterns or fingerprint chromatograms helpful in
   making identifications.  A palladium catalyst at 300°C (210) and reagents
   such as Na2C03, CuO, CdCl2» A1C13, and I^C^Oy at 240°C (214) have been
   applied to chlorinated and OP insecticides with EC detection of the
   reaction products.                                             .  .

          Gas chromatograms of 33 organochlorine pesticides after ultra-
   violet irradiation have been published.  These characteristic photo-
   decomposition patterns are also useful for conclusive residue confirma-
   tion  (215).

10Q   REFERENCES                      ,

   (1)  Analytical Methods for Pesticide Residues in Foods, Department of
        National Health and Welfare, Canada, 1973, Section 10.2.
                                   -434-

-------
                                                         Section 10Q


.(2)  Robinson,  J.,  Richardson,  A.,  and Elgar,  K.  E.,  Chemical Identity
     in Microanalysis,  presented at the ACS National  Meeting, New York
     City,  September 11-16,  1966;  Robinson, J.,  Chem. Br.,  7., 472 (1971).

 (3)  Elgar, K.  E.,  The  Identification of Pesticides at Residue Concentra-
     tions, Advances in Chemistry Series 104,  Chapter 10,  ACS, Washington,
     D.C.,  1971,  page 151.

 (4)  Onuska,  F.  I., and Comba,  Mi  E., J. Chromatogr., 119.  385 (1976).

 (5)  Ruzicka, J.  H. A., and Abbott, D. C., Talanta, 20, 1277 (1973).

 (6)  Bailey,  R.,  Health and Welfare Canada, Health Protection Branch,
     personal communication (1980).

 (7)  Aue,  W.  A.,  and Kapila, S., Anal. Chem..  50., 536 (1978); Kapila, S.,
     and  Aue, W.  A., J. Chromatogr.. 148(2). 343 (1978).

 (8)  Mallet,  V.,  J. Chromatogr.. 2i> 217 (1973).                     .

 (9)  Lawrence, J. F., and Frei, R, W., Chromatogr. Rev.. 18, 253  (1974).

(10)  Connors, K.  A., Anal.  Chem.. 46_, 53 (1974).      .

(11)  Dale, T., and Court, W. E.. Chromatographia. .13(1), 124  (1980).

(12)  Heinz, D. E., and Vitek, R. K., J. Chromatogr. Sci.. 135 570 (1975).

(13)  Dolan, J. W., and Seiber, J. N., Anal. Chem.. 49.  326  (1977).

(14)  Beroza, M., and Bowman, M. C., J. Assoc. Off. Anal. Chem., 48,  358  (1965).

(15)   Bowman, M. C., and Beroza, M... J. Assoc. Off. Anal. Chem.. 48,  943  (1965).
         •0                       '.''..
(16)   Beroza, M., and Bowman, M. C., Anal.  Chem.,  37,  291 (1965).

(17)   Beroza, M., and Bowman, M. C.s Anal.  Chem.. .38,  837 (1966).

(18)   Bowman, M. C., and Beroza, M.. Anal.  Chem..  38,  1427  (1966).

(19)   Crist, H. L., Harless, R. L.,  Moseman, R.  F., and  Callis, M. H.,
      Bull. Environ. Contam. Toxicol..  24(1),  231 (1980).

(20)   Cochrane, W.  P.,  J. Chromatogr.  Sci.. 17(3),  124 (1979).

(21)   Cochrane, W.  P.,  and  Chau, A.  S.  Y.,  Chemical Derivatization Techniques
      for Confirmation  of Organochlorine Residue Identity,  Advances in
      Chemistry Series  104,  Chapter 2, ACS, Washington,  D.C., 1971, page 11.

(22)  Cochrane, W.  P.,  J. Chromatogr.  Sci., 13,  246 (1975).
                                     -435-

-------
                                                          Section 10Q
 (23)

 (24)


 (25)


 (26)


 (27)

 (28)

 (29)


 (30)


 (31)


 (32)

 (33)


 (34)


 (35)


 (36)

 (37)




 (38)

 (39)


(40)

(41)
 Holdrinet, M., Bull.  Environ.  Contain.  Toxicol..  J21(l-2),  46 (1979).

 Lusby, W. R., and Hill, K.  R.,  Bull.  Environ.  Contam.  Toxicol..
 .22(4-5) , 576  (1979).                         "*"~	—

 Rosewell, K. T., and  Baker,  B.  E.,  Bull.  Environ.  Contam.  Toxicol.,
 21(4-5), 470  (1979).	

 De Beer, J. 0., Van Peteghem,  C. H.,  and  Heyndrickx, A. M., J. Assoc.
 Off. Anal. Chem.. 61, 1140  (1978).                           ~~	

 Glotfelty, D. E., Anal. Chem..  44,  1250 (1972).

 Banks, K. A., and Bills, D.  D., J.  Chromatogr..  33, 450 (1968).

 Cochrane, W. P., and Maybury, R. B., J. Assoc. Off. Anal.  Chem.,
 5£, 1324 (1973).              .       	"""~	—-

 Watts, R. W., Hodgson, D. W., Crist, H. L.,  and  Moseman, R.  F.,
 J.  Assoc. Off. Anal. Chem.. .63(5),  1128 (1980).

 Shafik, T.  M., Bradway, D.,  and Enos, H.  F., Bull. Environ.  Contam.
 Toxicol.. 6., 55  (1971).                          	
 McCully,  K.  A., J. Assoc. Off. Anal. Chem.. J50, 376  (1977).

 Singh,  J.,  and Lapointe, M* R., J. Assoc. Off. Anal. Chem., 57.
 1285 (1974).	 —

 Coburn, J.  A., and Chau, A. S. Y., J. Assoc. Off. Anal. Chem.. 57,
 1272 (1974);  Environ.  Lett.. 10, 225 (1975).—

 Forbes, M.  A., Wilson, B. P., Greenhaigh, R., and Cochrane, W. P.,"
 Bull. Environ. Contam. Toxicol.. 13. 141 (1975).

 McCully, K. A., J. Assoc. Off. Anal. Chem.. .62(2), 385 (1979).

