EPA-600/1-76-017
February 1976
MANUAL OF ANALYTICAL QUALITY CONTROL FOR PESTICIDES
AND RELATE) COMPOUNDS
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
Contract No. 68-02-1727
Project Officer - Editor
Jack F. Thompson
Environmental Toxicology Division
Health Effects Research Laboratory
Research Triangle Park, N.C. 27711
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.
11
-------
TABLE OF CONTENTS
Section In-Section
Number Section Page
PREFACE AND ACKNOWLEDGEMENTS
GENERAL DESCRIPTION OF PESTICIDE RESIDUE
ANALYTICAL METHODS
A. Extraction Procedures 1
B. Cleanup Procedures 2
C. Final Determination Methods 2
D. Confirmatory Techniques 4
E. References 6
INTERLABORATORY QUALITY CONTROL
A. Quality Control Program of the EPA Environ- 1
mental Toxicology Division (ETD) Laboratory
B. Program Objectives 2
C. Program Activities 2
D. Types and Preparation of Sample Media 3
E. Reporting Forms 4
F. Evaluation of Reported Data 6
G. Summary of Results Tables 6
H. Relative Performance Ranking 7
I. Private Critiques 13
J. Progression of Performance 13
K. Statistical Terms and Calculations 18
L. References 24
INTRALABORATORY QUALITY CONTROL
A. Purpose and Objectives 1
B. Purpose and Objectives of SPRM'S 1
C. Nature of SPRM'S 2
D. Frequency of SPRM Analysis 3
E. Record Keeping 4
F. Quality Control Charts 7
G. Benefits of the In-House SPRM Program 9
H. Analytical Balances 10
I. Purity of Reagents 11
J. Distillation of Solvents 13
K. Miscellaneous Reagents and Materials 13
L. Cleaning of Glassware 14
M. Pesticide Reference Standards 15
N. Calibration and Maintenance of the Gas Chromato-
graph and Accessories 24
O. Adherence to Official or Standardized Methodology 26
P. References 27
111
-------
Section In-Sectio|n
Number Section Page
4 EVALUATION, STANDARDIZATION, AND USE OF MATERIALS
FOR PESTICIDE RESIDUE ANALYSIS
A. Adsorbents 1
B. GC Column Technology-Introduction 7
C. Column Efficiency and Resolution 8
D. Sensitivity and Retention 10
E. Column Stability 11
F. Resistance to On-Column Decomposition 11
G. Homemade vs. Precoated Packings 15
H. Packing the Column 17
I. Column Conditioning 19
J. Evaluation of the Column 21
K. Maintenance and Use of GC Columns 22
L. References 25
5 OPERATION OF THE GAS CHROMATOGRAPH
A. Temperature Selection and Control 1
B. Selection and Control of Carrier Gas Flow Rate 5
C. Electron Capture Detector Operation 7
D. Microcoulometric Detector Operation 15
E. Alkali Flame lonization Detector 17
F. Flame Photometric Detector 19
G. Electrolytic Conductivity Detector 24
H. Other Detectors 29
I. Electrometer and Recorder 30
J. Sample Injection and Injection Port 30
K. Erratic Baselines 33
L. GC Columns Recommended for Pesticide Analysis 34
M. Sensitivity of the GC System 39
N. Qualitative Analysis 40
O. Quantitative Analysis 41
P. References 52
6 ADDITIONAL PROCEDURES IN PESTICIDE ANALYSIS
A. General Considerations 1
B. Representative vs. Biased Sampling 1
C. Sample Containers 2
D. Sample Compositing 2
E. Storage of Samples 3
F. Sampling of Agricultural and Food Products 4
G. Sampling of Biological Materials 4
H. Air Sampling 5
I. Water Sampling 6
IV
-------
Section In-Section
Number Section Page
J. Sampling of House Dust, Soil, and Stream
Bottom Sediment 9
K. Marine Biological Samples 10
L. Control of Procedures for Extraction of Residues 11
M. Control of Evaporation and Concentration of
Sample Solutions and Fractional Eluates 14
N. Introduction to High Performance Liquid Column
Column Chromatography (HPLC) 19
0. HPLC Instruments 20
P. Theory and Principles of LC 20
Q. Introduction to Thin Layer Chromatography (TLC) 22
R. Practical Considerations in TLC 23
S. Quantitative TLC 25
T. Thin Layer Systems 27
U. References 30
7 MULTIRESIDUE EXTRACTION AND ISOLATION PROCEDURES
FOR PESTICIDES AND METABOLITES
A. Tissue and Fat Analysis (Macro Technique) 1
B. Human or Animal Tissue and Human Milk (Micro
Technique) 3
C. Human Blood, Multiresidue 3
D. Pentachlorophenol (PCP) in Blood and Urine 4
E. p_,p_'-DDA in Urine 5
F. 2,4-D and 2,4,5-T in Urine 5
G. Analysis of Fatty and Nonfatty Foods by the
Mills, Onley, Gaither Method 6
H. Chlorophenoxy Herbicides in Fatty and Nonfatty
Foods 6
I. Carbon-Cellulose Column Cleanup 7
J. Cleanup on Deactivated Florisil 7
K. Low Temperature Precipitation 9
L. Miscellaneous Multiresidue Cleanup Procedures 9
M. Alkyl Phosphate Metabolites in Urine 12
N. Para-Nitrophenol (PNP) in Urine 13
O. Special Considerations in the Analysis of Fatty
and Nonfatty Foods Using Florisil Cleanup 13
P. Sweep Co-Distillation 14
Q. Charcoal Cleanup of Nonfatty Food Extracts 14
R. Miscellaneous Multiresidue Cleanup Procedures 16
Carbamate Pesticides and Miscellaneous Herbicides
S. 1-Naphthol in Urine 17
T. N-Methylcarbamate Insecticides in Blood and Fat 18
U. Other Indirect (Derivatization) Methods of
Analysis 18
-------
Section In-Section
Number Section Page
7 V. Direct Methods of Analysis 18
W. Analysis of Plant and Food Materials 19
X. Air Analysis 20
Y. Water Analysis 21
Z. Soil, House Dust, and Bottom Sediment 22
Polychlorinated Biphenyls (PCB's)
A.A. Pesticide-PCB Mixtures 24
A.B. Appearance of PCB Chromatograms 24
A.C. Methods of Separation and Analysis of Pesticides
and PCB's 28
A.D. Reviews of Analytical Methods For Pesticides 31
A.E. References 31
8 CONFIRMATORY PROCEDURES
A. Requirements for Positive Confirmation of
Pesticide Identity 1
B. GC Relative Retention Times 3
C. Selective GC Detectors 4
D. TLC RF Values 5
E. HPLC 6
F. Extraction p-Values 6
G. Derivatization Techniques 7
H. Spectrophotometry 12
I. Visible, UV, Fluorescence and Phosphorescence 12
J. Infrared (IR) 13
K. Nuclear Magnetic Resonance (NMR) 17
L. Mass Spectrometry (MS) and (GC-MS) 17
M. Biological Methods 25
N. Polarography 25
0. Miscellaneous Confirmatory Methods 26
P. References 28
9 MAINTENANCE, TROUBLESHOOTING AND CALIBRATION OF
INSTRUMENTS
A. Daily Operational Considerations for GC
Instrumentation 1
B. Check List When Instrumental Repair is Indicated 2
C. General Approach to Troubleshooting 3
D. GC Service Block Diagram 6
E. Gas Inlet System of the GC 7
F. Isolation of Problems in Flow Systems in GC
Equipped with EC Detector 9
VI
-------
Section In-Section
Number Section Page
9 G. Temperature Control and Indication in the GC 14
H. Detector and Electrometer 16
I. Observation of Problems on Chromatograms 17
J. Detector Background Signal Response (BGS) 18
K. Troubleshooting the Microcoulometric Detector
System 19
L. Troubleshooting the Coulson Electrolytic
Conductivity System 25
M. Troubleshooting the Flame Photometric Detector 28
N. Epilog 30
10 TRAINING OF PESTICIDE ANALYTICAL CHEMISTS 1-3
11 REVISIONS TO THE QUALITY CONTROL MANUAL 1
Vll
-------
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.
Vlll
-------
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. It
seems inappropriate to mention these laboratories by name as the quality
of the data from a few were somewhat less than ideal.
A note of thanks is due Mr. Frank Wilinski, Electronics Shop supervisor
at EPA, RTF, N.C. for the contribution of the raw data required for
the preparation of Section 9 covering instrumental troubleshooting.
A special word of gratitude is extended to Mrs. Nadine Vogel who pre-
pared the final typed draft of the manual and patiently endured the
many revisions.
IX
-------
Section 1
DESCRIPTION OF PESTICIDE
RESIDUE ANALYTICAL ETUODS
A pesticide residue analysis usually consists of four steps:
(1) Extraction of the residue from the sample matrix.
(2) Removal of interfering co-extractives ("cleanup").
(3) 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.
(4) 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.
l.A. EXTRACTION PROCEDURES
A solvent or mixture of solvents should be used which is at least 80
percent efficient, selective enough to require a minimum of cleanup,
and does not interfere with the final determination. 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 ex-
tractors, blenders, 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. Extraction in the presence
of sodium sulfate helps liberate more water-soluble compounds.
-------
i.B. 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. Extracts con-
taining fatty material are especially troublesome. Depending on the
extent and-nature of the co-extractives and the pesticide residue,
solvent partition, liquid chromatography (column adsorption or gel, or
TLC), and sweep co-distillation are most often used, alone or in com-
bination, for cleanup.
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. 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 acetoni-
trile, dimethylformamide, and dimethylsulfoxide. This 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.
Thin layer chromatography, sweep co-distillation, ion-exchange and gel
chromatography cleanup methods are applicable in some cases but have been
used less frequently than solvent partition and adsorption chromatography.
An automated instrument based on gel permeation chromatography has been
shown to efficiently separate lipids from chlorinated pesticides and PCB's.
l.C. FINAL DETERMINATION METHODS
Chromatographic methods are by far the most widely used for determination
of pesticide residues, followed by spectrophotometric and biological methods.
The latter include bioassay and enzymatic techniques which are simple,
since they require no cleanup, but non-specific. Enzyme inhibition, when
used as a detection procedure after thin layer chromatography, is a sensi-
tive (low ng detection limits) and selective method for certain organo-
phosphorus and carbamate pesticides.
- 2 -
-------
Spectrophotometric determination is generally less sensitive and selective
than gas or thin layer chromatography. Spectroscopic methods have their
main use as ancillary techniques to gas chromatography for confirmation
of residue identity. An advantage is that if selectivity and sensitivity
are adequate, colorimetric methods can be adapted to automated processes.
Fluorescent pesticides and metabolites may be determined by fluorimetry,
which is more sensitive than visible, UV, or IR methods. Since relatively
few pesticides are naturally fluorescent, fluorimetry is selective, but
removal of fluorescent impurities is often 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 deter-
minative procedure and thin layer chromatography (TLC) for screening, semi-
quantitation, and confirmation. TLC offers generally increased resolution,
shorter development times, and increased sensitivity compared to paper
chromatography. 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 spray-
ing with ethanolic AgNC>3 or incorporation of AgNO^ into the layer followed
by irradiation with ultraviolet light. Many organophosphorus and carba-
mate pesticides are detectable at low ng levels by enzyme inhibition
techniques or at higher levels by numerous chromogenic reagents. Fungi-
cides are detectable by bioautography. Many polar herbicides and heat-
labile, poorly-detectable carbamates, which require formation of derivatives
prior to gas chromatography, are particularly amenable to analysis by TLC.
Quantitative analysis by in situ scanning has been carried out in visible
absorption, UV absorption (quenching), and fluorimetric modes using
commercial densitometers.
Gas chromatography of pesticides is normally carried out on 3 to 7 foot
glass columns packed with single and mixed organosilicone and polyester
stationary phases ranging from low to high polarity. Among the most
used phases are SE-30, QF-1, DC-200, OV-210, OV-17, DECS, Carbowax 20M,
GE-XE60, Versamid 900, and DC-200/QF-1, OV-17/QF-1, or SE-30/OV-210
mixtures. Samples should be examined on two or three columns of markedly
different polarity before results are considered conclusive.
Organochlorine pesticides are usually analyzed with a tritium or Ni
electron capture detector. This detector is less specific than the
other common pesticide detectors but can detect low picogram amounts
of many halogenated compounds. 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 microcoulo-
metric and flame photometric (394nm) 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. The coulson detector can also be operated
- 3 -
-------
selectively for organochlorine compounds, and the Hall modification of
this detector is sensitive to 100-400 pg heptachlor. 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.
Currently of less importance than the above are the other proposed
pesticide detectors, including the phosphorus mode of the microcoulo-
metric detector, the sulfur and pyrolytic modes of the conductivity de-
tector, the halogen mode (negative response) of the thermionic detector,
the microwave plasma detector, the sulfur-phosphorus emission detector,
the Beilstein flame detector, and the flameless ionization (chemi-
ionization) detector. Selective detectors have the advantages of simpli-
fying cleanup procedures and aiding residue identification.
Modern, high performance liquid chromatography (LC) is being used increas-
ingly for the final determination of polar, involatile, or heat-labile
pesticide residues without derivative formation. Small particle porous
adsorbents and supports for the stationary liquid phase are packed into
narrow bore (down to 1 mm) columns and the mobile phase is forced through
at pressures up to 5000 psig. Pellicular (porous layer) and chemically
bonded packings are also used. Ultraviolet absorption, refractive index,
polarography, and fluorescence detectors have been used, the latter in
conjunction with fluorogenic labelling of pesticides which are not
naturally fluorescent by methods similar to those used earlier for thin
layer fluorodensitometry. The major disadvantage of LC at present is
the poorer sensitivity (ca. 10~6 to 10~9 grams) of commercially available
detectors, but 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 usually minimum sample cleanup is required, and 'that a.
greater number of separations of more complexity can be accomplished
because of the active role of the mobile phase in achieving resolution
and the wide range of stationary phases available for use in combination
with the great variety of possible mobile phases.
Other final determinative methods which have been applied to pesticide
residues include polarography for compounds containing an oxidizable
or reducible group, either naturally or after derivatization,, 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.
l.D. CONFIRMATORY TECHNIQUES
Three truly independent results are considered necessary for irrefutable
confirmation of the identity of a residue. Alternative methods which can
be combined are TLC and/or paper chromatography with sorbent-solvent
- 4 -
-------
systems of different polarity or different visualizing reagents, gas
chromatography on columns of different polarity, preparation of deriva-
tives to alter structure and thereby chromatographic properties, extrac-
tion JD-values, selective GC detectors, and mass spectroscopy. Unlike
NMR, IR, UV, etc., the latter method has sufficient sensitivity (as low
as 10 ng) for general application to residue identification as well as
the confirmation of pesticides in the presence of PCB's. 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.
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 where the final determination is
by colorimetry or UV absorption. Many of the colorimetric methods in-
volve 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 chromatography column, and a system for automatic cleanup of
samples by gel permeation chromatography have been designed. However,
a proven, completely automated procedure for residue analysis as it is
usually performed (ie., extraction, partition and adsorption chromato-
graphic cleanup, and gas chromatography) is not yet available.
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, of which there
are many hundreds of compounds with greatly differing chemical structures
and properties (e.g., organohalides, organophosphates, carbamates, ani-
lines, ureas, phenols, triazines, quinones, etc.), and the analysis may
involve traces of any of these materials alone or in combination in a
great variety of sample types, each with its own peculiar problems.
Further complications arise because metabolic degradation of certain
pesticides produces compounds which may be more; toxic and of different
polarity than the parent pesticide. 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. 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 analyzing for these related compounds,
if necessary. Trace contaminants contained in solvents or reagents, or
extracted from plastic apparatus, can give rise to GC peaks or TLC spots
which may be confused with pesticides. Positive confirmation of some
pesticides is especially difficult because of very similar chromato-
graphic properties, e.g., dieldrin and "photo-dieldrin".
- 5 -
-------
Since this introductory section is intended as a broad overview of
modern residue analytical methods, no details have been given. Much of
this same material will be discussed more completely in later sections,
and specific references to relevant sections of the EPA Pesticide Analy-
tical Manual or other sources will then appear. A general bibliography
of recent books and reviews on pesticide analysis follows.
I.E. REFERENCES
[1] Ruzicka, J. H. A., and Abbott, D. C. , Pesticide Residue Analysis,
Talenta, 2_0, 1261 (1973).
[2] Sherma, J., and Zweig, G., Pesticides, Chapter 25 of Chromatography,
Heftmann, E., editor, Van Nostrand Reinhold Co., 3rd edition, in
press.
[3] Sherma, J., Gas Chromatographic Analysis of Some Environmental
Pollutants, in Advances in Chromatography, Volume 12, Marcel Dekker
Inc., in press.
[4] Sherma, J., Chromatographic Analysis of Fungicides, Chromatographic
Reviews, 19_, 97-137 (1975).
[5] 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.
[6] 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.
[7] Sherma, J., and Zweig, G., Federal Regulations, Analysis of Insect
Pheremones, and Analytical Methods for New Individual Pesticides,
Volume VIII of Analytical Methods for Pesticides and Plant Growth
Regulators, Zweig, G., editor, Academic Press, in press.
[8] Sherma, J., and Zweig, G., 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, in preparation.
[9] Gaul, J. A., and McMahon, B. M., editors, Pesticide Analytical
Manual, Volumes I (multiresidues) and II (individual residues),
USDHEW, FDA, 5600 Fisher's Lane, Rockville, MD (yearly revisions).
Referred to throughout this manual as the FDA, PAM.
[10] Thompson, J. F., editor, Analysis of Pesticide Residues in Human
and Environmental Samples, USEPA, Chemistry Branch, Research Triangle
Park, N.C. (yearly revisions). Referred to throughout this manual
as the EPA, PAM.
- 6 -
-------
[11] 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).
[12] Cochrane, W. P., and Purkayastha, R., Analysis of Herbicides by Gas
Chromatography, Toxicological and Environmental Chemistry Reviews,
1, 137-268 (1973K
[13] SherL.a, J., Chromatographic Analysis of Pesticide Residues, CRC
Critical Reviews of Analytical Chemistry, August, 1973, pp. 299-354.
[14] Williams, I. H., Carbamate Insecticide Residues in Plant Material:
Determination by Gas Chromatography, Residue Reviews, 38, 1 (1971).
[15] Fishbein, L. , Chromatography of Triazines, Chromatographic Reviews,
12, 177 (1970).
[16] Fishbein, L., Chromatography of Organomercurial Fungicides,
Chromatographic Reviews, 13, 83 (1970).
[17] Malone, B., Analytical Methods for Determination of Fumigants,
Residue Reviews, 38, 21 (1971).
[18] Fishbein, L., Chromatographic and Biological Aspects of PCB's,
Journal of Chromatography, 68, 345 (1972).
- 7 -
-------
Section 2
^[TERLABORATORY QUALITY CONTROL
2.A. 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.
-------
indicates, on a mathematical basis, the degree of confidence which can be
placed in the results of sample analyses, and identifies analytical areas
needing further attention. The coordinating laboratory receives data firom
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 Intralaboratory control program, which will be treated in detail in
Section 3, assists a single laboratory in improving the accuracy and pre-
cision of data produced by its personnel by providing systematic guide-
lines for top quality analytical methodology and techniques. One feature
of this program is the continual, periodic analysis of standard reference
materials (SRM'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 and 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.
2.B. 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.
2.C. PROGRAM ACTIVITIES
The interlaboratory program includes the following activities,
a. The analyses of interlaboratory check samples by all participants.
b. A repository to provide analytical grade pesticide reference stan-
dards, over 400 of which are now available and listed in an index
available from the ETD Laboratory.
-2-
-------
c. Providing uniform, standard analytical methods in the form of an
analytical manual also available from the ETD Laboratory.
d. Quality controlling 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. The operation of an electronic facility for repair, overhaul, design
and calibration of laboratory instruments.
2.D. TYPES AND PREPARATION OF SAMPLE MEDIA
The check sample program is probably the most important aspect of the
interlaboratory activities since all other aspects are closely dependent
on it. Samples used in the program are mixtures of pesticides in a sub-
strate 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 samplers).
The preparation and handling of a blood interlaboratory check sample by
the coordinating laboratory will be detailed as an example. General popu-
lation serum samples are obtained from a local blood bank, typically in
300 ml bottles. The frozen samples are thawed in a refrigerator (2-3°C),
poured together into a stainless steel container which has been rinsed
with acetone, and mixed well. A quantity of approximately four liters
has been found sufficient for the needs of the program for about one
year. Experienced chemists analyze the pooled serum to establish the
base level profile and to be sure there is no gross contamination 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 in quantities large
enough to serve as their interlaboratory check sample and to also pro-
vide them with intralaboratory standard reference material (SRM) for a
six month period. Each laboratory supervisor requests in advance the
amount of sample required for the latter purpose based on his estimated
routine sample load (see Subsection 3.D.). A careful study has indi-
-3-
-------
cated no need to mail the samples frozen since neither pesticide nor
sample degradation has been observed in a 3 to 4 day period. After
removing the amount required for the interlaboratory check sample, per-
sonnel at each laboratory sub-divide the remainder into small vials
which are stored continuously in a freezer and individual vials removed
as needed to provide 2.0 ml intralaboratory SRM 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 test the participating laboratories on their recovery of
high pesticide levels, as might be encountered when analyzing 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 SRM'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 in the future to supply urine samples for
testing certain procedures.
With the check sample, each participant receives a covering letter pro-
viding the protocol for handling the sample. The time period allowed
for analyzing and reporting results corresponds to the normal time for
processing a similar routine sample.
Since the interlaboratory check sample is recognized as such by the
chemist at the time of analysis, it is likely that special care and
attention will be given it, and also that the best chemist in the labora-
tory 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 most often represent the very best work the labora-
tory can possibly turn out.
The importance of the interlaboratory check sample program is indicated
by the fact that on the basis of information obtained over the years a
number of actions have been initiated toward standardization. These in-
clude distribution of pretested Florisil cleanup adsorbent and GC column
packings and frequently updated standard analytical methods.
2.E. REPORTING FORMS
Laboratories are requested to report their results on special reporting
forms supplied for this purpose. The forms are designed to provide a
considerable amount of supplemental operating data in addition to numeri-
cal results of the analysis. The standard reporting form with detailed
instructions for completion on the reverse side is shown as Table 2-1.
It is seen that the data and information supplied by each laboratory
includes the sample size, extent of sample extract concentration, in-
jection volumes, elution cuts if column cleanup is required, all instru-
mental operating parameters, identity of the GC column, and the numerical
-------
0
O
0
4
5
E-
3 u.
Z M
I I r-i
CO
U b,
o -
a
5
U]
CU
S u
H M
OS
- o
E-«
U
»J
DU
U
^
|
U
i
5
o
o
5
W
^J
M
S
S
o
CU
4
1
I
5
c
0
-H
4J
3
e
o
3
0,
3
o
14
U
w
.H
M
U £
S E
Q <
C
-^
£
x
Q
&
H u
Z H
Q c<
J 3
0
8 .
c
1-1
V S
.
i-l
£
cu
£ S
EH -3
tu
*
tJ W
8 S
3
Q.
C
E
X.
r-t
£
g
,j
tM
2 tA
gut
M
j o;
o S
u <
U
i 0°'
3
o.
g
(/)
CU
§
H
u
S z
£ M
,j
u
o o:
o u
U b.
(0
z
o <
32
H
_
(B
r4
C
^j
U
M
3*
o:
S
§
EH
[O
M
X
B
OS
H
X
ii W
°^-7
J Z ^
D M "
> (M
-I
O
Z
3
0
a
Z
O
c.
)
u
Q
t-
p
1
t
O
M
{-
y-
ta
CU
i
i
(
i
i
i
i
i
)
i
,
i
S U
O o
a
S
0
CQ
C
0)
o ci
S-4 S
-P (U
H EH
z,
^_^
d)
O c!
3 0
O "O
CO g
0
d) pQ
P M
0 -H
P£
o
VJ Q)
rd r-\
OJ 0
^1 t | Ju
-M QJ
ox;
0) ^- -P
tn o
Oi O
-p
0) tfl O
E M rtf
3 4-1 >
O (D !H
> U O
C -P
O O
H >i o a;
C 03 ^-1
^ o
3 ^H id tn
n 4J -H
X M
4) U) (U -H
PC O
-H 0 LH CQ
U -r-l O
4J ni
« a; rn
t3 O h Oi
a) -H d E
0) UH -P -rH
3 -H 10 W
'O o a)
O E Oi
^ Ex;
P tU O) U)
d) 4J -P -H
E to c
-P O3 id
c
o
Si
(U
4) CU
tu £
ft O
o
o
Q
A
O
a)
C
-5-
-------
TABLE 2-1 (Reverse side data)
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.
-6-
-------
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 quanti-
tative results by the coordinating laboratory.
2.F. 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 2.H.).
After making these data evaluations and calculations, the completed
report forms and chromatograms of the laboratories with the poorer
rankings are subjected to detailed examination to determine, if possible,
the reasons for the inferior performance. Having the actual recorder
chromatograms available for study is a vital part of this operation since
it allows the coordinating laboratory to check such factors as column
efficiency, sensitivity of detection, instrumental problems such as base-
line noise and improperly adjusted recorder gain, proper operating para-
meters 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 sug-
gestions for corrections to improve performance.
The reports from all the other laboratories are then scanned to locate
any irregularities which might lead to future problems. A general letter
is drafted, a copy of which is mailed to all participating laboratories.
This letter discusses common analytical difficulties encountered by
several laboratories and offers suggestions which appear to have general
applicability for improving compound identification and quantitation.
With this general letter, each laboratory receives a copy of the Summary
of Results with each laboratory identified by an identifying code number,
a copy of the Relative Performance Table, and a private critique to
each laboratory exhibiting special need for help (Subsection 2.1.).
2.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 2.K.
-7-
-------
Table 2-2 shows data from a group of 34 laboratories participating in an
inter' .'Jburatory control program for the first time as a group entity.
The distributed sample consisted of a precise formulation of eight chlori-
nated pesticide and metabolite standards dissolved in pure solvent in a
sealed ampoule, for which no cleanup steps were required. The mean and
standard deviation values were calculated after rejection of the outlying
values designated by asterisks. (See Subsection 2.K.e for description
of fitness test). The precision (relative standard deviation) was con-
sidered "good" for this type sample only for the compounds lindane,
heptachlor epoxide, and p_,p'-DDT, "fair" for p_,p_'-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 beyond the "acceptable" level.
Table 2-3 shows, for comparison, results on the same sample (except for
a more difficult, lower endrin content) by a group of laboratories which
(except for one) have been in the quality control program for the past
several years. The calculated average RSD value is seen to be 7.7 per-
cent and the average Total Error 20 percent, both "excellent" performance
values. The average total time spent on the sample by this group of
laboratories was 1.5 days in each laboratory. During their earlier years
of participation, the data output of these laboratories was similar to
that shown in Table 2-2, but continuing participation in both Inter- and
Intralaboratory Programs resulted in gradual improvement to the levels
shown in Table 2-3. An example of a factor responsible for the poor
results in Table 2-2 is that 33 different GC columns were employed by
the 34 laboratories, while the experienced group represented in Table 2-3
used only the optimum GC columns and operating parameters recommended in
the EPA Pesticide Analytical Manual and Subsection 5.G. of this Manual.
Table 2-4 shows typical results for a more difficult blood serum check
sample by 18 laboratories with experience in the quality control programs.
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 labo-
ratories for a fat check sample. Average RSD and Total Error values are
obviously not so good in this case.
2.H. 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 this performance ranking, namely correct
identification of all pesticides present, correct quantitative assay of
the pesticides, and non-reporting of pesticides not present. The rank-
ing scheme incorporates all three of these criteria and provides a
numerical score for each, the laboratory with the best results receiving
8
-------
TABLE 2-2
INTERLABORATORY CHECK SAMPLE NO. 26, MIXTURE OF STANDARDS IN SOLVENT
SUMMARY OF RESULTS .
LAB CODE
NUMBER
FORMULA-
TION >
45.
47.
48.
52.
53.
54.
66.
68.
69.
71.
72.
73.
83.
84.
85.
87.
88.
89.
90.
92.
93.
95.
96.
97.
113.
113A.
130.
135.
137.
160.
161.
162.
163.
164.
Mean
Std.Dov.
Rel.Strt.
Dev. ,%
Total
£rror te/o
PESTICIDES REPORTED IN PICOGRAMS PER MICROLITER
Lindane
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
4.3*
9.0
10
11
12.4
9.4
9.0
9.2
11.2
9.8
11
11
9.7
10.1
14
4.5*
10.5
1.56
14.9
36
Aldrin
10
46*
11
9.2
11
14
3.0*
8.0
10
3*
5.0'
8.8
12
14.1
9.6
8.3
10.6
14
8.0
6.9
10
12
8.0
12.1
11
8
7.8
11.0
9.4
10
9.0
9.6
12.2
11
4.8
9.9
2.31
23.4
4.7
Kept.
Epoxide
10
53*
11
8.7
14
14
6.0*
9.2
8.5
10
9.4
10
13.5
8.4
8.1
9.6
12
10
5.1*
9.0
8l*
6.C
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
?8
70
45
90
22*
80
119»
89
69
72
86
74
73
102
70
77
78
18*
?4.C
12. S
17. l
36
Dieldrin
20
145*
31 -
16
18
14
21.6
92*
100*
19.5
19
29
10
18
22.8
22
10
11.5
20
32
21
21.7
23.4
16.4
16.8
23.8
14
23
16.9
24.3
19
13.4
19.6
5.64
28.8
58
Endrin
30
87*
29
15
87*
30
--.
6o»
42
27
39
15
2?
32
12
14
...
33
35.5
33.6
29
20
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
16
P.P1-
DDT
100
578*
195*
113
__-
91
2.2*
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
Por-
thane
0
379
34
Ethior
0
68
o.p'-
DDD
0
?0
P.P'-
DDD
0
4.
16
13.?
12
Mi-
rex
0
29
DCPrt
0
9.
Rej<
out:
B-BHC
0
2.5
tpd a:;
Pffi
-9-
-------
EH
I
O
Q
OS
I
OS
g
X
M
NO
CM
O
z.
P-.
W
O
3
o
>H
OS
K
o
m
CO
EH
W
CS
>H
OS
C/3
I
CM
OS
W
EH
M
O
O
M
S
OS
W
d.
m
o
o
o
M
o<
g
M
W
EH
K
O
W
PS
W
W
o
M
o
I-H
00
W
PH
>H
OS
O
LABORA1
i
"p,P
-Q
P.P
1
"P.EH
-o
ftQ
"PH'EH
0|Q
C
H
d
W
Dieldrin
i
°^Q
P|Q
Q)
3
-P X
P4 0
a> ft
KW
C
ti
0
H
<
0)
s
d
C
H
HH-
s
s
o
0
0
0
o
TH
o
W
VA
O
CO
vr\
CV
o
1-H
o
TH
o
1-H
it ion »
FormulE
CM
1-1
* *
^ONcnoNONCMOONOTHOONi-Hcn
ONONCOONONOOONOi-iO ON ON GO
TH TH rH i-H TH
fVC^- -3-OJvOOOcO J- vr\
r^N^}- T-I oj- C>-TH c^vr^vcw^r^csjo
CJCJCMr^CVOJCJOJCvJCMCVJCJCMi-)
*
CMCM ONOCOCMJ* C^- CV1
j- r^r^ivnoj- vnxAC^-oo c^^i^}- c^>
C^O\ \OOC\1 C^C^CVJ \O
OxOOCOWvAOTHVAOsOOOOOsC^-
rHTHT-IOjTHCvltMTH^CVlT-ICVjTMrH
>r»
* *
,-|XAxn-i>-c^-oco c^-c^-c^-i>-tvc»ooco'>r\
^H
*
r^c^- j-j-oJOO^Acor^vocMO
OO\OOOOONCOCTvO\OOO\C^
v-< r-< i-4 i-l i-l i-l
*
r-\o c^ic^-ONC^-O\oxr\i-ic^o\c\2
O ON O *-* ON ON ON ON ON ON O O CO !>-
iH i-l i-l i-( i-l
*
C^ ONVTN.T-1O-VACO CVlf\ C^\CO
ONOONOOOOONONOOOOONNO
T-I T-l TH 1-t f-t 1-1 1-1
>>
r-\ 6
p a
C -*fci
a> ijo
\ 0 0
So.
