EPA-600/1-79-008
January 1979
WNUAL OF ANALYTICAL QUALITY CONTROL FOR PESTICIDES
AND RELATED (IMPOUNDS
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 Officers - Editors
Randall R. Watts and Jack F. Thompson
Environmental Toxicology Division
Health Effects Research Laboratory
Research Triangle Park, N.C. 27711
Revised: 1979
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
HEALTH EFFECTS RESEARCH LABORATORY
RESEARCH TRIANGLE PARK, N.C. 27711
-------
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 fo our modern developing, inducstrial society are
accompanied by certain hazrds. 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 natianal
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
-------
TABLE OF CONTENTS
Section In-Section
Number Section Page
PREFACE AND ACKNOWLEDGMENTS
1 GENERAL DESCRIPTION OF PESTICIDE RESIDUE
ANALYTICAL METHODS
A. Sampling 1
B. Extraction Procedures 1
C. Cleanup Procedures 2
D. Final Determination Methods 4
E. Confirmatory Techniques 7
F. Automation 8
G. References 9
2 INTERLABORATORY QUALITY CONTROL
A. Quality Control Program of the EPA Environ- 13
mental Toxicology Division (ETD) Laboratory
B. Program Objectives 14
C. Program Activities 14
D. Types and Preparation of Sample Media 15
E. Reporting Forms 16
F. Evaluation of Reported Data 19
G. Summary of Results Tables 19
H. Relative Performance Ranking 24
I. Private Critiques 27
J. Progression of Performance 27
K. Statistical Terms and Calculations 31
L. References 40
3 INTRALABORATORY QUALITY CONTROL
A. Purpose and Objectives 41
B. Purpose and Objectives of SPRM'S 41
C. Nature of SPRM'S 42
D. Frequency of SPRM Analysis 43
E. Record Keeping 44
F. Quality Control Charts 47
G. Benefits of the In-House SPRM Program 50
H. Analytical Balances 51
I. Purity of Reagents 53
J. Distillation of Solvents 56
K. Miscellaneous Reagents and Materials 56
L. Cleaning of Glassware 58
M. Analytical Pesticide Reference Standards 59
N. Calibration and Maintenance of the Gas Chromato-
graph and Accessories 68
0. Adherence to Official or Standard Methodology 71
P. Implications of an Intralaboratory Quality Control
Program 72
Q. References 73
-------
Section In-Section
Number Section Page
4 EVALUATION, STANDARDIZATION, AND USE OF MATERIALS
FOR PESTICIDE RESIDUE ANALYSIS
A. Adsorbents 75
B. GC Packings, Introduction and Column Technology 82
C. Column Efficiency and Peak Resolution 84
D. Sensitivity and Retention 87
E. Column Stability 88
F. Resistance to On-Column Compound Decomposition 88
G. Homemade vs. Precoated Packings 92
H. Packing the Column 95
I. Column Conditioning 97
J. Evaluation of the Column 99
K. Maintenance and Use of GC Columns 101
L. References 104
5 OPERATION OF THE GAS CHROMATOGRAPH
A. Temperature Selection and Control 106
B. Selection and Control of Carrier Gas Flow Rate 111
C. Electron Capture Detector Operation 114
D. Microcoulometric Detector Operation 124
E. Alkali Flame lonization Detector 127
F. Flame Photometric Detector 130
G. Electrolytic Conductivity Detector 137
H. Other Detectors 145
I. Electrometer and Recorder 147
J. Sample Injection and Injection Port 148
K. Erratic Baselines 151
L. GC Columns Recommended for Pesticide Analysis 152
M. Sensitivity of the GC System 159
N. Qualitative Analysis ]60
0. Quantitative Analysis 163
P. References 175
6 ADDITIONAL PROCEDURES IN PESTICIDE ANALYSIS
A. General Considerations 181
B. Representative vs. Biased Sampling 182
C. Sample Containers 182
D. Sample Compositing 183
E. Storage of Samples 183
F. Sampling of Agricultural and Food Products 185
G. Sampling of Biological Materials 185
H. Air Sampling 186
I. Water Sampling 188
VI
-------
Section In-Section
Number Section Page
J. Sampling of House Dust, Soil, and Stream
Bottom Sediment 192
K. Marine Biological Samples 193
L. Control of Procedures for Extraction of Residues 194
M. Control of Methodology for Concentration of Sample
Solutions and Fractional Eluates 197
N. Introduction to High Performance Liquid Column
Chromatography (HPLC) . 203
0. HPLC Instruments 204
P. Theory and Principles of HPLC 207
Q. Columns and Solvents for HPLC 209
R. Practical Aspects of Successful HPLC Operation 211
S. Applications of HPLC to Pesticide Analysis 214
T. Introduction to Thin Layer Chromatography (TLC) 219
U. Practical Considerations in TLC 220
V. Quantitative TLC 223
W. Thin Layer Systems 225
X. References 230
MULTIRESIDUE EXTRACTION AND ISOLATION PROCEDURES
FOR PESTICIDES AND METABOLITES
A. Tissue and Fat Analysis (Macro Technique) 240
B. Human or Animal Tissue and Human Milk (Micro
Technique) 242
C. Human Blood or Serum 242
D. Pentachlorophenol (PGP) in Blood and Urine 243
E. jD,£'-DDA in Urine 245
F. 2,4-D and 2,4,5-T in Urine 245
G. Kepone in Human Blood and Environmental Samples 246
H. Analysis of Fatty and Nonfatty Foods by the Mills,
Onley, Gaither Method 247
I. Chlorophenoxy Herbicides in Fatty and Nonfatty
Foods 248
J. Carbon-Cellulose Column Cleanup 248
K. Cleanup on Deactivated Florisil 249
L. Low Temperature Precipitation 251
M. Miscellaneous Multiresidue Cleanup Procedures 251
N. Alkyl Phosphate Metabolites in Urine, Blood, Tissues 255
0. Para-Nitrophenol (PNP) in Urine 256
P. Analysis of Fatty and Nonfatty Foods Using Florisil
Cleanup 257
Q. Sweep Co-Distillation 258
R. Charcoal Cleanup of Nonfatty Food Extracts 260
S. Miscellaneous Multiresidue Cleanup Procedures 260
Carbamate Pesticides and Miscellaneous Herbicides
T. 1-Naphthol in Urine 263
vii
-------
Section In-Section
Number Section Page
U. N-Methylcarbamate Insecticides in Blood and Fat 263
V. Analysis of Amine Metabolites in Urine 263
W. Other Indirect (Derivatization) Methods of
Analysis 264
X. Direct Methods of Analysis 264
Y. Analysis of Plant and Food Materials 265
Z. Air Analysis 266
A,A. Water Analysis 267
A,B. Soil, House Dust, and Bottom Sediment 269
Polychlorinated Biphenyls (PCBs)
A,C. Pesticide-PCB Mixtures 271
A,D. Appearance of PCB Chromatograms 272
A,E. Methods of Separation and Analysis of Pesticides
and PCBs 276
A,F. Determination of PBBs 283
A,G. Separation and Determination of Dioxins 284
A,H. Determination of Ethylenethiourea 285
A,I. Determination of Conjugated Pesticide Residues 285
A,J. Reviews of Analytical Methods for Pesticides,
PCBs, and other Non-Pesticide Pollutants 287
A,K. References 288
8 CONFIRMATORY PROCEDURES
A. Requirements for Positive Confirmation of
Pesticide Identity 301
B. GC Relative Retention Times 303
C. Selective GC Detectors 305
D. TLC RF Values 306
E. HPLC 307
F. Extraction p-Values 308
G. Derivatization Techniques 309
H. Spectrophotometry 316
I. Visible, UV, Fluorescence and Phosphorescence 317
J. Infrared (IR) 320
K. Nuclear Magnetic Resonance (NMR) 325
L. Mass Spectrometry (MS) and (GC-MS) 326
M. Quality Assurance of GC-Low Resolution MS 339
N. Biological Methods 352
0. Polarography (Voltammetry) 353
P. Miscellaneous Confirmatory Methods 354
Q. References 355
viii
-------
Section In-Section
Number Section Page
9 MAINTENANCE, TROUBLESHOOTING AND CALIBRATION OF
INSTRUMENTS
A. Daily Operational Considerations for GC
Instrumentation 367
B. Check List When Instrumental Repair is Indicated 368
C. General Approach to Troubleshooting 369
D. GC Service Block Diagram 372
E. Gas Inlet System of the GC 373
F. Isolation of Problems in Flow Systems in GC
Equipped with EC Detector 375
G. Temperature Control and Indication in the GC 380
H. Detector and Electrometer 382
I. Observation of Problems on Chromatograms 384
J. Detector Background Signal Response (BGS) 385
K. Troubleshooting the Microcoulometric Detector
System 386
L. Troubleshooting the Coulson Electrolytic
Conductivity System 392
M. Troubleshooting the Flame Photometric Detector 395
N. Epilog 397
10 TRAINING OF PESTICIDE ANALYTICAL CHEMISTS 398
11 REVISIONS TO THE QUALITY CONTROL MANUAL 401
ix
-------
PREFACE
The analysis for multi-pesticide residues in biological media surely
ranks high among the more difficult and exacting areas of analytical
chemistry. The methods are generally complex, involving (a) a series of
steps to extract or "strip" the pesticide residue from the substrate,
(b) a "cleanup" or partitioning procedure to eliminate a significant
portion of contaminants as well as delegating certain pesticidal compounds
to certain elution fractions, (c) assay of the residue content via some
instrumental technique such as gas chromatography, spectrophotometry, or
spectrofluorimetry, and (d) interpretation of the analytical data.
Not unexpectedly, the complexity of a typical residue analysis poses a
number of pitfalls into which the analyst may unknowingly stray unless
extreme care is constantly exercised to avoid the problems. The purpose
of this Manual is to present the field of residue analysis as it is
practiced today and to provide guidelines to the analyst to assist him in
the implementation of a systematic program of control. Such a program,
which will help ensure that the final analytical data are truly valid and
representative of the residue profile of the media analyzed, is known
as Analytical Quality Control.
Nearly all analytical chemists apply various elements of quality control
whether or not it is recognized by this name. Any step taken to produce
valid results classifies as quality control. Unfortunately, relatively
few analytical chemists follow a program of systematic guidelines and
for a variety of reasons. One of the more prominent reasons relates
to the time required, and the fact that time is equated with cost and
work output. The comment is often heard to the effect that "we practice
quality control when we are not busy".
Analytical data are always produced for some kind of decision-making,
whether it be to determine that some agricultural commodity is in com-
pliance with established tolerances (regulatory), or to establish the
profile of pesticide pollution in a waterway (monitoring). Closely con-
trolled data provide management with reliable evidence suitable for use
in a court or sufficiently reliable to make accurate assessments of
the degree of pollution in some sector of the environment. Uncontrolled
or improperly controlled data, no matter how high the volume, are worse
than no data at all because of the misleading effects.
Not only will this Manual discuss the problems and precautions in residue
analysis so as to serve as a guide in obtaining valid analytical results,
but it will describe a quality control approach for monitoring the per-
formance of chemists and laboratories performing residue analyses. The
emphasis is on methods for analyzing residues in biological and environ-
mental samples such as tissue, blood, urine, feces, air, water, and soil.
Less attention is given to analysis of agricultural and food products,
but approaches to quality control and analytical methods are, for the
most part, very similar.
-------
PREFACE TO THE FIRST REVISION
This first revision (1978) of the Quality Control Manual has resulted in
significant changes in all chapters. As stated in the preface to the
first edition of the Manual (above), the purpose continues to be the
presentation of guidelines for a quality control program in pesticide
residue analysis as well as discussion of the latest pesticide analyti-
cal methods, with emphasis on biological and environmental samples. The
greatest changes have been made in the following topics. Recognizing
the increasing importance of high performance liquid column chromatog-
graphy, considerable new material on quality control, principles, and
methodology of this procedure has been added to Chapters 3 and 5. The
material in Chapter 5 on GC detectors has been extensively revised to
reflect current useage patterns and the availability of new commercial
detector models. The new technique high performance thin layer chroma-
tography is introduced and the coverage of quantitation by thin layer
densitometry is greatly increased in Chapter 6. The latest analytical
procedures for pesticide multiresidues and residues of important specific
compounds such as Kepone, PCBs, PBBs, dioxins, and ETU are outlined in
Section 7. Material on derivatization reactions and GC-MS for confir-
mation of pesticide residues has been expanded in Chapter 8. Literature
references have been increased and updated in all chapters and most
sections.
The author and editors believe that the current revision of the Quality
Control Manual is an invaluable and indispensible source of information
on proper laboratory techniques and the best available analytical meth-
odology for pesticide analysis. Any comments or suggestions for improve-
ments and additions in future revisions will be gratefully received from
users of this Manual.
ACKNOWLEDGEMENTS
The editor wishes to express appreciation to the many laboratories
throughout the United States and Canada whose analytical data provided
the basis for a number of illustrative tables used in the text.
A special note of thanks is due the following:
Mr. Frank Wilinski, Electronics Shop supervisor at EPA, RTP, NC
for the contribution of the raw data required for the preparation
of Section 9 covering instrumental troubleshooting.
Dr. Robert Lewis, Dr. Wayne Sovocool, Dr. Lynn Wright and Mr.
Robert Harless, Analytical Chemistry Branch, EPA, RTP, NC for
their extensive revision of the mass spectrometry portion of
Section 8.
xi
-------
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., 10~9 to 10~12 grams 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 substrate 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.
IB EXTRACTION PROCEDURES
Environmental and biological samples generally cannot be analyzed
directly for pesticide residues because the level of the desired
-------
Section 1C
residue may be too low and the levels of interfering constituents
too high. It is usually necessary to extract the desired consti-
tuent(s) from the substrate and to purify and concentrate it (them)
prior to determination.
A solvent or mixture of solvents should be used for extraction that
is at least 80 percent efficient, selective enough to require a
minimum of cleanup, and does not interfere with the final determina-
tion. Optimum extraction conditions should be found by recovery
studies for each analysis in terms of solvent polarity, and in the
time and manner of contact between sample and solvent. 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 non-polar, fat-
soluble organochlorine pesticides and benzene, chloroform, dichloro-
methane, or acetonitrile for the more polar compounds such as organo-
phosphates 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. Extraction efficiency can
be checked most accurately if the laboratory is equipped to apply
or incorporate radioactive-labeled pesticides in the sample substrate.
1C CLEANUP PROCEDURES
The amount of extract purification (cleanup) required prior to the
final determination depends upon 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 chromatograph 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, liquid chromatography (column or TLC),
distillation, sweep co-distillation, selective photodegradation,
chemical reactions, and reaction gas chromatography (alkaline pre-column)
are most often used for cleanup individually or in various combinations.
-------
Section 1C
For adsorption chromatography, extracts are concentrated to a small
volume, applied to the top of a Florisil, charcoal, alumina, silica
gel, or mixed-adsorbent column, and the pesticides are eluted in
fractions by passage of one or more solvents while the co-extractives
remain on the column or are elute-d in different fractions. Elution
of a residue in a certain fraction (selective adsorption) 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.
Solvents used for partition cleanup of hexane extracts include
acetonitrile, dimethylformamide, and dimethylsulfoxide. 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.
Of the procedures listed above, solvent partition followed by adsorption
chromatography is most often applied. An automated instrument based
on gel permeation chromatography has been shown to efficiently separate
lipids from chlorinated pesticides and PCB's and to be advantageous in
terms of convenience and speed of processing large numbers of samples.
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 interest at the desired sensitivity level, with
recovery of preferably 80-85% or better, and with removal or separation
of adequate levels of background interferences.
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 or degradation by carrying out
recovery studies on a spiked control (uncontJ"ninated) sample at
different fortification levels.
-------
Section ID
ID FINAL DETERMINATION METHODS
Chromatographic methods are by far the most widely used for determina-
tion of pesticide residues, followed by spectrophotometric and
biological methods. The latter include bioassay and enzymatic tech-
niques which 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 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; how-
ever, removal of fluorescent impurities is often necessary, and this
can be difficult.
Paper chromatography provided the analyst in the late 1950's 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. Most pesticide analyses have been performed
on 0.25 mm layers of alumina or silica gel, but polyamide and cellulose
are also used. Organochlorine compounds are detected at 5-500 ng levels
by spraying with ethanolic AgNO^ or incorporation of AgNOg into the
layer followed by irradiation with ultraviolet light. Many organo-
phosphorus 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 chroma t og: 3\>':>:-, .-. <:? 1.1 rj-.-f n
-------
Section ID
GE-XE60, Versamid 900, and DC-200/QF-1, OV-17/QF-1, SE-30/OV-210
mixtures. Samples should be examined on two or three columns of
markedly different polarity before results are considered conclusive.
Glass columns are preferred because they minimize decomposition and
are easier to pack for optimum efficiency. After packing, 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.
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 endrin, are subject to degradation
in columns under certain conditions, and these conditions should be
avoided.
Organochlorine pesticides are usually analyzed with a tritium or
63fti electron capture detector with DC or pulsed applied voltages.
Though this detector is less specific than the other common pesticide
detectors, it can detect low picogram amounts of many halogenated
compounds. 6%i 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 several decades
of concentration, and the pulsed wide-range "%i detector has become
especially popular because it can be used with automatic injection
systems. The microcoulometric detector can be operated in a mode
specific for halogenated compounds with a sensitivity of about 10~9 grams
of halogen. Organophosphorus pesticides are detected selectively at low
ng levels by the alkali-metal thermionic detector and the flame photometric
detector in the phosphorus mode (526 nm). Sulfur-containing pesticides
may be selectively detected by the microcoulometric and flame photo-
metric (394 nm) detectors. Nitrogen-containing pesticides are detected
selectively at high pg to low ng levels by the thermionic detector or
the Coulson electrolytic conductivity detector in their nitrogen modes.
Several highly selective flameless detectors for determining nitrogen-
containing compounds, which have become commercially available, appear
to have greater specificity and sensitivity. The Coulson detector can
also be operated selectively for organochlorine compounds, and the Hall
modification of this detector is sensitive to 100-400 pg heptachlor.
Labile, polar carbamate pesticides or their hydrolysis products are
often derivatized with a halogen-containing reagent and the resulting
derivative sensitively detected with the electron capture detector.
-------
Section ID
Currently of less importance than the above are the other proposed
pesticide detectors, including the phosphorus mode of the micro-
coulometric detector, the sulfur and pyrolytic modes of the con-
ductivity detector, the halogen mode (negative response) of the
thermionic detector, the microwave plasma detector, the sulfur-
phosphorus emission detector, and the Beilstein flame detector
Selective detectors have the advantages of simplifying cleanup
procedures and aiding residue identification.
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 changed regularly. Standards should be injected before
and periodically during the analysis of a series of samples to check
the established calibration curve. 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.
Modern high performance liquid chromatography (LC) is being used
increasingly for the final, room temperature determination of polar,
involatile, or heat-labile pesticide residues without derivative
formation. Analytical columns are commonly 25 or 30 cm in length
and 2-5 mm in internal diameter; they are packed with small particles
of a totally porous or pellicular (porous layer) adsorbent (usually
silica or alumina), reversed phase partition medium (e.g., silica to
which a liquid phase has been chemically bonded), or ion exchange
resin. The mobile phase is pumped through the column at pressures up
to 6000 psig. Ultraviolet absorption, refractive index, polarography,
and fluorescence detectors have been used, the latter in determining
non-fluorescent pesticides by fluorogenic labelling utilizing methods
similar to those employed earlier for thin-layer fluorodensitometry.
The major disadvantage of LC at present is the poor sensitivity (ca.
10~6 to 10~9 grams) of commercially available detectors; however, it
is often possible to improve detectability by injection of large volumes
of extract without adverse effect on column performance or linearity
of detector response. Other advantages of LC compared to GC are that
only minimum sample cleanup is usually required, and a greater number
of separations of greater complexity can be accomplished; the mobile
phase plays an active role in achieving resolution, and there is a
wide range of stationary phases available for use in combination with
a great variety of solvent mixtures and gradient elutions.
Quantitation of residues by scanning of thin layer chromatograms with
commercial densitometers is widely applied for analysis of nonvolatile
or unstable pesticides or where GC or HPLC equipment is not available.
Precision, accuracy, and selectivity are often comparable to those
techniques, and sensitivity is in the high pg-to-ng range for many
analyses in which detection is made with fluorescence, chromogenic, or
-------
Section IE
enzyme-inhibition reagents. For best quantitative results, spots are
automatically applied to precoated, hard-surface plates, detection
reagents are uniformly applied by dipping rather than spraying, and
samples and standards are developed together on each plate.
Other final determinative methods that have been applied to pesticide
residues include polarography for compounds containing an oxidizable
or reducible group, either naturally or after derivatization, atomic
absorption, activation analysis, and radiochemical techniques. The
latter are most often used in metabolism studies, for example thin-
layer chromatography of pesticides containing a radioactive isotope
combined with autoradiography or radio-scanning of the layers.
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 on columns of different polarity,
preparation of derivatives to alter structure and volatility and
thereby chromatographic properties, extraction ^-values, ultraviolet
photolysis, selective GC detectors, and mass spectroscopy. Unlike
conventional NMR, IR, UV, etc., the latter method has sufficient
sensitivity (as low as 10 ng) for general application to residue
identification as well as for confirmation of pesticides in the
presence of PCB's. 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.
-------
Section IF
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
organophosphorus 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 PCB's 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 con-
taminants contained in solvents or reagents, or extracted from plastic
apparatus, can give rise to GC peaks or TLC spots which 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".
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
Automation of pesticide analyses is presently in its early stages.
Totally automated procedures have been developed for analyses not re-
quiring column adsorption cleanup and those in which the final determina-
tion is by colorimetry or UV absorption. Many of the colorimetric
methods involve enzyme inhibition. For samples requiring more
extensive cleanup, manual procedures must be carried out before the
automatic technique. Several means for automatic transfer of manually
prepared samples onto a gas or liquid chromatography column and a
system for automatic cleanup of samples by gel permeation chromatography
have been designed. Data systems allow storage of large amounts of
data with computerized printouts that increase the speed and efficiency
-------
Section 1G
of analyses. 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-automated procedures is still required to carry
out successfully multiresidue analyses of complex samples containing
an unknown variety of pesticides and interferences. A proven, com-
pletely automated procedure for multiresidue analysis as it is usually
performed (i.e., extraction, partition and adsorption chromatographic
cleanup, and gas or liquid chromatography) is not yet available.
******
Since this introductory section is intended as a broad overview of
modern residue analytical methods and their quality control, no
details have been given. Much of the foregoing material will be dis-
cussed more completely in later sections, and specific references to
relevant sections of the EPA Pesticide Analytical Manual or other
sources will be given. A general bibliography of recent books and
reviews on pesticide analysis follows.
1G REFERENCES
(1) AOAC General Referee Reports: Subcommittee E, J. Ass. Qffjc.
Anal. Chem., 60, 368 (1977) (yearly reports).
(2) Berck, B., Analysis of Fumigants and Fumigant Residues,
J. Chromatogr. Sci., 13, 256 (1975).
(3) Bontoyan, W. R., chairman of editorial board, Manual of Chemical
Methods for Pesticides and Devices, Volumes I and II, U.S.
Environmental Protection Agency, TSD-Chemical and Biological
Investigations, Chemistry Laboratory, Beltsville, MD, (updates
every six months).
(4) Bowman, M. C., Analysis of Insect Chemosterilants, J. Chromatogr.
Sci., J.3, 307 (1975).
(5) Burchfield, H. P., and Storrs, E. E., Analysis for Organophosphorus
Insecticides and Metabolites, J. Chromatogr. Sci., 13, 202 (1975).
(6) Cochrane, W. P., Confirmation of Insecticide and Herbicide Residues
by Chemical Derivatization, J. Chromatogr. Sci., 13, 246 (1975).
9
-------
Section 1G
(7) Cochrane, W. P., and Purkayastha, Analysis of Herbicides by Gas
Chromatography, lexicological and Environmental Chemistry
Reviews, J., 137-268 (1973).
(8) Borough, H. W., and Thorstenson, J. H., Analysis for Carbamate
Insecticides and Metabolites, J. Chromatogr. Sci., 13, 202 (1975).
(9) Egan, H., Methods of Analysis: An Analysis of Methods,
J. Ass. Offlc. Anal. Chem.. J50, 260 (1977).
(10) Fishbein, L., Chromatography of Triazines, Chromatographic
Reviews. _1£, 177 (1970).
(11) Fishbein, L., Chromatography of Organomercurial Fungicides,
Chromatographic Reviews, 13, 83 (1970).
(12) Fishbein, L., Chromatographic and Biological Aspects of PCB's,
Journal of Chromatography, 68, 345 (1972).
(13) Fishbein, L., Chromatography of Environmental Hazards III,
Pesticides, Elsevier, N. Y., 1975, 830 pp.
(14) Frehse, H., Problems and Aspects of Present Day Residue Analysis,
Pure Appl. Chem., 42, 17 (1975).
(15) Lively, J. P. editor, Analytical Methods Manual, Inland Waters
Directorate, Water Quality Branch, Ottawa, Canada, 1974,
(periodic revisions).
(16) Magallona, E. D., Gas Chromatographic Analysis of Carbamates,
Residue Reviews, 56, 1 (1975).
(17) Malone, B., Analytical Methods for Determination of Fumigants,
Residue Reviews, 38, 21 (1971).
(18) McLeod, H. A., Systems for Automated Multiple Pesticide Residue
Analysis, J. Chromatogr. Sci., JJS, 302 (1975).
(19) 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.
10
-------
Section 1G
(20) McMahon, B. M. , and Sawyer, L. D. , editors, Pesticide
Analytical Manual, Volumes T (multiresidues) and TI (individual
residues), USDHEW, FDA, 5600 Fisher's Lane, Rockville, MD
(yearly revisions) . Referred to throughout this manual as the
FDA PAM.
(21) Oiler, W. L., and Cranmer, M. F. , Analysis of Chlorinated
Insecticides and Congeners, J_._ _Chroma_to£r_._^ ci . , JL3, 296 (1975).
(22) Ruzicka, J. H. A., and Abbott, D. C. , Pesticide Residue
Analysis, Talanta, 20, 1261 (1973).
(23) Sherma, J., Chroma tographic Analysis of Fungicides, Chromato-
graphic Reviews , j_9, 97-137 (1975).
(24) Sherraa, J., Gas Chroraatographic Analysis of Some Environmental
Pollutants, in Advances in Chromatography, Volume 12, Marcel
Dekker, Inc., 19 7 5~~~ Chapter sTpP- ~142-1 76.
(25) Sherraa, J., Chromatographic Analysis of Pesticide Residues,
, 1973,
.
pp. 299-354. ~ ~~ ~"
(26) Sherma, J., and Zweig, G. , Pesticides, Chapter 25 of Chromatography ,
Heftman, E. , editor, Van Nostrand Reinhold Co., 3rd edition, 1975,
pp. 781-814.
(27) 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, 1973, 729 pp.
(28) Slade, P., IUPAC Commission on Development, Improvement, and
Standardization of Methods of Pesticide Residue Analysis,
J. Ass. Offic. Anal. Chem. , 59. , 894-930 (1976).
(29) Thompson, J. F. , editor, Analy_sis_ of Pesticide Res idues in Hutnan
and Environmental Samples, US EPA, Chemistry Branch, Research
Triangle Park, NC (yearly revisions) . Referred to throughout
this manual as the EPA PAM.
(30) Tindle, R. C. , Handbook of Procedures fo_r Pest icide Residue
Ana ly sis , Technical Papers 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.
11
-------
Section 1G
(31) Williams, I. H., Carbamate Insecticide Residues in Plant
Material: Determination by Gas Chromatography, Residue
Reviews. 38, 1 (1971).
(32) Yip, G., Analysis for Herbicides and Metabolites, J. Chromatpgr.
Sci., 13, 225 (1975).
(33) Zweig, G., and Sherma, J., Gas Chromatographic Analysis,
Volume VI of Analytical Methods for Pesticides and Plant Growth
Regulators, Zweig, G., editor, Academic Press, 1972, 765 pp.
(34) 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, 1976, 509 pp.
(35) Zweig, G., and Sherma, J., Spectrometric Analysis and Analytical
Methods for New Individual Pesticides, Volume IX of Analytical
Methods for Pesticides and Plant Growth Regulators, Zweig, G.,
editor, Academic Press, 1977, 297 pp.
12
-------
Section 2
INTERLABORATORY QUALITY OWROL
2A QUALITY CONTROL PROGRAM OF THE EPA ENVIRONMENTAL TOXICOLOGY DIVISION
(ETD) LABORATORY
Quality control in the context of this Manual connotes procedures taken
to assure the accuracy and precision of analytical results. Qualitative
and quantitative determinations by residue analysts are utilized for such
important tasks as surveillance or monitoring of pesticide levels in some
segment of the environment or the food supply, and, if conclusions and
subsequent actions are to be valid, it is vital that the analytical data
be reliable. The complex nature and pitfalls of the analytical proce-
dures as outlined in Section 1 require a set of built-in controls to
prevent or detect incorrect results. This Manual is dedicated to a
program of quality control which will significantly minimize the output
of unreliable and invalid analytical data.
The Quality Assurance Section of the Chemistry Branch, EPA-ETD Labora-
tory 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, Florida,
before the Laboratory was moved to North Carolina. Originally, the pro-
gram was limited to Community Pesticide Studies, National Monitoring,
and State Services Laboratories operating under contract with the U.S.
Department of Health, Education, and Welfare, and more recently with the
EPA to conduct chemical monitoring for pesticide residues in man and his
environment. Parts such as the interlaboratory check sample program
have now been expanded to include other state and private laboratories
cooperating with the EPA.
The quality control program can be broadly divided into two classifi-
cations, both of which will be discussed in detail in this and the
following Sections. The Interlaboratory control program, which was the
first one formalized, involves analysis of uniform samples* by a number
of participating laboratories in order to assess the continuing capa-
bility and relative performance of each. In addition, this program
* The terms "check sample" and "blind sample" are used for the samples
prepared 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
procedure.
13
-------
Sections 2B, 2C
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 dis-
tributes the final results. Details of these procedures and typical
sample data are given in Subsections 2.D. through 2.H.
The Intralabo r at o ry_ 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 guide-
lines 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 per-
formance in terms of both precision and accuracy arid includes upper and
lower control limits to serve as criteria for remedial action or for
judging the significance of variations 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 statisti-
cal evaluation 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. The analysis of interlaboratory check samples by all participants.
b. Operation of a repository to provide to any non-profit laboratory
analytical grade pesticide reference standards, over 400 of which
are now available. These are listed in an index available from
the ETD Laboratory.
14
-------
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 mate-
rials 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-site 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 since 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
(ethylene glycol trapping solvent spiked at concentration levels found
in actual 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 four
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. Part
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 for later spiking.
Sub-samples are mailed to participating laboratories to serve as their
interlaboratory check sample and to provide sufficient intralaboratory
standard reference material (SPRM) for six months. Each laboratory
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 no need to mail the samples frozen
15
-------
Section 2E
since 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
remainder into small vials which are stored continuously in a freezer.
Individual vials are removed as needed to provide 2.0 ml intralaboratory
SPRM samples.
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 re-
coveries 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
interlaboratory 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 samples for testing certain
procedures will be supplied in the future.
With the check sample, each participant receives a covering letter pro-
viding 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.
Since the interlaboratory check sample will be recognized as such by the
chemist at the time of analysis, he is likely to give special care and
attention to it; also, 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 considered a very 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
pretested Florisil cleanup adsorbent and GC column packings and frequently
updated standard analytical methods.
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
detailed 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, in-
jection volumes, elution cuts if column cleanup is required, all instru-
mental operating parameters, identity of the GC column, and the numerical
16
-------
Of IKVERLABORATORY CHECK SAMPLE
Electron Capture Detection
LAD. CODE NO.
TRANSFER LINE TEMP.
PREVALENT STAND. CURRENT
COLUMN COL. TEMP. C. COL. LENGTH FT. DETECT. TEMP.
C. CARRIER.FLOW ml/mln. PURGE FLOW ml/min. CHART.SPEED in/min ORIG. SAMPLE SIZE
*f.s.d. at attenuat. of x ELECTROMETER SS E 2 (Circle)
°C. INLET TEMP °C
PESTICIDE COMPOUND
VOL. OF
FINAL EXTR.
(ml)
INJECT
VOLUME
(ul)
STANDARD INJECTION
PICO-
GRAMS
PEAK
HT. OR
AREA
1. Source of method used (Cite journal, volume, pa
2. Briefly state modifications, if any.
3. Method and temperature of extract concentratior
Kuderna-Danish Simple Boiling Roto-Vac
ATTENU-
ATION
SAMPLE INJECT.
PEAK
HT. OR
AREA
I
ATTENU-
ATION
RETENTION REL.
TO ALDRIN
STD.
SAMPLE
i
I
(ppb)
RESIDUE
ige , year or other source)
i (circle) Air Slowdown Nitrogen Slowdown
Other Temp. °C,
CHROMATOGRAM
NUMBERS
O
D
NJ
PI
Refer to back of sheet for instructions for
completion of report.
Do you wish your chromatograma returned?
YES NO'
(Circle)
Signature of Chemist Doing Work
-------
TABLE 2-1 (Reverse side data) Section 2E
INSTRUCTIONS FOR COMPLETION OF REPORT
1. This form is to be used only for determinations by electron capture detection
Results obtained by flame photometric detection to be reported on separate
form so designated.
2. 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.
3. All data requested on the report form are meaningful for a complete
evaluation of the analytical performance. Therefore, supply complete
information under all columns on the form.
4. In case any compound appears in more than one cleanup elution fraction,
report the quantitative value for each fraction separately. We will
sum to obtain your final total value.
5. Under the column headed VOL OF FINAL EXTR., ml, the numerical value
to be placed here should be the final volume and not necessarily the
volume achieved by concentration. For example, in handling a blood
sample, an extract concentration down to 1 ml might prove necessary for
the quantitation of dieldrin. In this case, the figure 1 would appear
in the column opposite dieldrin. However, for the determination of p,p'-DDE,
a dilution of the concentrated extract up to 10 ml may be indicated. In
this case, the figure 10 would appear opposite p,p'-DDE.
6. 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 chromato-
gram, the other representing the sample.
7. Include with the chromatograms a standing current profile for each E.G.
detector used. Label each step with the polarizing voltage for that
step. Record information (as to (1) column used during determination,
(2) flow parameters - purge and carrier, (3) temperature of column,
(4) attenuation used.
8. Mail only original chromatograms with your report. Fold chromatograms
for each column in accordion fashion from one continuous roll. Do
not mail photocopies. Chromatograms will be returned if you so indicate
in the space provided on the form.
9. With your report, include all chromatograms related to the sample work
whether used for quantitation or confirmation. However, for those
columns used for confirmation only, include all data on the report form
except a final quantitative value.
18
-------
Sections 2F, 2G
data and original chromatograms upon which all calculations are based.
The chromatograms must be clearly identified so that they may be re-
lated to the data on the reporting forms for easy checking of the
quantitative results by the coordinating laboratory.
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 sub-
section. 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 recording all results, a statistical analysis of the
results is made and recorded on the Summary of Results sheet. A rela-
tive 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 chromatograms for study is vital because it allows the
coordinating laboratory to check such factors as column efficiency,
sensitivity of detection, 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 detailed 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.
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
laboratories 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).
!G SUMMARY OF RESULTS TABLES
Typical summary tables are illustrated as Tables 2-2 through 2-5. Defi-
nitions and means of calculating the items included in the statistical
report at the bottom of each table are given in Subsection 2K.
19
-------
Section 2H
TABLE 2-2
INTERLABORATORY CHECK SAMPLE NO. 26, MIXTURE OF STANDARDS IN SOLVENT
SUMMARY OF RESULTS .
LAB CODE
NUMBER
FORMULA-
TION — >
<*5.
V?.
48.
52.
53.
54.
66.
68.
69.
71.
72.
73-
83.
84.
85.
87.
88.
89.
90.
92.
93.
95.
96.
97.
U3.
113A.
130.
135.
137.
160.
161.
162.
163.
164.
Mean
Std.Dev.
Rel.Std.
Dev. ,%
Total '
Error,?8
PESTICIDES REPORTED IN PICOGRAMS PER MICROLITER
Dindane
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
Tept.
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.P'-
DDE
75
379*
88
73
29*
47
66
76
_.-
...
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
12.5
17. <•
36
Dieldrin
20
145*
_—
31 '
16
18
14
?1.6
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
_._
—
60*
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
8.54
32.6
116
P.P1-
DDT
100
578*
195*
113
91
22*
115
99
---
150*
98
94
88
94
90
99
94
120
73
115
125
87
150*
110
94
96
112
104
103
97
95
101
96
56*
100
11. <
11. <
23
Per-
thane
0
379
34
Ethion
0
68
o,p'-
DDD
0
20
P-P'-
DDD
0
4.0
16
13.2
12
Mi-
rex
0
29
DCPA
0
9.C
*Rejc
outl
B-BHC
0
2.5
cted an
i.era
20
-------
Section 2H
I-H-
o
co
CO
p
§
EH
CO
o
w
Cri CO
X
s
VO
cvl
O
S
W
CO
w
§
o
>H
PS
CO
o
^
Pi
M
w
M
i
M
O
o
M
«
W
CO
*3*
K
o
o
o
M
PH
^
M
1
0
/*.
MH
w
PH
CO
W
o
M
O
M
Cj
t^
CO
W
JM
PS
O
C_l
tr1
<
PS
O
§
1-3
1
"ftQ
-Q
ftQ
l
"ft EH
-Q
ftQ
"ft! EH
r""*\
* \—4
0|Q
C
r,
>Hj
TJ
W
c
•rH
^
T3
H
ft
K W
c
•H,
f-l
•a
rH
0)
•a
c
•H
.
O
w
Q
O
0
o
0
o
rH
o
CM
u~\
•rH
0
C\!
\p^
fx.
0
rH
O
O
TH
T
c
o
•H
•P
tH
3
0
fr.
CM
*-4
* *
-C^ O^ C^ O^ O^> C\l ^D ON ^3 *~^ ^3 O^ *^ f^
ON ON cO ON ON ^5 ^D ON ^5 *~^ C3 ON ON CO
i— 1 T-* tH tH rH
O-O- -3-CMVQOOOO -^- V>
CNJCMCVJC^OJCXjCXlCVJCxJCNiCVJCNaCVjtH
*
CM C\i ON O OO CN! ^ D*- CM
^ r^a^xAO-d- xAvnc^-co r\ -^ ^ c^
^-f^^^t^vHTHtHTHiHTHrHTHrH^H
r*- ON \O r1^ cvj c^- c'A CM NO
• • ••• ••* •
ON CO GO CM ^f^\ C3 t — 1 ^J"N, ON C5 CO C5 ON C^
r-l tH rH CM T-l CM CM tH tH CM rH CM tH tH
xr\
* *
,Hxr\xncM tH o-^^ncM^)- co^j* OJNQ
C^- C>- C"- O CO CN- C^- C^- C^- Cs- CO CO CO V\
tH
^
C^iCN- ^-jj-CvlONON>ACOf^vOCMO
OvONOOOOONOOONONOOONl>-
rH rH rH TH rH TH
:|c
C^VON Cntv-ONIN-ONOirvrHC^ONCM
O ON O TH ON ON ON ON ON ON O O OO CN-
TH TH TH TH TH
*
ONOONOOOOONONOOOOONVO
TH TH rH TH rH rH TH
-P d
Q) tlO
^^ O O
TH| 0) f-t
TH rH TH rH TH ,D TJ
nJ CD
•f-l
*N. O
^' -^
0) CO
-P ^
o a>
(B -H
••"» H
- C**\
J- TH CM CN-
TH TH CM
VO
ON C^V
« •
ON »H o vr»
iH iH CM
C^-
t*"\ cv
• •
(N^ _H^- \r\ H^
O- TH
CM
<> vn C^
ON o vr\ TH
TH
o
ON VQ TH
ON O NO C*"%
TH
ON vr\ vr\
ON o vr\ CM
TH
S^v.
fH CD
•P O h
a] o
•H . b
CD -H W
Q W
H
£ - . at
Ctf T3 rH. 4^
CU -P OJ O
S W K &
21
-------
Section 2H
CM
g
w
CM
O
o
i_q
M
pq
M
m
03
S5
H
P
«
O
W
to
B
p
M
O
M
W
s
PH
K
EH
O
§
pq
o
s
m
i
Oi
en
p
p
3
p
p
"p)
p"i
en
p
p
1
"pj
o|
fi
•H
H
(U
•H
p
0
g
0
o
o
o
o
oo
VTA
0
NO
o
.*
1
o
•H
rt
H
o
fe
rH
V
ON O PA
• • •
TH TH TH
(x. QQ NO PA OO CM PA PA PA C*- 00 NO C^- OO
00 ONO-d-ONONONONOCOCO ONt>-OOONONOO O
TH rH «H TH
^•jd-^i-NO PAVJA^A^I-J- PA^J- PA\T\jd- PAJ- PAPA
OTHOONOWAVA-d-NOOOOONQ CNlONUAvH^}-
\r\vr\vr\o-NONONO XOX/A^- WAVT\^J- UA^}- VAUAOO
* * TH-VVArHNOON-d-
CNI vo ^J- NO PA ^J1 PA PA PA ^ ^1" 0~A PA ^ ^i" ^ ^" ^
THCNIPA^VANOO.COONOTHOJPA^VANO^CO^
•d w
0) b
-P Ql
O -H
0) H
• 1—3 +5
-
PA
CM
TH VfA
TH xr\ CM .4-
^}- TH CM
TH
NO TH O-
VA rH ON ^J-
TH ^ij"
TH ^}- CM
^ 0 TH VA
TH CM
o > •&.
'•P P f-)
a o
•H • 'H
> T) h
0) -P W
C • • co
rt -d H -P
0) -P 0) 0
S W « EH
22
-------
Section 2H
CM
w
p
w
w
M
CM
O CO
25 M
2
O
M
1^]
M
s
p-»
W
PH
W
EH
«
(^
^
M
Q
g
W
fe
«
r/)
W
O
M
O
M
OT
w
PW
>H
«
O
LA30RA1
o
W
?
ox
1
Pi
8
Q
I
H
Pt
§
O
1
w
f-H
PJ
EH
g
m
0|
a
•H
f-i
•d
a>
•H
O
CO
•o
• •H
PU 0
CD p
xw
c
•H
f-.
r-H
<
•
2
w
o
o
o
o
o
p-\
o
o
0
o
o
CO
o
o
t>-
o
0
(^
o
o
T-l
o
ition — )
Formulc
o
£)O-
(N-VO XTjO^XTvCO O .3- .3- .3- .=}• .3- -3- (M\Ac>,^t O-N^}- <^vtM r^c^i
*
vo rr>c\ij--d-»j->cv-o oco c^r^vcM c"-\ o O-CM
1>-XO OVO O-O-MD (N-COVD O-^3- !N-vr\xa\O'u^
rH
0 * *
iAcoot>-vr\,-i.4-xr\oJOooNONrr'\rn\oooj
^* W%MDVO O-C^-vp,ovPiCO CVcnvOVO C^t-< r-l
VH TH
* *
o o
cr^!^--d-^OO^-^^O^-oJO«^|^Jr-^c^j^x^vf^c^
c^cvcncMr^r^oojrAc^r^or^cvioJCMrH
0 OOOO OOOOOOO
1 Oscn 1 OO\I>-VO 1 OCO^OONOOXT\M3VO
lOiH I OOOO IT-IOOOOOOO
T-l^-VAvOC^-OOONT-I^J-VPi^O^J-xAMD-r-IO^jj-
•r-lT-lYH.c-ICMCMCVJC'^Cnr'^
O CO
01 rt , w
•rs ' f,
0) ^3 -F
iH <^>
0
o co
-=2- 0
0 0 0 0
CM -*
CO ^5
t-< ^t
0 J*
vO ^
O O C"» -*
CM vr»
CM
oo ^j-
CM O
O O V\ >A
TH <^\
O CM
00 CM
o o
O O CO VA
CM ^O
V.
0 ^ ^.
^ Q fc
0) P 0
•? - ^
g S w
° W H
e . oj
rf T( J -F
I s « a
23
-------
Section 2H
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
chlorinated 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 2Ke for de-
scription of fitness test). The precision (relative standard deviation)
was considered "good" for this type sample only for the compounds lindane,
heptachlor epoxide, and £,_p_'-DDT, "fair" for £,£*-DDE, and "poor" for the
other four. The overall average RSD for all compounds was 21.6 percent,
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
£,£'-DDT, aldrin is "fair", and the others "poor". The average Total
Error was 52 percent, just outside the "acceptable" level of <50 percent.
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 percent and the average Total
Error 20 percent, 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 con-
tinuing 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 Subsection 5G of this Manual.
Table 2-4 shows typical results for a more difficult blood serum check
sample reported by 18 laboratories with experience in the quality control
program. The average RSD of 13 percent and Total Error of 33 percent
are quite acceptable for this type sample. Table 2-5 shows results from
17 laboratories for a fat check sample. Average RSD and Total Error values
are obviously not so good in this case.
2H RELATIVE PERFORMANCE RANKING
A scheme has been developed and tested for the relative ranking of labo-
ratory results in the analysis of multiresidue check samples. There are
three essential criteria for a high score in this performance ranking,
namely, correct identification of all pesticides present, correct quantita-
tive 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.
24
-------
Section 2H
The maximum possible score is 200 points, 100 for correct identification
and 100 for quantitation. A detailed explanation of the calculation
procedure follows:
a. Identification
The 100 possible total points divided by the number of compounds actually
present yields the point value per compound. Correct identification 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 com-
pound 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 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.
b. Quantitation
The point value per compound is again the total possible points (100)
divided by the number of compounds present. Should all reported values
coincide exactly with the formulation values (an unlikely situation), the
full 100 points are awarded. When a reported value differs from the
formulation value, the difference between the two (the absolute error)
divided by the standard deviation (previously calculated for each com-
pound) gives a "weighted deviation". This value is subtracted from the
point value of the compound to give the quantitative score for that com-
pound :
Compound Quantitative _ Compound Point Absolute Error
Score ~ Value ~ Standard Deviation
The total score for this part is the sum of the individual compound
quantitative 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 other-
wise 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.
c. Total Score and Sample Results
The total score for laboratory performance is the sum of the identifi-
cation and quantitation point totals. Table 2-6 illustrates in detail
the method of calculation for a hypothetical analysis in which one com-
pound is missed and one extra is reported, resulting in an inferior
total score of 125.
25
-------
Section 2H
Table 2-6
CALCULATION OF TOTAL SCORE FOR RELATIVE PERFORMANCE RANKING
3-BHC
p_,p_'-DDE
Dieldrin
o,p_'-DDT
p_,p_'-DDT
a-BHC
Formulation Reported Analytical
pg/yl Values pg/yl
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
*Of all data from participating laboratories
Point value for
Identification
g-BHC
p_,p_'-DDE
Dieldrin
o ,£-DDT
£,£'-DDT
each compound is 100 -5- 5 = 20
20
20
20
0
2_2.
sum = 80
-20 Penalty for false identification of
-------
Sections 21, 2J
Tables 2-7 through 2-9 show Relative Performance Rankings for groups
of laboratories on check samples of different types. Table 2-7 shows
rankings for the laboratories reporting the data in Table 2-2. Labora-
tories 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 which should be quickly resolved.
Those with scores below 150 should suspend all routine pesticide analy-
tical work pending the resolution of very serious problems in both identi-
fication and quantitation, the effects of which place in doubt all rou-
tine analytical data output of the laboratories. The laboratories are
to initiate corrective action immediately based on the general and indi-
vidual 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 labora-
tory before deciding upon what, if any, action should be taken based on
the results. For example, Table 2-8 shows ranking data for 17 labora-
tories analyzing a fat sample (results reported in Table 2-5) and Table
2-9 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. Re-
ference to Table 2-5 shows that those below the break point had readily
apparent problems and these four laboratories 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.
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 cri-
tique depends upon the problems that are obvious from a careful analysis
of the submitted results and might include comments on incorrect stan-
dards, instrumental factors, calculation errors, poor choice of materials
or parameters, etc.
2J PROGRESSION OF PERFORMANCE
During 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.
27
-------
Section 2J
TABLE 2-7
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
0
1
0
0
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
I/ Values outside confidence limits
2/ Total possible score 200 points
28
-------
Section 2J
TABLE 2-8
RELATIVE PERFORMANCE RANKINGS - CHECK SAMPLE NO. 21, FAT
Lab. C<
Number
15.
16.
8.
25.
7.
4.
26.
33.
5.
34.
11.
9.
31.
6.
1.
14.
24.
I/
i/
Dde Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
Values outside confic
Total possible score
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
Jence limits
200
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
29
-------
Section 2J
TABLE 2-9
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
0
0
0
0
0
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
Score 2/
198
198
197
197
197
197
196
196
196
196
196
195
194
193
192
192
I/ Values outside confidence limits
2J Total possible score 200
30
-------
Section 2K
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 2-10. The method used in the first year was based on
gas chromatography of a concentrated 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 network. 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
acetonitrile/petroleum ether partition and Florisil column chromatography
(2), resulting in significant improvement in not only precision (Sample 9,
Table 2-10) but in accuracy as well. After several months' experience
with the method, results on another check sample (Sample 11, Table 2-10)
were even better, and with continued participation in the program, the
laboratories are making still further progress in their performance as
the latest (1974) figures show.
The results in Table 2-11 show progression of laboratory performance
on interlaboratory blood check samples between 1968 and 1974. In the
beginning the direct hexane extraction method of Dale et 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 (4) extraction method, which
utilizes a constant speed mixer (Subsection 7A). The results of the
blood check samples illustrate again the dual value of the Interlabora-
tory 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
31
-------
Section 2K
TABLE 2-10
PROGRESSION OF RESULTS
FAT CHECK SAMPLE
Interlaboratory
Sample Number
3
9
11
21*
24
28
Year
1967
1968
1969
1972
1973
1974
No. of
Labs
15
21
19
16
14
10
No. of
Compounds
7
7
7
7
7
7
Average
Recoveries,
**
**
108
89
95
96
Average
RSD,
50
38
24
19
14
12
*Complete data given in Table 2-5
**Unspiked samples for which precise pesticide levels were unknown
32
-------
Section 2K
TABLE 2-11
PROGRESSION OP RESULTS
BLOOD CHECK SAMPLE
Interlaboratory
Sample Number
6
10
16
17
22
23
25
27
Year
1968
1969
1970
1971
1972
1972
1973
1974
No. of
Labs
22
20
22
20
17
17
18
15
No. of
Compounds
6
5
4
4
4
4
4
3
Average Average
Recoveries, RSD,
* 36
* 29
* 21
* 17
96 14
91 12
100 13
* 16
* Unspiked samples for which the actual pesticide levels were not known
33
-------
Section 2K
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 uncertainties in measurements. An
example is the measurement of the height of a gas chromatographic peak
with a ruler, which even the most careful analyst can measure only to
the nearest one mm. 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.
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 + 15-20 percent or better are obtained on recovery samples
fortified at ppm levels, + 30 percent at ppb levels. One authority has
suggested that a coefficient of variation of less than 40 percent is
acceptable for precision between laboratories for a trace analytical
method.
Absolute error is the difference between the experimental result and
the true value. Relative error is absolute error divided by the true
value and multiplied x 100 to yield percent relative error or x 1000 to
yield parts per thousand relative error. As an example, an absolute
0.2 ul error in injection of a sample for GC corresponds to ? A • "
20% for a 1.0 ul sample but only °'2 x 10° = 4% for a 5.0 ul sample. It
is explained later in Section 3 that 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
recorded, and unless qualifying information is given this last digit is
assumed uncertain by +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
°-4-i- x 100 = 1%. Likewise, one should always be sure to record all certain
figures plus one uncertain figure in a measurement, these figures being
designated 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 pesti-
cide 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,
-------
Section 2K
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 num-
ber (in this case 1) and is, therefore, significant. All terminal zeroes
following a decimal point are significant. For example, 9.800 g indi-
cates a weight of 9.8 grams accurate to the nearest 1 mg, four signifi-
cant figures. The number 10,100 should indicate five significant figures,
but terminal zeroes in a whole number must be considered 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
uncertainty is +1 mm and the relative uncertainty
Tn-rnn— x 100 = 0.01%. If the measurement was actually made only to
XU f _LUw
the nearest 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 104 mm, to indicate an absolute uncertainty of +1 x 102 mm
and a relative uncertainty of 1 X 1Q2— x 100 = 1%.
10,100
Significant figures should be properly retained when performing mathe-
matical operations. Simplified rules which 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:
8Q1
80
Oc.
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 round-
ing 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 signi-
ficant figures. For example, in calculating the ng of pesticide repre-
sented by an unknown GC peak by comparison with the area of a standard
peak, the formula ngu - ngs |re|u is used if response is linear over
the range in question. If 1.0 ngsstandard 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% H.:^.,, - 1.3 ng with only two signi-
ficant 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.
35
-------
c. Average
Settion 2K
The average or mean (X) of a set of n values is calculated by summing
the individual values and dividing by n:
_CXi
n
d. Range
The difference between the highest and lowest values in a group.
e. Standard Deviation and Variation
Standard deviation (s) of a sample of n results is calculated by use
of the equation:
1/2
s =
n
n-1
Variance is equal to s . 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. Cf 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 u is the true mean, X is the experimental mean, and t is
a value obtainable in tables for different percentages of confidence
and numbers of trials (n). Values of r increase ae percentage
confidence desired increases and decreases for more replicates.
36
-------
Section 2K
The EPA Quality Assurance personnel have applied the following test for
determining "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 mathematical expression
in the next instruction from the following table.
Number of data points (n) Factor
in the data set
5 1.65
6 1.73
7 1.81
8 1.86
9 1.91
10 1.96
12 2.04
14 2.10
16 2.15
18 2.20
20 2.24
25 2.33
30 2.39
40 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 times the
standard deviation of the entire data set, the number is rejected as
lying outside a reasonable data set. The percent confidence interval
for the retained values would be given by:
(1 - v-)l°0 = %
2n
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 described by Hugh D. Young (5). 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.
37
-------
Section 2K
g. Total Error
Total error is a method proposed by McFarren et al. (6) for combining
precision and accuracy in one reporting expression:
Absolute Value of
m ^ i „ the Mean Error „ -\ no
Total Error = x 1UU
True Value
where s = standard deviation. In general, Total Error values < 25
percent are considered excellent, 5? 50 percent acceptable, and >50
percent unacceptable.