 Cochrane, W.  P., and Greenhaigh, R., Confirmation of Pesticide
 Residue Identity by Chemical Derivatization, presented at 166th
 National ACS  Meeting,  Chicago, August 29, 1973:  Khan, S. U.,
 Greenhaigh, R., and Cochrane, W. P., J.  Agr. Food Chem.. 23, 430  (1975).

 Lawrence, J.  F., J.  Agric.  Food Chem..  22,  936 (1974).

 Lawrence, J.  G., Lewis,  D., and McLeod,  H.  A*, J.  Agric.  Food Chem.,
.25,  1359 (1977).                               ~	   •	

 Lawrence, J.  F., J.  Aerie.  Food Chem..  2±,  1236 (1976).

 Lawrence, J.  F., Lewis,  D.  A., and McLeod,  H.  A.,  J.  Chromatogr.
 138, 143 (1977).	&—
                                 -436-

-------
                                                        Section 10Q

(42>  Miles, J. W. „ and Dale, W. E., J. Aerie. Food Chen., 26i, 480 (1978).

(43)  Greenhalgh, R., King, R. R., and Marshall, W. D., J. Agric. Food
      Chem.. 26, 475 (1978).

(44)  Magallona, E. D., and Gunther, F. A., Arch. Environ. Contam. Toxicol..
      5> 185 (1977).

(45)  Lawrence, J. F., J. Chromatogr.. 123, 287  (1976).

(46)  Worobey, B. L., and Webster, G. R. B.,  J.  Assoc. Off.  Anal. Chem..
      60, 213  (1977).

(47)  Noshe, N., Kobayoshi, S.,  Tanaka, A., Hirose, A.,  and  Watanabe,  A.,
      J. Chromatogr.. 130.  410 (1977).

(48)  Wein, R. G". ,  and Tanaka, F.  S., J. Chromatogr..  130.  55 (1977).

(49)  Bromilow, R.  H., and  Lord, K.  A., J.  Chromatogr..  125. 495  (1976).

(50)  Lawrence, J.  F., J. Assoc. Off. Anal.  Chem..  59. 1061 (1976).

(51)  Lawrence, J.  F., and  Sundaram, K. M.  S.,  J.  Assoc. Off. Anal.  Chem..
      59,  938  (1976).

(52)  Lawrence,  J.  F., and McLeod, H.  A.,  J.  Assoc.  Off. Anal. Chem.. 59.,
      637,  (1976).

(53)  Singh,  J.,  and Cochrane, W.  P.,  J.  Assoc. Off.  Anal.  Chem.. 62_(4),
      751 (1979).

(54)  Stoks,  P. G., and Schwartz,  A. W.,  J. Chromatogr.. 168(2), 455 (1979).

 (55)  Krause,  R.  J., J.  Chromatogr.. 185. 615 (1979).

 (56)  Aly, 0.  M., Faust, S. D., and Suffet,  I. H., Ultraviolet Spectro-
      photometry in Residue Analysis; Spectra-Structure Correlations,
      Advances in Chemistry Series 104, Chapter 7, ACS,. Washington, D.C.,
       1971, page 95.

 (57)   Gore, R. C., Hannah, R. W., Pattacini, S. C., and Porro, T. J.,
       J. Assoc. Off. Anal. Chem.. 54, 1040 (1971).

 (58)   Cyr, T., Cyr, N., and Haque,  R., Spectrophotometric Methods in
       Analytical Methods for  Pesticides and  Plant Growth Regulators,
       Zweig,  G., and Sherma,  J.,  eds., Volume  IX, Chapter 3,  Academic
       Press,  New York,  1977,  page 75.

 (59)  MacDougall,  D., Residue Rev.. JL, 24  (1962); 5.,  119  (1964).
                                   -437-

-------
                                                         Section 10Q
 (60)





 (61)

 (62)

 (63)



 (64)


 (65)





 (66)

 (67)

 (68)


 (69)


 (70)


 (71)


 (72)

 (73)


(74)


(75)

(76)
 Argauer, R. J., Fluorescence Methods for Pesticides, in Analytical
 Methods for Pesticides and Plant Growth Regulators. Zweig, G.,
 and Sherma, J., eds., Volume IX, Chapter 4, Academic Press, 1977,
 New York, page 101.

 Baeyens, W., Pharm. Weekbl.. Ill, 1075 (1976).

 Sloane, H.  J., and Obremski, R. J. t Appl. Spectrosc.. 31, 506 (1977).

 Blinn, R. C., The Identification of Pesticides at Residue Concentra-
 tions, Advances in Chemistry Series 104. Chapter 6, ACS, Washington,
 B.C.,  1971, page 81.

 Kettmann, R., Closset, J.-L., Copin, A'., Duculot, C., and Martens,
 P.-H., Anal. Lett., £, 1013 (1973).

 Cyr, T., Cyr, N., .and Haque, R., Infrared Spectrometry, in
 Analytical Methods for Pesticides and Plant Growth Regulators.
 Zweig, G.,  and Sherma, J., eds., Volume IX, Chapter 5,  Academic
 Press, New York,  1977, page 137.

 Crosby, N.  T.,  and Laws,  E. Q., Analyst. 89, 319 (1964).

 Neiswender, D.  D.,  J.  Gas Chromatogr.,  ,4, 426 (1966).

 Shearer, D. A.,  Stone, B.  C., and McGugan,  W.  A., Analyst, 88,
 147  (1963).                -                       	*—  —

 Bierl, B. A., Beroza,  M.,  and Ruth,  J.  M.,  J.  Gas Chromatogr., 6,
 286  (1968).                                	•—

 Scanlon, R. A., Arnold, R.  G.,  and Lindsay, R.  C., J. Gas Chromatogr.
 6., 372 (1968).                                      	:	—

 Nano,  G. M.,  Sancin, P.,  and  Martelli,  A.,  J.  Gas Chromatogr.. 3,
 85 (1965).                                  	—

 Hannah,  R.  W.,  and  Dwyer,  J.  L.,  Anal.  Chem.,  36. 2341  (1964).

 Hammerstrand, K., Juntunen, J.  M.,  and  Henes, A.  R., Anal.  Biochem.
.27, 172 (1969).	
Curry, A. S., Read, J. F., Brown, C.,  and Jenkins,  R. W.,  J.  Chromatogr.,
.38, 200  (1968).                                            "	5—

Hartman, K. T., J. Assoc. Off. Anal. Chem., 50,  615 (1967).