THCNlC^\^-\A\OC^-COOvOT-4CMC^\J-
1-1 T-I 1-1 1-1 1-1 fl tJ
0) 0)
i^ C
H
N °
v-^ -^
T3
Q) tQ
+^ ^
o o>
(U iH
-» ^
a) ra -p
PS nj 3
* o
o
c}- tv c^-
ON ^J? -£ 0
O\ i-(
ON
CS- NO
^- CM in 00
CN) iH ^1-
CO
vr\ C^- C'N
J- TH CN) Cv-
VO
ON cn
ON in O xr\
T-I iH CM
C^-
C"» C^
tv Jh >A ^~
C^- i-l
CM
ON u~\ c<>
ON O vr\ TH
T-I
O
ON NO i-l
ON O NO C^v
T-»
-*
ON xr, vr\
ON O V> CM
TH
^S.
C .
o > Mi.
H CO
P Q ^
a o
H . 14
CO -H W
Q W
1-t
C - oJ
ctf t3 ,-H -P
-------
I
CM
S
I
s
0^
w
w
.
\f*\
CM
O
EH
W H
(-1 t3
Pi W
S W
< K
w
X O
O
S OS
O <
**H §
K £>
CD C/2
EH
<£
O
pq
H
PS
o
EH
O
<3j
^
pq
o
|ij
o
pq
i
CO.
EH
R
R
I
"pM
«
PI
W
P
R
I
w
PI
PI
&-)
p
p
1
"PI
..
o|
C
H
-d
H
(U
H
R
0
W
P
O
o
o
o
o
CO
^f*\
_^J-
o
vO
o
^J.
T
c
o
H
-p
H
^
O
*-)
V
Q\ o r^\
j- xniA^t vn^}- \o\pioo
* *
CVi V^\ cj" ^O C^ _j- 0"^ 0^ O^N ^j- ^}" C°\ C°N -^}" -^ ^\ -3" -^3"
rH OJ O^ JJ" XP\ VQ C^ CO ON O *~H CSJ C1^ ^}" V^ NO C^- 00
T* CO
O $*t
-(-> ^.
H 0)
F R h
C(J O
H . ^
> -d ^
CD -P W
R CO
C co
oj n H -P
CD -P CD 0
S W PS EH
-11-
-------
EH
P
W
^:
M
CM
O
00
EH
i-l
M
W
EH
O
o
o
l
CM
CQ
<
EH
3
O
M
h3
M
^
PS
W
d.
w
EH
CC
FU
Ef,
M
P
W
EH
C£
&
(S
rn
CICIDE
w
w
CL.
>H
«
O
LABORAr
o
K
«
i
CO.
EH
0
P
I
Pi
P<
W
P
Q
1
ni
PI
P
P
P
1
QJ
P!
EH
P
P
I
0|
C
.H
SH
T)
0)
rH
p
-
o
o
0^1
o
o
xH
o
T
c
o
H.
p
Formula
o
CM
O
O
*
T-Hf^ONvoocOvOOC^-OP^OOCMOv-ic^i
i-HMDOCOvOOP^ONVOT-lOMDO-^-HCMC^-vo
r^cMCMCMCMC"^cMCMCMrr^JONOV£)(N-
C^-sQM^O^iACQ r-{ NVPvONCN3C^-C«^xr\c'Axr%o
COCO O-ONC^-CO CS-C^COCOOOvQ ONVOM3 C^-O\
o-^- [>-CMCOCOJ- ONcnc-^o CMCO c^\^}- CVION
r^^-J-^-^-J-^i- Csj\nc^^- r^\^j- r^cMc^vr^
*
vor^cMj--j-vnt>-oocor^c^cjc^iO!>-c\j
[>-\O ovo C^-C^vO tvco\O C^--d- [>-vr\xr\vovr\
rH
#
O * *
xnco o !>-vpi^Hj- xn-!>-^P\O V^OD C^C^\VOVO C^ivH ^)
T-) -rH
* *
0 O
f^it>--cJ-\OOT-HVOtV(MO^(Mi-t-vO 1 OOOVQONOOVTiMDvO
lOrH IOOOO IrHOOOOOOO
rH^-vr^vclC^-OOCJXr-IJ-invO-d-v/AVOTHC^^}-
rHrHTH'rHCMCMCMC'^C^NC^
1
O CO
0) (4 , -01
n ' ^
0) -c) 4^ 0>
K CO ^ -H
* +^ 0 rH
fO CO
C^- C~\
CM 0 -3- -3-
vH C°\
0^ O
0 0
CO rH C\J <*^
TH f°\
0
O 00
^ 0
0 0 O 0
CM -*
CO CM
\0 ^
0 0 CO ^0
fH ^t
0 J-
\O TH
O 0 C^ ^t
CM ^n
CM
co ^j-
CM O
O o vr\ vr\
TH r^
O CM
CO CM
0 0
O O CO xr\
CM M3
^.
0 ^ ^.
3* 2
's! ' £
0) ^ W
° 5 H
C . _ oJ
-------
the highest score. The maximum possible score is 200 points, divided
equally for correct identification and quantitation. A detailed expla-
nation 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. If all reported values should
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)
is calculated and divided by the standard deviation (previously calculated
for each compound) to obtain a "weighted deviation". This value is sub-
tracted from the point value of the compound, the difference representing
the point value for quantitation of that compound:
Compound Quantitative _ Compound Point Absolute Error
Score Value Standard Deviation
The total quantitation score is the sum of the individual compound point
values.
An important aspect of the quantitative portion of the ranking is the
use of the standard deviation for each compound. If the precision of
the group for the analysis of a particular pesticide is poor, the stan-
dard deviation for that compound will be a relatively high value. If a
laboratory has a large absolute error for this one compound but an other-
wise excellent performance, division of the error by the standard devia-
tion will lower the point loss so that an unduly heavy 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 a relatively
inferior total score of 125.
-13-
-------
Table 2-6
CALCULATION OF TOTAL SCORE FOR RELATIVE PERFORMANCE RANKING
B-BHC
£_,£_' -DDE
Dieldrin
o,£'-DDT
p,£'-DDT
a-BHC
*0f all data from
Formulation Reported Analytical
pg/yl Values pg/yl
30 27
40 40
20 50
10 Not Reported
50 47
None 10
participating laboratories
Standard
Deviation*
2.10
1.75
2.50
0.60
1.44
Point value for each compound is 100 + 5 = 20
Identification
3-BHC
p,£'-DDE
Dieldrin
o,p-DDT
p,£'-DDT
20
20
20
0
20
80
Quantitation
3-BHC
£,£'-DDE
Dieldrin
p_,£-DDT
£,p'-DDT
-20 Penalty for false identification of
60 Total identification points
30 - 27
-BHC
20 -
20 -
20 -
2.10
40 - 40
1.75
20 - J30
2.50
50 - 47
1.44
= 20
= 8
= 0
= 18
Total quantitation points = 65
Total laboratory score 60 + 65 = 125 (of a possible 200 points)
-14-
-------
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.
2.1. 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 which 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.
2.J. 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.
-15-
-------
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
137
181
169
168
168
164
159
158
157
146
133
128
127
123
115
84
25
_!/ Values r
2/ Total po:
Tide confidence limits
.j-ble score 200 points
-16-
-------
TABLE 2-8
RELATIVE PERFORMANCE RANKINGS - CHECK SAMPLE NO. 21, FAT
Lab. Code
Number
15.
16.
8.
25.
7.
4.
26.
33.
5.
34.
11.
9.
31.
6.
1.
14.
24.
Compounds
Missed
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
False
Identifications
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
No. of
Rejects I/
0
0
0
0
0
0
0
1
0
1
1
1
0
0
0
0
3
Total
Score 2/
198
198
198
197
195
193
191
191
191
189
188
187
185
168
168
167
165
I/ Values outside confidence limits
2/ Total possible score 200
-17-
-------
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
\f Values outside confidence limits
2/ Total possible score 200
-18-
-------
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 a fairly rapid and simple method, 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 pre-
cision 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
latest (1974) figures show that laboratories are making still further
progress in their performance.
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 utili-
zing a constant speed mixer (Subsection 7.A.). The results of the
blood check samples illustrate again the dual value of the Interlabora-
tory Control Program in upgrading laboratory performance and identifying
weak analytical methodology.
2.K. STATISTICAL TERMS AND CALCULATIONS
a. Accuracy and Precision
Precision refers to the agreement or reproducibility of a set of repli-
cate results among themselves. Precision is usually expressed in terms
of the deviation of a set of experimental results from the arithmetic
mean (average) of the set. Accuracy is the nearness of a result or the
mean of a set of results to the true value. Accuracy is usually ex-
pressed in terms of error.
Good precision often is an indication of good accuracy, but one can
obtain good precision with poor accuracy if a systematic (determinate)
error is present in the method used. Systematic errors are either
positive or negative in sign. The other general classification of
-19-
-------
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
-20-
-------
TABLE 2-11
PROGRESSION OF RESULTS
BLOOD CHECK SAMPLE
Inter laboratory
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
Recoveries ,
%
*
*
*
*
96
91
100
*
Average
RSD,
%
36
29
21
17
14
12
13
16
* Unspiked samples for which the actual pesticide levels were not known
-21-
-------
errors in analysis are indeterminate 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. It is for this reason
that 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 rectified 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 ppt 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) must be carried out with
this in mind. To the contrary, 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 quantitative and adequate when values +_ 25 percent or better
are obtained on recovery samples fortified at 0.01-1.0 ppm.
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 ppt relative error. As an example, an absolute 0..2 yl error
in injection of a sample for GC corresponds to .x. = 20% for a
1.0 yl sample but only ,X,. = 4% for a 5.0 yl sample. It is
O \J
explained later in Section 3 that low sample injection volumes are to
be avoided because of high potential errors.
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 uncer-
tainty is ^jrf 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 judgement 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,
-22-
-------
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
""10" "100 x l^O = 0.01%. If the measurement was actually made only to
the nearest 0.1 meter and the final zeroes only indicate the magniture of
the number in mm, the number would better be written in exponential
form, 1.01 x 1Q4 mm, to indicate an absolute uncertainty of +1 x 102 mm
and a relative uncertainty of 1 x 10 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:
90.7 90.7
8.81 8.8
+ 0.551 0.6
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 ng = ngs a'pag^ is used if response is linear over
the range in question. If 1.0 ng standard gives a peak of 9.0 cm
height (measured to the nearest 1 mm) and the unknown peak height is
12.0 cm, the ng of unknown is 1.00 ;*-? ? = 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.
-23-
-------
c. Average
The average or mean (X) of a set of n values is calculated by summing
the individual values and dividing by n:
x = E xi
n
d. Standard Deviation
Standard deviation (s) of a sample of n results is calculated by use of
the equation:
" (Exi)2
s =
n
n - 1
e. Fitness Test to Determine Outliers
The PTSE 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 of all results in the set including questionable
value(s).
(2) Compute the plus or minus arithmetic deviation of each value in
the set including the questionable value(s).
(3) Compute the average arithmetic deviation of all values including
the questionable value(s).
(4) Compute the arithmetic deviation of each questionable value by
subtracting it from the mean.
(5) If the arithmetic deviation of a questionable value is higher than
2.3 times the average arithmetic deviation of all values including
the questionable one(s), the questionable value is rejected as
lying outside the 97% probability range.
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 cri-
terion 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.
f. Total Error
Total error is a method proposed by McFarren et al. [6J for combining
-24-
-------
precision and accuracy in one reporting expression:
Absolute Value of _
the Mean Error
Total Error = x 100
True Value
where s = standard deviation. In £eneral, Total Error values < 25
percent are considered excellent, <50 percent acceptable, and >50
percent unacceptable.
Specifically, in the interlaboratory control program, Total Error is
calculated from the following equation:
x ~f" 2s
Total Error = x 100
where x = the arithmetic deviation of the overall mean obtained for a
given pesticide from its known formulation value and y = the formu-
lation value.
g. Numerical Conversions
1 g
1 mg =
1 yg =
1 ng =
1 pg
1 ml
1 yl
1000 mg
1000 yg
1000 nq
1000 pg
I0"12g
1000 yl
10~6 1
2.L. References
[1] Radomski, J. L., and Fiserova-Bergerova, V., Indust. Med. and
Surgery, 34, 12 (1965).
[2] Mills, P. A., Onley, J. H., and Gaither, R. A., J. Assoc. Offie.
Anal. Chem., 46, 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).
-25-
-------
Section 3
INTRAUTOATORY QUALITY CONTROL
3. A. 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.
3.B. 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.
-------
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.
3.C. NATURE OF 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 appriximating 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 3.F.
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.
- 2 -
-------
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 2.D.).
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 2.G.) 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.
3.D. 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 ninimum 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
- 3 -
-------
a team, e.g., with one chemist preparing extracts and another doing the
determination, SPRM samples should be handled in this same normal
fashion.
3.E. 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.
_ 4 _
-------
TABLE 3-1
RECORD OF ANALYSIS OF STANDARD REFERENCE MATERIAL
Laboratory
Analyst c
Sample
No.
Analyst c
i
c
Dr Tear
Date
y£ Teari
n
Aldrin
i
S-EHC
Kept.
Epox.
Diel-
drin
o,p'-DDT
p,p'-DDD
Media
p,p'-DDE
p,p'-DDT
-
J
1
Reporting units shojld be in ppb or ppm. Observe standing instructions for
F. reporting levels.
- 5 -
-------
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)
Blood (PCP)
Blood (Other) (specify)
Adipose Tissues
Other' Human Tissues
Air
Soils
Stream Sediment
Water (multiresidue)
Water (Other) (specify)
Urine (alkyl phosphate)
Urine (Other) (specify)
Housedust
Fish 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.
- 6 -
-------
3.F. QUALITY CONTROL CHARTS
In addition to recording numerical results of each analysis of an inter-
nal SPRM, each analyst or team constructs a Quality Control Chart
or Curve on which the result of each analysis is plotted as a point in
chronological sequence. 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
each 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 ppeparation 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 £,£' -DDE and £,£' -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, the value for one standard error unit was taken as 10 percent
of the spiking level for each pesticide since prior data showed the methodo-
logy in use should produce an in-house precision not greater than 10
percent RSD.
2 +
Figure 3-A. Laboratory A control curve* for blood SRM, three-month period.
If: ! - .1.: i i -i
#*-!
2-
2 +
p,p'-DDE
=^4^1
p,p'-DDT
-------
figure 3-B. Laboratory B control curves for blood SRM, three-month period.
2+"
The known formulation or spiking value is subtracted from the experimen-
tal value obtained for an analysis of the in-house standard sample to
obtain a (+) or (-) arithmetic deviation (difference). This difference
is then divided by the calculated standard deviation 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 SRM 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. The SD is calculated by multiplication of
the formulation value by the percent RSD to give a standard error value
which should be valid throughout the life of the specific SRM: 150 x 0.10
= 15 ppb. 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 135 ppb, the second point plotted along the horizon-
tal axis would be calculated as:
135 - 150
= -1.0 SEU
150 x 0.10
When constructed in this way, quality control charts readily show levels
of accuracy and precision for repetitive analyses by a given analyst.
Figure 3-A demonstrates excellent precision since the results all fall
along an essentially horizontal line. Accuracy is good because this
line is well within the control area of 2 SEU, all recovery values being
slightly low, probably due to an inherent negative determinative error
in the procedure being used. Figure 3-B, on the other hand, demonstrates
very poor analytical performance in both accuracy and precision. Nine
of the repetitive values for DDE and eleven for DDT are out of the
-------
acceptable control range of 2 SEU.
Control charts also highlight cases where errors are present exerting
similar effects on the analyses of several pesticides. The following
Figure 3-C, for example, demonstrates rather poor precision and also a
distinct correlation in the configuration or shape of the curves for both
compounds. This signals some common error proportionately affecting both
compounds, most likely the extraction step in this blood analysis.
Figure 3-C. Laboratory C control curves for blood SRM, three-month period.
2 +
p,p'-DDE
2-r
2-
p,p'-DDT
2 + -
2
__ rr.
From time to time, the following question is asked: "What is to prevent
an analyst from 'fudging1 the control chart points so that his curves
will appear significantly better than they should"? This can and has
occurred in very rare instances. The alert laboratory administrator,
however, should have little difficulty detecting the doctoring of curves.
When a chart is submitted that is virtually a straight line such as that
for p_,p_'-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 appre-
hensions. 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.
3.G. 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
- 9 -
-------
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 which
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.
3.H. ANALYTICAL BALANCES
Most laboratories contain balances of two types. Rough triple beam or
Dial-O-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 2.K.b.). This leads to
-10-
-------
a total accuracy and precision of
0.0002 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.
3.1. PURITY OF REAGENTS
The purity of reagents, solvents, adsorbents, distilled water, etc. is of
-11-
-------
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 which 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 cap-
ture detection. Sulfur and sulfur-containing compounds can be present in
solvents and column materials, as well as in certain substrates (onion,
cabbage, turnips), and can give rise to peaks easily confused with pesti-
cides [1] .
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 photochemical reac-
tions can produce compounds from pesticide-grade hexane which are detect-
ed by an electron capture detector and interfere with pesticide residue
determinations [2]. 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 evap-
orating a portion by as great a concentration factor as will ever be em-
ployed 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 yl into the gas chromatograph equipped with the detector of
choice. Detector response is recorded for 20 to 30 minutes. No peaks
which 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 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.
Solvents containing oxidants are especially troublesome in causing losses
of organophosphorus pesticides. Acetonitrile and ethyl ether are two
common solvents which 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. Ether is tested for the presence of per-
oxides by reaction with potassium iodide; if present, peroxides are re-
moved 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)].
-12-
-------
3.J. DISTILLATION OF SOLVENTS [3]
Distill reagent grade acetonitrile over reagent grade AgNO (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., and boiling chips for each distillation. Test the distillate for
interference. Distill acetone, hexane, benzene, chloroform, methylene
chloride, and ethyl acetate as above, without use of AgNO .
3.K. MISCELLANEOUS REAGENTS AND MATERIALS
Any other reagents used in the extraction or cleanup steps are also poten-
tial sources of contamination. These reagents, such as sodium sulfate
(Na SO ), 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, Na_SO.
is extracted in a reserved Soxhlet apparatus, the thimble of which is
pre-extracted before the first use. Methanol followed by hexane or petro-
leum ether are cycled for several hours each, after which the Na SO. is
dried and stored at 130°C in the Florisil oven in a glass container with
a glass cap. Plastic fiber pack liners have been found to contribute
PCB's and phthalates to Na SO. which must be removed by this procedure.
Filter paper should be checked by washing through the solvent to be used
and injecting a sample, after concentration, into the gas chromatograph.
Teflon and aluminum foil should be rinsed with an appropriate solvent.
Solvents in polyethylene wash bottles can become easily contaminated with
electron capturing and UV absorbing species and should be tested for im-
purities. Better still, avoid the use of plastic wash bottles, and use
all glass.
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 poten-
tial 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
-13-
-------
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 considera-
tion. Polyethylene bags are totally unsuitable for samples to be examined
by electron capture GC or TLC because of trace contaminants which may be
present. As an example, it has been reported [1] that polyethylene con-
tains a contaminant which reacts with AgNO chromogenic reagent, giving a
TLC spot close to that of p,p'-DDE and having similar GC retention times
to p_,p_'-DDE and p_,p'-DDE. Glass containers with aluminum foil or Teflon
lined caps are generally acceptable as sample containers.
Other examples of problems with reagent contaminants have appeared in the
literature. Bevenue et al. [4] reported on the contribution of contami-
nants by organic solvents, glassware, plastic ware, cellulose extraction
thimbles, filter paper, and silica gels to water samples causing inter-
ference with subsequent GC analysis in the ppb range. Prior to their use,
heat treatment of glassware and the silica gels was recommended to elim-
inate contaminants, while plastic ware and filter paper were excluded
from the procedure. Levi and Nowicki [5] found that cloth bags contained
residues which were absorbed by cereal grains stored in these bags and
gave spurious GC peaks with electron capture detection. The same workers
[6] found that Na SO , 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 [7] reported on the contribution
of extraneous components by high purity, analytical grade basic reagents
used for adjustment of pH during isolation steps in the analysis of chlo-
rophenoxy acid esters and ethyl or methyl derivatives of hexachlorophene
and PCP in plant and animal tissue and water samples. Baker et^ al_. [8]
found contamination of acetone with an impurity corresponding to CCl. and
interfering in the analysis of the latter pesticide (fumigant) by EC-GC.
It was shown that this contamination could be caused by CCl. in the lab-
oratory atmosphere, possibly arising from the use of aerosol propellent
cans for spraying thin layer chromatograms.
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 running through the entire procedure
without including the sample itself.
3.L. 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.
-14-
-------
b. Rinsing with tap water.
c. -vinsing 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 neoprene-coated 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
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.
3.M. 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.
-15-
-------
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 included weight of primary standard, concentration of all subse-
quent serial dilutions, and the dates of preparation of all dilutions.
In multiresidue analysis, it is common practice to prepare final working
standards as a mixture of pesticides of interest to the laboratory, this
subject to be treated in some detail later in this section. Such a
mixture should be assigned an identification number and so documented in
the 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 A 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.
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 orhigher 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.
Concentrated stock standard solutions are conveniently made up at a
200 ng/yl concentration by weighing 20.0 mg of pure standard and diluting
to 100 ml. If the primary standard is given as 99.0 percent pure,
20.0 - 20.2 mg; if the purity is given as 90.0 percent, the weight will
0.990
be 20.0 =22.2 mg.
0.900"
16
-------
-fure 3-D
PREPARATION OF CONCENTRATED STOCK STANDARDS
No.
Compound
Final Gro^ss Wt
*Tare Wt
Date / / Chemist
Lot No. Purity
g Dilution Vol.
g Concentr.
%
ml
ng/pl
Net Wt_
**Adj. Net Wt
ing
No,.
Compound
Final Gross Wt
*Tare Wt_
Net Wt_
**Adj. Net Wt
Date / /
Lot No.
Chemist
Purity
mg
Dilution Vol.
Concentr.
ml
_ng/yl
No,.
Compound
Final Gross Wt
Tare Wt
Net Wt_
**Adj. Net Wt
Date / /
Lot No.
Chemist
Purity
mg
Dilution Vol.
Concentr.
ml
ng/yl
No,.
Compound
Final Gross Wt
*Tare Wt_
Net Wt
**Adj. Net Wt
Date / /
Lot No.
Chemist
Purity
Dilution Vol
Concentr.
mg
ml
ng/yl
*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.
**Corrected for purity of primary standard.
-17-
-------
Figure 3-E
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/Pl
ml
ng/M.
Date / /
Chemist
Strength of Concentrated Stock
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
_ng/ui
ml
ml
NO.
Compound
Date / /
Chemist
Strength of Concentrated Stock
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
ng/ui
ml
ml
_ng/yl
NO..
Cornoound
Date / /
Chemist
Strength of Concentration Stock
Aliquot of Concentrated Stock
Dilution Volume
Final Concentration
ng/VH
ml
ml
ng/Pl
-18-
-------
Figure 3-F
PREPARATION OF FINAL WORKING STANDARD SOLUTIONS
,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/yl ml Vol. (ml) pg/yl
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
-19-
-------
Toxicity levels and relative stabilities are important factors which
dictate the methods of handling and storing various pesticide standards.
Highly toxic pesticides (low LD5Q 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
-20-
-------
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 3? 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 yl
in a 5.0 yl 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
-21-
-------
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_'-ODD 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 5.L.)
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 linciane, 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
-22-
-------
of each of these by separate pipets to individual 100 ml volumetric
flasks to prepare intermediate stock standards. Transfer 2.0 ml of
lindane, aldrin, and dieldrin and 1.5 ml each of o_,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 A3
Lindane Aldrm Dieldrin o,p'-DDT p,p'-DOT
EH O D D D Primary Standards
I I I 1 I
TT n n 20 mg eoch
_ ' -' ' Cone Stock Stds.
20ml I 20 ml (200 ng/^l eadl)
4ng/«l
nterm Stock Stds
Final Working Standard Mixture
e. Storage of Standards
Working standards may be stored in 30 ml inverted stopper bottles in a
refrigerator at ca. 5°C. They are warmed up to room temperature
before each use. Organochlorine standards are renewed monthly and
organophosphate standards semi-monthly. Alternatively, storage may be
in 1 dram vials with molded plastic screw caps cjid a Teflon liner. The
vials are stored in a freezer for periods up to two months, a fresh
vial being taken each week. The current organophosphate standards are
stored in a refrigerator during non-use periods; organochlorine stan-
dards are stable for at least one week at room temperature. If work-
ing standards are used frequently, evaporation of solvent caused by
repeated removal of the container cap will cause serious problems. In
-23-
-------
this case, a fresh portion of standard may be required each day. The
vial container protocol is outlined in the EPA PAM in Section 3,B.
3.N. CALIBRATION AND MAINTENANCE OF THE GAS CHROMATOGRAPH AND ACCESSORIES
It is essential that the entire gas chromatograph be maintained in top
operating condition if high quality analytical data is to be produced.
In appraising results of interlaboratory check samples, it is clear
from data and chromatograms that this is not the case in some labora-
tories. Section 5 will present details of proper operation of a gas
chromatograph. This section will offer guidelines for maJcing 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 each detector installed
(this will also serve as a record for Atomic Energy Commission
inspection).
b. Background current (BGC) profile furnished with the detector under
the EPA Interlaboratory 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.
-24-
-------
-c
£
Q
CO
UO
31
; £ £
i
JC C
) 4J n
<3 X (_
i
3 [
4->
i/t
+>
O
C
tj
>»
i- O
QJ 00
1 ** 1
X ra t/>
.a X
J> C C
- n3 O
I r^ £
J O U_ (.
n
-**
.C T3 T3 O 4.
I- 0 (, I- , H- C
*- 1* T3 V- *- c
QJ \O\ O 4J C iU !
+-* t/t Q) "O i
t/l ,= O U 1J I- C
t/l O- C r- *-> £
jo E u 1/1 c Q»
02: C QJ coc G *
* CO O CL -r- . O 4-»-r- 4-> 4
l/» 1- (U U 4-> O
13 i/l Ol OJ-C t- OJ «TJ i- I
W> £1 Q_ 4-1 -r- C r i
i nJ-t- -i +->C3 i C
T5 i V\ U 4-> QJ OJ (J C «»-
t- fO G t- * '
O -*l/> 3 E 3 -ii UirtO) O +
(Um fa i x cu O i- qj
zc jCQJ >o-»- jro-f- x:<
J -r- O U4 (J E tj i 5 C_) t
i
3
3
:L
D
1
3
rt
>
3 C
- O
-> r-»
= »-»
3 U
J tj
O
rO -r
Ul -U
i- iO .-O
Q> O> t.
> V.
O Ol t*~ (U
S w t: o
-M OJ OJ U
-*-> JT ->; o>
t. fl3 J C > .
O U ra CU T3
c c c «3 3; x: i-
O -r- jC t O* 3C t- ^3
O -X _-^ JD W O O
(13 ffl O
c
o ex u r i. -o i-
^- J3 3 0 fl
C C
a> i
L. 4-
U r-
3 C
u ;
-o c
c c
_J -
e.i:
CT> 1
^rf rt
3
3 O
' a.
c
> o
n 03
VI
1 <
-. ^^, C
3 ^| -
L.
i t)
X "U C
3 u "
; 3 H-
3 O >
1- tfl
Jlt-
*J +J
fO C
3 j^:
*-» O
IT) U")
TJ V-
C O
rO
j«:
c: TJ
o a»
i- O.
rQ ftt
L.
4-1 (1)
tr o
,
c:
«3
u
o
^
J<
ai
-C
0
*f-
c o
E O»
1 3 (rt 3
OJ * Ul ^»
> O OJ CL
QUO
£ X -M
Q>
r- +J 4J i/t
o ai w
CnjD U) 1- OT
4-» x; o
< *-» ^,
ill LJ 'O
J- O 6
>,4J O
C71 QJ 3 L.
C ^3 C f-
«~
O C T- CM
«J
LO
:5
,
o
c
o
CT
OJ
L.
H-
1/1
-C
t
en
c
3 CT
- C
3
L.
(U
cx
^
I/)
*+- +J
r- (J
fO
-0 i.
(U *-»
»-» X
rtJ (U
L>
r- Q>
T3 ^-
S t
«J
jQ ,
%.
>> O
13
E v*
t.
Jrf OJ
» 01
r- Ol
r- l-
ra nj
C3 r~
o
(U
O
C
X
OJ
t/l
c
o
a.
t/t
£
rw
U
(U
.C
u
3
O
2
a.
o
4->
QJ
-M
ro
'aJ
L.
o
c
rd r-
QJ
U
3
4-»
.25- ,
e "o c,
*£>
C C
J »- C
4-»
m
ij
>
4(
I/I
C
TJ
m
p
£
j rj
c
o
4-*
W
g
ai
4^
4->
rtJ
CTi
C
4J
13
'_
£
U
c
E
i
>>
rO C
ai o
* o
c
1- t,
ai a;
H
i_
0
4J
U
a*
4->
a>
T3
**-
0
t
3
4-*
m
L
QJ
d.
^
r
o
i ;
o
o
T3
C
*o
u
*f-
U)
*->
* 5
^ ^
10
3 *->
3 m
,-o
^J 01
^
c
rs
o
u
u~
c
-w:
1-
o
' S
c
o
I 01
; c
J O
4J
(O
^1
4->
c
o
+J
«3
3
c
a/
4-»
4->
fO
o
c
(TJ
c
0
rt3
V-
4->
C
QJ
(J
c:
o
u
TJ
f
o -
vo
C
o o
t-» -r-
O Li
m i»
**- **-
O OJ
"O
ftj QJ
>
I- JD l-
a* QJ
CLT3 4-»
Q O
a o
I I
CL) o
aj M
t C3
°?
U Qj
f QJ
^1
-25-
-------
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,p'-DDT
p,p'-DDT
Concentration ng/yl
0.010
.040
.010
.010
.020
.030
.040
.050
.080
.080
.080
.090
.100
3.O. 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
-26-
-------
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.
3.P. REFERENCES
[1] Ruzicka, J. H. A., and Abbott, D. C., Talanta, 20, 1282 (1973).
[2] Williams, I. H., J. Chromatog. Sci., 11, 593 (1973).
[3] Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Canada, 1973, Section 12.1 (b).
[4] Bevenue, A., Kelley, T. W., and Hylin, J. W., J. Chromatogr., 54,
71 (1971).
[5] Levi, I., and Nowicki, T. W., Bull. Environ. Contain. Toxicol.,
7, 133 (1972).
[6] Levi, I., and Nowicki, T. W. , Bull. Environ. Contain. Toxicol.,
7, 193 (1972).
[7] Bevenue, A., and Ogata, J. N., J. Chromatogr., 61, 147 (1971).
[8] Baker, P. B., Farrow, J. E., and Hoodless, R. A., Analyst, 98,
692 (1973).
-27-
-------
Section 4
EVALUATION, STANDARDIZATION, AND USE OF MATERIALS FOR
PESTICIDE. RESIDUE ANALYSIS
4. A. 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.
-------
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 because of degradation
by peroxides in ethyl ether and/or impurities in petroleum ether.
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 con-
stituent 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.
-2-
-------
>H
M
M
K
M
VI
S
O
M
O
W
W
t-l
W
c
0
H
-P
C)
oj
)H
pL,
O
xp\
j^j
0
H
P
o
«J
iH
ftH
^.
\|-^
,_)
£j
O
-p
o
s
JH
wo
M)"
f
0)
o
o
Q)
O
0
NO
\
o
0
\jT^
o
0
1
o
o
o
o
1
o
o
o
o
1
o
0
C
0
PH
o
0
<*>
o-
ON
o
0
r-t
0
X
\
d
0
t-r-"
1
ca
c>
[^.
ON
1
0
o
vH
0)
-------
Figure 4-A. The effects of polarity variation of editing solvent in Florisil
partitioning of 7 pesticides. Absolute ethyl ether mixed with
0, 2 ,and 4% absolute ethanol.
Elution Fraction*
Hept Epoxide
Dieldnn
Endnn
Diazinon
Methyl Parathion
Ethyl Porothion
Ma lath ion
No EthanoL
I
100
n
87
100
100
16
in
13
100
84
'Eluting mixtures:
Froct. I - 6% Et,O in pet ether
Froct.ll-15%
Froct. 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 4. A. a.
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.
-4-
-------
To verify the value obtained by the lauric acid method and to test for
-roper 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 7.A.). 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]:
-5-
-------
(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 7.A.) 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. Packing and Elution of Adsorbent Columns
Adsorbent is packed 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. The required
amount of column packing is added dry in increments with gentle tapping
to settle, followed by a layer of granular sodium sulfate (ca. 0.5 inches)
on top of the adsorbent. The column is prewashed with hexarie or petro-
leum ether, the level of liquid is brought to the top of the bed, sample
is added and washed into the bed with several small portions of the
first eluent, and various fractions are collected in separate containers.
Elution is carried out with a series of solvents and solvent mixtures
of increasing polarity. The polarity of the solvent series must be
selected in relation to the activity (polarity) of the adsorbent and
the polarity of the sample. The least polar solvent which will elute
the pesticides from the adsorbent should be used so that there is less
chance for polar impurities to be co-eluted.