Specifically, in the interlaboratory control program, Total Error is
calculated from the following equation:
x + 2s
Total Error = x 100
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 (7), indicating it may unnecessarily downgrade
a considerable portion of results. Alternate equations are recommended
which rigorously meet McFarren £t _al.'s 25 or 50 percent criterion with
at least 95 percent confidence. These equations are:
x + 1.7 s
x 100
y
to be used when x/s >0.3 and up to 44 results are available
x + 1.8 s
x 100
when x/s « 0.3 - 0.15 and number of results are 45-170
T - -— x 100
when x is not significantly different from zero with 95 percent
confidence.
38
-------
Section 2K
h. Numerical Conversions
1 g
1 mg
1 Pg
1 ng
1 Pg
1 ml
1 ul
•
•
=
•
n
-
SB
1000 mg
1000 ug
1000 ng
1000 pg
]_Q— 1^ g
1000 Ml
10-6 1
39
-------
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. Ass; Offic.
Anal. Chem.. 4£, 186 (1963).
(3) Dale, W. E., Curley, A., and Cueto, C., Life Sciences, _5, 47
(1966).
(4) EPA Pesticide Analytical Manual, Section 5, A, (3), (a).
(5) Young, H. D., Statistical Treatment of Experimental Data, Chapter
10, McGraw Hill, 1962.
(6) McFarren, E. F., Lishka, R. J., and Parker, J. H., Anal. Chem.,
42, 358 (1970).
(7) 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 in the
Food and Drug Administration's Pesticide Residue Analytical
Laboratories, Environ. Qual. Saf., Suppl., J3 (pesticides),
25-31 (1975).
(2) Egan, H., Methods of Analysis: An Analysis of Methods,
J. Ass. Offic. Anal. Chem., £0, 260-267 (1977).
(3) Eiduson, H. P., Applications of Tolerances, Standards, and Methods
in the Enforcement of the Food, Drug, and Cosmetic Act, J. Chem.
Inf. Comput. Sci., JJ (2), 102-105 (1977).
(4) Eiduson, H. P., Laboratory Quality Assurance, Bulletin of the
Association of Food and Drug Officials, pp. 151-156, 1976.
(5) Youden, W. J., and Steiner, E. H., Statistical Manual of the
Association of Official Analytical Chemists - Statistical Techniques
for Collaborative Tests, published by the AOAC, PO Box 540,
Benjamin Franklin Station, Washington, DC, 1975, 88 pp.
40
-------
Section 3
INTRAU\BORATORY 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 paral-
lel 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.
41
-------
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 exami-
nation by the chemist's supervisors.
3C NATURE OP SPRM'S
One possible approach is for a laboratory to prepare its own SPRM'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 care-
fully on a hot water bath to a temperature not .above 50°C with inter-
mittent 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 which are also stored in the
freezer. The weight of each unit is slightly larger than the intended
sample weight. These serve as unspiked SPRM's.
Sufficient analyses are made on the unspiked subsamples so as to be
satisfied 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, transferred to small bottles, and stored in a freezer. These
spiked samples serve to test the capability of the analyst for recovery
of higher pesticide levels.
For both the unspiked and spiked SPRM's, at least a dozen replications
of the analysis on the same sample should be conducted by chemists with
recognized competence. From this 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.
42
-------
Section 3D
The EPA-ETD Interlaboratory Program provides participating laboratories
with a sufficient supply of each interlaboratory check sample to serve
also as an intralaboratory SPRM for a six month period (Subsection 2D).
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 2G) 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 inter-
nal SPRM's with reliable results available to them without having to
prepare their own samples and establish residue levels and RSD 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) , III, 3]. Likewise, routine blood samples are
received as whole blood rather than as the serum form of the check
sample. It would be undoubtedly advantageous to devise SPRM's which
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 consistent analysis regardless of the portion taken. It might well
be feasible for some other sample substrate, such as urine or water.
3D FREQUENCY OF SPRM 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 corresponding
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
43
-------
Section 3E
a team, e.g., with one chemist preparing extracts and another doing the
determination, SPRM samples should be handled in this same normal
fashion.
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 is 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 laboratories and furnishes to each
statistical summaries for comparison of results.
44
-------
Section 3E
TABLE 3-1
RECORD OF ANALYSIS OF STANDARD REFERENCE MATERIAL
Laboratory
Analyst or Team
Sample
No.
Date
Aldrin
3-EHC
Hept.
Epox.
Media
Diel-
drin
o,p'-DDT
p,p'-DDD
p,p'-DDE
p,p'-DDT
Analyst or Team
Reporting units should be in ppb or ppm. Observe standing instructions for
minimum, reporting levels.
45
-------
Section 3E
TABLE 3-2
SUBJECT: Quarterly Report, Quarter Ending
TO: Chief, Quality Assurance Section, Analytical Chem. Branch, Health Effects
Research Laboratory, (MD-69), Research Triangle Park, NC 27711
FROM: Chief Chemist (Laboratory)
During the past quarter we have analyzed the following numbers of
routine* samples for pesticide residues:
Blood {multiresidue) L
Blood (PCP)
Blood (Other) (specify)
Adipose Tissues T
Other' Human .Tissues
Air
Soils
Stream Sediment _________
Water (multiresidue) __________
Water (Other) (specify)
Urine (alkyl phosphate)
Urine (Other) (specify)
Housedust
Pish or Shellfish
Wildlife
Other**
Are any spiked SPRM's prepared in-house? Yes No If Yes, 'list the
substrates on the reverse side of this sheet giving the spiking level
range of each compound spike.
Chief Chemist
*The term "routine" is intended to mean samples of local origin such as
donors, autopsies, etc.
**Specify substrates if 10 or more samples were analyzed during quarter.
46
-------
Section 3F
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 con-
struct 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 (coefficient of variation) (Subsection 2Ke) for
the particular analysis. Stated another way, this is a numerical
expression of the potential precision of the method in the hands of
competent analysts. Ideally, this value may be obtained by conducting
a collaborative study involving at least 8 laboratories. Since the
average small laboratory will not be able to carry out this rather
involved operation, a satisfactory alternative is for two or more
competent chemists of the staff to analyze six replications of the
same sample. From these data a reasonable value for percentage RSD may
be obtained as shown in 2Ke.
The preparation of the chart is illustrated by the following Figures
3-A and 3-B in which results for a serum intralaboratory SPRM analyzed
over a period of three months for £,j>'-DDE and p_,£?-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 10 percent of the spiking level for each
pesticide.
47
-------
Section 3F
Figure 3-A. Laboratory A control curves for blood SRM, three-nonth period*
1 •'•..i •'•!,• '• ;• TV .' ;--P '1
2 +
2-
'
_U_L
& | lfi» -t
p,p'-DDE
Ijiij.v''.: i =i|f L1 .' {• '
2+
ildH^H-H-H-r-!---!-=-i
p,p'—DDT
fill'll L-W 1 1. i >!' " :—J A-i^ : « ~ i- r 1 f ! 1 1
2+i
a-
l^uill^-^iilniijiti-i Ji'. |- rjpj-| Hi-.'!--;-;--I , j-jv
I'iV'1. :!!h'!l"|M'. L.|JuL^._: .:,.!.'' ' .„.! _|.4.--._L. !..!_;•... _;_.J_j_»L:l-:4. •_ -'.'.la
!niirl!;i:-.''! iiil'r'i :.;[• TV ! ~Hr' ;1:f 'iltT' f"-i :T"! •" .TTTTt "•".I'd" Tvid^r.'"! "•. "' '1
__ Figure 3-B. Laboratory B control curves for blood SRM, three-month period.
7_°"lr'll!" -A1 !' [! ' ' » 1
* ::,-,:i.,i'T, ^V , , , , ,X,
*r-.\~~', p,p'-DDT.
MBhfy-.il-.,.^; M '•••i:i!^-^TIL ;.-! :T>!.l-T>^ f !'i
B.JH:.[iv'"•! M -t " t* '-; ' •! I' i • '• . :^^ . .. t~i:
48
-------
Section 3F
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 SPKM 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 SPKM: 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 -25
or -TT = -1.7 SEU
one SEU •"
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 cas'es 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.
49
-------
Section 3G
_ Figur* 3-C. Laboratory C control curv«» (or blood SUM, thr««-nonth period.
EP AW Tn I * i li I »i ! - . i III. '. » ' i~ I "i I .1
'b_j_ Hi. .,..4 : _ , I , ;
2-6 ; ! -f ; ,—n ;rf-i —
tfc ., ,1— , -i ',~kr*-\-—-i~-;Tir-~~t -j "^n^.' ..~j •"!"
• i ^ :. „ .j
~.~:I ..! .I--it «-
3G
From time to time, the following question is asked: "What is to pre-
vent 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 adminis-
trator, however, should have little difficulty detecting the doctoring
of curves. When a chart is submitted that is virtually a straight line
such as that for £,£'-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 SPRM 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 SPRM is analyzed at the recommended minimum rate of
one SPRM per 10 samples, this would amount to at least 10 SPRM analyses
per month, a number sufficient to justify preparation of the chart.
BENEFITS OF THE IN-HOUSE SPRM PROGRAM
Analyzing in-house SPRM'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 analyti-
cal output in those laboratories involved. For example, chemists from
50
-------
Section 3H
regulatory 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 obvi-
ously 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 assurance 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 sam-
ples 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.
The following subsections are intended to highlight a number of in-house
factors which can lead to inaccurate analytical data in any laboratory
and to present guidelines for avoiding these pitfalls. Further details
of many of the areas mentioned will be given in appropriate later sec-
tions of this Manual or are covered in the cited sections of the EPA PAM.
H ANALYTICAL BALANCES
Most laboratories contain balances of two types. Rough triple beam or
Dial-0-Gram balances are used for weighing approximate amounts of mater-
ials 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)], 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 *0.0001 g (error and uncertainty),
the fourth decimal place being obtained by estimation and therefore the
final significant figure recordable (Subsection 2Kb.). This leads to
51
-------
Section 3H
a total accuracy and precision of
°'Q002 g x 100 = 1%
0.0200 g
in weighing 20.0 mg of pesticide standard by difference (two weighings),
as is usually done in preparing primary standard solutions (Subsection
3.M.). 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 main-
tenance 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 which 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, conven-
ient, 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:
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 obtained from someone
experienced in its proper use.
52
-------
Section 31
31 PURITY OF REAGENTS
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-n-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, aroclors, and PCBs
present in laboratory air and the cause of solvent, reagent, and glass-
ware contamination (1). Sulfur and sulfur-containing compounds can be
present in solvents and column materials, as well as in certain sub-
strates (onion, cabbage, turnips), and can give rise to peaks easily
confused with pesticides (2).
Commercial solvents designated "pesticide quality" 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 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 jjl into the gas chromatograph 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 con-
taining oxidants are especially troublesome in causing losses of
53
-------
Section 31
organophosphorus pesticides, 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 eluents. 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 percent ethanol
normally present in ether and also removed by the partitioning is
replaced (EPA PAM, Section 5,A (1)).
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
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, as well
as HC1 or oxidation products in chlorinated hydrocarbons, benzene in
hexane, and the afore-mentioned 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 T.-^ 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 which 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.
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 u membrane filter, and dissolved gases are removed by
heating, stirring under a vacuum, or ultrasonic 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 manufacturer.
54
-------
TABLS 3-3
Contaminant
SOLVENT 7URITY IN LC [53
Possible Source Effect
Section 31
Particulete During tran*f«r,
m«tt«r (duet* to} unclean veesels
Removal
May block in-lina filter*, Filtration through membrane
lodge in pump seals, or.ao- filter,
cumulate at column head.
Hater
Alcohol
»olv»nt
preparation or
manufacture
Stabiliser in
chloroform, impurity
In hydrocarbons
Variable column activity,
k* variation, stability of
Bilicato ester bonded phases.
Similar to water,
Drying over molecular sieve or
anhydrous sodium sulfate.
From hydrocarbons, pass through
activated silicai from CHC1,
extract with HgO, dry with
Hydrocarbons
(in water)
Organic matter
Baseline instability
during gradient elution.
Passage through porous
polymer column or C.g bonded
phase.
Peroxides
(in ethers)
Degradation
Oxidation of bonded phase
(e.g. , -NHg to -N02),,
reaction with sample,
column deactivation or
degradation (polystyrene-
based) -
Distillation or passage
through activated silica
gel or alumina.
HC1, HBr
(halogenated
solvents)
Degradation
Column degradation esp.
bonded phases, UV absortance
(bromide), stainless steel
attack.
Passage through activated
silica or calcium carbonate
chips.
BHT
Antioxidant in THP UV absorbing.
Dissolved oxygen Solvent preparation
Unknown
CV-absorbing
High boiling
compounds
Algae in water
From manufacture
From solvent
manufacture
Growth during
prolonged storage
Degrades polystyrene-based
packing, oxidizes S.P'-oxy-
dipropionltrlle, may react
with sample.
Dietillation.
Degas solvent with vacuum
or heat.
Baseline instability or drift Try activated silica or
during gradient elution, high alumina; distillation for
detector background. organics and reorystalli-
zation or passage Over ion-
Contaminates collected
sample in preparative 1C.
Can plug in-line filters,
column entrance frits.
exchange column for inorganics.
Distillation,
Distillation from alkaline
permanganate or discard.
55
-------
Sections 3J, 3K
3J DISTILLATION OF SOLVENTS (6, 7)
Distill reagent grade acetonitrile over reagent grade AgNOo (3 g/1)
with an all-glass fractionating column equipped with a water cooled
condenser. Discard about the first 10 percent of the distillate and
leave the last 20 percent of the solvent in the flask. Rinse the
flask and use fresh AgNO-j and boiling chips for each distillation.
Test the distillate for interference. Alternatively, to 4 liters of
acetonitrile add 1 ml of 85% H3P04, 30 g P205, and 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 percent
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 MISCELLANEOUS 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 Florisil oven. Plastic fiber pack liners have
been found to contribute PCBs and phthalates to Na2SO^ which 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 acid which
interfere with separations of pesticides from PCBs were reduced by
extraction of the adsorbent with solvent (9).
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 preextraction 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 UV absorbing species and should
be tested for impurities. Better still, avoid the use of plastic wash
bottles and use all-glass ones.
56
-------
Section 3K
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 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 = 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 significantly 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 appropriate immiscible solvent,
e.g., benzene or isooctane, followed by boiling, if necessary, to remove
the residual solvent. Aqueous salt solutions such as 2 percent 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.
Materials in which the initial sample is stored must be given consideration.
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 polyethylene
contains a contaminant which reacts with AgNO^ chromogenic reagent,
giving a TLC spot close to that of _p_,£/-DDE and having similar GC
retention times to jD,j3*-DDE and £,j3?-DDE. Glass containers with aluminum
foil or Teflon-lined caps are generally acceptable as sample containers
and for storing purified reagents.
Other examples of problems with reagent contaminants have appeared in
the literature. Bevenue ej^ al. (11) 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 (12) 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 (13) found that Na2SO< , 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 (14) reported on the contribution of extraneous components
57
-------
Section 3L
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 jst al. (15) found contamination
of acetone with an impurity corresponding to CC14 and interfering
in the analysis of the latter pesticide (fumigant) by EC-GC. It was
shown that this contamination could be caused by CC1» in the laboratory
atmosphere, possibly arising from the use of aerosol propellent cans
for spraying thin layer chromatograms. Trotter and Young (16) found
that impurities in SbCl5 reagent caused erratic recoveries of PCBs
in perchlorination procedures.
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.
3L 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.
Pipets 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
58
-------
Section 3M
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.
3M 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 laboratories' 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 solu-
tions or in their dilution would be reflected in the accuracy of analyti-
cal results for this entire period. Incorrect standards will result in
correspondingly incorrect analytical data even though first class tech-
nique 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 forma-
lized and standardized for all staff chemists within the laboratory group.
By so doing, it should be possible for any other staff chemist or a super-
visor 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 log book 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, concentra.tion 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 log book. 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. A code number such as
26 AI might be used, the number 26 indicating the sequential mixture
number, the letter A referring to a specified group of component com-
pounds, and the sub 1 to the concentration range.
59
-------
Section 3M
Sample sheets for maintenance of the reference standards log book are
shown in Figs. 3-D, 3-E, and 3-F. These forms are in routine use at
the EPA laboratory at Research Triangle, NC.
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, al-
though in the parlance of the pesticide chemist, analytical grade stan-
dards of 99 percent or higher purity are referred to as primary standards.
Purities of standards are commonly greater than 99 percent and seldom
less than 95 percent but may be lower in some cases. For example,
chlordane and toxaphene are available in technical grade with 60-70
percent 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
Laboratory 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-600/9-76-012 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.
Concentrated stock standard solutions are conveniently made up at a
200 ng/^il 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 percent, the weight will
0.990
be 20.0 _ oo o „
-6T90Q-- 22-2m§-
60
-------
Figure 3-D
Section 3M
PREPARATION OF CONCENTRATED STOCK STANDARDS
No,.
Compound
Final Gro.ss Wt_
*Tare Wt
Net Wt
**Adj. Net Wt_
Date
Chemist
Lot No.
mg
Purity
Dilution Vol._
Concentr.
ml
ng/yl
No,.
Con\pound_
Final Gross Wt_
*Tare Wt
Net Wt_
**Adj. Net Wt_
Date
Chemist
Lot No.
Purity
Dilution Vol._
Concentr.
_mg
ml
ng/Vl
No,
Net Wt
**Adj. Net Wt_
Date / /
Chemist
Compound
Final Gross Wt
*Tare Wt
Lot No. Purity %
g Dilution Vol. ml
g Concentr. ng/yl
mg
No,.
Compound
Final Gross Wt
*Tare Wt_
Net Wt
**Adj. Net Wt_
Date
Chemist
Lot No.
Purity
Dilution Vol_
Concentr.
_mg
ml
ng/pl
*If weighing into a beaker, this is the empty beaker weight. If weighing from a
dropping bottle, this is the initial weight of bottle and contents.
**Correcte<3 for purity of primary standard.
61
-------
Figure 3-E
Section 3M
PREPARATION OF STANDARDS OF INTERMEDIARY CONCENTRATION
NO,.
Compound
NO,.
Compound
Date / /
Chemist
Strength of Concentrated Stock_
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
ng/M.
ml
ml
ng/Vl
Date / /
Chewist
Strength of Concentrated Stock_
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
_ng/WL
ml
ml
ng/Vl
NO,.
Compound_
NO,.
Compound
Date
Chemist
Strength of Concentrated Stock_
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
_ng/Ul
ml
ml
Date / /
Chemist
Strength of Concentration Stock
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
ng/Vl
ml
ml
ng/Vl
62
-------
Section 3M
figure 3-F
PREPARATION OF FINAL WORKING STANDARD SOLUTIONS
MO, Date / / Chemist
1.
2.
3.
4.
5.
6.
7.
8.
1.
2.
3.
4.
5.
6.
7.
8.
1.
2.
3.
4.
5.
6.
7.
8.
Cone, of
Parent Sol. Parent Sol. Aliq. Vol. Dilution Final Cone.
Compound Number ng/pl ml Vol. (ml) pg/yil
Date / / Chemist_
Cone, of
Parent Sol. Parent Sol. Aliq. Vol. Dilution Final Cone.
Compound Number ng/yl ml Vol. (ml) pg/yl
NO, Date / / Chemist
Cone, of
Parent Sol. Parent Sol. Aliq. Vol. Dilution Final Cone.
Compound Number ng/yl ml Vol. (ml) pg/pl
63
-------
Section 3M
Toxicity levels and relative stabilities are important factors which
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 inha-
lation of vapors. The stable organochlorine compounds may be stored
at room temperature in tightly sealed containers, while organophos-
phates, 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 deep freeze, 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 of 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 glassware 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 esti-
mated amount of standard transferred directly into a volumetric flask,
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/yl. 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
64
-------
Section 3M
enough solvent (e.g., the solvent 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.
Benzene will dissolve virtually all pesticide primary standards, gentle
heating and stirring on a water bath being occasionally required. Hexane
and isooctane are other widely applicable solvents. Dioxane is useful
for triazines and methylene chloride for carbamate insecticides.
All concentrations of all standard solutions are preferably stored under
refrigeration (-15 to -18°C). Solutions should not be stored in volu-
metric flasks but transferred to more convenient containers such as
inverted $ 30 or 60 ml glass stoppered bottles (Corning No. 1560).
Organochlorine pesticide concentrated standards are safely kept at least
six months. Organophosphate solutions should be remade at least every
four months. Pesticide primary standards or solutions should never be
stored in the same refrigerator with raw samples awaiting analysis unless
extreme care is taken to carefully seal both samples and standards to
prevent contamination from spillage or vapors.
c. Intermediate Concentration Standards
It is usually necessary to prepare standards of intermediate concentration
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 vil
in a 5.0 ul transfer will be grossly magnified when a 5 to 10 yl in-
jection of the resulting solution is chromatographed.
The solvents for these standards are usually isooctane, hexane, or
methylene chloride, depending upon the solubility of the particular
pesticides. The former is preferred, if possible, because of its higher
boiling point and therefore smaller chance for solvent evaporation as
the solution is used. Changes in volume can occur on opening and closing
of the container or even due to volatile solvent leaking through closed
containers.
Separate solutions of each compound or a standard mixture can be pre-
pared 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
65
-------
Section 3M
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
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.
Details of storage periods and conditions are given in Section 3, B of
EPA PAM. Standard solutions should never be exposed to sunlight or
fluorescent lighting for extended periods, either at room or refrigeration
temperature.
d. Working Standards
Working standards are generally made up as mixtures, the actual com-
binations being dependent upon the compounds of interest and the ability
of the analytical method to resolve them. Each working standard mixture
should be made up at two or even three concentration levels, depending
on variations in pesticide concentrations in routine samples. No com-
pound should be present in such concentration that when injected into the
gas chromatograph the linear range of the detector will be violated. If
p_,p_'~DDT is present in a standard mixture, neither. p_,p_'-DDD nor p_,p_'-DDE
should be present since these compounds are breakdown products of 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 chlorinated pesticides for laboratories analyzing
tissue samples by EC-GC with the recommended columns (Subsection 5L)
are given in the EPA PAM, Section 3, B. A typical mixture, diagrammed
in Figure 3-G, is prepared as follows: weigh 20.0 mg each of primary
standard lindane, aldrin, dieldrin, o_,p_'-DDT, and p_,p_'-DDT into sepa-
rate 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
66
-------
Section 3M
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 p_,£'-DDT and p_,p_'-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.
Figure 3-G. Serial dilutions of pesticide standard mixture
PREPARATION OF WORKING STANDARD A.
Lindane Aldrin Dieldrin o,p'-DDT p,p'-DDT
D D D G D Primary Standards
* ' * * * 0 H
20 mg each
lo.sml I05ml I
_ Cone. Slock Stdi.
1.0 ml (200ng/^leooS)
409///I
Interm. Stock Stdl.
Final Working Standard Mixture
e. Storage of Standards
Over the last several years this laboratory has been attempting to
evaluate the factors that affect pesticide analytical standard
integrity. Briefly the following have been noted:
(1) Solutions of organochlorine and organophosphate standards
at working concentration and ambient temperature are stable with
respect to compound degradation for up to six months. This
tentative conclusion is based on observations to date of 24
organochlorine and 20 organophosphate compounds.
67
-------
Section 3M
(2) Standard solutions do concentrate by evaporation from glass
stoppered volumetric containers at ambient temperature. Hexane
evaporates 3.4 times faster than isooctane from closed
volumetric containers.
(3) The evaporation rate of isooctane from stoppered volumetrics
is independent of the container size. Therefore, the relative
loss of solvents is less with the increase in container size.
For example, both the 100 and the 10 ml volumetrics in the
experiment lost 280 mg of solvent in ten weeks. For the 100 ml
container this represented 0.4% of the solvent. But for the
10 ml container this represents 4% of the solvent.
(4) Solvent loss from screw capped prescription bottles or vials
at ambient temperature is solvent and seal dependent. Hexane
evaporates faster than isooctane. The solvent loss is dependent
upon the seal obtained between the cap and the mouth of the
bottle. (Some bottles leaked more than others.) If a Teflon
cap liner is added to the cap, it must be the correct size. If
the Teflon cap liner must be pushed into the cap, it is too large
and will crimp. Crimped cap liners will allow solvent evaporation
at any temperature, even in the deep freeze. Vial caps with
hard Teflon seals allow less solvent evaporation than caps with
a soft foam backing behind the Teflon. Solvent loss rate from
different screw cap bottles and vials ranged from negligible to
significant percentages. This variability was observed within
lots of supposedly identical containers.
It has been concluded that analytical reference standards should be
kept in relatively large volume isooctane solutions, due both to
the slower evaporation rate of the solvent and the slower relative
evaporation rate due to the larger volume. When not in use, standard
solutions should be refrigerated to reduce solvent loss. The use of
prescription bottles is cautiously recommended, provided that the
solvent evaporation is monitored. One method is to fill the bottle
and mark the meniscus with tape. When the solvent evaporation is
obvious, replace the standard with a new one. Organochlorine and
organophosphate standard solutions should not be kept longer than
six months.
If hexane is used, the recommended replacement time for solutions
is 1/3 of the above recommended times.
3N CALIBRATION AND MAINTENANCE OF THE GAS CHRQMATOGRAPH AND ACCESSORIES
It is essential that the entire gas chromatograph be maintained in top
operating condition if high quality analytical data is to be produced.
68
-------
Section 3N
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 mini-
mal down time. Correct procedures for the.operations mentioned (e.g.,
silylation and conditioning of columns, obtaining background profile)
will be described in Sections 4, 5, and 9 of this Manual.
Certain checks should be made daily, others on a weekly or monthly basis.
Table 3-3 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 ea.ch 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 Control 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 as:
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.
69
-------
Table 3-4
PERIODIC CHECK LIST FOR MS CHROHAT06RAPH (EC DETECTOR)
DAILY
WEEKLY
MONTHLY
1. Check response with standard
dixture and relate to previous
cay.
1. Change glass wool'plug at
column inlet.
1. Clean recorder slide wire
with Freon MS-180.
2. Change glass insert sleeve in
injection port (end of day).
2. Change septums. 3/
Z. Change pyrometer batteries and
clean battery contacts with
Freon MS-180.
3. Check recorder electrometer
zero and noise level at
operating attenuation.
3. Run background current profile
and polarizing voltage response
curve. 4/
3 Check recorder speed at settings
ir. normal use.
4. _Check carrier gas flow rate
'through each column with bubble
neter (early morning and
afternoon).
4. If endrin is a compound of interest^
chromatograph pure standard at a
concentration and attenuation that will
produce a peak gf 50 to 60 percent fsd. 5/
4. Check glass flow system for
leaks using "SNOOP". 6/
5. Check temperature of detector,
inlet, transfer'and column oven. V
Check for any shifting of column
packing resulting in forward move-
ment beyond the bottom of the column
exit nut and/or settling in excess of
1/2" from the glass wool inlet plug.
5. Evaluate performance of each
column with special standard
mixture.
6. Chroroatograph standard p_,p_'-DDT
on each working column used at a
concentration and attenuation that
will produce a peak of 50 to 60
percent full scale deflection (fsd). 2J
6. Check entire instrument for
loose connections and frayed
wire insulation.
7. Check all rotameters and flew
controllers for proper float
action.
CA
(D
O
]_/ Oven temperature should be
monitored by an outboard
thermometer.
3/ This frequency assumes the use
of improved silicone rubber. Old
type requires more frequent changing.
5/ The formation of one or two
additional peaks indicates
on-column breakdown.
ZJ The formation of £,£'-DDD and/or
£»£'-DOE indicates on-column breakdown.
4/ Daily check may be indicated if
large numoers of sample extracts
are injected.
6/ Checks should be made whenever a
new detector or new tank of gas
is installed, or whenever erratic
baselines are observed.
-------
Section 30
f. Background current obtained on newly installed column and subse-
quent 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 mix-
ture, recording absolute and relative retention data and efficiency
based on the p_,p_'-DDT peak.
Chlorinated Pesticide Mixture for GC
Column Evaluation
Compound
a-BHC
B-BHC
Lindane
Heptachlor
Aldrin
Kept . Epoxide
p_,p_'-DDE
Dieldrin
Endrin
o,p_'-DDD
p_,p_'-DDD
o_,£'-DDT
p_,p_'-DDT
Concentration ng/yl
0.010
.040
.010
.010
.020
.030
.040
.050
.080
.080
.080
.090
.100
30 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
71
-------
Section 3P
laboratories, it is essential that uniform standard analytical metho-
dology be used by all. In the EPA program, this methodology is devel-
oped, tested, and collected in the Analytical Manual by the coordinating
laboratory of the quality assurance program, in close cooperation with
the EPA methods development section. Individual laboratories in the
multi-laboratory system are encouraged to suggest improvements in exist-
ing 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 per-
formed by laboratories deviating in some way from the standard procedures.
It is important, therefore, that laboratories adhere to standard analyti-
cal methods, but also that they report any problems with them to the
coordinator so that these methods can be further researched 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 diffi-
culty rather than making revisions in the method which cannot be fully
studied and statistically evaluated by the individual.
3P 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 percent 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 control program
(Section 2), it can be stated without reservation that laboratories
lacking a systematic internal control program more than likely will do
72
-------
Section 3Q
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
which 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.
To cite a specific instance, one laboratory joining our program and
reporting the results of their first analysis of a spiked water sample
reported the presence of 2»£?~DDE, 2>2.'~DDT> £»2?~DDT> heptachlor
epoxide, C[,2'-DDE, and dieldrin. The actual spiking composition was
HCB, oxychlordane, 2>£'-DDE, 2>£*-DDT, and Aroclor 1254 (PCB). In other
words, the analyst found two of the compounds which were actually
present, four which were not present, and missed three which 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 than 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 shown
above should not be regarded as unusual.
3Q REFERENCES
(1) Singmaster, J. A., and Crosby, D. G. , Bull. Environ. Contain.
Toxicol., JL6, 291 (1976).
(2) Ruzicka, J. H. A., and Abbott, D. C., Talanta, 20, 1282 (1973).
(3) Williams, I. H., J. Chromatogr. Sci.. JLL, 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 Foods, Department
of National Health and Welfare, Canada, 1973, Section 12.1 (b).
73
-------
Section 3Q
(7) FDA PAM, Section 121.
(8) Wanchope, R. D., Anal. Chem., 4_7, 1879 (1975).
(9) Huckins, J. N., Stalling, D. L., and Johnson, J. L., J. Ass.
Of fie. Anal. Chem., 59_, 975 (1976).
(10) Kirchoff, J., Dt. Lebensmitt. Rdsch., 70., 284 (1974).
(11) Bevenue, A., Kelley, T. W., and Hylin, J. W., J. Chromatogr., 54,
71 (1971).
(12) Levi, I., and Nowicki, T. W., Bull. Environ. Contain. Toxicol., _7»
133 (1972).
(13) Levi, I., and Nowicki, T. W., Bull. Environ. Contam. Toxicol., ]_,
193 (1972).
(14.) Bevenue, A., and Ogata, J. N., J. Chromatogr., 61, 147 (1971).
(15) Baker, P. B., Farrow, J. E., and Hoodless, R. A., Analyst, 98,
692 (1973).
(16) Trotter, W. J., and Young, S. J. V., J. Ass. Offic. Anal. Chem., 58,
466 (1975).
74
-------
Section 4
EVALUATION, STANDARDIZATION, AND USE OF MATERIALS FOR
PESTICIDE RESIDUE ANALYSIS
4A ADSORBENTS
Cleanup and preliminary fractionation 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 pesti-
cide mixtures through a prepared column. Most materials may be acti-
vated by strong heating and some may be activated for a particular
purpose by pretreatment with an acid or base (e.g., alumina) or an
organic solvent (e.g., charcoal). Deactivation of an 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, requires careful pretesting of the adsorptive properties
of each batch 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
furnishing of uniform, standard quality Florisil to other EPA labora-
tories and to laboratories with direct contracts to conduct environ-
mental monitoring. Procedures specified by the program coordinator
and other available standardization methods will be described in the
following subsections.
a. EPA Procedures for Handling and Evaluation of Florisil
Details are given in Section 3,D 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. Upon receipt of the entire lot, 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 Ibs of Florisil.
75
-------
Section 4A
Evaluation of Florisil for use in a modified Mills, Onley, Gaither
procedure is made by heating Florisil in an Erlenmeyer flask overnight
or longer at 130°C in an oven which 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 Na2SO^ immediately prior
to use, as described on page 6 of the EPA PAM, Section 5, A, (1).
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 chlorinated and phosphate
insecticides are prepared at levels of 20-250 pg/yl and 5 ml of each is
added to separate columns and 5 ml hexane to the third as a control.
Elution is carried out with 200 ml 6 percent diethyl ether in petroleum
ether in two 100 ml portions and similarly with 15 percent and finally
50 percent 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 percent 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 percent v/v ethanol as commercially supplied, or if absolute
ether is used, exactly 2 percent v/v ethanol should be added to obtain
the correct polarity to result 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
compounds 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 percent ethanol in the second column
and petroleum ether with 4 percent 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 specific 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.
76
-------
ELUTION PATTERNS AND RECOVERY DATA FOR FLORISIL, LOT # 2961
BY METHOD SECTION 5,A,(1), EPA PESTICIDE ANALYTICAL MANUAL
FLORISIL COLUMN PREPACKED AND HELD IN 130°C OVEN AT LEAST 24 HRS BEFORE USE
RELATIVE HUMIDITY IN LABORATORY 60 %
ELUTION INCREMENTS (ml)
Fraction 15$ Fraction 50$ Fraction
Compound
a-BHC
3 -BHC
Lindane
Hentachlor
Aldrin
Kept. Epox.
Dieldrin
Endrin
p,p'-DD£
p , p ' -DDD
o , p ' -DDT
p,p'-DDT
Ronnel
Methyl
Parathion
Malathion
Ethyl
Parathion
Diazinon
Carbophenothion
0-100
100
100
100
100
38.9
100
100
100
100
100-200
61.1
i
200-300
300-400
22.8- ) ?7.2
49.8
4.1
14.4
71.2
50.2
95-9
85.6
28.8
— . - — - .. _
400-500
94.5
500-600
5-5
!
Recovery, %
97-3
97-3
104
97.7
108 i
96.9
102.3
100
105
102
86.9
103
86.6
91.7
46.7
CO
n>
o
rt
H-
O
0
Numerical values represent the percentage of each compound eluting in the given cut.
-------
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*
Hept. Epoxide
DJeldrin
Endrin
Diazinon
Methyl Parathion
Ethyl Parathion
Malothion
No EthanoL
I
100
n
87
100
100
B
16
ffl
13
100
84
"Eluting mixture!:
Fract. I • 6% Et,O in pet.ether
Froct.ll-15% " " " "
Fract. Ill-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 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
(20.0 ml, containing 400 mg) is added to 2.00 g 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
110
Lauric Acid Value of
batch
x 20 g
Lauric Acid Value = 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.
78
-------
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 yg 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 percent diethyl ether in petroleum ether. The three frac-
tions are concentrated prior to gas chromatography on an appropriate
column. Heptachlor, heptachlor epoxide, ethion, and carbophenothion
should elute with good recoveries in the 6 percent fraction; parathion,
dieldrin, and endrin in the 15 percent fraction; and malathion in the 50
percent 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 7J). 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 percent (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 percent methylene chloride
in hexane, 10 percent ethyl acetate in hexane, and 30 percent 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 percent. Late elution, especially of malathion, which
just barely elutes with 10 percent ethyl acetate, indicates insufficient
deactivation and the need for more polar solvents. Early elution indi-
cates 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 percent deactivated adsorbent on a small
scale has been conveniently and successfully carried out as follows [7]:
79
-------
Section 4A
(1) Woelm silica gel is activated for 48 hours at 170-175°C.
(2) Add two ml water to 10 g of adsorbent in a tightly capped Teflon-
lined screw top vial.
TM
(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. Standardization is
carried out, as above, by packing the required column, adding an
aliquot of standardization solution containing the pesticides 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 Holden and Marsden cleanup procedure (Subsection 7A) 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 6 M HC1 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
concentrated hydrochloric acid, and stir magnetically while boiling
for one hour. Add 500 ml of water, stir, and boil another 30 minutes.
Recover the charcoal by filtering through a Buchner funnel, wash with
water until washings are neutral, and dry at 130°C. (FDA PAM, Section
121). As an alternative procedure (9), add 225 g charcoal to 1.2 liters
of ethanol-conc. HCl-water (50:10:40) and reflux for one 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 charcoal 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.
80
-------
Section 4A
g. Magnesium Oxide (Sea Sorb 43)
Slurry 500 g with enough distilled water to cover it in a one 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 percent
water is absorbed in this procedure. Store in a closed jar (FDA PAM,
Section 121; 9).
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 required 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 eluent. 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
consistent with the activity (polarity) of the adsorbent and the
polarity of the sample. Use the least polar solvent that will elute
the pesticides 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
81
-------
Section 4B
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 chromatographic 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 inter-
laboratory quality control program described earlier in Section 2.
Some of the operational instructions, fully covered in Sections 4,A and
4,B of the EPA PAM, will be briefly reviewed in this subsection, 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 chromato-
graph, each compound is swept through the column at a rate that is
determined by the 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+fl)
~L
COLUMN
DETECTOR'
nm
CHROVATOGRAM
:DCI
_.^LJLZHD
irm
"Atfl
82
-------
Section AB
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 have been
used when mass spectroscopy and gas chromatography are used in com-
bination for identification or assay (Subsection 8H).
GC tubes are usually made of borosilicate glass. Copper and stainless
steel are best avoided because both can cause decomposition of com-
pounds unless special precautions are taken.
Commercial solid support materials are usually composed of flux-
calcined diatomaceous earth which may be treated by acid- or base-
washing or silanization. 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 interaction 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. Adsorption 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 (Section
4F and 41). As a general rule, column efficiency increases as the
particle size of the support decreases, but a greater carrier gas
pressure is required to maintain a given 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 pesticide 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.
83
-------
Section 4C
4C 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 chromatographer. 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 very high
efficiency (4,530 plates). It will be observed that on column .B, all
seven peaks give base-line 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
_P_,JD'-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 £,p_'-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.
84
-------
Section 4C
Figure 4-C. Calculation of column efficiency and resolution.
Efficiency: N=
Resolution: R=
L,
Figure 4-D. Effect of column efficiency on pesticide resolution.
85
-------
Section 4C
Factors that influence column 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 adsorption 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 percent, and also by the reduced capacity
for sample components.
Figure 4-E. Effect of stationary phase loading on column efficiency.
S'i, DC-200/7.SMQF-1
86
-------
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 increasing 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 _p_,_p_'-DDT has been found to approximate
these characteristics for a column. This retention range can be
obtained by operation of lower load columns (3-6 percent) 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 per-
centage 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.
87
-------
Sections 4E, 4F
4E COLUMN STABILITY
It is desirable to use columns that are heat-stable or "bleed" re-
sistant and that continue to function properly under injection loading
with dirty extract. Liquid phase bleed is evident from a persistently
drifting baseline and the inability to obtain a normal level of
standing current (Subsection 5C) from an electron capture detector.
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 conversion of pesticides,
erratic recoveries, and unsymmetrical peaks (see Figure 4-1).
4F RESISTANCE TO ON-COLUMN COMPOUND DECOMPOSITION
Unless a column is properly prepared, conditioned, and maintained, it
can cause such compounds as endrin and/or £,£f-DDT to undergo some
degree of decomposition. The main symptom of endrin decomposition is
a greatly reduced endrin peak with the formation of one or two
additional peaks arising from decomposition products. £,£T-DDT
decomposes to £,£'-000 and, in extreme cases, to £,£*-DDE.
Newly packed columns should be specially treated with a silylating
agent such as Silyl 8 to reduce the number of active adsorption sites
which 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.
88
-------
Section 4F
Figure 4-G. Reduction in breakdown of endrin resulting from column
silylation
BEFORE SILYLATION
12
24
28
32
36
40
I
44
B
AFTER SILYLATION
-------
Section 4F
Silylation 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 silylation did not im-
prove the situation. On the average, however, silylation clearly
improves the gas chromatographic behavior of endrin.
DDT breakdown is manifested by the appearance of p_,p_'-DDD and/or p, p' -DDE
on the chromatogram resulting from the injection of pure analytical grade
p_,p_'-DDT which is known to be free of these metabolites as impurities.
This problem is associated with overloading of the column packing adja-
cent 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) which caused a total of 25 percent
decomposition of the DDT peak to its two metabolites. This chromato-
gram was obtained in a laboratory where the injection insert had not
been changed for three weeks.
Figure 4-H. Breakdown of p_,p_'-DDT on 4% SE-30/6% QF-1 column.
FIGURE 4-H
A - No Conversion
BREAKDOWN OF DDT
SE-30/QF-1
B - Maximum Conversion
Figure 4-1 illustrates an extreme case of overloading of a column of
2% OV-1/3% QF-1. Chromatogram A is from a standard mixture of seven
pesticides on a freshly prepared column. The column was then discon-
nected from the detector so the exit end vented inside the oven. Eight-
een consecutive injections were then made of fatty tissue extract after
elution with 15 percent 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
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 p_,p_'-DDT to p_,p_'-DDD (in actuality, the ratio of these changed from
8:10 to 4:10). A clean Vycor glass insert was then installed in the
90
-------
Section 4F
injection port, the system re-equilibrated for 30 minutes, and another
equal volume of standard mixture 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 importantly, properly maintained and monitored columns should pro-
vide top performance without problems for many thousands of injections.
Figure 4-1. Chromatograms illustrating column overloading and subse-
quent rejuvenation.
91
-------
Section 4G
4G HOMEMADE VS. PRECOATED PACKINGS
The decision whether to make column packings or to buy them precoated
confronts 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 pro-
duce high quality packings. Notwithstanding, 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_,p_'-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 p_,p_'-DDT.
e. Final acceptance of each lot purchased to be based on buyer's
evaluation at time of delivery.
The final decision of whether to purchase or prepare column packing may
depend on the situation in a given laboratory. The successful formu-
lation of column packing in small batches requires a degree of expertise
somewhat beyond the purely scientific. The procedure has been described
as 50 percent science and 50 percent art. If some particular indivi-
dual 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 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 tech-
nique". The slurry in the beaker comprised of liquid phase, support,
92
-------
Section 4G
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.
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 con-
structed that a high volume of nitrogen can be blown 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.
As no preparation technique is presented in the EPA PAM, one method is
offered below for the benefit of laboratories which 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 production 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 gram total batch size, compute the amount of liquid
phase(s) to weigh in 30 ml beaker(s) on an analytical balance.
b. Weigh out liquid phase to two-place accuracy. If making mixed-
phase packing, weigh each liquid phase in a separate beaker,
c. With a 25 ml graduate, transfer 15 ml of the appropriate solvent into
each beaker. Stir with a 3 inch glass rod until the liquid phase
is completely dissolved.
d. 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 in that 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 appli-
cations of 7-9 ml of solvent, the exact amount depending on the
appropriate solvent/support ratio.
e. 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
93
-------
Section 4G
weigh out for a 21 gram batch is the difference, in grams., between the
total amount of liquid phase weighed and 21 grams. For example, with
a 21 gram batch of packing of 4% SE-30/6% QP-lt
SE-30 .040 x 21.0 = 0.84 grams
QF-1 .060 x 21.0 = 1.3 grams
2.1 grams
Total liquid phase:
21.0 - 2.1 = 18.9 grams of support
f. Attach the flask to a rotary (Rinco) evaporator.
g. Mix slowly for 10 minutes at room temperature with just enough
vacuum applied to hold the flask on the evaporator.
h. Advance the hot plate control sufficiently to raise 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 condition 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 suf-
ficiently to completely 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 col-
umns 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.
k. 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 two hours, or overnight.
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
determined, if required, by exhaustive Soxhlet extraction in glass
thimbles or standard low temperature or thermal ashing procedures.
94
-------
Section 4H
4H 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 preferred to prevent degradation often associated
with metal columns. U-shaped, 6-foot, glass columns are used in the
Tracer MT-200 gas chromatograph which 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 can be silanized prior to packing for chromatography of
especially labile compounds, but this has not been found necessary for
routine work.
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 chromato-
graphs such as the MT-220, one end of the column should be 1 inch
shorter than the other.
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. Note; 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.
95
-------
Section 4H
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 wool is packed by hand, the hands should be
carefully washed with soap or detergent, rinsed, and dried to
minimize skin oil contamination of the glass wool. Glass wool
can be silanized by treating with 10 percent dimethyldichlorosilane
in toluene for 10 minutes followed by rinsing with toluene and
treating for an additional 10 minutes with anhydrous methanol and
air drying, or the prepared material can be purchased commercially
(e.g., from Applied Science Laboratories). The column is now
ready for conditioning.
Straight metal columns are plugged at one end with glass wool. The
column is filled in portions with a minimum vibration and then the
plugged end is tapped on the floor to settle the packing material.
The other end is plugged after filling, and the column is formed
into the required shape. Coiled glass columns are packed by plugging
one end with glass wool and applying slight vacuum to this end to
suck the packing material in increments into the tube. The tube is
vibrated and tapped throughout to obtain a firm, uniform packing.
Stainless steel screens with 5 or 10 u openings (available from Supelco,
Inc.) have been recommended as replacements for glass wool in both GC
and LC columns to improve column efficiency and protect gas sampling
valves, if used, from column debris. Their use has not been evaluated
for pesticide analysis to our knowledge, but it is good practice not
to allow pesticides to contact hot metal surfaces during chromatography.
One excellent measure of a well packed column is the net weight of
packing per foot compared to a previous efficient column. Experienced
chromatographers can repeatedly prepare columns within ca. 2 mg/ft using
the same batch of packing.
96
-------
Section 41
41 COLUMN CONDITIONING
The column is conditioned, or made ready for routine use, by heat
curing, silylation or Carbowax treatment, and injection of a con-
centrated pesticide solution.
Heat curing of some EPA (Subsection 5L) and FDA prescribed GC columns
is carried out according to the following schedule in order to
eliminate liquid phase bleed, thereby improving the day-to-day stability
of sensitivity and baselines, quantitation, and the quality of chromato-
grams, and lowering the amount of detector cleaning needed:
Phase Oven Temp., °C —' Minimum Time, hr.
4% SE-30/6% OV-210 245 72
1.5% OV-17/1.95% QF-1 245 48
3% DECS -?-/ 235 20 -/
10% OV-210 245 48
10% DC-200 —t 250 16
10% DC-200/15% QF-1 (1:1) 250 72-120
15% QF-1/5% DC-710 (2:1) 240 120
— Carrier gas flow 60 to 70 ml per minute.
—' Shown for information only. Column not recommended for routine use.
Do not exceed this time period.
DC-200 columns are significan
out without carrier gas (11).
.
—' DC-200 columns are significantly improved if conditioning is carried
In general, it is desirable to heat cure the column at a temperature
ca. 50°C higher than its operating temperature with a flow of carrier gas
passing through the column in the normal direction. Details for the
97
-------
Section 41
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 may give a column with an altered phase ratio after conditioning
because of increased temperature stability. Although these columns
are more stable, conditioning before use is still required.
As mentioned previously, column efficiency and response, especially the
response of endrin, would slowly improve as new columns become
"seasoned" with use, but silylation 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 ul 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 silylation do not persist indefinitely,
and repeat treatment about once a month is recommended. Columns to be
used with flame photometric or thermionic detectors for detection of
organophosphorus pesticides should not be silylated 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 percent, with a 100
percent increase, or doubled response, being most likely. Silyl 8
conditioning has no beneficial effect 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 Guiffrida
to bleed from a two inch 10 percent precolumn heated in the chromatograph
oven at 230-235°C for 17 hours with a carrier gas flow of 20 ml/minute.
98
-------
Section 4J
The response characteristics of the column should be monitored with a
standard mixture of organophosphorus pesticides immediately after treat-
ment 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 silylated wool usually used.
Following the silanizing injections or Carbowax treatment and with the
oven temperature and carrier gas flow rate adjusted to the approximate
operating levels for the particular column, several successive injections
of a chlorinated pesticide priming mixture in the microgram range are
made onto the column with enough time between injections for all com-
pounds 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 which 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, azinphos-methyl, coumaphos,
and PGP.
4J EVALUATION OF THE COLUMN
Unfortunately, many chromatographers, after packing and conditioning the
column, proceed immediately to use it without making the effort to
systematically 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 two or three 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 raised 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 measure-
ments) . At least one hour, or preferably overnight, is allowed for the
chromatograph to equilibrate. The temperature and flow rate are
99
-------
Section 4J
rechecked after equilibration. Before making any injections, a back-
ground (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 which
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).
From the chromatogram of this mixture one can calculate the column
efficiency based on the peak from p,p_'-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 p,p'-DDT 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.
The absolute retention time of the p_,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 indicates 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 p,p_'-DDT and endrin. Columns indicating poor
resolution, efficiency, and/or retention characteristics which 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 p_,p_'-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 _lie main pesticide peaks, should not exceed 3
percent for DDT and 6 percent 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.
100
-------
Section 4K
Carbamate pesticides are especially difficult to chromatograph directly
on any current GC column without breakdown or tailing.
Reproducibility of the size of peaks when a compound is injected re-
petitively should be <2-3 percent. 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).
Priming injections of large amounts of a difficult compound, as mentioned
earlier, may allow maintenance of reproducibility for an adequate period
of time for an analysis. Difficult compounds should also be checked
for linearity of response (Section 50d) since one cause of non-linearity
may be on-column breakdown or adsorption.
4K 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 predominate samples. A
properly cared for column should provide service for many months. Off-
column injection of biological samples will enhance column life (Sub-
section 5J); frequent (daily) changing of the injection insert and
septum helps insure 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, chroma-
tograms showing excessive DDT breakdown, peak tailing, and depressed
peak height response will result. Changing the glass wool regularly
will usually restore proper performance.
The column packing near the inlet must also be replaced with fresh,
conditioned 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 (FDA 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
101
-------
Section 4K
•
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 silylation 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 organo-
phosphorus 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 percent 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 pre-
scribed 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 injection
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.
Further details of instrument maintenance, troubleshooting, and cali-
bration are given in Section 9.
102
-------
Section 4K
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
response 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 conditioned
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 described
in Section 2 has clearly indicated that in many laboratories the chroma-
tographer 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 Subsec-
tions 5A and 5B but are highlighted here because of the profound effects
on the day-to-day operation of GC columns. Figure 4-J 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_,p_'-DDE and dieldrin is normally as shown in Chromatogram A. One labora-
tory 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.
103
-------
Section 4L
Figure 4-J. Effect of temperature on resolution, 1.5% OV-17/1.95%
QF-1 column.
FIGURE 4-J EFFECT OF TEMPERATURE ON RESOLUTION
OV-17/QF-1
A. B.
17mm
Actual Oven Temp 201'C
Stated Carrier Flow 55 ml
Computed Efficiency 2,960 TP
23mm.
Actual Oven Temp 185'C
Stated Carrier Flow 61 ml
Computed Efficiency 3,310TP
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 percent 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.
4L 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. Ass. Offic. Anal Chem., 54, 1349 (1971).
[5] Mills, P. A., J. Ass. Offic. Anal. Chem., 51, 29 (1968).
[6] Analytical Methods for Pesticide Residues in Foods, Health Protection
Branch, Dept. of National Health and Welfare, Ontario, Canada, Pro-
cedure 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).
104
-------
Section 4L
(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) Ives, N. F., and Giuffrida, L., J. Ass. Offic. Anal. Chem.,
53, 973 (1970).
105
-------
Section 5
OPERATION OF THE GAS CHROMATOGRAPH
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
anything 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 is repetitive of
the instructions outlined in the EPA Pesticide Analytical Manual, but
because of its importance in analytical quality control, it was con-
sidered worthy of reemphasis. Section 4 should be consulted for material
on evaluation, standardization, and maintenance of GC columns and
Section 9 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 3-3 in
Section 3 outlines a series of periodic checks recommended for insuring
a continuing high level of chromatograph performance. Figure 5-A at
the top of the next page shows the MT-220 (Tracer, Inc.) gas chromato-
graph which is in widespread use throughout EPA laboratories. This is
a floor model chromatograph which 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 temperature and the carrier gas
flow rate (Subsection 5B) will have a great influence on the caliber
of performance 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
peak identification or too slow to be practical, or rising or erratic
baselines. Impaired resolution may preclude accurate quantitation of
two important pesticides which are not adequately separated, while
inaccurate retention values will make proper residue identification
difficult.
106
-------
Section 5A
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 should be checked monthly to determine if they are
delivering full voltage under load. A hint of 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
107
-------
Section 5A
by means other than the built-in instrument pyrometer. A precalibrated
dial thermometer with the stem 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
injection port temperature may lead to decomposition of heat-labile
pesticides, stripping of the partition liquid from the front end of
the column resulting in peak tailing, and increased 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 which 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 £,_p_'-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 chromato-
gram. There is a linear relationship between column temperature and
relative retention values so that comparison of this computed value
with those available for over 50 pesticides 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 £,_p_'-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 chromatograms 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
108
-------
Section 5A
values. These erroneous temperatures resulted from inaccurate
instrument 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_,j>'-DDT at a
temperature of 200°C and flow rate of 65 ml/minute. Under these
stated conditions the retention 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-J in 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
performance and suggestions for simple evaluation techniques for thermal
variables have been published (1).
109
-------
TABLE 5-1
RETENTION TIMES FOR jD.p/HDDT RELATIVE TO ALDRIN
Liquid Phase Temperatures, °C
170 174 178 182 186 190 194 198 202
1.5% OV-17/1.95% QF-1 5.57 5.39
4.0% SE-30/6.0% QF-1 4.04 3.92
5% OV-210 4.47 4.31
5.20 5.01 4.83 4.64 4.46 4.27 4.09
3.80 3.67 3.54 3.43 3.30 3.18 3.05
4.15 3.98 3.82 3.66 3.49 3.33 3.17
en
n>
o
o
3
-------
Section 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:
(1) a purge line running through the purge rotameter, flow controller,
and detector, and (2) the carrier gas flow line running through the
rotameters, the flow controllers, the column, and finally through the
transfer line into the detector. The choice of carrier gas is dictated
mainly by the requirements of the detector being employed. Nitrogen
is required for the usual pesticide detectors, except that the pulsed
mode of the electron capture detector usually employs argon with 5
percent methane. Flame detectors require gases such as hydrogen,
oxygen, and air for combustion.