Giang, P. A., Library of  Infrared Spectra of  Important Pesticides,
ia Analytical Methods for Pesticides and Plant Growth Regulators.
Zweig, G., and Sherma, J., eds., Volume IX, Chapter 6, Academic
Press, New York, 1977, page 153.
                                  -438-

-------
                                                        Section 10Q
(77)   Nicholas, M.  L.,  J. Assoc. Off. Anal. Chem.. 59, 197, 1071, 1266
      (1976).

(78)   Giang, P. A., Agric. Res. Serv. NE-91. 95 pp (1978).

(79)   Cournoyer, R., Shearer, J. C., and Anderson, D. H., Anal. Chem..
      .49, 2275 (1977).

(80)   Cyr, N., Cyr, T., and Haque, R., Nuclear Magnetic Resonance
      Spectroscopyr in Analytical Methods for Pesticides and Plant
      Growth Regulators. Zweig, G., and Sherma, J., eds., Volume IX,
      Chapter 2, Academic Press, New York, 1977, page 51.

(81)   Wilson, D. M., Olsen, R. W., and Burlingame, A. L., Rev. Sci.
      Instrum.. 45. 1095 (1974).

(82)   Biros, F. J., J.  Assoc. Off. Anal. Chem.. 53, 733  (1970).

(83)   Wilson, N. K., J. Am. Chem.  Soc., 94, 2431  (1972).

(84)   Ross, R* T.,  and Biros, J. F., Bjochem. Biophys. Res. Commun.,  39,
      723 (1970).

(85)   Ross, R. T.,  and Biros, F. J., Mass Spectrometry and NMR Spectroscopy
      in Pesticide Chemistry, Plenum Press, New York, N.Y., 1974, pages
      263-272.

(86)   McKinney, J.  D., Wilson, N.  K., Keith, L. H., and Alford, A. L.,
      Mass Spectrometry and NMR Spectroscopy in Pesticide Chemistry.
      Plenum Press, New York, .N.Y., 1974, pages 139-160.

(87)   Crombie, L.,  and Lown, J. W., J. Chem. Soc.. 1962» 775.

(88)   Wilson, N. K., J. Phvs. Chem.. 75, 1067 '(1971).

(89)   Keith, L. H., Garrison, A. W., and Alford, A. L., J. Assoc. Off.
      Anal. Chem..  51, 1053  (1968).

C90)r Keith, L. H., Alford, A. L., and Garrison, A. W., J. Assoc. Off.
   ""  Anal. Chem.,  52., 1074  (1969).

(91)   Keith, L. H., and Alford, A. L., J. Assoc.  Off. Anal. Chem.. 53,
      157 (1970).

(92)   Wilson, N. K., Nuclear Magnetic Resonance Spectroscopy in Pesticide
      Chemistry, in Pesticide Analysis, Das, K, G., ed., Marcel Dekker,
      Inc., New York, 1980.

(93)   Keith, L, H., and Alford, A. L., J. Assoc.  Off. Anal. Chem..53.
      1018  (1970).
                                    -439-

-------
                                                          Section 10Q

  (94)  Yamamoto,  0., Yanagisawa,  M.,  Hayamizu,  K.,  and Kotowycz, G.,
       J. Magn. Reson.. j>,  216  (1973).

  (95)  Wilson, N. K.,  and Anderson, M.,  Mass  Spectrometry and NMR
       Spectroscopy  in Pesticide  Chemistry. Plenum  Press,  New York,  N.Y.,
       1974, pp.  197-218.

  (96)  Wilson, N. K.,  J. Am.  Chem. Soc.. 97,  3573 (1975).

  (97)  Cox, R. H., and McKinney,  J. D.,  Ore.  Magn.  Reson.. 11(11), 54i
       (1978).

  (98)  Wilson, N. K.,  and Sovocool, G. W., Org. Magn.  Reson..  9., 536 (1977).

  (99)  Alley, E.  G., Layton,  B. R., and  Minyard, J. P.,  Jr.;  J.  Org.  Chem..
       41, 462 (1976).                                       	 	

(100)  Wilson, N. K.,  J. Phys.  Chem., 83_, 2649  (1979).

(101)  Wilson, N. K.,  and Zehr, R. D., J. Org.  Chem..  44.  1278 (1979).

(102)  Roll, D. B., and Biros,  F. J., Anal. Chem..  41,  407 (1969).

(103)  Gegiou, D., Anal. Chem.. 4£, 742  (1974).

(104)  Gegiou, D., Talanta. 21, 889 (1974).

(105)  Ross, R. T., and Biros,  J. F., Anal. Chem..  52,  139 (1970).

(106)  Gurley, T. W.,  and Ritchey, W. M., Anal. Chem..  48. 1137  (1976).

(107)  Benyon, J. H., Recommendations for Symbolism and Nomenclature  for
       Mass Spectroscopy, 26th Annual Conference on Mass Spectrometry
       and Allied Topics, IUPAC, May 28-Jutte  2, 1978, St.  Louis, MO.

(108)  Safe, S.,  Environmental Applications of Mass Spectrometry, in
       Mass Spectrometry. Vol.  5, Johnstone,  R. A. W., Sr. Reporter,    (
       The Chemical Society, London, 1979, Chapter  9, p. 234.

(109)  Skinner, S. I. M., and Greenhalgh, R., Mass  Spectra of  Insecticides.
       Herbicides and Fungicides and Metabolites. Chemistry and Biology
       Research Institute, Agriculture Canada, Ottawa, Ontario, Canada, 1977.

(110)  Ryan, J. F., Residue Analysis Applications of Mass  Spectrometry,
       in Analytical Methods for Pesticides and Plant Growth Regulators.
       Zweig, G., and Sherma, J., eds., Volume IX, Chapter 1, Academic   "-'-••'
       Press, New York, 1977, p. 1.
                                     -440-

-------
                                                           Section 10Q
(111)




(112)

(113)

(113A)


(114)

(115)



(116)

(117)




(118)


(119)



(120)

(121)


(122)


(123)


(124)


(125)

(126)
Burlingame, A. L. , Baillie, T. A., Derrick, P. J., and Chizov, 0. S.,
Anal. Chem.. 5.2(5), 214R (1980);  (see also earlier reviews of mass
spectrometry in The Fundamental Reviews issues of Anal. Chem.,
issue number 5 each even year).