The order of polarity for several common solvents is as follows:
Hexane (petroleum ether) - least polar
benzene
ethyl ether
-6-
-------
methylene chloride
ethyl acetate
acetonitrile
methanol - most polar
GAS CHROMATOGRAPHY PACKINGS
4.B. INTRODUCTION AND COLUMN TECHNOLOGY
At the risk of triteness, 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 opera-
ting problems will be discussed, many of which have come to light in the
interlaboratory quality control program described earlier in Section 2.
Some of the operational instructions, fully covered in Sections 4,A 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 chromato-
graphic 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. Upon injection of a mixture onto the column contained in a
chromatograph, each compound is swept through the column at a rate which
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, different compounds will migrate through the column at different
rates and separation will be achieved as diagrammed in Figure 4-B [9],
Figure 4-B- Schematic diagram for displacement analysis.
CHROMATOGRAM
COLUMN DETECTOR
SAMPLE U + B) H I'D
D I o IH * I M~1
- o i . lii-rnn
I " I4TI
& H ° i "rf~i
H~l
-7-
-------
Most applications of residue analysis are carried out with packed
columns of this type rather than wall-coated or support-coated capilla. y
(open tubular) columns, although the latter have been used, when mass
spectroscopy and gas chromatography are used in combination for identi-
fication or assay (Subsection 8.H.).
GC tubes are usually made of borosilicate glass. Teflon has also been
recommended, while copper and stainless steel are best avoided as both
can cause decomposition of compounds on the column unless special pre-
cautions are taken. Support materials are normally diatomaceous earth,
with firebrick, glass beads, and Teflon as other possibilities. A good
support material should be available in narrow and uniform ranges of
particle (mesh) size and have no active adsorption sites, high surface
area per unit volume, and good thermal stability and mechanical strength.
Finer particles produce columns of highest efficiency.
There are enormous numbers and types of liquid phases commercially
available, and the choice of a liquid phase is usually made on the basis
of the polarity of the compound to be separated. Phases recommended
for general use in pesticide analysis are described in Section 5.L.
Important column characteristics include efficiency and resolution
capability, sensitivity (in relation to the detector), retention,
compound elution pattern, stability against heat and injection loading,
and resistance to on-column compound decomposition. These will be
considered in light of their effect on the day-to-day operation of the
column.
4.C. EFFICIENCY AND PEAK RESOLUTION
Figure 4-C shows the equations used for calculating column efficiency
(in theoretical plates) and the resolution, or degree of separation
between peaks, from a chromatogram. Efficiency indicates the ability
of a column to provide good peak resolution, a value of at least 3,000
theoretical plates for a six foot column being considered acceptable.
Figure 4-D shows superimposed chromatograms of standard chlorinated
pesticide mixtures on two separate 6 foot 2% OV-1/3% QF-1 columns, one
with poor efficiency (A, 740 total plates) and the other with good
efficiency (B, 4530 plates) and the great difference in their inherent
peak resolution.
-8-
-------
Figure 4-C. Calculation of column efficiency and resolution.
Efficiency: N = 16(y)2
2z
Resolution: R = y+y
Figure 4-D. Effect of column efficiency on pesticide resolution.
-9-
-------
Factors which 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.
Figure 4-E. Effect of stationary phase loading on column efficiency.
3V. DC-200/73'AQF-l
4.D.
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 which increases the peak height for injection of a given
amount of pesticide will thereby increase detector response. The
columns recommended in this Manual (Subsection 5.L.) are designed for
adequate resolution and practical elution times, and an absolute retention
of 16-20 minutes for p,p'-DDT has been found to indicate these character-
istics. 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.
-10-
-------
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.
1.2
c
1 ]-°
"6
u
0 0-8
CO
0
"5
QC.
s.
&
\ M \
\ 1 i|
-U-L
r^
-p-p-H
^~^
tfifr
V
H
~H+
* ' I
is^
rrrbir
?T:-::H
^^
-::;::
= -.£':
rrfrr
_U- ^
3^~
^t
jjttr
~^
H
i
- ^
;:.
Figur* 4-F.
-H,-!4f
5vif
^3
* '
- - 1
- 1
~ "^_l. _^
-.-"
:::;:^:t:
t- -
- -_ r_
r^ipzi:
1
rzir:--
§3
Comparative rvcponst
8 GLC coluan* as r*l
to the r«f«r*nc« co]
of 3% OV-1, with «ac
coliuut op«r*t«d at :
optiaiM parajMtcri
3ix
:-^r
:H^^t
_ _,_!. - . .
. ., .
:::
:"
- -~:-
f-f-^
- ~
; ~^-
H i-t^
tra
:tv
-:
-^ t ^
::'T^!
' . " ' "*
- -
-
1
;
;
-1
~:
iiirr
of r
uan _
h
--
i^
' r± : ~
S - i - , .
:'.:'
:r:: '
---H H°-
~i - >
~-t~4-(-
z;!*
::-: «
Optimia P»r»*tt«r»
Carrier Flow
Loading* T«mp. ,°C. «l/»in
5 ISO 45
2/3 180 6O
3 ISO TO
3 180 60
1O 2OO 7O - 9O
4/6 2OO 7O - SO
1 . 5/1 . 95 2OO 60
5/7.5 2OO ISO
-DC-200"-:r4 :; -
.
:
v " - -
;
K
fc,
S
s
\
s
s,
s
)*
-r
-
3H :
OV17
~^r=r^nrr-.:
' _ -]. _|.
bp-1 -:- :
r:.: :
T
- ]
1
- . : | - -
4.E. COLUMN STABILITY
It is desirable to use columns which are heat-stable or "bleed" re-
sistant and which 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 5.C.) 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).
4.F. RESISTENCE TO ON-COLUMN COMPOUND DECOMPOSITION
Unless a column is properly prepared, conditioned, and maintained, it
can cause such compounds as endrin and/or p_,p_'-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 addi-
tional peaks arising from decomposition products. p_,p'-DDT decomposes
to p_,p_'-DDD and, in extreme cases, to p_,p_'-DDE.
-11-
-------
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 im-
proving response and minimizing conversion of endrin to breakdown pro-*
ducts is illustrated in Figure 4-G. Chromatogram A was obtained for an
aldrin-endrin mixture immediately after heat conditioning .and equili-
brating 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 quanti-
tated 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, arid Chromatogram B shows significant improvement in the
endrin response and complete disappearance of the two breakdown peaks.
Figure 4-G. Reduction in breakdown of endrin resulting from column
silylation.
A BEFORE SILYLATION
AFTER SILYLATION
10
if
ujO
20
-12-
-------
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-
grams was obtained in a laboratory where the injection insert had not
been changed for three weeks.
Figure 4-H. Breakdown of 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 £_/£_'-DDT to p_,p_'-DDD (in actuality, the ratio of these changed from
8:10 to 4:10). A clean Vykor glass insert was then installed in the
-13-
-------
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.
-14-
-------
4.G. 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,
-15-
-------
and solvent is removed by drawing air through the layer of packina 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
-16-
-------
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% QF-1:
SE-30 .040 x 21.0 = 0.84 grams
OF-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 302.12.
4.H. PACKING THE COLUMN
Columns for pesticide analysis are generally 4-7 feet in length and 1/8
-17-
-------
or 1/4 inch od metal or glass. Aluminum columns have been found suit-
able for chlorinated pesticides, but glass is usually preferred to pre-
vent 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, such as by hand vibration,
mechanical vibration, and vacuum. The method of choice may be dictated
by the configuration of the column. For example, vacuum is about the
only method for packing a coiled column. A U-shaped column may be
packed by any of the three methods. The recommended method is hand
vibrating, which has produced columns of consistently high quality:
a. The operator should be sure that the column, if intended as a 6-foot
column, is really 6 feet in total length, and not some lesser length.
Efficiency and retention time are both reduced in a shorter column.
For off-column injection in some chromatographs such as the MT-220,
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.
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 possi-
bility 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.
-18-
-------
If glass wool is packed by hand, the hands should be carefully washed
with soap or detergent, rinsed, and dried to minimize skin oil con-
tamination of the glass wool. Glass wool can be silanized by treating
with dimethyldichlorosilane in toluene for 10 minutes followed by
rinsing with toluene and treating for an additional 10 minutes with
anhydrous methanol, or the prepared material can be purchased com-
mercially (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 p 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.
4.1. COLUMN CONDITIONING
The column is conditioned, or made ready for routine use, by heat curing
silylation or Carbowax treatment, and injection of a concentrated pesti-
cide solution.
Heat curing of the prescribed GC columns (Subsection 5.L.) is carried
out according to the following schedule:
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
Shown for information only. Column not recommended for routine use.
2/
Carrier gas flow 60 to 70 ml per minute.
3/ Do not exceed this time period.
-19-
-------
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 con-
nection 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 de-
tector. 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.
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 5.L.). Four consecutive injec-
tions of 25 yl each of Silyl 8 (Pierce Chemical Co.) are made, spaced
30 minutes apart. Following the final injection, about 3 hours is
allowed for all traces of the silanizing material to elute from the
column. The syringe used for these injections should be used for no
other purpose and should be rinsed immediately after use to avoid damage.
The effects of silylation do not persist indefinitely, and repeat treat-
ment 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 photo-
metric detector since bleed will cause excessive fogging of the heat
shield. Details of the treatment and a special Swagelok assembly used
in the MT-220 chromatograph are given in Section 4,B,(2),IV of the EPA
PAM. This is a modification of the method reported by Ives and Giuffrida
[10]. The method involves vapor phase deposition of Carbowax caused
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.
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.
Following the silanizing injections or Carbowax treatment and with the
oven temperature and carrier gas flow rate adjusted to the approximate
-20-
-------
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, and PCP.
4.J. 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 5.A. and
5.B. 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 re-
checked after equilibration. Before making any injections, a background
(standing) current profile is run at the normal operating parameters
for the specific column being tested if an electron capture detector is
used. The polarizing voltage is set at its proper value. These opera-
tions are further discussed in Subsection 5.C. of this Manual.
A complex chlorinated pesticide mixture is now chromatographed to evalu-
ate efficiency, resolution, compound stability, and response characteris-
tics. 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 evalua-
tion (but not quantitation).
-21-
-------
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 5.A. 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 the 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.
4.K. 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 5.J.); 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
-22-
-------
must also be replaced if it becomes contaminated. Daily monitoring of
DDT breakdown is important for early indication of contamination of the
injection port and/or column. Improved cleanup of dirty extracts prior
to gas chromatography is an obvious aid in maintaining good column
performance.
The effects of 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 immedi-
ately 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 5.N.).
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
columns(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 recondi-
tioned if out of the instrument for more than a few days. A flow of
60 ml/minute carrier gas for several hours at a temperature ca. 25°C
above the prescribed operating temperature (venting into the oven) is
used for this operation.
Erratic and noisy baselines frequently indicate leaks in the column
connections or some other point in the flow system between the injection
port and the detector inlet. If the chromatograph oven can accomodate
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.
-23-
-------
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 5.C.).
b. Erratic Baselines
This phenomenon may be caused by a number of instrumental factors and
these will be treated in detail in Subsection 5,K. 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 grosslv inaccurate. These subjects will be discussed in Subsec-
tions 5. A. and 5.B. 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.
-24-
-------
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.
17mm
Actuol Oven Temp 201'C
Stated Carrier Flow 55 ml
Computed Efficiency 2,960 TP
23 mm
Actual Oven Temp 185'C
Stated Carrier Flow 61ml
Computed Efficiency 3,310 TP
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.
4.L. REFERENCES
[1] Tindle, R. C., and Stalling, D. L., Anal. Chem., 44, 1768 (1972).
[2] Brockmann, H., and Schodder, H., Chem. Ber., 74, 73 (1941).
[3] Bevenue, A., and Ogata, J. N., J. Chromatogr., 50, 142 (1972).
[4] Hall, E. T., J. Assoc. Offic. Anal Chem., 54, 1349 (1971).
[5] Mills, P. A., J. Assoc. 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. Contam. 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).
-25-
-------
[9] Figure from Dal Nogare, S., and Juvet, R. S., Jr., Gas-Liquid
Chromatography, p. 15, Interscience Publishers, N. Y., 1962.
[10] Ives, N. F., and Giuffrida, L., J. Assoc. Offie. Anal. Chem., 53,
973 (1970).
-26-
-------
Section 5
OPERATION OF THE GAS CHRONOGRAPH
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 instru-
mental 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 considered 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 (Tracor, Inc.) gas chromatograph 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 simul-
taneous installation of up to four different detectors.
5.A. TEMPERATURE SELECTION AND CONTROL
Proper adjustment of the column oven temperature and the carrier gas
flow rate (Subsection 5.B.) 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.
-------
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
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
- 2 -
-------
recommended. The instrument pyrometer must not be relied upon as the
only means of monitoring column temperature.
An injection port or transfer line at an excessive temperature may lead
to decomposition of heat-labile pesticides while a temperature lower
than optimum may cause incomplete sample volatization. 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
which will 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_,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 chromatogram.
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_,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 which 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
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_,p_'-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 illus-
tration of the effects of inaccurate column temperature on peak reso-
lution.
-------
Ul
dP
O
<
K>
H
O
>t>
,fc>
O
i(^
OJ
H1
*>.
H1
Ul
00
to
CO
00
oo
K>
OJ
en
Cfi
00
£>
U3
00
00
00
OJ
H1
-J
^ M
O Ul
*> OP
CO O
H <
1 1
OJ M
0 -J
\ \
CTi H
O >£i
dP Ul
rM
of
!O
^d O
i M
i ]ij
h-1 1
M
ip^ Ul
O Ul
Ji. -J
00 Ul
>£> 00
N) VO
00 Ul
00 K)
0 0
00 Ul
cn o
00
00 Ji.
rf^ cr>
OJ ifc.
OJ >ti
00 *>
o en
00 J^
M M
00 -J
OJ 4i
1 .
o o
Ul U3
tr1
H-
iQ
H-
a.
id
t?
01
in
CD
t1
^i
o
§
i-3
M
a
1-3
H
M O
w
o
n
H
i£>
O
VD
J^
t '
UD
00
NJ
O
M
h3
H
cn
0
»
w
w
i
1
8
S
H
P
a
w
H
2
IT"
H
Ul
I
- 4 -
-------
5.B. 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 10
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 of electron capture detectors. Each gas supply is fil-
tered through a filter-drier cartridge connected at the regulator out-
put of the cylinder. The filter may contain Linde 13X(1/16 inch)
molecular sieve pellets. 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 case when it is determined that
a contaminated tank of gas has been used. 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 H or ° Ni
electron capture detectors. If the baseline has been stable but
becomes erratic upon installation of a new column, a loose column con-
nection 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 is by spraying connections with Freon MS-180
with the instrument operating and observing any recorder response.
Short sprays are applied close to the connection, but not around the
injection port or the detector.
The carrier -gas flow velocities are checked using a soap bubble flow-
meter, which can be purchased commercially or easily constructed by
attaching a sidearm near the bottom of a 50 ml buret [Subsection 4,A,(6),
Figure 4(a), EPA PAM]. Since rotameters are installed ahead of the
columns, these cannot be relied upon when adjusting the carrier flow
or completely erroneous values may result. It is necessary to check
- 5 -
-------
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 at normal operating para-
meters 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 back-
ground current of an electron capture detector.
Carrier flow rates in excess of recommended values lead to lowered
absolute retention times and compressed chromatograms while rates which
are too low will have the opposite effect. Relative retention values
reflect only the operating temperature of the column (Subsection 5.A.)
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).
Figure 5-B. Effect of Flow Rate on GC Resolution
- 6 -
-------
GC Detectors
The reader is directed to reference [1] for a general review of the
e?ement selective pesticide detectors.
5.C. 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 rea-
gents. The detector consists of a radioactive source which emits low
energy 3-particles 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 col-
lector (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 de-
tectors such as the flame ionization detector. The magnitude of standing
current reduction, which depends upon both the number of electron cap-
turing species present and on their electronegativity, is measured on
the recorder and indicates the amount of material which captured the
electrons. After the component passes through the detector, the stand-
ing current recovers to the original value and a characteristic GC
peak is shown on the recorder, provided that the radioactive foil is
not overly contaminated.
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 g) amounts.V Advantage is taken of this sensitivity by preparing
halogenated derivatives of compounds (e.g., carbamate insecticides)
not normally well detected by EC. The response of EC detectors has been
studied and guidelines presented for predicting which derivatives
might best increase sensitivity [2].
jV Sensitivity of a GC detector is designated as that amount of pesticide
which will provide a peak whose height corresponds to some percentage of
full scale recorder deflection (usually 10 or 50 percent), while mini-
mum detectable amount is that quantity of pesticide giving a signal at
least four times background (noise) from the baseline.
- 7 -
-------
Sources of 6-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 temperature reducing the possibility of contamination of the
source from extract impurities or from bleeding of GC column liquid
phases and extending the number of compounds that can be detected and
the serviceability time of the detector.
The EC detector is used with either a constant negative DC voltage or
an intermittently pulsed voltage 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. When added as a make-up gas, introduced after the
column but prior to the ionization portion of the detector cell, nitro-
gen or helium can be used as the carrier gas and simultaneous dual
detector operation is possible. The pulsed and DC modes provide approxi-
mately equal sensitivity and linearity, but advantages have been claimed
for the former in terms of freedom from anomalous responses [3], repro-
ducibility of response, and 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 modulated) operation is a third EC detector mode.
A standing current is again achieved by applying voltage pulses, but
in this case the pulse sampling frequency is varied by a servo-mechan-
ism closed loop control circuit when an electron absorbing compound
enters the detector so that the standing current remains constant. Pulse
frequency is converted to a DC signal which is monitored in the usual
way. It is reported that this mode gives an increased linear range
without loss of detectability [4], but it has not been carefully eval-
uated for pesticide work.
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
63Ni cell is reputedly less than that of a tritium cell, but remains
relatively constant and may eventually equal or surpass the sensitivity
of a tritium cell after a period of use. Some compounds can show
increased sensitivity at the higher temperatures possible in the Ni
cell than in even a new tritium cell [3].
The linear range of the tritium detector has been traditionally three
to five times greater than for the ^-%i detector. Figure 5-C shows
typical linearity curves for £,p'-DDD using these detectors, both
operated in the DC mode. Notice that when a certain point in concen-
tration is reached, the linearity curve begins to "plateau". In
-------
performing quantitation, it is mandatory that pesticide concentrations
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 chromato-
graph, operation at an output attenuation of 10 x 8 or 16 on the E 2
electrometer or 102 x 8 or 16 on the SS electrometer will usually pre-
clude 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 response is avoided [see Subsection 5.0. on
quantitation].
Figure 5-C. Comparison of linear ranges of
p_,£'-DDD.
3H and bJNi detectors for
A linearizer has recently become available from Tracer, Inc. for the
6-%i detector operating in the pulsed mode. This modified °-%i detec-
tor extends the linear relationship between response and sample concen-
tration over a range of 1 to greater than 20,000 pg lindane (compared
to ca. 1-200 pg for the traditional detector) and allows operation with
argon-methane or nitrogen carrier gas.
The EPA analytical laboratories have used parallel-plate ^H EC detectors
exclusively because cleaning can be done in-house under an AEC permit.
Details are given in the FDA Pesticide Analytical Manual, Vol. I,
Section 311.12, for cleaning a tritium EC detector. Foils may be removed
- 9 -
-------
only by persons with an AEC license for this purpose. A cleaning solu-
tion 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. " Ni detector. The column
effluent entrance is shown on the left and the purge gas line, polari-
zing 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 ^3Ni 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 in the chroma-
tograph by injecting 100 yl 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.
Figure 5-D. Tracer, Inc. Ni EC Detector
Response of EC detectors depends upon temperature, gas flow, cell
dimensions, electrode positions, amount of radioactivity, and applied
potential. The adverse effects of even slight scoring on the EC col-
lector probe have been described [5]. Operating parameters must be
optimized for each manufacturer's detector. Electron capture detectors
containing tritiated scandium [6] and l^Pr gold foil [7] sources have
been described. The former can be operated at 325°C and has reportedly
three times the sensitivity of the usual tritium cell while the latter
is heat stable and lower in cost than 63IJi foils [3] . These detectors
- 10 -
-------
have not been evaluated in day-to-day use in the EPA laboratories.
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 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. In general, more significant information is obtained
by determining the background current at the normal operating parameters
for the column being used.
FIGURE 5-E. TYPICAL ELECTRON CAPTURE DETECTOR BACKGROUND
CURRENT PROFILES
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 3H detector or at 92 percent with the Ni detector.
- 11 -
-------
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 accomodate 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.
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.
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 OV silicones with low (1-5 per-
cent) phase loadings produce very favorable columns. The background
current determination is particularly important with a new column in
the instrument because background current that cannot be brought up to
the expected level indicates the probability of a bleeding column
requiring additional heat treatment to vaporize off the volatile
impurities.
Solvents and monomers can bleed from the septum and be swept through
the column into the detector. Glass inlet liners used for off-column
injection should be changed frequently. Proper handling of septums
and maintenance of the injection port are discussed in Sebsection 5.J.
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
- 12 -
-------
Chromatographer, Beware of Thy Detector!
We all know that the performance of the same types of GC
columns can vary with the quant-/ 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 colurnns are 6 ft x 4 mm ID glass U-tubes
packed with 10 wt % DC 200 on'a'silane treated support. Both
runs 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 endrm/dieldrin).
S o
1 5
j ! 5
r>
6
i 8
-* a
B U,
o o
o 2 r
LJ'U
< a
t* ft .£
e « t»
n I ^
b a 5
n ° i 5 *-
2 ? ? i g
0 Ul 0
1 60? a
1 c E i
Lfi ^ 1 ?
0 s =
.» | o
M A
Ul: A_
i i i i i
0 3 6 9 12 15 18 21
TIME (Minutes)
Figure* 1. Chromatogram of standard chlorinated pesticide mixture.
Column- 6 ft x 4. mm ID glass packed with 10% DC 200 on a
time-treated support. Column temperature: 200°C. Datector: Electron
capture at 1 x 10 8 AFS.
^ CN
X
0
o
" 1
Q
C
i, *
c
r*
^
_J
x
*N ,.
x X
S 1 E
i 0 *
2 a "O
< Ul *
LJ
? 1 »
J ^
i
?
pv
^
c ^
^ Q
U) O
9 -'
C *^
x a
* ?
1 =
UL-i
T i i r i i i
3 6 y 12 15 18 21
TIME (Minutes)
Figure 2 Chromatogram of standard chlorinated pestictde mixture.
Column: 6 ft x 4 mm ID glass packed with 10% DC 200 on a
silane-treatea uippcrt Column temperature. 200°C. Detector: Electron
capture at 1 x 10 a AFS.
You say that you would like a 6 ft column with the
efficiency shown in Figure 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 enctrin/dieldrin peak height ratio is good
(0.61). with signs of slight endnn decomposition, which is
normal. However, in Figure 2 the endrm/dieldrin ratio is only
0.31, indicating appreciably greater endnn decomposition, and
yet there are no signs of it in the Chromatogram. Something
appears to be radically wrong with the results in Figure 2.
Well, something is wrong. The column used in 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 DC voltage; i.e., any voltage can be 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
lifts below the knee leading to the plateau in the current vs.
voltage curve (see Figure 3). In our examples, this voltage
range is approximately 10 to 15 volts.
40-,
30-
UJ
oc
20-
O
C
o
O 10
10
20
VOLTAGE (Volts!
30
40
Figure 3. Plot of current v*. voltage! tor an EC detector.
At higher voltages, the response becomes non-linear and the
response-to-concentrdtion slope increases with increasing con
centration. This non linearity becomes extreme on the plateau
of the current vs voltage curve. Here, the response to
concentration slope is very small at low concentrations and
increases rapidly at high concentrations. This results in an
extreme contraction of the lower part of a GC peak and ar»
extension of the upper part of the peak When this occurs, a
- 13 -
-------
ehromatogram like the one in Figure 2 is produced. Figure 1
can be converted to an approximation of Figure 2. as shown in
Figure 4. A superficial baseline has been drawn which cuts out
the bottom part of the peaks. The similarity between Figures 2
nd 4 is obvious. If we extended the upper part of the peaks in
Figure 4, the chromatogram would resemble that in Figure 2
still more closely. This may appear extreme, but notice that in
Figure 2 we have lost not only the endnn decomposition
product, but also all the small impurity peaks that are seen in
Figure 1.
6 9 12 15
TIME (Minutes)
18 21
Figure 4. Same chromatogram as in Figure 1.
At voltages below the optimum range, the reverse occurs.
The response to concentration slope is high at low concentra-
tions and decreases with increasing concentration In this case,
the lower part of a peak will be extended and the upper part
contracted. This is observed as a wider peak, giving a low
theoretical plate calculation, and the peak maximum will tend
to be rounded. Also, 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 shift to lower
currents and voltages, as shown by the broken line curve in
Figure 3.
This problem of variable non-linear response with EC
detectors complicates quantitative analysis and is the reason
why frequent and careful use of calibration standards is so
important in pesticide analyses. However, when one is
interested in determining column efficiency, the effect of
non-linear detector response on the theoretical plate response
has not been so obvious. This effect can also be observed with
argon ionization detectors, where applied voltage also affects
linearity. Years ago we found we were consistently obtaining
about 150 more theoretical plates per foot from argon
ionization detectors at voltages above the optimum than from
flame ionization detectors, which have good linear response
characteristics.
To be fair to EC detectors, there are now one or more on
the market which operate at a fixed voltage and are claimed to
have good linear characteristics over a wide dynamic range.
Let it suffice to say: "Chromatographer, beware of thy
detector! Also, know thy detector1"
10 20
VOLTAGE (Volts)
Figure 5. Plot of peak height ratio of
haptachlor to dieldrin (h/d) vs. EC de-
tector voltage.
10 20
VOLTAGE (Volts)
30
Figure 6. Plot of peak height ratios of
endrin to dieldrin (e/d) and p,p'-DDT to
dieldrin (D/d) vs. EC detector voltage.
P 0.08 1
{!! 0.04 -
O
ui 0.02 -
x
10 20
VOLTAGE (Volts)
30
Figure 7. Plot of peak height ratio of
endrm decompostion product to dieldrin
(E/d) vs. EC detector voltage.
14-
13-
12-
11-
10-
7-
O
ui
I 5-
4-
10 20
VOLTAGE (Volts)
30
Figure 8. Plot of calculated theoretical
plates for dieldrin vs. EC detector volt-
- 14 -
-------
usually prevent this problem. These adsorbent traps must be regen-
erated regularly. If moisture has accumulated in the tubing and flow
controllers from a bad tank of gas, simply changing the tank may not
solve the problem. The entire system would have to be flushed out with
a low boiling solvent.
A good rule is to introduce only one variable at a time into the GC-
EC system and to run a background profile just before and just after
changing the variable. For example, a column and a tank of gas should
not both be changed at the same time. This rule makes the isolation
and correction of problems a much easier task.
A review of the operation and principles of the electron capture detec-
tor for pesticide analysis [8] and a review of its theory and charac-
teristics [3,9] have been published.
5.D. 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 yl are common with the
MC system, and this volume may represent 10-25 grams of original sample.
Cleanup procedures must accomodate 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 there-
fore be directly converted to a quantitative value by use of an equation,
without need for a calibration curve.
- 15 -
-------
Figure 5-F. Dohrmann Microcoulometric Detector
'"..", /J; « . .
mrtanf^ t ,
O > O
The various modes of detector operation are as follows:
Pesticide
Class
Products of
Furnace Reaction
Cell Reaction
Generation Reaction
Halogenated HX (oxidative, reduc- AG + X
tive, or catalytic)
AgX
Ag
Ag + e
Sulfur
SO (oxidative)
31
2e
Nitrogen NH (catalytic)
NH + H
-J
HN
H 2H + 2e
^-,
The coulometer and furnace systems require careful optimization to
obtain accurate and precise results. Selectivity depends on the speci-
ficity of the preliminary and coulometric reactions. For example, any
combustion 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. 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.
- 16 -
-------
Titration cells have two operational positions, termed position I and
position II. These differ in the gain setting, location of the elec-
trodes, and how the gas flow impinges on the electrodes. Position I
is operationally simple and provides dependable stoichiometric (coulo-
metric) response, but is relatively insensitive (ca. 100 ng). Position
II is ca. 20 times more sensitive (less than 10 ng), but response is not
stoichiometric (calibration is needed) and more careful operation is
required.
The response of the MC detector is linear, it operates over a wide
temperature range, and its sensitivity is not affected by carrier gas
flow rate, column bleed, or temperature changes. The detector is,
therefore, satisfactory for use with temperature programming. This is
not generally recommended in pesticide analysis, however, because of
poor precision of retention times from run to run. Temperature program-
ming with the MC detector is possible only under special conditions [10].
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 pesticides (e.g., carbofuran [11]) has predominated. Use
of the detector for phosphorus compounds has rarely been reported. The
detector 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 FPD and Coulson conductivity detectors,
manual venting of the solvent introduced during injection is required
unless special automatic venting equipment [13] is installed.
As with other selective detectors, the MC detector is useful for con-
firmation of residues tentatively identified with the EC detector.
Details of the operation of the MC detector as recommended by the U.S.
F.D.A. for Cl and S detection have been published [13].
5-E- ALKALI FLAME IONI7.ATION (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 contained different alkali metal salts, the response
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.
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 Na SO . Other salts such as KC1, KBr, Kb SO ,
- 17 -
-------
and K^SC^ 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. Sensitivities for organophosphorus pesticides are around
0.2-5 ng, compounds with short retention times and narrow peaks being
more sensitively detected. A typical linear range is 0.5-25 ng.
Minimum detection levels of triazine herbicides (N mode, T^SO^ tip)
were reported as 0.2-0.5 ng with a 920 selectivity vs. carbon compounds
but a poor selectivity vs. phosphorus [14]. A detector with a KC1-
Rb SO (1:1) tip was used to determine ca. 30 ng of carbamate herbicides
[15].
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 alteration 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.
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 (compared to the conventional
FID), while the increase for N is 100 and As 30 [16] . Selectivity for
P with respect to hydrocarbons 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.
The alkali metals Na, K, Rb, and Cs increase selectivity for P about
equally. The same is true for N response except for Na, which is not
so great. Rb is usually used, however, for detecting N compounds.
Halogen response can be depressed by KC1 and KBr, but Na salts increase
response to Cl and Br compounds. Response 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 [16].
A Varian Aerograph Rb2S04 detector modified to allow increased carrier
gas flow up to 60 ml/minute and injection volumes up to 7 yl has been
described in detail [17]. With this detector, greater than 1/2 fsd
- 18 -
-------
was achieved upon injection of 0.2-0.5 ng of six common organophos-
phorus insecticides. Details for preparation and use of a KC1 thermionic
detector providing 1/2 fsd for one ng parathion have also been given [18].
The stability and sensitivity of a RbCl nitrogen thermionic detector
was carefully compared with a conventional FID for the pesticide feni-
trothion [19]. A new design detector with a glass bead containing non-
volatile rubidium silicate at a negative potential as the thermionic
source can be operated as a specific nitrogen plus phosphorus detector
or for phosphorus detection alone. This detector has been preliminari-
ly tested using standard malathion but has not been evaluated for routine
residue analysis [20].
In summary, the thermionic detector is extremely simple, quite inexpensive,
and very useful. Its optimum, reproducible use, however, is more of an
art than science, and it is recommended mainly for experienced gas
chromatographers. Reviews of the thermionic detector have been pub-
lished [21, 22].
5.F. FLAME PHOTOMETRIC DETECTOR
This detector operates by monitoring HPO and 82 emission bands, which
result from burning the column effluent in a 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, Cr) have also
been made with limits of ca. 10 - 10 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 cir-
cular 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 phos-
phorus output from a single injection, as well as normal flame ioni-
zation output 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 the FPD mounted to the MT-220
gas chromatograph.
- 19 -
-------
Figure 5-G. Tracer flame photometric detector.
Figure 5-H. Cross- section of a flame photometric detector.
Photomultiplier tube
Glass window
Swageiock
fitt rig
olumn effluent
(N2)
- 20 -
-------
Figure 5-1. Dual flame photometric detector.
The details of detector operation are as follows: oxygen 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 transmits a specific wavelength
of the element to be monitored. A potential is applied to the PM tube
and its output is amplified by the electrometer and read-out on a
recorder.
Solvent in the injected sample will extinguish the flame unless modifi-
cation 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 been shown
to give a "hyperventilated" flame [23] which allows injection of up
to 25 yl of solvent with no flame blowout and reportedly also better
signal/noise ratio, baseline stability, and linearity but an approxi-
mate 20 fold loss in selectivity for most detectors. Carbowax-treated
GC columns (Subsection 4.1) sometimes are advised with the FPD, depending
on the column support used.