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 percent
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 hydro-
carbons, and C02- 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 completely 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
°%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 detector, provided the
connections are at room temperature. Do not attempt to use "Snoop" on
a hot column. Another means of detecting leaks when using an electron
capture detector is by spraying connections with Freon MS-180 with the
111
-------
Section 5B
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
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.
Likewise, 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 rate (A).
112
-------
Section 5B
Figure 5-B. Effect of Flow Rate on GC Resolution
1.5%OV-17/1.95%QF-1
Column Oven Temp , °C
Retent lime p.p'-DPT,
Gamer gas flow, ml/mi
Oiart Speed, in./nun
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 chromatogratns
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.
The reader is directed to references (2-5) for general reviews of the
element selective pesticide detectors.
113
-------
Section 5C
5C ELECTRON CAPTURE
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
|3 -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 (6-8).
The EC detector is selective in principle for highly electronegative
compounds, but in practice it is the least selective of the widely
used pesticide detectors. Rigorous cleanup of pesticide extracts is
required to eliminate extraneous peaks due to compounds containing
halogen, phosphorus, sulfur, nitrogen dioxide, lead, and some hydro-
carbons. 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 pre-
paring 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 (9).
jj[/ Sensitivity of a GC detector is designated as the amount of pesticide
that will provide a peak whose height corresponds to some percentage
of the full scale recorder deflection (usually 10 or 50 percent).
Minimum detectable amount is that quantity of pesticide giving a
signal at least four times the background noise (random fluctuations)
at baseline.
114
-------
Section 5C
Sources of /3-radiation have usually been either tritiated titanium on
copper or a ^^Ni 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 detector with extract impurities or the bleeding of GC liquid
phases. It also extends the number of compounds that can be detected
and greatly reduces detector maintenance.
The EC detector is used with either a constant negative DC voltage
or an intermittently pulsed voltage (constant frequency or "plain
pulsed" mode) imposed across the anode-cathode. The former mode, which
is most used, requires nitrogen carrier gas, while argon plus 5-10
percent 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 responses (10), reproduci-
bility of response, independence of response to voltage, and when
working 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-modulated 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. This mode of operation provides an increased linear
response range without loss of detectability and a high degree of base-
line stability, but it has not been carefully evaluated for routine
pesticide work (11-13). Linearized ^%1 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^. This 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 pulsed EC cells require argon-methane carrier gas,
115
-------
Section 5C
but at least one commercial version has been designated to permit
efficient electron collection (0.1 pg detection of aldrin and
heptachlor) with nitrogen carrier gas (14). This model functions as
a concentration-type detector, as indicated by a dependence of
linearity range on flow rate and retention time, and has sufficient
baseline stability for temperature programming at the 10 pg level. A
constant current tritiated scandium EC detector with high frequency
pulsed operation (6) had an extended linear range to 10" and reduced
nonlinearities compared to those previously reported (15) for lower
frequency detectors with strongly electrophilic compounds such as
aldrin. The tritium source is more sensitive than nickel for a short
period of time, reaching maximum sensitivity after a few days of
operation. Then there is typically a constant loss in sensitivity,
requiring frequent recalibration and eventual foil replacement. The
sensitivity of the 63^1 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 (10).
Traditionally, the linear range of the tritium detector has been three
to five times greater than that of the DC 6%i detector. Figure 5-C
shows typical linearity curves for £,£r-DDD using these detectors, both
operated in the DC mode. When a certain point in concentration is
reached, the linearity curve begins to "plateau". For quantitation,
it is mandatory that concentrations of pesticide injected be within the
linear or straight line range of the detector. Since this range may
change with the age and use of the detector, standard curves must be
determined regularly. With the Tracer MT-220 chromatograph, operation
at an output attenuation of 10 x 8 or 16 on the E 2 electrometer or
Ifl2 x 8 or 16 on the SS electrometer will usually preclude violation
of the linear range of the EC detector simply because peaks generated
by excessive concentrations of pesticides will go off-scale. Samples
should be diluted to give quantifiable peaks at these settings so that
non-linear responses are avoided (see Subsection 50 on quantitation).
116
-------
Figure 5-C. Comparison of linear ranges of and i detectors for
Section 5C
6
Iltiiiililliiliillltittllillllil
f IMiifliMiiiiiiiiiii
& SiMiSSiSiSiBlIll
mm
; mm
11
11
I
I
•
llli
1
I
I
11111
11
11
I
1
:li
I
1
1
iiiiiii iiiiiiii
flillillllllHB
[I'mimlltimltlytlilltllillillmili:
ill
fl t i ILttil llltyif 1 1 ! IUWllil ! iUtMtii
llllIIIilH
1111111
ill ill
111111 1
iiiii llli
/to
Bo ,
•to •
The EPA analytical laboratories have mainly used parallel-plate % EC
detectors because cleaning can be done in-house under an NRC permit.
Details are given in the FDA Pesticide Analytical Manual, Vol. I,
Section 311.12, for cleaning a tritium EC detector. Foils may be re-
moved only by persons with an NRC license for this purpose. A cleaning
solution of 5 percent KOH in absolute ethanol is recommended for
cleaning radioactive foils. Tritium foils should not be exposed for
more than one hour, and aqueous solutions or traces of water should be
avoided. Mildly abrasive cleaning compounds and ultrasonic cleaning
apparatus may also be used. High temperature "-%i detectors are
probably in greatest general use, and concentric-design tritium
detectors are also used. Figure 5-D shows a Tracer, Inc. "-%i detector.
The column effluent entrance is shown on the left and the purge gas line,
polarizing voltage connector, electrometer input connector, and gas
effluent outlet (top to bottom) on the right. The heater-limit switch
fits into the nearest large hole seen on the top, front. The °%i foil
is a sealed source and is usually sent back to the manufacturer for
cleaning. It is frequently possible, however, to clean a nickel foil
117
-------
Section 5C
in the chromatograph by injecting 100 pi of water a few times Into a
300°C system employing an empty column. Purging the detector at 400°C
overnight may also be helpful. It has been reported (6) 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.
Figure 5-D. Tracer, Inc. 63N1 EC Detector
Response of EC detectors depends upon temperature (16); type, flow rate,
and pressure (17) 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 (18). The unpredictable
nature of these parameters causes anomalous responses, drifting base-
lines, variable sensitivity, and a limited, variable dynamic range in
the DC mode. Operating parameters must be optimized for each manufacturer's
detector. Electron capture detectors containing tritiated scandium (19)
and 147pr g0i
-------
Section 5C
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 an attenuation
setting of 12.8 x 10~9 A.F.S. when using a 1 mv recorder, a good
detector should produce a response of 60-80 percent full scale deflec-
tion, and with aging this will approach 30 percent, when the foil should
be replaced. A profile which 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 operat-
ing parameters for the column being used.
FIGURE 5-E, TYPICAL ELECTRON CAPTURE DETECTOR BACKGROUND
CURRENT PROFILES
119
-------
Section 5C
From the background current profile, the optimum polarizing voltage 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 percent of the total profile with
the % detector or at 92 percent with the 63Ni detector.
A more reliable method, especially for older, partially dirty detectors,
is to run a polarizing voltage/response curve as described in Section
4,A,(3) of the EPA PAM. Selection of the proper polarizing voltage is
very important so as to (1) produce maximum peak height (response)
with minimum electrical overshoot on the backside of the peak (Figure
5-S,A), and (2) ensure maximum possible efficiency and peak resolution.
The polarizing voltage must be adjusted to accommodate a slowly deteri-
orating 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, a reprint is presented
on the following two pages, courtesy of Applied Science Laboratories,
State College, PA.
Performance curves for pulsed operation are obtained by increasing the
applied pulse voltage while maintaining a constant pulse rate and
width. Rather than the typical stepped curves for DC operation (Figure
5-E), the detector response (recorder deflection) rises smoothly to a
maximum at a relatively low pulse voltage (e.g., 25 V) and then con-
tinues at a nearly constant response plateau with higher voltages. An
operating voltage is chosen well above that necessary to just give
maximum response (i.e., 50-55 V) in order to allow a considerable margin
for response reduction as the detector is used with contaminating sample
extracts. To evaluate the progressive contamination of a detector in
pulsed-mode operation, the decrease in the peak response level for a
given amount of a standard pesticide at a certain pulse voltage is
monitored and compared to that obtained when the detector was first
installed.
120
-------
Section 5C
Chromatographer, Beware of Thy Detector!
We all know that the performance of the same types of GC
columns can vary with the quality of the packings and the
columns themselves. Figures 1 and 2 are examples of extreme
differences in column performance when used for pesticide
analysis. Both columns are 6 ft x 4 mm 10 glass U-tubes
packed with 10 wt % DC 200 on a'silane treated support. Both
njns were made at the same operating conditions. Resolution
in Figure 1 is good, but the resolution in Figure 2 is far
superior because of the unbelievable 13.000 theoretical plates
obtained. Yet, the separation factors are the same in both
cases (e.g., 1.12 for endnn/dieldrin).
0
1 1
\
3
I i
a uj
n ~ o
e
LjL
\\
n I 2
"J i.i s
O e c 3 n
0 Ul 0 w
d ? | 2
! it n 3 <•
1 1 °' 0 n
« S b
I'bUlA-LJU
3 6 9 12 15 18 21
TIME (Minutes)
Figur* 1. Chromatogram of standard chlormattd pasticida mixture.
Column: S ft x 4 mm IO glan packed with 10% OC 200 on a
ulana-traatad support. Column tamparatura: 200°C. Oataetor: Elact/on
capture at 1 x 10 a AFS.
o*
5
I !
1
o
: ^
c %
3 r
5 ?
4
^
1
r*
LJ
X
M •»
s" 1 :
-S ° "
2 a TI
< UJ *
OH Q
u.
c a
^ -J C X —
I *
c
!
1 5
ui O
" •_
c ^
•x a
* ?
Uv — |
3 S S 12 15 IS 21
TIME IMinutail
2. Chromatogram of standard chiormatad paiticida mixtura.
Column: 8 ft x 4 mm IO glasi ujchM wi'rt 10% OC 200 on a
tilana-traataa MIPPCM Column tvmparatura. 200°C. Oataetor: Elactron
ea»tunat t x 10,d AFS.
You say that you would like a 6 ft column with the
efficiency shown in Figurs 2? Well, we can make such columns
and we can obtain results like those in Figure 2 anytime we
want to - providing we use an EC detector. In fact, we make
these columns all the time, but do not obtain the Figure 2
results - because we do not want to. You ask why? A very
good question.
Let's compare Figures 1 and 2 more closely. There is a
noticeable difference in absolute detector response between
the two results and also in relative responses among the peaks.
In Figure 1, the endnn/dieldrin peak height ratio is good
(0.611, with signs of slight endrin decomposition, which is
normal. However, in Figure 2 the endnn/dieldrin ratio is only
0.31, indicating appreciably greater endrm decomposition, and
yet there are no signs of it m the Chromatogram. Something
appears to be radically wrong with the results in Figure 2.
Well, something is wrong. The column used m Figure 2 does
not produce 13.000 theoretical plates. Actually, it is the exact
same column which gave the Figure 1 results. Everything was
the same for the two runs except for one thing - the EC
detector voltages were different. A voltage of 10 volts
produced Figure 1, while a voltage of 30 volts produced Figure
2. The detector is an old Barber-Colman tritium EC detector
with variable OC voltage; i.e., any voltage can oe applied. A
plot of background current (baseline) vs. detector voltage is
shown in Figure 3. The problem is non-linear detector
response. EC detector response is linear or approaches linearity
only over a small range of voltages. This voltage range usually
lies below the knee leading to the plateau m the current vs.
voltage curve (see Figure 3). In our examples, this voltage
range is approximately 10 to 15 volts.
30-
20-
O
cc
o
O 10
10 20
VOLTAGE (Volts!
30
40
Figura 3. Plot of currant n. voltag* lor in EC datactor.
At higner voltages, the response Decomes non-linear and the
response-to-concentration slope increases with increasing con-
centration. This non linearity becomes extreme on the plateau
of the current vs. voltage curve. Here, the resoonse to
concentration slope is very small at low concentrations and
increases rapidly at high concentrations. This results in an
extreme contrdction of the lower part of a GC peak and an
extension of the upper part of the peak. When this occurs a
121
-------
Section 5C
cfwomatogiam like the one in Figure 2 is produced. Figure 1
can be converted 10 an approximation of Figure 2, at shown in
Figure 4. A superficial uaselme has been drawn which cult out
th») bottom part of the peaks The similarity between Figuie* 2
and 4 is obvious If we extended ihe upper part of the peaks in
Figure 4, the chromatogram would resemble that in Figure 2
Will moct closely. This may appear exireme. but notice that in
Figure 2 we have lost nol only the endrm decomposition
product, but also *U the small impurity peaks that are seen in
Figure 1.
Figure 4. Sam* chromatogram M in Figure 1.
At voltage* below the optimum range, the reverse occun.
The response to concentration slop* 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 d wider peak, giving a low
theoretical plate calculation, and the peak maximum will tend
to be rounded Also, small 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 shirt 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 ionieation detectors, where applied voltage also affects
linearity Years ago we found we were consistently obtaining
about 150 more theoretical plates per foot from argon
lonization detectors at voltages above the optimum than from
flame tonization 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!"
tO 20
VOLTAGE IVoln)
Figure 5. Plot of peak height r«no of
iMpUchlor to dieldrm Ih/dl >i. EC de-
Mclor voltage.
°-8 1
5 0.4 H
0.2-1
o-f-
10 20
VOLTAGE IVoln)
30
Figure 6. Plot of peak height ratios of
endrm to dieldrm (e/dl and p.p'-UOT to
dieldrin (O/d) vi EC detector voltage.
I 006 -I
004 -
O
0.02 -
10 20
VOLTAGE (Volti)
30
Figure 7. Plot of paak height ratio of
andrin decompaction product to dialdnn
(E/d) vt. EC datactor voitaga.
14 -
13-
12-
| 11-
9 -
UI '
t-
<
J
S 7-l
o
ui
I 6-
4 -
10 20
VOLTAGE (Volt.)
30
Figure 8. Plot of calculated theoretical
plate* for dieldrm VL EC detector volt
122
-------
Section 5C
c. Detector Contamination
Contamination of the detector by deposition of a coating of low vapor
pressure materials on the electrodes seriously affects detector perfor-
mance by consuming a portion of the detector capability, leading to
loss of sensitivity, trailing peaks, or erratic chromatographic base-
lines. Contaminant sources may be a bleeding column, contaminated
carrier gas, bleeding septum, dirty sample inlet, unclean samples,
dirty carrier gas flow controller, or dirty tubing. Many of these
factors are treated in detail elsewhere in this Manual, but a review
is presented below.
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 (12, 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, and QF-1 are such phases and should be avoided where possible.
The high temperature 0V silicones with low (1-5 percent) phase loadings
produce very favorable columns. The background current determination is
particularly important with a new column in the instrument because back-
ground 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
prevent 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
123
-------
Section 5D
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 de-
tector for pesticide analysis (22) and a review of its theory and
characteristics (10, 23) have been published.
5D MICROCOULOMETRIC DETECTOR
The original detector for organochlorine pesticides was the micro-
coulometric detector, but it has now been largely replaced by the
electron capture detector and/or the electrolytic conductivity detector
because of their greater sensitivity and easier operation and mainten-
ance. Whereas the EC detector responds to all electron-capturing
species, the MC detector can be operated to be specific for halogenated
pesticides (or for pesticides containing S, P, or N). The latest modi-
fied MC detector has detection limits for halogen, sulfur, or nitrogen
of 1-3 ng. The sensitivity for pesticides depends upon the percentage
of these elements in the pesticide. With this relatively low absolute
sensitivity, it is necessary to concentrate sample extracts to a greater
extent and/or to inject larger sample volumes compared with the electron
capture detector. Injection volumes of 50-100 pi are common with the
MC system, and this volume may represent 10-25 grains of original sample.
Cleanup procedures must accommodate the larger samples, but because of
the increased detector specificity, cleanup often does not have to be
as rigorous as for EC detection.
The MC detector, pictured in Figure 5-F on the following page, operates
in the following manner. The pesticide is pyrolyzed in a furnace and
the reaction product automatically titrated with an electrochemically
generated reagent in the coulometric cell. The amount of ions generated
can be related theoretically by Coulomb's Law to the absolute amounts of
pesticides passing through the GC column, and peak areas can therefore
be directly converted to a quantitative value by use of an equationj
without need for a calibration curve.
124
-------
Section 5D
Figure 5-F. Dohrjnann-^Enviroteck Microcoulometric Detector
The various modes of detector operation are as follows:
Pesticide
Class
Products of
Furnace Reaction
Cell Reaction
Generation Reaction
Haloganated
(Cl~, Br~,
or I )
HX (oxidative, reduc- Ag+ + X~ —*• AgX
tive, or catalytic) (T-200-S cell)
Sulfur S02 (oxidative)
I3 + S02 —*• S03+
Ag° —+• Ag+
31 —»• I3 + 2e"
Nitrogen NH~ (catalytic)
(T-300-P cell)
NH3 + HT —*• N
(T-400-H cell)
H2 —»• 2H+ + 2e"
The coulometer and furnace systems require careful optimization to obtain
accurate and precise results. Selectivity depends on the specificity
125
-------
Section 5D
of the preliminary and coulometric reactions. For example, any com-
bustion products in addition to SO? (such as HX or nitrogen oxides)
which can reduce iodine titrant will give a response in the sulfur
mode with the iodine cell. Large amounts of elemental sulfur, not
completely converted to S02, can enter the halogen mode cell and react
with silver ions. Selectivity compared to hydrocarbons is good in all
modes when reaction conditions are carefully chosen. Selectivity may
be improved by using appropriate subtraction (absorption) tubes.
Titration cells have two operational positions, termed position I and
position II. These differ in the microcoulometer settings, position of
the titration cell electrodes, and the rate of stirring the electrolyte.
The chemical reaction and electrical responses within the cell and
microcoulometer are basically the same in these two positions; the
rate of reaction of the generated ion at the surface of the sensor
electrode is greater in position II. Position I is operationally simple
and provides dependable stoichiometric (coulometric) response but is
relatively insensitive Cca. 100 ng). Position II is ca. 20 times more
sensitive (less than 10 ng), but response is not stoichiometric (calibra-
tion is needed) and more careful operation is required. Different
models of the cell and coulometer and individual detectors of the same
model will vary in sensitivity.
The MC detector operates over a wide temperature range, and its sensi-
tivity is not affected by carrier gas flow rate, column bleed, or
temperature changes. The detector is, therefore, satisfactory for use
with temperature programming under special conditions (24). This is
not generally recommended In pesticide analysis, however, because of
poor precision of retention times from run to run. Peaks produced by
the MC detector tend to be nonsymmetrical and tailing, leading generally
to response that is not strictly linear in position II. In position I,
constant recovery, which is the critical parameter rather than con-
ventional linearity because of the application of Faraday's Law, is
achieved above 500 ng of chlorinated pesticides.
In the period between 1960 and 1966, the MC detector was extensively
used for pesticides containing Cl and S atoms, but since 1966 detection
of nitrogen (e.g., carbofuran (25)) and chlorinated pesticides has pre-
dominated. The major use of the detector is for confirmation of halo-
genated pesticides in position II with the T-200-S cell. It is also
advantageous for multicomponent chemicals such as PCBs because the
chromatogram, unlike the response of the EC detector, will consist of
peaks whose sizes are proportional to the amount of halogen present.
The analysis of environmental residues of such compounds, the amounts
of whose various components may change drastically from the original
126
-------
Section 5E
material being used for reference, is therefore facilitated. Use of the
detector for phosphorus compounds has rarely been reported. The de-
tector requires a relatively large amount of day-to-day maintenance
and cleaning In order to keep it operating with sensitivity, selectivity,
and stable baselines. Like the Coulson conductivity- detector,
manual venting of the solvent introduced during Injection is required
unless special automatic venting equipment (26) is installed. U.S.F.D.A.
has published details of design and hints for successful MC detection of
Cl and S C27).
5E ALKALI FLAME IONIZATION (THERMIONIC) DETECTOR
This detector was discovered about simultaneously by Giuffrida and
Karmen in 1964. These workers showed that when the cathode of a flame
ionization detector (FID) contained different alkali metal salts, the re-
sponse for halogen- and especially phosphorus-containing pesticides was
enhanced. Since then, improved response for compounds containing S, N,
As, and B has also been noted. At least eight theories have been
advanced for the phenomenon, but none completely explains the observed
responses. Since response of the FID is directly related to thermally
excited ions, the term "thermionic" has been used for the detector.
The original Giuffrida design consisted of a standard FID, the cathode
of which was replaced with a Pt-Ir coiled wire coated with Na2S04. An
early commercial version had a salt tip made from a pressed pellet of
CsBr added to the regular FID. The Cs tip enhances stability and
working life compared to ^2804. Other salts such as KC1, KBr, Rb2S04,
and K2S04 are also used. Response values for various compounds depend
upon carrier and fuel gas flow rates, electrode geometry, distance from
the flame, potential and polarization of the electrodes, carrier gas
flow configuration, and the nature and purity of the salt. The type of
metal and the hydrogen flow rate (flame temperature) seem especially
important. Sensitivity and selectivity to particular elements is
achieved by optimizing the above parameters. For example, phosphorus
selectivity depends very sensitively on the fuel gas ratio, and altera-
tion of geometry permits selective detection of compounds containing
halogen and nitrogen at the expense of those containing phosphorus.
Optimization can be time consuming, and frequent replacement of the salt
tip is required if baseline stability, sensitivity, and quantitative
response are to be maintained. Calibration should be performed with
each analysis for reliable results. Large solvent injections disturb
performance and should be avoided, and aromatic or halogenated solvents
and acetonitrile are never used.
127
-------
Section 5E
The greatest selectivity seems to be achievable for phosphorus, which
can be detected about 1000-fold more sensitively than compounds con-
taining N, S, or halogen. As an example, the KC1 detector has an
enhanced response for P of 10,000 or better Ccompared to the con^
ventional FID), while the increase for N is 100 and As 30 (28). Other
sources list selectivity ratios as high as 75,000/1 and 35,000/1 vs C
for P and N, respectively (6), Selectivity for P with respect to hydro-
carbons is usually reported as being poorer than that of the flame
photometric detector, but detection limits are commonly stated as being
potentially at least one order of magnitude better than for the FPD.
Discrimination among N, S, and halogen compounds is generally quite poor.
Although the alkali metals Na, K, Rb, and Cs all increase selectivity
for P markedly, the greatest sensitivity is produced by use of a
potassium salt. For detection of nitrogen, the best combination of
sensitivity and stability is obtained when a rubidium salt is used.
Halogen response can be depressed by KC1 and KBr, but sodium sulfate,
carbonate, or nitrate increase response to Cl and Br compounds. Re-
sponse to S and halogen compounds can be negative (inverted) but still
related to the amount of compound present. The negative response has
been used to detect chlorinated pesticides in soil at levels between
0.01 and 10 ppm without extensive cleanup of an exhaustive hexane
extract (29). The amount of hydrogen in the flame must be optimized for
each particular salt.
Sensitivities for organophosphorus pesticides are generally in the low
ng range (e.g., 0.2-2 ng parathion for 1/2 fsd (28)), compounds with
short retention times and narrow peaks being more sensitively detected.
Reports of the detector's linear range have varied from 3-5 decades
(13) to 0.5 to 20 ng (28). Minimum detection of triazine herbicides
(N mode, Itt^SO^ tip) was reported as 0.2-0.5 ng with a 920 selectivity
vs carbon compounds but a poor selectivity vs P (30). A 9:1 RbCl:RbBr
coil gave about 1/2 scale deflection for 10 ng of various triazines
(28). A detector with a KCl-Rb2S04 (1:1) tip was used to determine
ca. 30 ng of carbamate herbicides (31). Triazines, which contain a
greater percentage of nitrogen per molecule than carbamates, would be
expected to show a proportionately greater response.
In order to maintain the highest possible sensitivity of detection for
N or P compounds and to avoid anomalies, such as spurious peaks, de-
tector contamination must be minimized. The use of cyano-substituted
silicone liquid phases (such as OV-225, OV-275, XE-60, and the SP-2300
series), H3P04~treated supports and glass wool, phosphate-containing
detergents, and Snoop leak detector should be avoided. Rapid deterio-
ration of sensitivity will result from the loss of the alkali salt, most
commonly caused by overheating the collector due to a major disturbance
in gas flows (especially carrier and air). Care should be taken to turn
128
-------
Section 5E
off the collector voltage while changing the septum, columns, gas
cylinders, etc. Cleaning with Freon Cas ts often done with the standard
FID) will adversely affect the alkali salt of the thermionic detector
(32),
A Varian Aerograph Rb2SC>4 detector modified to allow increased carrier
gas- flow -up to 60 ml/minute and injection volumes- up to 7 ul has been
described in detail C33), With this detector, greater than 1/2 fsd
was achieved upon injection of 0,2^-0,5 ng of sin common organophosphorus
insecticides. Details for preparation and operation of a KC1 thermionic
detector providing 1/2 fsd for two ng of parathion have also been
given (28), The stability and sensitivity of a RbCl nitrogen thermionic
detector were carefully compared with those of a conventional FID for
the pesticide fenitrothion (34), A Pye 3-electrode HbCl thermionic
detector has been tuned to give higher detection sensitivity for
sulfur Cpositive detector response) compared to nitrogen but less than
that for phosphorus. The relationship between hydrogen flow rate and
optimum detection of the three elements was shown (35). The sulfur re-
sponse of this detector, optimized in terms of electrode assembly position
and flame conditions, was found to be more linear and as sensitive as
the S-mode of the FPD, but selectivity against P was not as good (36),
A KC1 detector with changeable geometric arrangement of the source and
the polarization electrode to vary the selectivity factor between P and
N was described (37).
A new design of the thermionic detector provides increased stability by
replacing the normal FID ion-collection assembly with a collector that
has a ceramic cylinder coated with an alkali salt. The cylinder,
centered about 1,25 cm above the flame jet, is heated electrically to
a dull red color while a negative polarizing voltage is applied to the
species collector. The ion current generated by the thermionic emission
in the presence of P and N is measured with a conventional FID electro-
meter. The flame in this detector is not ignited. It is adjusted to
give a low temperature plasma used for production and dissociation of
ions from the organic compounds but not to heat the collector. The
linear dynamic range is >10^ for both N and P compounds, selectivity
>lCn grams C/gram N or P, and minimum detectable level <1 x 10~13
gram N/second and <5 x lO"^ gram P/second with H2~He (8:92) support gas,
The detector has been tested with folpet, dyrene, and atrazine (38),
An earlier modified design of the thermionic detector utilizes a small,
nonvolatile rubidium silicate glass bead between the flame and collector
electrode. The bead is heated electrically and is, therefore, inde^-
pendent of the flame as a source of thermal energy, This detector,
operable in phosphorus or nitrogen plus phosphorus modes, was preliminarily
tested using standard malathion (39) and further studied (40) for optimi^
zation of the nitrogen response. The main advantage of this model is
129
-------
Section 5F
the ability to partially compensate for aging of alkali volatiliza-
tion by continuously increasing the heating current. The latest
version of this detector (41) has rapid, automatic changeover among
three modes (normal FID, P, P + N), performed by operating push-
botton electrical switches and pneumatic valves on a control unit.
Neither of these detectors has been evaluated for routine pesticide
analyses. A still earlier flameless detector for P pesticides,
termed a "chemi-ionization" detector, utilized an electrically
heated CsBr atmosphere. It was found less sensitive than the flame
thermionic detector but quite selective (42).
In summary, compared to the FPD and electrolytic conductivity de-
tector for detecting P- and N-containing pesticides, the flame
thermionic detector is generally the cheapest, most sensitive, least
selective, but least rugged (12). Its optimum, reproducible use is
more of an art than science, and it is recommended mainly for
experienced gas chromatographers. Reviews of the thermionic de-
tector have been published (43, 44).
5F FLAME PHOTOMETRIC DETECTOR (FPD)
This detector operates by monitoring HPO and 82 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"' - 10~H grams.
Figure 5-G 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-H. Figure 5-1 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 chromato-
graph,. At least three other companies produce FPDs that differ in a
number of construction aspects and performance.
130
-------
Section 5F
Figure 5-G. Tracor flame photometric detector.
Figure 5-H. Cross section of a flame photometric detector
Photomultiplier tube
Burner
SwogelocK
fitting
"^
Column effluent
131
-------
Section 5F
Figure 5-1. Dual flame photometric detector
The original operating conditions for the FPD were as follows:
oxygen (+ air) is mixed 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 which shields the flame
from view by the PM tube. Emission occurs above the flame tip, and
the light is transmitted to the PM tube through a filter which trans-
mits a specific wavelength of the element to be monitored. A potential
is applied to the PM tube and its output is amplified by the electro-
meter and read-out on a recorder.
When the FPD is operated in this manner, 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, is recommended for this purpose. Reversal of the
hydrogen and air/02 inlets has now been shown to give a "hyperventilated"
flame (45) that allows injection of up to 50 ul 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.
Hydrogen is introduced through the air/02 inlet and air (only) into
the port marked for hydrogen; the use of oxygen is not recommended with
this new plumbing arrangement, and typical gas flows are hydrogen
50-60 ml/minute and air 8n 100 ml/minute (Tracor Instruments FPD
Operation Manual). A new design dual-channel FPD with a separate com-
bustion chamber, in which the sample is mixed with hydrogen before
introduction to the detector, also resists flame-out with up to 50 pi
of liquid sample (46).
132
-------
Section 5F
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. A small addition of S(>2 to the hydrogen of the FPD
flame caused an 8-fold increase in the sensitivity of the S-mode to
aldicarb and malathion (47).
The degree of sensitivity of the FPD relates to 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 detector interior. Each of these
factors should be checked and optimized 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
pesticide (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. Typical operating parameters
for FPD in the P-mode are in the table on page 29, (These values and
most of the information in this section are for the Tracer FPD, used
in most EPA laboratories, operated as originally designed. See the
earlier paragraph on operating conditions when support gases are re-
versed and reference (48) for optimization of gas flow rates for the
Pye FPD of different design.)
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 usually produces the most favorable signal to noise ratio.
133
-------
Section 5F
Temperatures (°C)
Flow Rates (ml/min.)
Co lumn
Injection Block
Detector**
Transfer Line
Switching Valve*
200
210-225
165-250
235
235-240
Purge*
Carrier
Hydrogen
Oxygen
Air
70-80
70-80
150-200
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 80 ml/minute or
more may be required, depending on the detector
Although the FPD Is not as sensitive to gas leaks as the electron
capture detector, the flow system of the chromatograph should be
tight. Leaks in the hydrogen, oxygen, or air supply can be
hazardous from an explosion standpoint.
Optimum response voltage for the PM tube is determined using a variable
power supply which allows the voltage to be increased with little
Increase in electronic noise. Raising the voltage from the electro-
meter will increase electronic noise inordinately. With optimum
flow rates, the power supply is set at 750 V and a sample of diazinon
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 re-
quire 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 of at
least 1 x 10~6 amps. PM tubes are heat sensitive and should be well
insulated from sources of heat.
134
-------
Section 5F
Interference filters may be changed at any time, 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 chromatograph will increase the noise level
and decrease sensitivity. 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. The pyrex or quartz glass
windows in the optical system can become etched through the action
of alkali vapor and must be replaced or polished.
The FPD response to P is linear over a concentration range of about
3-5 decades, e.g., 0.4-400 ng for parathion in the P-mode. Nonlinear
response of the FPD (526 nm filter) to oxygen analogs of OP pesticides
is often noted and is thought to be caused by degradation of these
P ±rrO 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 ng 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 (49). An 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 ys concentration are
linear within + 5 percent. This linear response facilitates easy
interpretation and allows electronic integration and data acquisition
not possible without 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 (50) .
In order to operate in the dual mode, it is necessary to optimize
combustion 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 detector 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 response.
The proper attenuation for a given sample will depend upon the sensi-
tivity achieved, but, in general, it is best to operate at the minimum
detection level and to dilute the sample as necessary. Selectivities
135
-------
Section 5F
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 selectivity of the FPD have been
studied (51).
Maximum utility of the FPD is afforded by the dual photomultiplier
arrangement (Figure 5-1) 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 confirmation of residue identity. The response ratio
(xlO^) ranges from 5.0-5.8 for PS compounds, 2.5-3.4 for PS2 compounds,
and 1.6-2.4 for PS^ compounds (52). 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 con-
ditions.
Errors have been noted (53) 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 (54).
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 appears to have advantages over the normal flame thermionic
detector for routine analysis in terms of ease of operation, better
stability, less maintenance time, independence of response to gas
flow rates, and need for less frequent injection of standards.
Sensitivity of the FPD may be about one order of magnitude less for
P compounds than with a fully optimized thermionic detector. Applica-
tions and limitations of the FPD in atmospheric analysis have been
reviewed (5).
Varian has developed a new FPD with dual flame design that is re-
portedly (55) superior, but it has not yet been carefully evaluated for
136
-------
Section 5G
routine residue analysis. Two hydrogen-air flames are used to
separate the regions of sample decomposition and emission, so that
the (second) emission flame is more efficient, and sensitivity is
improved compared to the single-flame FPD. Besides reduced peak
quenching, interference from hydrocarbon solvents is reduced because
much of the C-H emission takes place in the first flame, which is not
viewed by the photomultiplier. Reported selectivities are 10^ grams
C/gram P and 1(H 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 pi of solvent can be injected
without extinguishing the flame, and a pushbutton linearizer for the
exact quadratic response of the S-mode is included.
5G ELECTROLYTIC CONDUCTIVITY DETECTOR
a. Original Coulson Detector
This detector operates by combustion of pesticides eluting from the
GC column to form simple molecular species that readily ionize and
contribute to the conductivity of water deionized by passage through
a mixed-bed ion-exchange resin column. Changes in conductivity are
monitored with a DC bridge circuit and recorded on a strip chart
recorder. The detector is pictured in Figure 5-J. Components include
the pyrolyzer assembly with a separately heated vent valve, pyrolyzer
furnace, heater and gas controls; the glass conductivity cell including
the gas-liquid contactor; the water circulation cell consisting of
water reservoir, water pump, and ion-exchange bed; and the conductivity
bridge assembly. The conductivity cell is shown in Figure 5-K.
137
-------
Section 5G
Figure 5-J. Tracer (Coulson) electro-
lytic conductivity detector
Figure 5-K, Electrolytic
conductivity
cell
Gas
Water-
Electrodes
Table 5-2 shows the various modes of operation of the conductivity
detector as described by Cochrane. 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.
138
-------
TABLE 5-2
CONDUCTIVITY DETECTION MODES OF OPERATION
Mode
Conditions
Selectivity
Sensitivity
Cl
Pyrolytic
H2 reduction gas - 850°C -
Ni wire catalyst - NH3 formed
and detected as NH^— He or
Ar carrier gas
empty tube reduction using
H2 gas - HC1 produced - 600
or 850°C - N2 carrier gas
S—compounds oxidized to S02
and 863 in 02 stream in
empty tube - Cl converted to
HC1 - N2 carrier gas - Pt
gauze catalyst may be used -
Furnace temperature
850°C
empty tube pyrolysis
400-600° or 850°C
Sr(OH)2 scrubber
removes acidic
products — responds
only to N compounds
responds to Cl and
N-contg. OGP insect.
such as diazinon
responds to S, Cl,
and N «- Ag wire
scrubber increases
selectivity with
respect to Cl, but
at expense of sensi-
tivity and reproduci—
bility - CaO scrubber
for St>2 has also been
used
responds to N, Cl, S
100 pg N; 100-200 ng subst. ureas,
7-15 ng triazines, 50-80 ng car-
bamates, 35-50 ng thiol-carbamates;
150 ng parathion, 25 ng diazinon.
500 pg Cl; 6-40 ng Cl insecticides.
1 ng S; 30-2200 ng N compounds gave
1/2 fsd, depending on Qy flow.
500 pg S and/or Cl, 1 ng N; linear
range 2-1000 ng S; 6-100 ng S and
Cl pesticides.
From: Cochrane, W. P., presented at May, 1973 Symposium on Pollutant Analysis, Athens, Ga.: See also
J. Chromatogr., 75, 207 (1973); Int. J. Environ. Anal. Chem.. _3, 199 (1974); and (59).
OT
o>
o
o
a
o
-------
Section 5G
The best selectivity and sensitivity have generally been obtained for
N compounds, and detection of these has been the major use of the de-
tector up to now. A comprehensive study of 95 organonitrogen pesticides
with 5 and 10 percent DC-200 columns gave a range of 1-1000 ng for 1/2 fsd
(56). Sensitivities for several herbicides determined in foods (terbacil,
propanil, linuron, DCNA, benzoprop-ethyl) ranged from 7-100 ng (1/2 fsd),
and there were fewer extraneous peaks than with electron capture GC or
HPLC (UV) (57). If the Sr(OH)2 pledget and the catalyst are removed,
the N-mode will detect organochlorine pesticides. Sensitivity for N
compounds is better than with the microcoulometric detector, and similar
to that of the rubidium thermionic detector. Selectivity for N compounds
is reported to be superior to the thermionic detector (58). The S-mode
of the conductivity detector is of comparable or somewhat better (59)
sensitivity but less selective than the S-mode of the FPD. The con-
ductivity detector is less complex than the MC detector and requires
less maintenance, but reports differ on which requires cleaner extracts.
The conductivity detector is much more selective but generally one or
two orders of magnitude less sensitive than the electron capture de-
tector, and analyses can be performed without cleanup using the former,
whereas cleanup would be required for the latter (e.g., for determination
of triazine herbicides (60, 61)). Heptafluorobutyryl and other perfluoro
derivatives of pesticides were detected with a sensitivity of 1-5 ng in
the halogen-mode, compared to 40-400 pg for EC detection. However, 1-2 ppb
of residues in foods were determined by either detector because the
greater selectivity of the conductivity detector permitted a greater
quantity of sample to be injected (62).
Variables affecting the detection of nitrogen pesticides have been studied
(63). Response was not affected by changes in gas flows from 10-100
ml/minute, while water flow rate through the conductivity cell of 1-2
ml/minute gave maximum response. Skewed chromatographic peaks and re-
duced sensitivity resulted from adsorption of NHg on contaminated quartz
pyrolysis tubes or the pyrex glass detector assembly and even some
brands of new quartz tubes. Cooling the deionized water through the
detector increased sensitivity by reducing its background conductivity
and background electronic noise. (A simple refrigeration unit has been
designed to prevent temperature increases in the circulating deionized
water and thereby increase uniformity of response throughout a day's
operation of the detector (64)). Besides the GC and detector operating
parameters, sensitivity for various compounds depended upon the number
of nitrogen atoms and their molecular arrangement. As a specific example,
the minimum response for methomyl oxime was 3 ng with linearity to
ca. 38 ng with the following detector operating parameters:
140
-------
Section 5G
Conductivity bridge
Pyrolysis furnace temperature
Vent block temperature
Gas flows
Voltage 30
- Attenuation 1
800°C
220°C
- helium 80 ml/min.
- hydrogen 80 ml/min.
A study of the effect of operating parameters on the response of
triazine herbicides (65) indicated maximum response with equal water
flow rates in the mixing chamber and siphon arm. Tailing of peaks
increased at slow water rates, and 1.2 ml/minute produced the highest
response. Sensitivity increased with increasing furnace temperature,
due to increased combustion of the pesticide.
A general review of the electrolytic conductivity detector has been
published (66).
141
-------
Section 5G
b. Modified Conductivity Detectors
Using the reductive mode of operation, Hall (67) enhanced the selectivity
and sensitivity of the conventional conductivity detector for organo-
chlorine pesticides by modifying the reaction tube and furnace exit
and increasing the maximum cell voltage. Heptachlor was detected by
Hall at 0.1-0.4 ng levels compared to 2-5 ng for the original model, and
chlorinated insecticides were detected with a 1000:1 selectivity over
PCBs. Factors most affecting selectivity were furnace temperature
and reaction gas composition and flow rate,.while sensitivity was
affected most by adsorptive surfaces, electrode polarization, system
stability, and furnace temperature.
The commercial version of the Hall detector is shown in Figure 5-L.
This detector features miniaturized one-package design, an improved
furnace temperature controller, and an AC conductivity bridge. A low
volume combustion tube and alcohol conductivity solvent instead of
water -eliminate need for solvent venting. (A vent valve has been
described that reportedly eliminates solvent interference and increases
reproducibility (68)). Response for Cl, N, and S is linear over a range
of greater than 10^, and peak broadening is greatly minimized. Se-
lectivity v_s carbon is 10^-10" (69, 70). Chromatograms produced by
Hall (71) illustrate the detection (15-90 percent fsd) of nine
chlorinated insecticides at 0.9-6.5 ng levels with the chloride mode,
and one ng simazine and 10 ng atrazine, diazinon, ethyl parathion, and
carbofuran with the nitrogen mode (Ni catalyst, isopropanol-water solvent)
The improved sensitivity of the Hall detector is seen by comparison of
these values with Table 5-2.
142
-------
Section 5G
Figure 5-1. Tracer Model 310 Hall electrolytic conductivity detector
A further comparison of the Hall and Coulson detectors has been carried
out. An approximate 7-fold increased sensitivity was found for the Hall
detector relative to the Coulson detector for nitrogen-containing pesti-
cides. Values obtained on a 4% OV-101/6% OV-210 column at 205°C were
as follows (72):
Atrazine
Bladex
Chloropropham
Diazinon
Ramrod
Parathion
ng for 1/2 fsd
Coulson
7
15
75
25
50
150
Hall
1.1
1.1
6.0
4.5
6.5
20
143
-------
Section 5G
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 percent fsd
peaks with <1 percent fsd background level under the following typical
operating conditions: 6 feet x 1/4 inch 3% OV-1 column, 200°C, helium
carrier 50 ml/minute, hydrogen reaction gas 50 ml/minute, 50 percent
n-propanol in deionized water electrolyte flowing -at 0.8 ml/minute,
850°C furnace temperature (73). Bayer reports easy detection of
0.5 ng of lindane in the reductive mode with an attenuation of
1 x 10 (32).
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 (74).
The following are some operating characteristics and maintenace
instructions for the Hall detector as outlined by Bayer (32): 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 percent 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.
A second modified detector was recently described by Lawrence and Moore
(75) with a new, compact cell and optimized flow conditions and a water
144
-------
Section 5H
jacket for temperature control. A five-fold increase in sensitivity
was achieved over the conventional detector with a water flow of
1.2 ml/minute and temperature less than 20°C. Two modifications were
reported for this detector (76): An improved water flow control, con-
sisting of a fine stainless steel wire inserted into the capillary
tubing leading to the mixing chamber to reduce flow to the optimum
value, provided 1/2 fsd for 2 ng of atrazine. Modified bridge
circuitry allowing analyses to be carried out at 100 V increased the
peak height for 5 ng of atrazine ca. 3 times compared to operation at
30 V.
5H OTHER DETECTORS AND DETECTOR COMBINATIONS
Other detectors have been employed on occasion for pesticide residue
analysis, but none is as prominent as those covered in Subsection 5C
through 5G. Bowman, Beroza, and co-workers (77-79) positioned a
copper screen or pellet of Na2S04 or In metal at the flame tip of a
FID as in the thermionic detector. Compounds containing halogen
modify the metallic atomic line spectral emissions which are monitored
at characteristic wavelengths in a manner similar to the FPD. Detect-
ability for aldrin was about 3 orders poorer than for the EC detector
and selectivity with respect to hydrocarbons was 10,000:1 (inferior to
the EC detector) and 100:1 to S- and P-compounds (better than EC).
Response was linear with concentration on a log-log scale. Thorough
extract cleanup was required with these detectors. The indium detector
was combined with a conventional FPD to give a dual flame photometric
detector for simultaneous and selective detection of P-, S-, and
Cl-containing pesticides eluted from a GC column (80). In the lower
flame, compounds emerging from the column are pyrolyzed and monitored
for S and P. Between the two flames is a stainless steel net holding
In pellets. The Cl-containing compounds interact with In and the re-
sulting emission is monitored at 360 nm. Sensitivity for Cl is ca. 100
times less than the electron capture detector, linearity is 100 times
greater, and selectivity is greater. The GC detection of halogenated
pesticides based on the Beilstein response with Cu has been questioned
because certain halogen-free organic compounds also release volatile
Cu species in the flame and give a positive green flame (81).
The microwave plasma detector of Bache and Lisk is much like a FPD
except that the relatively low energy flame is replaced by a much higher
energy microwave plasma excitation. The plasma is established by
irradiating an inert carrier gas such as Ar or He flowing through a
tuned microwave cavity at either atmospheric or reduced pressure. All
elements of organic compounds can be detected, and more than one element
may be examined simultaneously. Detection limits for most elements are
0.1-1 ng/second (3 ng/second for N) with an adequate linear range.
Organophosphorus pesticides, carbamates and triazines, and organo-
mercurials have been determined with this detector (82).
145
-------
Section 5H
The sulfur-phosphorus emission detector (SPED) is similar to the FPD
except that fibre 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 (59) :
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 lO^-KP, and the minimum detectable amounts were 5 x 10~H
to 2.3 x 10-!3 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. 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 NH-j, PH-j, and AsHg to which the FID does
not respond. The detector (83) 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 10.2 ev, such
as C^-C^ hydrocarbons including many common pesticide solvents (methanol,
methylene chloride, carbon tetrachloride, acetonitrile), do not re-
spond. 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 250°C. Although this detector has not been evaluated
for pesticide analysis, its high sensitivity suggests that it may be
promising, especially in automated systems because of its great linear
range.
Principles of the use and details of the arrangement of multiple de-
tectors for efficient GC determination and confirmation of pesticides
of different chemical types in a single sample are discussed in Section
320 of the FDA PAM, 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. When 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 com-
patible with both. A GC system with one column, a three-way effluent
146
-------
Section 51
splitter, and five different detectors (electrolytic conductivity
(N-mode), FPD (FID, S-, and P-tnodes), and electron capture) operating
simultaneously was described (84). 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 (85).
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 electro-
meter 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 manufacturing the chromatograph.
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 percent nois-e 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 which 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.
147
-------
Section 5J
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 £,_p_'-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 percent 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 p,p'-DDT breakdown
problems. The label instructions were followed.
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 alcoholic 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. Septurns
A large number of different types of septums are available commercially
ranging from inexpensive plain black or grey silicone rubber to more
sophisticated and expensive sandwich types. Excellent results have
been obtained with the blue silicone rubber materials marketed by
Applied Science Laboratories as their "W" series.
148
-------
Section 5J
Bleed from seven types of freshly installed septums has been reported
(86) for their first 2-3 hours of use. Although this did not cause
loss of aldrin and endrin studied as model pesticides, precautions
should be taken to avoid bleed by rinsing the septums with acetone,
preconditioning in a vacuum oven for two hours at 250°C, and storing
until use in an oven at 130°C. Another option is to routinely change
the septum at the end of each day to allow overnight purging of the
system of any contaminant bleed. Still another approach is to install
a septum shield, which in effect removes the septum from the injection
system except at the very moment of injection. The paper describing
this device as a remedy for septum bleed also studied the appearance
of extraneous peaks due to the flow controller and the carrier gas as
well as design of the injection system (87). Handling of septums with
the fingers should be avoided.
Applied Science Laboratories HT blue septums are made from an especially
low bleed silicone rubber and require no extensive preconditioning
steps. The septum is rinsed with acetone, wiped with a Kimwipe tissue,
air dried, and inserted using the tissue or clean forceps. Both the
W and HT septums are indefinitely stable below 250°C and will typically
withstand at least 50 injections with a 26-gauge needle in good condition.
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 ul syringe for the electron capture, thermionic, and FPD
detectors or a larger capacity if required for other less sensitive detectors
(e.g., microcoulometric). Automatic injection devices are available for
use with some chromatographs and detectors. (See Section 50j).
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 repeat-
edly 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 ul 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.
149
-------
Section 5J
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 ul syringe,
the needle will retain ca. 0.2-0.3 pi 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 pi of
solvent is first drawn into the syringe followed by a 1-2 pi air pocket
and then 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.
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.
150
-------
Section $K
(2) Preferred Volume Range
Injection of 1-3 pi samples is not recommended because of the
large increase in error probability resulting from these small volumes.
For example, a typical absolute injection error of 0.2 pi in a 1.0 ul
injection would produce a ^g x 100 = 20% relative error, while the
'
same 0.2 ul error in a 5 ul' injection volume reduces the relative error
to a tolerable 4 percent level. The analyst is strongly urged to inject
volumes between 5 and 8 ul from a 10 ;ul 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) . A capable analyst should be able to
reproduce a series of 3-8 pi injections to within 1-5 percent of
average peak area or height when response is ca. 1/2 fsd with use of
proper techniques.
The preceding paragraph describes the conventional wisdom concerning
normal use of a common 10 pi GC syringe. Data have been presented,
however, leading to recommended volumes between 2-4 pi. Below 2 pi,
the error of injection increased above the + 4-5 percent range. Re-
producibility decreased for samples greater than 4 pi, supposedly due
to the difficulty in quantitatively 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 (87) ,
)K 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 septums which are not changed often enough or
loose column connections. When a column is installed, fresh 0-rings
should be used. Satisfactory results have been obtained with heat-
resistant silicone rubber column 0-rings (temperature limit 300°C) used
with brass ferrules. A reversed back ferrule has traditionally been
used between two 0-rings to cushion the glass from thermal shock, but
one 0-ring with no back ferrule is recommended by one commercial source
of the silicone rings. Even though the Swagelok nuts are tightened
securely on a cold column, an overnight period at normal operating
151
-------
Section 5L
temperature may result in sufficient loosening to cause a leak. When
fresh 0-rings are installed, it is good practice to open up the oven
after the first overnight heating period and retighten the nuts.
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-
graphite combination ferrules are all available in one piece design for
use without a brass ferrule or any 0-rings (also reducing ferrules for
use without reducing unions).
All are reuseable if carefully removed from old columns. Temperature
limits are 250°C for Teflon, 450°C for graphite, and 350-450°C for
Vespel. As with other Swage-type fittings, the nut is tightened until
finger tight and then a further 1/4 to 1/2 turn with a wrench until
leak tight. The need for further tightening should be checked after
the initial heating period.
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 labora-
tory. 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-Dolar 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, £,£'-DDT and p,p'-DDD, and the isomers of BHC (Figure 5-M,A) .
152
-------
Section 5L
(2) A high efficiency column is desirable if injected extracts
contain extraneous materials and detection of low pesticide concentra-
tions 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 5-3 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, ODD, 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 percent
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 /3-BHC but good separation
between dieldrin and j3,_p_'-DDE on an efficient column.
153
-------
TABLE 5-3
RECOMMENDED LIOUID PHASES FOR PESTICIDE ANALYSIS
Phase
1.5% OV-17/
1.95% OV-210
4% SE-30/ */
6% OV-210
5% OV-210
3% DECS
Solid Support
Chromosorb W, H.P,
or Gas-Chrom Q,
100/120 mesh
Chromosorb W, H.P.
or Gas-Chrom Q,
80/100 mesh
Chromosorb W, H.P.
or Gas-Chrom Q,
100/120 mesh
Gas-Chrom P
80/100 mesh
Approx. Operating
T, °C
Approx. Flow Rate,
ml/min.
200
50-70
200
70-90
200
45-60
195
70-90
*/
Individual liquid phases premixed prior to coating on silanized support
n>
o
o
3
-------
Section 5L
The single phase OV-210 gives a full separation of the common BHC
isomers, but only fair separation between the compound pairs of hepta-
chlor, epoxide/£,£'-DDE and £,£'-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 ( ft -BHC after £,£'-DDT and £,£*-DDT before £,£*-DDD) which
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 therefore not recommended as a routine working column,
but only as a special purpose identification tool.
Chromatograms of standard chlorinated pesticide mixtures on these
columns and a single phase nonpolar DC-200 column are shown in Figure
5-M. Relative retention times of over 60 chlorinated and phosphate
pesticides on OV-17/QF-1 I/, SE-30/OV-210 and OV-210 columns between
170 and 204°C are listed in Subsection 4,A(6) of the EPA PAM and data
on these columns also appeared in the literature (89, 90).
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,
and a survey of the literature indicates that these are still the most
widely used today. However, mixed phase columns which combine polar
and nonpolar liquids in varying degrees and newer single phases with
varying polarities (the 0V series) have become increasingly important
as the range of pesticide types has drastically grown. Some of the
more widely used additional phases include nonpolar SE-30, DC-200, DC-11,
and Apiezon L; intermediate polarity 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.
-/ equivalent to OV-210
155
-------
Section 5L
Figure 5-K. Peak elution patterns of 13 pesticides on five columns.
10% DC-200
v ;
% OV-210
3% PEGS
I *
|i
1 /
156
-------
Section 5L
The U.S.F.D.A. 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 numerous compounds under these
conditions are listed in the FDA 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 con-
tains 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 percent
trifluoropropyl substituted silicone, similar to QF-1 and OV-210),
HI-EFF-1BP (similar to DECS), and OV-11 (35 percent phenyl substituted
silicone, similar to 50 percent phenyl substituted silicone DC-710).
The Canadian Department of National Health and Welfare (91) specifies
6 foot x 1/4 inch columns of the following single phases coated at a
3 percent level on Chromosorb W, AW, or HP for their multiresidue
monitoring procedures: OV-1 (nonpolar), OV-17 (slightly polar),
0V—225 (medium polar), ethylene glycol adipate (polar), and DECS (very
polar). The relative polarities were calculated from McReynolds
constants (92). 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 (93) 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
organochlorine insecticides, and tables of relative retentions were
given for 14 compounds (94). A column packed with three silicone
stationary phases, namely 2.5% OV-11 + 1% QF-1 +0.5% XE-60 (phases
premixed 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 (95).
Other extensive compilations of pesticide relative retention data appear
in References (96-98).
In addition to the packed GC columns already discussed in this subsection
5L, the use of capillary columns is growing rapidly in analytical
157
-------
Section 5L
situations where high resolving power is required. 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 accomplished by an injector-splitter where
typically 1 ul is injected, 0.01 ul enters the capillary column, and
0.99 pi is vented. Carrier gas flow through such columns is ca. 1 ml/
minute, and a make-up gas system is required 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 environmental samples. The thin
liquid film thickness provides fast analysis times, often at relatively
low temperatures, and sharp peaks. Also available 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 splitting 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. See Reference (99) 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 7AC). In
addition, retention data were reported for 60 organophosphorus pesticides
and for 27 chloro-, bromo-, and nitrophenols on SE-30 capillary columns
(54). 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 (100).