Alford, A., Biomed. Mass Spectrom..4. 1  (1977);  5_,  259 (1978).

Biros, F. J., Residue Rev.. 4jO, 1 (1971).

Kuehl, D. W., Whitaker, M. J., and Dougherty, R.  C., Anal. Chem..
52., 935  (1980).

Grayson, M. A., J. Chromatogr. Sci.. 15.  539  (1977).

McFadden, W. H., Techniques of Combined Gas Chromatography/Mass
Spectrometry;  Applications in Organic Analysis,  John Wiley  & Son,
New York, 1973.

McFadden, W. H., J. Chromatogr. Sci.. 17, 2  (1979).

Wright,  L. H.,  and Oswald, E. 0., Liquid  Chromatography/Mass
Spectrometric Methods  of Analyses for Garbamate  Pesticides,
presented at  the 26th  Annual  Conference on Mass  Spectrometry and
Allied Topics,  St. Louis,  MO, May-June', 1978.

Fales, H. M., Milne, G. W. A., Winkler, H. U., Beckey,  H.  D.,
Damico,  J. N.,  and Barren, R., Anal. Chem.,  47,  207  (1975).

Hunt, D. F.,  Stafford, G.  C., Crow,  F., and  Russell, J. W.,
24th Annual Conference on  Mass Spectrometry  and  Allied  Topics,
Meeting  Abstract No. AA5,  1976.

Stan,.H.-J.,  Z. Anal.  Chem..  287. 104  (1977).

Holmstead, R. L.,  and  Casida, J.  E., J. Assoc.  Off.  Anal.  Chem..
57,  1050 (1974).
 Damico,  J.  N., Barren, R. P., and Sphon, J. A., Int. J. Mass Spectrom.
 Ion Phys.,  2± 161 (1969).

 Heck, H. , Dyer, R. L., Scott, A. C., and Anbar, M., J. Agric. Food
 Chem.. 25,  901 (1977).

 Schulten, H. R., and Beckey, H. D., J. Agric.'Food Chem.. 21, 372
 (1973);  24, 743 (1976).

 Schulten, H. R., 2. Anal. Chem.. 293(4), 273 (1978).

 Yamato,  Y., Suzuki, M., and Watanabe, T., Biomed. Mass Spectrom..
 £(5), 205 (1979).
                                     -441-

-------
                                                            Section 10Q


 (127)   Carroll,  D.  I., Dzidic, I., Stillwell, R. N., Horning, M. G., and
        Horning,  E.  C., Anal. Chem.. 46. 706 (1974).

 (128)   Carroll,  D.  I., Dzidic, I., Stillwell, R. N., Haegele, K. D.,
        and Horning, E. C.,  Anal. Chem.. £7, 2369 (1975).

 (129)   Horning,  E.  C., Carroll, D. I.,  Dzidic, I., Lin,, S.-N., Stillwell,
        R.  N.,  and Thenot, J.-P., J.  Chromatogr., 142. 481 (1977).

 (130)   Hunt, D.  F., Stafford,  G. C., Crow, F.  W., and Russell, J. W.,
        Anal. Chem.. 4£,  2098 (1976).

 (131)   Haas, J.  R., Friesen, M. D.,  Harvan, D. J., and Parker, C. E.,
        Anal. Chem., 50.  1474 (1978).
                                                   •

 (132)   Biros,  F.  J., Dougherty, R. C.,  and Dalton, J., Qrg.  Mass. Spectrom..
       .6,  1161 and  1171  (1972).                                 *~~~^	"

 (133)   Dougherty, R. .C.,- Roberts,  .J. D., and Biros, F. J., Anal.  Chem..
        47,  54  (1975).                                      '	~	'

 (134)   Dougherty, R. C., and Piotrowska, K. K.,  J.  Assoc. Off.  Anal.  Chem..
        59_,  1023  (1976).                          —____^	.	


 (134A) Hass, J. R., and Friesen, M.  D.,  Ore. Mass Spectrom..  14
        9 (1979).                        —*	*—	   —'

 (134B) Hunt, D. F., Harvey,  T.  M., and  Russell,  S.  W., J. Chem.  Soc.  Chem.
        Commun., 151 (1975).                             	"^	

 (135)  Milne, G. W. A., and  Lacey, M. J.,  CRC  Grit. Rev. Anal.  Chem..  4,
       45-104  (1974).                      ~'	'	—:	  ~

 (136)  Milberg, R.  M., and Cook, J.  C.,  Jr., J.  Chromatogr. Sci., 17,
       17  (1979).                            	   	 —

 (137)  Busch, K. L., Bursey, M.  M.,  Haas,  J. R.,  and  Sovocool,  G. W.,
       Appl. Spectrosc.. 3^(4),  388  (1978).

 (138)  Abramson, F.  P., Anal. Chem.. <44_(14), 28A (1972).

 (139)  Oswald,  E. 0., Albro, P. W.,  and McKinney, J. D.. J.  Chromatogr..
       98,  363 (1974).                                    	£~

 (140)  Vander Veld,  G., and Ryan,  J. F., J. Chromatogr. Sci.. 13_, 322  (1975),

(141)  Dougherty, R. C., in Biochemical Applications of Mass  Spectrometrv.
       First Supplementary Volume. Waller, G.  R., and Dermer, 0.  C., eds.,
       Wiley-Interscience,  New York, 1980, Chapter 32A.
                                     -442-

-------
                                                           Section 10Q


(142)   Sovocool, G.  W.,  and Lewis, R. G., The Identification of Trace.
       Levels of Organic Pollutants in Human Tissues:  Compounds Related
       to Chlordane/Heptachlor Exposure, in Trace Substances in Environ-
       mental Health, IX, 1975, Hemphill, D. D., ed., University of
       Missouri Press,  1976, pp. 265-280.