The minimum detectable quantities of elements S and P are about 200 pg
and 40 pg, respectively. In routine operation, 2.5 ng of ethyl para-
thion 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.
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 H , O ,
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
- 21 -
-------
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 as follows:
Temperatures (°C) Flow Rates (ml/min.)
Column
Injection Block
Detector*
Transfer Line
Switching Valve**
200
210-225
165-250
235
235-240
Purge**
Carrier
Hydrogen
Oxygen
Air
70-80
70-80
150-200
10-30***
0-100
* High temperature model is never heated above 250°C, low temperature
model never above 170°C
** With Valco switching valve
*** To ignite the flame, an oxygen flow of 80 ml/minute or more may
be required, depending on the detector
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.
- 22 -
-------
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 require different
voltages for best performance, a value of about 850 V being typical.
A suspect PM tube may be checked with one of known sensitivity to give
indication of its condition. Satisfactory operation of the FPD over
its full dynamic range requires both a highly stabilized 750 V power
supply plus an electrometer with a bucking capability of at least
1 x 10~° amps.
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 voltage is applied.
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 re-
placed 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 with the P mode. Sulfur
response for compounds containing a single S atom increases approximately
as the square of the concentration, the S mode is inherently less
sensitive than the P mode. 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
for P and S are about 1,000-10,000:1 compared to hydrocarbons and
halogenated hydrocarbons. Large amounts of sulfur impurities give a
response in the P mode (P:S response ratio 4:1 at 526 nm) whereas
- 23 -
-------
phosphorus impurities cause negligible response in the S mode (S:P
response ratio 100:1 at 394 nm). As the degree of sulfur oxidation
in the molecule increases, there is usually a decrease in sulfur
response.
Maximum utility of the FPD is afforded by a 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 is important
information for distinguishing between PS, PS2» and PSo compounds for
confirmation of residue identity. The atomic ratio of P:S in a molecule
is the P-response divided by the V S-response. As mentioned earlier,
dual operation will not be practical for analysis of low amounts of
residue barely detectable by the P mode of an FPD optimized for this
mode because of the much lower sensitivity of the S mode under these
conditions.
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 thermionic detector for routine analysis
in terms of 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.
5.G. 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 contri-
bute 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.
- 24 -
-------
Figure 5-J. Tracer (Coulson) electro-
lytic conductivity
detector
Figure 5-K.
Electrolytic
conductivity cell
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. The best
selectivity and sensitivity have generally been obtained for N compounds,
and detection of these has been the major use of the detector up to now.
A comprehensive study of 95 organonitrogen pesticides with 5 and 10%
DC-200 columns gave a range of 1-1000 ng for 1/2 fsd [24], If the Sr(OH)2
pledget and the catalyst are removed, the N mode will detect organo-
chlorine 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 [25]. The S mode of the conductivity
detector is of comparable sensitivity but less selective than the S mode
of the FPD. The conductivity detector is less complex than the MC
- 25 -
-------
I
w
co
i
§
H
B
U
EH
H
>
H
EH
§
§
U
CN
in
H
^1
4->
rl
H
H
01
d
CD
CO
>i
rl
H
4->
u
CU
H
CU
CO
01
d
o
-H
H
d
o
u
0)
0
S
^
U3
£j
£2
-P
01
§
01
tn
d
o
o
CM
1
O
o
r-H
,»
2
tn
ft
0
0
rH
^
X!
3
SH
U
tn
0)
I
SH
l
U
0
O
in
co
i
01
tn
£j
o
-rl
4-1
u
a
CU
1
id
o
ty
d
o
00
1
0
m
^
01
CU
d
H
N
id
SH
-P
tn
d
in
rH
i
r-
0
H
T)
H
0
id
tn
CU
>
g
CU
a
0)
0
(V
PH
J2J
|
4-1
0)
>i
rH
id
4-1
id
u
CU
H
CN-H
g;
3
.,
to
0) .
-P d
id 0
g r^
id -H
X! N
SH id
id -H
u -O
rH tn
0 d
rl
XI "1
4-1 CN
tn -
d d
O
o -H
in ,G
1 4-1
in id
M SH
id
- ft
01
cu tn
-P d
id
g 0
a in
X! rH
01 01
d d
O 3
ft 0
in ft
Q) g
S-) O
u
1
2
01
4J 0
U 4->
O >i
O H
ft 0
SH
cu
l
_l_
"d1
K
2
01
id w
id
cu
-P SH
O cu
CU -H
-p SH
cu SH
13 id
o
C
i3 <;
.
to
cu
T3
H
U
rl
4-1
0
CU
01
d
rl
rH
U
tn
d
0
i
vO
*.
rH
U
D1
QJ
0
o
in
1 1
T3 O
d cu
id co
H -H
u
o o
P O
01 .
*O tn
d 4-1
o d
ft 0
01 u
cu i
o
tn o
rH
tn i
0
13
d a)
o u
H 3
u o
3 SH
Ti ft
tt)
O
CU W
S i
p
01
>i id
-P C7*
ft
g CN
CU ffi
r~H
u
d
0
d
rl
id
H
T3
in
IB
X!
u
0
w
w
ffl
tn
^|
cu
H
^-1
^
iB
U
CM
|
U
o
m
00
o
(U
j>
id
tn
tn
-O S
d O
O 4H
ft
g CN
0 0
0
d
2 0
tn tn
d d
-H
0 T3
o d
CN CU
CN ft
1 CU
O 'D
ro
^
* i3
co tn
MH
tn
d CN
rH rH
*.
H
U
CU
CO -H
O
4-> tn
<
tn
T) 1
Is
QJ C
O
CO
0 -H
4-1
^ id
Q) CU
N SH
H 4-1
ti 01
H
X CM
0 O
01 d
T3 -H
S (V)
0 O
ft co
g
0 'O
? §
CO
co
01
cu
w
id
£
d
H
J^
0)
XI
}*4
Q
o
a
0)
-p
^_l
cu
d
o
o
rH
u
|
cu
g
-p
^1
-p
i4
cu
(~J
4-)
*
Ss,
4->
H
>
H
4->
0)
t |
CU
4-1
ft
1
01
id
tn
^|
cu
H
JH
^
iB
U
CN
2
l
rH
U
K
i M d
1 -rl CU CU
H U XI i >, o
ft CU -rl -H T3
M > rH SH CU
CU 4-J *H *H O in
Q)
01
£3
<1)
X!
>i
id
g
4J
W
>l
i-H
id
ti
u
Q)
3
tn
^J
id -O
m c
d S
rl
rH CO
.. tn
2 d
tn o
d o
H
H 1
xO
.
rH »
CJ CO
SH tji
o d
N^
Ti O
d o
IB 0
1 1
01 1
CM
tn
ft Q)
tn
o d
o io
in rl
co
«.
r-H
U
ss
o
01
o
d
o
ft
01
0)
in
rH
01 U
>1 °
rH O
o m
SH oo
^t
ft M
O
CU
P 0
CD
^i ^.D
P i
ft 0
g o
CU "J1
u
rl
-p
rH
O
tn
0)
a
rl
U
rl
4J
to
CU
ft
rH
O
O
01
01
0)
w
IB
O
I
- 26 -
-------
detector and requires less maintenance, but reports differ on which
requires cleaner extracts. The conductivity detector is much more
s"1active than the electron capture detector, and analyses can be per-
formed without cleanup using the former, whereas cleanup would be
required for the latter (e.g., for determination of triazine herbicides
[26, 27]).
Variables affecting the detection of nitrogen pesticides have been
studied [28]. 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 reduced sensitivity resulted from adsorption of NH, 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 con-
ductivity and background electronic noise. Besides the GC and detec-
tor 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:
Conductivity bridge - Voltage 30
- Attenuation 1
Pyrolysis furnace temperature - 800°C
Vent block temperature - 220°C
Gas flows - helium 80 ml/min.
- hydrogen 80 ml/min.
A study of the effect of operating parameters on the response of tria-
zine herbicides [29] 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.
b. Modified Conductivity Detectors
Using the reductive mode of operation, Hall [30] 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
- 27 -
-------
PCB's. 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 elimi-
nate need for solvent venting. Response for Cl, N, and S is linear
over a range of greater than 10 , and peak broadening is greatly mini-
mized. Selectivity vs. carbon is about 10^ [31]. Chromatograms pro-
duced by Hall [32] illustrate the detection (15-90% 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.
Figure 5-L. Tracer Model 310 Hall electrolytic conductivity detector.
- 28 -
-------
A second modified detector was recently described by Lawrence and
Moore [33] with a new, compact cell and optimized flow conditions and
a water jacket for temperature control. A five-fold increase in sensi-
tivity was achieved over the conventional detector with water flow and
temperature of 1.2 ml/minute and less than 20°C, respectively.
5.H. OTHER DETECTORS
Other detectors have been employed on occasion for pesticide residue
analysis, but none is as prominent as those covered in Subsections 5.C
through 5.G. Bowman, Beroza, and co-workers [34-36] positioned a copper
screen or pellet of Na SO. 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 charac-
teristic wavelengths in a manner similar to the FPD. Detectability 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 simulta-
neous and selective detection of P-, S-, and Cl-containing pesticides
eluted from a GC column [37] . 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 resulting emission is
monitored at 360 nm.
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 [38].
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 flameless
ionization (chemi-ionization) detector detects the ionization produced
when phosphorus compounds pass from the column into a heated atmos-
phere of CsBr vapor and inert gas. The latter was found less sensitive
than the thermionic detector but quite selective [39].
- 29 -
-------
5.1. ELECTROMETER AND RECORDER
The electrometer is primarily a device for amplifying the electrical
signal from the detector prior to its introduction to the recorder.
Units may be single channel, designed to operate with one detector
and recorder, or dual channel. Customary controls on the electrometer
include input and output attenuators, output polarity, and controls for
the recorder zero and bucking current. Servicing of electrometers
is generally a function of a trained electronic technician or representa-
tive of the company manufacturing the chromatograph.
To check electrometers on the Tracer MT-220 chromatograph, set attenua-
tors to the off position and zero the recorder. Set the output atten-
uator at xl and record the baseline. A steady baseline with less than
1 percent noise should be obtained.
Recorders may be of single or dual channel design, the latter being
capable of receiving two separate voltage signals supplied from the
electrometer to two pens 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 re.sponse 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.
5.J. SAMPLE INJECTION AND THE INJECTION PORT
a. On-Column and Off-Column Injection
Some gas chromatographs have injection ports designed to accomodate
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
-30 -
-------
change the glass wool inlet plug. The frequency of change is deter-
mined by daily monitoring of the extent of p_,p_'-DDT conversion (see
4.F. 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_,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 4.F. Glass injection sleeves are cleaned in chromic acid
cleaning solution, rinsed with water and acetone, and stored in a 130°C
oven until use.
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. Septums
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.
Bleed from seven types of freshly installed septums has been reported
[40] 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. 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 forcepts. 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.
- 31 -
-------
c. Injection Techniques
(1) Handling the Syringe
When a sample is injected into the chroma to graph, it is essential it be
entirely vaporized without loss. Injections are usually made using a 10
yl syringe for the electron capture detector or a larger ceipacity if
required for other detectors. Automatic injection devices are available
for use with some chromatographs and detectors, but their use in pesti-
cide analysis has not yet been extensive.
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 yl more than necessary
into the barrel. Hold the syringe vertically with the needle pointing
up, put the needle through a tissue to absorb expelled liquid, and push
the plunger until it reads the desired value. The excess air should
now have been expelled. There should be no delay between filling and
injection of the sample. After injection, the syringe is rinsed clean
by filling with and expelling 5 portions of ethyl acetate or acetone,
and the syringe is pre-rinsed with the next sample to be injected in
the same manner. Be sure to follow carefully all manufacturer's sugges-
tions for proper use of each particular syringe.
When a sample is injected in this normal manner from a 10 yi syringe,
the needle will retain ca. 0.2-0.3 yl of sample. It is usually safe
to ignore this volume since standards are injected for comparison and
the errors due to retained volumes will cancel out if equal volumes are
used and concentration differences are negligible. Alternatively, the
syringe may be filled by drawing the entire sample into the barrel,
noting the final volume by reading each end of the column of liquid.
After injection, the plunger is pulled back and the small volume of
retained solvent now in the syringe barrel is read and applied as a
correction. This will correct for nonreproducible injection technique
but not, however, for the error encountered if the retained volume has
a composition different from the original sample, as would happen if
nonuniform distillation had occurred in the needle. Then the remaining
liquid would be richer in high boiling sample components.
This can be overcome by using the solvent flush injection technique, the
most reliable and reproducible method available. About 2 yl of solvent
is first drawn into the syringe followed by a 1-2 yl air pocket and then
the required volume of sample. The sample is brought completely into
the barrel so its volume can be read. On injection, the flush solvent
- 32 -
-------
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 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.
(2) Preferred Volume Range
Injection of 1-3 yl 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 yl in a 1.0 yl in-
jection would produce a ,'Q x 100 = 20% relative error, while the same
0.2 yl error in a 5 yl injection volume reduces the relative error to
a tolerable 4 percent level. The analyst is strongly urged to inject
volumes between 5 and 8 yl from a 10 yl syringe for analyses with
electron capture detectors. A syringe filled close to capacity is more
difficult to manipulate. Proper dilution of standards and samples will
provide on-scale peaks upon injection of optimum volumes. Standard and
sample solutions are prepared so that peaks of approximately the same
area are produced (Subsection 5.O.). A capable analyst should be able
to reproduce a series of 5-8 yl injections to within 1-3 percent of
average peak area or height when response is ca. 1/2 fsd with use of
proper techniques.
5.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 O-rings
should be used. Satisfactory results have been obtained with heat-
resistant silicone rubber Viton column O-rings used with brass ferrules.
Even though the Swagelok nuts are tightened securely on a cold column,
an overnight period at normal operating temperature may result in suf-
ficient loosening to cause a leak. When fresh O-rings are installed,
it is good practice to open up the oven after the first overnight heating
period and retighten the nuts.
- 33 -
-------
The new commercially available graphite ferrules seem ideally suited for
sealing glass columns to injectors and detectors. They can withstand
temperatures above 400°C, do not adhere to instrument parts, and are
reuseable. A Swagelok brass ferrule is installed backwards behind the
graphite ferrule and ahead of the nut. 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. No further tightening of this
graphite unit should be required regardless of temperature since cold
flow is absent in the graphite material, but this should be verified by
checking. 'Teflon ferrules, also used without 0-rings, are commercially
available but not as preferable as graphite.
5.L. 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-polar methyl silicone
columns are of limited value to the analyst determining the more common
chlorinated insecticides in environmental or animal samples. Partially
or completely overlapping peaks are obtained for several pesticides
generally detected in these sample substrates, e.g., p_,p_'-DDE and
dieldrin, o_,p_'-DDT and p_,p_'-DDD, and the isomers of BHC (Figure 5-M,A) .
(2) A high efficiency column is desirable if injected extracts contain
extraneous materials and detection of low pesticide concentrations is
required. This will provide sharper separation between the peaks of
interest and extraneous peaks of biological origin.
(3) Retention time or elution speed may be a primary consideration if
the analyst is concerned only with quantitation of certain, specific
pesticides. For example, in a project where the sole interest is to
routinely determine residues of a late-eluting pesticide such as methoxy-
chlor, the column selection and operating parameters would be tailored
to elute methoxychlor in a minimal time period consistent with its se-
paration 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 5.N.).
- 34 -
-------
(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_,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 p_,p_'-DDE on an efficient column.
The single phase OV-210 gives a full separation of the common BHC isomers
but only fair separation between the compound pairs of heptachlor,
epoxide/p_,p_'-DDE and p_,p_'-DDD/p_,p_'-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 (6-BHC
after o_,p_'-DDT and p_,p_' -DDT before p_,p'-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.
Chronatograms 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 V, 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 [41], [50].
"_/ equivalent to OV-210
- 35 -
-------
n
F
in
W
CO
H
CO
>H
J
S
S
w
Q
H
U
H
EH
CO
W
ft
rt
g
CO
w
co
<
ffi
ft
Q
H
D
O<
H
iJ
O
g
1 1
a
§
S
0
o
w
K
co
0
H
Q
*
m
0
rH
CN
£
O
*
in
\.
« 1
\ 0
0 rH
m CN
i >
co O
dP *
31 \D
\.
* 1
O
\ H
r- CM
rH 1
£> 0
K* V '
O
dP
dp in
in en
rH rH
CU
W
(d
£1
ft
ft .C
(0
g CO
0 g
rl
,C O
U 0 O
i H en
co \ in i
(do en o
O co H r-
ft
X -
Oi
S g £1
O W
X! M (U
tl ^ g
O U
W 1 O
ecu o
(d CN O
O CJ \ ii)
MO O 1
£1 M 0 O in
U O rH CN "31
ft
as" >
Oi
s g
0 £
X! M 0)
M £, (U
O U g
CO 1
o w o
g (d o o
o o rH ft rH
W SC EH < g
^J
M
O
ft
ft
3
W
T3
0)
N
H
c
13
i-H
H
m
G
O
tn
G
rl
p
rd
O
U
0
-P
rH
o
H
M
ft
a
cu
X
rl
g
0)
M
ft
w
0)
tn
(d
X!
ft
a
H
r-«
P
W
rl
rH
rH
rd
P
TJ
rl
>
H
'O
c
H
\.
* 1
- 36 -
-------
Figure 5-M. Peak elation patterns of 13 pesticides on five columns.
10% DC-200
SV. OV-210
3% PEGS
45bSE-30/6?bOV-210
m.CV-17/1.95-/.Qf-J
- 37 -
-------
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 OV 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 ABS, Supelcoport, and Diatoport S. In a
particular analytical situation, any one of these or some other column
might possibly be equal or even superior to one of those recommended
previously.
The U.S.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 concen-
tration 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.
The Canadian Department of National Health and Welfare [42] 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), OV-225
(medium polar), ethylene glycol adipate (polar), and DECS (very polar).
The relative polarities were calculated from McReynolds constants [43].
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 [44] for over 100 pesticides on OV-1,
OV-210 (intermediate polarity), DECS, and mixed phase columns.
Other extensive compilations of relative retention data appear in
References [45, 46, 47].
- 38 -
-------
5.M. 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 5.L.)
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 Mode), 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 sensitivity
(and concurrently in column efficiency) can be achieved by simply lowering
the carrier gas flow rate.
A sample extract volume of 10 ml from a 5 gram sample contains the tissue
equivalent of 0.5 mg/yl. A 5 yl injection of this extract (2.5 mg of
sample) into an electron capture detector should easily produce quanti-
fiable peaks at pesticide concentrations of at least 0.1 ppm provided
sensitivity is adequate and attenuation is appropriately adjusted. The
high sensitivity capability of the chromatograph should be utilized by
optimization of parameters to permit operation at low output attenuation.
It is poor practice to operate the electrometer at high attenuations
(10 x 32 or 10 x 64 on the Tracer MT-220) while adjusting standard and
sample concentrations to fit this attenuation range. With a new detector
foil, high attenuation may be necessary, but in general this practice,
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
which is particularly important when injecting the 15 percent ether-
petroleum ether Florisil column eluate from a fat sample (Subsection
7.A.a.). If the instrument is functioning properly, it should be pos-
sible to have a noise level not exceeding 2 percent full scale at a
low signal attenuation (10 x 8 or 10 x 16).
- 39 -
-------
5-N. 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^
p,p'-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 5.O. 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.
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
- 40 -
-------
column has separated all pesticides present in an unknown mixture, and
'_f 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 DEGS.
Elgar [48] 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. Con-
versely, when two dissimilar columns are used, the plotted points show
a wide scatter, enhancing the probability of reliable identification.
Figure 5-N shows the plots of three column pairs for 17 pesticidal com-
pounds 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% DEGS, and C is 5% OV-210
against 1.5% OV-17/1.95% QF-1. It will be observed that the KRT 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 identifi-
cation 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 5.L., 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 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.
5.O. 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 of the peak for each pesticide in the
sample and the size of a peak from a similar, known amount of each com-
pound 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.
- 41 -
-------
10 % DC-200 (200°c.)
STSp'j'?''-0 ° 7 ? * f * ~
CD
Ui
0
ft
cn
0
HI
H
fD
rt
(P
3
rt
H-
§
H
P)
ft
H-
o
M
n>
M
ft)
rf
H-
rt
O
5%OV-2K>(180°C.)
10 % DC-200 (200°c >
,in
-0.0-
2,3"
jf
-f
II1
^ a
O
Hi
O
CD
01
rt
H-
O
H-
O
I
O
S
(D
(D
O
O
0)
H-
4
(0
- 42 -
-------
The exploratory chromatogram of the sample extract used to obtain rela-
tive retention data will provide a tentative indication to the analyst^
of the proper standard mixture to be used. This mixture should con^n
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. In-
jection 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 facili-
tate the analysis (Subsection S.O.f.).
b. Comparison of External and Internal Standardization
Internal standardization is widely used, general analytical and gas
chromatographic technique which, however, is not recommended for multi-
residue pesticide determinations. Since multiresidue methods can detect
and measure a large number of different compounds, choice of a suitable
standard with appropriate structural and chromatographic properties in
terms of all compounds to be quantitated would be an impossible, or at
least 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 chromato-
graphy at several dilutions to quantitate all residues, so different
quantities of internal standard would be required. Detector response to
sample coextractices further complicates the choice of an internal stan-
dard. 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
depends upon the ability to inject exact amounts of samples and standards
reproducibly and having all instrumental parameters under tight control
so that data is comparable from run to run.
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:
- 43 -
-------
abe
R =
cd
where a = nanograms of pesticide represented by the standard peak
b = height (or 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:
- ml of extracting solvent x volume of final extract*
aliquot taken of original extract (ml) x yl injected
* This value is in yl 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 (yl or ml)
yl 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 yl)
The equation can be simplified to match these parameters and would read
R- abf x 0.12
c
where f = volume of final extract in yl or ml. The numerical factor
0.12 represents e for the specific illustration above.
- 44 -
-------
Units frequently used in pesticide analyses are the following:
-6
yg = 10 grams
-9
ng = 10 grams
-12
pg = 10 grams
ml = 10~ liters
-6
yl = 10 liters
ppm = parts per million = yg/g, yg/ml, ng/mg, or pg/yg
ppb = parts per billion = ng/g, ng/ml, or pg/mg
d. Detector Linearity
Linearity may be defined as the range of concentration over which a
detector maintains a constant sensitivity. If a detector has a linear-
ity of 10-3 an^ the minimum detectable quantity (MDQ) of a certain pesti-
cide is 1 pg, the upper limit for analysis is 1 ng. If the MDQ is 0.1 pg,
the pesticide can be determined only up to 100 pg. MDQ 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 detector, the linear
range varies somewhat between pesticides. For example, the isomers of
BHC exhibit a more restricted EC linear range than p_,p_'-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'-ODD with these two detectors.
Before any attempt is made to try quantitation with a new or newly reno-
vated 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 concen-
tration 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.
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).
- 45 -
-------
FIGURE 5-O QUANTITATION BY PEAK HEIGHT
C METHOD
Peak Height =CD
e. 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 pesticide
concentration within .linear boundaries.
Electron capture electrometer attenuation should be adjusted to obtain
a minimum sensitivity level equivalent to a 50% fsd peak from the
injection of 100 pg aldrin.
There is no objection to using different instrumental attenuation set-
tings for standards and samples provided that concentrations are within
the linear detection range and checks are made to insure that the
- 46 -
-------
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 Tracor
MT-220 chromatograph electrometer is convenient to assure operation
within the range of the detector.
f. Injection Volumes and Standards
As described in Subsection 5.J., injection of small volumes such as 1-3
yl can lead to large relative errors and should be avoided. A common
reason for low injection volumes is to provide on-scale peaks of sample
for reference against peaks from a particular single standard or standard
mixture. To circumvent this problem, standards should be made up at
several concentrations, each succeeding level being twice -the previous
concentration. For example, a typical set of three standard mixtures in
pg/yl for electron capture GC might be:
Mixture A,
Mixture A,
Mixture A.
Lindane
Aldrin
Dieldrin
o^p'-DDT
p,p'-DDT
5
5
10
10
20
10
10
20
20
40
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.
It should be unnecessary to reiterate that accuracy of analysis is
limited by the accuracy of the standard quantitating solutions. Consis-
tently 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.
- 47 -
-------
g. 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 para-
meters are optimally set, the 10 or 20 percent fsd minimum peak require-
ment 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 wnen the
concept of different standard concentrations (Subsection S.O.f.) 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 attemp-
ting 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 \..1, 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.
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 siae 1.67 grams;
injection volume 5 yl.
Standard
0.2 ng
Final
Extract
Volume
4000 U\ 400
Samp. InjecM: p.n.b - 0.2x13x4000 , o .
130x5 * "-a- v.ou
Samp. Inject. 1- ff.b.- 0.2x16x4000 vn, ., so
130x5 X0.6-II.no
Deviation 2.20
or 23% error j
Samp. Inject. 1 : p.pi. = 0.2 x 130x400 _
130x5 * u-° - *'*"
Samp. Inject. 2: D.p.b.- 0.2x133x400....
130X5 * u-° - 9-»'_
Deviation .
or 2.2% error
- 48 -
-------
h. 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 un-
known sample would be bracketed between standard injections made immedi-
ately before and after the sample.
i. 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 tall, symmetrical, fairly narrow peaks which have no
obscuring overlaps. These are chracteristic 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, sym-
metrical, and fairly wide peaks such as produced by mirex, methoxychlor,
Guthion, etc. 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 and in all measurements.
When peak heights are used, the assumption is necessarily made that
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 con-
trol, the second peak may instead elute later or earlier than the first,
resulting in a wider or narrower peak. However, the peak areas should
be the same in both cases. For this reason, peak area or peak height
x retention time is considered by many operators to be more reliable than
poak height alone since slightly shifting peak positions will not be so
important.
- 49 -
-------
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-O are measureable by the peak
height method because their overlap does not obscure the height of either
peak. Peaks on a sloping baseline but too narrow to be triangulated can
be measured by the peak height method. The distance KI would be used as
seen in Figure 5-R. Peaks can be widened by using a faster recorder
chart speed. Use of the planimeter is an alternate method for measuring
unsymmetrical peaks or peaks on a sloping baseline. Precision may be
improved by tracing the peak several times and taking an average value.
FIGURE 5-Q QUANTITATION BY PEAK AREA METHOD
FIGURE 5-R QUANTITATION BY TRIANGULATION
METHOD
Peak Area = CD*AB
Quantitation of peaks indicating heavy electrical overshoot (Figure 5-S,A)
or nonlinear response (5-S,B) will lead to unreliable quantitation. Peak
overshoot is influenced by foil contamination and by improper EC detector
polarizing voltage (Subsection S.C.b.).
- 50 -
-------
FIGURE 5-S EXAMPLES OF GAS CHROMATOGRAPHIC
PEAKS
A.
Distorted
Peaks
Non-linear
Indication
Linear
Indication
Automatic (disc) integration is a convenient procedure which can be used
in place of manual procedures whenever baselines are steady. This method
should be used for late eluting peaks only.
Gaul [49] compared five methods for quantitation of aldrin, heptachlor
epoxide, and dieldrin with a tritium electron capture gas chromatograph.
The methods were disc integration, triangulation, peak height times width
at half height, retention time multiplied by peak height, and peak height.
No significant differences were found among the five methods in this study.
The same author described methods for properly placing baselines for
typical overlapping and unsyiranetrical gas chromatographic peaks, and sug-
gested procedures for quantitating multipeak chromatograms of pesticides
which are mixtures of isomers, e.g., DDT, BHC, chlordane, and toxaphene.
j. Automation
Digital computer systems are available today which perform peak and base-
line detection, area integration, baseline correction, area allocation
of fused peaks, and postrun calculations. These systems are relatively
expensive, but their increased use in the future is certain as tech-
nology is simplified and prices decrease, especially in laboratories
with many chromatographs and a heavy sample load.
- 51 -
-------
5.P. References
[1] Natusch, D. F. S. , and Thorpe, T. M. , Anal. Chem., 45,, 1184A (1973)
*t «v
[2] Sullivan, J. J., J. Chromatogr., 87, 9 (1973).
[3] Burgett, C.A., Research/Development, 25, 28 (1974).
[4] Maggs, R. J., Joynes, P. L., and Lovelock, J. E., Anal. Chem., 43,
1966 (1971).
[5] Holden, E. R., and Hill, K. R. , J. Assoc. Offic. Anal.. Chem., 57,
1217 (1974).
[6] Hartmann, C. H., Anal. Chem., 45, 733 (1973).
[7] Lubkowitz, J. A., Montoloy, D., and Parker, W. C., J. Chromatogr.,
76, 21 (1973).
[8] Aue, W. A., and Kapila, S. , J. Chromatog. Sci., 11, 255 (1973).
[9] Pellizzari, E. D., J. Chromatogr., 98, 323 (1974).
[10] McCullough, P. R., and Aue, W. A., J. Chromatogr., 82, 269 (1973).
[11] 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.
[12] Karlhuber, B., Ramsteiner, K., Hermann, W. D., and Simon, W.,
J. Chromatogr., 84, 387 (1973).
[13] FDA Pesticide Analytical Manual, Vol. I, Section 312.
[14] Tindle, R. C., Gehrke, C. W., and Aue, W. A., J. Assoc. Offic. Anal.
Chem., 51, 682 (1968).
[15] Onley, J. H., and Yip, G., J. Assoc. Offic. Anal. Chem., 54, 1366
(1971).
[16] Lakota, S., and Aue, W. A., J. Chromatogr., 44, 472 (1969).
[17] Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Section 8.3.
[18] FDA Pesticide Analytical Manual, Vol. I, Section 313.
[19] Gough, T. A., and Sugden, K., J. Chromatogr., 86, 65 (1973).
[20] Kolb, B., and Bischoff, J., J. Chromatog. Sci., 12, 625 (1974).
- 52 -
-------
[21] Brazhnikov, V. V., Gurev, M. V., and Sakodynsky, K. I., Chromatogr.
Rev., 12, 1 (1970).
[22] Aue, W. A., Advances in Chemistry Series, No. 104, ACS, Washington,
D. C., 1971, page 39.
[23] Burgett, C. A., and Green, L. E., J. Chromatog. Sci. , 12, 356 (1974),
[24] Laski, R. R., and Watts, R. R., J. Assoc. Offic. Anal. Chem., 56,
328 (1973).
[25] Greenhalgh, R. , and Cochrane, W. P., J. Chromatogr., 70, 37 (1972).
[26] Cochrane, W. P., and Purkayastha, R., J. Agr. Food Chem., 21, 93
(1973).
[27] Young, H. Y., and Chu, A., J. Agr. Food Chem., 21, 711 (1973).
[28] Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Section 8.5.
[29] Lawrence, J. F., J. Chromatogr., 87, 333 (1973).
[30] Dolan, J. W., and Hall, R. C., Anal. Chem., 45, 2198 (1973).
[31] Hall, R. C., J. Chromatog. Sci., 12, 152 (1974).
[32] Hall, R. C., in Retention Times, Tracer Instruments, Inc., November,
1974.
[33] Lawrence, J. F., and Moore, A. H., Anal. Chem., 46, 755 (1974).
[34] Bowman, M. C., and Beroza, M., J. Chromatog. Sci., 7, 484 (1969).
[35] Bowman, M. C., Beroza, M., and Nickless, G., J. Chromatog. Sci., 9,
44 (1971).
[36] Bowman, M. C., Beroza, M. , and Hill, K. R., J. Chromatog. Sci.,
9, 162 (1971) .
[37] Versino, B., and Rossi, G., Chromatographia, 4, 331 (1971).
[38] Bache, C. A., and Lisk, D. J., Anal. Chem., 37, 1477 (1965); 38,
783 and 1757 (1966); 43, 950 (1971); J. Assoc~ Offic. Anal.Chem.,
50, 1246 (1967); J. Gas Chromatogr., 6, 301 (1968).
[39] Scolnick, M., J. Chromatog. Sci., 8, 462 (1970).
[40] Smith, E. D., Sorrells, K. E., and Swinea, R. G., J. Chromatog.
Sci., 12, 101 (1974).
- 53 -
-------
[41] Thompson, J. F., Walker, A. C., and Moseman, R. F., J. Assoc. Offie.
Anal. Chem., 52, 1251 and 1263 (1969).
[42] Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Section 12.3.
[43] McReynolds, W. O., J. Chromatog. Sci. , 8, 685 (1970).
[44] Analytical Methods for Pesticide Residues in Foods, Department of
National Health and Welfare, Ottawa, Canada, Table 8,. 1.