158
-------
Section 5M
5M SENSITIVITY OF THE GC SYSTEM
For analysis of pesticides in environmental media, concentrations of
residues are 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 temperature
parameters must be optimized to produce a peak at least 50 percent full
scale deflection (fsd) (with minimal baseline noise) from injection of
100 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: 2.5 ng ethyl parathion for the
FPD (P Model), 25 ng diazinon and 35 ng parathion for the Coulson con-
ductivity 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 sensi-
tivity (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/ul. A 5 jul 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,
although giving chromatograms 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 percent ether-
petroleum ether Florisil column eluate from a fat sample (Subsection
7Aa). If the instrument is functioning properly, it should be possible
to have a noise level not exceeding 2 percent full scale at a low
signal attenuation (10 x 8 or 10 x 16).
159
-------
Section 5N
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 (101). 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
detection 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 pro-
cedure (102). The value depends upon the pesticide and the substrate
and is expressed in relative units such as ppm or ppb.
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^.
£,£T-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 indicates 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 chromatographic (e.g., TLC)
or non-chromatographic procedure (Section 8.).
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 detectors
such as the microcoulometer which 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 which elute before the reference,
the relative retention time will be less than 1.0; for those which
elute after the reference, the relative retention time will be greater
than 1.0. When reporting relative retention data, the absolute reten-
tion time and relevant instrumental parameters should be given.
160
-------
Section 5N
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 8, 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 (103) 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 identifica-
tion. Figure 5-N 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-210
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 plots. 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.
High column efficiency is a distinct advantage for compound identification
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 _p_,_p_'-DDT elutes in 16 to 18 minutes
with a carrier gas flow of 50-80 ml/minute. The recorder chart speed
161
-------
Figure 5-N. Plots of retention ratios, relative to aldrin, of 17 pesticidal compounds on three column pairs.
S3
10
o
o
,of 17
- . n.
•;.;;•-.
•y/T*
-L_J*.
"' Ml
ion ratk
pesticide
i j
^^
»
M.
3s,retot
al comp
ii:::-..
/
!«/
I!:-:;-:'
ive to -
ounds :
f
*f
lj
i^, _
,::.
Vriij
)
i
i
3
4
5
«
7
•
9
10
11
12
13
14
15
16
17
1 ^
M'ni:
iJ|!;H
/ii:i;:
! "Mi
i « • t • M
j ,'
i :
rrrtifi
a-iHc
Lindoiu
S-BHC
KxHachlor T
Aldrin (Refnene*)
rbpt. Epoxid.
p.p'-DDE
Dieldrin
o,p'-DDT
p,p'-DDD
p,p'-DDT
Caplan
I-Hydro«yclilanleii«
Dim*lhoate
M«lhyl Pomlhion
Elbyl
Mo loth ion
l'1'^
5% DC--200/7.5% QF-1 (200t.)
Plot of retention ratios, relative to
. aldrin, of 17 pesticidal compounds
3% DEGSCzotfc.)
Plot of retention ratios, relative to
aldrin, of 17 pesticidal compounds
U%OV-17/1.95%Qf-l(200*e.)
-------
Section 50
is set to permit adequate peak spacing and a total retention distance
such that an absolute measurement error of 1.0 mm will correspond 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 (104). Chromatograms
are reproduced with corrected baseline drift and solvent peak elimina-
tion, 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.
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 concentration is required for each unknown if injections are
made at concentration 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 concen-
tration 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 50f).
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
163
-------
Section 50
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 chromatography 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 the
analysis of pesticide residues of one or a small, definite group of
pesticides.
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 being conducted within the linear concentration range.
A recent study (87) concluded that generally unrecognized systematic
errors were inherent in the accepted procedures of both 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
A universal equation applicable for any GC analysis where an unknown
peak is calculated against a peak resulting from injection of a standard
of known concentration is given below. This equation is equally appli-
cable to external standardization procedures based on comparison of
standard and unknown liquid chromatography peaks or thin layer chroma-
tography spot sizes. The equation is:
164
-------
R =
Section 50
abe
cd
where a = nanograms of pesticide represented by the standard peak
b = height Cor area) of sample peak
c = height (or area) of standard peak
d = grams (or ml) of original sample
R = residue concentration in parts per million or billion
e = Dilution Factor derived as follows:
e = nil of extracting solvent x volume of final extract*
aliquot taken of original extract (ml) x ul injected
* This value is in ul for ppb and in ml for ppm
In calculating the Dilution Factor (e) for a situation where the final
extract concentrate contains the entire original sample (where no
initial aliquot is taken), the values for the "ml of extracting solvent"
and the "aliquot taken of original extract" would cancel out, so that
the e value would become
volume of final extract (ul or ml)
ul injected
The basic equation can be simplified when conducting routine, repetitive
analyses. For example, considering a method such as the blood procedure
outlined in Subsection 5,A,(3),(a) of the EPA PAM,
(1) d is always a constant size (2 ml)
(2) the volume of extracting solvent is always the same (6 ml)
(3) an aliquot of uniform volume is always taken (5 ml)
(4) a constant injection volume is always used (5 ul)
The equation can be simplified to match these parameters and would read
R = abf x 0.12
where f = volume of final extract in ul or ml. The numerical factor
0.12 represents e for the specific illustration above.
165
-------
Section 50
d. Reporting of Results
The method for reporting analytical results wll.l 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.
(A) 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
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 = micrograms per gram
ppb = nanograms per gram
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.
166
-------
Section 50
The following Is a summary of units frequently used in pesticide
analyses:
jug -
ng -
Pg =
ppm -
ppb =
ppt =
10""6 grams
10~9 grams
10~12 grams
parts per million
parts per billion
pg/g, pg/ml (PPt
mean
ml - 10~3 liters
Ail - 10"6 liters
= jug/g, »g/ml, ng/mg,
= jug/1, ng/g, ng/ml,
is used frequently in
"parts per thousand") .
or pg/jug
or pg/mg
other books to
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 1CP and the detector sensitivity for a certain pesticide
is 1 pg, the upper limit of analysis is 1 ng. If the detector
sensitivity 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.
Quantitation 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 de-
tector, the linear range varies somewhat between pesticides. For
example, the isomers of BHC exhibit a more restricted EC linear range
than £,£'-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. Figure 5-C compares
typical linearity curves for p_,p_'-DDD with these two detectors.
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.
167
-------
Section 50
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 5-0, distance CD). The linear range is observed by
plotting height vs. amount of pesticide injected (Figure 5-C).
FIGURE 5-O QUANTITATION BY PEAK HEIGHT
C METHOD
Peak Height ;
f. Sensitivity Control
When analyses are performed in the ppb or even ppt (parts per trillion)
range, electrometer attenuation is required to attain maximum sensi-
tivity consistent with an acceptable baseline noise level. Electro-
meter 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 atten-
uation been set for high sensitivity, chromatographic peaks would have
gone off-scale, requiring serial dilutions of final extract to a pesti-
cide concentration within linear boundaries.
168
-------
Section 50
Electron capture electrometer attenuation should be adjusted to obtain
a minimum sensitivity level equivalent to a 50% fsd peak from the
injection of 100 pg aldrin.
There is no objection to using different instrumental attenuation
settings for standards and samples provided that concentrations are
within the linear detection 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
ul 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/ul for electron capture GC might be:
Lindane
Aldrin
Dieldrin
£,2T-DDT
£,£/-DDT
Mixture A^
5
5
10
10
20
Mixture ^
10
10
20
20
40
Mixture A.,
20
20
40
40
80
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 stan-
dard mixture(s) if pesticides are present in the sample at widely
different levels.
169
-------
Section 50
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 percent for highest accuracy and at
most by 25 percent. If all GC modules are operating properly and
parameters are optimally set, the 10 or 20 percent 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.
The requirement of referencing samples against standards differing
by no more than 25 percent in peak height causes no inconvenience
when the concept of different standard concentrations (Subsection 50f)
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 arising from instrumental sources and/or
from small injection errors. Figure 5-P 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 percent. On the other
hand, when the extract is further concentrated down to an assumed final
volume of 400 jul, 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 percent, a very acceptable value.
170
-------
Section 50
Figure 5-P. 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 Ail.
Standard
0.2 na
Final
ract
lunw
4000 U< 400—-
Samp. ln|«c».2i d".b. a
Samp. ln|o<>. It P.p.h. - 0.2 « 13x4000
Samp. ln|«ct. 1 i p.pi. .,0.2X130X400 ft Q6 _ 9 M
fa! x0-6 ""'*"
Deviation 2.20
or 23% *rror
Samp. ln|.o. J , P.O.B..
X 0.6 =9.81
Deviation 0.21
or 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.
The proper procedure for quantitation 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.
171
-------
Section 50
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 measurement of very small peaks or tall, symmetrical, fairly
narrow peaks ( < 10 mm at the base) which 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. Triangulation is used for
separated, unsymmetrical peaks or peaks on a sloping baseline.
Triangulation should never be used on very narrow peaks. Extreme
care must be taken in the construction of inflectional tangents in
all measurements.
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, re-
sulting 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 5-0, 5-Q, and 5-R illustrate the peak height, peak height x
width at half height, and triangulation measurement methods, respectively.
The two right-hand peaks in Figure 5-0 are measurable by the peak height
method because their overlap does not abscure 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-R. 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.
172
-------
FIGURE 5-Q QUANTITATION BY PEAK AREA METHOD
Section 50
FIGURE 5-R QUANTITATION BY TRIANGUIATION
METHOD
Quantitation of peaks indicating heavy electrical overshoot (Figure
5-S,A) or nonlinear response (5-S,B) will lead to unreliable quantita-
tion. Peak overshoot is influenced by foil contamination and by
improper EC detector polarizing voltage (Subsection 5Cb).
FIGURE 5-S EXAMPLES OF GAS CHROMATOGRAPHIC
PEAKS
B.
r~\
Distorted
Peaks
Non-linear
Indication
Linear
Indication
173
-------
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 integrator, 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 (105) compared five methods for quantitation of aldrin, hepta-
chlor 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 chromatograms 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. Automation
Digital computer systems are available today that perform peak and
baseline detection, area integration, baseline correction, area
allocation of fused peaks, and postrun calculations. These relatively
expensive systems provide the fastest and probably the most accurate
means of quantitation. Their increased use in the future is expected
as technology is simplified and prices decrease, especially in
laboratories with many chromatographs and a heavy sample load. How-
ever, small laboratories may find complete automation impractical
and inadvisable for complex biological samples. The various approaches
for automation of sample introduction in gas chromatography have been
reviewed and a typical autosampler described in detail (106). 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 percent of aldrin and dieldrin in a commercially automated
dry capsule injector were reduced by silylation of the capsules (107).
174
-------
Section 5P
P References
(1) Welsh, P., Amer. Lab.. 9_ (5), 36 (1977).
(2) Natusch, D. F. S., and Thorpe, T-. M., Anal. Chem.. 4.5, 1184A (1973).
(3) Sevclk, T., Detectors in Gas Chromatography, Elsevier Publishing
Co., N.Y., 1976, 192 pp.
(4) Pigliucci, R., Averill, W., Purcell, J. E., and Ettre, L. S.,
Chromatographia. J5, 165 (1975).
(5) Farwell, S. 0., and Rasmussen, R. A., J. Chromatogr. Sci., 14,
224 (1976).
(6) Karasak, F. W., and Field, L. R., Research/Development, p. 42,
March, 1977.
(7) Siegel, M. W., and McKeown, M., Research/Development, p. 101,
July, 1977.
(8) Aue, W. A., and Kapila, S., J. Chromatogr., 112, 247 (1975).
(9) Sullivan, J. J., J. Chromatogr., 87_, 9 (1974).
(10) Burgett, C. A., Research/Development, 25, 28 (1974).
(11) Maggs, R. J., Joynes, P. L., and Lovelock, J. E., Anal. Chem., 43,
1966 (1971).
(12) Aue, W. A., J. Ass. Offie. Anal. Chem.. _59, 909 (1976).
(13) Aue, W. A., J. Chromatogr. Sci.. 2.3, 329 (1975).
(14) Varian Instrument Applications, JLO (3), 4 (1976); Patterson, P. L.,
J. Chromatogr.. 134. 25 (1977).
(15) Sullivan, J. J., and Burgett, C. A., Chromatographia, J5, 176 (1975).
(16) Poole, C. F., J. Chromatogr.. 118, 280 (1976).
(17) Kapila, S., and Aue, W. A., J. Chromatogr.. 118. 233 (1976).
(18) Holden, E. R., and Hill, K. R., J. Ass. Offic. Anal. Chem., _57,
1217 (1974).
175
-------
Section 5P
(19) Hartmann, C. H., Anal. Chem.. 4-5, 733 (1973).
(20) Lubkowitz, J. A., Montoloy, D., and Parker, W. C., J. Chromatogr.,
J76, 21 (1973).
(21) Van de Wiel, J. H., and Thoramassen, P. J., J. Chromatogr., 71,
1 (1972).
(22) Aue, W. A., and Kapila, S., J. Chromatogr. Sci., II, 255 (1973).
(23) Pellizzari, E. D., J. Chromatogr., j>8, 323 (1974).
(24) McCullough, P. R., and Aue, W. A., J. Chromatogr., 82, 269 (1973).
(25) Cook, R. F., in Analytical Methods for Pesticides and Plant
Growth Regulators, Sherma, J., and Zweig, G., eds., Volume VII,
Academic Press, N.Y., 1973, page 187.
(26) Karlhuber, B., Ramsteiner, K., Hermann, W. D., and Simon, W.,
J. Chromatogr., 84, 387 (1973).
(27) FDA Pesticide Analytical Manual, Vol. I, Section 312.
(28) FDA Pesticide Analytical Manual, Vol. I, Section 313.
(29) Lakota, S., and Aue, W. A., J. Chromatogr., 44, 472 (1969).
(30) Tindle, R. C., Gehrke, C. W., and Aue, W. A., J. Ass. Offic. Anal.
Chem. , .51, 682 (1968).
(31) Onley, J. H., and Yip, G., J. Ass. Offic. Anal. Chem., _54, 1366 (1971)
(32) Bayer, F. L., J. Chromatogr. Sci., 15, 580 (1977).
(33) Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Section 8.3.
(34) Gough, T. A., and Sugden, K., J. Chromatogr., 86, 65 (1973).
(35) Mellor, N., J. Chromatogr., 123. 396 (1976).
(36) Hoodless, R. A., Sargent, M., and Treble, R. D., Analyst, 101,
757 (1976); J. Chromatogr.. 136. 199 (1977).
(37) Verga, G. R., and Poy, F., J. Chromatogr., 116, 17 (1976).
176
-------
Section 5P
(38) Burgett, C. A., Smith, D. H., and Bente, H. B., J. Chromatogr.,
134, 57 (1977).
(39) Kolb, B., and Bischoff, J., J. Chromatogr. Sci.. JL2, 625 (1974).
(40) Lubkowitz, J. A., Glajch, J. L., Semonian, B. P., and Rogers, L. B.,
J. Chromatogr., 133, 37 (1977).
(41) Kolb, B., Auer, M., and Pospisil, P., J. Chromatogr.. 134. 65
(1977); J. Chromatogr. Sci.. 15, 53 (1977).
(42) Scolnick, M., J. Chromatogr. Sci., _8, 462, (1970).
(43) Brazhnikov, V. V., Gurev, M. V., and Sakodynsky, K. I., Chromatogr.
Rev.. JJ2, 1 (1970).
(44) Aue, W. A., Advances in Chemistry Series, No. 104, ACS, Washington,
B.C., 1971, page 39.
(45) Burgett, C. A., and Green, L. E., J. Chromatogr. Sci.. 12, 356
(1974); Spectrochim.Acta B, JO, 55 (1975).
(46) Joonson, V. A., and Loog, E. P., J. Chromatogr.. 120, 285 (1976).
(47) Zehner, J. M., and Simonaitis, R. A., J. Chromatogr. Sci., 14,
348 (1976).
(48) Greenhalgh, R., and Wilson, M. A., J. Chromatogr., 128, 157 (1976).
(49) Attar, A., Forgey, R., Horn, J., and Corcoran, W. H., J. Chromatogr.
Sci.. JL5, 222 (1977).
(50) Burnett, C. H., Adams, D. F., and Farwell, S. 0., J. Chromatogr. Sci.,
15, 230 (1977).
(51) Sevcik, J., Thao, N. t. P., Chromatographia, 10, 559 (1975).
(52) Bowman, M. C., and Beroza, M., Anal. Chem.. 40, 1448 (1968).
(53) Zehner, J. M., and Simonaitis, R. A., Anal. Chem.. 47, 2485 (1975).
(54) Krijgsman, W., and Van De Kamp, C. G., J. Chromatogr., 117, 201
(1976) and 131, 412 (1977).
(55) Varian Instrument Applications, JL1 (1), 15 (1977); Patterson, P. L.,
Anal. Chem., 50, 339 and 345 (1978).
177
-------
Section 5P
(56) Laski, R. R., and Watts, R. R., J. Ass. Offie. Anal. Chem.. 56,
328 (1973).
(57) Lawrence, J. F., J. Chromatogr. Sci.. 14, 557 (1976).
(58) Greenhalgh, R., and Cochrane, W. P., J. Chromatogr., 70, 37 (1972).
(59) Cochrane, W. P., and Greenhalgh, R., Chromatographia. £, 255 (1976).
(60) Cochrane, W. P., and Purkayastha, R., J. Agr. Food Chem., 21, 93
(1973).
(61) Young, H. Y., and Chu, A., J. Agr. Food Chem.. 21, 711 (1973).
(62) Lawrence, J. F., and Ryan, J. J., J. Chromatogr., 130, 97 (1977);
135, 117 (1977).
(63) Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Section 8.5.
(64) Winnett, G., and Illingsworth, W. L., J. Chromatogr. Sci., 14, 255 (x97(
(65) Lawrence, J. F., J. Chromatogr., 87, 333 (1973).
(66) Pope, B. E., Rodgers, D. H., and Flynn, T. C., J. Chromatogr.,
134, 1 (1977).
(67) Dolan, J. W. , and Hall, R. C. , Anal. Chem., 45_, 2198 (1973).
(68) MacDonald, J., and King, J. W., J. Chromatogr.. 124. 364 (1976).
(69) Hall, R. C., J. Chromatogr. Sci.. _12, 152 (1974).
(70) Tracor Instruments Retention Times, January, 1976.
(71) Hall, R. C., Tracor Instruments Retention Times, November, 1974.
(72) Wilson, B. P., Cochrane, W. P., J. Chromatogr.. 106. 174 (1975).
(73) Tracor Instruments Model 700 Hall Electrolytic Conductivity Detector
Operation Manual 115008A, pp. 23-24.
(74) Moseman, R., and Thompson, J. F., personal communication.
(75) Lawrence, J. F., and Moore, A. H., Anal. Chem., 46, 755 (1974).
178
-------
Section 5P
(76) Lawrence, J. F., J. Chromatogr.. 128. 154 (1976); Anal. Chem.. 47,
367 (1975).
(77) Bowman, M. C., and Beroza, M., J. Chromatogr. Sci.. _7> 484 (1969).
(78) Bowman, M. C., Beroza, M., and Nickless, G. , J. Chromatogr. Sci., 9^,
44 (1971).
(79) Bowman, M. C., Beroza, M., and Hill, K. R., J. Chromatogr. Sci., 9.,
162 (1971).
(80) Versino, B., and Rossi, G., Chromatographia. 4^ 331 (1971);
Versino, B., and Vissers, H., Chromatographia, J5, 5 (1975).
(81) Moore, C. E., Kara, D., and Marks, G. E., Appl. Spectrosc., 29,
531 (1975).
(82) Bache, C. A., and Lisk, D. J., Anal. Chem.. _37_, 1477 (1965); .38,
783 and 1757 (1966); ^3, 950 (1971); J. Ass. Offic. Anal. Chem..
5.0, 1246 (1967); J. Gas Chromatogr.. £, 301 (1968).
(83) Driscoll, J. N., J. Chromatogr.. 134. 49 (1977).
(84) McLeod, H. A., Butterfield, A. G., Lewis, D., Phillips, W. E. J.,
and Coffin, D. E., Anal. Chem.. 47,'674 (1975).
(85) McLeod, H. A., and Lewis, D., J. Ass. Offic. Anal. Chem., 61,
18 (1978).
(86) Smith, E. D., Sorrells, K. E., and Swinea, R. G., J. Chromatogr. Sci.,
12, 101 (1974).
(87) Purcell, J. E., Downs, H. D., and Ettre, L. S., Chromatographia, JJ,
605 (1975).
(88) Shatkay, A., and Flavian, S., Anal. Chem.. .49, 2222 (1977).
(89) Thompson, J. F., Walker, A. C., and Moseman, R. F., J. Ass. Offic.
Anal. Chem., J52_, 1251 and 1263 (1969).
(90) Thompson, J. F., Mann, J. B., Apodaca, A. 0., and Kantor, E. J.,
J. Ass. Offic. Anal. Chem.. _58, 1037 (1975).
(91) Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Section 12.3.
179
-------
Section 5P
(92) McReynolds, W. 0., J. Chromatogr. Sci., J5, 685 (1970).
(93) Analytical Methods for Pesticide Residues in Foods, Department
of National Health and Welfare, Ottawa, Canada, Table 8.1.
(94) Suzuki, M., Yamato, Y., and Watanabe, T., J. Ass. Offic. Anal.
Chem. , _58, 297 (1975); _59, 1180 (1976).
(95) Rus, V., Funduc, I., Crainiceanu, A., and Trestianu, S., Anal.
Chem., 49, 2123 (1977).
(96) Zweig, G., and Sherma, J., Analytical Methods for Pesticides and
Plant Growth Regulators, Volume VI, Gas Chromatographic Analysis,
Academic Press, N.Y., 1972.
(97) Zweig, G., and Sherma, J., editors, Handbook of Chromatography,
Volume I, CRC Press, Cleveland, Ohio, 1972.
(98) Watts, R. R., and Storherr, R. W., J. Ass. Offic. Anal. Chem., 52,
513 (1969).
(99) Hammarstrand, K., Varian Instrument Applications, _11 (2) , 8 (1977).
(100) Rejthar, L., and Tesarik, K., J. Chromatogr., 131. 404 (1977).
(101) Skogerboe, R. K., and Grant, C. L., Spectroscopy Letters, _3 (889),
215-220 (1970).
(102) Sutherland, G. L., Residue Reviews, H), 85-96 (1965).
(103) 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, B.C., 1971, Chapter 10, page 151.
(104) Glaser, E. R., Silver, B., and Suffet, I. H., J. Chromatogr. Sci.,
15, 22 (1977).
(105) Gaul, J. A., J. Ass. Offic. Anal. Chem., 49, 389 (1966); Section
300.6, FDA Pesticide Analytical Manual.
(106) Baumann, F., Pecsar, R. E. , Wadsworth, B., and Thompson, B.,
Amer. Lab., J3 (10), 97 (1976).
(107) Lines, D. S., Brain, K. R., and Ross, M. S. F., J. Chromatogr., 117,
59 (1976).
180
-------
Section 6
ADDITIONAL PROCEDURES IN PESTICIDE ANALYSIS
This section treats a number of miscellaneous topics important in
residue analysis. These include general considerations for collection
and extraction of samples, methods for concentration of extracts, and
determinations by thin layer and high performance liquid chromatography.
Specific procedures for extraction and cleanup of pesticide and metabolite
residues are discussed in Section 7.
SAMPLE COLLECTION, PREPARATION, AND STORAGE
6A GENERAL CONSIDERATIONS
Special consideration must be given to the procurement, storage, and
transportation of samples to be analyzed for pesticide residues. Pro-
cedures should insure, as well as possible, that the pesticides origi-
nally present have not undergone degradation or concentration and that
contaminating impurities which might interfere with the analysis 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 polyethylene may produce spurious responses. Similarly,
metal containers may contain trace impurities such as oil films, lacquers,
or rosin from soldered joints that will cause interference 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 absorbed 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 proven 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
181
-------
Sections 6B, C
Methodologies for Ambient Pesticide Monitoring" (Monitoring Panel,
Federal Working Group on Pesticide Management, Washington, DC, 1974).
This 60-page manual contains chapters on statistics and study design,
air, soil, the hydrologic environment, estuaries, fresh water fish,
wildlife, foods and feeds, and human tissues.
6B 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 which is likely to contain only metabolites. Similar-
ly, 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 deciding on the sample to be collected so
that it is valid for the purpose of the analysis and valuable time is
not wasted on a worthless sample.
6C 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 foil lined screw caps for autopsy tissue samples
of less than 25 grams, 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 outlined 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,(4),(a),IV of the
EPA PAM.
182
-------
Sections 6D, E
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 (1).
6D
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 6-A shows typical steps in reduction
of a gross sample of an agricultural product collected in the field,
during processing or at the market.
FIGURE 6-A. TYPICAL STEPS IN REDUC-
TION OF A GROSS SAMPLE
Alternative step
Remove peel or husk (if necessory) and
reduce size of large units by cutting or chopping
If necessary, reduce
size of large units by
cutting or chopping
I
6E 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
183
-------
Section 6E
conditions which 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. Extended storage in a freezer can cause de-
hydration or freeze-drying of the sample and possibly affect the subse-
quent analysis (2). 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.
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
organophosphates 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.
Water samples should be extracted at once if at all possible or stored
just above the freezing point 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.
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.
Comments pertinent to collecting samples of different types will be
presented in the Subsections 6F to 6K. Methods for the analysis of
the various sample types are surveyed in Section 7 of this Manual.
184
-------
Sections 6F, G
6F SAMPLING OF AGRICULTURAL AND POOD 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.
It is suggested that readers interested in analysis of sample substrates
of this type for legal compliance to tolerance levels should refer to
these two excellent sources of information. If the purpose of an
analysis is to obtain information on maximum residue levels in a particu-
lar 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.
6G 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 which 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 DDA from DDT. If body tissues or fluids are analyzed
quickly in cases of high exposure, the chance of finding the parent pesti-
cide 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 pesti-
cides or metabolites are expected to be small, samples must be larger,
e.g., morning urine samples or 24-hour pooled specimens.
185
-------
Section 6H
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 screw cap. Plastic bags 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 [EPA PAM, Section 5,A,(1),V].
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 in 24 hours or in a deep freeze (-15 to -25°C) for longer
periods. The analysis of chlorinated pesticides is not adversely
affected by such storage for periods up to one month [EPA PAM, Section
5,A,(2),(a),IV].
6H 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 hydro-
carbon 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
(m3) being controlled by means of a flow meter and timer.
The usual sample consisted of 400 ml ethylene glycol, representing
the contents of four impingers containing 100 ml each. Two impingers
were operated simultaneously for 12 hours, then the other two for an
additional 12 hours. The total air volume sampled in a 24 hour period
was generally ca. 80 m^. The ethylene glycol samples and the filters
used in the machine to collect particulate matter were shipped to the
laboratory in labeled glass containers. Refrigeration of the collected
air samples was unnecessary except in cases of long delay between
sampling and analysis.
The ethylene glycol sampling procedure has not proven to be wholly
acceptable in terms of convenience or reliability. Although ethylene
glycol is an efficient medium for dissolving pesticide residues, it is
difficult to obtain glycol of adequate purity to reduce background in
the subsequent analysis. The EPA national air sampling program has been
discontinued as of this writing pending the development of sampling
and analysis procedures of higher volumes of air than has been possible
in the past. However, the ethylene glycol impinger continues to be
used by some laboratories in local monitoring programs (1, 3, 4).
186
-------
Section 6H
Research (5) in the EPA methods development laboratories has indicated
that a small cartridge packed with ca. 1 gram Chromosorb 101 adsorbent
allows analysis of chlorinated hydrocarbons in air down to 15 ppt after
2 hours of sampling at a rate of 1 to 2 liters/minute. The compounds
are efficiently trapped on these small plugs, which can be very con-
veniently transported to the laboratory for analysis. The short path
length of the collector is possible because of the low air volume used.
Chromosorb 101 was used uncoated after purification by extraction and
activation by heating. Collected pesticides were eluted from the
adsorbent with an appropriate solvent (e.g., hexane for RGB) p'rior to
analysis.
Other proposed media for collecting chlorinated and phosphorus pesticides
and other organic pollutants in air include glass beads coated with
cottonseed oil in a high-rate sampler (6), support-bonded chromatographic
phases such as 24% (Ci8H37sl°3/2)n on Chromosorb A or 31% (CisH37Si03/2)n
on Diatom W (7), and columns of Tenax GC and carbon adsorbents (8).
Adsorption and desorption (at 250°C) of biphenyl at greater than 90
percent efficiency was demonstrated with Tenax (9). 2,4-D acid and
its ester and amide derivatives were trapped from air and recovered
with 86-96 percent efficiency using tubes containing XAD-2 resin (10).
Another approach is exposure of a nylon chiffon cloth screen, saturated
with 10 percent ethylene glycol in acetone, to the atmosphere for a
number of days, followed by Soxhlet extraction of the pesticides from
the cloth with hexane-acetone (1:1 v/v) (11, 12). Results with the
screen are only qualitative, as the amount of air passing through is
not rigorously measured. A similar method involved adsorption on layers
of polyethylene glycol 400 spread on stainless steel nets followed by
extraction of the pesticides into petroleum ether or benzene before
GC determination. Inclusion of glass fiber filters plus coated nets
in the collection tubes differentiated aerosol-bound pesticides and
vapor fractions (13). Composite filters composed of layered glass
fiber and Porapak R have been shown to efficiently collect malathion
(13a), methyl parathion (13b), and chlordane/heptachlor (13c) at flow
rates of 150 to 170 liters/minute. The carbamate insecticide propoxur
(Baygon) was collected from air in a Greenburgh-Smith impinger containing
NaOH. The solution was acidified and the pesticide extracted with benzene
prior to analysis (14). Hexylene glycol contained in glass scrubbers has
been used to recover dieldrin and heptachlor from air at 0.1 ng/m-* (15).
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. Polyurethane foam vapor traps following
a particle filter have been evaluated (16) for sampling of pesticides,
PCBs and polychlorinated naphthalenes. Collection rates up to 360 m^
of air per 24 hours and sensitivities as low as 0.1 ng/m^ for some
compounds were achieved. The filters and plugs were Soxhlet extracted
187
-------
Section 61
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 cleanup. Collection was generally
satisfactory but was poor for the more-volatile OC1 compounds. Recovery
was ca. 75 percent 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 an optimization
method given for the plug size and shape for any chosen sampling rate.
Trapping efficiency depended upon pesticide vapor pressure and the
flow rate of air. The quantitation limit was ca. 0.1 ng/m in a
5 m-^ air sample. It was crucial that the plugs were carefully protected
from laboratory contamination and Soxhlet extracted with pesticide-
grade acetone and hexane prior to use if clean blanks and highest
sensitivity were to be achieved (17).
Because of possible pesticide degradation, air sampling apparatus should
be shielded from light during sample collection.
61 WATER SAMPLING
The design of a comprehensive pesticide sampling program for environ-
mental waters is a specialized topic which is covered in publications
available from the Water Quality Control Division of the USEPA, National
Environmental 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 which would be least representative 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.
188
-------
Section 61
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 sensitivity 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 two liters or more may be needed for a sur-
veillance 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, suspected con-
taminants, 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
couple of hours. If this is impractical because of distance from
the sampling site to the laboratory and/or the laboratory work load,
storage should be made 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 (18) . 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.
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 proceeding 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 (5-90 percent to 46-68 percent) of DDT in water
analysis upon storage has also been noted due to adsorption on
suspended matter in the sample (19).
189
-------
Section 61
The assumption made is that a grab sample is at least representative
of the immediate water mass from which it was taken and, hopefully,
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
upon the study objectives and the variability within the "population"
being sampled.
After sampling, pesticides are extracted from water, cleaned-up and
concentrated as necessary, and determined by GC or an alternate
method. Pesticides 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 cleanup. 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
organics in natural waters since they were developed and introduced
by the U.S. Public Health Service in 1951 (20). 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 combina-
tion 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 bacteria]
and oxidizing attack on the sorbed pesticides have caused problems
with carbon columns (21, 22).
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. Carbo-
wax 4000 (5 g) and ja-undecane (23), silicones chemically bonded
to diatomaceous earth support (24), covalently bonded aromatic
and alkyl chlorosilanes on Celite (25), porous polyurethane foam
columns (for pesticides and PCBs) (26, 27), polyethylene film
(20-25 jam thickness) (28), and polyurethane foam coated with
selective adsorbents (29) have all been used with varying success.
190
-------
Section 61
The XAD macroreticular adsorbent resins (XAD-1, -2, -4, and
-7) have been used to collect organics from both potable (30,
31) and sea (32) water. Optimum conditions for use with XAD-4
resin were found to be 2 grams of adsorbent, a flow rate through
the resin of 8 ml/minute, and 100 ml hexane-diethyl ether
(10:1 v/v) as eluting solvent. Among 10 chlorinated insecticides
studied, only aldrin and £,£'-DDE were not quantitatively recovered,
and recovery of PCBs was 76 percent (33). Details for use of
XAD-2 and -4 resins for many classes of trace organic water
contaminants have been published (34) and recoveries between
81 and 96 percent reported for 20 ppt levels of atrazine, lindane,
dieldrin, DDT, and DDE (47 percent for aldrin). An EPA report (35)
recommends XAD-2 resin for routine monitoring of sea water for
chlorinated insecticides and PCBs. Average recovery for XAD-2
extraction of fortified natural waters collected across Canada
was 85 percent for the 10-100 ng/liter levels of 10 OC1 pesticides
(recovery of mirex was unacceptably low) and 82 percent for 250
ng/liter levels of PCBs; blanks from the resin were a low 4 ng
PCBs/liter (36). Amberlite XAD-4, porous polyurethane foam, and
undecane plus Carbowax 4000 on Chromosorb were found to be com-
parable for extracting 10 OC1 insecticides from environmental
water samples (23).
Continuous liquid-liquid extractors are an alternative to the
filter-adsorbent processes preferred by some analysts. A multi-
chamber extractor with internal solvent renewal replenishing (37)
allowed extraction of 135 liters of water at rates of 0.5-1.0 liter/
hour and recovered greater than 97 percent of ppb levels of pesticides,
Subsequently, a similar modified apparatus permitted use of both
heavier-and lighter-than-water solvents (38). A simple and rugged
field version of the Kahn and Wayman apparatus (37) excluded solvent
recycling and was based on mixed settling (39). 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
(40). A Teflon helix, continuous liquid-liquid extractor, plus a
continuous evaporative concentrator recovered jjg to ng per liter
amounts of OP pesticides from river and sea water or from secondary
sewage effluent with >80 percent efficiency (41). 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 (42).
191
-------
Section 6J
The theory for extracting chlorinated pesticides continuously
from water with a stationary immiscible solvent is discussed
in reference (39).
See Section 10,A,III-V of the EPA PAM for updated material
on water sampling.
6J SAMPLING OF HOUSE DUST, SOIL, AND STREAM 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 one square meter in size. Ten to twenty cores repre-
senting a surface area of at least 200 cm^ are recommended. The first
two 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
(1) involves collection of cores 1-3 inches deep and 3 inches in
diameter with a hand-operated auger; on a 1/4 acre site (105 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 fifteen
feet from the original sampling line, until a total of 49 cores is
collected. The cores are sieved through a hardware cloth screen into
a 3 gallon galvanized pail and thoroughly mixed. The sample is trans-
ferred 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 reproducibility 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 con-
cerning 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 contamination profile of the body of water. Bottom sediment
varies with respect to both particle size composition (surface adsorp-
tive power) and organic content. Therefore, sample sites should be
selected at random in an effort to collect samples representing a range
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.
192
-------
Section 6K
About a quart of sediment is a typical sample size. Actual collection
is accomplished with one of a variety of core samplers or dredges.
A diagram of a dredge-type device for collecting sediment samples
has been published (1). 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 shipment 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 (43). Samples
may be preserved with formalin or a variety of other sterilants
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.
6K MARINE BIOLOGICAL SAMPLES AND OTHER MISCELLANEOUS SAMPLES
A problem sometimed encountered when collecting plankton and bottom
organisms is to obtain the minimum weight necessary for successful
analysis. As a general rule, a minimum of about 10 grams will be
required. Collected organisms can be frozen at once or preserved
with 5-10 percent formalin or 70 percent 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 preservative must
be extracted and analyzed to determine if exchange of pesticides 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 hand picked from the remaining debris.
Fish are collected utilizing seines, gill nets, traps, electrocution
devices, 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 percent ^28203 showed
no loss of Abate (temephos) residues (>1 ppb) for up to three weeks (44).
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
193
-------
Section 6L
dying fish may be valuable for determining probable cause of death
where pesticide exposure is suspected. The blood is obtained by
cutting off 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
Wojick £t ad. (45). 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 (1).
Some of the material in the sections on sampling was adapted from an
EPA training course manual (46). A review that includes some of the
above sampling procedures and additional methods for collection of
environmental samples has been published (1).
6L 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 7 of this Manual.
This subsection discusses general considerations of pesticide extraction
from collected samples.
Many solvents are employed for extracting residues, depending upon 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 (acetoni-
trile plus ca. 35 percent water for low moisture samples), and hexane/
acetonitrile systems are widely recommended for partition cleanup.
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
194
-------
Section 6L
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
masceration 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 Blendor (Figure 6-B),
omni-mixer, or Wiley mill is probably the most usual extraction pro-
cedure 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 five 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
described (47). 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 pesti-
cides has also been devised (48).
Figure 6-B. Waring Aseptic Dispersall Model AS-1. (Shown on 702-CR Base)
195
-------
Section 6L
Blending with a solvent followed by filtering or centrifuging is
particularly efficient for most vegetable samples. The water in
the sample may give rise to emulsions with nonpolar solvents, and
this can often be avoided by use of a drying agent such as anhydrous
Na2SO^ 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 Na2SO^
and sharp sand to help break down some connective tissue. More
volatile pesticides (e.g., lindane) might be lost in this way.
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)
(49). Preliminary steps such as drying, grinding, or chopping
normally precede the extraction. (Soxhlet thimbles may require
exhaustive extraction prior to use so they do not contribute interferences
to the analysis (50)). Even this procedure may not give complete
extraction in all cases, and only studies with samples to which radio-
active 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.
Water samples (100-500 ml) are generally extracted by shaking with an
appropriate solvent (3 x 100 ml) in a separatory funnel. 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
[51], 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 [52]-
An apparatus that simultaneously Soxhlet extracts pesticides and con-
centrates the resulting extract has been designed [53]. 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.
196
-------
Section 6M
The most efficient solvent and parameters for extraction of pesticides
from water can be determined using the jD-values originally suggested
by Beroza and co-workers for use in residue confirmation (Subsection
8F in Section 8). The _p_-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 (54), and the best solvents were found to be
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 (55).
Using 2,4-D as a model, it was shown that a pesticide with a £-value
equal or greater than 0.90 can be 95 percent extracted from 1000 ml
aqueous phase by up to five extractions of >50 ml with a total volume
of solvent up to 500 ml. The best solvents and conditions were then
worked out on a practical scale for extraction of 2,4-D and 2,4,5-T
and their esters (56). Diethyl ether or ethyl acetate was found best
for 2,4-D acid and esters and benzene for 2,4,5-T acid and esters, and
a 99 percent 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.
6M CONTROL OF METHODOLOGY FOR CONCENTRATION OF SAMPLE SOLUTIONS AND
FRACTIONAL' 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 concentrate final solutions according to the least sensitive
pesticide in the method's scope.
Purified extracts or eluate solutions are concentrated to a volume of
ca. 5 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 6-C).
197
-------
Section 6M
Figure 6-C. Kuderna-Danish
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 percent 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 con-
centrate is collected in the lower tube. Boiling chips are re-
quired 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.
198
-------
Section 6M
For concentration from 5 ml to smaller volumes (as low as 0.1-0.5 ml),
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 6-D) 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 agitation. The apparatus is
cooled 3-5 minutes, and condensate is allowed to drain down into the
tube before removing the column. 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 which 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 PAM, Section 5,A,3,a.
Figure 6-D. Semi-Micro Kuderna
Danish Apparatus, Kontes Glass
Co., No. K-569250.
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. Extracts contained in a beaker or
a centrifuge tube immersed in a water bath at 40°C can be evaporated
199
-------
Section 6M
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 which could
cause pesticide degradation. Warming a tube by holding it in the
hand is a useful, gentle evaporation aid during nitrogen blow-down.
An evaporation assembly combining an evaporative concentrator tube,
a Kuderna-Danish flask, and a rotary vacuum evaporator (Figure 6-G)
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 confiries 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.
An excellent multitube apparatus for nitrogen evaporation is available
from Kontes Glass Co. (57). 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
(58) are avoided, and a minimum of analyst attention is required
(Fig. 6-E).
Figure 6-E. Ebullator,
Kontes Glass Co.
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 nitrogen through the solution to initiate and maintain
ebullation. Recoveries of seven chlorinated pesticides after concen-
tration for two hours in this apparatus were greater than 94 percent
with both hexane and benzene solvents.
200
-------
Section 6M
It is important to avoid pesticide loss or decomposition during
evaporation steps. Numerous reports have been made [e.g., 58, 59]
of severe pesticide loss during concentration steps, even in the
presence of sample coextractives. 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 pesticides can be concentrated to
small volumes without loss by the K-D evaporative procedures described
at the beginning of this subsection. Evaporation to dryness should
never occur. If the complete removal of a particular solvent is re-
quired, 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 relatively high, but significant losses have been documented of
even the more stable pesticides at low concentration levels [58].
A commercial tube heater which avoids evaporation to dryness with micro
K-D apparatus was originally described by Beroza and Bowman [60]
(Figure 6-F). Six extended-tip K-D concentrator tubes are accomo-
dated and simultaneous evaporation to less than 1 ml can be carried
out without attention.
Figure 6-F. Tube Heater,
Kontes Glass Co., K-720000
201
-------
Section 6M
Other reports of pesticide loss include dieldrin and DDT when an
extract was evaporated in the presence of light [61] and carbamate
pesticides when evaporated in a K-D apparatus 162]. In the latter
case, rotary vacuum evaporation (Figure 6-G) at 50-558C with addition
of a keeper solution was recommended. Many carbamates can be success-
fully evaporated under a nitrogen stream without loss after adding a
keeper. A satisfactory general purpose keeper is 5 drops of 1 percent
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.
Figure 6-G. Rotary Evaporator,
Kontes Glass Co., No. K-570160
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.
202
-------
Section 6N
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 ultra-
violet LC monitor.
HIGH PERFORMANCE LIQUID COLUMN CHROMATOGRAPHY (HPLC)
6N 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 which commonly operates at ambient
temperature. It has advantages ovfer gas chromatography where compounds
are -not naturally volatile and derivatization is difficult or unsatis-
factory, and for thermally labile compounds. Sample preparation may be
much simpler for LC since complicated matrices can often be injected
directly where GC would require a prior cleanup procedure. Compounds
which may not gas chromatograph well unless derivatized (e.g., phenoxy
acid herbicides, which require methylation prior to GC) often give
excellent liquid chromatograms [633. The two most widely used detectors
in LC, the differential refractometer and ultraviolet (UV) photometer,
are nondestructive to the sample, making fraction collection a routine
matter. LC is ideally suited to sample scale-up, and because of high
loadability, detection of the sample is usually not a problem. The
separated compounds emerge from the chromatograph dissolved in solvent
which 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 determination [64] as well as for the final
determination itself.
203
-------
Section 60
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~-> to 10~9 gram range. In one comparative study of GC and HPLC
(63), detection limits 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 pi) without loss of linearity or
peak symmetry. The ability to introduce large volumes in LC can
make sensitivity between the two methods comparable. Derivatization
methods can increase the sensitivity of detection, e.g., by formation
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 determinations.
60 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 6-H). 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 together 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.
204
-------
Section 60
FIGURE 6-H HIGH PERFORMANCE LIQUID CHROMATOGRAPH INCLUDING TWO MODEL 6000 PIUPS,
A M30EL 660 SOLVEKT PRQGRAKM3!, AND MODEL ¥0 UV ABSORBANCE AND
WlOl DIFFERENTIAL REFRACTOMETER DETECTORS, MATERS ASSOCIATES.
The pumping system must provide the pressure required to achieve
the correct flow rate through the column. Although most instruments
permit 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. HPLC 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). The
continuous-displacement pumps deliver a smooth, pulseless flow of
solvent but have a limited solvent reservoir, which may require
interrupting a run for refilling. They are inconvenient for
gradient elution because the composition of solvent in the pump
cannot be changed during the run, and the solvent gradients must
be formed at high pressure downstream from the pump. Intermittent
displacement pumps operate from an open, unlimited solvent reservoir
at ambient pressure, and solvent refill or formation of solvent
gradients upstream at low pressure is no problem. These pumps
have a pulsating output which must be damped, electronically filtered,
or cancelled-out by a two pump head arrangement.
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 columns (either by stopping the pump or by using
a shut-off valve), removing a cap from the head of the column, and
205
-------
Section 60
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
resumed without significant loss of efficiency. Two different
injector systems allow sample introduction from a syringe without
stopping solvent flow. With 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 (65)).
Most commonly, HPLC utilizes the ultraviolet absorbance detector
(fixed 245 and 280 nm or continuously variable wavelength), re-
fractive index detector, fluorescence detector, or traveling wire
flame ionization detector. The mercury source of the UV absorption
detector has a maximum output at 254 nm, so sensitivity is greatest
for compounds that absorb at or near this wavelength. Other wave-
lengths are obtained, at the cost of some sensitivity loss, by
interposing a filter between the lamp and cell or using a continuously
variable, spectrophotometer type detector. Refractive index detectors
are either based on the Christiansen effect or are of the prism type.
These are universal, relatively insensitive (10~6 -10~7 g) detectors
that require close temperature control and cannot be used with solvent
programming. Fluorescence detectors are highly sensitive «10~^ g)
and selective because of the choice of excitation and emission wave-
lengths. Derivatization (pre- or postcolumn) of the pesticide of
interest is often required. 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.
206
-------
Section 6P
6P THEORY AND PRINCIPLES OF HPLC
The principles and theory governing 60 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 thermodynamic 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:
Vl - Vo
Retention: k =
Vo
where V]_ is the retention volume for the peak of interest from the
point of injection, and V0 is the retention volume of a non-retained
peak measured at the peak apex. Times or distances measured along
the recorder chart can be used more conveniently than volumes if
the flow rate is constant.
207
-------
Section 6P
Selectivity for k'2 V2 - V0
compounds 1 and 2: a = " =
- Vo
.2
No. of Plates: N - 16 / Vl
where W^ is the width of the peak for component 1 (see Figure 4-C)
in terms of V.
Resolute: !-
~ ) (
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 k' is ca. 2-6. a is the separation or selectivity term,
which is affected by the chemistry of the entire system, including
the functionality of the sample components. Values of 1.1-2 are
more typical in HPLC.
N is a measure of band broadening; typically, a highly efficient
small-particle, porous 25-cm silica gel column will show approx. 10,00i
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 unsatisfactc
there are three different ways to try to improve the separation. The
solvent can be changed (solvent strength lowered) 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
selectivity (a). As a specific example, an inferior reversed phase
separation on a 37-75 u bonded phase C^Q column with methanol/water
solvent might be improved by changing to a lower percentage methanol
(increase k'), changing to a 10 ^i column (increase N), or changing to
acetonitrile/water solvent (increase a).
208
-------
Section 6Q
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
CIQ bonded column. If there is intermediate 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 C-^g column
is recommended. Use the best available column to increase N,
choose a solvent mixture and vary its proportions to alter k!,
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 increasing the plate number is to reduce the solvent flow
rate with a constant column length, but again we pay for this by
increased separation 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
(66).
6Q COLUMNS AND SOLVENTS 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 ^un can be successfully dry packed in the laboratory
while microparticulates (5-10 pn) must be slurry packed. Because of
the difficulty of this operation, commercial, pre-packed columns are
usually used.
209
-------
Section 6Q
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 p.) and also have higher capacity than pellicular packings.
They require high inlet pressure for acceptable flow rates and
must be slurry packed. Present use of pellicular packings is
mostly for situations where highest resolution and speed are not
required, cost is a factor, or dry packing is desirable (micro-
particulates are at least one order of magnitude more efficient
than pellicular packings). An exception to this is pellicular
ion exchangers, which are often preferred to the microparticulate
type.
For liquid-solid adsorption chromatography, microparticulate silica
gel of irregular or spherical shape is used most. 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 Cg or C^g linear hydrocarbon covalently bonded to
silica gel particles is by far the most widely used LC mode. Aqueous
solvent mixtures are used as the mobile phase. This method owes its
popularity to its ability to separate nonionic, ionic, and ionizable
substances in partition, ion suppression, or ion pairing (67) modes;
the stability of the bonded-phase columns (if properly used); and the
simple, inexpensive solvent systems, utilized such as methanol or
acetonitrile/water mixtures. Other reversed-phase bonded packings
have phenyl and cyclohexyl groups, while normal phase bonded packings
contain polar functionalities (e.g., nitrile, amine). A detailed
discussion of HPLC columns and column technology has been published (68]
Solvents chosen for adsorption HPLC should have a low viscosity (for
high efficiency), 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 detector).
A widely versatile set of solvents with a range of chromatographic
properties is hexane, methylene chloride, diethyl ether, acetonitrile,
and methanol (most polar — strongest eluent for normal phase adsorptioi
chromatography). These solvents are usually used in mixtures with a
210
-------
Section 6R
solvent composition and strength that optimize capacity and selectivity
(Section 6P). Solvent mixtures for adsorption HPLC of polar samples
often contain at least a small concentration (0.01-1 percent) of a
polar modifier (e.g., water, alcohol, acetonitrile) (69).
Many of the same considerations apply to the selection of solvents
for the other important HPLC mode, reversed-phase chromatography
on chemically bonded packings. Because the phases are reversed,
water, the most polar solvent, is the weakest eluent, while neat
methanol and acetonitrile are the strongest eluents used in most
applications. The most polar solutes are the least retained on the
column. Eluents of intermediate polarity are usually obtained by
mixing methanol or acetonitrile with water or an aqueous acid, base,
or buffer solution (70) . One instrument manufacturer strongly
recommends ternary solvent mixtures of varying composition, e.g.,
methanol, acetonitrile, and water (71). In general, silica gel
reversed phase packings are stable only with solvents in the pH
range of 2-7.5. 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 com-
pared to isocratic elution (72)).
Solvent selection for a particular separation is aided by two funda-
mental 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 (73).
6R PRACTICAL ASPECTS OF 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 (this can require up to several hundred
void volumes with an adsorption column), 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 de-
grades they will not overlap. The test mixture can contain pesticides
to be analyzed in real samples with the same solvent system that is to
211
-------
Section 6R
be employed, or it can be a mixture specified by the column manu-
facturer 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 (kf) and relative (d) retention
plate number (N), asymmetry (tailing), void volume, and pressure drop.
Small differences in ot and k' usually reflect normal differences in
solvent composition, but large decreases in these parameters or an
increase in asymmetry are indications of column degradation or de-
activation. 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 usually be restored by
regenerating the column with a series of solvents of increasing
eluting strength (adjacent solvents must be miscible). 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
methylene chloride-dry hexane (20 column volumes each). Acetonitrilc
is an effective solvent for regenerating reversed phase bonded
packings. The pump and connecting tubing are pre-washed with each
new solvent that is put through the column, high flow rates are used
to speed the operation, and the detector is left connected, if
possible, to also clean it.
The inevitable, permanent column degradation that occurs with pro-
longed 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 recommendations 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 discussed along with other aspects
of solvent purity in Section 31. Sample extracts or solutions should
also be free of insoluble particles (74) and should be filtered, if
necessary, with a hypodermic syringe fitted with a Swinny-type filter
(0.5-1/i), Irreversibly sorbed compounds can irrepairably 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 p. pellicular
212
-------
Section 6R
sorbents, but some efficiency may be lost if the analytical column
contains microparticulates. Columns damaged by plugging, bad com-
pression, 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, alternatively,
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.
HPLC columns should never be bumped, dropped, jarred, bent, tapped,
or vibrated. All connection fittings must be clean (use an ultra-
sonic 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 never stored in aqueous solution but are flushed and
stored in methanol or acetonitrile. If the column has been used
with a buffer, it is flushed with water and then 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 (69) follow. Baseline drift can be caused by strongly
adsorbed 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 can be suppressed by changing the solvent pH or ionic strength.
Injection of the sample in a solvent stronger than the eluent can
also cause tailing; a solvent weaker than the eluent, or more pre-
ferably the eluent itself, should always be used, if possible, or
the sample may be injected in a very low volume of a stronger solvent.
Drifting retention times can be caused by differences among solvent
batches, changes in composition of a batch of solvent on standing,
changes in flow rate or temperature, or inconsistent adsorbent
activity. Adsorbent activity is maintained constant by using clean
samples, pure solvents with an adequate level of modifier, and
213
-------
Section 6S
frequent column regeneration. Poor reproducibility of retention
times and peak areas in gradient elution may arise from inadequate
column regeneration between runs, nonreproducible pump delivery,
or poor mixing in the mixing chamber.
Dirty detector flow cells may be cleaned by using a 10 ml syringe
to rinse the cell successively with methylene chloride, methanol,
and water. The cell is then filled with 50 percent nitric acid
and allowed to stand for 30 minutes, followed by flushing with
filtered, distilled water (75). If this operation does not 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) (76).
Readers using HPLC for analysis of pesticide residues are strongly
urged to study the excellent discussion of many practical aspects
of the field given in reference (69), from which much of the
material in this subsection was taken.
6S 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 revisions 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 (77-79), column packings (68),
and general principles and equipment (80, 81), and to books covering
theory, principles, and practice (82, 83). 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 (84).
A novel method with great promise for the simplified monitoring
of residues in water samples down to ppt levels has been termed
214
-------
Section 6S
"trace enrichment". The procedure combines concentration, separation,
detection, and quantitation 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 eluent, so the organic compounds will be
concentrated in a tight band at the head of the column). A gradient
elution from 100 percent water to 100 percent 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 (85).
Table 6-1 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 determinative step. For example,
1977 issues of the Journal of the AOAC have reported HPLC formula-
tion determinations of Folpet, esters and salts of 2,4-D, methyl
parathion, and methyl pirimiphos. Reviews of earlier applications
of HPLC to pesticide analysis are cited in References (63, 108-111).