(143)   Wright, L. H., Lewis, R. G., Grist, H. L., Sovocool, G. W.., and
       Simpson, J. M.,  The Identification of Polychlorinated Terphenyls
       at Trace Levels in Human Adipose Tissue by Gas Chromatography/Mass
       Spectrometry, J.  Anal. Toxicol., 2., 76-79 (1978).

(144)   Weil, L., Frimmel, F., and Quentin, K.-E., Z. Anal. Chem.. 268,
       97 (1974).

(145)   Rosen, J. D., and Pareles, S. R., Mass Spectrometry and NMR Spectro-
       scopy in Pesticide Chemistry. Plenum Press, New York, N.Y., 1974,
       pp. 91-98.

(146)   Chapman, R. A.,  and Robinson, J. R., J. Chromatogr.. 140, 209  (1977).

(147)   Bergstedt, L., and Widmark, G., Chromatographia. 3_, 59 (1970).

(148)   Fenselau, C.  Anal. Chem.. j49_(6), 563A (1977).

(149)   Budde, W. L., and Eichelberger, J. W., J. Chromatogr.. 134. 147  (1977).

(150)   Kuehl, D. W., Anal. Chem.. 49. 521 (1977).

(151)   .McLafferty, F. W., and Veiikataraghavan, R., J. Chromatogr. Sci.. 17,
       24 (1979).               '

(152)   Damico, J. N., Barren, R. P., and Ruth, J. M., Organic Mass Spectro-
       metry, JL, 331 (1968); Damico, J. N., J. Assoc. Off. Anal. Chem.. 49,
       1027  (1966); Benson, W. R., and Damico, J. N., J. Assoc. Off.  Anal.
       Chem., 48, 344 (1965); 51, 347 (1968).

(153)   Mestres, R., Chevallier, C., Espinoza, C., and Corner, R., Ann.  Falsif.
       Expert. Chim.. .70(751), 177 (1977).

(154)   Zeman, A., and Woerle, R., Org. Mass Spectrom.. 13(1), 43 (1978).

(155)   Leclercq, P. A.,  and Pacakova, V., J. Chromatogr.. 1.78(1), 193 (1979).

(156)   Stan, H.-J., Abraham, B., Jung,  J., Kellert,  M., and Steinland,  K.,
       Z. Anal. Chem.. 287. 271 (1977).

(157)   Biros, F. J., and Sullivan, H.,  cited in (113).

(158)   Biros, F. J., Applications of Combined GC/MS  to Pesticide Residue
       Identification, presented at the ACS-Chem. Inst. Canada International
       Meeting, Toronto, May, 1970.
                                    -443-

-------
                                                            Section 10Q
(159)  Van Peteghem,  C. H.,  and Hendrickx,  A.  M.,  J.  Agric.  Food Chem..
       24, 635  (1976).                                    ~——*-

(160)  Shafik,  T. M., Biros,  F.  J.,  and Enos,  H. F.,  J.  Aerie.  Food Chem..
       18_, 1174 (1970).             '         . .

(161)  Biros, F. J.,  and Ross,  R.  T.,  Fragmentation Processes  in the Mass
       Spectra  of Trialkylphosphates,  Phosphorothionates, Phosphorothiolates,
       and Phosphorodithiolates, Presented  at  18th Conference  on Mass
       Spectrometry and Allied  Topics, San  Francisco, CA, June, 1970.

(162)  Stan, H.-J., Z. Lebensm.  Unters.  Forsch.. 164. 153  (1977).

(163)  Biros, F. J.,  and Walker, A.  C.,  J.  Agric.  Food.Chem..  18, 425 (1970).

(164)  Bellman, S. W., and Barry,  T. L.,  J.  Assoc. Off.  Anal.  Chem.. 54,
       499 (1971).                        'i    ~"      ~"   ~~     "  ~~

(165)  Suzuki, M., Yamato, Y.,  and Koga,  M., Biomed.  Mass Spectrom..
       5.(9), 518 (1978).                            '."'-.

(166)  Harless, R. L., and Oswald, E.  0., Gas  Chromatography/Mass Spectro-
       metric Methods of Analysis  for  Toxaphene and Dioxins  in Human and
       Biological Samples, presented at  the 26th Annual  Conference on
       Mass Spectrometry & Allied  Topics, St.  Louis,  MO, May-June, 1978.

(167)  Harless, R. L., Harris, D.  E.,  Sovocool, G. W., Zehr, R. D.,
       Wilson, N. K., and Oswald,  E. 0.,  Biomed. Mass Spectrom..  5(3),
       232 (1978).

(168)  Sovocool, G. W., Lewis, R.  G.,  Harless, R. L.,  Wilson,. N.  K., and
       Zehr, R. D., Anal. Chem.. 49, 734  (1977).

(169)  Van den Heuvel, W. J. A., Wood, J. S.,  DiGiovanni, M.,  and Walker,
       R. W., J. Agric.  Food Chem..  25.  386 (1977).

(170)  Lee, Y. W., and Westcott, N.  D., J.  Assoc. Off. Anal. Chem..  62(4).
       782 (1979).                      ~~~~~          . ' '  '   |~~*'

(171)  Laseter, J. L., DeLeon, I. R.,  and Remele, P.  C., Anal.  Chem..
       50_(8), 1169  (1978).                                         "~

(172)  Bagley, E. G., Reichel, W. L. ,  and Cromartie,  E., J. Assoc. Off.
       Anal. Chem.. 5.3,  251  (1970)'.                                '

(173)  Bonelli, E. J., Anal. Chem..  44, 603  (1972).  ,

(174)  Canada, D. C., and Regnier, F.  E., J. Chromatogr. Sci..  14, 149 (1976).
                                      -444-

-------
                                                           Section 10Q


(175)   Neher,  M.  B.,  and Hoyland, J.  R., Specific Ion Mass Spectrometric
       Detector for Gas Chromatographic Pesticide Analysis, U.S. Environ-
       mental Protection Agency, Washington, DC, Report No. EPA-660/2-  .
       74-004, January, 1974.

(176)   Thruston,  A.  D., Jr., A Quantitative Method for Toxaphene by
       GC/CIMS Specific Ion Monitoring,  U.S. Environmental Protection
       Agency, Washington, DC,. Report No. EPA-600/4-76-010, March, 1976.