[45] Zweig, G., and Sherma, J., Analytical Methods for Pesticides and
Plant Growth Regulators, Volume VI, Gas Chromatographic Analysis,
Academic Press, N. Y., 1972.
[46] Zweig, G., and Sherma, J., editors, Handbook of Chromatography,
Volume I, CRC Press, Cleveland, Ohio, 1972.
[47] Watts, R. R., and Storherr, R. W. , J. Assoc. Offie. Anal. Chem.,
52, 513 (1969).
[48] Elgar, K. E., paper presented to joint meeting of ACS and CIS,
Toronto, Canada, May, 1970; published in Advances in Chemistry
Series No. 104, ACS, Washington, D. C., 1971, Chapter 10, page 151.
[49] Gaul, J. A., J. Assoc. Offic. Anal. Chem., 49, 389 (1966); Section
302.43 and 302.44, FDA Pesticide Analytical Manual.
[50] Thompson, J. F., Mann, J. B., Apodaca, A. 0., and Kantor, E. J.,
J. Assoc. Offic. Anal. Chem., 58, 5, 1037 (1975).
- 54 -
-------
Section 6
ADDITIONAL PROCEDURES IN PESTICIDE ANALYSIS
This section treats a number of miscellaneous topics important in resi-
due 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
6.A. 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 polyethelene may produce spurious responses. Similar-
ly, metal containers may contain trace impurities that will cause inter-
ference in GC analysis. In general, glass jars or bottles with aluminum
foil or Teflon lined lids are the most suitable sample containers,
although it is sometimes possible for pesticides in stored extracts to
be adsorbed onto the glass surfaces. Glass containers should be care-
fully precleaned as outlined in Subsection 3.L. in Section 3. Plastic
containers may be used, if necessary, only when non-interference with
the subsequent analysis has been proven.
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.
6.B. 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
be usually more useful for determining any pesticide contamination rather
than a dead species 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.
6.C. 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 la.rger
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.
It is good procedure to clearly label collected samples with all perti-
nent information such as a code number, date of collection, type of
sample, place and method of collection, description of collection site,
size of sample, etc.
6.D. 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 ssmple collected.
The exact steps in the compositing procedure will depend on the parti-
cular sample involved. Figure 6-A shows typical steps in reduction
- 2 -
-------
of a gross sample of an agricultural product collected in the field or
during processing or at the market.
FIGURE 6-A, TYPICAL STEPS IN REDUC-
TION OF A GROSS SAMPLE
Alternate step
Remove peel or husk (if necessary) and
reduce size of large units by cutting or cnopping
If necessary, reduce
size of large units by
cutting or chopping
6.E. STORAGE OF SAMPLES
As a general rule, samples should be analyzed as soon as possible after
their collection. If storage is necessary, it should be under prescribed
conditions which preserve the integrity of the original sample.
Tissue samples which 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 which 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. Under no circumstances should
extraction be deferred longer than an overnight period, even when the
samples are frozen.
- 3 -
-------
Water samples should be extracted at once if at all possible and stored
just above the freezing point if necessary. Pesticides can be adsorbed
on the glass container during storage.
If lengthy storage is required prior to analysis, a good alternative to
storage of samples is to extract at once and store the stripped solvent.
Decomposition in samples which must be stored can be evaluated by storing
spiked controls along with the samples.
Comments pertinent to collecting samples of different types will be pre-
sented in the Subsections 6.F. to 6.K. Methods for the analysis of the
various sample types are surveyed in Section 7 of this Manual.
6.F. SAMPLING OF AGRICULTURAL AND FOOD PRODUCTS
Procedures for sampling, sample preparation, sample compositing, and
sample reporting, as required by Federal law, for all commodity types
are outlined in detail in Sections 140-143 of the FDA Pesticide
Analytical Manual, Volume I. Section 3 of the Canadian Department of
National Health and Welfare Analytical Manual for Pesticide Residues
in Foods gives guidelines for systematically obtaining representative
samples of processed and packaged foods, bulk foods, and field crops
and for handling, shipping, and storing samples. Recommended minimum
sizes are tabulated for different samples, with a general sample
requirement of n product units, where n equals the square root of the
total but need not exceed 10-15 separate units.
Section 4 of the same Canadian PAM covers laboratory preparation of
analytical samples from gross samples of fresh, frozen, and canned
vegetables, fruits, and juices; dry cereal grains, flakes, dehydrated
fruits and vegetables; animal tissues; eggs; butter and margarine; milk
and cream; cheese and nuts; fats and oil; and fish and fish products.
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.
6.G. 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
- 4 -
-------
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
OGP 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.
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].
6.H. 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 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 . Refrigeration of the collected air samples
was unnecessary except in cases of long delay between sampling and
analysis.
- 5 -
-------
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 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.
Research [1] 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 conveniently
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 HCB) prior to analysis.
Other proposed media for collecting chlorinated and phosphate pesticides
in air include glass beads coated with cottonseed oil in a high-rate
sampler [2], support-bonded chromatographic phases such as 24%
(C H SiO ) on Chromosorb A or 31% (C H SiO . ) on Diatom W [3],
and columns of Tenax GC adsorbent. Adsorption and desorption (at 250°C)
of biphenyl at greater than 90 percent efficiency was demonstrated
with the latter [4].
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) [5]. Results with the screen
are only qualitative as the amount of air passing through is variable
and not 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 [6]. The carbamate insecticide propoxur (Baygon) was col-
lected from air in a Greenburgh-Smith impinger containing NaOH. The
solution was acidified and the pesticide extracted with benzene prior
to analysis [7].
6.1. WATER SAMPLING
The design of a comprehensive pesticide sampling program for environmental
waters is a specialized topic which is covered in publications available
from the Water Quality Control Division of the USEPA, National Environ-
mental Research Center, Cincinnati, Ohio. Important considerations
include the objective of the study, frequency of sampling, location of
- 6 -
-------
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
I
1
-------
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 as being present. Especially unstable pesti-
cides 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 deter-
minative steps. One disadvantage of glass sample bottles is possible
in sh pment, and care should be exercised in proper packaging
-------
(for pesticides and PCB's) [15, 16], polyethylene film (20-25 pm
thickness) [17], and polyurethane foam coated with selective
adsorbents [18] have all been used with varying success.
The XAD macroreticular adsorbent resins (XAD-1, -2, -4, and -7) have
been used to collect organics from both potable [19] and sea [20]
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 (0:1 v/v) as eluting sol-
vent. Among 10 chlorinated insecticides studied, only aldrin and
p_,p_'-DDE were not quantitatively recovered, and recovery of PCB's
was 76 percent [21]. Another study on XAD resins [12] found efficient
adsorption of chlorinated pesticides but poor recoveries on re-
extraction from the resin. Details for use of XAD-2 and -4 resins
for many classes of trace organic water contaminants have been pub-
lished [22], and recoveries between 81 and 96 percent reported for
20 ppt levels of atrazine, lindane, dieldrin, DDT, and DDE (47
percent for aldrin). A recent EPA report [23] recommends XAD-2
resin for routine monitoring of sea water for chlorinated insecti-
cides and PCB's.
Continuous liquid-liquid extractors are an alternative to filter-
adsorbent processes which are preferred by some analysts. A multi-
chamber extractor with internal solvent refreshing was constructed
[24] which 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. A similar modified apparatus was presented
later for use of both heavier- and lighter-than-water solvents [25].
A simple and rugged field version of the Kahn and Wayman apparatus
[24] was designed excluding solvent recycling and based on mixed
settling [26]. This apparatus, which consisted of an extraction
unit, magnetic stirrer, and pump, provided quantitative recovery of
pesticides and PCB's at levels of 0.1-1,0 ng/liter of river water.
The theory for extracting chlorinated pesticides continuously from
water with a stationary immiscible solvent is discussed in reference
[26].
6.J. SAMPLING OF HOUSE DUST, SOIL, AND STREAM BOTTOM SEDIMENT
House dust is collected with a vacuum cleaner, air dried, and sieved
prior to analysis. To sample soil, cores or borings of a known diameter
are 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 representing 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 com-
bined, quartered, and divided into 2 Ib samples for analysis. Soils
are analyzed in an air-dry state after sieving to remove foreign
material. There is no way to collect a truly representative soil
- 9 -
-------
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 which 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 it would be likely to find the maximum amounts of pol-
lution from considerations such as currents and industrial effluent
discharges.
About a quart of sediment is a typical sample size. Actual collection
is accomplished with one of a variety of core samplers or dredges.
Samples may be preserved with formalin or a variety of other sterilants
so long as the integrity of the sample is not affected. Samples are
air dried and ground prior to analysis. Storage, if necessary, is in
a freezer if volatile compounds such as 2,4-D may be present.
6.K. MARINE BIOLOGICAL SAMPLES
A problem sometimes encountered when collecting planktonic and bottom
fauna 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, using 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 organism 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. Preservation by quick freezing in
dry ice is most desirable. When this is not possible, liquid preser-
vatives are used. Larger fish should be injected with preservative
from a syrirtge to prevent decomposition of internal organs. Fish can
be analyzed whole to yield data on gross contamination, or the fish
can be portioned to obtain information on edible and non-edible parts.
Analyses of individual organs and tissues yield information on
- 10 -
-------
distribution of pesticides in the fish. Analysis of blood from a
dying fish may be valuable for determining probable cause of death
where pesticide exposure is suspected. The blood is obtained by cutting
off the tail at the caudal peduncle and collecting and freezing the
blood in a small vial.
Some of the material in the sections on sampling was adapted from a
training course manual entitled "Analysis of Pesticides in the Aquatic
Environment", U.S. Department of the Interior, Federal Water Pollution
Control Administration, March, 1968. Some of the above sampling proce-
dures and additional methods for collection of environmental samples
have been reviewed [27].
6.L. 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
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
- 11 -
-------
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 homogeneous sample has been described
[28]. The device, consisting of a handle and shaped aluminum sheet,
fits inside a blender jar and serves to gently push bulky samples into
the cutting blades during the blending operation. A liquid-nitrogen
cooled freeze grinder for biological materials containing labile
pesticides has also been devised [29].
Figure 6-B. Waring Aseptic Dispersall Model ASJ-1. (Shown on 702-CR Base)
. rigm 6-B. wMwe Asamc wswswi moan *#-
' '""" "
Blending with a solvent followed by filtering or centrifuging is particu-
larly 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 Na^SO 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 Na.SO 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
- 12 -
-------
or longer with a solvent such as methanol-chloroform (1:1 v/v) [30].
Preliminary steps such as drying, grinding, or chopping normally precede
the extraction. Even this procedure may not give complete extraction in
all cases, and only studies with samples to which radioactive tracers
have been applied can indicate the absolute extraction efficiency in
any particular case. The usual evaluation procedure of spiking a
sample with pesticide and looking for quantitative extraction is less
reliable than the radiotracer method because the spiked chemical will
not be naturally incorporated in the sample 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
[31], 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 [32].
An apparatus that simultaneously Soxhlet extracts pesticides and con-
centrates the resulting extract has been designed [33]. Advantages of
this cyclic extraction-evaporation system are that distillation of
solvents prior to extraction can often be omitted and excess solvent
is re-utilized for extraction.
The most efficient solvent and parameters for extraction of pesticides
from water can be determined using the p_-values originally suggested
by Beroza and co-workers for use in residue confirmation (Subsection
8.F. 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 or organo-
phosphates from water [34], 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. p-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 [35]. Using
2,4-D as a model, it was shown that a pesticide with a p-value equal or
greater than 0.90 can be 95 percent extracted from 50 ml aqueous phase
by up to five extractions 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 [36].
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.
- 13 -
-------
6.M. CONTROL OF METHODOLOGY FOR EVAPORATION AND 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 deter-
mine this and concentrate final solutions according to the least sensi-
tive 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).
Figure 6-C. Kuderna-Danish
Evaporative Concentrator, Kontes
Glass Co. No. K-570000.
The tube is heated in a vigorously boiling water bath in a hood with the
draft off so the column becomes hot enough to allow the escape of vola-
tile solvent vapors. The water level of the bath should be maintained
just below the lower S joint and the apparatus 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 required
for smooth operation of the K-D evaporator, and carborundum, checked for
absence of contamination, is recommended in preference to porcelain,
vanadium, or glass chips.
- 14 -
-------
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 tip of the tube in a steam or hot water bath by
means of a spring test tube holder. 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 100 yl syringe is
useful for performing this rinse. Finally, further fresh solvent is
added to dilute up to the desired volume, if necessary. A 2-ball micro-
Snyder column or a modified (3-ball) column can be used, the latter
being preferred for solvents boiling above ca. 65°C since problems from
bumping and column flooding are reduced.
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.
- 15 -
-------
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
under a stream of nitrogen adjusted to such pressure 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 excellent multitube apparatus for nitrogen evaporation is available
from Kontes Glass Co. [37]. 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
[38] are avoided, and a minimum of analyst attention is required.
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.
- 16 -
-------
It is important to avoid pesticide loss or decomposition during
evaporation steps. Numerous reports have been made [e.g., 38, 39]
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 in small volumes of acetone
as evaporation continues until all hexane is eliminated.
The use of air to enhance 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 [38].
A commercial tube heater which avoids evaporation to dryness with micro
K-D apparatus was originally described by Beroza and Bowman [40]
(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
- 17 -
-------
Other reports of pesticide loss include dieldrin and DDT when an
extract was evaporated in the presence of light [41] and carbamate
pesticides when evaporated in a K-D apparatus [42]. In the latter
case, rotary vacuum evaporation (Figure 6-G) at 50-55°C 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.
- 18 -
-------
HIGH PERFORMANCE LIQUID COLUMN CHROMATOGRAPHY (HPLC)
6.N. INTRODUCTION TO HPLC
High performance liquid chromatography is becoming increasingly important
as a powerful technique for the separation and analysis of complex mix-
tures. HPLC is a very gentle technique which commonly operates at ambient
temperature. It has advantages over gas chromatography where compounds
are not naturally volatile and derivatization is difficult or unsatis-
factory, and for 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 [43]. 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 pure
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 [44] as well as for the final
determination itself.
A disadvantage of LC is 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 gram range. In one comparative study of GC and HPLC [43],
detection limits by EC-GC were 100 to 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 yD without loss of linearity or peak symmetry. The
ability to apply large volumes in LC (usually with a sample injection
valve) can make sensitivity comparable between the two methods. Deriva-
tization methods can increase the sensitivity of detection, e.g, by
formation of UV adsorbing or fluorescent derivatives, but only at the cost
of more complicated sample preparation. Another advantage of GC is the
widespread knowledge of techniques and ready availability of numerous
relatively inexpensive, simple, reliable instruments. Many laboratories
purchasing HPLC instrumentation for the first time are disappointed
because reliable, consistent results are not as easy to obtain as with
their current GC instrumentation. The field of HPLC is only now in
its infancy, and as improved equipment is developed and analysts obtain
a better knowledge of the field, it will undoubtedly take its place
beside GC as an indispensable tool for pesticide residue determinations.
- 19 -
-------
6.0. HPLC INSTRUMENTS
The basic elements of a complete, automated HPLC instrument include a
solvent reservoir Cca. 50-300 ml) and gradient forming device, high
pressure pump and pulse damper, injection port, column, detector, and
recorder (Figure 6-H). The instrument components must be joined by
capillary tubing so as to reduce dead volume and 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.
Fieas 6-H. HIGH »
-------
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 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 resolution that occurs in a GC
separation. 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 sufficiently great that the trace will return
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 dif-
ferentiate between two solutes).
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.
As of this writing, HPLC is not yet an adequately sensitive, practical,
or reliable multiresidue procedure for the average residue analyst in
a field situation. Therefore, details of column packings, injection
methods, solvents, detectors, and other aspects are not included in this
Manual. Interested readers are referred to reviews describing LC
detectors [45-47], column packings [48], and general principles and
equipment [49, 50] and to books covering theory, principles, and practice
- 21 -
-------
of HPLC [51, 52]. The analysis of pesticides has been reviewed in
several papers [43, 53-55]. Certain pitfalls of HPLC such as effects
of column temperature on reversed phase separations, injection of samples
at high pressures, and column blockage when running samples with extrane-
ous material have been elaborated [56].
THIN LAYER CHROMATOGRAPHY
6.Q. INTRODUCTION
The first multiresidue method available to the pesticide analyst for
identification and estimation was based on paper chromatography [57-59].
Paper chromatography has now been largely replaced by thin layer chromato-
graphy (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 [59-62].
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
8.E. 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 concentration steps normally precede TLC
determination. Often more stringent cleanup is required for TLC than for
GC if streaked zones are to be avoided. For example, the 15 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 [63].
Major advantages of TLC are simplicity, rapidity, and low cost. Sensi-
tivity ranges from about 5-500 ng for most pesticide detection methods.
Rapid semi-quantitative estimation can be achieved by visual comparison
of sample and standard spot sizes and/or intensities, and more accurate
quantitation can be carried out by in situ scanning of spots with a
spectrodensitometer.
This section will briefly survey important aspects of TLC for screening
and quantitation of pesticide residues. General techniques of TLC were
described in detail in an extensive treatise [64], while specific
procedures for pesticide TLC were covered in several papers [60, 65].
Applications of TLC to pesticide analysis have been reviewed [66, 67],
- 22 -
-------
6.R. 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 ym thick. Pre-coated
layers are of high purity and uniformity and are used almost exclusively
in most laboratories. It should be mentioned here that pre-coated plates
in which AgNO is incorporated in the coating should be avoided. The
response of a number of pesticides is far less than when the AgNO
solution is applied at the time of developing. 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 layers (e.g., 120°C for 2 hours) prior to spotting and
development is often required. Silica gel and alumina usually give
the best results, but polyamide, microcrystalline cellulose, kieselguhr,
and magnesium oxide among other adsorbents have also been used.
FlflWE 6-1, ItedSH/fttrNCTW* AailJST-
(OLE WLICATOR FOR
CtWTIW, REGULAR OR FK«-
DIBTT UWB PLATES.
BSINWW* IKTHKNTS,
IHC.
- 23 -
-------
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 exactly follow 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 source of variation from laboratory to laboratory.
FIGURE 6-J, KEHB* maimus IX T«*S,
INSTRUMENTS, INC,
The following have proved to be generally useful solvent systems 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 moities of urea herbicides by coupling with
N-ethyl-1-naphthylamine [FDA PAM, Vol. II, Sec. 120.216]. Colorless
- 24 -
-------
spots can be detected by applying a chromogenic reagent, either by
spraying or dipping, after developing and drying the layer. A com-
mercial aerosol spray device is shown in Figure 6-K. Dipping is the
preferred method of application, if feasible, because of the uniformity
achieved and the hazards involved in careless spraying of corrosive,
toxic, or carcinogenic reagents. A Thomas-Mitchell dipping tank
is recommended. Sometimes the reagent can be incorporated in the layer
prior to development or included in the developing solvent. Naturally
fluorescent spots can be detected under short (254 nm) or long (366 nm)
wave UV light, or fluorescence may be induced by application of fluori-
genic reagents after development or preparation of fluorescent deriva-
tives (e.g., dansyl compounds) prior to spotting [68]. Spots which
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.
FjSURE W(. TIC AEROSOL SPRAYER, BtUMttttM
s, INC.
6.S. QUANTITATIVE TLC
Quantitation of separated spots may be 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. [69], Radioactive spots
can be quantitated by scintillation counting after scraping or by
automatic scanning of radioactivity on the layer.
- 25 -
-------
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 ul 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 [70]. 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 [71].
W_. FIBER OPTICS WIN LAYER SWNNER AND AUTOWTIC
SPOT APPLICATOR, KOTOS fiuASS
- 26 -
-------
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. Dipping is not always possible, however,
depending on the reagent solvent, adsorbent, and type of compounds on
the layer.
When developing a new densitometric method, the spot should be scanned
in all possible modes and directions and at a variety of wavelengths
in order to obtain the best signal to noise ratio and selectivity
for the compound of interest. These optimum conditions are then used
to obtain the calibration curve (linear range) and perform the analysis.
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 and given in a recent
book [72], and the quantitative TLC of pesticides has been reviewed
[73]. A fiber optics scanner specifically designed for pesticide
analysis [74] is available from the Kontes Glass Co. at a modest price
(Figure 6-L).
6.T. 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 pro-
vided by spraying with AgNO -2-phenoxyethanol reagent in ethanol or
acetone and exposing to hign 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 des-
cribed in the Canadian PAM, Section 14.10. Thin layer media must be
very low in chlorine content and other precautions and care taken or
large areas of the background may turn brown or grey and reduce the
contrast of the spots. A sensitivity in the 5-500 ng range is possible
with AgNO reagent, 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 R values for numerous
compounds in these two solvents, as well as for an alternate system
consisting of immobile dimethylformamide on alumina and isooctane
solvent, are given in Sections 410, 411, and 413 of the FDA PAM. Silver
nitrate can instead be 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].
- 27 -
-------
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 [75] and the latter [76] for the separation
of 13 common pesticides.
Extensive listings of additional solvent systems and corresponding R.,
values for chlorinated pesticides will be found in references [66], T77] ,
and [78]. The thin layer densitometry of chlorinated pesticides after
spraying with silver nitrate reagent is described in Chapter 15 of refe-
rence [72] .
b. ORGANOPHOSPHATES
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 p_-nitrobenzylpyridine and tetraethylpentamine spray [FDA PAM,
Section 432].
A two dimensional procedure [79] has the significant advantage of
specificity, obtained by oxidation of the phosphate pesticides with
bromine before development in the second direction. Silica gel layers
with toluene, 25 percent heptane in ethyl acetate, or ethyl acetate as
solvents were used along with the Storherr charcoal column cleanup
procedure [Subsection 7.Q. in Section 7] and enzymatic detection to
identify 18 pesticides in crops at 0.01 ppm levels. The same procedure
- 28 -
-------
should be well suited to phosphate pesticides in human and environmental
samples after appropriate cleanup.
Enzyme inhibition techniques are becoming increasingly important for the
selective and sensitive (pg-ng amounts) detection of enzyme inhibitors
such as organophosphate 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 ym), 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 sur-
rounded by enzyme free to hydrolyze the substrate and thus produce color.
While many phosphate 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 treat-
ment for phosphate and carbamate pesticides, and diagrams of mobilities
in hexane-acetone (8:2 v/v), a generally useful solvent for TLC on
450 pm silica gel layers. The preparation of these layers is detailed
in Section 12.4 of the Canadian PAM.
TLC enzyme inhibition methods and applications to pesticides have been
reviewed [80-82], as have the merits of TLC for analysis of organo-
phosphate residues [83]. The separation and detection of 42 phosphate
compounds using five ternary solvent solvent systems on three adsorbents
and three selective chromogenic sprays was reported [84].
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 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 pro-
duces 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 R^ 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
[85],
d. OTHER PESTICIDE CLASSES
The TLC of other classes of pesticides including carbamates, ureas,
phenols, dithiocarbamates, triazines, and organomercurials was reviewed
- 29 -
-------
in references [66] and [67]. 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 GC without derivative formation. Polyamine layers are especially
superior for the TLC of many carbamate pesticides [86].
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 dimethyIdithiocarbamates) or acetic acid-
methanol-benzene (1:2:12 v/v) for ethylenebisdithiocarbamates are used,
with detection as yellow, brown, or green spots by a cupric chloride-
hydroxylamine hydrochloride spray.
6.U. REFERENCES
[I] Mann, J. B., Enos, H. F., Gonzalez, J. , and Thompson, J. F,, , Environ
Sci. and Technol. , 8_, 584 (1974) .
[2] Compton, W., Analysis of Pesticides in Air, 13th Eastern Analytical
Symposium, N. Y. C., November 10-12, 1971, Abstract No. 70.
[3] Aue, W. A., and Teli, R. M., J. Chromatogr., 62, 15 (1971).
[4] Versino, B. , de Groot, M. , and Geiss, F. , Chromatographia, 7_,
302 (1974) . ~
[5] Tessari, J. D., and Spencer, D. L., J. Assoc. Offic. Anal. Chem.,
_54_, 1376 (1971) .
[6] Beyermann, K., and Eckrich, W., Z. Anal. Chem., 265, 4 (1973);
269, 279 (1974).
[7] Miller, C. W., Shafik, T. M., and Biros, F. J., Bull. Environ.
Contam. Toxicol., 8_, 339 (1972) .
[8] Eichelberger, J. W., and Lichtenberg, J. J., Environ. Sci. and
Technol., 5_, 541 (1971) .
[9] Braus, H. , Middleton, F. M. , and Walton, G. , Anal. Chem. , 2_3_,
1160 (1951).
[10] Sproul, O. J., and Ryckman, D. W., J. Water Pollut. Contr. Fed.,
33, 1188 (1961).
[11] Ruzicka, J. H. A., and Abbott, D. C., Talanta, 20, 1261 (1973).
- 30 -
-------
[12] Ahling, B., and Jensen, S., Anal. Chem., 42, 1483 (1970).
[13] Aue, W. A., Kapila, S., and Hastings, C. R., J. Chromatogr., 73,
99 (1972).
[14] Ito, T., Water Resource Research Institute of the University of
North Carolina, Report No. 54, August, 1971.
[15] 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).
[16] Musty, P. R., and Nickless, G., J. Chromatogr., 100, 83 (1974).
[17] Weil, L., Quentin, K -E, and Gitzowa, S., Gas -u. WassFach, 113,
64 (1972).
[18] Uthe, J. F. , Reinke, J. , and Gesser, H. D. , Environ. Lett. , 3_, 117
(1972).
[19] Burnham, A. K., Calder, G. V., Fritz, J. S., Junk, G. A., Svec, H. J.,
and Willis, R., Anal. Chem., 44, 139 (1972).
[20] Riley, J. P., and Taylor, D., Anal. Chim. Acta, 46, 307 (1969).
[21] Musty, P. R., and Nickless, G., J. Chromatogr., 89, 185 (1974).
[22] 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).
[23] Harvey, G. R. , U.S.E.P.A. Report R2-73-177, 1973, 32 pp.
[24] Kahn, L., and Wayman, C. H., Anal. Chem., 36, 1340 (1964).
[25] Goldberg, M. C., DeLong, L., and Sinclair, M., Anal. Chem., 45,
89 (1973).
[26] Ahnoff, M., and Josefsson, B., Anal. Chem., 46, 658 (1974).
[27] Ford, J. H. , McDaniel, C. A., White, F. C., Vest, R. E., and Roberts,
R. E., J. Chromatog. Sci., 13, 291 (1975).
[28] Analytical Methods for Pesticide Residues in Foods, Canadian Depart-
ment of National Health and Welfare, Section 14.4.
[29] Howells, K. J., Shaw, T. C., Rogers, P. P., and Galbraith, K. A.,
Lab. Pract., 23, 248 (1974).
- 31 -
-------
[30] U.S.F.D.A. Pesticide Analytical Manual, Vol. I, Section 253.
[31] Hill, B. D., and Stobbe, E. H., J. Agr. Food Chem., 22, 1143 (1974).
[32] Johnsen,R.E., and Starr, R. I., J. Agr. Food Chem., 20, 48 (1972).
[33] Voss, G., and Blass, G. , Analyst, 98_, 811 (1973).
[34] Suffet, I. H., and Faust, S. D., J. Agr. Food Chem. , 2_0, 52 (1972).
[35] Suffet, I. H., J. Agr. Food Chem., 21_, 288 (1973).
[36] Suffet, I. H. , J. Agr. Food Chem. , 2l_, 591 (1973).
[37] Beroza, M., Bowman, M. C., and Bierl, B. A., Anal. Chem., 44, 2411
(1972).
[38] Burke, J. A., Mills, P. A., and Bostwick, D. C., J. Assoc. Offie.
Anal. Chem., 49, 999 (1966).
[39] Chiba, M., and Morley, H. V., J. Assoc. Offic. Anal. Chem., 51, 55
(1968).
[40] Beroza, M., and Bowman, M. C., Anal. Chem., 39, 1200 (1967).
[41] McKinley, W. P., and Savary, G. , J. Agr. Food Chem., 10, 229 (1962).
[42] Storherr, R. W., J. Assoc. Offic. Anal. Chem., 55, 283 (1972).
[43] Eisenbeiss, F., and Seiper, H., J. Chromatogr., 83, 439 (1973).
[44] Larose, R. H., J. Assoc. Offic. Anal. Chem., 75, 1046 (1974).
[45] Conlon, R. D., Anal. Chem., 41, 107A (1969).
[46] Munk, M. N. , J. Chromatog. Sci. , 8_, 491 (1970).
[47] Veening, H., J. Chem. Educ., 43, 36A (1971).
[48] Majors, R. E., American Laboratory, May, 1972, p. 27.
[49] Taggart, W. P., Industrial Research, February, 1974, p. 76.
[50] Karasak, F. W., Research/Development, June, 1973, p.52.
[51] Brown, P. R., High Pressure Liquid Chromatography, Academic Press
N. Y., 1973.
[52] Kirkland, J. J., ed., Modern Practice of Liquid Chromatography, Wiley-
Interscience, N. Y., 1971.
- 32 -
-------
[53] Morgan, 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.
[54] Moye, H. A., J. Chromatog. Sci., 13, 268 (1975).
[55] Frei, R. W., and Lawrence, J. F., J. Chromatogr., 83, 321 1973).
[56] Chamberlain, W. J., Walters, D. B., and Chortyk, 0. T., Anal. Chim.
Acta, 76, 213 (1975).
[57] Mills, P. A., J. Assoc. Offic. Agr. Chem., 42, 734 (1959); 44, 171
(1961).
[58] Mitchell, L. C., J. Assoc. Offic. Agr. Chem., 40, 999 (1957).
[59] Sherma, J., and Zweig, G., Paper Chromatography, Academic Press,
N. Y. , 1971, Chapter 12 (Pesticides), pp. 355-396.
[60] Getz, M. E., Advances in Chemistry Series 104. ACS, 1971, Chapter 8.
[61] Getz, M. E., Residue Reviews, 2_, 9 (1963).
[62] Coffin, D. E., J. Assoc. Offic. Anal. Chem., 49, 1018 (1966).
[63] Heatherington, R. M., and Parouchais, C., J. Assoc. Offic. Anal.
Chem. , 53_, 146 (1970).
[64] Stahl, E., Thin Layer Chromatography, 2nd ed., Springer-Verlag, N. Y.,
1969.
[65] Kovacs, M. F., Jr., J. Assoc. Offic. Agr. Chem., 46, 884 (1963); 47,
1097 (1964).
[66] Sherma, J. , Chapter 1 in Analytical Methods for Pesticides and Plant
Growth Regulators, Vol. VII, Sherma, J., and Zweig, G., eds.,
Academic Press, N. Y., 1973.
[67] Sherma, J., Critical Reviews of Analytical Chemistry, CRC Press,
Aug. 1973, pp. 333-338.
[68] Lawrence, J. F., and Frei, R. W., Chromatogr. Rev., 18, 253 (1974).
(This is a review of theory, factors affecting emission, and
applications of fluorimetric derivatization for pesticide analysis
by HPLC and TLC.)
[69] Vitek, R. K., Seul, C. J., Baier, M., and Lau, E., American
Laboratory, February, 1974, p. 109.
[70] Emanuel, C. F., Anal. Chem., 45, 1568 (1973).
[71] Getz, M. E., J. Assoc. Offic. Anal. Chem., 54, 982 (1971).
- 33 -
-------
[72] Touchstone, J. C., ed., Quantitative Thin Layer Chromatography,
Wiley-Interscience, N. Y. , 1973.
[73] MacNeil, J. D. , and Frei, R. W. , J. Chromatog. Sci. , 13_, 279 (1975).
[74] Beroza, M., Hill, K. R., and Norris, K. H., Anal. Chem., 40, 1608
(1968).
[75] Eliakis, C. E. , Coutselinis, A. S. , and Eliakis, D. C.., Analyst,
93_, 368 (1968) .
[76] Szokolay, A., and Madaric, A., J. Chromatogr., 42, 509 (1969).
[77] Walker, K. C., and Beroza, M., J. Assoc. Offic. Agr. Chem., 46, 250
(1963).
[78] Thielemann, H. , Z. Chemie Lpz. , 14_, 292 (1974).
[79] Gardner, A. M. , J. Assoc. Anal. Chem., 54, 517 (1971),.
[80] Mendoza, C. E., Residue Reviews, 43, 105 (1972) and 50, 43 (1974);
J. Chromatogr., 78, 29 (1973).
[81] Gardner, A. M., FDA By-Lines, 2_(3) , 173, Jan., 1972.
[82] Villeneuve, D. C., Advances in Chemistry Series 104, ACS, 1971,
Chapter 3.
[83] Abbott, D. C., and Thomson, J., Residue Reviews, 11, 1 (1965);
Abbott, D. C., and Egan, H., Analyst, 92, 475 (1967).