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 ioniza-
tion; (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 hydrocarbon and subsequent
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 and some applications of HPLC-MS have been reviewed in
detail (112). A simple off-line coupling procedure has also been
described (113).
215
-------
Table 6-1
SELECTED SEPARATIONS AND DETERMINATIONS OF PESTICIDES BY HPLC
ON
Compound
determined
TH-6040 insect
growth regulator
Carbamate
insecticides and
metabolites
Carbofuran and
3-OH and 3-keto
metabolites
17 Carbamate and
urea pesticides
5 fungicides
(benomyl, phenyl-
phenol, biphenyl,
thlabendazole,
carbendazim)
Difenacoum and
Warfarin
Methyl
parathion
Sample
matrix
bovine manure
vegetables
crops
standards
only
fruits
liver,
plasma,
urine
runoff water
Column
type
CIB reversed
phase
GIS reversed
phase
Sum silica
gel,
adsorption
Partasil-10
ODS reversed
phase
adsorption,
reversed
phase, ion
exchange
Corasil II
pellicular,
adsorption
Paiiasil ODS,
reversed
phase
Solvent
system
acetonitrile-
water (57:43)
3-8% isopro-
panol in
isooctane
different
polarity
solvents
various ,
compatible
with column
isopropanol-
CHCl3-isooctane
(1:2:397)
acetonitrile-
water (50:50)
Detector
UV, 254 nm
fluorescence,
after dynamic
hydrolysis and
labeling
UV, 254 or
280 nm
UV, 254 nm
UV, 254 or
288 nm
UV
UV, 270 nm
Cleanup ' Recovery
Florisil 97%
column
partition 69-90%
partition and 68-110%
Florisil column
— —
steam distillation, 64-102%
partition,
chemical reaction
GPC or 62-90%
partition
XAD-2 for 99%
preconccntration
Residue Reference
level
0.5 and 2 ppm (86)
0.01-1 ppm (87)
0.1 and 1 ppm (88)
(89)
0.8-100 ppm (90)
0.025-5 ppm (91)
2-3 ppm (92)
-------
Table 6-1 (continued)
Warfarin
Azinphos methyl oxon
Phenylurea
herbicides
Karbutilate
and metabolites
NO
•~j Triazine, urea,
uracil herbicides
Warfarin and
metabolites
biological
fluids
foliage
soil
soil,
river water
water,
soil, grass
standards
blood, plasma
and liver
microsomes
Car hamate pesticides foods
(Swep, Landrin,
carbof uran , amlnocarb ,
Micropak
CH-10
Corasil-
uC^g, reversed
phase
Spherisorb
ODS, reversed
phase
uPorasil,
adsorption
PE C18 Sil-
X-ll,
reversed
phase
PC18>
reversed phase
5 pn silica
gel,
adsorption
methanol-
0.5% acetic
acid (1:1)
acetonitrile-
water gradient
methanol-
water-NH3
3-7% ethanol
In ethylene
dichloride
2.5-25%
methanol in
water
1.5% acetic
acid-aceto-
nitrile (69:31)
57, isopropanol
in isooctane
UV, 308 nm
UV or on line
colorimetric
reactor
UV
UV, 254 nm
UV,
200-300 nm
UV, 313 nm
UV, 254 nm
freezing of
aqueous phase
none
none
Florisil
column (for
grass extracts)
— ~
on- column
concentration
partition and
FlorisJl column
Banol, carbaryl,
Zectran, methiocarb,
propoxur, Mobam)
PCBs
commercial 5 urn silica dry hexane
mixtures gel adsorption
UV
2-5%\
C.V.J
96%
>70%
0.1-4 jig/nil
1-10 ppm
(soils)
(93)
(94)
90-100% 0.5-1 ppm (95)
80-103% 0.1-0.2 ppm (96)
Sensitivity (97)
0.05 ppm In soil
and 0.001 ppm in
water stated
-100% 0.04-0.08 pg/ul (98)
0.1-0.3 ppm (99)
-------
Table 6-1 (continued)
Urea herbicides
(linuron,
chlorbromuron,
chlroxuron,
fenuron, etc.)
Benomyl and
TBZ
Paraquat or
diquat
£f Dioxins
00
vegetables
and wheat
plant
tissues
urine
PCP
5 jim silica
gel,
adsorption
20-23 pm
silica gel,
adsorption
ot-amino-
propyltri-
ethoxysilane
bonded to
20 pm alumina
Perraaphase-
ODS, reversed
phase
5-20% iso- UV, 254 nm
propanol in
isooctane
0.1% acetic fluorescence
acid in
methanol
pH 2.45 UV, 258 nm
phosphate
buffer-
raethanol (11:14)
14Z H20 in UV
methanol
partition
and Florisil
column
benomyl
hydrolyzed to
2-AB
none
extraction,
ion exchange
column
Terbutryne
Chlortoluron
Methazole, DCPMU,
and DCPU
water
Permaphase- 20% methanol
ETH, aliphatic in water
ether, chemi-
cally bonded
pellicular
UV
soil
p lant and
animal
samples
10 fim silica
gel,
adsorption
Zorbax SIL,
adsorption
15% isopropanol UV, 240 nm
in hexane
0.05% methanol UV, 254 nm
in dichlorome
thane or CHC1,-
silica gel
column
partition and
silica gel
column
petroleum ether-
methanol
(6:3:0.5)
>80Z 0.01-1 ppa (101)
90-103% ca. 0.01- (102)
.2 ppm
ca. 100% 650 pg/L (103)
93-104% 0.7-41 ppn (104)
89-100Z 0.001-0.1 ppm (105)
71-95Z 0.25-2.5 pp. (106)
75-105Z 0.1-0.2 ppn (107)
-------
Section 6T
THIN LAYER CHROMATOGRAPHY
6T INTRODUCTION
The first multiresidue method available to the pesticide analyst
for identification and estimation was based on paper chromatography
(114-116). Paper chromatography has now been largely replaced by
thin layer chromatography (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 (116-119).
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 Subsection 8E in Section 8 of 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 concen-
tration 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 percent diethyl ether fraction
from the Florisil column cleanup of a fat sample contains a large
amount of lipids. Although adequate for GC, further TLC or column
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 (120).
Major advantages of TLC are simplicity, rapidity, and low cost.
Sensitivity 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 (121), while
specific procedures for pesticide TLC were covered in several papers
(117, 122). Applications of TLC to pesticide analysis have been
reviewed (123, 124).
219
-------
Section 6U
6U 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
spotted to allow for detection after the run. Care should be taken
that the spotting pipette 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 6-1) or purchased pre-coated on glass, plastic, or
aluminum backing. Analytical layers are usually 250 um 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. Activation
of adsorbent layers (e.g., 80-110°C for 30-60 minutes) prior to
spotting and development is often required. Once the adsorbent
has been activated, it must either be used promptly or stored under
desiccating conditions, or activation must be repeated. Silica gel
and alumina layers usually give the best results, but polyamide,
microcrystalline cellulose, kieselguhr, zinc carbonate (125) and
magnesium oxide among other adsorbents have also been used. For
reversed phase TLC, hydrophobic C18 chemically bonded silica gel
plates are commercially available.
FIGURE 6-1. lfesM*/BRrNHWH ADJUST-
ABLE APPLICATOR POR
COATING RESULAR OR GRA-
DIENT LAYER PLATES,
BRINWUNN INSTWBITS,
INC,
220
-------
Section 6U
Chromatography is carried out in a development chamber, most often
a rectangular, glass, paper-lined tank saturated with solvent vapors
(Figure 6-J). Low volume "sandwich" chambers are also used. Both
saturated and unsaturated atmospheres have been 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 or 5 x 5 cm layers of silica gel with a smaller particle
size and a narrower particle size distribution than in conventional
TLC plates, and thereby gives improved resolution and sensitivity of
detection. Volumes no larger than 1 microliter 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 rectangular chamber or in a commercial, automatic U-chamber
device producing radial zones (126) (a special radial scanner is
needed to quantitate these separations). High resolution is achieved
rapidly with short development distances. 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 ul
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 (127). 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 (128). HPTLC plates are available from Merck and
Quantum Industries, and HPTLC equipment from Camag, Inc.
221
-------
Section 6U
FIGURE 6-J, IESAGA REOTWSJUR TIC TANKS,
BRIHWINN INSTRUMENTS, INC.
The following solvent systems have proved to be generally useful for
separation of a wide range of pesticides on 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 components 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
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 cr.n be detected by applying a chromogenic reagent, either
by spraying or dipping. A commercial aerosol spray device is
shown in Figure 6-K. Dipping is the preferred method of applica-
tion, 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
222
-------
Section 6V
fluorescent spots can be detected under short (254 nm) or long (366 nm)
wave UV light, or fluorescence may be induced by application of fluoro-
genic reagents after development or preparation of fluorescent deriva-
tives (e.g., dansyl compounds} prior to spotting (129). Spots that
absorb UV light are detected as quenched (dark) spots on layers con-
taining phosphor activated by UV light (usually 254 nm). Radioactive
(labeled) pesticides are detected by autoradiography and some fungi-
cides by direct bioautography.
FIGURE S-K. TIC AEROSOL SPMVER, BWNMWNN
) INC.
6V QUANTITATIVE TLC
Quantitation of separated spots may he achieved by "eyeball" comparison
between sample and standard spots run on the same plate or by some
independent analytical method (e.g., spectrophotometry or GC) after
scraping the spot, collecting, and eluting the pesticide from the adsor-
bent. Manual elution is simply carried out by scraping the area con-
taining the pesticide spot, collecting the scrapings in a vial or tube,
adding solvent and agitating (vortexing), filtering the adsorbent, and
concentrating the filtrate containing the pesticide. An automated
elution system is available from Camag, Inc.(130). Radioactive spots
can be quantitated by scintillation counting after scraping or by
automatic scanning of radioactivity on the layer.
223
-------
Section 6V
Colored, fluorescent, or quenched spots may be scanned on the layer when
a spectrodensitometer is available. Quantitation is achieved by scan-
ning 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 versa-
tile densitometer is capable of scanning in single or double beam and
reflectance or transmission modes, and has monochrometers or filters for
selection of the best wavelengths of incident and emitted (for fluo-
rescence) energy. Important considerations for densitometry are precise
and accurate spotting, uniform layers, R^ values between 0.3 and 0.7,
uniform application of detection reagents, and optimum use of a good
densitometer.
Manual spotting is best performed with 1 or 2 yl Microcap disposable
pipettes, using repeated spotting with drying in between for larger
volumes. It has been shown that sample delivery errors below 1 percent
are feasible with Microcaps (131). Larger volumes of sample extracts are
conveniently and reproducibly spotted with a device such as the Kontes
automatic spotter which 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 6-L
and described in reference (132).
FIGURE 6~L, fiten antes THIN LW sssme* «w nurwuttc
SPOT sm,!CATa>, tents rt*ss Cww, INC.
224
-------
Section 6W
Pre-coated layers are recommended for quantitative TLC since it is
very difficult to hand-coat layers with adequate uniformity. 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 com-
pounds on the layer.
When developing a new densitometric method, the spots of interest should
be scanned in all possible modes and directions and at a variety of
wavelengths in order to obtain the best signal to noise ratios and
selectivities. The optimum conditions are then used to obtain the
calibration 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 percent on a
routine basis and can rival GC for determination of certain pesticide
residues in the hands of an experienced operator. Principles and
experimental details of thin layer densitometry have been published (133),
and quantitative TLC of pesticides has been reviewed (134). Table 6-2
contains some selected applications of thin layer densitometry reported
since this review was published. A fiber optics scanner specifically
designed for pesticide analysis (135) is available from the Kontes
Glass Co. at a modest price (Figure 6-L).
6W 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 percent diethyl ether-petroleum ether Florisil eluate) or
2 percent acetone in heptane (15 percent diethyl ether fraction).
Detection is provided by spraying with AgNC>3-2-phenoxyethanol reagent
in ethanol or acetone and exposing to high intensity short-wave UV
light to produce brown to purplish-black spots. The construction of
a UV 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 grey and thereby reduce the contrast of the spots
with background. A sensitivity in the 5-500 ng range is possible with
AgN03 reagent, with a light steaming before spraying often aiding the
detection. Conventional 8 inch by 8 inch 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 Rp values for numerous
225
-------
TABLE 6-2
Compound*
Captan, captafol
Benomyl, carben-
dacltn, and 2-AB
Herbicldea contain-
ing NHj or OH groupa
s-Triniines
Fenitrothion
Chlorophenoxy
acid herbicides
OC1 paaclcldes
OP pesticides
Acidic herbicides
OP Insecticides
Coumaphoa
OP pesticides
Maretln
Qulnonethionate
Coumaphos and
0-analog
Thlabendazole
Chloramben
Bayrusll
DDT
Carbaryl
PESTICIDES QUANTITATET) BY '
Sample matrix
apple, potato,
fruits, vegetables
water, aoll
standards only
water
water
human aufop«y samples
tiisuea
standards only
standards only
water
water
milk, egga
crops
egga
fruits
bean, tomato
foods
water
potato
THIN LAYER DENSITOMETRY
Scanning mode
fluorescence
quench
fluorescence
quench
fluorescence
visible
visible
visible
fluorescence
visible
fluorescence
fluorescence
fluorescence
fluorescence
fluorescence
fluorescence
visible
fluorescence
visible
visible
Detection method
NaClOj
fluorescent layem
dsnsyl chloride
fluorescent layers
SnClj/fluoreacamlne
AgN03
AgN03
palladium chloride
4-brnmoethyl-7-
methoxycoumarln
AgNOj or enzyme
inhibition
heating
hydrolysla/dansyl
chloride
~
~
hentlng
—
Brat ton-Marshal 1
reagent
heating
AgN03
j>-nitrobenzenedi-
nzoniumfluoborate
Referenct
(13fi)
(137)
(138)
(139)
(140)
(141)
(142)
(143)
(144)
(145)
(146)
(147)
(148)
(149)
(150)
(151)
(152)
(153)
(154)
(154)
226
-------
Section 6W
compounds in the aforementioned two solvent systems, as well as for
an alternate 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 coating the plates so
that only exposure to UV light is required for spot visualization
(FDA PAM, Section 412). The AgN03 detection method has recently
been studied in detail for the determination of chlorinated insecticides
and herbicides (155).
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 grams 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 gram sample, concentrated to 500
yl, will give a readable spot at 10 ppb. The method involves TLC of
the 6 percent and 15 percent Florisil column eluates as above, with
additional prior cleanup of the 15 percent fraction on an alumina micro-
plate developed with acetonitrile.
Silica gel layers developed with hexane, 1 percent acetone-hexane, 10-50
percent benzene-hexane, or 1 percent ethanol-hexane are recommended for
screening chlorinated pesticides in foods at 0.1 ppm levels [Canadian
PAM, Procedures 0.1 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.
For complex pesticide mixtures, two dimensional or multiple development
techniques may be helpful. The former was used to identify organochlorine
pesticides in blood and tissues (156) and the latter (157) for the separation
of 13 common pesticides.
Extensive listings of additional solvent systems and corresponding tL,
values for chlorinated pesticides will be found in references (123, 158,
and 159). The thin layer densitometry of chlorinated pesticides after
spraying with silver nitrate reagent is described in Chapter 15 of refe-
rence (133).
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, AgNO_, 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 £-nitrobenzyl pyridine and tetraethylpentamine spray (FDA PAM,
Section 432].
227
-------
Section 6W
A two dimensional procedure (160) has the significant advantage of
specificity, obtained by oxidation of the OP pesticides with bromine
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 (Subsection 7Q in Section 7) and enzymatic detection
to identify 18 pesticides in crops at 0.01 ppm levels. 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 urn), 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 substrate 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 details, tables of
sensitivities and effects of bromine and UV treatment for OP and
carbamate pesticides, and diagrams of mobilities with hexane-acetone
(8:2 v/v), a generally useful development solvent for TLC on 450 um
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 OP and carbamate pesticides in
vegetables and fruits (161).
TLC enzyme inhibition methods and applications to pesticides have been
reviewed (162-164) as have the merits of TLC for analysis of residues
(165). The separation and detection of 42 phosphate compounds using
five ternary solvent systems on three adsorbents and three selective
chromogenic sprays have been reported (166). Twenty five solvent
systems and several visualization reagents were evaluated for detection
of 12 OP insecticides in tissues (167).
228
-------
Section 6W
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 Rp 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 (168).
d. Other Pesticide Classes
The TLC of other classes of pesticides including carbamates, ureas,
phenols, dithiocarbatnates, triazines, and organomercurials was reviewed
in references (123) and (124). 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
(139, 169) and polyamide (170); determination of 11 urea herbicides
in water (171); detection of dithiocarbamate fungicides with congo red
(172); and separation of carbamate and phenylurea pesticides <->n
polyamide (173).
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 with benzene for dimethyldithiocarbamates or acetic
acid-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.
229
-------
Section 6X
6X REFERENCES
(1) Ford, J. H., McDaniel, C. A., White, F. C., Vest, R. E., and
Roberts, R. E., J. Chromatogr. Sci., 13, 291 (1975).
(2) Kawar, N. S., deBatista, G. C., and Gunther, F. A., Residue Rev.
4J5, 45-77 (1973).
(3) Arthur, R. D., Cain, J. D., and Barrentine, B. F., Bull. Environ.
Contam. Toxicol.. L5, 129 (1976).
(4) 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, West Palm Beach, FL.
(5) Mann, J. B., Enos, H. F., Gonzalez, J., and Thompson, J. F.,
Environ. Sci. and Technol., J3, 584 (1974).
(6) Compton, B., Bazydlo, B., and Zweig, G., Field Evaluation of Methods
of Collection and Analysis of Airborne Pesticides, prepared under
contract CPA 70-145 for the U.S. Environmental Protection Agency,
Research Triangle Park, NC, May 1972 (NTIS no. PB-214008); Compton, B.,
and Bjorklund, J., Design of a High-volume Sampler for Airborne
Pesticide Collection, presented at the 163rd Natl. ACS Meeting,
Boston, MA, April, 1972.
(7) Aue, W. A., and Teli, R. M., J. Chromatogr., j>2_, 15 (1971).
(8) Ciccioli, P., Bertoni, G., Brancaleonic, E., Fratarcangeli, R.,
and Bruner, F., J. Chromatogr., 126, 757 (1976).
(9) Versino, B., de Groot, M. , and Geiss, F., Chromatographia, J7_,
302 (1974).
(10) Johnson, E. R., Yu, T. C., and Montgomery, M. L., Bull. Environ.
Contam. Toxicol. , .17, 369 (1977).
(11) Tessari, J. D., and Spencer, D. L., J. Ass. Offic. Anal. Chem.,
54., 1376 (1971).
(12) Winnet, G., and Siewierski, M., Bull. Environ. Contam. Toxicol.,
3.4, 681 (1975).
(13) Beyermann, K., and Eckrich, W., Z. Anal. Chem., 265, 4 (1973);
269, 279 (1974).
(13a) Bosin, W. A., and Lissick, M. 0., Air Sampler Field Tests, Malathion,
Cavalier, ND, September 24-25, 1974, Environmental Research Corporation
Technical Report No. 8-7582, St. Paul, MN, November 1974.
(13b) Jackson, M. D., and Lewis, R. G., Effect of Physical Form on Volatili-
zation of Methyl Parathion from Treated Fields, presented at the 172nd
Natl. ACS Meet., San Francisco, CA, August 1976 (in press, Bull. Environ.
Contam. Toxicol.)
(13c) Thomas, T. C., and Jackson, J. W., An Evaluation of a Commercial Instrumen
for Chlordane and Heptachlor Sampling, U.S. Air Force Environmental
Health Laboratory, Report No. 75M-12, McClellan Air Force Base, CA,
August 1975.
230
-------
Section 6X
(14) Miller, C. W., Shafik, T. M., and Biros, F. J., Bull. Environ.
Contain. Toxicol.. j}, 839 (1972). ~
(15) 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).
(16) Lewis, R. G., Brown, A. R., and Jackson, M. D., Anal. Chem.. 49,
1668 (1977). ~
(17) Turner, B. C., and Glotfelty, D. E., Anal. Chem.. 49. 7 (1977).
(18) Eichelberger, J. W., and Lichtenberg, J. J., Environ. Sci. and
Technol.. 5., 541 (1971).
(19) Wilson, A. J., Bull. Environ. Contam. Toxicol.. 15., 515 (1976).
(20) Braus, H., Middleton, F. M., and Walton, G., Anal. Chem.. .23,
1160 (1951).
(21) Sproul, 0. J., and Ryckman, D. W., J. Water Pollut. Contr. Fed.,
.33, 1188 (1961).
(22) Ruzicka, J. H. A., and Abbott, D. C., Talanta. 20, 1261 (1973).
(23) Musty, P. R., and Nickless, G., J. Chromatozr., 120, 369 (1976).
(24) Aue, W. A., Kapila, S., and Hastings, C. R., J. Chromatogr., _73,
99 (1972).
(25) Ito, T., Water Resource Research Institute of the University of
North Carolina, Report No. 54, August, 1971.
(26) Gesser, H. D., Chow, A., Davis, F. C., Uthe, J. F., and Reinke,
J>» Anal. Lett.. .4» 883 (1971); Gesser, H. D., Sparling, A. B.,
Chow, A., and Turner, C. W., J. Am. Wat. Wks. Ass., 65. 220 (1973).
(27) Musty, P. R., and Nickless, G., J. Chromatogr., 100, 83 (1974).
(28) Weil, L., Quentin, K.-E., and Gitzowa, S., Gas -u. WassFach. 113,
64 (1972).
(29) Uthe, J. F., Reinke, J., and Gesser, H. D., Environ. Lett.. _3, 117
(1972).
(30) Burnham, A. K., Calder, G. V., Fritz, J. S., Junk, G. A., Svec, H. J.,
and Willis, R., Anal. Chem.. ^4, 139 (1972).
(31) McNeil, E. E., Otson, R., Miles, W. F., and Rajabalee, F. J.,
J. Chromatogr.. 132 277 (1977).
231
-------
Section 6X
(32) Riley, J. P., and Taylor, D., Anal. Chim. Acta, 46. 307 (1969).
(33) Musty, P. R., and Nickless, G., J. Chromatogr., 89, 185 (1974).
(34) 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).
(35) Harvey, G. R., U.S.E.P.A. Report R2-73-177, 1973, 32 pp.
(36) Coburn, J. A., Valdmanis, I. A., and Chau, A. S. Y., J. Ass.
Of fie. Anal. Chem.. j>0, 224 (1977).
(37) Kahn, L., and Wayman, C. H., Anal. Chem., _36L, 1340 (1964).
(38) Goldberg, M. C., DeLong, L., and Sinclair, M., Anal. Chem. , 45,
89 (1973).
(39) Ahnoff, M., and Josefsson, B., Anal. Chem., 46. 658 (1974).
(40) Ahnoff, M., and Josefsson, B., Anal. Chem.. ^8, 1268 (1976).
(41) Wu, C., and Suffet, I. H., Anal. Chem., 49, 231 (1977).
(42) Brodtmann, N. V., J. Am. Wat. Wks. Ass., 67_, 558 (1975).
(43) Miles, J. R. W., Pestic. Monit. J.. 10, 87 (1976).
(44) Miles, J. W., Dale, W. E., and Churchill, F. C., Arch. Environ.
Contain. Toxicol.. .5, 29 (1976).
(45) Wojcik, D. P., Banks, W. A., Hicks, D. M., and Plumley, J. K.,
Florida Entomologist. 55, 115 (1972).
(46) Pesticide Residue Analysis in Water, U.S.E.P.A., Office of Water
Programs, September, 1974, distributed by NTIS (No. PB-238 072).
(47) Analytical Methods for Pesticide Residues in Foods, Canadian
Department of National Health and Welfare, Section 14.4.
(48) Howells, K. J., Shaw, T. C., Rogers, P. P., and Galbraith, K. A.,
Lab. Pract.. 23, 248 (1974).
(49) U.S.F.D.A. Pesticide Analytical Manual, Vol. I, Section 253.
232
-------
Section 6X
(50) Telling, G. M. , Sissons, D. J., and Brinkman, H. W., J. Chromatogr.,
137. 405 (1977).
(51) Hill, B. D., and Stobbe, E. H., J. Agr. Food Chem. , 22, 1143 (1974).
(52) Johnsen, R. E. , and Starr, R. I., J. Agr. Food Chem., 20, 48 (1972).
(53) Voss, G., and Blass, G., Analyst, 98, 811 (1973).
(54) Suffet, 1. H., and Faust, S. D., J. Agr. Food Chem., 2,0, 52 (1972).
(55) Suffet, I. H., J. Agr. Food Chem., 21, 288 (1973).
(56) Suffet, I. H., J. Agr. Food Chem., 21, 591 (1973).
(57) Beroza, M. , Bowman, M. C., and Bierl, B. A., Anal. Chem., 44, 2411
(1972).
(58) Burke, J. A., Mills, P. A., and Bostwick, D. C. , J. Ass. Of fie.
Anal. Chem., ^9, 999 (1966).
(59) Chiba, M. , and Morley, H. V., J. Ass. Of fie. Anal. Chem., 51, 55
(1968).
(60) Beroza, M. , and Bowman, M. C., Anal. Chem.. _39, 1200 (1967).
(61) McKinley, W. P., and Savary, G., J. Agr. Food Chem., 10, 229 (1962).
(62) Storherr, R. W., J. Ass. Offic. Anal. Chem., _55, 283 (1972).
(63) Eisenbeiss, F. , and Seiper, H. , J. Chromatogr.. j53, 439 (1973).
(64) Larose, R. H., J. Ass. Offic. Anal. Chem., 75, 1046 (1974).
(65) Baumann, F. , Pecsar, R. E. , Wadsworth, B., and Thompson, B.,
^ Lab . , p. 97, September (1976).
(66) Snyder, L. R. , J. Chromatogr. Sci. , 15, 441 (1977).
(67) Horvath, C. , Melander, W. , Molnar, I., and Molnar, P., Anal. Chem. ,
14, 2295 (1975).
(68) Majors, R. E. , J. Chromatogr. Sci.. 15_, 334 (1977).
(69) Saunders, D. L., J. Chromatogr. Sci., r>, 372 (1977).
233
-------
Section 6X
(70) Horvath, C., and Melander, W., J. Chromatogr. Sci.. 15, 393
(1977).
(71) Greenwood, J. V,, Spectra-Physics Chromatography Review, _3 (2),
1 (1977).
(72) Snyder, L. R., and Saunders, D. L., J. Chromatogr. Sci.. 2» 195
(1969).
(73) Snyder, L. R., J. Chromatogr.. £2, 223 (1974).
(74) Chamberlain, W. J., Walters, D. B., and Chortyk, 0. T.,
Anal. Chim Acta. j£, 213 (1975).
(75) Berg, J., Varian Instrument Applications. 11(2), 5 (1977).
(76) Bakalyar, S. R., and Henry, R. A., J. Chromatogr.. 126, 327 (1976).
(77) Conlon, R. D., Anal. Chem.. 41., 107A (1969).
(78) Munk, M. N., J. Chromatogc. Sci.. JJ, 491 (1970).
(79) Veening, H., J. Chem. Educ.. .43, 36A (1971).
(80) Taggart, W. P., Industrial Research, February, 1974, p. 76.
(81) Karasak, F. W., Research/Development, June, 1973, p. 52.
(82) Brown, P. R., High Pressure Liquid Chromatography. Academic Press,
N.Y., 1973.
(83) Parris, N. A., Instrumental Liquid Chromatography, Elsevier
Scientific Publishing Co., New York, 1976, 329 pp.
(84) Jackson, M. T., Jr., and Walker, J. Q., J. Chromatogr. Sci..
15, 490 (1977).
(85) Creed, C. G., Research/Development. 27(9). 40 (1976).
(86) Oehler, D. D., and Holman, G. M., J. Agr. Food Chem., 23.
590 (1975).
(87) Moye, H. A., and Scherer, S. J., oral presentation at the
91st AOAC Meeting, October 17-20, 1977, Arlington, VA, Abstract No. 79.
234
-------
Section 6X
(88) Lawrence, J. F., and Leduc, R. , J. Agr. Food Chem., 25, 1362 (1977).
(89) Aten, C. F., and Bourke, J. F., J. Agr. Food Chem.. 15, 1428 (1977).
(90) Farrow, J. E., Hoodless, R. A., Sargent, M., and Sidwell, J. A.,
Analyst. 102, 752 (1977).
(91) Mundy, D. E., Machin, A. F., J. Chromatogr.. 139. 321 (1977);
Mundy, D. E., Quick, M. P., and Machin, A. F., J. Chromatogr.,
121, 335 (1976).
(92) Paschal, D. C., Bicknell, R., and Dresbach, D., Anal. Chem., .49.,
1551 (1977).
(93) Bjornsson, T. D., Blaschke, T. F., and Meffin, P. J., J. Pharm.
Sci., 66., 142 (1977).
(94) Kvalvag, J., Ott, D. E., and Gunther, F. A., J. Ass. Offie. Anal.
Chem.. 60, 911 (1977).
(95) Farrington, D. S., Hopkins, R. G., and Ruzicka, J. H. A.,
Analyst. 102, 377 (1977).
(96) Selim, S., Cook, R. F., and Leppert, B. C., J. Agr. Food Chem., 25,
567 (1977).
(97) Byast, T. H., J. Chromatogr., 134. 216 (1977).
(98) Fasco, M. J., Piper, L. J., and Kaminsky, L. S., J. Chromatogr., 131,
365 (1977).
(99) Lawrence, J. F., J. Agr. Food Chem.. 25_, 211 (1977).
(100) Brinkman, U. A. Th., DeKok, A., DeVries, G., and Reymer, H. G. M.,
J. Chromatogr.. 128. 101 (1976).
(101) Lawrence, J. F., J. Ass. Offie. Anal. Chem., 59, 1066 (1976).
(102) Maeda, M., and Tsuji, A., J. Chromatogr., 120. 449 (1976).
(103) Pryde, A., and Darby, F. J., J. Chromatogr.. 115. 107 (1975).
(104) Pfeiffer, C. D., J. Chromatogr. Sci.. JL4, 386 (1976).
(105) Byast, T. H., and Cotterill, E. G., J. Chromatogr., 104. 211 (1975).
235
-------
Section 6X
(106) Smith, A. E., and Lord, K. A., J. Chromatogr.. 107. 407 (1975).
(107) Whitacre, Atallah, Y. H., Forrette, J. E., and Suzuki, H. K.,
in Analytical Methods for Pesticides and Plant Growth Regulators
Volume 10, p. 367 (1978).
(108) Horgan, D. F., Jr., Chapter 2 in Analytical Methods for Pesticides
and_ Plant Growth Regulators, Vol. VII, Sherma, J., and Zweig, G.,
eds., Academic Press, N.Y., 1973.
(109) Moye, H. A., J. Chromatogr. Sci., 13, 268 (1975).
(110) Frei, R. W., and Lawrence, J. F., J. Chromatogr., JJ3, 321 (1973).
(Ill) Self, C., McKerrell, E. H., and Webber, T. J. N., Proc. Anal.
Div. Ghent. Soc., _12, 288 (1975).
(112) McFadden, W. H., Bradford, D. C., Games, D. E., and Gower, J. L.,
Amer. Lab., October, 1977, p. 55; McFadden, W. H., Schwartz, H. L.,
and Evans, S., J. Chromatogr., 122, 389 (1976).
(113) Elbert, S., Gruhn, B., Wipfelder, E., and Heusinger, H., Anal.
Chem.. 4_8, 1270 (1976).
(114) Mills, P. A., J. Ass. Offie. Agr. Chem., 4_2!, 734 (1959); 44, 171 (1961)
(115) Mitchell, L. C., J. Ass. Offic. Agr. Chem.. 40, 999 (1957).
(116) Sherma, J., and Zweig, G., Paper Chromatography, Academic Press,
N.Y., 1971, Chapter 12 (Pesticides), pp. 355-396.
(117) Getz, M. E., Advances in Chemistry Series 104. ACS, 1971, Chapter 8.
(118) Getz, M. E., Residue Reviews, 2:, 9 (1963).
(119) Coffin, D. E., J. Ass. Offic. Anal. Chem., 4_9_, 1018 (1966).
(120) Heatherington, R. M., and Parouchais, C., J. Ass. Offic. Anal. Chem.,
J53, 146 (1970).
(121) Stahl, E., Thin Layer Chromatography, 2nd ed., Springer-Verlag, N.Y.,
1969.
(122) Kovacs, M. F., Jr., J. Ass. Offic. Agr. Chem., 46. 884 (1963); ^7,
1097 (1964).
236
-------
Section 6X
(123) Sherma, J., Chapter 1 in Analytical Methods for Pesticides and
Plant Growth Regulatorst Vol. VII, Sherma, J., and Zweig, G.,
eds., Academic Press, N.Y., 1973.
(124) Sherma, J., Critical Reviews of Analytical Chemistry, CRC Press,
Aug. 1973, pp. 333-338.
(125) Reichling, J., Z. Anal. Chem.. 281. 139 (1976).
(126) Vitek, R. K., and Kent, D. M., Amer. Lab.. 10(1), 71 (1978).
(127) Halpaap, H., and Krebs, K.-F., J. Chromatogr., 142. 823 (1977).
(128) Faber, D. B., J. Chromatogr.. 142. 422 (1977).
(129) Lawrence, J. F., and Frei, R. W., Chromatogr. Rev., 18, 253 (1974).
(This is a review of theory, factors affecting emission, and
applications of fluorometric derivatization for pesticide
analysis by HPLC and TLC).
(130) Vitek, R. K., Seul, C. J., Baier, M., and Lau, E., Amer.
Lab., February, 1974, p. 109.
(131) Emanuel, C. F., Anal. Chem.. 45, 1568 (1973).
(132) Getz, M. E., J. Ass. Offic. Anal. Chem.. 54, 982 (1971).
(133) Touchstone, J. C., ed., Quantitative Thin Layer Chromatography,
Wiley-Interscience, N.Y., 1973.
(134) MacNeil, J. D., and Frei, R. W., J. Chromatogr.Sci., JL3, 279 (1975),
(135) Beroza, M., Hill, K. R., and Norris, K. H., Anal. Chem., ^0, 1608
(1968).
(136) Francoeur, Y., and Mallet, V., J. Ass. Offic. Anal. Chem., 60,
1328 (1977).
(137) Polzhofer, K., Z. Lebensm. -Unters. Forsch., 163, 109 (1977).
(138) Pribyl, J., and Herzel, F., Z. Anal. Chem., 286, 95 (1977).
(139) Jork, H., and Roth, B., J. Chromatogr., 144, 39 (1977).
(140) Berkane, K., Cassie, G. E., and Mallet, V. N., J. Chromatogr., 139,
386 (1977).
237
-------
Section 6X
(141) Sherma, J., and Koropchack, J., Anal. Chim. Acta. 91, 259 (1977).
(142) Tewari, S. N., and Sharma, I. C., J. Chromatogr., 131, 275 (1977).
(143) Tewari, S. N., and Harpalani, S. P., J. Chromatogr., 130. 229 (1977).
(144) Dunges, W., Chromatographia, £, 624 (1976).
(145) Stefanac, Z., Stengl, B., and Vasilic, Z., J. Chromatogr., 124,
127 (1976).
(146) Volpe, Y., and Mallet, V. N., Anal. Chim. Acta, 81, 111 (1976).
(147) Lawrence, J. F., Renault, C., and Frei, R. W., J. Chromatogr., 121,
343 (1976).
(148) Zakrevsky, J. G., and Mallet, V. N., J. Agr. Food Chem., 23^, 754 (1975)
(149) Francoeur, Y., and Mallet, V. N. , J. Ass. Offic. Anal. Chem., 59_,
172 (1976).
(150) Zakrevsky, J. G., and Mallet, V. N., J. Ass. Offic. Anal. Chem., _58,
554 (1975).
(151) Ottender, H., and Hezel, U., J. Chromatogr.. 109. 181 (1975).
(152) Sherma, J., and Touchstone, J. C., Chromatographia, JJ, 261 (1975).
(153) Zakrevsky, J. G., and Mallet, V. N., Bull. Environ. Contain. Toxicol.,
13, 633 (1975).
(154) Sherma, J., Klopping, K., and Getz, M. E., Amer. Lab., December, 1977,
p. 66.
(155) Ludwick, A. G., Lau, A. N. K., and Ludwick, L. M., J. Ass. Offic.
Anal. Chem.. 6>0, 1077 (1977).
(156) Eliakis, C. E., Coutselinis, A. S., and Eliakis, D. C., Analyst. £3,
368 (1968).
(157) Szokolay, A., and Madaric, A., J. Chromatogr.. .42, 509 (1969).
(158) Walker, K. C., and Beroza, M., J. Ass. Offic. Agr. Chem., ^6, 250 (1963.
(159) Thielemann, H., Z. Chemie Lpz.. 14, 292 (1974).
238
-------
Section 6X
(160) Gardner, A. M., J. Ass. Offic. Anal. Chem., J54, 517 (1971).
(161) Ernst, G. F., Pieterse, C., and Marterns, L. J. H., J. Chromatogr..
133, 245 (1977).
(162) Mendoza, C. E., Residue Reviews. 43, 105 (1972) and 50, 43 (1974);
J. Chromatogr.. 78, 29 (1973).
(163) Gardner, A. M., FDA By-Lines. _2(3), 173, Jan., 1972.
(164) Villeneuve, D. C., Advances in Chemistry Series 104. ACS, 1971,
Chapter 3.
(165) Abbott, D. C., and Thompson, J., Residue Reviews, 11, 1 (1965);
Abbott, D. C., and Egan, H., Analyst. £2, 475 (1967).
(166) Getz, M. E., and Wheeler, H. G., J. Ass. Offic. Anal. Chem., 51,
1101 (1968).
(167) Tewari, S. N., and Harpalani, S. P., J. Chromatogr.. 130. 229 (1977)
(168) Thielemann, H., Z. Anal. Chem., 272. 286 (1974).
(169) Van den Heede, M., and Heyndricks, A., Meded. Fac. Landbouwvet.
Rijksuniv. Gent. 41, 1457 (1976) (Analytical Abstracts, abstract
5G20, November, 1977; Chem. Abst. 86, 84568c, 1977).
(170) Reichling, J., and Fischer, H., Z. Anal. Chem.. 115. 670 (1975).
(171) Deleu, R., Barthelemy, J.-P., and Copin, A., J. Chromatogr., 134,
483 (1977).
(172) Ragab, M. T. H., Anal. Lett.. 9^, 295 (1976).
(173) Reichling, H., and Fischer, H., Z. Anal. Chem.. 276, 301 (1975).
239
-------
Section 7
MULTIRESIDUE EXTRACTION AND ISOLATION PROCEDURES
FOR PESTICIDES AND fOABOLITES
This section will present brief descriptions and problem areas of widely
used multiresidue analytical procedures for different sample substrates.
A few methods for important individual residues will also be included.
Many of these problems are treated elsewhere in this Manual but are
highlighted again here to alert the analyst using the different methods.
References will be given in each case to sources of detailed methodology.
Control of procedures for collection of samples for these methods is
covered in Subsections 6A - 6K in Section 6, while general comments
on sample extraction and extract concentration will be found in Sub-
sections 6L and 6M in Section 6.
CHLORINATED PESTICIDES
7A TISSUE AND FAT ANALYSIS BY A MODIFIED MILLS, ONLEY, GAITHER PROCEDURE
The macro method described in Section 5,A,(1) 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 car-
bamate pesticides are not recovered. This method involves grinding of
a 5 gram sample with sand and anhydrous Na SO , isolation of fat by
repeated extraction with petroleum ether, re-extraction of residues
into acetonitrile and then partitioning back into PE after adding 2
percent NaCl, drying by elution through a column of Na SO , concentration
of the eluate, cleanup on a Florisil column, and EC-GC after reconcen-
tration of column eluates. If necessary, further cleanup of the 15
percent 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].
In addition to being complex, this method is highly empirical in that
failure to follow the methodology as described may lead the analyst into
problems of residue identification and quantitation. Some of the method
pitfalls observed during the operation of the interlaboratory quality
control program (Section 2) are as follows:
240
-------
Section 7A
a. Some analysts, with the mistaken notion of saving time, have combined
6 percent and 15 percent ethyl ether-petroleum ether Florisil column
fractions, and have then attempted gas chromatography on the mixture.
With some luck this might prove successful, but there is a high pro-
bability of being unlucky. 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 but
in the 15 percent eluate, negating the possibility of aldrin since
this compound elutes wholly in the 6 percent fraction. By combining
the fraction extracts, the analyst inadvertently neglected one valuable
identification tool, that of selective adsorption.
b. The polarity of the ethyl ether-petroleum ether eluting solutions
exerts a profound effect on the elution pattern of several pesticidal
compounds. The amount of ethanol, 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 no ethanol, dieldrin
would be expected to yield only 87 percent recovery in Fraction II
with the balance being retained on the column. If twice the proper
amount of ethanol is present, approximately 7 percent should elute
in Fraction I, giving a 93 percent recovery in Fraction II. If 2
percent 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 percent fraction.
c. The activity characteristics of Florisil may vary somewhat from lot to
lot. Each lot, when received at a laboratory, should be carefully
evaluated to be certain the compound elution characteristics are
satisfactory.
d. Storage and holding temperature of Florisil are critical. The oven
used for holding this (and other adsorbents) should be confined exclu-
sively to this usage and not used as an all-purpose drying oven.
Florisil will readily pick up air-borne contaminants which may result
in spurious^ chromatographic peaks. If the oven temperature varies
more than - 1°C, considerable influence may be observed in the reten-
tion characteristics. The recommended activation temperature is
130°C.
e. Anhydrous Na SO 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
situation, it is good practice to Soxhlet extract all lots of this
salt before use.
f. The presence of peroxides in ethyl ether can result in extremely
low recoveries of organophosphorus compounds and also poses a serious
241
-------
Sectons 7B,C
safety hazard. Methods have been set forth for the removal of
peroxides from ether but have not proven wholly satisfactory.
The purity of petroleum ether is also critical and may exert a
profound effect on the recovery of certain of the organophosphorus
compounds.
g. Glassware must be meticulously cleaned to remove electron capturing
contaminants. Reagent blanks must be run with each set of samples.
h. Most chlorinated pesticides should be recovered in the range of 85-100
percent. 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 percent.
See also Subsection 7K for further comments on pesticide elution
from Florisil.
7B HUMAN OR ANIMAL TISSUE AND HUMAN MILK ANALYSIS BY THE 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 percent methanol in hexane; eluate II: 12 ml of 1 percent
methanol in hexane), and concentrated fractions determined by EC-GC.
Several pesticides, including a-BHC, lindane, diazinon, DDD, and toxa-
phene, split between fractions. Florisil columns must be conditioned at
130°C at least overnight before using. Precautions concerning use of
Florisil are similar to those outlined in Subsection 7A. Virtues of
the micromethod include a low background level and savings in the volume
of solvent required.
7C HUMAN BLOOD OR SERUM
A 2 ml aliquot of serum is extracted with 6 ml hexane for 2 hr 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 Table 2-11 in Section 2). Since all pesticides
242
-------
Section 7D
will be present in one extract, a GC column must be chosen which will
separate the expected pesticides. Certain serum samples will yield a
very late eluting extraneous peak (probably a phthalate) which 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 (1). Microcoulometric GC determination
after sulfuric acid extraction was successfully applied to 24 organo-
chlorine pesticides in blood at 1 ppb levels with no cleanup (2).
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 (3).
Aaother 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 (4).
7D PENTACHLOROPHENOL (PGP) 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 made alkaline and extracted with
hexane to remove basic and neutral interfering compounds (e.g., DDT,
BHC), followed by acidification, hexane extraction, methylation, and
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,
diazopentane) have been widely used in pesticide residue analysis
and are cited in several procedures in this Manual and the EPA PAM.
These compounds 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 7AA). 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.
243
-------
Section 7D
b. 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 PGP methyl esters are
identical and these pesticides would not be differentiated.
c. 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.
d. Glassware should be washed with dilute NaOH followed by deionized
water and acetone.
e. 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 PGP.
f. Other ether derivatives (e.g., ethyl, propyl, amyl, etc.) can be
prepared and characterized for confirmation of PGP identity.
g. A Vortex-Genie or similar mixer should be available for efficient
extraction and quenching of the derivatization reaction.
An improved EC-GC method for determination of PGP in urine and blood
has been worked out as part of a multi-phenol analytical procedure. A
hydrolysis step is included to increase recovery of PCP compared to
extraction with hexane only. A brief outline of the full procedure is
given below; further details can be obtained from Dr. R. Moseman, U.S.E.P.A.
Health Effects Research Laboratory, Research Triangle Park, NC 27711:
a. Acidify 2 ml of urir.e with 0.5 ml cone. HC1 or 1 gram of blood
serum with 3 ml of 6N HC1 in a 125 mm screw cap culture tube.
b. Seal the tube and place in a boiling water bath for 1 hour with
periodic shaking.
c. Extract the cooled sample twice for one-hour periods on a mechanical
rotator at 30-50 rpm with 5 ml portions of benzene for urine or with
hexane/ethyl ether (1:1) for blood.
d. Centrifuge after each extraction and concentrate the combined
extracts to 0.3-0.5 ml.
244
-------
Sections 7E, F
e. Methylate with 5 ml of diazomethane reagent. (See footnote on
page 4 when using diazoalkane reagents.)
f. Allow the methylated extract to stand for one hour and concentrate
to 0.3 ml under a gentle stream of nitrogen.
g. Add 2 ml of hexane and reconcentrate to 0.3 ml.
h. Chromatograph the methylated sample through a size 22-9 Kontes
Chromaflex column containing 4 grams of acid alumina topped by
1.6 grams of Na2SO^. Elute the column with a total of 5 ml of
hexane (including the volume - ca. 3 x 0.5 ml - used for rinsing
the sample onto the column) and discard eluate; then elute the PGP
methyl ester with 20 ml of benzene-hexane (10:90 v/v). (The
prepared column is washed with 30 ml of benzene-hexane (40:60 v/v)
to remove impurities, air dried, and stored in an oven at 130°C
overnight prior to use.)
i. Concentrate the eluate to an appropriate volume and compare to
similarly treated PGP standards using EC-GC with standard columns
and conditions described in Section 5 of this Manual and Sections
5,A,(3),(b) and 5,A,(4),(a) of the EPA PAM.
7E BIS (p_-CHLOROPHENYL) ACETIC ACID (p_,p_'-DDA) IN HUMAN URINE
The excretion level of this metabolite is a sensitive indicator of
exposure to p_,p_'-DDT. Urine is extracted three times with an equal
volume of 2 percent 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 BF -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 7B) is required when the
poorly selective EC detector is used. DDA should elute completely in
Fraction II. Concentration and injection volumes depend upon the sensi-
tivity 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.
7F 2,4-D AND 2,4,5-T IN URINE
A method is described in the EPA PAM, Section 5,A,(4),(c), for deter-
mining 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 percent water), and determination of concentrated eluates by EC-GC
on a 4% SE-30/6% OV-210 column.
245
-------
Section 7G
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 com-
pounds of interest elute in a later fraction than that indicated in
the detailed procedure, or, the percentage of benzene in the benzene-
hexane eluent 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 4 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.
A multiresidue scheme for phenol metabolites and including 2,4-D,
2,4,5-T, and silvex is discussed in Subsection 70.
7G KEPONE IN HUMAN BLOOD AND 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 et al. (5). Samples are
extracted, and the extracts are cleaned up by chromatography on a
micro Florisil column, base partitioning, or gel permeation chroma-
tography. Kepone is determined by EC-GC and confirmed by chemical
conversion to mirex (6), detection with a Hall conductivity detector
in the Cl-mode, or chemical ionization mass spectrometry.
It is mandatory to use 1-2 percent methanol in benzene for all sample
and standard solutions injected for EC-GC to obtain the maximum re-
producible 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.
The analysis of field-collected avian tissues and eggs for Kepone
residues has been reported (7). Samples were extracted with benzene-
isopropanol (2:1 v/v) and extracts cleaned up with fuming H2S04~con-
centrated 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 percent at 1 ppm. Procedures for
246
-------
Section 7H
determination of Kepone in serum, plasma, urine and fat have been re-
ported. After addition of H2S04, 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) (8).
7H ANALYSIS OF FATTY AND NONFATTY FOODS BY THE MOG METHOD
The macro Florisil column method for determining nonionic chlorinated
pesticides in fatty foods is similar to that outlined in Subsection 7A
and is described in detail in Sections 211 and 231 of the FDA PAM.
Eluents are 6, 15, and 50 percent ethyl ether in petroleum ether. The
method for nonfatty foods (FDA PAM, Sections 212 and 232) involves
extraction of pesticides with acetonitrile or water-acetonitrile 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 179 pesticides and other chemicals are given in
Table 201-A, and over 200 compounds have been tested (9)), crops to
which they are applicable, and supplemental cleanup procedures for the
Florisil column fractions. This AOAC multiresidue method is currently
official for 25 OC1 and OP pesticides and PCBs (9). The problem areas
are the same as those given in Subsections 7A and 7K. The elution
pattern of more than 150 pesticides from the U.S.F.D.A. Florisil column
eluted with 6, 15, 20, 30, 50, and 65 percent diethyl ether in
petroleum ether is tabulated in Section 7.2(b) of the Canadian PAM.
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 eluents. These
eluent 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 (10). A rapid screening
method for residues in milk is another modification of the standard
procedure. Pesticides are cleaned-up by partioning, but Florisil
chromatography is excluded (11).
247
-------
Sections 71, J
71 DETERMINATION OF CHLOROPHENOXY HERBICIDES IN FATTY AND NONFATTY FOODS
Section 221 of the FDA PAM covers the analysis of chlorophenoxy acid
herbicides in fatty foods. The pesticides are extracted from the foods,
partitioned into alkaline solution, the extract is washed with or-
ganic solvents and acidified, and the pesticides are extracted into
chloroform, methylated with diazomethane, and determined by microcoulo-
metric GC. The recovery of eight compounds from oils, dairy products,
and animal tissues has been verified. The cleanup procedure is not
sufficient for determination by EC-GC. Methylated acids are easily
lost during evaporation steps, especially from standard solutions.
Solvents should be evaporated very carefully in a gentle stream of nitroger
Traces of soap which are not rinsed away can give interference peaks
even with the selective microcoulometric detector.
Chlorophenoxy acids and esters are extracted from nonfatty foods with
mixed ethers under acid conditions, extracts are shaken with sodium
carbonate solution, and esters are extracted with ether and deter-
mined by microcoulometric GC after cleanup on a Florisil column. The
carbonate solution is acidified to reform the free acids, which are
extracted into chloroform, methylated (see the footnote on page 4
concerning the hazards of diazoalkanes), and determined by microcoulo-
metric GC without column cleanup (FDA PAM, Section 222). The re-
covery of residues of eight compounds in grains and vegetables has
been verified. Sodium sulfate used in drying steps may adsorb some
free chlorophenoxy acids, and the reagent should be checked for
complete recovery. An alternative alumina and Florisil column cleanup
prior to EC-GC for the ester fraction from vegetable samples is given
in the Canadian PAM, Section 7.3.
A more recent method for determining the acid herbicides in foods and
including Florisil cleanup has been reported (12). Residues are ex-
tracted with acetonitrile (nonfatty foods) or chloroform (fatty foods
and liquids); the acids are converted to their sodium salts, acidified,
methylated, cleaned-up on a Florisil column, and determined by EC- or
microcoulometric GC. Severe emulsions form during cleanup of the basic
phase by extraction with chloroform and must be broken by centrifugation.
7J CARBON-CELLULOSE COLUMN CLEANUP
Section 7.1 of the Canadian PAM describes a method for cleanup of
residues of organochlorine and organophosphorus insecticides, herbicides,
and fungicides in foods following acetonitrile extraction (blending) and
hexane partition. Sequential elution with three solvents (1.5 percent
acetonitrile in hexane, chloroform, and benzene) separates the pesticides
into three fractions which are suitable for EC-GC and TLC determination.
248
-------
Section 7K
Some 40 pesticides have been quantitatively recovered from a variety
of foods with this system. A carbon-cellulose (4:10 w/w) minicolumn
eluted with ethyl acetate has been used for cleanup and separation
of phorate and metabolite residues in crops determined by FPD-GC (13).
The Canadian PAM specifies that carbon (e.g., Darco G60) is prepared
by heating for 12 hours at 300°C and then extracting for two one hour
periods with hexane on a mechanical shaker. See also Section 4Af
in this Manual. Cellulose (e.g., Solka Floe BW40) is extracted twice
with acetone in a similar manner without prior heating.
7K CLEANUP ON DEACTIVATED FLORISIL
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 pesti-
cides on Florisil deactivated with 2 percent water has been deter-
mined 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 percent methylene chloride in hexane to 5-30 percent
ethyl acetate in hexane (Table 7-1). If the analyst wishes to screen
a sample extract for a larger number of pesticides in one or two GC
injections, the less polar eluents 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,
chlorobenside, disulfoton, and demeton have losses ranging from
20-100 percent. 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 percent deactivated Florisil column can tolerate up to one
gram of fat or oil (30 percent methylene chloride in hexane or
less polar eluents) 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 percent methylene chloride in hexane
to recover BHC isomers, the DDT group, PCBs and HCB.
249
-------
Table 7-1
Section 7K
ORDER OK ELVTION OP PESTICIDES PROM FLORISIL PARTIALLY
DEACTIVATED WITH 2* WATER USINO 300 ml VOLUME OP ELUENTS
(Prom the Canadian PAM)
PESTICIDES
CHgC^ in Hexane
% EtOAo inHexRne
Hexane 551
Aroolor 1254 PCB +
Chlordane +•
Toxaphene +
Strobane +
trans-Chlordana +
CHlordene +
Aldrin +
Hexachlorobenzene +
Heptaohlor +
p. p1 -DDE +
o.p'-DDT +
Mirex +
leobenzan + 2
p.p'-DDT S(90Jt) +
a-BHC S(45#) +
Perthane S(10J6) S(85#
p.p'-DDD +
Chlorbenside M + 2
IT-BHC S(80J5
PCNB S(80#
TCNB S(4555
P-BHC S(10#
f-BHC
Dicofol
Ronnol OP
Hepachlor epoxide
Dichlofenthion OP
Phorate OP^l
Carbophenothion O^M
Endosulfan I
Dieldrin
Chlorpyrifoo OP
Endrin
Methoxychlor
Parathion OP
Ethion2 OP
2,4-D methyl ester
2A5-T methyl ester
Anilazine
Ovex
Fenitrothion OP
Tetradifon
Diazinon OP
Chlorothalonil
Methyl ParathionJ OP
Sulphenone
Dioxathion OP
Malathion OP
Atrazinej*
Simazine^
Endosulfan II
Captan M
Phosmet OP
DCPA
Azinphoismothyl OP
m iflt 2055 30* & 1055 ^ m Reooverioa
>95
s o e
773
' yj
>95
>95
v O t.