(177)   Dyer, R. L.,  Heck, H. d'A., Scott, A. C., and Anbar, M., Feasibility
       of Applying Field lonization Mass Spectrometry to Pesticide Research,
       U.S.  Environmental Protection Agency, Washington, DC, Report No.
       EPA-600/1-76-037, November/ 1976.

(178)   Ryan; J. F.,  Harless, R.L., and Lewis, R. G., Application of Field
       lonization Mass Spectrometry to Environmental Analysis, Proceedings
       of, the 23rd Annual Conference on Mass Spectrometry and Allied
       Topics, Houston, TX,  May 25-30, 1975, pp. 46-48.

(179)   Horning, E.  C.,  Carroll, C. I., Dzidic, I., Stillwell, R., and
       Thenot, J.-P,, J. Assoc. Off.  Anal. Chem., 61, 1232 (1978).

(179A)'Hunt, D. F.,  Shabanowitz, J.,  and Giordani, A. B., Anal. Chem. .52.
       386-390 (1980).

(179B) Yost, R. A.,  and Enke, C. G.,  Anal. Chem., 51(12), 1251A (1979).

(179C) Cooks, R.  G.,  Amer. Lab., p. Ill, October (1978).

(180)   Lewis, R.  G.,  Research Report of the Analytical Chemistry Branch,
       ETD,  HERL, RTP Program Element No. 1EAG15, Sovocool, G. W., and
       Wright, L. H., principal investigators.

(181)   Heller, S. R., and Milne, G. W. A., Environ. Sci. Technol., 13(7),
       798-803 (1979).

(182)   Budde, W.  L.,  and Eichelberger, J. W., Organic Analysis Using Gas
       Chromatography/Mass Spectrometry, EPA 600/8-79-006, Environmental
       Monitoring and Support Laboratory, Office of Research and Development,
       U.S.  EPA,  Cincinnati, Ohio, March, 1979.  Ann Arbor Science
       Publishers (Wiley), Ann Arbor, MI, 241 pp (1979).

(183)   Sun,  Y. P., Analytical Methods for Pesticides and Plant Growth
       Regulators, Zweig, G., ed., Vol. 1, Academic Press, New York, N.Y.,
       1963, page 571.

(184)   Allen, P.  T., Analytical Methods for Pesticides and Plant Growth
       Regulators, Zweig, G,, ed. Vol. V, Chapter 3, Academic Press,
      • New York, 1967, page 67.
                                     -445-

-------
                                                            Section 10Q


 (185)   Gajan,  R.  X.,  Residue Rev.  5, 80 (1964); £, 75 (1964).

 (186)   Smyth,  M.  R..,  and Smyth, W.  F., Analyst. 103(1227), 529 (1978).

 (187)   Smyth,  M.  R.,  and Osteryoung, J. G., Anal. Chim.  Acta.  96, 335  (1978).

 (188)   Gajan,  R.  T.,  J.  Assoc.  Off.  Anal.  Chem.. 5_2, 811 (1969).

 (189)   Zietek, M., Mikrochim.'' Acta.  j21(l-2) , 75 (1979).

 (190)   Hance,  R.  J.,  Pestic.  Sci.. ^, 112  (1970).

 (191)   Booth,  M.  D.,  and Fleet, B.,  Talanta. 17, 491 (1970).

 (192)   Kutlukova, V.  S.,  Toropov, A. P., and Lozovatskaya, M. A., Anal. Abstr..
       .27, Abstract No.  2940  (1974).

 (193)   Farwell, S. 0., Beland,  F. A., and  Geer, R.  D.,  Bull.  Environ. Contam.
        Toxicol..  10,  157 (1973).                             "   '.      '

 (194)   Zietefc, M., Mikrochim. Acta.  II (5-6), 549 (1976).

 (195)  Marchidan, S., Rev. Roum. Chim., 22^  127 (1977).

 (196)   Filimonova, M. M., Zh. Anal.  Khim., 32_,  140  (1977).

 (197)  Budnlkov, G. K.,  Zh. Anal. Khim., 32_,  212 (1977);  3£ 2275  (1975).

 (198)  Nangniot, P.,  Zenon-Roland, L.,  and Berlemont-Frennet, M.,
       Analusis, 6(6), 273 (1978).

 (199)  Southwick, L. M., Willis, G.  H., Dasgupta, P. K.,  and  Kesztheli,
       C. P., Anal. Chim. Acta, 82,  29  (1976).

 (200)  Osteryoung, J. G., Anal. Chem.,  49, 2310 (1977).

 (201)  Filimonova, M. M., Zh. Anal.  Khim., 32,  812  (1977).

 (202)  Smyth, M.  R., and Osteryoung,  J. G., Anal. Chem.,  50(12),  1632 (1978).

 (203)  Wade, A. L., Hawkridge,  F. M., and  Williams, H. P.,  Anal.  Chim. Acta.
       105(1), 91 (1979).                                   	

 (204)  Davidek, J., Nemethova, M., and  Seifert,  J., Z. Anal.  Chem., 287,
       286 (1977).                                  	'	 	

(205)  Franke, G., Pietrulla, W., and Preussner, K., Z. Anal.  Chem.,  298(1),
       38 (1979).                                           ~~	
                                     -446-

-------
                                                       Section 10Q


(206)  Beroza, M., and Inscoe, M. N., in Ancillary  Techniques of Gas
       Chromatography. Ettre, L. S., and McFadden,  W.  H.,  eds.,  Wiley-
       Interscience, N.Y., 1969, pages 89-144.

(207)  Beroza, M., and Goad, R. A.,  J. Gas Chromatogr.,  4.,  199 (1966).

(208)  Beroza, M., J. Org. Chem.. £8,. 3562 (1963).

(209)  Beroza, M., J. Gas Chromatogr.. 2., 330  (1964).

(210)  Asai, R. I., Gunther, F. A. ,  and Westlake, W. E., Residue Rev.,
       J.9, 57 (1967).

(211)  Asai, R. I., Gunther, F. A., Westlake, W. "E., and Iwata,  Y.,
       J. Agr. Food Chem.. 20, 628  (1971).

(212)  Zimmerli, B., J. Chromatogr., 88., 65  (1974).