[84] Getz, M. E., and Wheeler, H. G., J. Assoc. Offic. Anal. Chem.,
51., 1101 (1968) .
[85] Thielemann, H., Z. Anal. Chem., 272, 286 (1974).
[86] Nagasawa, K., Yoshidome, H., and Kamata, F., J. Chromatogr., 52,
453 (1970).
- 34 -
-------
Section 7
MULTIRESIDUE EXTRACTION AND ISOLATION PROCEDURES
FOR PESTICIDES AND METABOLITES
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 6.A.-6.K. in Section 6, while general comments
on sample extraction and extract concentration will be found in Sub-
sections 6.L. and 6.M. in Section 6.
CHLORINATED PESTICIDES
7.A. 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 phosphate 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 (PE), 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 quantisation. Some of the method
pitfalls observed during the operation of the interlaboratory quality
control program (Section 2) are as follows:
-------
a. Some analysts, with the mistaken notion of saving time, have combined
6 percent and 15 percent ethyl ether-petroleum ether Plorisil 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 pose a serious
- 2 -
-------
safety hazard. Methods have been set forth for the removal of
peroxides from ether but have not proven wholly satisfactory. If
the impure ether is in a small unit such as 1 pound, it is recommended
that it be discarded. If a larger volume is involved, it may be pre-
ferable to redistill it. Extreme caution must be used in the latter
course as the presence of the peroxides enhances instability. 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
impurities. 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 7.J. for further comments on pesticide elution
from Florisil.
7.B. 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 ot-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 7.A. Virtues of
the micromethod include a low background level and savings in the volume
of solvent required.
7.C. HUMAN BLOOD OR SERUM
A 2 ml aliquot of serum is extracted with hexane for 2 hours on a slow
speed rotary mixer. After concentration, the hexane layer is analyzed
by EC-GC [EPA PAM, Section 5,A,(3),(a)]. The procedure involves no
cleanup, but, if carefully handled, it is capable of yielding recoveries
of chlorinated pesticides comparable to that obtained from a full MOG
cleanup technique (see Table 2-11 in Section 2). Since all pesticides
- 3 -
-------
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 some-
times 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. Microcoulo-
metric GC determination after sulfuric acid extraction was successfully
applied to 24 organochlorine pesticides in blood at 1 ppb levels with no
cleanup [1].
7.D. PENTACHLOROPHENOL (PCP)IN BLOOD AND URINE
Acidified blood is extracted with benzene on a Roto-Rack for 2 hours fol-
lowed 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.
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 PCP methyl ester 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 PCP.
f. Other ether derivatives (e.g., ethyl, propyl, amyl, etc.) can be
prepared and characterized for confirmation of PCP identity.
g. A Vortex-Genie or similar mixer should be available for efficient
extraction and quenching of the derivatization reaction.
- 4 -
-------
7.E. 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 7.B.) 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_,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.
7.F. 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.
Deactivated silica gel (Subsection 4.A.d. 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 percen-
tage water added for deactivation should be increased if the compounds of
interest elute in a later fraction than that indicated in the detailed
procedure, or, the percentage of benzene in the benzene-hexane eluent can
be increased. Early elution would be remedied by less deactivation or
less polar solvents. Spiked control urine, rather than standard com-
pounds, should be used to determine the elution pattern.
The ethylating reagent is extremely hazardous and should be used with
extreme caution. It should not be stored or prepared in ground glass
stoppered or etched glassware. 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 7.N.
- 5 -
-------
7.G. 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 7.A.
and is described in detail in Section 211 of the FDA PAM. Eluents are
6, 15, and 50 percent ethyl ether in petroleum ether. The method for
nonfatty foods [FDA PAM, Section 212] 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 procedures, crops to which
they are applicable, and supplemental cleanup procedures for the Florisil
column fractions. Collaborative studies have been carried out with a
number of pesticides [2, 3]. The problem areas are the same as those
given in Subsections 7.A. and 7.J. 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 tabu-
lated 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., organophos-
phates), a new elution system consisting of different mixtures of methy-
lene chloride, hexane, and acetonitrile was devised as replacement for
the traditional diethyl ether-petroleum ether eluents. At least 50
pesticides and related chemicals have been recovered, in groupings
different from the mixed ether systems, with these new solvents [4].
A rapid screening method for residues in milk is another modification
of the standard procedure. Pesticides are cleaned-up by partitioning,
but Florisil chromatography is excluded [5].
7.H. 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 air.
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
- 6 -
-------
extracted into chloroform, methylated, and determined by microcoulo-
metric GC without column cleanup [FDA PAM, Section 222]. The recovery
of residues of eight compounds in grains and vegetables has been veri-
fied. Sodium sulfate used in drying steps may adsorb some free chloro-
phenoxy acids, and the reagent should be checked for complete recovery.
An alternate 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 [6]. Residues are ex-
tracted with acetonitrile (nonfatty foods) or chloroform (fatty foods and
liquids), the acids are converted to their sodium salts, acidified, methy-
lated, cleaned-up on a Florisil column, and determined by EC- or micro-
coulometric GC. Severe emulsions form during cleanup of the basic
phase by extraction with chloroform and must be broken by centrifugation.
7.1. CARBON-CELLULOSE COLUMN CLEANUP
Section 7.1 of the Canadian PAM describes a method for cleanup of organo-
chlorine and organophosphorus insecticides, herbicides, and fungicides
following acetonitrile extraction (blending) and hexane partition of
residues in foods. Sequential elution with three solvents (1.5 percent
acetonitrile in hexane, chloroform, and benzene) separates the pesticides
into three fractions which are suitable for FC and TLC determination.
Some 40 pesticides have been quantitatively recovered from a variety of
foods with this system.
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 mechani-
cal shaker. Cellulose (e.g., Solka Floe BW40) is extracted twice with
acetone in a similar manner without prior heating.
7-J. 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 4.A.C.
in Section 4 of the 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
- 7 -
-------
Table 7-1
ORDER OF ELVTION OF PESTICIDES FROM FLORISIL PARTIALLY
DEACTIVATED WITH 2# WATER USING 300 ml VOLUME OF ELUENTS
(From the Canadian PAM)
1 # CH?Clp in Hexane
PESTICIDES ^
Hexane y/° 109& 157° 2.(y/h ~}%
Aroclor 125^ PCB +
Chlordane +
Toxaphene +
Strobane +
trans-Chlordane +
Chlordene +
Aldrin +
Hexachlorobenzene +
Heptachlor +
p.p'-DDE +
O.p'-DDT +
Mirex +
Isobenzan +2
p,p'-DDT S(90$) -*
a-BHC SC*5#) +
Perthane s(lOJS) S(85#) +
p.p'-DDD +
Chlorbenside M +2
f-BHC S(80J8) +
PCNB S(80J5) +
TCNB S(^5%) +2
g-BHC S(10J{) +
f-BHC +
Dicofol +
Ronnel OP 8(65^) +
Hepachlor epoxide S(60J4) +
Dichlofenthion OP +
Phorate OP,M +
Carbophenothion OEM S(25^) +
Endosulfan I S(10#) +
Dieldrir. +
Chlorpyrifos OP «
Endrin *
Methoxychlor *
Parathion OP
Ethion2 OP
2,^-D methyl ester
2A5-T methyl ester
Anilazine
Ovex
Fenitrothion OP
Tetradifon
Diazinon OP
Chlorothalonil
Methyl Parathion3 OP
Sulphenone
Dioxathion OP
Malathion OP
Atrazinej4'
Simazine4
Endosulfan II
Captan M
Phosmet OP
DCPA
Azinphosmethyl OP
% EtOAc in Hexane Percent
rj% IVfc 2Ofi ')W/° Recoveries
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
>95
^95
>95
>95
>95
>95
>95
>95
>95
>95
~ 20
>95
>95
>95
>95
>95
+ >95
+ >95
+ >95
+ >95
+ >95
+ >95
S(90?S) + >95
S(90?8) + >95
S(90£) + >95
S(70?J) + >95
S(50J5) + >95
s(l5^) + >95
+ >95
* >95
+ >90
»- ^90
S(75#) + >95
+ "BO
+ >95
+ > 90
f2 >95
Notet A 30% CH2C12 fraction was eluted prior to all ethyl acetate fractions. All others were
+ - mostly elutes in first 25° ml
* - large amount in 250-300 ml fraction
S - some (as percent)
Footnotesi 1 - OP = organophosphorusi M = mercaptani PCB = polychlorinated biphenyl
2 - Higher recoveries are obtained by elution with more polar eJuents
3 - Remaining methyl parathion elvites in another 50 ml of 5% EtOAc
k - Detected by alakli flame detector
-------
by this procedure or mutually interfering residues may occur in the
same fraction.
The following factors affect the success of this Plorisil 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, PCB's, and HCB.
7.K. 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 clean-
up 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 supernatant
and can be determined by EC-GC. Forty pesticides have been quantita-
tively (80+ percent) recovered from a variety of plant and animal pro-
ducts at levels greater than 0.05 ppm.
7.L. MISCELLANEOUS MULTIRESIDUE CLEANUP PROCEDURES
Other frequently applied multiresidue procedures include the following.
The method of de Faubert Maunder [7] 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
- 9 -
-------
extracted three times with hexane-saturated hexane and then shaken
with a large volume of 2 percent Na SO 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 [8] 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 [9]. No gross general differences
were found in results, but one method might be advantageous for a parti-
cular sample type.
Hexane extracts or 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 [10]. 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.
A gel permeation chromatography (GPG) system using Bio-Beads S-X2
crosslinked polystyrene gel has been designed by Stalling et_ al. [11]
for removal of lipids from extracts of samples such as fish, before
EC-GC determination of commonly occurring pesticide and PCB residues.
In a typical cleanup, a fatty tissue sample may be desiccated with
anhydrous Na?SO (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 PCB's 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 [12] 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 significant
change in elution volumes or recoveries.
- 10 -
-------
Table 7-2
ORDER OF ELUTION OF ORGANOCHLORINES FROM DEACTIVATED SILICA GEL
ACCORDING TO THE METHOD OF HOLDEN AND MARSDEN [10]
Eluted in order
by hexane
Eluted in order by 10%
diethyl ether in hexane
Hexachlorobenzene
Aldrin
PCB's
£_,£_' -DDE
Heptachlor
£,£'-MDE (DDMU)
0_,£'-DDT
£,£' -DDT
Endrin
Chlordane
£,£'-DCBP
Toxaphene
p_,p_'-TDE
Telodrin
Heptachlor epoxide
a-BHC
Perthane
3-BHC
Kelthane
Y-BHC
Dieldrin
Methoxychlor
Thin layer chromatography (TLC) on 1-5 mm layers can provide cleanup if
a minimal amount of fatty material is present in the extract. Sample is
applied as a streak and developed along with standard marker compounds
on the same plate to allow location of the pesticide zones. These bands
are removed by scraping and are extracted to recover the separated
pesticides. Modified layers have been devised with capability for
increased sample loading, e.g., multiband or wedge-layer chromatoplates
[13]. With the latter, cleanup and determination can be combined on
the same layer without intervening elution.
- 11 -
-------
The use of ion exchange resins for cleanup of ionic pesticides has been
reviewed [14]. 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 [15].
ORGANOPHOSPHORUS (OGP) PESTICIDES
7.M. DETERMINATION OF METABOLITES OR HYDROLYSIS PRODUCTS IN HUMAN URINE
The determination of alkyl phosphate metabolites in urine provides a
measure of the extent of human exposure to the parent OGP pesticide.
Section 6,A,(2),(a) of the EPA PAM contains a sensitive and selective
analytical procedure for six phosphate, phosphorothioate, and phosphoro-
dithioate metabolites (hydrolysis products) of important pesticides.
Alkyl phosphates in acidified urine are extracted with acetonitrile-
diethyl ether (1:1 v/v), converted to dialkyl phosphates by reaction
with diazopentane, and the derivatives are pre-fractionated and separated
from inorganic phosphate and other impurities on a 2 gram silica gel
column deactivated with 1-2 percent water. Determination and confir-
mation 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 pesti-
cide (but not the exact compound) involved in the exposure may be deduced
by characterizing the metabolite(s) excreted.
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), VIII 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
4.1. in Section 4). Further confirmation of any particular metabolite
can be accomplished by preparing its hexyl derivative.
Alkyl phosphates are only minor urinary metabolites of malathion, the
major metabolites being the corresponding mono- and dicarboxylic acids.
-12-
-------
An unpublished silica gel cleanup FPD-GC method for determining these
acids as a measure of exposure to malathion has been devised by Shafik
and Bradway in the Methods Development Section of the Analytical
Chemistry Branch, EPA Health Effects Research Laboratory, Research
Triangle Park, N. C., from which details can be obtained.
7.N. DETERMINATION OF p_-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 [16]. 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). Elution with
various concentrations of benzene in hexane purifies and fractionates
the ethers, which are finally determined by EC-GC.
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 plus 80 percent fractions.
The phenoxy acids are detected intact along with 2,4-dichlorophenol and
2,4,5-trichlorophenol, their potential mammalian metabolites.
7.0. ANALYSIS OF FATTY AND NONFATTY FOODS USING FLORISIL CLEANUP
The MOG Florisil procedures described in Subsection 7.G. are adequate
for determination of some OGP pesticides in fatty and nonfatty foods
[FDA PAM, Sections 231 and 232], Malathion and some other phosphate
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 to be occasional-
ly inconsistent. As was mentioned earlier, OGP pesticides can be lost
through degradation on the Florisil column and during subsequent eva-
porations, or when water dilution of the acetonitrile extract for
- 13 -
-------
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 phosphate 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 [17-19].
Beckman and' Garber [20] recommended the solvent series benzene, diethyl
ether-benzene (1:2), acetone, and methanol for elution of Florisil columns.
The elution pattern and recovery of 65 OGP 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 phosphate pesticides [21].
7.P. SWEEP CO-DISTILLATION
In the sweep co-distillation technique [FDA PAM, Section 232.2], samples
are extracted by blending with ethyl acetate or another suitable solvent,
and an aliquot is concentrated and cleaned-up by injection into a 24 cm
heated glass tube packed with glass wool, followed at three minute inter-
vals by repeated injections of ethyl acetate. Nitrogen carrier gas sweeps
the vaporized volatile components through the tube to the condensing bath
and then through a short Anakrom scrubber tube to a concentration col-
lection tube. The organic interferences remain on the glass wool while
the pesticides are collected for analysis by FPD- or thermionic-GC. The
cleanup may not be suitable for EC-GC. Phosphate pesticides in a variety
of fatty [18] and nonfatty crops have been determined after sweep co-
distillation cleanup. Figure 7-A is a schematic diagram of the required
apparatus, and Figure 7-B shows the appearance of the commercial model.
Sweep co-distillation is a simple, useful technique which eliminates the
need for specialized adsorbents and large volumes of purified solvents.
It has been extended to organochlorine residues with superheated hexane
vapor as the carrier in place of nitrogen [22].
7.Q. CHARCOAL CLEANUP OF NONFATTY FOOD EXTRACTS
A general determinative method for organophosphorus pesticide residues
in nonfatty foods is based on the FDA acetonitrile extraction procedure
followed by dilution with methylene chloride to free extracted 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-Sea Sorb
- 14 -
-------
Figure 7-A. Sweep co-distillation apparatus, schematic diagram.
Copper
tube
Insulation
Asbestos
-Heating tape
Storherr tube
containing 5-6 inches
silonized glass wool
Nitrogen
inlet
Figure 7-B. Sweep co-distillation apparatus, Kontes Glass Co., K-500750,
- 15 -
-------
43-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 satisfac-
tory recovery of 41 pesticides and alteration products from kale and
9 typical pesticides from other low and high sugar content crops was
demonstrated [23]. A collaborative study [24] of this method for resi-
dues of six OGP compounds in apples and green beans verified recoveries
between 86 and 125 percent when either a thermionic or FPD detector was
employed.
Charcoal is acid washed by boiling 100 to 200 grams with 500 ml concen-
trated HC1 with magnetic stirring for one hour in a covered container.
Then 500 ml water is added and boiling repeated for an additional 30
minutes. The charcoal is collected in a Buchner funnel, washed with
water until neutral to indicator paper, and dried in an oven at 130°C.
Although becoming increasingly popular for multiresidue analytical pro-
cedures, problems are to be expected with the reproducibillty of
cleanup columns containing charcoal mixed with other adsorbents and
"mixed columns" in general.
7.R. MISCELLANEOUS MULTIRESIDUE CLEANUP PROCEDURES
Nine extraction procedures were compared for efficiency of removal of
six OGP pesticides and metabolites from field treated crops, and Soxhlet
extraction of the finely chopped crops with chloroform-10 percent
methanol proved most reliable and efficient [25].
Alumina has proven unreliable as a cleanup adsorbent for phosphates
since recovery of the more polar compounds is not complete [26]. Using
alumina (activity II to III) and petroleum ether and 3 percent aicetone-
petroleum ether as eluents, Renvall and Akerblom [27] eluted only 13 of
the 31 OGP compounds they tested.
Cleanup by solvent partition without column chromatography has proven
adequate for analyses of seven types of foods for 39 pesticides and
metabolites when detection was made with a thermionic detector [28].
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 [29]. A multiresidue
analysis of 14 pesticides in natural waters at ppb levels involving
extraction and concentration before FPD-GC has been reported [30] .
The elution pattern of a series of representative phosphate 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:
- 16 -
-------
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 columns have been used for cleanup of animal, plant, soil,
and water extracts prior to GC determination of OGP pesticides [31] and
to separate OGP pesticides and metabolites into groups to facilitate
their identification by GC [32].
The cleanup of 22 phosphate pesticides in 12 vegetable extracts was
achieved by gel filtration chromatography on Sephadex LH-20, Residues
were determined by GC with thermionic detection at levels of 0.05 to
0.5 ppm [33].
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 [34].
CARBAMATE PESTICIDES AND MISCELLANEOUS HERBICIDES
7.S. 1-NAPHTHOL IN URINE
Humans exposed to the N-methyl carbamate 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-naphthol 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
- 17 -
-------
reaction and must be avoided. Derivatized standards are stable for
about 6 months if stored in a refrigerator.
7.T. DETERMINATION OF 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-
flurorpropionic anhydride and determined by EC-GC. Details are available
from the above address.
7.U. 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 hydroly-
sis, 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, l-fluoro-2,4-dinitrobenzene, and penta-
fluorobenzyl bromide. 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 [35] 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-
trifluoro-3,5-dinitrotoluene, l-fluoro-2,4-dinitrobenzene, and penta-
fluoropropionic anhydride. These reactions are also reviewed in
reference [35].
7.V. DIRECT METHODS OF ANALYSIS
Determinations of intact, underivatized N-methyl carbamate insecticides
are hampered by their decomposition on GC columns under ordinary operating
- 18 -
-------
conditions [36]. Losses can be minimized by the use of specially pre-
pared, conditioned, and maintained columns. Presilanized supports do
not provide sufficient deactivation to prevent degradation of carbamates
so it is necessary to employ in situ silanization both during initial
conditioning and thereafter to restore column performance. Examples
of direct analyses of crop extracts include a multiresidue method [37] on
5-6% DC-200.after acetonitrile partition and charcoal cleanup as for
organophosphates [23] and determination of carbofuran and other carba-
mates on 20% SE-30 [38]. Highly deactivated GC column packing prepared
according to Aue [39] from acid washed Chromosorb W support, surface
modified with Carbowax 20M, has also been successfully used for chroma-
tography of intact N-methylcarbamates without degradation on the column
[40]. Such columns are extremely promising for performing analyses
without required derivatization.
Urea and N-aryl carbamate 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 and 10% DC-200/15% QF-1 (1:1) at 160°C have been
successfully used, the former for multiresidues of urea herbicides [41]
and the latter for carbamate herbicides [42] in foods. However, decom-
position of compounds of these types has been noted under certain conditions,
and determinations are therefore often made via thermally stable deriva-
tives of hydrolysis products as mentioned earlier.
7.W. ANALYSIS OF PLANT AND FOOD MATERIALS
Extraction of urea and carbamate pesticides from plant materials usually
involves blending with methylene chloride, acetone, chloroform, acetoni-
trile, or an alcohol (or these solvents plus anhydrous Na SO ). If
the presence of possible conjugated hydroxy metabolites is suspected,
hydrolysis with a strong acid during extraction may be included.
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 [42], while Florisil
was employed after acetonitrile-petroleum ether partition for the multi-
residue, multiclass determination of carbamate, urea, and amide residues
[43]. Methods for extraction, cleanup, and GC of carbamates, ureas, and
other classes of herbicides (triazines, uracils, phenols) have been
reviewed [44,45]. 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 [46].
- 19 -
-------
7.X. AIR ANALYSIS
Section 8,B of the EPA PAM contains details for analysis of the 400 ml
ethylene glycol sample from four air-sampling impingers (Subsection
6.H. in Section 6). The ethylene glycol is transferred to a large
separatory funnel and diluted with water. Pesticides are extracted
with hexane, the extract concentrated in a K-D evaporator, and a prelimi-
nary GC evaluation made for organochlorine and phosphate pesticides
using EC and FPC detectors, respectively. Background interferences may
preclude EC appraisal at this point, but the selectivity of the FPD may
allow tentative identification and quantitation of phosphates without
further cleanup. The electrolytic conductivity detector might be suitable
for determination of organochlorines at this stage without additional
cleanup. If required, based on preliminary results, the pesticide extract
is cleaned-up and fractionated on a Florisil column prior to detailed
analysis by EC-GC. FPD-GC can be repeated on the column eluates to
assure identifications obtained from the unpartitioned extract. Detection
limits [EPA PAM, Section 8,B, Table 1] range between 0.2-4 ng/m for
most pesticides recovered through the procedure. A total of 80 m
of air is sampled in 24 hours in four impingers, represented by 400 ml of
ethylene glycol. All precautions and comments under Subsection 7.A. are
pertinent to this methodology.
Although the above analytical method has been used for routine moni-
toring of some twenty-four organochlorine and chlorophenoxy acid ester
pesticides and four phosphate insecticides, it has the limitation that
only those pesticides extracted from ethylene glycol by hexane and sub-
sequently eluted from the Florisil column by the mixed ether eluents may
be identified and quantitated. Carbamates and numerous polar phosphate
pesticides would not be determined by this system because of these
limitations.
A multiresidue, multiclass analytical method for chlorinated, phosphate,
and N-methyl carbamate insecticides in ethylene glycol trapping solvent
was developed in the EPA laboratories [47] to overcome these difficulties.
Although the method has been tested with only a limited number of repre-
sentative compounds present in ethylene glycol, it is likely to be suit-
able for the determination of many other pesticides.
The new procedure involves addition of 600 ml of 2 percent sodium sulfate
solution to 100 ml ethylene glycol, 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, phosphates (fractions II and III) by FPD-GC, and carbamates
(fraction III) by EC-GC after derivatization with pentafluoropropionic
- 20 -
-------
anhydride. Recoveries of 90-110 percent of most of the 27 compounds
studied were obtained at 1-4 ng/m levels. Problems were encountered
with propoxur, aminocarb, and mexacarbate at low levels because of
background interference from the ethylene glycol.
This new multiresidue approach should be adaptable for the determination
of pesticide residues in samples other than air, although additional
cleanup steps may be necessary for difficult samples (e.g., gel permeation
chromatography for human fat). This general applicability was indicated
by the fact that all the pesticides studied were quantitatively extract-
able by methylene chloride from acetonitrile, which is a common solvent
for extractions from human and environmental samples.
7.Y. WATER ANALYSIS
To determine chlorinated and nonpolar phosphate pesticides, the water
sample (Subsection 6.1. in Section 6) is extracted with 15 percent methy-
lene chloride-hexane in a separatory funnel and an aliquot of extract
concentrate injected for an exploratory electron capture gas chromatogram.
The sample size is dictated by the expected residue levels and the require-
ments of the confirmatory procedures to be used. If the initial chroma-
togram indicates the presence of excessive amounts of interfering materials,
cleanup by acetonitrile partition and Florisil adsorption is carried out
as for tissue samples (Subsection 7.A.). Details of these water analysis
procedures are found in the EPA PAM, Section 10,A. The method, with
some modifications, is essentially the same as that presented in the
Federal Register in June, 1973 [48].
Cleanup is often not required for EC-GC analysis of surface water samples
[49] and is usually not required for any type of water if a selective GC
detector is employed. For example, the multiresidue analysis of 14 OGP
pesticides in natural waters has been carried out at ppb levels by
extraction, concentration, and direct GC with a FPD detector in the P
mode [30]. Where needed, cleanup and separation of common chlorinated
and phosphate insecticides extracted from water have been successfully
carried out in silica gel microcolumns [50, 51] and columns of deacti-
vated (5 percent HO) silica gel and alumina [52].
Extracts of water, sediment, and soil often contain large amounts of
elemental sulfur, which interferes in the GC analysis of early eluting
pesticides with the EC or FPD detectors. Chemical desulfurization
with Raney copper powder [53] or precipitation with metallic mercury
[54] have been used to remove such interference (Subsection 7.Z.).
Polar phosphate, 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
- 21 -
-------
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 derivatization steps.
Chlorophenoxy herbicides and their esters are 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 microcoulornetric detector
[55].
TLC determinations of carbamate, urea, triazine, and uracil herbicide
residues in water have been reviewed [35, 56], as have the extraction,
cleanup, GC determination, and confirmation of chlorinated insecticides
in water and soils [57].
7.Z. 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.
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). 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 dehy-
drated 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 GC-GC. Details of the entire procedure
are presented in Section 11,B of the EPA PAM.
Air drying of the sample required 1-3 days, depending on the soil type.
Such samples will contain at least 50 percent water. Pesticide concen-
trations 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.
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
- 22 -
-------
this treatment (e.g., organophosphates, 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
[58] with recoveries of 75-99 percent from suspended sediment and bottom
material. Extraction was with acetone and hexane added separately,
coextractives (including PCB's) were isolated by alumina and silica gel
column chromatography, and EC-GC was applied to 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 [59]. The shake extraction method with
hexane-acetone after moistening the soil with 0.2 M NH Cl was studied
collaboratively using standard AOAC analytical methods (Florisil cleanup
and EC-GC [60] and found to give excellent recoveries for six insecti-
cides in three different soils [61].
A multiresidue GC procedure for the herbicides dichlobenil, dinitramine,
triallate, and trifluralin in soils was described by Smith [62], Extrac-
tion was carried out with acetonitrile-water (9:1) 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. Anilide herbicides were determined by GC
after extraction from soil by blending with acetone [63]. Urea and
carbamate herbicides were recovered from soils by shaking with methanol
[64] or acetone [65] and by alkaline hydrolysis and steam distillation
[66], lodinated [66] and 2,4-dinitrophenyl [65] derivatives were used
for EC-GC determination of the herbicides. Triazines were extracted
with diethyl ether from soil treated with ammonia [67] and uracils with
1.5 N NaOH [68], Nineteen acidic, neutral and basic herbicides have
been determined in soils by two dimensional TLC [69].
The electrolytic conductivity detector has been used to determine nitro-
gen-containing residues in crude soil extracts. A detector maintenance
program for decontamination of the transfer lines and vent valve provided
reliable operation with little "down time" even though lengthy extract
cleanup was not carried out [70].
The analysis of pesticides of many classes in soils and plants has been
reviewed [71].
- 23 -
-------
POLYCHLORINATED BIPHENYLS (PCB'S)
7.A.A. PESTICIDE-PCB MIXTURES
PCB's 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 PCB's but with
their effect' on the reliable determination of pesticide residues. PCS
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, p_,p_'-DDT, £,£'-DDD, and p_,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.
PCB's 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).
7.A.B. 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 PCB's 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 PCB's 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), £,£f-DDE (3), p_,£'-DDD (5), £,£'-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
- 24 -
-------
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
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
- 25 -
-------
1 AldVj
i niurjri
Figure 7-E.Aroclor 1254 (solid line) 2 Kept. Epo.id.
and pesticide mixture * n'P,jDDE
4 Dieldrin
(dotted line). Column SP.P'-DDD
4% SE-30/6% QF-1, 200°C, 6P,P'-DDT
carrier flow 10 ml/min. 'D;'-" I *»»*«*'«'
7 a Oilan H
8 12 16 20
Retention, minutes
24
28
32
Confusing chromatograms also result when PCB's 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 PCS 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
- 26 -
-------
Figure 7-G shows a mixture of Aroclor 1254 with toxaphene. Analyses of
toxaphene, chlordane, and PCB's 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 PCB's 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
28
36
The actual effect of PCB's 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 PCB's 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 PCB's 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.
- 27 -
-------
7.A.C. METHODS FOR SEPARATION AND ANALYSIS OF PESTICIDES AND PCB'S
a. Published Procedures and Data
The EPA Pam contains a detailed discussion of PCB interference with
£,£'-DDT, p_,£'-DDT, £,£'-DDD and £,£'-DDE and specific recommen-
dations for quantitation of these pesticides on the recommended
GC columns (Section 9,B). The EPA Manual also describes the
separation of PCB's from DDT and its analogs by the method of
Armour and Burke [72] (Section 9,C), and a thin layer method for
semiquantitative estimation of PCB's 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.
b. Cleanup and Separation Systems
PCB's are eluted with 6 percent ethyl ether-petroleum ether in
the modified MOG procedure described in the EPA PAM, Section
5,A,(1). A study by Lieb and Bills [73] 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.
The method of Armour and Burke [72] has been most used for
pesticide-PCB separation. The 6 percent ethyl ether-petroleum
ether Florisil column eluate is concentrated to an appropriate
volume and a 5 ml aliquot applied to a column of partially deacti-
vated silicic acid and Celite. 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. PCB's and polychlorinated terphenyls
split between the two fractions [EPA PAM, Section 9,C, Table 1]
as do the pesticides aldrin and P_,P_'-DDE [Canadian PAM, Section
7.5]. Polychlorinated naphthalenes [74] 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
- 28 -
-------
perform a sufficient number of recovery trials with spiked
samples to gain confidence in its reliability. A slightly
modified version of the Armour-Burke method is detailed in
the Canadian PAM Section 7.5, and the method has been minia-
turized for determining chlorinated pesticide and PCB residues
in fish. In the latter method, the sample is dried with Na SO
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 [75].
Aroclor 1254 residues in blood have been determined by extraction
with hexane-saturated acetonitrile and cleanup on an alumina
column. Eluates were determined by EC-GC on an OV-1 column [76].
c. GC Quantitation
One of the most difficult aspects of PCB quantitation is to
obtain a match between the sample and a standard. Environ-
mental samples seldom have a GC peak pattern that will match any
Aroclor standard, and even commercial preparations of PCB's vary
from batch to batch in abundance of minor components. 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 of detector response for the residue to the total area of
response obtained under the same conditions for a known weight of
the commercial Aroclor standard with the most similar pattern.
If the presence of more than one Aroclor is clearly indicated,
the residue may be quantitated using standards judged appro-
priate for different portions of the sample chromatogram. Other
quantitation approaches which have been attempted include per-
chlorination of all PCB compounds with SbCl to a single derivative
(decachlorobiphenyl) [77], estimation of the weight of PCB in-
jected by dividing the retention time x peak height for all PCB
peaks by the product of peak height and retention time for 1 ng
p_,p_'-DDE on the same GC column [78], and peak resolution and
matching by a computer [79]. GC-MS with individual mass moni-
toring using a minicomputer controlled spectrometer has been
reported [80]. This method provided sensitive qualitative and
quantitative analysis of sediment extract without the need for
elaborate column adsorption separations prior to GC.
Beroza p_-values have been applied to quantitation of p_,p_'-DDT in
the presence of non-resolved PCB peaks and results within +11
percent of actual were reported [81].
d. Estimation by TLC
The semiquantitative TLC procedure [82] for determination of PCB's
in adipose tissue makes use of a 6 percent Florisil cleanup
- 29 -
-------
column eluate. Concentrate is treated with KOH to dehydro-
chlorinate DDT and ODD to their olefins, thereby eliminating
the problem of separating the pesticides from the PCB's. Any
interfering DDE is then oxidized to p_,p_'-Dichlorobenzophenone
which has an FL, value different from the PCB's on a silver nitrate-
impregnated alumina layer developed with 5 percent benzene in
hexane. The PCB's 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 [83] .
e. Confirmation of PCB's
Confirmation of PCB's has been obtained by perchlorination [77]
and alkaline hydrolysis [84]. The stability of PCB's to alkali
makes the latter reaction useful for confirming the identity of
PCB residues, and at the same time, conversion of DDT to DDE
removes some interference to quantitation of PCB's. Alkaline
hydrolysis also provides additional cleanup for many sample types.
Resistance to oxidation with chromic acid-acetic acid reagent
is also useful evidence for identifying PCB's in the presence of
reactive pesticides such as DDE and DDT [82] and chlorinated
naphthalenes [85].