>95
>95
>O C
95
>95
>95
>95
>95
>95
>95
+ soe
>95
+ \e\f
. '95
T' *O C
+ 2 '*
T >95
T * rt £
. >95
T ^Q C
T >95
T" v Q C
_,,.,-,. >95
«\oj^j f ^95
S( 60^) + ^ 9 5
* ~ 20
S(25j8) + /^80
S(10J5) + >95
+ >95
•* >95
* >95
* >95
-1- >95
+ >95
+ >95
+ >95
S(90?5) + >95
S(90?5) + >95
S(50j5) + >95
S(15^) + >95
•t- >95
* >95
+ >90
+ >90
S(75?5) + >95
+ >95
+ >90
+ 2 >95
Notei A
fraction was eluted prior to all ethyl acetate fractions. All others were single
elutions.
+ - mostly elutes in first 250 ml
* - large amount in 250-300 ml fraction
S - some (as perceive)
Footnotesi 1 - OP = orfjanophosphorusi M = mercaptam PCB = polychlorinated biphenyl
2 - Higher recoveries arc obtained by elution with more polar e3uents
3 - Remaining methyl parathion elxitos in another 50 ml of y%> EtOAc
4 - Detected by alkali flame detector
250
-------
Sections 7L, M
The latter approach with hexane elution has been used to determine
as low as 5 ppb of HCB in adipose tissue (ref 14; and Section 12,D,(2)
of the EPA PAM). The recovery of HCB and mirex in fish and butterfat
by elution with acetonitrile from a column composed of the fat or
oil distributed on unactivated Florisil has been collaboratively
studied (15). 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 (16).
7L 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 precipitation 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 quantita-
tively (80+ percent) recovered from a variety of plant and animal prod-
ucts 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
(17).
7M MISCELLANEOUS MULTIRESIDUE CLEANUP PROCEDURES
Other frequently applied multiresidue procedures include the following:
The method of de Faubert Maunder (18) employs partition with dimethyl-
formamide (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 percent Na2$04 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.
Wood (19) proposed the use of dimethyl sulfoxide (DMSO) as another good
solvent for chlorinated pesticides because it dissolves low amounts
of oil or fat. The fatty sample is mixed with Celite (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 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 (20). No gross general differences
were found in results, but one method might be advantageous for a parti-
cular sample type.
251
-------
Section 7M
Table 7-2
ORDER OF ELUTION OF ORGANOCHLORINES FROM DEACTIVATED SILICA GEL
ACCORDING TO THE METHOD OF HOLDEN AND MARSDEN [21 ]
Eluted in order
by hexane
Eluted in order by 10%
diethyl ether in hexane
Hexachlorobenzene
Aldrin
PCBs
p_,p_'-DDE
Heptachlor
p_,£'-MDE (DDMU)
£,£' -DDT
Endrin
Chlordane
p_,p_'-DCBP
Toxaphene
p_,p_' -TDE
Telodrin
Heptachlor epoxide
ct-BHC
Perthane
6-BHC
Kelthane
Y-BHC
Dieldrin
Methoxychlor
252
-------
Section 7M
Hexane extracts of animal tissues are cleaned-up and prefractionated
on narrow bore columns dry-packed with partially deactivated alumina
and silica gel by the method of Holden and Marsden (21). The initial
alumina column eluted with hexane provides removal of lipids, while
the second column affords pre-GLC separation of residues plus further
cleanup. Table 7-2 shows the elution order of chlorinated insecticides
with hexane and 10 percent diethyl ether-hexane eluents. Alumina is
activated at 800°C and silica gel at 150°C before deactivation with
5 percent (w/w) water. Another alumina-silica column scheme (22) 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 percent water were used.
Organochlorine insecticide residues in fatty foodstuffs were determined
(23) using a cleanup technique based on a single 22 gram column of
activity-4 alumina. Concentrated hexane extracts of samples, con-
taining 0.4-0.5 grams of fat, were transferred to the column, and
pesticides were eluted with 150 ml of hexane prior to determination
by EC-GC. Recoveries of 15 insecticides from vegetable oil samples
spiked at levels of 5-250 ug/kg were between 70-124 percent. Routine
determinations were carried out for cyclodienes, BHC isomers, and HCB
at the 5-10 ug/kg level and DDT-type compounds at the 20-30 ug/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, j3,jp_'-DDE and £,£'-DDT. The second fraction
contained all the BHC isomers, heptachlor epoxide, dieldrin, endrin,
_p_,jp_'-DDD, methoxychlor, and Endosulfan A. Compounds splitting between
fractions included methoxychlor, toxaphene, perthane, chlordane, and
strobane.
In a comparative study (24), 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 organochlorine pesticides from lipids than
activated adsorbents and more polar eluents. Saponification with
ethanolic NaOH followed by alumina column chromatography provided
efficient removal of lipids prior to GC determination of several OC1
insecticides (DDT was converted to DDE) (25).
A gel permeation chromatography (GPC) system using Bio-Beads S-X2
crosslinked polystyrene gel has been designed by Stalling et_ al. (26)
for removal of lipids from extracts of samples such as fish before
EC-GC determination of commonly occurring pesticide and PCB residues.
253
-------
Section 7M
In a typical cleanup, a fatty tissue sample may be desiccated with
anhydrous Na2SC>4 (e.g., 1:3 w/w) and then extracted with 6 percent
diethyl ether-petroleum ether. The extract is concentrated, brought
to volume with cyclohexane or another suitable solvent, and an aliquot
injected onto the GPC column. The lipids, because of their high
molecular weights, are excluded from the gel pores and elute in the
first fraction which is discarded. The fraction containing the lower
molecular weight pesticides and PCBs elutes later and is collected
and concentrated for GC analysis. Up to 500 mg of lipid can be in-
jected as a 5 ml aliquot onto the GPC column. An evaluation of this
GPC system (27) with different sample types indicated that ca. 98
percent 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.
Analyses can be automated since an important feature of GPC is that
the same column can be used repeatedly over long periods without signifi-
cant change in elution volumes or recoveries. Details of GPC cleanup
are presented in Subsection 8M in connection with the quality assurance
of tissue analysis by GC-MS.
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 percent.
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 (28).
A reuseable, macroporous silica gel column provided fractionation and
88-105 percent recoveries of 0.1-1 ppm levels of different classes of
pesticides when eluted with a series of solvents of increasing polarity
(29).
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 com-
pounds on the same plate to allow location of the pesticide zones.
These bands are removed by scraping and are extracted to recover the
separated pesticides. Modified layers have been devised with capability
for increased sample loading, e.g., multiband or wedge-layer chromato-
plates (30). 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 (31). 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 eluent and determined by EC-GC after appropriate derivati-
zation reactions (32).
254
-------
Section 7N
The results of international cooperative studies of OC1 pesticide,
PCB, and Hg residues in wildlife have been reported (33). The
analytical methods were based on extraction, cleanup, and GC
determination, but no two laboratories used exactly the same pro-
cedure. Nonetheless, there was reasonable agreement among laboratories
in analysis of test samples, the coefficient of variation for
different chlorinated compounds ranging from 10-17 percent.
Collaborative testing of a multiresidue method for chlorinated
hydrocarbon and other fumigant residues among 8 foreign laboratories
was successfully completed, and results were reported (34).
ORGANOPHOSPHORUS (OP) PESTICIDES
7N DETERMINATION OF METABOLITES OR HYDROLYSIS PRODUCTS IN HUMAN
URINE, BLOOD, AND OTHER TISSUES
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 (35) 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 Interfering compounds. The compounds are eluted from the resin,
derivatized with diazopentane (see footnote on page 4 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 percent 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 which 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 OP and phosphonate pesticides (36).
255
-------
Section 70
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 occa-
sional analyses are performed, or the ratio of SPRM to routine analyses
is at least 10 percent when larger numbers are involved. Because of the
instability of urine samples spiked with alkyl phosphates, large samples
of SPRM cannot 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.
Underivatized compounds may accumulate on the GC column after periods
of extended use. Injection of 1 yl 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.
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 (37). Urine
is extracted, the extract is alkylated, and derivatized carboxylic acids
are cleaned up according to a previously published (38) alkyl phosphate
method. Additional cleanup by solvent partitioning with ether and
silica 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.
70 DETERMINATION OF £-NITROPHENOL (PNP) IN URINE
Urinary PNP, the phenolic metabolite of ethyl and methyl parathion, EPN,
nitrofen, etc., can be 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 deriva-
tive [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 pesti-
cides [39]. A one to five ml sample is treated with a 1/5 volume of
concentrated hydrochloric acid and the mixture refluxed at 100°C for
one hour. The phenols are extracted with diethyl ether, ethylated by
reaction with diazoethane, and the ethyl ethers chromatographed on a
silica gel column (2 grams, 2 percent water deactivation). (See the
footnote on page 4 concerning precautions when using diazoalkanes).
Elution with various concentrations of benzene in hexane purifies and
fractxonates the phenolic ethers, which are finally determined by EC-GC.
256
-------
Section 7P
Ten phenols, including the pesticides PCP and DNOC, plus the herbicides
2,4-D, 2,4,5-T, and silvex can be determined by this scheme in one sample.
All halogenated phenols are eluted with 20 percent benzene-hexane,
while nitrophenols and phenoxy acids elute in the 60 and 80 percent
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 (40). After extraction, the sample is reacted with
diazomethane (see footnote on page 4 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 percent were obtained from samples fortified at 0.1-30 ppm levels.
7P ANALYSIS OF FATTY AND NONFATTY FOODS USING FLORISIL CLEANUP
The MOG Florisil procedures described in Subsection 7G are adequate
for determination of some OP pesticides in fatty and nonfatty foods
(FDA PAM, Sections 231 and 232). Malathion and some other OP
pesticides require 50 percent diethyl ether-petroleum ether for
elution from Florisil. This elution, which must be preceded by
elution with the 6 percent and 15 percent eluents, has been found
occasionally to be inconsistent. As was mentioned earlier, OP
pesticides can be lost through degradation 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 substrate.
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 (41-43).
Beckman and Garber (44) 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
(45).
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 (46). The technique was
useful for cleanup of pesticides having favorable £-values (Section 8F)
257
-------
Section 7Q
in a hexane-acetonitrile system, which included dimethoate, temephos
methyl paration, fenitrothion, crufomate, malathion, and parathion.
7Q SWEEP CO-DISTILLATION
Sweep co-distillation has proven to be a simple time saving cleanup
technique which eliminates the need for specialized adsorbents and
large volumes of purified solvents (42, 47-52). 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 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.
Figure 7A 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 (53, 54). 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 rain sweep operation.
Figure 7B shows the appearance of the current 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 one gram samples of fats or oils (operated in vertical
position at 250°C with 600 ml N2/min). 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 with 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.
258
-------
Section 7Q
Figure 7-A. Sweep co-distillation apparatus, schematic diagram.
Copper
tube
Septurru
Septum
Scrubber
tube
Woter
ond ice
both
_ Glass wool
(silomzed)
— 4cm Anakrom
Insulation
^-Asbestos
' ^^E^^Sf*™^*™-—™*™****
\ To pyrometer
JL
Stc
CO
Sll
py
Storherr tube
containing 5-6 inches
silanized glass wool
1L
Nitrogen
-Cooling inlet
coil
7— Adapter
I—Glass wool
(silomzed)
I 19/22
Concentration
tube
Beoker
of woter
Figure 7-B. Sweep co-distillation apparatus, (oven positioned for fruit
and vegetable cleanup), Kontes Glass Co., K-500750.
259
-------
Sections 7R, S
7R 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
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 (55). A collaborative study (56) of
this method for residues of six OP compounds in apples and green beans
verified recoveries between 86 and 125 percent 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.
7S MISCELLANEOUS MULTIRESIDUE 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-10 percent
methanol proved most reliable and efficient (57).
Alumina has not proven totally satisfactory for cleanup of OP
compounds since recovery of the more polar compounds is not complete
(58). Using alumina (activity II to III) and petroleum ether and
3 percent acetone-petroleum ether as eluents, Renvall and Akerblom
(59) 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 (60) and monocrotophos
in tobacco (61) by FPD-GC.
The Abbott et al. method (62), involving cleanup by solvent partition
without column chromatography, has proven adequate for analyses of
seven types of foods for 39 pesticides ami iueiabr-1 jtes 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
260
-------
Section 7S
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 (63). Azinphosmethyl and dimethoate residues in apple
leaves were determined by FPD-GC following ethyl acetate extraction
and cleanup by methylene chloride-water and hexane-acetonitrile
partitionings (64). A multiresidue analysis of 14 pesticides in
natural waters at ppb levels involving extraction and concentration
before FPD-GC has been reported (65).
The elution pattern of a series of representative OP pesticides from
a column (Kontes, Size 22) containing one gram Woelm silica gel
deactivated with 1.5 percent water and prewashed with 8 ml hexane before
applying the sample mixture is as follows:
Eluent Pesticides Eluted
7 ml hexane
8 ml 60% benzene-hexane carbophenothion
8 ml benzene ethyl parathion
8 ml 8% ethyl acetate-benzene malathion, diazinon
8 ml 50% ethyl acetate-benzene paraoxon
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 (66-68) and to separate OP pesticides and metabolites
into groups to facilitate their identification by GC (69). 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 (70).
The cleanup of 22 OP pesticides in 12 vegetable extracts was achieved
by gel filtration chromatography on Sephade'x LH-20. Residues were
determined by GC with thermionic detection at levels of 0.05 to 0.5 ppm
(71). Recoveries ranging from 88-106 percent were reported for disulfoton,
diazinon, methyl parathion, malathion, parathion, dichlorvos, and
fensulfothion in an evaluation study of the automated gel permeation
chromatographic cleanup technique (Section 7M). Biobeads S-X3 gel and
a toluene-ethyl acetate (1:3 v/v) elution solvent were used (72). A
solvent composed of methylene chloride-cyclohexane (15:85 v/v) with
Biobeads S-X3 provided adequate cleanup for EC-GC (no liquid-liquid
partitioning) of 9 OP pesticides and 2 metabolites and 16 nonionic
OC1 pesticides in vegetable oils at 0.05-1.0 ppm. Elution character-
istics of 12 other OP compounds, 12 carbamates, 2 triazines, and
trifluralin were also investigated (73).
261
-------
Section 7S
A rapid, simple approach has been developed for approximating 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 the oxygen analog sul-
fone with m-chloroperbenzoic acid, followed by removal of the acid on an
alumina column and determination of the sulfone by FPD-GC. Quantitative
recoveries of parent pesticides and metabolites from corn, milk, grass,
and feces have been demonstrated [74] • Metasystox-R and its sulfone were
determined in plant and animal tissues and water at 10 ppb levels as the
sulfone after oxidation by KMn04 (75).
A collaborative study by 12 laboratories of the Abbott e£ _al. (62; see
above, this Subsection), Watts et^ _al., and Sissions and Telling methods
was conducted for OP pesticides in fruits and vegetables (76). The
Abbott et al. method was found satisfactory for determination of
malathion, dichlorvos, dimethoate, omethoate, and parathion in 6 fruit
and vegetable crops (>-90 percent 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 Watts et_ _al. method (77), 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 7R), was found satisfactory for the
same pesticides plus azinphosmethyl in 6 crops (>90 percent average
recovery at 0.5-2 mg/kg) and was also judged to be much more
widely applicable. The Sissions and Telling method (78), 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
collaborative study.
A tabulation has been made (79) of the validated applicability of five
multiresidue analytical procedures 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 oscillographic polarographic confirmatory
method; the Abbott e£ al. method for total diet studies (62); and an
undescribed German procedure (80). In addition, individual determinations
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.
262
-------
Sections 7T, U, V
CARBAMATE PESTICIDES AND MISCELLANEOUS HERBICIDES
7T 1-NAPHTHOL IN URINE
Humans exposed to the N-methyl oarbamate insecticide carbaryl excrete
in urine relatively large quantities of the metabolite 1-naphthol con-
jugated as either the sulfate or glucuronide. Determination of 1-naph-
thol is made by subjecting 5 ml of urine to acid hydrolysis under reflux
to break conjugates, extracting the 1-naph'thol with benzene, and
derivatizing with chloroacetic anhydride solution. After cleanup on a
small silica gel column (1 gram, 1,5 percent 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
f.bout 6 months if stored in a refrigerator.
7U DETERMINATION OP N-METHYLCARBAMATE INSECTICIDES IN BLOOD SERUM AND
HUMAN FAT
An unpublished procedure developed in the Methods Development Section
of the Analytical Chemistry Branch, EPA, Health Effects Research Labora-
tory, Research Triangle Park, N.C., simultaneously determines 0.04 to
1 ppm of propoxur, Landrin, carbofuran, aminocarb, mexacarbate, carbaryl,
and methiocarb in human serum and fat. The pesticides are extracted
with acetonitrile, partitioned into methylene chloride, and cleaned-up
by gel permeation and adsorption (silica gel or charcoal) column chroma-
tography. The purified, intact carbamates are derivatized with penta-
fluoaopropionlc anhydride and determined by EC-GC. Details are available
from the above address.
7V . 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 (81). Pentafluoropropionic anhydride was the pre-
ferred derivatization reagent for the aniline compounds, with cleanup
on 1 gram deactivated (3 percent water) silica gel columns. Determina-
tion was by EC-GC on a 3% OV-1 column. Recoveries ranged from 85-90
percent at 1.0 and 0.1 ppm.
263
-------
Sections 7W, X
7W 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
1-fluoro-2,4-dinitrobenzene. The latter reagent is used for derivatization
of carbamate insecticides in the methods for air analysis (Section 7Z) and
water analysis (Section 7AA) discussed in this Manual and described in
detail in the EPA PAM.
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 (82) 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, 1-fluoro-2,4-dinitrobenzene, and penta-
fluoropropionic anhydride. These reactions are also reviewed in
reference (82).
A fluorogenic labeling derivatization reaction with dansyl chloride has
been combined with HPLC for the determination of N-methylcarbamate
insecticides in soil and water. No preliminary cleanup was required,
and detection limits were 1-10 ng/4 ul injection (83).
7X DIRECT METHODS OF ANALYSIS
Determinations of intact, underivatized N-methylcarbamate insecticides
are hampered by their decomposition on GC columns under ordinary operating
conditions (84). 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
264
-------
Section Y
of crop extracts include a multiresidue method (85) on 5-6% DC-200
after acetonitrile partition and charcoal cleanup as for OP pesticides
(55), determination of carbofuran and other carbamates on 20% SE-30
(86), and determination of 0.2-15 ng of carbaryl and 1-naphthol on a
short column of 3% SE-30 (87). Highly deactivated GC column packing
prepared according to Aue (88) from acid washed Chromosorb W support,
surface modified with Carbowax 20M, has also been successfully used
for chromatography of intact N-methylcarbamates without degradation
on the column (89). Such columns are extremely promising for performing
analyses without required derivatization.
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 (90), 10% DC-200/15% QF-1 (1:1) at 160°C (91), and
5 and 10% DC-200 (92) have been successfully used, the former for
multiresidues of urea herbicides and the latter two for carbamate
herbicides in foods. However, decomposition of compounds of these
types has been noted under certain conditions, and determinations are
therefore often made via thermally stable derivatives of hydrolysis
products as mentioned earlier.
7Y ANALYSIS OF PLANT AND FOOD MATERIALS
Extraction of urea and carbamate pesticides from plant materials usually
involves blending with raethylene 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 7A,I).
Cleanup steps include solvent partition and/or liquid column chromato-
graphy, the exact nature of which are pesticide- and sample-dependent.
For example, a column of 4:1 MgO-cellulose was used for cleanup of
carbamate herbicide residues from a variety of foods (91) , while Florisil
was employed after acetonitrile-petroleum ether partition for the multi-
residue, multiclass determination of carbamate, urea, and amide residues
(93). Methods for extraction, cleanup, and GC of carbamates, ureas, and
other classes of herbicides (triazines, uracils, phenols) have been
reviewed (94-96). 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 (97). Residues of 15 organonitrogen herbicides and fungicides were
screened in foods by acetone extraction, partition and Florisil (2 percent
water) column cleanup, and CCD-GC determination (98). 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 (99).
265
-------
Section 7Z
Total residues of Mesurol and its sulfoxide and sulfone metabolites
in plant and animal tissues were determined by oxidation of the extract
with KMnC>4 to convert all residues to the corresponding sulfone, which
was detected at a limit of 0.03 ppm by an S-mode FPD (100).
7Z AIR ANALYSIS
Section 8,B of the EPA PAM contains details of a multiresidue, multiclass
analytical method for chlorinated, organophosphorus, and N-methylcarbamate
insecticides in ethylene glycol trapping solvent from air sampling
impingers (Subsection 6,H) (101). Although the method has been tested
with only a limited number of representative compounds present in
ethylene glycol, it is likely to be suitable for the determination
not only of these but of many other pesticides if they can be efficiently
trapped by this solvent.
The procedure involves addition of 600 ml of 2 percent sodium sulfate
solution to 100 ml ethylene glycol (equivalent to 20 m3 of sampled air),
extraction of pesticides with 40 ml methylene chloride, and fractionation
and cleanup by elution through a one gram silica gel (20 percent water)
column with 10 ml hexane, 15 ml 60 percent benzene in hexane, and 15 ml
5 percent acetonitrile in benzene. Chlorinated pesticides (found in
fractions I and II) are determined by EC-GC, OPs (fractions II and III)
by FPD-GC, and carbamates (fraction III) by EC-GC after derivatization
with pentafluoropropionic anhydride or 1-fluro-2,4-dinitrobenzene.
Recoveries of 90-110 percent of most of the 25 compounds studied were
obtained at
-------
Section 7A,A
effectively recovered (e.g., 102). This former method, which includes
Florisil column cleanup and GC with EC and FPD detectors, was the basis
of the former EPA National Air Pesticide Monitoring Programs (102a).
Details can be found in earlier editions of the EPA PAM, but use of the
new, more widely applicable procedure described above is generally
recommended.
7A,A 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. Recovery studies were conducted on 42 halogenated compounds,
38 OP compounds, and 7 carbamates. Recoveries of >80 percent were
achieved for 58 of the 87 compounds, 60-80 percent recovery for 13
compounds, and <60 percent for the remaining 16 compounds (concentration
levels 0.09-400 ppb). Pesticides are extracted from water with methylene
chloride, and the concentrated extract is chromatographed on a 1 gram
deactivated (20 percent 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 chromatography 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 concentrated 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. A number of the OP compounds require considerable column pre-
conditioning by repetitive injection of high-concentration standards in
order to obtain linearity of response and accurate quantitation. 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 re-
quired for a 2 or 3 day period. Larger 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.
Confirmation of pesticide identity should be made by several techniques
outlined in Section 8.
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 percent BC13 in
2-chloroethanol, and the resulting esters are extracted with hexane,
267
-------
Section 7A,A
concentrated, and determined by EC-GC. If cleanup is required, chroma-
tography on silica gel deactivated with 20 percent water is employed.
This procedure is a further extension of the multiclass, multiresidue
procedure described directly above. When preparing the BCl3-chloro-
ethanol esterification reagent, work in an efficient exhaust hood and
wear disposable vinyl gloves because 2-chloroethanol is toxic 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
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 (103) as the best derivatization
reagent for the determination of 8 phenylalkanoic acid herbicides in
water (0.01-2.5 ug/L); solvent partition and silica gel (5 percent water
deactivated) minicolumn cleanup and EC-GC with an OV-17/QF-1 column were
employed.
The Manual of Analytical Methods 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,
and pentachlorophenol in waters. The method for organochlorines,
employing Florisil column cleanup and EC-GC, has detection limits
ranging from 0.005-0.1 ppb. The first OP procedure determines dimethoate,
fenitrothion, and phosphamidon and the second determines 14 other OP
pesticides, all at 0.01-0.1 ppb levels by FPD-GC without cleanup. Phenoxy
acid herbicides are extracted with chloroform from acidified water and
converted to their methyl esters utilizing BFo-methanol prior to cleanup
on a Florisil column 2 cm in height and EC-GC determination at 0.01-0.05 ppb
levels. PCP is detected at 0.01 pg/L by benzene extraction from acidified
water, partition into potassium carbonate solution, acetylation with
acetic anhydride, partition into hexane, and EC-GC.
Cleanup is often not required for EC-GC analysis of surface water samples
(104) 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 (65). Results of an interlaboratory study of the analysis of
15 water samples for 10 OC1 pesticides without any column cleanup have
been reported (105). Where needed, cleanup and separation of common
chlorinated and OP insecticides extracted from water have been successfully
carried out in silica gel microcolumns (106-107) and columns of deacti-
vated (5-20 percent 1^0) silica gel (above) and alumina (108).
268
-------
Section 7A,B
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
desulfurization with Raney copper powder (109) or copper ribbon (110),
precipitation with metallic mercury (111), reaction with CN~ (112),
and treatment with tetrabutylammonium sulfite to produce an ion pair
with sulfur as 8263 = (113) have been used to remove such interference.
(See also Subsection 7A,B).
Polar phosphorus, urea, and carbamate pesticides are extracted from
water with more polar solvents such as chloroform, acetone, or methylene
chloride. Extraction of acidic or basic compounds is aided by adjusting
the water sample to a controlled pH value. Determination by GC is
carried out using an appropriate selective detector after extract
concentration and any required cleanup and/or derlvatization steps.
As an example, carbaryl and 1-naphthol have been determined in
natural water at 2.5-10 ppb levels (82-102 percent recovery at 5 ppb)
by EC-GC after methylene chloride extraction, cleanup on an XAD-8
column, and derivatization with HFBA reagent (114).
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 diazomethane or BF^-methanol,
followed by gas chromatography with an electron capture or micro-
coulometric detector (115) (see the footnote on page 4 concerning the
hazards of diazomethane).
TLC determinations of carbamate, urea, triazine, and uracil herbicide
residues in water have been reviewed (82, 116), as have the extraction,
cleanup, GC determination, and confirmation of chlorinated insecticides
in water and soils (117).
7A,B 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
determined by EC-GC. A similar AOAC method has been declared official
final action for residues of aldrin, £,£'-DDE, £,£f-DDT, £,£?-DDT,
£,£*-TDE, dieldrin, endrin, heptachlor, heptachlor epoxide, and lindane
(118).
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 percent 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 procedure are presented in Section 11,B of the EPA PAM.
269
-------
Section 7A,B
Air drying of the sample required 1-3 days, depending on the soil type.
Such samples will contain at least 50 percent water. Pesticide con-
centrations 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 (119).
Sediment samples may contain elemental sulfur which 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,VI
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 percent
ethyl ether eluates from the Florisil column are concentrated.
Nine chlorinated insecticides were determined by a modified GC procedure
(120) with recoveries of 75-99 percent from suspended sediment and
bottom material. Extraction was with acetone and hexane added separately,
coextractives (including PCBs) were isolated by alumina and silica gel
column chromatography, and EC-GC was used to analyze the various column
eluates.
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 among the three
methods for the majority of pesticides, but shake extraction was
significantly more efficient for BHC isomers (121). The shake extraction
method with hexane-acetone after moistening the soil with 0.2 M NH4C1
was studied collaboratively using standard AOAC analytical methods
(Florisil cleanup and EC-GQ (122) and found to give excellent recoveries
for six insecticides in three different soils (123). Some soil analyses
have been carried out by EC-GC with no required column cleanup (124), but
this is not common.
Soil residues of chlorfenvinphos, Chlormephos, disulfoton, phorate, and
pirimiphos-ethyl were determined by GC with thermionic detection.
Extracted compounds were cleaned-up on a carbon-cellulose column.
Recoveries ranged from 95-101 percent (125).
270
-------
Section 7A,C
A multiresidue GC procedure for the herbicides dichlobenil, dinitramine,
trlallate, and trifluralin in soils was described by Smith (126).
Extraction 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 percent 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 (127).
Anilide herbicides were determined by GC after extraction from soil
by blending with acetone (128). Urea and carbamate herbicides were
recovered from soils by shaking with methanol (129) or acetone (130)
and by alkaline hydrolysis and steam distillation (131). lodinated
(131) and 2,4-dinitrophenyl (130) derivatives were used for EC-GC
determination of the herbicides. Triazines were extracted with diethyl
ether from soil treated with ammonia (132) and uracils with 1.5 N NaOH
(133). Nineteen acidic, neutral and basic herbicides have been
determined in soils by two dimensional TLC (134). Carbofuran residues
in soil were determined at the 0.1 mg/kg level without cleanup by
EC-GC after ammonium acetate extraction and formation of the dinitrophenyl-
ether derivative (135). Uracils have been recovered by elution with
water from a column prepared by mixing soil with Celite and Ca(OH)25
the eluate was acidified and extracted with CHCl^, and uracil determina-
tion was by RbCl thermionic-GC (136).
The electrolytic conductivity detector has been used to determine nitrogen-
containing residues 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 (137).
The analysis of pesticides of many classes in soils and plants has been
reviewed (138).
POLYCHLORINATED BIPHENYLS (PCBs)
C PESTICIDE-PCB MIXTURES
PCBs are among the most ubiquitous and persistent chlorinated pollu-
tants 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
£,£'-DDT, c_,p_'-DDT, £,£'-DDD, and £,p_'-DDE, as well as with early eluting
pesticides such as BHC isomers, aldrin, heptachlor, and heptachlor
epoxide, since prominent PCB peaks have retention times similar to these
pesticides on the recommended GC columns.
271
-------
Section 7A,D
PCBs are frequently detected in human adipose tissue, often at con-
centrations similar to those of chlorinated pesticides. Occasionally,
however, PCB interference is significant in human tissue analysis,
depending upon the columns and operating parameters used. These inter-
ferences 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 8).
7A,D 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 7-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 conse-
quence has an absolute retention of about 6 minutes and the final peak
about 38 minutes. Figure 7-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.
The type of confusion evident when pesticides and PCBs are present in
the same substrate is illustrated in Figure 7-E showing Aroclor 1254
co-chromatographed with a mixture of eight chlorinated insecticides.
Aldrin (peak 1), £,j3'-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,
and complete overlap of DDE, dieldrin, ODD, and DDT.
Figure 7-C. Aroclor 1254. Column
4% SE-30/6% QF-1, 200°C,
carrier flow 70 ml/min.
8 12 16 20
Retention, minutes
24
28 32 36 40
272
-------
Section 7A,D
Figure 7-D. Aroclor 1260. Column
4% SE-20/6% QF-1, 200°C,
carrier flow 70 ml/min.
16 24 32 40
Retention, minutes
48
1 Aldrin
Figure 7-E. Aroclor 1254 (solid line) 2 H«pt. Epo
and pesticide mixture lo'Zurin*
(dotted line). Column SP.R'-DDD
4% SE-30/6% QF-1, 200°C, 6P,p--DDT
carrier flow 70 ml/min. ™'|™J
8 12 16 20
Retention, minutes
24
28
32
273
-------
Section 7A,D
Confusing chromatograms also result when-PCBs are mixed with the multi-
peak pesticides chlordane or toxaphene. 'Figure 7-F shows the co-
chromatogram 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.
Figure 7-F. Aroclor 1254 (solid line)
and chlordane (dotted line}
Column 4% SE-30/6% QF-1,
200°C, carrier flow 70 ml/
min.
Retention , minutes
20
24
274
-------
Section 7A,D
Figure 7-G shows a mixture of Aroclor 1254 with toxaphene. Analyses of
toxaphene, chlordane, and PCBs are further confused because the chroma-
tograms 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 7-G. Aroclor 1254 (solid line)
and toxaphene (dotted line)
Column 4% SE-30/6% QF-1,
200°C, carrier flow 70 ml/
min.
12 20
Retention , minutes
36 40
The actual effect of PCBs on quantitation of chlorinated pesticides
is highly dependent on the levels involved, the pesticide of interest,
and the attenuation being used. For example, if the ratio of PCBs to
pesticides is 10 ppm to 3 ppm, an attenuation can be used which 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, quanti-
tation of DDT will definitely be affected, and with 100 ppm PCB to
3 ppm pesticide and attenuation to keep DDE on scale, determination of
the latter would be affected.
275
-------
Section 7A,E
7A,E 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 con-
ductivity 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 acetonitrile, residues are partitioned into hexane, the
hexane is concentrated, and the PCBs are eluted through a 1
gram deactivated (3 percent water) Florisil column. The eluted
PCBs are further separated from chlorinated pesticides on a
micro silicic acid column. 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 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.
276
-------
Section 7A,E
Analysts inexperienced with the method should be guided
through the procedures 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 (139)
(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
response 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) (140). HPLC and
capillary column GC have also been used to separate PCB
mixtures (141, 142).
Crist and Moseman (143) of the EPA reported a simplified micro
perchlorination method for determination of PCBs in biological
samples. A sample was cleaned-up by the modified MOG procedure
(Section 7A), and the PCBs were perchlorinated with SbCl5 to
decachlorobiphenyl (DCS), which was cleaned-up by hexane
partitioning and chromatography on a 1.6 gram column of
activated Florisil. DCB was eluted with 7 ml of hexane.
The amount of DCB determined in the sample by EC-GC was con-
verted to the desired Aroclor through multiplication by the
quotient obtained from the average molecular weight of Aroclor
divided by the molecular weight of DCB. Recovery of as little
as 0.1 ppm of PCBs in 500 mg of tissue extract was 79-99 percent.
Human milk and fortified goat milk samples were also analyzed.
The presence of impurities in SbCl5 reagent that can cause
erratic recoveries of PCBs was noted by Trotter and Young (144),
and DCB impurity was detected in various brands of the reagents
used in the Crist and Moseman procedure (143).
277
-------
Section 7A,E
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 quantitation of each without
prior separation, others will require a separation before
determination, and still others may require a separation
procedure that destroys or converts some of the compounds
to permit quantitation of those remaining unchanged.
PCBs are eluted with 6 percent ethyl ether-petroleum ether
in the modified MOG procedure described -in the EPA PAM,
Section 5,A,(1) and in the FDA PAM multiresidue procedures,
Sections 211 and 212. They elute with eluent I of the
alternative methylene chloride elution system (Section 7H
of this Manual and Section 252 of the FDA PAM). A study
by Lieb and Bills (145) found that the storage temperature
of Florisil after 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 (146).
Polychlorinated terphenyls (PCTs) are also recovered by the
multiresidue procedure described in Section 5,A,(1), but
these compounds elute from the GC column much later than OC1
pesticides and PCBs and therefore do not interfere. To
determine PCTs, GC parameters must be altered to provide more
rapid elution and greater sensitivity. The spectrometric and
GC properties of 22 PCTs have been reported (147).
The method of Armour and Burke (139) has been most used for
pesticide-PCB separation. The 6 percent ethyl ether-petroleum
ether Florisil column eluate or eluate 1 of the alternative
278
-------
Section 7A,E
procedure (Section 7G 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_,jD'-DDE and Aroclor 1254. Petroleum ether followed
by acetonitrile-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 _p_,£r-DDE (Canadian
PAM, Section 7.5). Polychlorinated naphthalenes (148) and
dioxins 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 (149). Pesticide-PCB
separations were found reproducible only for individual
batches of adsorbent.
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 (150).
Other columns used in various multiresidue cleanup procedures
provide at least partial separations of organochlorine pesti-
cides from PCBs. These include columns of activated alumina
(23; Section 7M), silica plus alumina (21; Section 7M), and
charcoal (151). 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
279
-------
Section 7A,E
DDT to DDE and IDE to its olefin; PCBs are unchanged. Subse-
quent 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 dichlorobenzophenone 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 percent (0.4-56 ug amounts), while DDT,
DDE, and TDE were recovered (as dichlorobenzophenone) between
5-86 percent (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 1,5-diazobicyclo(5.4.0)undec-5-ene
reagent (152); sodium dichromate in acetic acid plus sulfuric
acid for conversion of DDE to dichlorobenzophenone without
affecting DDT, TDE, or PCBs (153); a silica gel tube with MgO
catalyst for conversion of DDT to DDE and TDE to DMU without
effect on PCBs (154); and heating cleaned-up fish or serum
extracts with KOH/ethanol to convert OC1 pesticides to alkenes,
oxidation with Cr203 to more polar compounds, and separation
from PCBs on a Florisil column (155).
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
(156). PCT, PCB, and DDT residues in blood (5-11 ppb) were
determined by heating with ethanol and KOH to dehydrochlorinate
DDT, extracting with hexane, washing with H2S04, and chromatographing
on a mixed silica + Florisil + Na2SO^ column. The hexane eluate
was concentrated and analyzed by EC-GC on a 2% OV-1 column, and
confirmation was by MS (157).
c. GC Quantitation of PCBs
One of the most difficult aspects of PCB quantitation is to obtain
a match between the sample and a standard. 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 to different PCB isomers can vary by as much as 10,000-
fold, so-that accurate quantitation is impossible. The most
widely used practical approach is to compare the total area
or height of detector response for the residue peaks to the total
280
-------
Section 7A,E
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 chromato-
gram.
The procedure of Webb and McCall (158) has an advantage com-
pared to this method 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 de-
tector for Cl measurement and chemical ionization MS with single
ion monitoring for molecular weights (159). A ten laboratory
collaborative 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 (160).
Other quantitation approaches that have been attempted include
perchlorination of all PCB compounds with SbCl3 to a single
derivative (decachlorobiphenyl) (143, 161; Subsection 7A,E.a.),
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 jD,£'-DDE on the
same GC column (162), and peak resolution and matching by a
computer (163). GC-MS with individual mass monitoring using
a minicomputer-controlled spectrometer has been reported (164).
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
2~values have been applied to quantitation of jD,£f-DDT in
the presence of non-resolved PCB peaks and results within 11
percent of actual were reported (165). The USFDA approach to
chemical profiling of PCB content in a sample to select the
most suitable quantitation standard has been discussed.
d. PCB Estimation by TLC
The semiquantitative TLC procedure (167) for determination of
PCBs in adipose tissue utilizes the 6 percent eluate of the
Florisil cleanup. The concentrate is treated with KOH to dehydro-
281
-------
Section 7A,E
chlorinate DDT and ODD to their olefins, thereby eliminating
the problem of separating the pesticides from the PCBs. Any
interfering DDE is then oxidized to p_,p_'-dichlorobenzophenone
which has an R^ value different from the PCBs on a silver nitrate-
impregnated alumina layer developed with 5 percent 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, with a precision of
roughly +50 percent 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 confirmation by com-
bined GC-MS (168).
e. Confirmation of PCBs
Confirmation of PCBs has been obtained by perchlorination (143,
161) and alkali treatment (169). The stability of PCBs to
alkali makes the latter reaction useful for confirming the
identity of PCB residues, and at the same time, conversion of
DDT to DDE by the alkali removes some interference to quantita-
tion 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 (167) and chlorinated naphthalenes (170) .
Two-dimensional (171) or multi-development reversed phase (172)
TLC systems, which separate PCBs from DDE, DDT, and other
pesticides, can aid identification. PCBs are destroyed by UV
irradiation, but many pesticides may be altered as well.
Toxaphene survives UV treatment that destroys PCBs and can
be confirmed in mixtures in this way. Mirex, a late eluting
pesticide which usually is not interfered with by PCBs, also
withstands UV irradiation and can thus be confirmed. Irradiation
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 wavelengths >290 nm; subsequent irradiation
with wavelengths ^230 nm yielded photodechlorinated products
of PCBs with shorter retention times (173). Most organochlorine
282
-------
Section 7A,F
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/minute (174). Mirex and PCBs have
been differentiated based on the low sensitivity of the Hall
detector for the latter (175). A collection of spectra
helpful in confirming isolated residues of PCBs has been
published.
7A,F DETERMINATION OF POLYBROMINATED BIPHENYLS
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 FDA and the EPA.
One 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 (177).
A rapid method has been developed for analysis of plasma, feces,
milk, and bile using all disposable glassware to reduce the amount
of laboratory background and cross-contamination of samples (178).
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
5% OV-17 column. Mean recoveries for the six major components of a
commercial PBB mixture were approximately 96 percent for plasma,
59 percent for feces, 98 percent for milk, and 89 percent 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.
283
-------
Section 7A,G
The separation and characterization of PBBs by chromatography and
spectroscopy have been studied (179). Columns containing 1% SE-30
or 2% OV-17 were used for GC-FID-MS, 5 urn 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.
PBB residues in dairy products were determined at 7 ppb levels by
co-extraction from the sample along with fat, separation by GPC,
EC-GC on an OV-101 column, and confirmation by TLC (180).
7A,G SEPARATION AND DETERMINATION OF DIOXINS
Laboratories that have conducted environmental monitoring projects
for 2,3,7,8-tetrachlorodibenzo-p_-dioxin (TCDD) have developed and
applied analytical cleanup systems and mass spectrometric methods
of analysis for sub-ppt to 100 ppt levels of TCDD residues in environ-
mental, biological, and human samples and chemical formulations
(181-192).
The cleanup methods involve acid, base, and/or neutral extraction pro-
cedures followed by column chromatography using Celite, alumina, and
silica gel columns. The mass spectrometric analysis methods include
direct probe high resolution single and double ion monitoring, low
resolution (packed column) GC/low mass resolution MS analyses, low
resolution GC/high mass resolution double ion monitoring, high resolution
(capillary) GC/low mass resolution analyses and high resolution (capi-
llary) GC/high mass resolution multiple (three ion) monitoring MS
techniques.
The preparation and characterization by GC-MS of a series of dioxin
standards useful for residue quantitation have been described (91).
284
-------
Sections 7A,H, I
7A,H DETERMINATION OF ETHYLENETHIOUREA (ETU)
Because of its toxicological significance, constant occurrence as
a terminal residue following crop treatments with ethylenebisdithio-
carbamate fungicides, and its actual presence in technical ethylenebis-
dithiocarbamates, analytical methods for determination of ETU are
becoming increasingly important. A method for ETU in apples (193)
was based on reaction with benzyl chloride to give 2-benzylmercapto-
imidazoline, which is subsequently treated with trifTuoroacetic
anhydride to yield 2-benzylmercapto-ttf-trifluoroacetylimidazoline.
This derivative is measured by EC-GG.
ETU residues were measured in various crops by methanol extraction,
alumina column cleanup, and derivatization 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 (194). A similar method that determines ETU in milk, fruits,
and vegetables as the same derivative has been collaboratively studied,
and tentative evaluation indicates successful results (195).
EC-GC as well as S-mode FPD-GC has been used to determine ETU residues
from crops after derivatization with mr-trifluoromethylbenzyl chloride
(196). The trifluoroacetylated S-benzyl derivative has also been
used to determine ETU residues on tomatoes (197). ETU residues on
fruit and vegetable crops were determined at 0.01-0.1 ppm levels
without derivatization (198). After methanol extraction and cleanup
by hexane/aqueous NH^Cl partition and alumina column chromatography,
GC was performed on a 3% Versamid 900 column with S-mode FPD detection.
Recoveries ranged from 62-95 percent.
The occurrence, chemistry, and metabolism of ETU and analytical methods
for its determination have been reviewed (199).
7A, I 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
285
-------
Section 7A,I
conjugates 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 8L), 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 8K), as
are traditional IR and UV absorption spectrometry and micro-IR (Section
8J).
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-dried 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 anion-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 G gel columns, Biogel P columns, cation-
exchange resin amino-acid analyzer columns, liquid anion-exchange
paper chromatography, TLC, and GG 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
radiotracer 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 (200). This volume also
contains information on the nature and analysis of "bound" or
unextractable 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 (201); pyrolysates containing
intact chloroanilines were collected and derivatized as trifluoro-
acetanilides, which were purified and determined by EC-GC.
286
-------
Section 7A,J
Relatively little attention has been given to the recovery of pesti-
cide 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 carbofuran residues in plant
and animal samples (202). The major carbofuran metabolite produced
in animals is 3-hydroxy-carbofuran, present as the water soluble
glucuronide conjugate. The analytical procedure for intact carbo-
furan residues was modified to include a mild acid hydrolysis, the
purpose of which was to free the conjugated forms of the pesticide
metabolite and allow an organic solvent extraction that included the
unchanged parent compound.
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 determina-
tion (e.g., hydrolysis of carbamate insecticides to the corresponding
phenol, which is derivatlzed, cleaned-up, and determined by GC).
Conjugates of 2,4,5-T in biological samples have been broken and
the free acids released by a basic hydrolysis step (203). 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 (204).
7A,J REVIEWS OF ANALYTICAL METHODS FOR PESTICIDES, PCBs, AND OTHER
NON-PESTICIDE POLLUTANTS
The analysis of residues of various classes of pesticides has been
reviewed in special issues of the Journal of Chromatographic Science
edited by M. C. Bowman (205). Organophosphorus insecticides and
metabolites, carbamate insecticides and metabolites, herbicides and
metabolites, insecticide synergists, chlorinated insecticides and
congeners, and fumigant residues have been covered. A review of
the determination of fungicide residues (206, 207) and more recent
coverage of herbicide residue determinations (89) have been published.
Published reviews have covered the extraction, cleanup, and chromato-
graphic detection and quantitation of PCBs and other non-pesticide
environmental pollutants such as chlorinated naphthalenes, dibenzofurans,
and dibenzodioxins (208, 209); USDA studies on the sources, analysis,
and environmental monitoring of dibenzodioxins and dibenzofurans (210);
the utilization of GC coupled with chemical ionization and electron
impact MS for analysis of these compounds and pesticides (211); the
chromatography of PCBs (212); and PCB interference and analysis in
terms of the AOAC-FDA multlresidue method for chlorinated pesticides (213)
287
-------
Section 7A,K
7A,K REFERENCES
(1) Franken, J. J., and Luyten, B. J. M., J. Ass. Offic. Anal. Chem.,
J>9, 1279 (.1976).
(2) Griffith, F.D., and Blanks, R. V., J. Ass. Offic. Anal. Chem.,
57_, 595 (1974).
(3) Kurtz, D.A., and Kim, K. C. , Pestie. Monit. J., J.O, 79 (1976).
(4) Reiner, E., Krauthacker, B., Stipcevic, M., and Stefanic, Z.,
Pest. Monit. J.. LI, 54 (1977).
(5) Moseman, R. F., Crist, H. L., Edgerton, T. R., and Ward, M. K.,
Arch. Environ. Contam. Toxicol., j6, 221 (1977).
(6) Moseman, R. F., Ward, M. K., Crist, H. L., and Zehr, R. D.,
J. Agr. Food Chem., JZ6, 965 (1978).
(7) Stafford, C. J., Reichel, W. L., Swineford, D. M., Prouty, R. M.,
and Gay, M. L., J. Ass. Offic. Anal. Chem., 61, 8 (1978).
(8) Blanke, R. V., Farris, M. W., Griffith, F. P., and Guzelian, P.,
J. Anal. Toxicol., I, 57 (1977).
(9) McMahon, B.M., and Burke, J. A., 91st National AOAC Meeting,
1977, Abstract No. 73.
(10) Mills, P. A., Bong, B. A., Kamps, L. R., and Burke, J. A.,
J. Ass.- Offie. Anal. Chem.. 55, 39 (1972); Mitchell, L. E.,
J. Ass. Offic. Anal. Chem.. 5J^, 209 (1976); FDA PAM, Section 252
and Table 201-G.
(11) Gabica, J., Watson, M., and Benson, W. W., J. Ass. Offic. Anal.
Chem.. 57., 173 (1974).
(12) Yip, G., J. Ass. Offic. Anal. Chem.. _54, 966 (1971).
(13) Brown, M. J., J. Agr. Food Chem.. ,23, 334 (1975).
(14) Crist, H. L., Moseman, R. F., and Noneman, J. W., Bull. Environ.
Contam. Toxicol., 14, 273 (1975).
(15) Bong, R. L., J. Ass. Offic. Anal. Chem., £0, 229 (1977).
288
-------
Section 7A,K
(16) Rourke, D. R., Mueller, W. F., and Yang, R. S. H., J. Ass. Offic.
Anal. Chem., 60, 233 0-977).
(17) Solomon, J., and Lockhart, W. L., J. Ass. Offic. Anal. Chem., 60,
690 (1977).
(18) de Faubert Maunder, M. J., Egan, H., Godley, E. W., Hammond, E. W.,
Roburn, J., and Thomson, J., Analyst, 89, 168 (1964).
(19) Wood, N. F., Analyst, 94., 399 (1969).
(20) Smart, N. A., Hill, A. R. C., and Roughan, P. A., J. Ass. Offic.
Anal. Chem., 57., 153 (1974).
(21) Holden, A.V., and Marsden, K., J. Chromatogr., 4^, 481 (1969).
(22) Wells, D.E., and Johnstone, S. J., J. Chromatogr., 140, 17 (1977).
(23) Telling, G. M., Sissons, D. J., and Brinkman, H. W., J. Chromatogr.,
137, 405 (1977).
(24) Claeys, R. R., and Inman, R. D., J. Ass. Offic. Anal. Chem., 57,
399 (1974).
(25) Lea, R., J. Ass. Offic. Anal. Chem., 59, 224 (1976).
(26) Stalling, D. C., Tindle, R. C., and Johnson, J. L., J. Ass. Offic.
Anal. Chem. , .55, 32 (1972).
(27) Griffitt, K.R., and Craun, J. C., J. Ass. Offic. Anal. Chem., 57,
168 (1974).
(28) Friestad, H. 0., J. Ass. Offic. Anal. Chem., j>0, 268 (1977).
(29) Getz, M. E. , Talanta. 22_, 935 (1975).
(30) Abbott, B.C., and Thomson, J., Chem. Ind. (London), p. 481 (1964);
Analyst, 89_, 613 (1964).
(31) Calderbank, A., Residue Reviews, 12, 14 (1966).
(32) Renberg, L., Anal. Chem., 4£, 459 (1974).
(33) Holden, A. V., Pestic. Monit. J.. 7_, 37 (1973); Environ. Qual.
Saf... Suppl.. 3, 40 (1975).
289
-------
Section 7A,K
(34) Report by the Panel on Fumigant Residues in Grain, Analyst. 99,
570 (1974).
(35) Lores, E. M., and Bradway, D. E., J. Agr. Food Chem., 25. 75 (1977).
(36) Bradway, D. E., Shafik, T. M., and Lores, E. M., J. Agr. Food
Chem.. 21» 1353 (1977).
(37) Bradway, D. E,, and Shafik, T, M., J. Agr. Pood Chem., 2j>, 1342 (1977).
(38) 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).
(39) Shafik, T. M., Sullivan, H. C., and Enos, H. F., J. Agr. Food
Chem. . .21, 295 (1973).
(40) Edgerton, T. R., and Moseman, R. F., J. Agr. Food Chem., 26, 425
(1978).
(41) Wessel, J. R., J. Ass. Offie. Anal. Chem.. j>0, 430 (1967).
(42) McCully, K. A., J. Ass. Offic. Anal. Chem.. .55, 291 (1972).
(43) Pardue, J. R. , J. Ass. Offic. Anal. Chem.. 5.4, 359 (1971).
(44) Beckman, H., and Barber, D., J. Ass. Offic. Anal. Chem., 52. 286 (1969).
(45) Versino, B., van der Venne, M. Th., and Vissers, H., J. Ass. Offic.
Anal. Chem.. 54., 147 (1971).
(46) Dale, W. E., and Miles, J. W. , J. Ass. Offic. Anal. Chem., .59, 165
(1976).
(47) Hladka, A., Kovac, J., and Krampl, V., Z. Anal. Chem., 274. 371 (1975).
(48) Kovac, J., Batova, V., Hankova, A., and Szokolay, A., Bull. Environ.
Contain. Toxicol., 1.3, 692 (1975).
(49) Storherr, R.W., and Watts, R. R., J. Ass. Offic. Anal. Chem., 48.
1154 (1965).
(50) Watts, R. R., and Storherr, R. W., J. Ass. Offic. Anal. Chem., 48,
1158 (1965).
(51) Watts, R. R., and Storherr, R. W., J. Ass. Offic. Anal. Chem., 50.
581 (1967).
290
-------
Section 7A,K
(52) Storherr, R.W., and Watts, R. R., J. Ass. Offic. Anal. Chem.. 51,
662 (1968).
(53) Storherr, R. W., Murray, E. J., Klein, I., and Rosenberg, L. A.,
J. Ass. Offic. Anal. Chem.. 50 605 (1967).
(54) Malone, B., and Burke, J. A., J. Ass. Offic. Anal. Chem., 52, 790
(1969).
(55) Storherr, R. W., Ott, P., and Watts, R. R., J. Ass. Offic. Anal.
Chem.. ,54, 513 (1971).
(56) Laski, R. R., J. Ass. Offic. Anal. Chem.. ^7, 930 (1974).
(57) Bowman, M. C., Beroza, M., and Leuck, D. B., J. Agr. Food Chem.,
JL6, 796 (1968).
(58) Sissons, D. J., and Telling, G. M., J. Chromatogr., 47, 328 (1970).
(59) Renvall.S., and Akerblom, M., Residue Reviews, 34, 1 (1971).
(60) Jennings, D. M., Bunyan, P. J., Brown, P. M., Stanley, P. I.,
and Jones, F. J. S., Pestic. Sci.t j6, 245 (1975).
(61) Domanski, J. J., Haire, P. L., and Sheets, T. J., Arch. Environ.
Contain. Toxicol. . ^, 377 (1976).
(62) Abbott, D. C., Crisp, S., Tarrant, K. R., and Tatton, J. O'G.,
' Pestic. Sci.. JL, 10 (1970).
(63) Kadoum, A. M., Bull. Environ. Contain. Toxicol., _3, 247 (1968).
(64) Pree, D. J., Butler, K. P., Kimball, E. R., and Stewart, D. K. R.,
J. Econ. Entomol., 69, 473 (1976).
(65) Ripley, B. D., Wilkinson, R. J., and Chau, A. S. Y., J. Ass. Offic.
Anal. Chem.. _57, 1033 (1974).
(66) Kadoum, A. M., Bull. Environ. Cqntam. Toxicol., ^3, 65 (1968) and
2, 264 (1967).
(67) Ivey, M.C., and Mann, H. D., J. Agr. Food Chem.. 23, 319 (1975).