(213)  Cooke, M., Nickless, G., Prescott, A. M., and Roberts,  D.  J.,
       J. Chromatogr.. .156(2) , 293  (1978) ; Prescott, .A.'M., and  Cooke,  M.,
       Proc. Analyt.. Div. Chem. Soc., 16_(1), 10  (1979).

(214)  Minyard, J. P., and Jackson, E. R., J. Agr.  Food  Chem., 13, 50  (1965),

(215)  Erney, D. R., Anal. Lett.. 12(A5). 501  (1979).
                                    -447-

-------
                               Section 11
                TRAINING OF PESTICIDE ANALYTI01 CHEMISTS
 This chapter is by far the shortest in the manual,  and the reader nay
 question the logic of devoting a special section to this subject.  During
 the years of operating the interlaboratory quality  control program des-
 cribed in Section 2,  the editors observed overwhelming evidence  that
 participating laboratories with chemists who had formalt specialized
 training demonstrated far superior analytical performance than did those
 laboratories which lacked this advantage.  We, therefore,  regard training
 as  a highly important subject and deserving of special treatment.

 Although many good programs are available in undergraduate and graduate
 schools for the training of analytical chemists,  few,  if any, specifically
 train pesticide analysts.   An undergraduate or graduate student  on a
 research project with a professor interested in development of residue
 analytical methods does receive valuable training and  experience, but
 such, professors are few and far between in American education institutions.
 A number of companies in the private sector offer short courses  particularly
 designed for training users of company-produced equipment.  A certain few
 universities and private educational organizations  run short courses
 touching upon a few of the highlights of pesticide  residue analysis.
 Some governmental agencies operate similar short  training  courses.

 The residue chemist must not only be familiar with  the technique of trace
 analysis in general and of residue analysis in particular, but he must be
 able to  perform routine service and adjustments and preventative main-
 tenance,  such as module replacements  and replumbing, on his instruments.
 In  order to achieve these  abilities,  a generally  trained analytical chemist
 should be given on-the-job training by an experienced  residue chemist when
 he  is hired,  if at  all  possible.   Since this  is often not possible,
 especially in smaller laboratories, this Manual is  designed to substitute,
 in  small part,  for  such training  and  to help  the  analyst recognize certain
 pitfalls  and to better  perform analyses of  biological and environmental
 media.   There is, however, no  really  satisfactory substitute for intensive,
 practical bench training of  the type  formerly provided by  the EPA Perrine
 Primate Laboratory  Training  Program, Perrine, Florida.  During the years of
 conducting the  interlaboratory  quality control program described in  Section
 2,  it was very  apparent  that those  laboratories which took most advantage
 of  the Perrine  training  facility recorded far better analytical performances
 on  round  robin  samples  than  laboratories not participating in the training
 program.  As  a  specific  illustration of  this, the reader is referred  to
Table 11-1  (copied  from  Table 2-15  in  Section 2) which lists the relative
performance ranking of  34  laboratories  in one interlaboratory check sample
 exercise.
                                   -448-

-------
                                                         Section 11
The eight laboratories with top performance had previously sent personnel
to the Perrine training program.  Of the 17 laboratories in the top half
of the table, 10 of these laboratories had Perrine-trained chemists.  Of
the 17 laboratories in the lower quality half of the table, only one
laboratory near the top of the lower half had sent personnel for training.
All laboratories which had Perrine-trained personnel are check-marked
next to their identifying code numbers.

The editors feel that the data shown by this table provide  most conclusive
evidence of the value of a proper training program in the potential
quality output of a pesticide analytical chemist.  Unfortunately, however,
the agency saw fit to discontinue the Perrine training program, the only
one of its kind in existence, and, as stated in Section 2J, some recent
results on interlaboratory quality assurance fat check samples (see Table
2-23) indicate the need for a training program.

It is hoped that some educational institutions or governmental agency will
recognize the need and set up programs to provide such training, and that
laboratory supervisors will take advantage of these in urging their residue
chemists to obtain and refresh, on a continuing basis, their training and
knowledge in analytical and instrumental areas.  Rapid developments in
instrumentation and new techniques, and the need to analyze at lower and
lower levels for an ever increasing number of pesticides and metabolites,
dictate a constant need for training and retraining in a field as highly
complex as that of residue analytical chemistry.  Furthermore, recent
disclosures of pollution of the nation's air and water by a wide variety
of organic compounds, Including pesticides, point up the need for scientists
with a sound background of analytical expertise.
                                  -449-

-------
TABLE 11-1
Section 11
                     RELATIVE PERFORMANCE RANKINGS
               CHECK SAMPLE NO. 26, MIXTURE IN SOLVENT
Lab; Code
Number
/161.
/137,
/135.
/162.
/ 87.
/113A.
/113.
/ 85.
48.
130.
/ 66.
73.
/ 72.
84.
89.
88.
83,
96.
97.
164.
/ 68.
92.
93.
90.
53.
163.
95.
160.
45.
71.
52.
47.
69.
54.
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
-1
1
1
2
2
0
3
2
3
4
4
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.
0
. 0
0
1
1
- 0
0
1
0
0
1
6
1
0
0
1
0
0
4
No. of ..
Rejects -'
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
2
0
4
1
0
2
2
1
0
0
0
6
3
2
1
2
4
Total <,/
Score -'
198
198
197
197
197
197
196
196
. 195
195
195
194
194
192
192
189
189
187
181
169
168
. 168
164
159
158
157
146
133
128
127
123
115 "'
84
25
       If Values outside confidence limits
       2J Total possible score, 200 points
                                    -450-