Two-dimensional [86] or multi-development reversed phase [87] TLC
systems which separate PCB's from DDE, DDT, and other pesticides
can aid identification. PCB's are destroyed by UV irradiation,
but many pesticides may be altered as well. Toxaphene survives
UV treatment that destroys PCB's and can be confirmed in mixtures
in this way. Mirex, a late eluting pesticide which usually is
not interfered with by PCB's, also withstands UV irradiation and
can thus be confirmed. Most organochlorine pesticides are
destroyed by reaction with HNO ~H_SO. whereas PCB's and toxaphene
are unaffected. Chlorinated pesticides were selectively detected
in the presence of PCB's by use of a modified Coulson conductivity
detector at 600°C with a hydrogen flow of 1-2 ml/minute [88].
f. Reviews of PCB Analysis
Published reviews have covered the extraction, cleanup, and
chromatographic detection and quantitation of PCB's and other
non-pesticide environmental pollutants such as chlorinated
naphthalenes, dibenzofurans, and dibenzodioxins [89], the
utilization of GC coupled with chemical ionization and electron
- 30 -
-------
impact MS for analysis of these compounds and pesticides [90],
the chromatography of PCB's [91], and PCB interference and
analysis in terms of the AOAC-FDA multiresidue method for
chlorinated pesticides [92].
9.A.D. REVIEWS OF ANALYTICAL METHODS FOR PESTICIDES
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 [93]. Organophosphorus insecticides and metabo-
lites, carbamate insecticides and metabolites, herbicides and metabolites,
insecticide synergists, chlorinated insecticides and congeners, and
fumigant residues have been covered. The determination of fungicide
residues [94, 95] has also been reviewed.
7.A.E. REFERENCES
[1] Griffith, F. D., and Blanke, R. V., J. Assoc. Offic. Anal. Chem.,
57_, 595 (1974).
[2] Official Methods of Analysis, (1970), llth Edition, AOAC, Washington,
D.C.
[3] Krause, R. T., J. Assoc. Offic. Anal. Chem., 56, 721 (1973).
[4] Mills, P. A., Bong, B. A., Kamps, L. R., and Burke, J. A., J. Assoc.
Offic. Anal. Chem., 55_, 39 (1972); FDA PAM, Section 252.
[5] Gabica, J., Watson, M., and Benson, W. W., J. Assoc. Offic. Anal.
Chem. , 57_, 173 (1974) .
[6] Yip, G., J. Assoc. Offic. Anal. Chem., 54, 966 (1971).
[7] de Faubert Maunder, M. J., Egan, H., Godley, E. W., Hammond, E. W. ,
Roburn, J., and Thomson, J., Analyst, 89, 168 (1964).
[8] Wood, N. F., Analyst, 94, 399 (1969).
[9] Smart, N. A., Hill, A. R. C., and Roughan, P. A., J. Assoc. Offic.
Anal. Chem., 57, 153 (1974).
[10] Holden, A. V., and Marsden, K., J. Chromatogr., 44, 481 (1969).
[11] Stalling, D. C., Tindle, R. C., and Johnson, J. L., J. Assoc.
Offic. Anal. Chem., 55, 32 (1972).
- 31 -
-------
[12] Griffitt, K. R., and Craun, J. C., J. Assoc. Offie. Anal. Chem.
57_, 168 (1974) .
[13] Abbott, D. C., and Thomson, J., Chem. Ind. (London), p. 481, (1964);
Analyst, 89, 613 (1964).
[14] Calderbank, A., Residue Reviews, 12, 14 (1966).
[15] Renberg, L., Anal. Chem., 46, 459 (1974).
[16] Shafik, T. M., Sullivan, H. C., and Enos, H., J. Agr. Food Chem.,
21_, 295 (1973).
[17] Wessel, J. R., J. Assoc. Offic. Anal. Chem., 50, 430 (1967).
[18] McCully, K. A., J. Assoc. Offic. Anal. Chem., 55, 291 (1972).
[19] Pardue, J. R., J. Assoc. Offic. Anal. Chem., 54, 359 (1971).
[20] Beckman, H., and Barber, D., J. Assoc. Offic. Anal. Chem., 52^, 286
(1969).
[21] Versino, B. , van der Venne, M. Th., and Vissers, H., J. Assoc. Offic.
Anal. Chem., 5_4, 147 (1971).
[22] Kim, J. J. S., and Wilson, C. W., J. Agr. Food Chem., 14, 615 (1966).
[23] Storherr, R. W., Ott, P., and Watts, R. R., J. Assoc. Offic Anal.
Chem. , 54_, 513 (1971) .
[24] Laski, R. R., J. Assoc. Offic. Anal. Chem., 57, 930 (1974).
[25] Bowman, M. C., Beroza, M., and Leuck, D. B., J. Agr. Food Chem.,
16_, 796 (1968) .
[26] Sissons, D. J. , and Telling, G. M. , J. Chromatogr. , 47,, 328 (1970).
[27] Renvall, S., and Akerblom, M., Residue Reviews, 34, 1 (1971).
[28] Abbott, D. C., Crisp, S., Tarrant, K. R. , and Tatton, J.O'G.,
Pestic. Sci., .L, 10 (1970.
[29] Kadoum, A. M. , Bull. Environ. Contain. Toxicol. , 3_, 247 (1968) .
[30] Ripley, B. D., Wilkinson, R. J. and Chau, A. S. Y., J. Assoc. ,
Offic. Anal. Chem. , 57_, 1033 (1974).
[31] Kadoum, A. M. , Bull. Environ. Contam. Toxicol. , 3_, 65 (1968) and
2, 264 (1967).
- 32 -
-------
[32] Bowman, M.C., and Hill, K. R., J. Agr. Food Chem., 19, 342 (1971).
[33] Pflugmacher, J., and Ebing, W., J. Chromatogr., 93, 457 (1974).
[34] Bowman, M. C., and Beroza, M., J. Assoc. Offic. Anal. Chem., 52,
1231 (1969).
[35] Sherma, J., Critical Reviews in Analytical Chemistry, (CRC Press)
3_ (3) , pp. 299-354 (1973) .
[36] Strother, A., J. Gas Chromatog. , 6_, 110 (1968).
[37] Storherr, R. W., J. Assoc. Offic. Anal. Chem., 55, 283 (1972).
[38] Cook, R. F., Stanovick, R. P., and Cassil, C. C., J. Agr. Food Chem.,
r?, 277 (1969).
[39] Aue, W. A., Hastings, C. R., and Kapila, S., J. Chromatogr., 77,
299, (1973).
[40] Lorah, E. J., and Hemphill, D. D., J. Assoc. Offic. Anal. Chem.,
_57, 570 (1974).
[41] McKone, C. E., and Hance, R. J., J. Chromatogr., 36, 234 (1968).
[42] Onley, J. H., and Yip, G. , J. Assoc. Offic. Anal. Chem ., 54, 1366
(1971).
[43] Thier, H. P., Angew. Chem. Int. Ed., 10, 860 (1971).
[44] Williams, I. H., Residue Review? 38, 1 (1971).
[45] Cochrane, W. P., and Purkayastha, R., Toxicological and Environmental
Chemistry Reviews, .L, 137 (1973) .
[46] Ramsteiner, K., Hermann, W. D., and Eberle, D. 0., J. Assoc. Offic.
Anal. Chem., 57_, 192 (1974).
[47] Sherma, J. , and Shafik, T. M. , Arch. Env. Contam. Toxicol. , 3_, 55
(1975).
[48] Analysis of Pollutants-Proposed Test Procedures, Federal Register,
Vol. 38, 125, Part II, 29 June, 1973.
[49] Engst, R., and Knoll, R., Nahrung, 17, 837 (1973).
[50] Johnson, L. G. , Bull. Environ. Contam. Toxicol., 5_, 542 (1970).
[51] Leoni, V., J. Chromatogr., 62, 63 (1971).
- 33 -
-------
[52] Law, L. M., and Goerlitz, D. F., J. Assoc. Offic. Anal. Chem., 53,
1276 (1970).
[53] Schutzmann, R. L., Woodham, D. W., and Collier, C. W., J. Assoc.
Offic. Anal. Chem. , 54_, 1117 (1971) .
[54] Goerlitz, D. F., and Law, L. M., Bull. Environ. Contain. Toxicol. ,
6_, 9 (1971).
[55] Devine, J. M., and Zweig, G., J. Assoc. Offic. Anal. Chem., 52, 187
(1969).
[56] 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.
[57] Pionke, H. B., and Chesters, G., Soil Sci. Soc. Am. Proc., 32, 749 (1968)
[58] Goerlitz, D. F., and Law, L. M., J. Assoc. Offic. Anal. Chem., 57, 176
(1974).
[59] Nash, R. G., Harris, P. D., Ensor, P. D., and Woolson, E. A., J. Assoc.
Offic. Anal. Chem., 56, 728 (1973).
[60] Official Methods of Analysis, (1970), llth Edition, AOAC, Washington,
D.C., Sections 29.014 and 29.017.
[61] Woolson, E. A., J. Assoc. Offic. Anal. Chem., 57, 604 (1974).
[62] Smith, A. E., J. Chromatogr. , 97_, 103 (1974).
[63] Bartha, R., J. Agr. Food Chem., 16, 602 (1968).
[64] McKone, C. E., J. Chromatogr., 44, 60 (1969).
[65] Cohen, I. C., and Wheals, B. B., J. Chromatogr., 43, 233 (1969).
[66] Baunok, I., and Geissbuhler, H. , Bull. Environ. Contain. Toxicol. , 3^,
7 (1968).
[67] Abbott, D. C., Bunting, J. A., and Thomson, J., Analyst, 90, 357 (1965).
[68] Jolliffe, V. A., Day, B. E., Jordan, L. S., and Mann, J. D., J. Agr.
Food Chem., 15_, 174 (1967) .
[69] Smith, A. E., and Fitzpatrick, A., J. Chromatogr., 57, 303 (1971).
[70] Sarnj L. P., and Webster, G. R. B., J. Assoc. Offic. Anal. Chem.,
57, j,,79 (1974).
- 34 -
-------
[71] Thier, H. P., Angew. Chem. Int. Ed. , 13^ 217 (1974).
[72] Armour, J. A., and Burke, J. A., J. Assoc, Offic. Anal. Chem.,
53_, 761 (1970) .
[73] Lieb, A. J., and Bills, D. D., Bull. Environ. Contain. Toxicol., 12
328 (1974).
[74] Armour, J. A., and Burke, J. A., J. Assoc. Offic. Anal. Chem., 54,
175 (1971).
[75] Erney, D. R., J. Assoc. Offic. Anal. Chem., 57, 576 (1974).
[76] Welborn, M. E., Borchard, R. E., Hansen, L. G., and Metcalf, R. L.,
J. Assoc. Offic. Anal. Chem., 57, 1248 (1974).
[77] Armour, J. A., J. Assoc. Offic. Anal. Chem., 56, 302 (1973).
[78] Collins, G. B., Holmes, D. C., and Jackson, F. J., J. Chromatogr.,
71, 443 (1972).
[79] Zobel, M. G. R., J. Assoc. Offic. Anal. Chem., 57, 791 (1974).
[80] Eichelberger, J. W., Harris, L. E., and Budde, W. L., Anal. Chem.,
46_, 227 (1974).
[81] Zelenski, S. G., Tashiro, J., Worthen, L. R., and Olney, C. E.,
J. Chromatogr., 84, 67 (1973).
[82] Mulhern, B. M., Cromartie, E., Reichel, W. L., and Belisle, A. A.,
J. Assoc. Offic. Anal. Chem., 54, 548 (1970).
[83] 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.
[84] Young, S. J. V. , and Burke, J. A., Bull. Environ. Contam. Toxicol.,
7_, 160 (1972) .
[85] Holmes, D. C., and Wallen, M., J. Chromatogr., 71, 562 (1972).
[86] Fehringer, N. V., and Westall, J. E., J. Chromatogr., 57, 397 (1971).
[87] de Vos, R. H. , and Peet, E. W. , Bull. Environ. Contam. Toxicol., 6_, 164
(1971).
[88] Dolan, J. W., Hall, R. C., and Todd, T. M., J. Assoc. Offic. Anal.
Chem. , _55_, 537 (1972).
[89] Sherma, J., Advances in Chromatography, Volume 12, Marcel Dekker,
Inc., 1975.
- 35 -
-------
[90] Oswald, E. O., Albro, P. W., and McKinney, J. D., J. Chromatogr.,
98, 363 (1974); Oswald, E. O., Levy, L., Corbett, B. J., and Walket,
M. P., J. Chromatogr., 93, 63 (1974).
[91] Fishbein, L., J. Chromatogr., 68, 345 (1972).
[92] Burke, J. A., J. Assoc. Offic. Anal. Chem., 55, 284 (1972).
[93] Bowman, M. C., editor, J. Chromatog. Sci., 13, June and July issues
(1975).
[94] Sherma, J., J. Chromatogr., 113, 97-137 (1975).
[95] Baker, P. B., and Hoodless, R. A., Pestic. Sci., 5, 465 (1974).
- 36 -
-------
Section 8
CONFIRMATORY PROCEDURES
8. A. REQUIREMENTS FOR POSITIVE CONFIRMATION OF PESTICIDE IDENTITY
One of the major tasks facing the pesticide analyst is to obtain con-
vincing identification of a trace residue. The identity of pesticide
residues should always be confirmed by a different method from that
used in the initial determination since interpretation of results
(e.g., decisions of a legal or health nature) as well as reliable
quantitation (selection of standards) depend upon correct identifica-
tion. 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 PCB
compounds, which are easily mistaken for pesticide residues such as
p_,p_'-DDE and p_,p_'-DDT. Another example, of importance when screening
foods for tolerance levels, involves the fact that a 4% SE-30/6% QF-1
column may give essentially identical peaks for endrin and o_,p_'-DDT,
for Endosulfan I and p_,p_'-DDE, and for 6-BHC and lindane. Both the
DDT and DDE are very common pesticides with rather high tolerance
levels. Thus, if the analyst is uanaware that endrin and endosulfan
may produce corresponding GC responses, he may conclude that observed
peaks indicate only insignificant quantities of DDT and DDE relative
to tolerance levels and do no further work on the sample. However,
what appears to be insignificant response for DDT and DDE is very
substantial response for endrin and Endosulfan I because of lower GLC
response characteristics of these compounds and lower permitted
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, but 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 upon 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.
-------
Considerations of set theory [2] indicate that three independent
"equivocal" results are required in order to be confident of the identi-
ty of a pesticide residue. These might be eluation in a certain
fraction from a liquid chromatography cleanup column, a GC retention
time, and positive response of a selective GC detector. Another possible
combination that would be basis for confidence is the GC retention times
from a polar column and a nonpolar column plus a R value from PC or
TLC or an extraction p-value. Still another would be a GC retention
time, a PC or TLC R, value, and the GC retention time of a derivative
formed by a chemical or photochemical reaction.
The dependence or independence of measured values was studied by Elgar
who reported [3] that many widely used confirmatory methods may not give
truly independent evidence of identity since they are measuring the same
physical property, and that care must be taken when deciding which
methods to use in combination so that a great deal of work is not done
without gaining additional useful information. Examples of highly cor-
related (not independent) values include GC retention times on certain
stationary phases (Figure 5-N in Section 5); PC or TLC R values using
certain adsorbent/solvent systems; p_-values in different solvent pairs;
and PC, TLC, and p-values. These combinations will not provide inde-
pendent 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 R^ 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 GC 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 alternate 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 p_-values, C = PC vs TLC, D = TLC vs GC
Extraction p-value v
r^ l.u
:= 0 8
t
SO 6
V
t>
<
>> 0 4
c
f-
X
I02
O
O
o
, . o
f/''J
C' >
( ^
C9o
^/
0 02 04 OG 08 1
fsooc[;:ne 'D .'f
MH
1 0
.§08
i;
|06
<
^o 0. 4
c
re
>c
£0.2
0
thin-layer chromatog
O
00
o
o
o
0
0 0204 06 0310
Silica gel/H«.ine (R-)
Paper clirornatotjraphy v
thin-layer chromatography
GLC v TLC
E '-u
"c
g 0 8
£"06
o
c
O
£ 0 4
o
c
I02
o
"
° 0 0
0 °
0 °
o
0
0 02 04 06 08 1.0
Silicagsl/Mexano (Rf)
b.u
5.0
4.0
3.0
2.0
1.0
0
DC 200
O
o
o o °
0 o
0
: °°
02 04 0.6 08 1"
I 0
Silica3ol/Hex?.ne (R(.)
-------
The good use of common sense should not be overlooked when making
pesticide identifications. Examples of misapplied common sense include
reporting of 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 as body metabolism should normally
convert it to heptachlor epoxide. Chromatography of human adipose tissue
from the general population using EC detection often produces peaks with
retention characteristics very close or identical to the RRT values for
a-BHC and/or p_,p_'-DDE. However, the presence of these compounds has
rarely, if ever, been confirmed. In these instances, the peaks in
question represent artifacts which happen to have the same retention
times as these pesticides, and careful confirmation by ancillary tech-
niques would provide the proper identification.
8.B GC RELATIVE RETENTION TIMES
Certain guidelines are useful for the proper utilization of retention
times in making compound identifications.
a. The use of relative retention rather than absolute retention is
more reliable (Subsection 5.N. in Section 5).
b. Be highly suspicious of any peak whose calculated relative retention
value (RRT) 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 you use cleanup 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
alternate 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 obtaining 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 [4] of the utility of at least two different GC
columns for sample diagnosis include the following. Identity of certain
early eluting BHC isomers, particularly the alpha isomer, may be
-------
hindered by the presence of hexachlorobenzene (HBC). The latter is co-'
eluted with a-BHC on silicone columns and with 3-BHC on apiezon, but all
three compounds are resolved on a pclar cyano-silicone column. Dieldrin
and p_,p/ -DDE are difficult to resolve on a number of single phase *
silicone columns but are separated on Apiezon, cyano-silicone, and
trifluoromethyl silicone (QF-1, OV-210, SP-24Q1). 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
p,p_'-DDE, to the extent of about xl.4 at 180°C column temperature.
8.C. 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 chroma-
tograms are run using one or more of the highly selective detectors.
The microcoulometric (MC), flame photometric, or conductivity detectors,
described in Subsections 5.D., 5.F., and 5.G., 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 con-
clusive than is positive response. For example, if a peak on an electron
capture chromatogram suspected of being a chlorinated pesticide does not
show up when 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.
If the peak does show up in the MC or conductivity chromatogram, this
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 some other halogenated contaminant.
Because of the selectivity of its filters, the flame photometric detec-
tor (FPD) greatly simplifies confirmation of sulfur- and/or phosphorus-
containing residues. Identification of a thiophosphate is unequivocally
accomplished 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 (526 mm) to phosphorus (394 mm) response ratio of the FPD
matches the standard (Subsection 5.E.)
The MC detection system (Subsection 5.D.) provides additional evidence
of identity for compounds containing halogen, sulfur, or nitrogen by
use of appropriate specific titration cells. The detector has rela-
tively low sensitivity, therefore requiring a greater extent of extract
concentration, injection of larger samples, and/or a larger initial
sample than for the EC detector.
-------
8.D. THIN LAYER CHROMATOGRAPHY (TLC) R VALUES
F
Experimental aspects of TLC and its use for screening and quantitation
of residues were covered earlier in Subsections 6.Q. through 6.T. 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 detec-
table 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 coextractives
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 R values
for making confirmations because differences in development conditions
from run to run will cause these values to be non-reproducible.
Standards and samples should always be run on adjacent areas of the
same plate if at all possible. If R values must be used, the value
relative to the R of a standard compound X run on the same plate (R
value) will be more reliable than the absolute R 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.
X
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 that can be achieved on the sample extract and the level of
detection. Oils and waxes will particularly interfere with TLC,
causing streaked zones and/or distorted R values which may completely
negate its value for confirmation. The 15 percent diethyl 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 being produced on the layer after spraying and oven heating
[EPA PAM, Section 6.2]. Identification of naturally fluorescent pesti-
cides is aided by heating the chromatogram, causing specific alterations
in recorded spectra [5]. This heating procedure may, however, increase
background fluorescence from co-extracted compounds also present in the
sample. TLC after fluorigenic labelling [6] of pesticide residues is a
combination of chromatography with chemical derivatization (Subsection
-------
8.G.) which can provide very specific detection of certain residues. If
sufficient pesticide is present in the thin layer spot, scraping, collec-
tion of the adsorbent, and elution of the compound followed by mass
spectrometry (Subsection 8.L.) 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 R, values for identification purposes
was studied by Connors [7], who found that useful, uncorrelated data can
be obtained in several ways, as by pairing aqueous with nonaqueous
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 upon 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 nec-
essarily 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. Where available, the latter appears to be
the preferred procedure.
8.E. HIGH PERFORMANCE LIQUID CHROMATOGRAPHY (HPLC)
See Subsections 6.N. through 6.P. in Section 6 for a discussion of this
topic. HPLC has been used mainly for quantitation of residues in situ-
ations where GC is either not applicable or convenient to use. A HPLC
retention time can serve as evidence to confirm GC in the same way as a
PC or TLC R^ value. The liquid chromatographic system should be care-
fully 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.
8.F. EXTRACTION p_-VALUES
Extraction p_-values [8-12] are a tool for identifying pesticides at the
low ng level. The p_-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 de-
rived 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 is required, and sensitivity
is at the level of ED-GC.
-------
Details including experimental procedures, formulas for calculating
p_-values and the fractional amount extracted after repeated extractions,
graphs for determining specificity in a giyen system, and p_-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, refer-
ence [9], and Section 12.C of the EPA PAM (data for 88 pesticides in the
latter). A device and method for determining p-values with unequili-
brated solvents or unequal phase volumes are given in the FDA PAM, Section
622.1 and reference [12].
As mentioned earlier, the general technique of determining p-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 is not
correlated, it is best to use either a PC or TLC R value or an extrac-
tion p_-value as one independent criterion of identity. The great
advantage of jo-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.
8.G. DERIVATIZATION (CHEMICAL REACTION) TECHNIQUES
Derivatives of pesticides are prepared for various reasons, as to de-
crease volatility or increase detectability for TLC; increase volatility,
stability, and/or detectability and avoid tailing peaks for gas chroma-
tography; and to alter the structure in order to aid characterization.
It is this latter topic that will be discussed in this subsection.
Comparison of retention times on a given GC stationary phase before and
after chemical derivatization is a relatively recent innovation which
is becoming increasingly important for corroboration of residue identity.
Desirable characteristics of any chemical derivatization technique in-
clude:
a. Formation of a product with at least as much or hopefully more re-
sponse 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 and employ highly pure
reagents and solvents.
d. A cleanup method should be available to remove any background in-
terferences introduced by the reaction.
e. If product structures and reaction mechanisms and limitations are
identified, misidentifications can be avoided since the analyst can
elucidate the extent and probably sources of error in the procedure.
-------
f. Sensitivity should be at least in the Q.01 to 0.1 ppm range in terms
of the parent pesticide, which is lowej: than the established tolerance
values for most pesticides.
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 [13] . Another review of
reactions for chlorinated pesticides is found in reference [3].
It must be realized that these reactions destroy some pesticides (and
artifacts) in addition to forming pesticide derivatives.
-------
Table 8-1
Pesticide
CONFIRMATORY DERIVATIZATION TESTS FOR
PESTICIDE AND METABOLITE RESIDUES [13]
Reaction Utilized
DDT
DDE
ODD
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
-------
Table 8-1 Continued
Parent Pesticide
Heptachlor
trans-Chlordane
cis-chlordane
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 Utilized
1} Allylic hydrogen
2) Double bond
1) Allylic hydroxy
2) Double bond
1) Hydroxyl
2) Epoxide
Double bond or
gem-dichloro group
Chloro epoxide or
Ti-dichloro group
n-Dichloro
Hydroxyl
Derivative
1-Bromochlordene
Chlordene epoxide
Silyl ether
Chloroacetate;
epoxide
Silyl ether
Trihydroxy
chlordane
Epoxide or
hexachloro
Chloroacetate
or heptachloro
Pentachloro
Acetate or silyl
ether
10
-------
Degradation products arising from UV treatment of chlorinated insecticides
detected by EC-GC can provide identification of these pesticides 114] at
75-100 pg levels. Depending on the length of irradiation (often ca. 1,0
minutes), all of the parent pesticide may not be degraded. Solvent and
sample blanks should be run to prove if conversion of background occurs
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
treatment dehydrochlorination method for use in multiresidue analysis.
This procedure produces derivatives for identity confirmation and pro-
vides supplemental cleanup for some troublesome extracts after Florisil
chromatography. Alakali reactions carried out on a GC pre-column rather
than in solution have proven advantageous in some instances [15].
Section 11 of the Canadian PAM gives complete details for the following
tests:
Pesticide(s)
o_, £_'-DDT, p_,p_'-DDT, p_,£'-TDE,
methoxychlor
p,p_'-DDT, endrin
dieldrin, endrin
chlordane, heptachlor
epoxide
aldrin, heptachlor,
p_,p' -DDE
aldrin
endosulfan
chlorophenoxy acid
herbicides
captan
Reagent
sodium methylate
chromous chloride
BC1 in 2-chloroethanol
K-tert butoxide/tert-butanol,
silylation
chromic acid
m-chloroperbenzoic acid
alcoholic KOH
n-propanol
resorcinol
A special two stage, mixed phase 6 foot column consisting of 5 foot
6 inches of 4% OV-1/6% QF-1 and 6 inches 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. All these
are not resolved on the 4% SE-30/6% QF-1 working column.
One method for differentiation of PCB's and organochlorine pesticides
is by observing the reactivity of the residues to HNO - H SO..
Organochlorine pesticides are destroyed whereas PCB's (and toxaphene)
11
-------
are unaffected. Other confirmation methods for PCB's are covered in
Subsections 7.A.A.-7.A.C. in Section 7.
Recent studies of confirmatory reactions have been made for the chlori-
nated pesticides endosulfan 116-18], dieldrin [19], heptachlor [18, 20,
21], endrin [21], and HCB [reference 22 and Section 12,D,(2) of the EPA
PAM] .
b. OTHER PESTICIDE CLASSES
Residues of organophosphorus pesticides may be confirmed by alkaline
hydrolysis followed by esterification of the resulting dialkylphosphates
to trialkylphosphates [23]. This procedure does not distinguish between
two pesticides that produce the same hydrolysis product. Other reactions
used to identify organophosphorus pesticides include oxidation with
KMNO [24] or m_-chloroperoxybenzoic acid [25] to corresponding sulfones,
reduction with chromous chloride or Zn/HCl [26], derivatization with
pentafluorobenzyl (PFB) bromide to produce PFB ethers of phenolic
hydrolysis products [27] , and oxidation by neutralized sodium hypochlo-
rite to corresponding oxons (oxygen analogs) [28]. The products are
normally gas chromatographed using a FPD or thermionic detector.
Triazine herbicides have been confirmed by silylation, methoxylation
(in sodium methoxide-methanol), methylation (CH I - NaH), and hydroly-
sis - DNFB reactions [26, 29], and linuron also by alkylation (with
alkyl halide - NaH) [26].
8.H. SPECTROMETRY (SPECTROPHOTOMETRY)
Spectrophotometric methods for residue determination (quant:itation)
usually do not achieve the sensitivity or selectivity of GC or TLC, and
this is why they are not so widely used as in the early days of pesti-
cide analysis before chromatographic methods were developed. The
applicability of spectrometry is especially limited for multiresidue
determinations or analyses of a parent compound, metabolites, and hydrol-
ysis products.
Spectrometry can be very valuable, however, in conjunction with chroma-
tography as a confirmatory tool, and it is this aspect that will be
stressed in the following subsections.
8.1. 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.
12
-------
The correlation between UV spectra and pesticide structure and the use-
fulness of UV spectroscopy in confirming identification have been
reviewed I30]. 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, as IR, NMR, and MS. In i>ome cases, extinction coefficients
(absorptivities) are sufficiently large to permit identifications at
submicrogram levels. Since 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 im-
poses 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 achieve
detection of pesticides by HPLC UV detectors and in TLC.
If a pesticide is naturally fluorescent or can be made fluorescent by
derivatization, fluorescence spectrophotometry is likely to be more
selective and sensitive than either visible or UV absorption methods.
Concurrence of fluorescence excitation and emission spectra between
samples and standards, recorded either in solution or directly on thin
layer chromatograms, has served as a valuable confirmatory aid for
certain pesticides. Fluorescence characteristics are dependent on a
number of experimental conditions which must be closely controlled,
e.g., solvent and pH effects [31]. Removal of naturally occurring
fluorescent interferences from biological samples can pose serious
cleanup problems.
A survey of the phosphorescence of 32 pesticides and presentation of
excitation and emission spectra, decay times, analytical curves, and
detection limits have been made [32]. Laser-excited fluorescence and
Raman spectra 01 organophosphate, carbamate, and chlorinated pesticides
have been measured [33].
8.J. INFRARED (IR)
IR spectroscopy with micro sampling techniques is generally sensitive at
the 1 yg level but has been used down to the 0.1 yg level in some appli-
cations. 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, collected, and the pesticide 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 going to the detector for monitoring purposes
while the remainder goes to a collecting device.
13
-------
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 employing 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. Just this very tip is
then cut or broken from the wedge and pressed into a micro-pellet.
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 inconsequen-
tial 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 re-
agents and carefully cleaned equipment should be used. Inevitable
losses due to handling and processing require starting the isolation
procedure 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 [34] , including discussion
of micro multiple internal reflectance. Advantages of internal reflectance
include ease of applying (by dotting or streaking) sample to the surface
of the reflectance plate (crystal), minimizing of interferences from
handling and reagents, and ease of recovery after IR evaluation (samples
made into pellets are essentially lost for further scrutiny). A disadvan-
tage 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 identification of Thiram
residues at 0.1 ppm on lettuce after extraction, Florisil chromatography,
and TLC [35]. An alternate micro KBr technique with sensitivity levels
similar to the method in the EPA PAM is detailed in the FDA PAM, Section
631.
14
-------
Figure 8-B. EftuidRent for nreparation
of micro ICBr pellets (Dtioto courtesy
of R, C. 81inr!)
Fl^are S-C. Illystration of techntoue
of Curry, et al. (Photo courtesy of
R. C. Blinn)
A. S. Curry, J. F Read, C Brown, and '? W Jenkins
J Chromatog 38, 206 (1968)
STAINLESS STEEL \^ /__
HOLDER
STAINLESS
STEEL CAP
SAMPLE CONCENTRATED
AT TIP
GLASS VIAL
KBr "WICK-STICK"
VOLATILE SOLVENT
15
-------
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 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 coridensates 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 yg quantities for
IR) and the destructive nature of pesticide GC detectors.
Several different types of IR detectors directly coupled with gas chroma-
tographs 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:
a. Passing column effluent through solvent [36, 37].
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 on column packing [38, 39, 40]. This procedure is very
efficient, and fractions are easily collected for subsequent IR evalu-
ation; reagent interferences are possible.
e. Collection on a TLC plate for further cleanup prior to IR [41].
f. Trapping on Millipore or siliconized filter material [42, 43].
g. Various types of liquid nitrogen or dry ice cold surface traps [44].
h. A cool or cold small id tubing at the GC vent [45].
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 it is
free of sources of interferences.
The choice of trapping procedure will depend on the amount of compound
available, IR technique to be employed, purity of the compound eluted
from the GC column, and equipment available to the chemist.
16
-------
IR and UV spectra of 76 reference pesticides have been published [46J to
aid the analyst in matching spectra of unknown pesticides.
8.K. 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 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 sensitivity of C is
lower; with current commercial instrumentation, a practical sample size
is greater than 20 mg, although C spectra of as little as 300 yg have
been obtained on modified instruments [47] . Useful information is pro-
vided by NMR in many areas relevant to the analysis of pesticides, their
metabolites, and degradation products, such as identification and
structural characterization, molecular geometries, conformations and
stereochemistry, chemical kinetics and equilibria, complex formation and
binding, and electronic charge distributions.
Residues of p_,p_'-DDT and p_,p_'-DDE isolated from adipose and liver tissue
samples have been analyzed by NMR [48], with semi-quantitative determi-
nation 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 p_,p'-DDT [49, 50], p,p_'-DDA [51],
aldrin and dieldrin derivatives and other chlorinated pesticides [52],
rotenoids [53], and dithiocarbamates [54]. Other H reference spectra
of organophosphorus [55], diphenylmethane (DDT type) [56], and carbamate
[57] pesticides have been published and are useful for identity con-
firmation. The application of NMR to pesticide analysis has been re-
viewed [58] .