(68) Ivey, M. C., J. Ass. Offic. Anal. Chem.. j>9, 261 (1976).
(69) Bowman, M. C., and Hill, K. R., J. Agr. Food Chem., 1£, 342 (1971).
291
-------
Section 7A,K
(70) Aharonson, N. , and Ben-Aziz, A., J. Agr. Food Chem., 22, 704 (1974),
(71) Pflugmacher, J., and Ebing, W., J. Chromatogr., S>3, 457 (1974).
(72) Johnson, L. D., Waltz, R. H., Ussary, J. P., and Kaiser, F. E.,
J. Ass. Offic. Anal. Chem., 5£, 174 (1976).
(73) Leicht, R. W., Schofield, M. G., Johnson, L. D., and Waltz, R. H.,
91st National AOAC Meeting, 1977, Abstract No. 77.
(74) Bowman, M. C., and Beroza, M., J. Ass. Offic. Anal. Chem., 52,
1231 (1969).
(75) Thornton, J. S., Olson, T. J., and Wagner, K. , J. A_gr_. Food Chem.,
25_, 573 (1977).
(76) Report of the Ministry of Agriculture, Fisheries, and Foods,
Analyst, 102, 858 (1977).
(77) Watts, R. R., Storherr, R. W., Pardue, J. R., and Osgood, T.,
J. Ass. Offic. Anal. Chem., .52, 522 (1969).
(78) Sissons, D. J., and Telling, G. M. , J. Chromatogr.. 47, 328 (1970);
48, 468 (1970).
(79) Frehse, H., Anal. Chem.. 59., 896 (1976).
(80) Becker, G., Method 58 in Deutsche Forschungsgemeinschaft (Lieferung
1974) Ruckstands Analytik von Pflanzenschutzmitteln, No. 3., Verlag
Chemie, Weinheim/Bergstr., W. Germany.
(81) Bradway, D. E. , and Shafik, T. M. , J. Chromatogr. Sci., 1.5, 322 (1977)
(82) Sherma, J., Critical Reviews in Analytical Chemistry, (CRC Press)
J3 (3), pp. 299-354 (1973).
(83) Frei, R. W., Lawrence, J. F. , Hope, J., and Cassidy, R. M.,
J. Chromatogr. Sci., 12, 40 (1974).
(84) Strother, A., J. Gas Chrom<:Lo;:,,-., •' 10 'r°^8).
(85) Storherr, R. W., J. Ass. Offic. Anal. Chem., _55_, 283 (1972).
292
-------
Section 7A,K
(86) Cook, R. F., Stanovick, R . P., and Cassil, C. C., J. Agr. Food
Chem., 17., 277 (1969).
(87) Lewis, D. L. and Paris, D. F., J. Agr. Food Chem., 22, 148 (1974).
(88) Aue, W. A., Hastings, C. R., and Kapila, S., J. Chromatogr., 77,
299 (1973).
(89) Lorah, E.J., and Hemphill, D. D., J. Ass. Offic. Anal. Chem., 57,
570 (1974).
(90) McKone, C.E., and Hance, R. J., J. Chromatogr., 36, 234 (1968).
(91) Onley, J.H., and Yip, G., J. Ass. Offic. Anal. Chem., 54_, 1366 (1971)
(92) Laski, R. R., and Watts, R. R., J. Ass. Offic. Anal. Chem., 56,
328 (1973).
(93) Thier, H. P., Angew. Chem. Int. Ed., JLO, 860 (1971).
(94) Williams, I. H., Residue Reviews, 3<8, 1 (1971).
(95) Cochrane, W. P. and Purkayastha, R., Toxicological and Environmental
Chemistry Reviews, 1, 137 (1973).
(96) 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.
(97) Ramsteiner, K., Hermann, W. D., and Eberle, D. 0., J. Ass. Offic.
Anal. Chem.. 57., 192 (1974).
(98) Lawrence, J.F., and McLeod, H. A., J. Ass. Offic. Anal. Chem., 60,
979 (1977).
(99) Lawrence, J. F., J. Chromatogr., 121, 85 (1976).
(100) Thornton, J.S., and Draeger, G., Int. J. Environ. Anal. Chem., 2^,
229 (1973).
(101) Sherma, J., and Shafik, T. M., Arch. Environ. Contam. Toxicol., _3,
55 (1975).
(102) Arthur, R. D., Cain, J. D., and Barrentine, B. F., Bull. Environ.
Contam. Toxicol., 15, 129 (1976); Pestic. Monit. J.? 10, 168 (1977).
(102a) 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.
293
-------
Section 7A,K
(103) Agemian.HL., and Chau, A. S. Y., J. Ass. Offic. Anal. Chem., 60,
1070 (1977). — —
(104) Engst,R., and Knoll, R., Nahrung, 17, 837 (1973).
(105) Aspila, K. I., Carron, J. M., and Chau, A. S. Y., J. Ass. Offic.
Anal. Chem.t 60, 1097 (1977).
(106) Johnson, L. G., Bull. Environ. Contain. Toxicol.. _5, 542 (1970).
(107) Leoni, V., J. Chromatogr.. j>2, 63 (1971).
(108) Law, L.M., and Goerlitz, D. F., J. Ass. Offic. Anal. Chem.. 53,
1276 (1970). ~~—
(109) Schutzmann, R. L., Woodham, D. W., and Collier, C. W., J. Ass.
Offic. Anal. Chem.. ,54, 1117 (1971).
(110) Johnson, R.E., and Munsell, L. Y., Bull. Environ. Contain.
Toxicol., _17, 573 (1977). ~
(111) Goerlitz, D.F., and Law, L. M., Bull. Environ. Contain. Toxicol..
.6, 9 (1971). ~~ ~~
(112) Mattson, P.E., and Nygren, S., J. Chromatogr.. 124. 265 (1976).
(113) Jensen, S., Renberg, L., and Reutergardh, L., Anal. Chem.. 49,
316 (1977).
(114) Nagasawa, K., Uchiyama, H., Ogamo, A., and Shinozuka, T.,
J. Chromatogr.. 144. 77 (1977).
(115) Devine, J.M., and Zweig, G., J. Ass. Offic. Anal. Chem.. 521, 187 (1969).
(116) Sherma, J., "Thin Layer Chromatography: Recent Advances", Chapter
1 in Analytical Methods for Pesticides and Plant Growth Regulators.
Volume VII, Sherma, J. and Zweig, G., editors, Academic Press, N.Y.,
1973, pp. 1-88.
(117) Pionke, H. B.,and Chesters, G., Soil Sci. Soc. Am. Proc.. _32, 749 (1968)
(118) Burke, J. A., J. Ass. Offic. Anal. Chem.. 59_, 339 (1976).
(119) Maini, P., and Collina, A., Bull. Environ. Contain. Toxicol., 14.
593 (1975).
294
-------
Section 7A,K
(120) Goerlitz, D. P., and Law, L. M., J. Asa. Offic. Anal. Chem.. 57, 176
(1974).
(121) Nash, R. G., Harris, P. D., Ensor, P. D., and Woolson, E. A.,
J. Ass. Offic. Anal. Chem.. _56, 728 (1973).
(122) Official Methods of Analysis, (1970), llth Edition, AOAC,
Washington, D.C., Sections 29.014 and 29.017.
(123) Woolson, E. A., J. Ass. Offic. Anal. Chem.. ,57, 604 (1974).
(124) Suzuki, M., Yamato, Y., and Watanabe, T., Pestio Monit. J., 11,
54 (1977). ~
(125) Suett, D. L., Pestic. Sci.. _6, 385 (1975).
(126) Smith, A. E., J. Chromatogr.. J97, 103 (1974).
(127) Smith, A. E., J. Chromatogr.. 129. 309 (1976).
(128) Bartha, R., J. Agr. Food Chem., 16_, 602 (1968).
(129) McKone, C. E., J. Chromatogr.. 44, 60 (1969).
(130) Cohen, I.C., and Wheals, B. B., J. Chromatogr., 43, 233 (1969).
(131) Baunok, I., and Geissbuhler, H., Bull. Environ. Contam. Toxicol.. _3,
7 (1968).
(132) Abbott, D. C., Bunting, J. A., and Thomson, J., Analyst, 90. 357 (1965)
(133) Jolliffe, V. A., Day, B. E., Jordan, L. S., and Mann, J. D.,
J. Agr. Food Chem.. 15, 174 (1967).
(134) Smith, A. E., and Fitzpatrick, A., J. Chromatogr.. 57, 303 (1971).
(135) Caro, J. H., Glotfelty, D. E., Freeman, H. P., and Taylor, A. W.,
J. Ass. Offic. Anal. Chem.. _56, 1319 (1973).
(136) Caverly, D. J., and Denney, R. C., Analyst. 102. 576 (1977).
(137) Sarna, L.P., and Webster, G. R. B., J. Ass. Offic. Anal. Chem., 57,
1279 (1974).
(138) Thier, H. P., Angew. Chem. Int. Ed.» 13, 217 (1974).
(139) Armour, J.A., and Burke, J. A., J. Ass. Offic. Anal. Chem.t 53,
761 (1970); FDA PAM Section 251.1.
295
-------
Section 7A,K
(140) Albro, P. W., Haseman, J. K., Clemmer, T. A., and Corbett, B. J.,
J. Chromatogr.. 136. 147 (1977).
(141) Krupcik, J., Kriz, J., Prusova, D., Suchanek, P., and Cervenka, Z.,
J. Chromatogr.. 142. 797 (1977).
(142) Brinkman, U. A. Th., Seetz, F, L., and Reymer, H. G. M.,
J. Chromatogr.. 116. 353 (1976).
(143) Crist, H. L., and Moseman, R. F., J. Ass. Offic. Anal. Chem., 60,
1277 (1977).
(144) Trotter, W. J., and Young, S. J. V., J. Ass. Offic. Anal. Chem.,
^8, 466 (1975).
(145) Lieb,A. J., and Bills, D. D. , Bull. Environ. Contain. Toxicol.,
12, 328 (1974).
(146) Gardner, A. M., Righter, H. F., and Roach, J. A. G., J. Ass. Offic.
Anal. Chem.. _59_, 273 (1976).
(147) Chittam, B., Safe, S., Ruzo, L. 0., Hutzinger, 0., and Zitko, V.,
J. Agr. Food Chem., 25_t 323 (1977).
(148) Armour, j. A. , and Burke, J. A., J. Ass. Offic. Anal. Chem., 54,
175 (1971).
(149) Huckins, J. N., Stalling, D. L., and Johnson, J. L., J. Ass.
Offic. Anal. Chem.. 59, 975 (1976).
(150) Erney, D. R., J. Ass. Offic. Anal. Chem.. 57_, 576 (1974).
(151) Berg, 0. W., Diosady, P. L., and Rees, G. A. U., J. Ass. Offic.
Anal. Chem., _53_, 761 (1970).
(152) Miles, J. R. W., J. Ass. Offic. Anal. Chem., 55., 1039 (1972).
(153) Jensen, S., and Sundstrom, G., Ambio., _3, 7° (1974).
(154) Luckas, B., Pscheidl, H., and Haberland, D., Nahrung, 20,
K1-K2 (1976).
(155) Trotter, W. J. , J. Ass. Offic. Anal. Chem.. _58., 461 (1975).
(156) Welborn, M. E., Borchard, R. E., Hansen, L. G., and Metcalf, R. L.,
J. Ass. Offic. Anal. Chem.. 5_7, 1248 (1974).
296
-------
Section 7A,K
(157) Doguchi,M., and Fukano, S., Bull. Environ. Contain. Toxicol., 13,
57 (1975).
(.158) Webb, R. W., and McCall, A. C., J. Chromatogr. Sci., 11, 366 (1973).
(.159) Sawyer, L. D., J. Ass. Of fie. Anal. Chem. . 61, 272 (1978).
(160) Sawyer, L. P., J. Ass. Offic. Anal. Chem.. 61, 282 (1978).
(161) Armour, J. A., J. Ass. Offic. Anal. Chem.. 56., 302 (1973).
(162) Collins, G. B., Holmes, D. C., and Jackson, F. J., J. Chromatogr.,
_71, 443 (1972).
(163) Zobel, M. G. R., J. Ass. Offic. Anal. Chem., _5_Z, 791 (1974).
(164) Eichelberger, J. W., Harris, L. E., and Budde, W. L., Anal. Chem. ,
_46, 227 (1974).
(165) Zelenski, S. G., Tashiro, J., Worthen, L. R., and Olney, C. E.,
J. Chromatogr., 84, 67 (1973).
(166) Cairns, T., and Jacobson, R. A., J. Chem. Inf. Comput. Sci., 17,
105 (1977).
(167) Mulhern, B. M., Cromartie, E., Reichel, W. L., and Belisle, A. A.,
J. Ass. Offic. Anal. Chem.. 5>4, 548 (1970).
(168) 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 York City, August, 1972.
(169) Young, S. J. V., and Burke, J. A., Bull. Environ. Contain. Toxicol.,
_7, 160 (1972).
(170) Holmes, D. C. and Wallen, M., J. Chromatogr., 71, 562 (1972).
(171) Fehringer,N. V., and Westall, J. E., J. Chromatogr.. 5]_, 397 (1971).
(172) de Vos, R. and Peet, E. W., Bull. Environ. Contain. Toxicol.,
^, 164 (1971).
(173) Mansour,M., and Parlar, H., J. Agr. Food Chem., 26, 483 (1978).
297
-------
Section 7A,K
(174) Dolan, J. W., Hall, R. C., and Todd, T. M., J. Ass. Offie. Anal.
Chem.. 55., 537 (1972).
(175) Hallet, D. J., Noratrom, R. J., Onuska, F. I., Comba, M. E.,
and Sampson, R., J. Agr. Food Chem., 24, 1189 (1976).
(176) MacNeil, J. D., Bull. Environ. Contam. Toxicol.. 15. 66 (1976).
(177) Hass, J. R., McConnell, E. E., and Harvan, D. J., J. Agr. Food
Chem.. 26., 94 (1978).
(178) Willett, L. B., Brumm, C. J., and Williams, C. L., J. Agr. Food
Chem.. 26_, 122 (1978).
(179) DeKok, J. J., J. Chromatogr.. 142. 367 (1977).
(180) Fehringer, N. V., J. Ass. Offic. Anal. Chem.. 58., 798 (1975).
(181) O'Keefe, P. W., Meselson, M. S., and Baughman, R. W., J. Ass.
Offic. Anal. Chem.. 61, 621 (1978).
(182) Shadoff, L. A., and Hummel, R. A., Biomed. Mass. Spectrom. 5_, 7
(1978)
(183) Shadoff, L. A., Hummel, R. A., Lamparski, L., and Davidson, J. H.,
Bull. Environ. Contam. Toxicol. 18. 478 (1977).
(184) Smith, D. L., Presented at 24th Annual Conference on Mass Spectro-
metry and Allied Topics. PSB ILF (1976).
(185) Taylor, M. L., Presented at 24th Annual Conference on Mass Spectro-
metry and Allied Topics. H5 (1976).
(186) Ryan, J. F., Presented at 22nd Annual Conference on Mass Spectrometry
and Allied Topics. C-2 (1974).
(187) Harless, R. L., Presented at 24th Annual Conference on Mass Spectro-
metry and Allied Topics. PSB ILG (1976).
(188) Harless, R. L., Presented at 26th Annual Conference on Mass Spectro-
metry and Allied Topics. TP-10 (1978).
298
-------
Section 7A,K
(189) Hummel, R. A., J. Agr. Food Chem.. _25_, 1049 (1977).
(190) Buser, H. R., Anal. Chem. 4£, 918 (1977).
(191) Buser, H. R., J. Chromatogr. 129. 303 (1976).
(192) Rappe, C., and Marklund, S., Chemosphere 3, 281 (1978).
(193) Newsome, W. H., J. Agr. Food Chem.. j!0, 967 (1972).
(194) Raines, L. D., and Adler, I. L., J. Ass. Offie. Anal. Chem., 56,
333 (1973).
(195) Yip, G., J. Ass. Offic. Anal. Chem.. j>0, 370, 1105, and 1111 (1977),
(196) King, R. R., J. Agr. Food Chem.. 25_, 73 (1977).
(197) Newsome, W. H., J. Agr. Food Chem.. ^4, 999 (1976).
(198) Otto, S., Keller, W., and Drescher, N., J. Environ. Sci. Health,
Part B. 12, 179 (1977).
(199) Committee on Terminal Pesticide Residues, Applied Chemistry
Division, IUPAC, Pure Appl. Chem., 4£, 675 (1977).
(200) Kaufman, D. D., Still, G. G., Paulson, G. D., and Bandal, S. K.,
Bound and Conjugated Pesticide Residues, ACS Symposium, Series
No. 29, 1976.
(201) 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.
(202) Cook, R. F., Carbofuran, in Analytical Methods for Pesticides
and Plant Growth Regulators. Vol. VII, Sherma, J. and Zweig, G.,
editors, 1973, p. 187; J. Agr. Food Chem., J25, 1013 (1977).
(203) Nony, C. R., Bowman, M. C., Holder, C. L., Young, J. F., and
Oiler, W. L., J. Pharm. Sci.. 65, 1810 (1977).
(204) Ivey, M. C., and Oehler, D. D.,, J. Agr. Food Chem., _24, 1049 (1976).
(205) Bowman, M. C., editor, J. Chromatogr. Sci., 13, June and July
issues (1975).
(206) Sherma, J., J. Chromatogr., 113. 97-137 (1975).
299
-------
Section 7A,K
(207) Baker, P. B., and Hoodless, R. A., Pestle. Sci., _5, 465 (1974).
C208) Sherma, J., Advances in Chromatography, Volume 12, Marcel Dekker,
Inc., 1975, p. 141.
(209) Brinkman, U. A. Th., and Reymer, H. G. M., J. Chromatogr., 127,
203 (1976).
(210) Helling, C. S., Isensee, A. R., Woolson, E. A., Ensor, P. D. J.,
Jones, G. E., Plimmer, J. R., and Kearney, P. C. , J. Environ.
Qual., 2, 171 (1973).
(211) Oswald, E. 0., Albro, P. W., and McKinney, J. D., J. Chromatogr.,
98, 363 (1974); Oswald, E. 0., Levy, L., Corbett, B. J., and
Walker, M. P., J. Chromatogr. ,• 93, 63 (1974).
(212) Fishbein, L., J. Chromatogr., JJ8, 345 (1972).
(213) Burke, J. A., J. Ass. Offic. Anal. Chem. , 55_, 284 (1972).
300
-------
Section 8
CONFIRMATORY PROCEDURES
8A 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 quantita-
tion (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
determination of the PCBs which are easily mistaken for pesticide
residues such as £,_p_'-DDE and _p_,j3'-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 j3,£'-DDT, for Endosulfan I
and £,jDT-DDE, and for 3-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 corres-
ponding 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
nonspecific 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.
301
-------
Section 8A
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 chromatography 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
a 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 Rp value, amd the GG 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 (Figure 5-N in Section 5);
PC or TLC RF values from certain adsorbent/solvent systems; £-values
in different solvent pairs; and PC, TLC, and £-values. These combina-
tions will not provide independent information for confirming residue
identity.
In Figure 8-A, the correspondence between extraction _p_-values in hexane/
acetonitrile and isooctane/DMF solvent pairs (A), and _p_-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 8-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 confirmation.
Figure 8-A. Degree of correspondence between different types of data
for residue confirmation. A = extraction p_-values,
B = TLC vs £-values, C = PC vs TLC, D = TLC vs GC
. 1.0
I 0.8
I 0.6
|o.2
O
o
r
g Extraction p-value v
•"• thin-layer chromatocjraphy
1.0
Q Paper chromatography v
thin-layer chromatography
GLC v TLC
I 0.6
0.2
0 0.2 0.1 0.6 0.8 1.0
Isooctano/DMF
Oo
<
bc°" ...
o 0.2 o'.i 0.6 "bis 1.0
Silicagel/Hexane (Rr)
1.0
0.8
06
0.4
0.2
° 0 0
0 °
0 °
o
o
O.U
6.0
4.0
23.0
<
C£
2.0
1.0
0 0.2 0.4 0.6 0.8 1.0
Silica gol/Hexane (Rf)
DC 200
O
0
o o °
• 0 o
o
o o
0 °
0.2 0.1 0.6 0.8 1
Silica gol/Hcxsna (Fij.)
302
-------
Section 8B
When the analyst is making pesticide identifications, common sense
is necessary. An example of misapplied common sense would be re-
porting 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 RRTA values for a-BHC and/or £,£?-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 identifi-
cation.
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 this data and all other available
information, including pesticide usage, the chemistry and metabolism
of residues, common artifacts from sample substrates 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;
sample type; purpose of the analysis; and practical considerations
such as time, cost, number of samples, and available instrumentation.
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.
8B GC RELATIVE RETENTION TIMES
In most laboratories, the initial, tentative identification of a
pesticide residue results from a multiresidue procedure involving
303
-------
Section 8B
extraction, cleanup, 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 Kovats' retention indicies (4)
rather than absolute retention is more reliable (Subsection 5N in
Section 5).
b. Be highly suspicious of any peak with a calculated relative
retention value (RRT) that does not precisely match that of your
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 your 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 com-
bined with GC retention characteristics provides a valuable identifica-
tion 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-N 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
304
-------
Section 8C
by the presence of hexachlorobenzene. The latter is co-eluted with
a-BHC on silicone columns and with g-BHC on Apiezon, but all three
compounds are resolved on a polar cyano-silicone column. Dieldrin
and £,j>/-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 column, dieldrin elutes far later
than j>,j3f-DDE, to the extent of about 1.4x at 180°C column temperature.
8C 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 microcoulometric (MC), flame photometric, or con-
ductivity detectors, described in Subsections 5D, 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 MC or Coulson
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 MC or 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) greatly simplifies confirmation of sulfur- and/or
phosphorus-containing residues. Identification of a thiophosphate is
unequivocal if (1) its retention ratio on a given column matches with
a standard, (2) the compound elutes in the correct fraction from a
cleanup column, (3) it is detected by the FPD detector, and (4) the
sulfur (394 nm) to phosphorus (526 nm) response ratio of the FPD
matches the standard (Subsection 5F).
Use of appropriate specific titration cells in the MC detection system
(Subsection 5D) provides additional evidence of identity for compounds
305
-------
Section 8D
containing halogen, sulfur, or nitrogen. The detector has relatively
low sensitivity, therefore requiring a greater concentration of the
extract, larger sample injection, and/or a larger initial sample than
for the EC detector.
The selectivity of an EC detector can be improved if the products
formed in an EC detector are allowed to pass to a second column with
another EC detector; the resultant distinctive peak pattern can provide
identification of OC1 pesticides and PCBs (6).
8D THIN LAYER CHROMATOGRAPHY (TLC) RF VALUES
Experimental aspects of TLC and its use for screening and quantitation
of residues were covered earlier in Subsections 6Q through 6T in Section 6.
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 RF values
for confirmations because differences in development conditions from
run to run cause these values to be non-reproducible. Standards and
samples should always be run on adjacent areas of the same plate if
possible. If RF values must be used, the value relative to the Rp
of a standard compound X run on the same plate (Rx value) will be more
reliable than the absolute Rj? 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 cleanup on the sample extract and the level of detection. Oils and
waxes will particularly interfere with TLC, causing streaked zones and/or
306
-------
Section 8E
distorted Rp values which may completely negate its value for confirma-
tion. The 15 percent 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 (7). This heating procedure may, however, increase
background fluorescence from co-extracted compounds also present in
the sample. TLC after fluorogenic labeling (8) of pesticide residues
is a combination of chromatography with chemical derivatization (Sub-
section 8G) which 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 8L) 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 (9), 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 which 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.
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, densi-
tometry, or color photography (10). Where available, the latter appears
to be the preferred procedure.
8E HIGH PERFORMANCE LIQUID CHROMATOGRAPHY (HPLC)
See Subsections 6N through 6S in Section 6 for a discussion of this
topic. HPLC has been used mainly for quantitation of residues in
307
-------
Section 8F
situations where GC is either not applicable or not convenient to use.
AnHPLC 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 wavelength of maximum absorption for each pesticide. Detection
of HPLC effluents with a Cl-selective Coulson electrolytic conductivity
detector (11) can also provide useful confirmatory evidence.
8F EXTRACTION £-VALUES
Extraction _p_-values (12-16) 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 _p_-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. _p_-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
(13), and Section 12.C of the EPA PAM (data for 88 pesticides in the
latter) . A device and method for determining p-values with unequilibrated
solvents or unequal phase volumes are given in the FDA PAM, Section 622.1
and reference (16).
As mentioned earlier, the general technique of determining £-values has
much in common with the use of several GC columns, PC, and TLC in
identification studies since all systems may share the same partition
mechanisms. Unless the analyst, assures himself that the data are not
correlated, it is best to use either a PC or TLC % value or an extraction
_p_-value as one independent criterion of identity. The great advantage of
_p_-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.
308
-------
Section 8G
8G DERIVATIZATION (CHEMICAL REACTION) TECHNIQUES
Derivatives of pesticides are prepared for various reasons, such as to
decrease volatility or increase detectability for TLC; to increase
volatility, stability, and/or detectability and avoid tailing peaks
for gas chromatography; and to alter the structure to aid characteriza-
tion. 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 recent innovation which
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.
d. A cleanup method should be available to remove any background inter-
ferences 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.
Derivatization reactions are 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 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
309
-------
Section 8G
into a gas chromatograph is not possible, cleanup by solvent parti-
tioning and/or column chromatography and concentration steps may be
applied. Solid matrix reactions are generally carried out by intro-
duction 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 not
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.
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 in GC contains a section on pesticides (17).
a. ORGANOCHLORINE PESTICIDES
Most of the effort to date in the development of confirmatory derivatiza-
tion tests has been confined to the organochlorine insecticides. For
these compounds, addition, oxidation, epoxidation, rearrangement, de-
chlorination, hydrolysis, reduction, and dehydrochlorination are the
most commonly used reactions. Examples of specific tests are shown
in Table 8-1, as reviewed by Cochrane and Chau (18). A later review
(19) contains more recent references to the reactions shown in Table 8-2.
Table 8-3 lists selected references for OC1 pesticide derivatization
methods published since 1975. Still another review of reactions for
chlorinated pesticides is found in reference (3). Section 7A,E,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.
310
-------
Table 8-1
Section 8G
Pesticide
CONFIRMATORY DERIVATIZATION TESTS FOR
PESTICIDE AND METABOLITE RESIDUES (l8)
Reaction
DDT
DDE
DDD
Methoxychlor
a) Dehydrochlorination
b) Dechlorination (of p_,p_' -isomer)
Oxidation
Dehydrochlorination
Dehydrochlorination
Aldrin
Dieldrin
Endrin
Endosulfan
Heptachlor
Heptachlor epoxide
cis- and trans-Chlordane
Nonachlor
a) Addition
\tert-BuOCl
b) Epoxidation
Epoxide
cleavage
rearrangement
acetylation
a) Epoxide rearrangement
b) Dechlorination
Sulfite reduction
acetylation
a) Allylic ^ hydroxylation
dechlorination
b) Addition
c) Epoxidation
Epoxide rearrangement
Dehydrochlorination
a) Dechlorination
b) Dehydrochlorination
311
-------
Table 8-2
Section 8G
CONFIRMATORY TESTS FOR ORGANOCHLORINE PESTICIDES (19)
Pesticide Class
Reagent or Reaction Type
General
Hexachlorobenzene (HCB)
BHC isomers
Cyclodiene insecticides
Mirex
Kepone
PCBs
Chlorobiphenyls and PGP
DDT
CrCl2 reduction (26)
KOH dehydrochlorination (46)
Base/alcohol
KOH hydrolysis/diazomethane
NaOMe/MeOH or GC alkaline precolumn
Comparison of 8 methods (D)
10 various reactions (D)
BCl3/2-chloroethanol (D/E)
UV irradiation (D/E/H)
H2S04 or 60% KOH (E/M)
_t-BuOK/t-BuOH or CrCl2 (E/M)
Acid or base-A^Os 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
313
-------
Section 8G
Table 8-1 Continued
Parent Pesticide
Heptachlor
trans-Chlordane
cis- Jilordane
cis- and trans-
Chlordane
Endrin
Endosulfan
Metabolite
a) Chlordene
b) 1-Hydroxychlordene
c) l-Hydroxy-2,3-
epoxychlordene
2-Chlorochlordene
3-Chlorochlordene
1,2-Dichlorochlordene
epoxide
Photo-endrin
Endosulfan diol
Group Reacted
1} Allylie hydrogen
2} Double bond
1) Allylic hydroxy
2) Double bond
1} Hydroxyl
2) Epoxide
Double bond or
gem-dichloro group
Chloro epoxide or
gem-dichloro group
gem-Dichloro
Hydroxyl
Derivative
1-Bromochlordene
Chlordene epoxide
Silyl ether
Chloroacetate
Epoxide
Silyl ether
Trihydroxv
chlordane
Epoxide or
hexachloro
Chloroacetate
or heptachloro
Pentachloro
Acetate or silyl
ether
312
-------
Section 8G
Table 8-3
CONFIRMATORY DERIVATIZATION REACTIONS FOR OC1
PESTICIDES PUBLISHED SINCE 1975
Compounds Studied
Reagent (sensitivity)
Reference
HCB, PCBs and
naphthalenes as
interferences
Heptachlor and
its epoxide
Mirex, PCB inter-
ference
Phenoxyacetic
acid herbicides
HCH, DDT, PCBs
MCPA and MCPB
herbicides
10 Herbicidal acids
Dechlorination in GC flow-through
reactor containing carrier and H2
gases + Ni catalyst (0.1-1 ng)
UV light (0.01 ppm)
Diethylamine-assisted photo-
decomposition of PCBs (0.2-5 ppm)
2,2,2-Trichloroethanol (0.1 ppb)
MgO in GC microreactor
l-Bromomethyl-2,3,4,
5,6-pentafluorobenzene (0.1 ppb)
Pentafluorobenzyl bromide +
K2C03 catalyst (100-1000 pg)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
314
-------
Section 8G
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
pesticides (28) at 75-100 pg levels. Depending on the length of
irradiation (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
under these reaction conditions. Alkali reactions carried out on a
GC precolumn rather than in solution have proved advantageous in some
instances (29). Section 12,D,(2) of the EPA PAM describes the confirmation
of HCB in fatty tissues by formation of bis-isopropoxytetrachlorobenzene.
Section 11 of the Canadian PAM gives complete details for the following
tests:
Pesticide(s) Reagent
jo,£'-DDT, £,£*-DDT, £,£*-TDE, Sodium methylate
methoxychlor
£,£*-DDT, endrin Chromous chloride
Dieldrin, endrin BC13 in 2-chloroethanol
Chlordane, heptachlor K-tert butoxide/tert-butanol,
epoxide silylation
Aldrin, heptachlor, Chromic acid
£,£!-DDE
Aldrin m-Chloroperbenzoic acid
Endosulfan Alcoholic KOH
Chlorophenoxy acid n-Propanol
herbicides
Captan Resorcinol
315
-------
Section 8H
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 an HNC^-t^SO^ mixture. Organochlorine
pesticides are destroyed whereas PCBs (and toxaphene) are unaffected.
Other confirmation methods for PCBs are covered in Subsection 7A,E,e
in Section 7 of this Manual.
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 (30). This procedure does not distinguish
pesticides that produce the same hydrolysis product. According to
McCully (31), the three most practical methods for confirmation of
OP pesticides are oxidation to oxygen analogs (32), pentafluorobenzyl
bromide derivatization of hydrolyzed phenols or thiophenols (33), and
chromous chloride reduction (34). The sodium hypochlorite oxidation
method has the widest applicability, but it suffers from low sensi-
tivity, 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 8-4, along with information on triazine,
carbamate, and urea pesticides. This table is from a review article
(19) that gives the original references for these reactions. Table 8-5
contains a selection of more recent references.
Triazine herbicides have been confirmed by silylation, methoxylation
(in sodium methoxide-methanol) , methylation (C^I-NaH) , and hydrolysis-
DNFB reactions (35, 36); and linuron has also been confirmed by alkylation
(with alkyl halide - NaH)(35).
8H 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
determinations or analyses of a parent compound, metabolites, and
hydrolysis products.
316
-------
Section 81
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.
81 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 (51). Spectra-structure correlations can be of value
to the analyst in identifying chromophores and therefore making
confirmations, 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 identifications 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 great 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 spectral
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 (52). Visible and UV
spectrophotometric methods for pesticides have been reviewed (53),
including development of color by azo coupling and Si* complexing and
recent instrumental developments.
If a pesticide is naturally fluorescent or can be made fluorescent by
derivatization, fluorescence spectrophotometry is likely to be more
317
-------
CONFIRMATORY TESTS FOK OWiANOHlOSHlOKUS, TR1AZ1ME, UREft. AND CflflflftMATfi COMPOUHDh [ IV J
PeBtloide Class Compound Type
Organophospliates
Trlazinea
a) General
b) Phenol-generating compounds
c) Aryl-N0? and Aryl-CN
containing compounds
d| P= S compounds
e) -Nil end -NH, containing compounds
£.
f) OH compounds (dlazinon metabolites)
g) Cruf ornate
a) Chloro-s-triazlnes
b) Hydroxy-s-trlazinee
Hydrolysls/methylatlon
llydrolysis/PFB ether formation
Reduction (CrCl-.PdCl,,, Zn/IICl)
Z Z
Oxidation (to P =0)
l)Alkyatlon (NaH/Mel/DMSO)
i i) Deaini nation/me thy la t ion
Silylation or alkylation
UV deohlorination
1) Alkylation
ilj Silylation
ill Methoxylatlon
iv HydrolyslB/llNP formation
i S|lylatlon
11) Alkylation
ill) Chlorinatlon
Carbamates and
ureas
c) Cyanatine and Metabolites
a) Intact compound
b) Phenol-generating compounds
o) Amine-generatlng compounds
Chlorophenoxy a) Esters
acids
'Includes trifluoroaoetylation, pentnfluoropropylation,
Acid catalyzed oyalieation
i
il
Hi
Iv
i
11
ill
iv
v
vl
vil
1
11
Hi
Iv
V
vi
Acetylation
Silylation
Alkylation
Perfluorlnatlon*
Bromlnatlon
Chloroacetylatlon
Thiophosphorylation
Silylabion
Dichlorobenzene sulfonylation
DNT/DNP
Pentafluorobenzylation
Inclination
Dromination
p-Bromobenzoylatlon
2,4-DNP
DNT
Pentafluorobenr.ylatlon
(Amines in general)
i) Transesterlfloation
il)
Dromination
and heptafluorobutylatlon
318
-------
Table 8-5
Compounds Studied
CONFIRMATORY DERIVATIZATION REACTIONS FOR PESTICIDES
OF VARIOUS CLASSES PUBLISHED SINCE 1975
Reagent or Derivative (Senaltlvity)
Reference
-N0_-containing herbicides
and fungicides
Organonitrogen fungicides
and herbicides
Carbofuran and metabolites
OP pest)olden
Sulfoxlde-oontalnlng pesticides
Carbaryl
Carbamate Insecticides
Dlmllin (TH 60'tO)
Thlabendazole
N-Aryl carbamates
S-Contalnlng oarbamates
Carbamate and urea herbicides
Dlmllin (TH 60^0)
Azodrln
CrCl, reduction to -NH, followed by CCD-OC [37]
(0.5*1.0 ppm)
Alkylation, mathoxylatlon, trifluoroalkylation 1,38]
(0.1 ppm)
Heptafluorobutyrio anhydride plus trimethylamine [39.1
catalyst (10 pg)
In-blook methylatlon with THAM (ug levels)
Trlfluoroaoetio anhydride (1 ppm)
JJ-mono- and trlchloroacetyl, and g-nltroso derivatives
Heptafluorobutyryl derivatives (0.1 ppm)
Trifluoroacetyl derivative (0,02 ppm)
Pentafluorobenzyl chloride (0.01 ppm)
Flash heater reaction with trimethylanillnlum t^J
hydroxide (ng levels)
Trlmethylphenylammonlum hydroxide injected with
compound into gas chromatograph (20 ng)
Alkylation by NaH/CH,I (O.I ppm)
Conversion to N,N'-dimethyl analog with NaH/CH,I
(0.25 ng) }
Trlfluoroacetylatlon (2 ppb) [50]
319
-------
Section 8J
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 (54). Removal of naturally occurring
fluorescent interferences from biological samples can pose serious
cleanup problems. Fluorescence and phosphorescence methods for
pesticides have been reviewed by Argauer (55) and the phosphorimetry
of pesticides has been reviewed by Baeyens (56).
8J INFRARED (IR)
IR spectroscopy with micro sampling techniques is generally sensitive
at the 1 jig level but has been used as low as the 0.1 jug 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 8-B, and the technique for transfer of
the sample-KBr mixture adhering to the syringe needle is pictured in
Figure 8-C. The method using a commercially available "wick stick"
is probably the most reliable and foolproof for preparing micro KBr
pellets (Figure 8-D). 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.
320
-------
Section 8J
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 which 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 which is in effect 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 (57), 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 (58). An alternative micro-KBr technique
with sensitivity levels similar to the method in the EPA PAM is detailed
in the FDA PAM, Section 631.
321
-------
zze
-------
Section 8J
Finure 8-0. Illustration of
VHck-St1ck Method y
STAINLESS STEEL-
HOLDER
STAINLESS
STEEL CAP
SAMPLE CONCENTRATED
AT TIP
GLASS VIAL
KBr "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 for 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
containing isolated pesticide could be totally unsuitable for IR charac-
terization. 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 re-
sulting 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 jug quantities for
IR) and the destructive nature of pesticide detectors.
Several different types of IR detectors directly coupled with gas chroma-
tographs (59) have become available commercially but have not proven
especially useful for pesticide residue work because of various disad-
vantages. Trapping procedures have been used almost exclusively, in-
cluding the following methods:
323
-------
Section 8J
a. Passing column effluent through solvent (60, 61).
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 (62-64). 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 (65).
f. Trapping on Millipore or siliconized filter material (66, 67).
g. Using various types of liquid nitrogen or dry ice cold surface
traps (68).
h. Using a cool or cold small i.d. tubing at the GC vent (69).
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 (70) to
aid the analyst in matching spectra of unknown pesticides. The ASTM
FIRST-1 computer search program (59) 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 (71).
An important recent development in IR analysis that is capable of
sensitivity at subnanogram (72) residue levels is the Fourier transform
(FT) or interferometric m:Lhod. In FT-IR, a Michelson interferometer is
used instead of the prism or grating and slits in a conventional
spectrometer. The slitless spectrometer has an advantage in energy
throughput, in addition to the so-called Fellget's or multiplex advantage
324
-------
Section 8K
that allows all wavelengths to be detected simultaneously throughout
the spectral range. The signal to noise ratio increases with con-
secutive 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 conventional 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
(59). There is no doubt that much use will be made of FT-IR spectroscopy
for pesticide determination and confirmation as the principles, tech-
niques, and instrumentation become more familiar.
8K 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 (73) allow
routine acquisition of useful data on as little as 10 ;ug of a proton
NMR sample in a few minutes of experimental time. The NMR sensi-
tivity of l^c 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 (74).
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 distri-
butions.
Residues of _p_,j>'-DDT and _p_,£'-DDE isolated from adipose and liver tissue
samples have been analyzed by NMR (75), 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
!H spectra useful for identification of £,£'-DDT (76, 77), j3,£'-DDA (78),
aldrin and dieldrin derivatives and other chlorinated pesticides (79),
325
-------
Section 8L
rotenoids (80) , and dithiocarbamates (81) . Other ^H reference spectra
of organophosphorus (82), diphenylmethane (DDT type) (83), and carbamate
(84) pesticides have been published and are useful for identity con-
firmation. The application of NMR to pesticide analysis has been
reviewed (73, 85, 86).
Carbon-13 NMR spectra have been published for a-BHC (87) and for several
chlorinated biphenyls (88, 89). Studies of technical chlordane com-
ponents (90) , Mirex (91) , and Kepone and its photo-products (92) contain
13c 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,
aldrln, endrin, endosulfan, and dieldrin (93-95). 31P-NMR chemical
shifts have been correlated with structures of some organophosphorus
pesticides (96), and 31p Fourier transform NMR has been used for the
determination of malathion at ppm levels (97) .
8L MASS SPECTROMETRY (MS)
The mass spectrometer is a very sensitive spectroscopic tool for
pesticide residue analysis, providing useful data on ng or less
material. Ions are produced from neutral sample molecules and are
then sorted according to their mass-to-charge ratio. 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 characteristic of the unknown compound. Careful inspection of
key fragment ions in the spectrum can also assist in elucidating the
structure of an unknown compound even without comparison to a reference
material. These procedures are feasible with both low and high resolu-
tion spectrometers, but the exact elemental compositions of the parent
ion (if obtainable) and fragments provided by high resolution MS is a
valuable aid to the analyst.
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 addition, a vacuum must be maintained throughout
the spectrometer from inlet to detector so that ions formed in the
326
-------
Section 8L
source will not be lost from collisions with atmospheric gas
molecules. A sample is introduced to the ion source where it
is ionized via the inlet system. The function of the inlet system
is to transfer the sample from a high pressure (i.e., 1 atm) region
into the high vacuum of the spectrometer without seriously un-
balancing the spectrometer operation. The generated beam of ions
is focused and separated in the mass analyzer according to the
m/e 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. Excellent references for pesticide residue analyses
by MS have been written by Ryan (98), Safe and Hutzinger (99),
and Skinner and Greenhalgh (100).
(2) Inlet Systems: Combined GC-MS
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 raised until the sample vaporizes
and a spectrum is obtained. Techniques for trapping GC fractions for
introduction into an independent mass spectrometer have been published
(63, 101-103), and pesticides cleaned-up on and eluted from thin
layer plates have also been studied by MS.
For impure samples such as biological extracts, the gas chromatograph
of a coupled GC-MS instrument serves as an efficient inlet system for
introduction of samples into the spectrometer. The resolution pro-
vided by gas chromatography offers extra sample cleanup in addition
to any partition and liquid column 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 problem in GC-MS, since bleeding liquid phase
is also detected by the mass spectrometer 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 (10A).
327
-------
Section 8L
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 - 10~6 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. 20 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 instru-
ments. 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, jet, and silicone membrane; each has its own advantages
and limitations. 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 percent of the sample in the GC effluent to the mass spectro-
meter. The theory and operation of separators is described in
detail by McFadden (105). 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 spectrometer have been explored (106-111). As of this writing,
only one manufacturer offers a commercial liquid chromatograph-mass
spectrometer (LC-MS) interface (112). This interface consists of a
continuous belt that accepts the LC effluent in 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 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
belt before its return to the atmospheric chamber via the vacuum
locks.
328
-------
Section 8L
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/min.
The use of water as an LC solvent generally requires an 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
Q.I ml/min. The LC-MS system has been successfully applied to the
analysis of a large number of carbamate pesticides (112).
The field of LC-MS is still under development. Other commercial
LC-MS interfaces are anticipated in the near future.
(3) lonization Processes
The most widely used ionization source is the electron impact (El)
type wherein gaseous molecules are ionized by electrons emitted
from a 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.
At low electron energy levels (0-20/electron volts or eV), much 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 rearrangement 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. Because mass spectra are more reproducible
when compounds are ionized by 60-80 eV electrons, most mass spectro-
meters are operated in this energy range. It is noteworthy 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) spectra are obtained by adding methane,
helium, or other reagent gas (at relatively high pressures of about
1 mm 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 MS 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
329
-------
Section 8L
attachment. The mass spectra obtained with CI are quite different
from those formed on electron impact and are, in general, simple,
easy to interpret, and complementary to electron impact spectra for
unequivocal pesticide confirmation. Although CI usually provides
molecular ion (or M+ + 1 or M+ - 1) peaks of high intensity, a
study (113) of a series of chlorinated and organophosphorus pesti-
cides found no molecular ion produced from electron impact or
chemical ionization. CIMS has sensitivity at least as good as
that of El (114) and offers the advantage of allowing characteriza-
tion of a sample's chemical reactivity through the choice of the
reagent gases. In addition to methane and helium, isobutane,
hydrogen, argon-water, ammonia, nitric oxide, and other gases have
been used successfully to produce CI spectra. The positive and
negative methane (115) and isobutane (116) CI mass spectra of
selected chlorinated insecticides of several types and methane
positive ion CI data for 29 organophosphorus insecticides and
metabolites (117) have been determined and published. Negative
CIMS with methylene chloride reagent gas was the basis of a
multiresidue screening procedure of OC1 residues in environmental
substrates at 1 ng levels (118).
Field ionization (FI) involves passing a gaseous compound between an
anode (usually a thin wire or sharp blade) and a cathode. An
extraordinarily high electric field, approximately 10& V/cm, is
impressed on the anode, permitting (according to one theory) valence
electrons of the sample to "tunnel" to the metal of the wire or
blade. A positive ion results, which can be separated according
to mass-to-charge ratio and detected. This is a relatively low
energy or soft 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 (119), and FIMS has been
combined sequentially with HPLC for the determination of trifluralin
(120).
Field desorption MS is a modification of FI in which the sample is
applied directly to the anode. As with FIMS, field desorption depends
on application of very high electric fields 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
is a soft ionization process usually resulting in minimal sample
fragmentation. Unlike FI, field desorption has no requirement 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. Strong molecular ion peaks are produced for most
330
-------
Section 8L
pesticides (113) including highly polar pesticides and metabolites
(121). However, assignment of molecular ions in biological samples
can be complicated by the presence of (M + H)+, (M + Na)+, or other
ion adducts.
A novel method with unsurpassed sensitivity involves generation
of ions with an atmospheric pressure ionization (API) source. This
source uses 63fli 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 63^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"
(.122). The 63fli source has been replaced by a corona discharge (123),
producing identical API mass spectra and limits of detection but a
greater dynamic response range. The API source is not yet commercially
available and has not been evaluated for residue analysis, but it
obviously has great promise in this regard.
A discussion of 13 methods for ionization of organic compounds in MS
has been published, including detailed consideration of chemical
ionization and field ionization and pesticide spectra (124).
(4) Mass Analyzer Systems
Low resolution magnetic analyzer systems depend on deflection of the
ion beam in a magnetic field. The magnetic field classifies and
segregates the ions into beams, each of a different m/e. To obtain
the mass spectrum, the magnetic field is varied and each m/e ion from
light to heavy is successively brought to focus on the exit silt.
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/e ratio.
Quadrupole analyzers are based on mass separation in a radio frequency
(RF) electric field. This field is established on a set of four
precision parallel, usually circular, rods, with both a d.c. voltage
and an RF alternating voltage being applied to these electrodes.
331
-------
Section 8L
Ions are accelerated gently C5-30 V) into the analyzer or filter
region and begin to oscillate between the rods of rapidly alternating
polarity. At a given d.c. and RF level, ions of a specified m/e
value undergo stable oscillation and pass through the length of
the analyzer tube to the detector. Ions of lower or higher mass
will undergo ever increasing oscillations that eventually result
in their striking one of the rods. The spectrum is obtained by
sweeping the applied RF alternating frequency 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 AM/M, 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 percent, with
each peak contributing 5 percent 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 percent level. Low resolution mass spectrometers typically
show maximum resolution values between 300 and 1000 while high
resolution instruments are capable of attaining resolutions well
in excess of 10^. The advantage of a high resolution spectrometer
is the capability of resolving ions with very little differences in
mass and obtaining the molecular ion and fragment ion masses
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 (125). Once
the exact mass of a key ion (often the molecular ion) is known, the
elemental composition or formula of the fragment is obtained, again
by using a computer or by consulting tabulations of the masses of
different combinations of atoms indicated to be present by the mass
spectral pattern or prior information about the unknown sample.
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 a high resolution mass spectrometer reduces the
332
-------
Section 8L
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 sensitivity, only as much resolution as is
necessary to perform the required analysis should be used, since
the accuracy of an exact mass measurement is independent of resolu-
tion as long as any mass doublets are separated.
Descriptions of Finnigan and Hewlett Packard GC-MS systems and
detailed operating parameters are given in Section 8M. References
(98, 99, 101, 126, 127) review methods and applications of MS and
combined GC-MS to pesticide residue analysis, and references
(105, 128-130) give a more general survey of GC-MS instrumentation
and principles.
b. EXAMPLES OF GC-MS CONFIRMATION
Figure 8-E shows the electron capture gas chromatogram obtained
by injection of an aliquot of the 6 percent ethyl ether Florisil
column eluate from cleanup of a human adipose tissue extract (131).
Figure 8-F 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
superior specificity of the mass spectrometer. In general, chroma-
tograms traced by the total ion monitor are similar, but not
necessarily identical, in response and sensitivity to that traced
concurrently 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.
333
-------
Section 8L
FIGURE 8-E ELECTRON CAPTURE CHROMATOGRAM OF HUMAN ADIPOSE
TISSUE EXTRACT, 6X ETHER FLORISIL COLUMN ELUATE
INJECTIOf
FIGURE 8-F TOTAL ION CURRENT CHROMATOGRAM OF SAME HUMAN ADIPOSE
TISSUE EXTRACT
334
-------
Section 8L
Figure 8-G is the mass spectrum of the key £,£'-DDE GC peak evident
in both chromatograras in Figures 8-E and 8-F. The molecular ion
(M+) peak at 316 amu and the characteristic fragments at 281, 280,
and 246 would confirm the tentative identification of this peak as
_p_,£*-DDE obtained from comparing the retention time of the GC peak
to standard tables. Figure 8-H illustrates the total spectrum of
standard 2,£'-DDE for comparison to Figure 8-G.
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 percent 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 percent
ethyl ether eluate. Gel permeation chromatography has also been
successfully applied to the 6 and 15 percent fractions (132).
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 detector. These procedures require that the analyst
know what compound or compounds he 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 (133). To
use SIM effectively, one should know the kind of compound sought and
its MS characteristics. Sensitivity of detection for SIM can some-
times 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 reflects the extent of fragmentation and fraction of the
total ion current carried by the selected ions. Identification 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.
335
-------
RELATIVE INTENSITY
RELATIVE ABUNDANCE
8-
~
s-.
•
-
—
O ,
CO ~
u?
8-
o .
-
8:
-
-
=~
•-
— —
== —
fl 6,
^ . C'r^§
? ? V
fe, - S
ij-i
1 1 1
g
§•
? K 5 ? I
O
5
NOUVZINOI 1W01 *
CO
(D
NOIlVZINa 1V101
O
O
-------
Section 8L
Total ion current plots (TICP) cannot be generated by the SIM
technique because data from only certain masses are collected. A
plot of the change in ion abundance as a function of time using
abundances measured by SIM is termed a selective ion current
profile (SICP). These and other nomenclature problems have been
discussed along with suggested remedies by Budde and Eichelberger
(129). Compounds not resolved by gas chromatography can still be
detected with certainty if their molecular (or other selected)
ions can be resolved 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 (134)
and to carbofuran and metabolites in crops (135).
Repetitive scanning through a narrow mass range generates quanti-
fiable spectral envelopes from several ions at once. This procedure,
generally sensitive at low ng levels, has been applied to pesticide
analysis (136).
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 decrease when they react with material eluted from
the GC column, providing a chromatogram that is distinctive from
those produced by other detectors (130).
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. Computer control of data acquisition may enable the
operator to generate 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 discontinuous mass ranges may be sampled.
(2) Column bleed and other background can be conveniently subtracted
by the computer.
337
-------
Section 8L
(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 de-
tected in each scan but excluding carrier gas ions or other inter-
fering ions. Total ion profile chromatograms obtained resemble
those traced in real time by a conventional total ion monitor of
a magnetic deflection spectrometer.
(4) The computer can trace the intensities of selected characteristic
masses from among the great quantity of data acquired by continuous
repetitive scanning. The resulting mass chromatograms or extracted
ion current profiles (EICP) (129) 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 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 integration time
characteristic of the latter method (129). Reference (127) illustrates
computer-generated individual ion chromatograms.
The limited mass range chromatogram (130) is a variation of mass
chromatography 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 3?C1
from each of the 12 chlorine atoms, is spread over a range of more
than 27 amu. Instead of treating a single ion (e.g., C^o^Sci-^*,
nominal m/e 540), the entire cluster can be summed to provide increased
sensitivity with some sacrifice in specificity. The method has been
used to identify dieldrin and HCB residues in lake trout (137).
(5) Quantitation of peak areas in the selected ion profiles and
ratios of these peaks can be provided. Computer coupled GC-MS equip-
ment is extremely expensive, highly qualified personnel is needed for
its operation and maintenance and to interpret data, and a significant
338
-------
Section 8M
amount of "down-time" is to be anticipated because of its complexity.
Computerized data acquisition and processing for magnetic instruments,
quadrupole instruments, and selected ion monitoring have been de-
scribed (98).
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 (96, 99, 138,
139). Applications of GC-MS include confirmation of the 1-naphthyl
chloroacetate derivative of 1-naphthol (a carbaryl metabolite)
extracted from urine (140); 2,4-D, 2,4,5-T, and 2,4,5-TCP in urine
(141, 142); organophosphorus pesticides in blood and urine (143, 144);
multiple chlorinated insecticides in human adipose and liver tissue
(131, 132, 145) or foods (146); toxaphene in human and biological
samples (147); Kepone in human and environmental samples (148);
chlordane-related residues in human samples (131); and thiabendazole
and 5-hydroxythiabendazole in animal tissue (on-column methylation
plus SIM) (149). An important application of GC-MS has been mutual
determination and identification of PCBs in the presence of chlorinated
pesticides (150). 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 (151). GC-MS has been
successfully applied to the detailed analysis of complex pesticide
mixtures, such as technical chlordane (152). 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 (153). 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 (154, 155), glass capillary techniques (156),
field ionization (157) , and field desorption MS (158). Methods have
also been developed for the determination of carbamates by combined
liquid chromatography-mass spectrometry (112). References (98, 126,
159, 160) contain further reviews of applications to residue analysis.