-------
 a
AFID
AFS
API
AR
                                Section 12

                              ABBREVIATIONS*

              Selectivity
              alkali flame ionization detector
              amperes full scale
              atmospheric pressure  ionization
              analytical  reagent
BGS
BHC
BHT
             detector  background  signal
             hexachlorocyclohexane
             butylated hydroxytoluene
°C
CCD
CDEC
CI
cm
cone.
DCNA
DDA
CSC
CV
             degrees centigrade
             Coulson conductivity  detector
             sulfallate
             chemical ionization
             centimeter
             concentrated
             dichloran
             bis Cp_-chlorophenyl) acetic acid
             carbon skeleton chromatography
             coefficient of variation
2,4-D  •
DC or dc
DCS
ODD
DDE    ,
DDT
DDMU
DEF
DECS
DEPP   ,
DEPTP
DFTPP  •
DMF
DMSO
DNBP
DNFB   '
DNOC
             2,4-dichlorophenoxyacetic.acid
             direct current
             decachlorobiphenyl
             See TDE
             dichlorodiphenyldichlo'roethylene
             dichlorodiphenyltrichloroethane
             £.»£.' -DDD» olefin
             S,S,S-tributyl phosphorotrithioate
             diethylene glycol succ/nate
             (C2H50)2-PO-0-C6H5
             decafluorotriphenyl phosphine
             dimethylformamide
             dimethyl sulfoxide
             dinoseb
             2,4-dinitrofluorobenzene
             4,6-dinitro-o-cresol
EC
El
EICP
EPA
EPN
ETD
ETU
eV
             electron capture
             electron impact                .  .. • .
             extracted ion current profile
             Environmental Protection Agency
             0-ethyl 0-p_-nitrophenyl phenylphosphonothioate
             Environmental Toxicology Division
             ethylenethiourea
             electron volt
                                     -451-

-------
                                                            Section 12
 FD
 FDA
 FI
 FID
 FPD
 fsd
 FT
 8
 GC/MS
 GC
 GPC
 field desorption
 Food and Drug Administration
 field ionization
 flame ionization detector
 flame photometric detector
 full scale deflection
 Fourier transform
 gram
 gas chromatography coupled with mass spectrometry
 gas chromatography
 gel permeation chromatography
 HCB
 HECD
 HP
 HPLC
 HPTLC
 Hz
 hexachlorobenzene
 Hall electrolytic conductivity detector
 high performance
 high performance liquid chromatography
 high performance thin .layer chromatography
 hertz
 id
 IR
 •'•sat
 inside  diameter
 infrared     .                   .    ,     .
 maximum- current  from a  saturated  detector
k'
K-D
kg
capacity factor
Kuderna-Danish
kilogram
1 or L
LC
liter
liquid chromatography
MC
MCPA
MCPB
m/z
mg
MID
MIS
ml
mm
MOG
MS
MT
molecular ion
microcoulometric
[(4-chloro-p_-tolyl)oxy] acetic acid
4-[(4-chloro-o_-tolyl)oxy] butyric acid
mass to charge ratio
milligram
multiple ion detection
multiple ion selection
milliliter
millimeter
Mills, Onley, Gaither
mass spectrometry
Microtek
                                   -452-

-------
                                                            Section 12
N
ng
Ni
nm
NMR
N-P
number of theoretical plates
nanogram
nickel
nanometer
nuclear magnetic resonance
nitrogen-phosphorus
OC1
od
OP
organochlorine
outside diameter
organophosphorus
PAM          pesticide analytical manual
PBB          polybrominated biphenyl
PC           paper chromatography
PCB          polychlorinated biphenyl
POP          pentachlorophenol
PCT          polychlorinated terphenyl
PFTBA        perfluorotri-n-butylamine
pg           picogram
pH           measure of acidity; negative log of H+ concentration
PID          photoionization detector
PLOT         porous layer open tubular
PM           photomultiplier
PNP          4-nitrOphenol
ppb          parts per billion
ppm          parts per million
ppt          parts per trillion
psi          pounds per square inch
^-values     partition ratio of a solute between immiscible solvents
QA
QC
quality assurance
quality control
R
RF
Rj«
RRT
RSD
resolution
radiofrequency
ratio of distance moved by TLC spot to distance of solvent front
relative retention time
relative standard deviation
Rp value relative to that of a standard compound
s or SD      standard deviation
SCOT         support coated open tubular
SEU          standard error unit
SICP         selective ion current profile
SIM          selected ion monitoring
SIS          selected ion summation
SPED         sulfur-phosphorus emission detector
SPRM         standard reference material
                                  -453-

-------
                                                             .Section 12
T            total error
2,4,5-T      2,4,5-trichlorophenoxyacetic acid
TC           to contain
TCDD         2,3,7,8-tetrachlorodibenzo-p_-dioxin
TD           to deliver
TDE          DDD; 2,2-bis(£-chlorophenyl)-l,l-dichloroethane
THF          tetrahydrofuran
TIC          total ion current
TICP         total ion current plot
TLC          thin layer chromatography
P
VS
yi
ym

UV
V
vs.
WCOT
micron; also atomic mass units
microgram
microliter
micrometer

ultraviolet
volts
versus
wall coated open tubular
* For abbreviations, names, and formulas of pesticides not listed,
  see the U.S. EPA Analytical Reference Standards Manual (EPA-600/9-78-012).
                                     -454-

-------
                              Section  13
                  QUALITY CONTROL MANUAL REVISIONS

     This manual will be revised biennially dnd all persons  on  the
mailing list will automatically be mailed copies of the revisions.   The
question then for each manual holder is whether his name  is  in  fact
on the list.  Consider the following points:

1.  If you received this manual or a set of revisions in  response to a
    mail or phone request, you. are definitely on the list.     ... .

2.  If you received the manual as a handout at some training course,
    and your name and affiliation were not recorded, you  are probably
    not on the list and, therefore, will not automatically receive
    revisions.

3.  If you obtained your copy of the manual from some individual not
    associated with the Laboratory at Research Triangle Park, NC, you
    are probably not on the list and, therefore, will not automatically
    receive revisions.
     If, after reading the foregoing, there is a doubt that you may not
be on the mailing list, please clip off the section below, complete it
in full and mail it as shown to ensure that you will receive all future
revisions.                            ......
TO:  Quality Assurance Section, Anal. Chem. Br.  (MD-69)
     Environmental Toxicology Division
     EPA, Health Effects Research Laboratory
     Research Triangle Park, NC  27711

        This is to request that your record be reviewed to be certain the
     undersigned is on your mailing list to receive copies of all future
     quality control manual revisions.

                         (Print or type name and full business address)
                                   455
*U.S. GOVERNMENT PRINTING OFP1CE: 19 9 2 . 618 . 0 0 3/10 7 9 0

-------

-------