1 QUALITY ASSURANCE OF GC-LOW RESOLUTION MS
This section reviews procedures to be followed for quality assurance of
data derived from the mass spectrometer in the identification, confirmation,
and quantitative determination of chlorinated insecticides, PCBs, hexa-
chlorophene, and PBBs in human tissues and fluids. These methods were
339
-------
Section 8M
developed at the Health Effects Research Laboratory, U.S. EPA, Research
Triangle Park, NC (161) 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:
oi,£'-DDT Aldrin
£,£?-DDT Dieldrin
£,£?-DDE Heptachlor
£,£*-DDE Heptachlor epoxide
os£'-DDD Endrin
£,£'-DDD Mirex
0,-BHC Oxychlordane
3-BHC trans-Nonachlor
Lindane (y-BHC) Polychlorinated biphenyls
6-BHC Hexachlorobenzene
Polybrominated biphenyls
Polychlorinated terphenyls
INTRODUCTION TO THE QUALITY ASSURANCE PROCEDURES
Correct identification of organic pollutants from gas chromatography-
mass spectrometry (GC-MS) data require valid mass spectra of the com-
pounds detected. This is independent of the actual method of interpreta-
tion 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 fragmentation. A properly operating and well tuned GC-MS
is required to obtain valid mass spectra.
The purpose of the following procedure is to permit a check of the
performance of the total operating computerized GC-MS system. Thus,
with a minimum expenditure of time, an operator can be reasonably sure
that the GC 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. 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.
340
-------
Section 8M
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. (See Subsections g and
h for descriptions and operating parameters of these instruments.)
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 spectro-
meter, 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 amu 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.
Assurance of mass spectral data is obtained through a set of two levels
of functionality tests. The first test requires establishment of
production, 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/e = 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 heptachlor
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/e 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/e cluster may be examined
at 100, 10 and 1 ng levels, as instrument sensitivity requires, with
341
-------
Section 8M
respect to appearance of the six-chlorine cluster versus statistical
appearance. Resolution of ^C 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 rdutinely 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
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/e ions are present and
in reasonable relative abundance according to the following
tabulation:
Genera.]. Desired Appearance of the Mass Spectrum of PFTBA
Mass Relative
(m/e) Abundance %
69 100.0
100 24.3
114 9.5
119 20.3
131 70.3
219 68.0
264 16.2
414 5.4
426 2.7
502 2.7
614 0.3
Isotope Abundance Checks, Percentage Ratio of
Ion Signal Abundances
(70)/(69) - 1.1%
(219)/(220) - 4.4%
(502)/(503) =10.3%
342
-------
Section 8M
(b) 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 reconstructed
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.
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/ul 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 DECAFLUOROTRIPHENYL PHOSPHINE (DFTPP) STANDARDS
Prepare a stock solution of DFTPP at 1 rag/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 iig/ml 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.
343
-------
Section 8M
(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.
(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 provide good
mass spectra.
Reference Aldrin Mass Spectrum
(5-chlorine cluster check)
Reference Heptachlor Epoxide Mass Specti
(6-chlorine cluster check)
m/e
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
(%)
m/e
351
352
353
354
355
357
359
361
363
(351)
(81) *//0
Abund. (%)
51.2
5.6
100.0
11.2
81.2
35.2
8.5
1.1
0.06
344
-------
Section 8M
Reference Mass Spectrum of DFTPP
Mass Ion Abundance Criteria
51 30-60% of mass 198
68 Less than 2% of mass 69
70 Less than 2% of mass 69 (1.1% theoretical)
127 40-60% of mass 198
197 Less than 1% of mass 198
198 Base peak, 100% relative abundance
199 5-9% of mass 198 (6.6% theoretical)
275 10-30% of mass 198
365 1% of mass 198
441 , Less than mass 443
442 (M*) 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.
443 (M+l) 17-23% of mass 442 (19.8% theoretical)
444 (M+2) 1.86% (theoretical)
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
individuals sampled. The blood is drawn into evacuated ampoules,
allowed to clot, and centrifuged, and the serum is decanted into a
clean vial.
345
-------
Section 8M
(2) Clean-up
Tissues are normally extracted and cleaned up according to a modified
Mills-Olney-Gaither procedure (Subsection 7A) by laboratories under
contract to the National Human Monitoring Program. Concentrated
extracts, corresponding to the 6 percent and 15 percent 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 Subsection j. Blood sera samples
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 ul, but it may be smaller if levels of compounds sought
are particularly low. Qualitative analysis is performed in the electron
impact (El) mode. Aliquots of 5 to 50 /il are co-injected with aldrin
(e.g., 250 ng in hexane) as an internal standard into the GC-MS system.
A total ion chromatogram 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., 96, 99, 126, 145, 159), through the Cyphernet™
reference library, or generated from authentic laboratory standards.
Relative retention times are also compared to those of the reference
material for further confirmation. After identification^ quantitative
analyses are usually performed by selected ion monitoring (SIM) . An
authentic reference sample is used for direct comparison (See Sub-
section i). Identification may be further confirmed by chemical
ionization GC-MS where available (Subsection h).
g. DESCRIPTION OF HEWLETT PACKARD GC-MS SYSTEM AND OPERATING PARAMETERS
The Hewlett Packard 5930A quadrupole-focusing mass spectrometer is
interfaced with a Hewlett Packard 5700A gas chromatograph and Hewlett
Packard 5933A data system. The mass spectrometer is operated in the
electron impact (El) ionization mode typically at 70 eV electron
energy and 120 yamps filament emissions with a target current of
100 pamps. Selected ion monitoring (SIM) runs are normally made at
100 yamps and high electron multiplier voltages. The ion source
temperature is usually maintained at 180°C and the mass filter
346
-------
Section 8M
temperature at 110°C. Ions are detected by a Bendix continuous dynode
electron multiplier with voltages varied between 1.5 and 3.0 Kv, as
required for sufficient signal strength. Data are collected, stored,
and plotted by the 5933A data system. Selected ion monitoring is
controlled by a suitable program provided by the manufacturer and by
the data system. Data output also involves the use of a light beam
oscillograph, as well as plotted mass spectra and reconstructed total
ion chromatograms from tha data system.
The gas chromatograph is often equipped with 1.8m x 2mm (i-d.) glass
columns packed with 1.5% OV-17/1.95% OV-210 or 3% OV-17 on 80-100 mesh
Gas Chrom Q. The GC inlet temperature is normally 200°C and the helium
carrier gas flow rate 40 to 45 ml/min. Isothermal runs are typically
made at 180° to 190°C, while temperature programming is usually from
80°C (2 min) to 210°C (or higher) at 8°C/min. The transfer line to
the mass spectrometer is maintained at 210°C, and the silicone membrane
separator is kept at 180°C.
Detailed operating instructions for the Hewlett Packard mass spectro-
meter, gas chromatograph, and data system are given in the appropriate
manufacturer's operating manuals.
h. DESCRIPTION OF FINNIGAN GC-MS SYSTEM AND OPERATING PARAMETERS
(1) Equipment
The Finnigan Model 3200 quadrupole mass spectrometer is equipped with
both electron impact and chemical ionization vacuum consoles. Each
vacuum console is interfaced to a dedicated Finnigan Model 9500 gas
chromatograph. Mass spectra and chromatographs may be obtained
"manually" by the use of a light beam oscillograph and strip chart
recorder, respectively. Alternatively,, the collection and storage
of data may be placed under the control of a Finnigan Model 6100
data system. The data system is capable of data collection in either
scanning or a selected ion monitoring (SIM) mode.
Detailed operating instructions may be found in the appropriate
Finnigan manuals.
(2) GC Conditions
Column: 150 cm x 2 mm i.d. glass column silanized before packing.
Packing: 1.5% OV-17/1.95% OV-210 on 80-100 mesh Gas Chrom Q.
347
-------
Section 8M
Freshly packed columns are baked at 275°C for at least 48 hrs with
a carrier gas flow. The oven temperature is then reduced to 250°C.
The column is then treated with five 20 ul injections of a silating
reagent such as Silon-CT at 10 minute intervals. The testing of
standards for column evaluation is begun not less than 30 minutes
after the final injection.
Carrier: Carrier gas flow is regulated to approximately 20 ml/min.
Helium is used as a carrier gas for electron impact ionization studies
and for those chemical ionization studies in which the reagent gas is
not suited for gas chromatography. Methane is the routinely used
chemical ionization reagent gas, and it is suitable for gas chromatography
Column oven: In order to obtain sharp, resolved chromatographic peaks
for most of the mixture components, the column oven is normally
operated under a temperature program. The initial run is generally
made with a 1 min isothermal hold at 150°C followed by a programmed
increase of 2°/min to 275°C.
Injector: 200-225°C
Transfer line: 200°C
Separator oven: 250°C
Subsequent runs are made under conditions which, in the opinion of
the operator, will enhance regions of interest in the chromatogram.
(3) MS Conditions
For qualitative studies, the mass spectrometer is operated in an
electron impact (El) ionization mode under control of the data system.
The operator variable controls are set to the following values:
Preamp: 10~° amp/volt
Electron multiplier: 2 kV
Emission current: 0.5 ma
Electron energy: 70 eV
Analyzer heater: 50°C
The lens, repeller, ion energy, and resolution controls are adjusted
on a day-to-day basis to obtain an optimum spectrum of the reference
compound perfluorotri-n-butylamine (PFTBA or FC43) as per the detailed
instructions in the Finnigan "3200/3300 GC/MS Systems Operation Manual".
348
-------
Section 8M
Chemical ionization with methane as the reagent gas is often used
for confirmation. Operator variables are as follows:
Q
Preamp: 10 amp/volt
Electron multiplier:2 kV
Emission current: 1.0 ma
Electron energy: 110 eV
Analyzer heater: 70°C
Source temperature: 120°C
Source pressure: 900-1000 microns
The instrument is optimized for CI studies as per the instructions
in the Finnigan "Chemical Ionization System: Models 3100E, 3200E,
3300E; Operation and Maintenance Manual".
(4) Data Systems Parameters
The data system is generated in a scanning mode for qualitative studies.
Scan time is held to approximately 5 sec/scan in order to retain as
much chromatographic resolution plus sensitivity as possible. In
order to achieve this, the scan is generally restricted to 450 amu.
For chlorinated compound analyses (e.g., PCBs, chlordane, DDE, etc.),
the following parameters are used:
Mass range: 100-150, 151-230, 231-430, 431-550
Integration time: 3, 6, 12, 18 milliseconds/amu
Seconds per scan: 5
Threshold: 1
Maximum run time: 90 minutes
Instrument range setting: E
For quantitative analyses, selected ion monitoring is normally used.
The exact parameters are sample and component dependent. The masses
to be monitored are chosen at the time of the analysis.
i. QUALITY CONTROL OF SELECTED ION MONITORING DATA
(FOR HEWLETT PACKARD SYSTEM)
Operation in the selected ion monitoring mode is verified by injection
of known levels of the compound to be determined and examination of
the resulting signal levels. Tuning of such instrumental parameters
as the peak width repeller voltage and ion focus is usually required,
and high voltage on the electron multiplier must be used for maximum
sensitivity. Usually several test SIM runs are required to obtain
sensitivity. In addition, the following precautions are to be
routinely exercised:
349
-------
Section 8M
(1) Run blank checks involving all solvents, glassware (evaporators,
etc), and microliter syringes to verify the absence of the compound
in question. Blanks may commonly consist of ca 10 ml of solvent
evaporated to 0.1 ml in an evaporator tube in a stream of nitrogen,
followed by injection of 20 ul into the GC-MS system.
(2) Study the sample using scanning mass spectrometry in an attempt
to ascertain if interfering ions are present from sample components
that may tend to "swamp" or interfere in retention time with the
compound to be monitored. The scanned spectra obtained add to the
qualitative evidence of the presence of the compound.
(3) Bracket samples with levels of known standards.
(4) Co-inject standards of approximately equal levels with the
sample to be monitored and observe the signal to see if it is doubled.
Also determine whether the identity is consistent with the GC re-
tention time.
(5) GC column resolution may decrease as sequential fat injections
are made. This may be partially corrected by removing the glass
wool plug from the on-column injection port of the gas chromatograph.
However, the GC column will usually have to be changed as a study
progresses.
(6) Intersperse blank solvent injections after standard injections
in order to verify the absence of "hang-up" of any of the compounds
to be monitored in any component of the system, with subsequent release
by the solvent "burst" of the following injection.
Careful attention to the above protocol will produce results as "real"
as can be obtained by low resolution GC-MS. Caution must be exercised,
however, because low resolution SIM data combined with a correct re-
tention time is not always sufficiently specific to define the compound
qualitatively. For example, tetrachlorodibenzo-_p_-dioxin monitored ions
may be interfered with by PCBs.
Readers interested in further information on this topic are referred to
reference (154).
j. GEL PERMEATION CHROMATOGRAPHIC CLEANUP OF ADIPOSE TISSUE
(1) Theory
Gel permeation chromatography (GPC) is a form of liquid chromatography
by which compounds are separated according to molecular size. It is
350
-------
Section 8M
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 (162) and has the added advantages that removal of fat
is more complete and recoveries of pesticides are nearly quantita-
tive. 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., Bio-Beads SX-3) are used as gel particles
and organic solvents (e.g., toluene, ethyl acetate, or cyclohexane)
are used for the mobile phase. The elution process is very simple
(isocratic only); the same solvent system is used for column prepara-
tion, elution, and washing. Macromolecules cannot permeate the
porous gel and are rapidly eluted or "dumped" from the column.
Molecules which can enter the pores of the beads are temporarily
retained to greater or lesser extents depending on their molecular
volumes. Hence, large-volume pesticides 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 3-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
(i.d.) x 60 cm glass column (ChromaflexR R-422350/6025, Kontes, Vineland,
NJ, or equivalent) packed with 200- to 400-mesh Bio-Beads SX-3 (Bio-Rad
Laboratories, Richmond, CA).
(3) Column Preparation and Operation
(a) Prepare a slurry of ca 60 g of Bio-Beads 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.
351
-------
Section 8N
Add only toluene-ethyl acetate (1:3 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/min 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
toluene-ethyl acetate (1:3 v/v) until it is purged of entrained
air.
(b) Introduce £l gram of sample in 8 ml of eluting solvent into
the sample introduction valve. Rotate to the next sample loop
and introduce the next sample. After the last sample is loaded,
rinse the sample valve with clean solvent.
(c) Set "dump" and "collect" cycles to previously established
rates for elution of the pesticides of interest (usually 100-125 ml
and 125-225 ml, respectively).
(d) For multiple sample cleanup, an appropriate "wash" cycle of
ca 50 ml toluene-ethyl acetate (1:3 v/v) should be included after
every sample.
8N BIOLOGICAL METHODS
Bioassay techniques, which include insecticidal activity, enzymatic,
and immunological methods, have been described as providing an inde-
pendent criterion of identity when combined with GC, chemical reactions,
etc. (2). These methods, which depend on the measurement of a physio-
logical response of a test organism induced by exposure to the pesticide,
352
-------
Section 80
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 (163).
Specificity of enzyme inhibition is greatly enhanced by combination
with TLC for detection and confirmation of organophosphate and certain
carbamate pesticides. The Rp value plus biological response provides
important identity information at levels typically in the range of
500 pg to 10 ng for these compounds.
80 POLAROGRAPHY (VOLTAMMETRY)
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 determination of pesticide residues have been reviewed (164, 165).
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.
A collaborative study confirmed the usefulness of single sweep oscill—
ographic polarography for identifying such residues in non-fatty foods
(166). Thirty-eight herbicides have been studied by single sweep
derivative polarography (167), methylcarbamate insecticides by AC
polarography and cyclic voltammetry (168), and urea herbicides by
anodic polarography (169). Published voltammetric reduction potentials
for about 100 organochlorine insecticides, PCBs and naphthalenes
(3 electrode potentiostat, DMSO solvent) are a useful aid in identifi-
cation of residues (170). Parathion and related insecticides and
metabolites were polarographically determined in blood without
extraction (171). The polarography of 1,3,5-triazines (172), propachlor
353
-------
Section 8P
herbicide (in soil) (173), dithiocarbamates (174), dinitroaniline
herbicides (175), thiourea-containing pesticides (176), and
trifluralin (in soils) (177) has been reported.
8P 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 precolunm containing a hot (ca 300°C) catalyst
attached to a GC 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 instrument
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 derivatization 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 (178-181) and Asai et al
(182, 183). Identification of 5-10 ng amounts of polychlorinated
biphenyls, terphenyls, naphthalenes, dioxins, and dibenzofurans in
biological samples has been demonstrated (184).
b. FRAGMENTATION PROCEDURES
GC fragmentation procedures are similar to CSC except that the reaction
in the precolumn decomposes the pesticides, yielding characteristic
fragment peak patterns or fingerprint chromatograms helpful in making
identifications. A palladium catalyst at 300°C (182) and reagents
such as Na2C03, CuO, CdCl2, A1C13, and K2Cr207 at 240°C (185) have
been applied to chlorinated and OP insecticides with EC detection of
the reaction products.
354
-------
Section 8Q
8Q REFERENCES
(1) Analytical Methods for Pesticide Residues in Foods, Depart-
ment of National Health and Welfare, Canada, 1973, Section
10.2.
(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
Concentrations, Advances in Chemistry Series 104, Chapter 10,
ACS, Washington, B.C., 1971, page 151.
(4) Onuska, F. I., and Comba, M. E., J. Chromatogr., 119,
385 (1976).
(5) Ruzicka, J. H. A., and Abbott, D. C., Talanta, 20, 1277 (1973).
(6) Aue, W. A., and Kapila, S., Anal. Chem., 50, 536 (1978).
(7) Mallet, V., J. Chromatogr.. 7^, 217 (1973).
(8) Lawrence, J. F., and Frei, R. W., Chromatogr. Rev., 18, 253 (1974).
(9) Connors, K. A., Anal. Chem., 46, 53 (1974).
(10) Heinz, D. E., and Vitek, R. K., J. Chromatogr. Sci. . .13, 570 (1975)
(11) Dolan, J. W. , and Seiber, J. N., Anal. Chem., 49_, 326 (1977).
(12) Beroza, M., and Bowman, M. C., J. Ass. Offic. Anal. Chem., 48,
358 (1965).
(13) Bowman, M. C., and Beroza, M., J. Ass. Offic. Anal. Chem., 48,
943 (1965).
(14) Beroza, M., and Bowman, M. C., Anal. Chem., 37, 291 (1965).
(15) Beroza, M., and Bowman, M. C., Anal. Chem., 38. 837 (1966).
(16) Bowman, M. C., and Beroza, M., Anal. Chem., 38. 1427 (1966).
(17) Drozd, J., J. Chromatogr.. 113, 303 (1975).
355
-------
Section 8Q
(18) 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.
(19) Cochrane, W. P., J. Chromatogr. Sci., 13, 246 (1975).
(20) Kapila, S., and Aue, W. A., J. Chromatogr. Sci., 15, 569 (1977).
(21) Ward, P. M. , J. Ass. Offic. Anal. Chem. , j>0, 673 (1977).
(22) Lewis, R. G., Hanisch, R. C., McLeod, K. E., and Sovocool, G. W.,
J. Agr. Food Chem., 2A_, 1030 (1976).
(23) Mierzwa, S., and Witek, S., J. Chromatogr., 136, 105 (1977).
(24) Luckas, B., Pscheidl, H., and Haberland, D., J. Chromatogr.,
147. 41 (3978).
(25) Agemian, H., and Chau, A. S. Y., Analyst, 101, 732 (1976).
(26) Chau, A. S. Y., and Terry, K., J. Ass. Offic. Anal. Chem., 59,
633 (1976).
(27) Glotfelty, D. E., Anal. Chem., 44^ 1250 (1972).
(28) Banks, K. A., and Bills, D. D., J. Chromatogr., 33, 450 (1968).
(29) Cochrane, W. P., and Maybury, R. B., J. Ass. Offic. Anal. Chem.,
5£, 1324 (1973).
(30) Shafik, T. M., Bradway, D., and Enos, H. F., Bull. Environ.
Contain. Toxicol. , j5, 55 (1971).
(31) McCully, K. A., J. Ass. Offic. Anal. Chem., 60, 376 (1977).
(32) Singh, J., and Lapointe, M. R., J. Ass. Offic. Anal. Chem., 57,
1285 (1974).
(33) Coburn, J. A., and Chau, A. S. Y., J. Ass. Offic. Anal. Chem., 57,
1272 (1974); Environ. Lett.. 10, 225 (1975).
(34) Forbes, M. A., Wilson, B. P., Greenhalgh, R., and Cnchrane, W. P.,
Bull. Environ. Contarn. To:;Acol., 13, 141 (1975;,
(35) Cochrane, W. P., and Greenhalgh, R., Confirmation of Pesticide
Residue Identity by Chemical Derivatization, presented at 166th
National ACS Meeting, Chicago, August 29, 1973: Khan, S. U.,
Greenhalgh, R., and Cochrane, W. P., J. Agr. Food Chem., 23, 430 (1975)
356
-------
Section 8Q
(36) Lawrence, J. F. , J. Agr. Food Chem. , _2jZ, 936 (1974).
(37) Lawrence, J. G., Lewis, D., and McLeod, H. A., J. Agr. Food
Chem.. 25, 1359 (1977).
(38) Lawrence, J. F., J. Agr. Food Chem., 24, 1236 (1976).
(39) Lawrence, J, F., Lewis, D. A., and McLeod, H. A., J. Chromatogr.
138, 143 (1977).
(40) Miles, J.W. , and Dale, W. E. , J. Agr. Food Chem., 2j±, 480 (1978).
(41) Greenhalgh, R., King, R. R., and Marshall, W. D., J. Agr. Food
Chem., ^6_, 475 (1978).
(42) Magallona, E. D., and Gunther, F. A., Arch. Environ. Contam.
Toxicol.. 1, 185 (1977).
(43) Lawrence, J. F., J. Chromatogr., 123, 287 (1976).
(44) Worobey, B. L., and Webster, G. R. B., J. Ass. Offic. Anal. Chem.,
60, 213 (1977).
(45) Noshe, N., Kobayoshi, S., Tanaka, A., Hirose, A., and Watanabe, A.,
J. Chromatogr., 130. 410 (1977).
(46) Wein, R. G., and Tanaka, F. S., J. Chromatogr.. 130, 55 (1977).
(47) Bromilow, R. H., and Lord, K. A., J. Chromatogr., 125, 495 (1976).
(48) Lawrence, J. F., J. Ass. Offic. Anal. Chem., 59, 1061 (1976).
(49) Lawrence, J. F., and Sundaram, K. M. S., J. Ass. Offic. Anal.
Ch_em. , 59_, 938 (1976).
(50) Lawrence, J. F., and McLeod, H. A., J. Ass. Offic. Anal. Chem.,
59^, 637, (1976).
(51) 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.
(52) Gore, R. C., Hannah, R. W., Pattacini, S. C., and Porro, T. J.,
J. Ass. Offic. Anal. Chem., 54, 1040 (1971).
357
-------
Section 8Q
(53) 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.
(54) MacDougall, D., Residue Rev.. 1., 24 (1962); 5_, 119 (1964).
(55) 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.
(56) Baeyens, W., Pharm. Weekbl.. 111. 1075 (1976).
(57) Blinn, R. C., The Identification of Pesticides at Residue
Concentrations, Advances in Chemistry Series 104, Chapter 6,
ACS, Washington, D.C., 1971, page 81.
(58) Kettmann, R., Closset, J.-L., Copin, A., Duculot, C., and
Martens, P.-H., Anal. Lett.. J5, 1013 (1973).
(59) 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.
(60) Crosby, N. T., and Laws, E. Q., Analyst, J39, 319 (1964).
(61) Neiswender, D. D., J. Gas Chromatogr., _4, 426 (1966).
(62) Shearer, D. A., Stone, B. C., and McGugan, W. A., Analys t, 88,
147 (1963).
(63) Bierl, B. A., Beroza, M., and Ruth, J. M., J. Gas Chromatogr.,
j6, 286 (1968).
(64) Scanlon, R. A., Arnold, R. G., and Lindsay, R. C., J. Gas
Chromatogr.. j>, 372 (1968).
(65) Nano, G. M., Sancin, P., and Martelli, A., J. Gas Chromatogr., _3>
85 (1965).
(66) Hannah, R. W., and Dwyer, J. L., Anal. Ghent., 36, 2341 (1964).
(67) Hammerstrand, K., Juntunen, J. M., and Henes, A. R., Anal. Biochem.
21, 172 (1969).
358
-------
Section 8Q
(68) Curry, A. S., Read, J. F., Brown, C., and Jenkins, R. W.,
J. Chromatogr. ,J38, 200 (1968).
(69) Hartman, K. T., J. Ass. Offic. Anal. Chem.. 50., 615 (1967).
(70) Giang, P. A., Library of Infrared Spectra of Important Pesticides,
in Analytical Methods for Pesticides and Plant GrowthRegulators,
Zweig, G., and Sherma, J., eds., Volume IX, Chapter 6, Academic
Press, New York, 1977, page 153.
(71) Nicholas, M. L., J. Ass. Offic. Anal. Chem.. 59., 197, 1071, 1266
(1976).
(72) Cournoyer, R., Shearer, J. C., and Anderson, D. H., Anal. Chem.,
49, 2275 (1977).
(73) Cyr, N., Cyr, T., and Haque, R., Nuclear Magnetic Resonance
Spectroscopy, 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.
(74) Wilson, D. M., Olsen, R. W., and Burlingame, A. L., Rev. Sci.
Instrum.. 45, 1095 (1974).
(75) Biros, F. J., J. Ass. Offic. Anal. Chem.. ,53, 733 (1970).
(76) Wilson, N. K., J. Am. Chem. Soc. . 94., 2431 (1972).
(77) Ross, R. T., and Biros, J. F., Biochem. Biophys. Res. Commun.,
.39, 723 (1970).
(78) 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.
(79) 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.
(80) Crombie, L., and Lown, J. W., J. Chem. Soc., 1962, 775.
(81) Wilson, N. K., J. Phys. Chem.. 75, 1067 (1971).
359
-------
Section 8Q
(82) Keith, L. H., Garrison, A. W., and Alford, A. L., J. Ass. Offie.
Anal. Chem., 5J., 1053 (1968).
(83) Keith, L. H., Alford, A. L., and Garrison, A. W., J. Ass. Offie.
Anal. Chem.. 5>2, 1074 (1969).
(84) Keith, L. H., and Alford, A. L., J. Ass. Offie. Anal. Chem..
j>3, 157 (1970).
(85) Wilson, N. K., Nuclear Magnetic Resonance Spectroscopy in
Pesticide Chemistry, in Pesticide Analysis. Das, K. G., ed.,
Marcel Dekker, Inc., New York, 1978.
(86) Keith, L. H., and Alford, A. L., J. Ass. Offie. Anal. Chem.,
_53, 1018 (1970).
(87) Yamamoto, 0., Yanagisawa, M., Hayamizu, K., and Kotowycz, G.,
J. Magn. Reson. . 9^, 216 (1973).
(88) Wilson, N. K., and Anderson, M., Mass Spectrometry and NMR
Spectroscopy in Pesticide Chemistry, Plenum Press, New York, N.Y.,
1974, pp. 197-218.
(89) Wilson, N. K., J. Am. Chem. Soc., 9^7, 3573 (1975).
(90) Wilson, N. K., and Sovocool, G. W., Org. Magn. Reson., J3,
536 (1977).
(91) Alley, E. G., Layton, B. R., and Minyard, J. P., Jr., J. Org.
Chem.. 41, 462 (1976).
(92) Wilson, N. K., and Zehr, R. D., J. Org. Chem., in press (1978).
(93) Roll, D. B., and Biros, F. J., Anal. Chem.. _41, 407 (1969).
(94) Gegiou, D., Anal. Chem.. _46, 742 (1974).
(95) Gegiou, D., Talanta, _21, 889 (1974).
(96) Ross, R. T., and Biros, J. F., Anal. Chem., 52. 139 (1970).
(97) Gurley, T. W., and Ritchey, W. M., Anal. Chem.. _48, 1137 (1976).
(98) Ryan, J. F., Residue Analysis Applications of Mass Spectro-
metry, in Analytical Methods for Pesticides and Plant Growth
Regulators, Zweig, G., and Sherma, J., eds., Volume IX, Chapter 1,
Academic Press, New York, 1977, page 1.
360
-------
Section 8Q
(99) Safe, S., and Hutzinger, 0., Mass Spectroscopy of Pesticides
and Pollutants, CRC Press, Cleveland, OH, 1973.
(100) 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.
(101) Biros, F.J., Residue Rev.. 40, 1 (1971).
(102) Burson, K. R., and Kenner, C. T., J. Chromatogr. Sci. , _7,
63 (1969); Amy, J., Chait, E. M., Baitinger, W. E., and
McLafferty, F. W., Anal. Chem.. 3_7, 1265 (1965).
(103) Orland, R. K., Clock, E., and Bodenhamer, N. L., J. Chromatogr.
Sci., 1_, 187 (1969).
(104) Grayson, M. A., J. Chromatogr. Sci., 15, 539 (1977).
(105) McFadden, W. H., Techniques of Combined Gas Chromatography/
Mags Spectrometry; Applications in Organic Analysis, John
Wiley & Son, New York, 1973.
(106) McLafferty, F. W., Knutti, R., Venhataraghavan, R., Arpino, P. J.,
and Dawkins, B. G., Anal. Chem., 47, 1503 (1975) and references
therein.
(107) Jones, P. R., and Yang, S. K., Anal. Chem., _47, 1000 (1975).
(108) Carroll, D. I., Dzidic, I., Stillwell, R. N. , Haegle, K. D., and
Horning, E. C., Anal. Chem., _47, 2369 (1975).
(109) Scott, R. P. W., Scott, C. G., Munroe, M., and Hess, J., Jr., J_.
Chromatogr., 99_, 395 (1975).
(110) Blakley, C. R., and Vestal, M. L., paper No. D7, 25th Annual
Conference on Mass Spectrometry and Allied Topics, Washington,
DC, 1977.
(Ill) McFadden, W. H., U.S. Patent 4,055,987 (1977).
(112) Wright, L. H., and Oswald, E. 0., Liquid Chromatography/Mass
Spectrometric Methods of Analyses for Carbamate Pesticides,
/resented at the 26th Annual Conference on Mass Spectrometry
and Allied Topics, St. Louis, MO, May-June 1978.
(113) Fales, H. M., Milne, G. W. A., Winkler, H. U., Beckey, H. D.,
Damico, J. N., and Barron, R., Anal. Chem., 4_7, 207 (1975).
361
-------
Section 8Q
(114) 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.
(115) Biros, F. J., Dougherty, R. C., and Dalton, J., Org. Mass.
Spectrom.. j6, 1161 and 1171 (1972).
(116) Dougherty, R. C., Roberts, J. D., and Biros, F. J., Anal. Chem.,
47, 54 (1975).
(117) Holmstead, R. L., and Casida, J. E., J. Ass. Offie. Anal. Chem.,
57_, 1050 (1974).
(118) Dougherty, R. C., and Piotrowska, K. K., J. Ass. Offie. Anal.
Chem.. 59., 1023 (1976).
(119) Damico, J. N., Barren, R. P., and Sphon, J, A., Int.J. Mass
Spectrom. Ion Phys., _2, 161 (1969).
(120) Heck, H., Dyer, R. L., Scott, A. C., and Anbar, M., J. Agr.
Food Chem.. 25., 901 (1977).
(121) Schulten, H. R., and Beckey, H. D., J. Agr.Food Chem., 21, 372
(1973); 24, 743 (1976).
(122) Carroll, D. I., Dzidic, I., Stillwell, R. N. , Horning, M. G.,
and Horning, E. C., Anal.r Chem., 46, 706 (1974).
(123) Carroll, D. I., Dzidic, I., Stillwell, R. N., Haegele, K. D.,
and Horning, E. C., Ana_lv_Chem. , 47, 2369 (1975).
(124) Milne, G. W. A., and Lacey, M. J., CRC Grit. Rev. Anal. Chem.,
4., 45-104 (1974).
(125) Abramson, F. P., Anal. Chem., .44(14), 28A (1972).
(126) Oswald, E. 0., Albro, P. W., and McKinney, J. D., J. Chromatogr.,
28, 363 (1974).
(127) Vander Veld, G., and Ryan, J. F., J. Chromatogr. Sci., _13, 322 (1975)
(128) Flath, R. E., Gas Chromatography-Mass Spectroscopy, in Guide to
Modern Methods of Instrumental Analysis, Gouw, T. H., ed., Wiley-
Interscience, New York, 1972, Chapter IX, pages 323-350.
(129) Budde, W. L., and Eichelberger, J. W., J. Chromatogr., 134, 147
(1977).
362
-------
Section 8Q
(130) Fenselau, C. , Anal. Chem.. 49(6), 563A (1977).
(131) Sovocool, G. W., and Lewis, R. G., The Identification of Trace
Levels of Organic Pollutants in Human Tissues: Compounds Re-
lated to Chlordane/Heptachlor Exposure, in Trace Substances
in Environmental Health. IX, 1975, D. D. Hemphill, ed.,
University of Missouri Press, 1976, pages 265-280.
(132) Wright, L. H., Lewis, R. G., Crist, 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). ~
(133) Weil, L., Frimmel, F., and Quentin, K.-E., Z. Anal. Chem.,
268, 97 (1974).
(134) Rosen, J. D., and Pareles, S. R., Mass Spectrometry and NMR
Spectroscopy in Pesticide Chemistry, Plenum Press, New York,
1974, pages 91-98.
(135) Chapman, R. A., and Robinson, J. R., J. Chromatogr., 140,
209 (1977).
(136) Bergstedt, L., and Widmark, G., Chromatographia, J3> 59 (1970).
(137) Kuehl, D. W. , Anal. Chem., 4-9_, 521 (1977).
(138) Damico, J. N., Barren, R. P., and Ruth, J. M., Organic Mass
Spectrometry, 1, 331 (1968); Damico, J. N., J. Ass. Offic.
Anal. Chem., 4-9, 1027 (1966); Benson, W. R., and Damico, J. N.,
J. Ass. Offic. Anal. Chem.. 48, 344 (1965); 51, 347 (1968).
(139) Mestres, R., Chevallier, C., Espinoza, C., and Corner, R.,
Ann. Falsif. Expert. Chim.. .70(751), 177 (1977).
(140) Biros, F. J., and Sullivan, H., cited in (101).
(141) Biros, F. J., Applications of Combined GC-MS to Pesticide
Residue Identification, presented at the ACS-Chem. Inst. Canada
International Meeting, Toronto, May 1970.
(142) Van Peteghem, C. H., and Hendrickx, A. M., J. Agr. Food Chem.,
J24, 635 (1976).
(143) Shafik, T. M., Biros, F. J., and Enos, H. F., J. Agr. Food Chem.,
JL8, 1174 (1970).
363
-------
Section 8Q
(144) 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.
(145) Biros, F. J., and Walker, A. C., J. Agr. Food Chem., 18, 425
(1970).
(146) Bellman, S. W., and Barry, T. L., J. Ass. Offic. Anal. Chem.,
54, 499 (1971).
(147) Harless, R. L., and Oswald, E. 0., Gas Chromatography/Mass
Spectrometric 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.
(148) 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).
(149) Van den Heuvel, W. J. A., Wood, J. S., DiGiovanni, M., and
Walker, R. W., J. Agr. Food Chem., 25_, 386 (1977).
(150) Bagley, E. G., Reichel, W. L., and Cromartie, E., J. Ass. Offic.
Anal. Chem., 53, 251 (1970).
(151) Bonelli, E. J. , Anal. Chem., ^4_, 603 (1972).
(152) Sovocool, G. W., Lewis, R. G., Harless, R. L., Wilson, N. K.,
and Zehr, R. D., Anal. Chem., ^9_, 734 (1977).
(153) Canada, D. C., and Regnier, F. E., J. Chromatogr. Sci., 14,
149 (1976).
(154) Neher, M. B., and Hoyland, J. R., Specific Ion Mass Spectrometric
Detector for Gas Chromatographic Pesticide Analysis, U.S.
Environmental Protection Agency, Washington, D.C., Report No.
EPA-660/2-74-004, January 1974.
(155) Thruston, Jr., A. D., A Quantitative Method for Toxaphene by
GC-CI-MS Specific Ion Monitoring, U.S. Environmental Protection
Agency, Washington, B.C., Report No. EPA-600/4-76-010, March 1976.
(156) Harless, R. L. , Ellis, P. R., and Oswald, E. 0., Revised Gas
Chromatography-Mass Spectrometry Interface Techniques for Glass
Capillary Columns, Biomed. Mass Spjectrom., in press (1978).
364
-------
Section 8Q
(157) 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, B.C., Report No. EPA-600/1-76-037, November 1976.
(158) 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, pages 46-48.
(159) Alford, A., Biomed. Mass Spectrom., 4_(1), 1-22 (1977).
(160) Biros, F. J., in Pesticides Identification at the Residue
Level, Advances in Chemistry Series 104, American Chemical
Society, Washington, D.C., 1971, pages 132-150.
(161) Lewis, R. G., Research Report of the Analytical Chemistry
Branch, ETD, HERL, RTF Program Element No. 1EAG15, Sovocool,
G. W., and Wright, L. H., principal investigators.
(162) Johnson, L. D., Waltz, R. H., Ussary, J. P., and Kaiser, F. E.,
J. Ass. Offic. Anal. Chem., ,59, 174 (1976).
(163) Sun, Y. P., Analytical Methods £ orJ?es ticides and Plant Growth
Regulators'•_, Zweig, G. , ed. , Vol. 1, Academic Press, New York,
1963, page 571.
(164) 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.
(165) Gajan, R. T., Residue Rev. .5, 80 (1964); J5, 75 (1964).
(166) Gajan, R. T., J. Ass. Offic. Anal. Chem., 52, 811 (1969).
(167) Hance, R. J., Pestic. Sci., I, 112 (1970).
(168) Booth, M. D., and Fleet, B., Talanta, .17, 491 (1970).
(169) Kutlukova, V. S., Toropov, A. P., and Lozovatskaya, M. A.,
Anal. Abstr., J27, Abstract No. 2940 (1974).
(170) Farwell, S. 0., Beland, F. A., and Geer, R. D., Bull. Environ.
Contarn. Toxicol., 10, 157 (1973).
(171) Zietek, M., Mikrochim. Acta, II (5-6), 549 (1976).
365
-------
Section 8Q
(172) Marchidan, S., Rev. Roum. Chim.. 22_, 127 (1977).
(173) Filimonova, M. M., Zh. Anal. Khim.. _32, 140 (1977).
(174) Budnikov, G. K., Zh. Anal. Khim., J2, 212 (1977); 30i 2275 (1975)
(175) Southwick, L. M., Willis, G. H., Dasgupta, P. K., and Kesztheli,
C. P., Anal. Chim. Acta. JJ2, 29 (1976).
(176) Osteryoung, J. G., Anal. Chem., 49, 2310 (1977).
(177) Filimonova, M. M., Zh. Anal. Khim., J32, 812 (1977).
(178) 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.
(179) Beroza, M., and Coad, R. A., J. Gas Chromatogr., 4^ 199 (1966).
(180) Beroza, M., J. Org. Chem.. ^8, 3562 (1963).
(181) Beroza, M., J. Gas Chromatogr., 2_, 330 (1964).
(182) Asai, R. I., Gunther, F. A., and Westlake, W. E., Residue Rev.,
19, 57 (1967).
(183) Asai, R. I., Gunther, F. A., Westlake, W. E., and Iwata, Y.,
J. Agr. Food Chem.. ^0, 628 (1971).
(184) Zimmerli, B., J. Chromatogr.. 88, 65 (1974).
(185) Minyard, J. P., and Jackson, E. R., J. Agr. Food Chem., 13,
50 (1965).
366
-------
Section 9
MAINTENANCE, TROUBLESHOOTING, AND
CALIBRATION OF INSTRUMENTS
The EPA Pesticide Analytical Manual, Appendix I, contains comments
on the maintenance and repair of instruments primarily intended for
laboratories which are part of the EPA or have contractural arrange-
ments with EPA allowing them to make use of the electronic repair
facility located at Research Triangle Park, N. C. The Instrument
Shop in RTP is equipped to handle repairs, modifications, and
calibrations on various gas chromatographs, recorders, and GC de-
tectors.
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.
9 A 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
profile (BGS), and polarizing 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 final
hold control in a safe position to avoid accidental overheat operation?
(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?
367
-------
Section 9B
(h) 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?
9 B CHECK LIST WHEN INSTRUMENTAL REPAIR IS INDICATED
A systematic check-out routine is recommended to determine whether
instrumental 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
Tracor 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
indication that serious trouble is imminent. Keep in mind the type
of detector, column, carrier gas, and temperature range being used.
Recall, or preferably, 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.
Check List
(a) Is there proper insulation packing in the detector compartment?
Improper packing can lead to variation in signal due to ambient
temperature 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?
368
-------
Section 9C
(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, ie., 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? Refer to previous chromato-
grams obtained at like settings.
(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, electro-
meter , etc., plugged into the proper power source?
(n) Check that thermometers and thermocouples in the detector com-
partment 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 detector 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.
9 C GENERAL APPROACH TO TROUBLESHOOTING
The following pyramid approach represents a logical, general method for
instrumental troubleshooting
369
-------
Section 9C
g
\
f / Use alternate tests
-------
Section 9C
e.g.
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
Abnormal
Observe and measure for
normal/abnormal readings
Injection
point
Column
Detector
Abnormal
Bracket Placement
e.g., no flow: Rotameter indicates flow.
371
-------
Section 9D
(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 alternate tests.
(g) Speed, availability, and complexity determine replacement or
repair procedure.
9 D GAS CHROMATOGRAPH SERVICE BLOCK DIAGRAM
Gas Flow
control
a
Injection
Ports
b
Vs
\
X
Column Oven
c
)
x.
/
Support Electronics
g
Detectors
d
S
Detector
Electronics
e
'i
,
Recorder
f
(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
components, 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.
372
-------
Section 9E
9 E GAS INLET SYSTEM OF THE GAS CHROMATOGRAPH
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 is then 5 or 10 percent methane in argon.
The carrier gas must be dry and contain less than 5 ppm O . Con-
tamination 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 con-
tamination from entering the gas chromatographic flow system. Mole-
cular 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 the cleaned with hexaneacetone before reloading. It may also be
necessary to flame the dryer frit to expel all contamination.
373
-------
Section 9E
(d) Copper tubing, 1/8 inch instrument grade is used between the
filter dryer and the instrument. This tubing should be rinsed with
CH Cl and then acetone before installation. If the old tubing is
used, it should be also flamed. It is also advisable to clean all
Swagelok 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. Con-
tamination of the rotameters may cause the float to stick. Moisture 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-180. 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 pro-
gramming, 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 out-
let of the flow controller. This is a Brooks type 8501/8502 unit
that will protect the controller from flashback and trap solid con-
taminants 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
374
-------
Section 9F
unions with a Swagelok 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 con-
trolling 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 (Sub-
section 5.J. 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.
9 F 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 activated 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/min.
375
-------
Section 9F
(5) Adjust the detector temp-set controller to maintain a
detector temperature of approximately 200°C. If a faulty temp-set
is encountered, substitute a Superior type B variac.
(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 provided
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
which 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 septums.
(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 common 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.
(d) Corrective measures for locating leaks in the carrier flow
system.
376
-------
Section 9F
(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) "Snoop" all lines 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.
(2) Follow the procedures outlined in (a) (6) and (a) (7) for
each port.
(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 trans-
fer 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 new septums 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 de-
pression 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.
377
-------
Section 9F
(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 RTF 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
378
-------
Section 9F
lines can be checked for leaks. These valves may also be used if repairs
and operation of the instrument are to be carried on simultaneously.
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 com-
pletely 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 con-
tamination and to some extent from flashback.
When installing columns in the instrument, they can be "set-up" as shown
in the diagram following. The upper "0" ring should seat firmly on the
column and should not move freely. In no case should it be necessary to
use two "0" rings to obtain a seal. The rear furrule is installed in a
reversed procedure than would be used in a metal to metal fitting
installation. 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 re-
commended between the column packing and the fittings after installation
is completed.
379
-------
Section 9G
-"0" Ring
—Rear furrule (Reversed)
-1/4" Swagelok Nut
-"0" Ring
l~
Glass Wool Plug
— Column Packing
9 G 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 con-
ditions 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.
380
-------
Section 9G
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 whenever temperature indications appear faulty. (Usually one thermo-
couple is placed in the open air to rough-check the pyrometer against
ambient temperature). Always be sure that battery contacts are clean.
Oven temperature and control is 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 eg_uilibration.
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
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
381
-------
Section 9H
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.
9 H DETECTOR AND ELECTROMETER
The detector and electrometer are integral parts of the gas chromatograph.
They are connected together 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
( H) or ceramic (63Ni) 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 H (tritium) detector is used, an adjust-
able limit switch in series with the heat cartridge prevents the de-
tector 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
Ni 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
amps. This is obtained at the 0.1 setting. A minimum attenuation of
5 x 10 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 chart 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 for electron capture operation
or 10 for flame. A 10 turn potentiometer is located on the front
panel for fine bucking control.
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 electrometer
with its cover correctly in place (otherwise noise will be excessive
due to stray field pick up).
Electrometer Problems:
(a) Can't zero and/or can't buck out - possible cure:
(1) Check zero and bucking pot on front panel for continuity.
382
-------
Section 9H
(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 50
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.
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 chromatographer 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.
383
-------
Section 91
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 appeal
to function normally. If this occurs, refer to the operational manual
schematics and:
(1) Change R£ and R3 to
(2) Change VR2 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 %W 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 Tracor service manual procedure "2" states
"Remove clip lead." This is incorrect. The clip lead should be retained.
In procedure "3", do not clip a shunt lead from E^ to ground and do not
adjust to zero but to 30 MV.
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.
9 I OBSERVATION OF PROBLEMS ON CHROMATOGRAMS
(a) Peaks return below baseline: dirty or partially depleted detector
foil - clean or replace foil (See Figure 5-5, A in Section 5) .
(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.
(e) Stepping baseline (may be observed on peaks) : check for dirty
recorder slide wire, line voltage changes, recorder drive, recorder
gain control.
384
-------
Section 9J
(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 on substitute, 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.
9 J DETECTOR 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.
385
-------
Section 9K
Troubleshooting:
(1) Leave column cold. Eliminates (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).
9 K TROUBLESHOOTING THE MICROCOULOMETRIC SYSTEM
(a) No flow from pyrolysis tube.
Symptom: no flow indicated on bubble meter.
Probable causes:
(1) Empty gas tank.
(2) Pressure not on.
(3) Broken pyrolysis tube.
(4) Broken column.
(5) Broken flow meter.
(6) Broken flow control.
Troubleshooting procedure:
(1) Check tank pressure.
(2) Set proper output pressure from tank.
(3) Remove and check pyrolysis tube.
(4) Remove and check column.
386
-------
Section 9K
(5) Remove and check flow meter.
(6) Check output from flow control.
(b) Furnace will not heat.
Symptom: no temperature increase on pyrometer.
Probable cause:
(1) Heat control not on.
(2) Thermocouple open.
(3) Furnace heater open.
(4) Heat control defective.
Troubleshooting procedure:
(1) Turn heater control on.
(2) Check to see if furnace is red on inside, check thermocouple
with ohm meter.
(3) Check furnace heater resistance with ohm meter.
—{4) Check 0-110 VAC output from heat control.
(c) Inlet block will not heat.
Symptom: no temperature increase on pyrometer.
Probable cause:
(1) Heat control not on.
(2) Thermocouple open.
(3) Furnace heater open.
(4) Heat control defective.
Troubleshooting procedure:
(1) Turn heater control on.
(2) Feel block, check thermocouple with ohm meter.
— Caution: use only volt-ohmeter. Do not use vacuum volt meter.
387
-------
Section 9K
(3) Check resistance of block heater with ohm meter.
— (4) Check 0-110 VAC output from heat control.
(d) No output from A-100 test box.
Symptom: will not read 1 VDC on volt meter.
Probable cause:
(1) Duo-Dial not set at 1000.
(2) Batteries defective.
(3) A-100 defective.
Troubleshooting procedure:
(1) Set Duo-Dial at 1000 (10 turns).
(2) Check batteries 1.35 VDC. Replace if necessary.
(3) Substitute A-100.
(e) Recorder will not zero.
Symptom: microcoulometer output range at 000, recorder will not zero.
Probable cause:
(1) Cable not connected from coulometer to recorder.
(2) Coulometer batteries defective.
(3) Coulometer defective.
(4) Recorder defective.
Troubleshooting procedure:
(1) Connect cable from coulometer to recorder.
(2) Check batteries 1.35 VDC. Replace if necessary.
(3) Substitute coulometer.
(4) Substitute reeorde1
—Caution: use only volt-ohmeter. Do not use vacuum volt meter.
388
-------
Section 9K
(f) No bias output
Symptom: meter does not move when bias control is adjusted.
Probable cause:
(1) Function switch not in proper position.
(2) Batteries defective.
(3) Coulometer defective.
Troubleshooting procedure:
(1) Set function switch to Bias Read.
(2) Check batteries 1.35 VDC. Replace if necessary.
(3) Substitute coulometer.
(g) Coulometer will not balance in Generator Read.
Symptom: meter is either extreme left or right.
Probable cause:
(1) A-100 not connected.
(2) Batteries defective.
(3) Coulometer defective.
(4) A-100 defective.
Troubleshooting procedure:
(1) Connect A-100 to coulometer.
(2) Check batteries 1.35 VDC. Replace if necessary.
(3) Substitute coulometer.
(4) Substitute A-100.
(h) Baseline noisy when connected to A-100 test box, Operate mode.
Symptom: recorder baseline 1-1 1/2 percent noise at range 200/Lo
Gain, range 50/Hi Gain.
Probable cause:
(1) Ground not connected from A-100 to coulometer.
389
-------
Section 9K
(2) A-100 defective.
(3) Coulometer defective.
(4) Recorder defective.
Troubleshooting procedure:
(1) Connect ground from A-100 to coulometer.
(2) Substitute A-100.
(3) Substitute coulometer.
(4) Substitute recorder.
(i) Unable to balance coulometer, cell connected.
Symptomi meter extreme left or right.
Probable cause:
(1) Function switch in wrong position.
(2) Batteries defective.
(3) Cell needs flushing.
(4) Cell defective.
(5) Coulometer defective.
Troubleshooting procedure:
(1) Place function switch in Operate position.
(2) Check batteries 1.35 VDC. Replace if necessary.
(3) Flush cell.
(4) Substitute cell.
(5) Substitute coulometer.
(j) Recorder baseline noisy, cell connected.
Symptom: baseline 1 1/2 - 2 percent noise, range 200/Lo Gain,
range 50/Hi Gain.
390
-------
Section 9K
Probable cause:
(1) Cell contaminated.
(2) Gas contaminated.
(3) Coulometer defective.
(4) Recorder defective.
(5) Flow system has leak.
Troubleshooting procedure:
(1) Clean cell.
(2) Open vent, substitute gas.
(3) Substitute coulometer.
(4) Substitute recorder.
(5) Test system for leaks.
(k) Poor sensitivity.
Symptom: standard injected produces little or no response.
Probable cause:
(1) Wrong standard.
(2) Wrong attenuation on coulometer.
(3) Cell needs replating.
(4) Pyrolysis tube contaminated.
(5) Leak in flow system.
Troub1eshoot ing procedure:
(1) Check standard.
(2) Check range and gain setting.
(3) Replate cell.
(4) Replace pyrolysis tube.
(5) Test flow system for leaks.
391
-------
Section 9L
9 L 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.
(6) Test system for leaks.
(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.
392
-------
Section 9L
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 dis-
tilled water.
(6) Recondition column.
(d) Loss of gas flow.
Symptom: bubbles not present in cell.
Probable cause:
(1) Gas tank empty.
(2) Broken pyrolysis tube.
393
-------
Section 9L
(3) Broken column.
(4) Valve blocked.
(5) Broken flow control.
Troubleshooting procedure:
(1) Check tank pressure.
(2) Remove and inspect pyrolysis tube.
(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.
(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.
•^(4) Check variable 1-110 VAC output of heater control with volt
meter.
i'Caution: use only a volt-ohmeter. Do not use a vacuum tube volt meter.
394
-------
Section 9M
(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.
9 M 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 footenote on page 28.
t
In any case of noisy base
adjusted and the slidewlre is clean.
o /
*-* In any case of noisy baseline, make certain the recorder gain is properly
395
-------
Section 9M
(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.
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 grey 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.
396
-------
Section 9N
9 N EPILOG
The reader is cautioned against leaping to a conclusion 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 con-
tamination of the GC column (see Figure 4-1 in Section 4). Because of
the visible similarities on the chromatograms of electronic vs. other
problem sources, the operator should not proceed post haste to tear
the instrument apart 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.
397
-------
Section 10
TRAINING OF PESTICIDE ANALYTICAL CHEMISTS
This chapter is by far the shortest in the manual, and the reader may
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 formal, specialized
training demonstrated far superior analytical performances 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 techniques 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 10-1 (copied from Table 2-7 in Section 2) which lists the relative
performance ranking of 34 laboratories in one interlaboratory check
sample exercise.
398
-------
Section 10
The top eight laboratories, performancewise, 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 provides most con-
clusive 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, but it is hoped that some edu-
cational institutions or governmental agency will recognize the need
and set up programs to provide such training, and that laboratory super-
visors 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 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.
399
-------
TABLE 10-1
Section 10
RELATIVE PERFORMANCE RANKINGS
CHECK SAMPLE NO. 26, MIXTURE IN SOLVENT
Lab . Code
Number
•l61.
/ 137.
• 135.
• 162.
* 87.
•'USA.
/ 113.
V 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
0
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
400
-------
Section 11
REVISIONS TO THE QUALITY CONTROL MANUAL
This manual will be revised biennially and 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)
401
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/1-79-008
3, RECIPIENT'S ACCESSION>NO.
4. TITLE AND SUBTITLE
MANUAL OF ANALYTICAL QUALITY CONTROL FOR PESTICIDES
AND RELATED COMPOUNDS In Human and Environmental
Samples - First Revision
5, REPORT DATE
January 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Joseph Sherma
Revisions by Joseph Sherma and Morton Beroza
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemistry,
Easton, Pennsylvania
Lafayette College,
10. PROGRAM ELEMENT NO.
1ED868
11. CONTRACT/GRANT NO.
Association of Official Analytical Chemists
Washington. DC
68-02-2474
12. SPONSORING AGENCY NAME AND ADDRESS
Health Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Manual
14. SPONSORING AGENCY CODE
EPA 600/11
16. SUPPLEMENTARY NOTES
16. 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 standardization of materials used, and the
operation of the gas chromatograph ate 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.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Quality control
Chemical analysis
Chemical tests
Pesticides
Bioassay
Environmental samples
07, C
07, B
13. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
412
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
402
*U.S. GOVERNMENT PRINTING OFFICE: 1979 - 640- 0 1 3 U202REGIONNO. 4
------- |