Carbon-13 NMR spectra have been published for cx-BHC [59] and for
several chlorinated biphenyls [60, 61]. Chlorine nuclear quadrupole
resonance spectrometry has been employed to study the structures of
several chlorinated pesticides including BHC, aldrin, endrin, endosulfan,
and dieldrin [62-64] , and P-NMR chemical shifts have been correlated
with structures of some organophosphorus pesticides [65].
8.L. MASS SPECTROMETRY (MS)
The mass spectrometer is a very sensitive spectroscopic tool for pesti-
cide residue analysis, providing useful data on less than 10 ng of
material. Charged 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
abundances. A mass spectrum is usually quite characteristic of an
individual pesticide, sometimes even providing data which will
differentiate among geometric isomers. Pesticide identifications can be
made by matching a mass spectrum of the unknown sample with the mass
17
-------
spectrum of a known material. Inspection of key fragment ions in the
spectrum can also assist in elucidating the structure of the unknown
material. These procedures are feasible both with low and high resolu-
tion spectrometers, but the exact elemental compositions of the parent
ion and fragments provided by high resolution MS is a valuable aid to
the analyst.'
a. MS INSTRUMENTATION AND OPERATION
(1) Introduction
Four components are common to most mass spectrometers: the inlet system,
the ion source, the mass analyzer, and the detector and readout system.
In addition, a high vacuum must be maintained throughout the spectro-
meter from inlet to detector so that ions formed in the source will not
be lost from collisions with atmospheric gas molecules. A sample is
introduced to the ion source through an inlet system. There the sample
is ionized. The generated beam of ions is directed into the mass
analyzer where the ions are separated and focused according to their 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 which can be interpreted by the analyst.
(2) Inlet Systems: Combined GS-MS
Very clean samples may be directly introduced 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 furance 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 arrange-
ment. The sample temperature is then raised until a spectrum is obtained.
Techniques for trapping GC fractions for introduction into an independent
mass spectrometer have been published [39, 66-68] , 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 in-
troduction of samples into the spectrometer. The resolution provided by
gas chromatography offers extra sample cleanup in addition to any parti-
tion and liquid column chromatography steps. Temperature and flow rate
programming have proven useful for achieving high chromatography reso-
lution with the combined instrument. Column bleed can be a sei'ious
problem in GC-MS, since bleeding liquid phase is ionized .by the mass
spectrometer and contributes spurious ions to the analytical spectra.
Columns stable at high temperatures should be used whenever possible.
Compatability of the gas chromatograph and mass spectrometer is a
problem because of the large volume of carrier gas eluting from the
former and the need to operate the latter at high vacuum (lo - 10
torr). The simplest approach is to connect the two instruments directly
18
-------
and to use a large pumping system to maintain the required vacuum in the
mass spectrometer. This approach has been used successfully with capil-
lary GC columns having flow rates up to ca. 20 ml/minute. Introduction
of samples from packed columns into the mass spectrometer requires that
most of the gas be diverted in an interface between the two instruments
so that only a small fraction actually enters the source. The interface
should divert the carrier gas and not the sample, so that the gas flowing
into the spectrometer is enriched in sample. Several designs and many
variations of frit, jet, and membrane type molecular separators or
sample-enriching devices have been reported for this purpose, each with
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. The mass spectrometer in a combined
instrument must be able to scan through an appropriate mass range, from
mass 10 to mass 800, in less time than it takes to elute the peaks from
the gas chromatograph.
(3) lonization Source
The most widely used ionization source is the electron impact type
wherein gaseous molecules are ionized by energetic electrons emitted
from a glowing filament. These positive ions are accelerated and
focused by a high voltage field maintained between a pair of positively
charged slits.
Electron-impact ionization can cause cleavage of molecules to such an
extent that the molecular ion is absent from the mass spectrum or of
very low intensity. Chemical ionization (CI) spectra are obtained by
adding methane or isobutane (at relatively high pressures of about
1 mm Hg) to the GC effluent before it enters the source or in the source
itself. Electrons produce methane or isobutane reagent gas ions which
react further with neutral sample molecules. The mass spectra obtained
with CI are often quite different than those formed on electron impact
and are in general simple, easy to interpret, and complementary to
electron impact spectra for unequivocal pesticide confirmation. Al-
though CI often provides molecular ion (or M +1 or M -1) peaks of high
intensity, a recent study [69] of a series of chlorinated and phosphate
pesticides found no molecular ion produced from electron impact, CI, or
field ionization. The positive and negative methane [70] and isobutane
[71] CI mass spectra of selected chlorinated insecticides of several
types and methane positive ion CI data for 29 organophosphorus insecti-
cides and metabolites [72] have been determined and published.
Field ionization (FI) occurs when a molecule is in close proximity to a
carbon needle wire in the presence of a high voltage electric field.
This is a relatively low energy ionization method which often produces
enhanced molecular ion intensities and a cleaner spectrum for compounds
with poor thermal stability. The FI MS of a number of pesticides has
been studied [73]. Field desorption MS is a modification of FI in
which the wire anode is coated with the sample. Field desorption has no
requirement that the compound be volatile and produces minimal
19
-------
fragmentation. Strong molecular ion peaks are produced for most pesti-
cides 169] including highly polar pesticide metabolites [74] .
A disucssion 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 [75].
(4) 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 slit. Such analyzers
are reffered to as single- or direction-focusing analyzers.
Quadrupole analyzers are based on mass separation in a quadrupole radio
frequency electric field. This field is established on a set of four
precision parallel rods, with both a d.c. voltage and a RF alternating
voltage being applied to these electrodes. By varying the RF component
of the field, ions of various masses inserted into the field undergo
stable oscillations and pass through the length of the analyzer tube to
a detector. At a specific value of the d.c. and RF voltages, ions with
only a certain m/e ratio can go through the analyzer tube without
striking an electrode. The spectrum is obtained by sweeping the applied
RF alternating frequency and measuring the detector current as a function
of time.
(5) Resolution
The term 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 reso-
lution 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 Lons with
very little differences in mass and obtaining the molecular ion and
fragment ion masses accurately to 0.001 units. These high resolution
instruments have an analyzer region with an electrostatic sector for
velocity focusing plus a magnetic sector for separation of fragments
according to m/e ratio. 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 [76]. Once the exact mass of a
key ion (often the molecular ion) is known, the elemental composition
20
-------
or formula of the fragment is obtained again using a computer or by don-
suiting 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.
Descriptions of MS instruments and techniques are beyond the scope of
this Manual References [66, 77, 78] review methods and applications of
MS and combined GC-MS to pesticide residue analysis, and references
[79, 80] give a more general survey of GC-MS instrumentation and prin-
ciples.
b. EXAMPLE OF GC-MS CONFIRMATION
Figure 8-E shows the electron capture gas chromatogram obtained by in-
jection of an aliquot of the 6 percent ethyl ether Florisil column
eluate from cleanup of a human adipose tissue extract. Figure 8-F shows
the total ion current chromatogram of the same eluate from GC-MS. Al-
though the curves are drawn to different scales and are not directly
comparable, it is evident that the electron capture chromatogram does not
contain nearly the number of peaks as does the total ion chromatogram
because of the selectivity of the EC detector. In general, chromatograms
traced by the total ion monitor are similar, but not necessarily
identical, in response and sensitivity to that which would be 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. Figure 8-G is the mass spectrum
of the key p_,p_' -DDE GC peak evident in both chromatograms in Figures 8-E
and 8-F. The molecular ion (M ) peak at 316 units and the characteristic
fragments at 281, 280, and 246 would confirm the tentative identification
of this peak as p_,p_'-DDE obtained from comparing the retention time of
the GC peak to standard tables. Figure 8-H illustrates the total
spectrum of standard p_,p_'-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.
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 ether eluate.
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
21
-------
FIGURE 8-E ELECTRON CAPTURE CHROMATOGRAM OF HUMAN ADIPOSE
TISSUE EXTRACT, b% ETHER FLORISIL COLUMN ELUATE
FIGURE 8-F TOTAL ION CURRENT CHROMATOGRAM OF SAME HUMAN ADIPOSE
TISSUE EXTRACT
-8
S.
22
-------
FIGURE 8-G MASS SPECTRUM OF p,p'-DDE PEAK,
Ions below 200 not shown
100-
80-
60-
40-
20-
2
ions D«OW t\jv nor snown
i
1 , ,1 1
&-* l] ft' " -alo '"
M'=3toKCI
M'-C!-281 3O1
M'- 2CI = 246 2CI
MI46-CI = 211 1 Ct
L
350 400 45
FIGURE 8-H TOTAL MASS SPECTRUM OF p,p'-DDE
oTo
i^^UWMJ.^A^
in ' i' i' n i' i'
250 300
23
-------
the analyst know what compound or compounds he is looking for and are
not applicable to totally unknown samples.
In single ion monitoring, the spectrometer is focused on one ion known
to be characteristic of the compound or class of compounds sought. A
response is seen only when compounds which fragment to that particular
ion are eluted from the GC column. In most cases single ion monitoring
is more sensitive than total ion monitoring and is also more selective
because ions formed from background are not detected if their masses
are different from that being monitored. Subnanogram amounts of material
have been quantitated by constructing peak height vs. sample size cali-
bration curves. Identification of eluted species can be aided by moni-
toring the mass specified to several decimal places, but this is done
at the expense of sensitivity.
Multiple ion monitoring (MIM), also called mass fragmentography, in-
volves automatic, continuous, and simultaneous monitoring of several
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 detection of one or several compounds can be achieved with
characterization of each being based on the formation of one or more
selected ions [81]. To use MIM effectively, one should know the kind
of compound sought and its MS characteristics. Sensitivity of detection
is approximately 10 pg - 10 ng, considerably more sensitive than con-
ventional scanning because of the longer sampling time at each selected
mass. Sensitivity for a particular compound reflects the extent of
fragmentation and the fraction of the total ion current carried by the
selected ions. As an example, mass fragmentography has been applied to
some organophosphate insecticides [81].
Repetitive scanning through a narrow mass range generates quantifiable
spectral envelopes from several ions at once. This procedure, generally
sensitive at low ng levels, has been applied to pesticide analysis [83].
d. COMPUTERIZATION OF GC-MS
Combination of a computer with a GC-MS system can serve several very use-
ful functions:
(1) Column bleed and other background can be conveniently subtracted by
the computer.
(2) Continuous repetitive scans can be made during the entire chromato-
graphic separation, for example, a spectrum is scanned every 2-4 seconds.
All spectra are stored, and chromatograms may later be reconstructed by
the computer, summing and plotting the total ion current detected in
each scan but with exclusion of carrier gas ions or other interfering
ions. Chromatograms obtained resemble those traced by a conventional
total ion monitor.
24
-------
(3) The computer can trace only the intensities of selected character-
istic peaks from among the great mass of data acquired by continuous
repetitive scanning. The resulting mass chromatograms resemble the
selected ion profiles described earlier and permit compounds and spectra
of interest to be located and the appropriate spectrum to be retrieved
and plotted. Mass chromatography has an advantage over MIM 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 than MIM.
(4) Quantitation of peak areas in the selected ion profiles and ratios
of these peaks can be provided. Computer coupled GC-MS equipment is
extremely expensive, highly qualified personnel is needed for its opera-
tion and maintenance and to interpret data, and a significant amount of
"down-time" is to be anticipated because of its complexity.
e. APPLICATIONS OF GC-MS TO PESTICIDE ANALYSIS
Damico and co-workers have published reference spectra for pesticides of
several types [84] . Applications of GC-MS include confirmation of the
1-naphthyl chloracetate derivative of 1-naphthol (a carbaryl metabolite)
extracted from urine [85] 2,4-D, 2,4,5-T and 2,4,5-TCP in urine [86],
organophosphate pesticides in blood and urine [87, 88], and multiple
chlorinated insecticides in human adipose and liver tissue [89] or foods
[90]. An important application of GC-MS has been mutual determination
and identification of PCB's in the presence of chlorinated pesticides
[91]. Insecticides mixed with PCB's have been identified at levels
below 10 ng without complete separation on a GC column by peak monitor-
ing MS as described earlier [91].
8.M. BIOLOGICAL METHODS
Bioassay techniques, which include insecticidal activity, enzymatic, and
immunological methods, have been described as providing an independent
criterion of identity when combined with GC, chemical reactions, etc.
[2]. These methods, which depend upon the measurement of a physiological
response of a test organism induced by exposure to the pesticide, have
advantages of simplicity and sensitivity but are relatively non-specific
so that their utility for confirmation is rather poor. The insect
bioassay technique has been reviewed [93].
Specificity of enzyme inhibition is greatly enhanced by combination with
TLC for detection and confirmation of organophosphate and certain carba-
mate pesticides. The R value plus biological response provides impor-
tant identity information at levels typically in the range of 500 pg to
10 ng for these compounds.
8.N. POLORAGRAPHY (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
25
-------
and applications of polarography for both identification and determination
of pesticide residues have been reviewed 194, 95].
Polarographic identification of a pesticide residue is based on the de-j
termination 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, additions of the stan-
dard 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 identi-
fications based on retention times. Instrumentation for such modern
voltammetric techniques as fast sweep oscillography provide sensitivity
comparable to colorimetry. Pesticides not containing an oxidizable or
reducible functional group can be made amenable to polarography by for-
mation 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 Ccirbophenothion.
A collaborative study confirmed the usefuleness of a single sweep oscill-
ographic polarography for identifying such residues in non fatty foods
[96]. Thirty-eight herbicides have been studied by single sweep deriva-
tive polarography [97]/ methylcarbamate insecticides by AC polarography
and cyclic voltammetry [98], and urea herbicides by anodic polarography
[99]. Published voltammetric reduction potentials for some 100 organo-
chlorine insecticides/ PCB's, and naphthalenes (3 electrode: potentiostat,
DMSO solvent) are a possible aid in identification of residues [100].
8.0. MISCELLANEOUS CONFIRMATORY METHODS
a. CARBON SKELETON CHROMATOGRAPHY
Carbon skeleton chromatography (CSC) is useful in characterizing insec-
ticide residues in amounts down to 5-100 ng. Apparatus for CSC consists
of a precolumn containing a hot (ca. 300°C) catalyst attached to a 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 characteris-
tics relative to standards. This identification method, which is in
effect a derivatization procedure, 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, applications to many pesticide classes, and charac-
terization of products of CSC (as well as some other precolumn reaction
confirmatory methods) have been reported by Beroza and co-workers
26
-------
[101-104] and Asai et. al. [105, 106]. Identification of 5-10 ng
amounts of polychlorinated biphenyls, terphenyls, naphthalenes, dioxins,
and dibenzofurans in biological samples has been demonstrated 1107].
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 of fingerprint chromatograms helpful in making
identifications. A palladium catalyst at 300°C [105] and reagents such
Na CO , CuO, CdCl , A1C1 , and K^r^O at 24o°C [108] have been applied
to chlorinated and phosphate insecticides with EC detection of the reac-
tion products.
27
-------
8. P. 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., ]_, 472 (1971) .
[3] Elgar, K. E., The Identification of Pesticides at Residue
Concentrations, Advances in Chemistry Series 104, Chapter 10,
ACS, Washington, D. C., 1971, page 151.
[4] Ruzicka, J. H. A., and Abbott, D. C., Talanta, 20, 1277 (1973).
[5] Mallet, V., J. Chromatogr., 79, 217 (1973).
[6] Lawrence, J. F. , and Frei, R. W., Chromatogr. Rev., 18, 253
(1974).
[7] Connors, K. A., Anal. Chem., 46, 53 (1974).
[8] Beroza, M., and Bowman, M. C., J. Assoc. Offic, Anal. Chem.,
48, 358 (1965).
[9] Bowman, M. C., and Beroza, M., J. Assoc. Offic. Anal. Chem.,
4_8, 943 (1965) .
[10] Beroza, M., and Bowman, M. C., Anal. Chem., 37, 291 (1965).
[11] Beroza, M., and Bowman, M. C., Anal. Chem., 38, 837 (1966).
[12] Bowman, M. C., and Beroza, M., Anal. Chem., 38, 1427 (1966).
[13] 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.
[14] Banks, K. A., and Bills, D. D. , J. Chromatogr. , 33_, 450 (1968).
[15] Cochrane, W. P., and Maybury, R. B., J. Assoc. Offic. Anal. Chem.,
5_6, 1324 (1973) .
[16] Chau, A. S. Y., J. Assoc. Offic. Anal. Chem., 55, 1232 (1972).
28
-------
[17] Chau, A. S. Y., and Terry, K., J. Assoc. Offic. Anal. Chem.,
55_, 1228 (1972).
[18] Chau, A. S. Y., and Terry, K., J. Assoc. Offic. Anal. Chem.,
57_, 394 (1974) .
[19] Maybury, R. B., and Cochrane, W. P., J. Assoc. Offic. Anal.
Chem., 56, 36 (1973).
[20] Chau, A. S. Y., Demayo, A., Apsimon, J. W., Buccini, J. A.,
and Fruchier, A., J. Assoc. Offic. Anal. Chem. 57, 205 (1974).
[21] Chau, A. S. Y., J. Assoc. Offic. Anal. Chem., 57, 585 (1974).
[22] Baker, B. E., Bull. Environ. Contain. Toxicol. , 10, 279 (1973).
[23] Shafik, T. M., Bradway, D., and Enos, H. F., Bull Environ.
Contamn. Toxicol., 6_, 55 (1971).
[24] Thornton, J. S., and Anderson, C. A., J. Agr. Food Chem., 16,
895 (1968).
[25] Bowman, M. C., and Beroza, M., J. Assoc. Offic. Anal. Chem.,
52, 1231 (1969); Schutzmann, R. L., and Barthel, W. F., Anal. Chem.,
52_, 151 (1969) .
[26] Cochrane, W. P., and Greenhalgh, R. Confirmation of Pesticide
Residue Identity by Chemical Derivatization, presented at
166th National ACS Meeting, Chicago, Aug. 29, 1973; Khan, S.
U., Greenhalgh, R., and Cochrane, W. P., J. Agr. Food Chem.,
£3_, 430 (1975) .
[27] Coburn, J. A., and Chau, A. S. Y., J. Assoc. Offic. Anal. Chem.,
57, 1272 (1974).
[28] Singh, J., and Lapointe, M. R., J. Assoc. Offic. Anal. Chem.,
57, 1285 (1974).
[29] Lawrence, J. F., J. Agr. Food Chem., 22, 936 (1974).
[30] Aly, O. M., Faust, S. D., and Suffet, I. H., Ultraviolet
Spectrophotometry in Residue Analysis; Spectra-Structure
Correlations, Advances in Chemistry Series 104, Chapter 7,
ACS, Washington, D. C., 1971, page 95.
[31] MacDougall, D. , Residue Reviews, .1, 24 (1962); 5_, 119 (1964).
[32] Moye, H. A., and Winefordner, J. D., J. Agr. Food Chem., 13,
516 (1965).
29
-------
[33] Vickers, R. S., Chan, P. W., and Johnsen, R. E., Spectrosc.
Lett. , 6_, 131 (1973).
[34] Blinn, R. C., The Identification of Pesticides at Residue
Concentrations, Advances in Chemistry Series 104, Chapter 6,
ACS, Washington, D. C., 1971, page 81.
[35] Kettmann, R., Closset, J.-L., Copin, A., Duculot, C., and
Martens, P.-H., Anal. Lett., 6_, 1013 (1973).
[36] Crosby, N. T., and Laws, E. Q., Analyst, 89, 319 (1964).
[37] Neiswender, D. D., J. Gas Chromatog., 4_, 426 (1966),
[38] Shearer, D. A., Stone, B. C., and McGugan, W. A., Analyst,
88, 147 (1963).
[39] Bierl, B. A., Beroza, M., and Ruth, J. M., J. Gas. Chromatog.,
6_, 286 (1968) .
[40] Scanlon, R. A., Arnold, R. G., and Lindsay, R. C., J. Gas
Chromatog. , 6_, 372 (1968) .
[41] Nano, G. M., Sancin, P., and Martelli, A., J. Gas Chromatog.,
3_, 85 (1965) .
[42] Hannah, R. W., and Dwyer, J. L., Anal. Chem., 36, 2341 (1964).
[43] Hammerstrand, K., Juntunen, J. M., and Henes, A. R. , Anal. Biochem.
27, 172 (1969).
[44] Curry, A. S., Read, J. F., Brown, C., and Jenkins, R. W.,
J. Chromatogr., 38, 200 (1968).
[45] Hartman, K. T., J. Assoc. Offic. Anal. Chem., 50, 615 (1967).
[46] Gore, R. C., Hannah, R. W., Pattacini, S. C., and Porro,
T. J., J. Assoc. Offic. Anal. Chem., 54, 1040 (1971).
[47] Wilson, D. M., Olsen, R. W., and Burlingame, A. L., Rev. Sci.
Instrum., 45, 1095 (1974).
[48] Biros, F. J., J. Assoc. Offic. Anal. Chem., 53, 733 (1970).
[49] Wilson, N. K., J. Amer. Chem. Soc., 94, 2431 (1972).
[50] Ross, R. T., and Biros, J. F., Biochem. Biophys. Res. Commun.,
39, 723 (1970).
30
-------
[51] Ross, R. T., and Biros, F. J., Mass Spectrometry and NMR
Spectroscopy in Pesticide Chemistry, Plenum, N. Y., 1974,
pp. 263-272.
[52] McKinney, J. D., Wilson, N. K. , Keith, L. H., and Alford, A. L.,
iMass Spectrometry and NMR Spectroscopy in Pesticide Chemistry,
Plenum, N. Y., 1974, pp. 139-160.
[53] Crombie, L., and Lown, J. W., J. Chem. Soc., 1962, 775.
[54] Wilson, N. K., J. Phys. Chem., 75, 1067 (1971).
[55] Keith, L. H., Garrison, A. W., and Alford, A. L., J. Assoc.
Offie. Anal. Chem., 51, 1063 (1968).
[56] Keith, L. H., Alford, A. L., and Garrison, A. W., J. Assoc.
Offie. Anal. Chem., 52, 1074 (1969).
[57] Keith, L. H., and Alford, Al L., J. Assoc. Offic. Anal. Chem.,
5_3, 157 (1970).
[58] Keith, L. H., and Alford, A. L., J. Assoc. Offic. Anal. Chem.,
53, 1018 (1970).
[59] Yamamoto, 0., Yanagisawa, M., Hayamizu, K., and Kotowycz, G.,
J. Magn. Reson. , 9_, 216 (1973).
[60] Wilson, N.K., and Anderson, M., Mass Spectrometry and NMR
Spectroscopy in Pesticide Chemistry, Plenum, N. Y., 1974,
pp. 197-218.
[61] Wilson, N. K., J. Amer. Chem. Soc., 97, In press (1975).
[62] Roll, D. B., and Biros, F. J., Anal. Chem., 41, 407 (1969).
[63] Gegiou, D., Anal. Chem., 46, 742 (1974).
[64] Gegiou, D., Talanta, 21, 889 (1974).
[65] Ross, R. T., and Biros, J. F., Anal. Chem., 52, 139 (1970).
[66] Biros, F. J., Residue Reviews, 40, 1 (1971).
[67] Burson, K. R., and Kenner, C. T., J. Chromatog. Sci.,
7_, 63 (1969); Amy, J. , Chait, E. M. , Baitinger, W. E. , and
McLafferty, F. W., Anal. Chem. , 37_, 1265 (1965).
[68] Orland, R. K., Clock, E., and Bodenhamer, N. L., J. Chromatog.
Sci., 7, 187 (1969).
31
-------
[69] Fales, H. M. , et_._ aO._, Anal. Chem. , 47_, 207 (1975).
[70] Biros, F. J. , Dougherty, R. C., and Dalton, J. , Org. Mass.
Spectrom. , 6_, 1161 and 1171 (1972) .
[71] Dougherty, R. C., Roberts, J. D., and Biros, F. J., Anal. Chem.,
47_, 54 (1975) .
[72] Holmstead, R. L. , and Casida, J. E., J. Assoc. Offie. Anal. Chem.,
57_, 1050 (1974) .
[73] Damico, J. N., Barron, R. P., and Sphon, J. A., Int. J. Mass.
Spectrom. Ion Phys. , 2_, 161 (1969) .
[74] Schulten, H.R., and Beckey, H. D. , J. Agr. Food Chem. , 2_1_, 372
(1973).
[75] Milne, G. W. A., and Lacey, M. J., CRC Grit. Rev. Analyt. Chem.,
4_, 45-104 (1974) .
[76] Abramson, F. P., Anal. Chem., 44(14), 28A (1972).
[77] Oswald, E. 0., Albro, P.W., and McKinney, J. D., J. Chromatogr.,
98_, 363 (1974) .
[78] Safe, S., and Hutziner, 0., Mass Spectroscopy of Pesticides and
Pollutants, CRC Press, Cleveland, Ohio, 1973.
[79] Flath, R. E., Gas Chromatography-Mass Spectrometry, in Guide
to Modern Methods of Instrumental Analysis, Gouw, T. H., ed.,
Wiley-Interscience, N. Y., 1972, Chapter IX, pp. 323-350.
[80] McFadden, W. H., Techniques of Combined Gas Chromatography/
Mass Spectrometry: Applications in Organic Analysis, John Wiley
& Sons, N. Y., 1973.
[81] Weil, L. , Frimmel, F., and Quentin, K.-E., Z. Anal. Chem., 268,
97 (1974) .
[82] Rosen, J. D., and Pareles, S. R., Mass Spectrometry and NMR
Spectroscopy in Pesticide Chemistry, Plenum, N. Y., 1974,
pp. 91-98.
[83] Bergstedt, L., and Widmark, G., Chromatographia, 3, 59 (1970).
[84] Damico, J. N., Barron, R. P., and Ruth, J. M., Organic Mass
Spectrometry, !_, 331 (1968); Damico, J. N., J. Assoc. Of fie.
Anal. Chem., 49, 1027 (1966); Benson, W. R., and Damico, J. N.,
J. Assoc. Offic. Anal. Chem., 48, 344 (1965) and 51, 347 (1968).
32
-------
[85] Biros, F. J., and Sullivan, H., cited in [66].
[86] Biros, F. J., Applications of Combined GC-MS to Pesticide
Residue Identification, presented at the ACS-Chem. Inst.
Canada International Meeting, Toronto, May, 1970.
[87] Shafik, T. M., Biros, F. J., and Enos, H. F., J. Agr. Food Chem.,
18_, 1174 (1970) .
[88] Biros, F. J., and Ross, R. T., Fragmentation Processes in the
Mass Spectra of Trialkylphosphates, Phosphorothionates,
Phosphorothilates, and Phosphorodithiolates, Presented at
18th Conference on Mass Spectrometry and Allied Topics,
San Francisco, June, 1970.
[89] Biros, F. J., and Walker, A. C., J. Agr. Food Chem., 18, 425
(1970).
[90] Bellman, S. W., and Barry, T. L., J. Assoc. Offic. Anal. Chem.,
54_, 499 (1971) .
[91] Bagley, G. E., Reichel, W. L., and Cromartie, E., J. Assoc.
Offic. Anal. Chem., 53, 251 (1970).
[92] Bonelli, E. J., Anal. Chem., 44, 603 (1972).
[93] Sun, Y. P., Analytical Methods for Pesticides and Plant Growth
Regulators, Zweig, G., ed., Vol. 1, Academic Press, N. Y.,
1963, page 571.
[94] Allen, P. T., Analytical Methods for Pesticides and Plant
Growth Regulators, Zweig, G., Ed. Vol. V, Chapter 3, Academic
Press, N. Y., 1967, page 67.
[95] Gajan, R. T. , Residue Reviews, 5_, 80 (1964); £, 75 (1964).
[96] Gajan, R. T., J. Assoc. Offic. Anal. Chem., 52, 811 (1969).
[97] Hance, R. J., Pest. Sci., ^, 112 (1970).
[98] Booth, M. D., and Fleet, B., Talanta, 17, 491 (1970).
[99] Kutlukova, V. S., Toropov, A. P., and Lozovatskaya, M. A.,
Analytical Abstracts, 27, Abstract No. 2940 (1974).
[100] Farwell, S. O., Beland, F. A., and Geer, R. D., Bull. Environ.
Contain. Toxicol., 10, 157 (1973) .
[101] 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, pp. 89-144.
33
-------
[102] Beroza, M. , and Goad, R. A., J. Gas Chroma tog. , 4_, 199 (1966).
[103] Beroza, M., J. Org. Chem., 28, 3562 (1963).
[104] Beroza, M. , J. Gas Chroma tog, 2_, 330 (1964).
[105] Asai, R. I., Gunther, F. A., and Westlake, W. E., Residue Reviews,
19_, 57 (1967) .
[106] Asai, R. I., Gunther, F. A., Westlake, W. E. , and Iwata, Y.,
J. Agr. Food Chem., 20, 628 (1971).
[107] Zimmerli, B., J. Chromatogr., 88, 65 (1974).
[108] Minyard, J. P., and Jackson, E. R., J. Agr. Food Chem., 13,
50 (1965). "
34
-------
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 RTF 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?
-------
(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?
- 2 -
-------
(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
- 3 -
-------
f / Use alternate tests
Pause and think
\
b
M-l
Isolate problem area
VOM-Meter checks
Front Panel checks
Define the problems
(a) Write down the symptoms or difficulties observed. Diagnosis
should include observations of instrument settings.
(b) Examine the problem by using the controls, meters, and lights
on the instrument.
(c) Use a VOM to check voltages or resistances.
(d) Bracket the system path - electrical, mechanical, flow, and
chemical. The overall system and local systems must be considered.
Isolation principles:
Before starting: Read the instrument manual - understand the
instrument, know how to use test equipment, and understand test
results.
What to do first: Use logical approach, bracket the problem to a
system, bracket the area in the system, use checks to isolate faulty
area.
Signal paths: Use a block diagram and locate where in the signal
path normal and abnormal output occurs.
_ 4 _
-------
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
Abnormal
Observe and measure for
normal/abnormal readings
Normal
Injection
point
Detector
Abnormal
Bracket Placement
e.g., no flow: Rotameter indicates flow ,
- 5 -
-------
(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
^x
\
X
Column Oven
c
j.
/
Support Electronics
g
Detectors
d
S
Detector
Electronics
e
*.
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 (chromatograin) .
(g) Support electronics: Controls the temperature of the injection
ports, column oven, detectors; indicates temperatures, indicates
control voltage.
- 6 -
-------
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 hexane^cetone before reloading. It may also be
necessary to flame the dryer frit to expel all contamination.
- 7 -
-------
(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
- 8 -
-------
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.
- 9 -
-------
(5) Adjust the detector temp-set controller to maintain a
detector temperature of approximately 200°C.
(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.
- 10 -
-------
(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.
- 11 -
-------
(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 am accept-
able BGS is obtained. If, after the addition of the in-line filters, a
proper BGS is obtained, it may be assumed that the problem is in the
rotameter or flow controller area. Do not operate with these filters
permanently. Install a new carrier module and replace all tubing from
the bulkhead fittings to the in-line filters.
vi. Install well conditioned columns and allow the system to
equilibrate. If in-line filters remain in the carrier loop, allow
additional equilibration time because of the greater volume in the loop.
Flow control changes will take approximately 30 minutes to equilibrate
with these filters in the system. Obtain a BGS profile from all ports.
Insure that the columns are not filled to the point where the packing
will come in contact with the metal inlet and outlet fittings. The
higher temperature at these points may cause the column packing to bleed.
Slowly raise the oven temperature to 100°C and obtain a BGS profile.
If it is satisfactory, raise the oven temperature to full operating
temperature and obtain a BGS profile. If it is again satisfactory, the
instrument should now be in full operating condition.
It has been the experience of the RTP Laboratory that in-line filters
loaded with 13X molecular sieve are superior in performance to those
loaded with type 5A. It is therefore recommended that all filters used
be charged with 13X and conditioned as prescribed. The addition of in-
line filters at the rear of the instrument between rotameters and column
inlets is not meant to be a permanent change. The installation of these
filters allows operation of the instrument until it is convenient to
obtain materials to make repairs. As an additional means of rapidly
checking the system for leaks, it is suggested that the unions on the
instrument rear where the copper tubing meets the stainless steel tubing
be replaced with Swagelok slide valves (#200-1/8 SV-6). These valves
will enable the chromatographer to totally shut off any individual
carrier loop so that the float in the rotameter, flow controller, and
- 12 -
-------
lines 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.
- 13 -
-------
-"0" Ring
-Rear furrule (Reversed)
-1/4" Swagelok Nut
-"0" Ring
Column Packing
Glass Wool Plug
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.
- 14 -
-------
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 will be used in newer
types for temperature sensing.*
(c) Fan motor and squirrel cage blower assembly.
(d) Damper system.
The oven is generally a stainless steel unit insulated with micro fiber
insulation cover. The design permits rapid heating and cooling
dynamically to desired temperature equilibration.
The top sheathed thermocouple is used as the temperature indicating unit
and the bottom sheathed thermocouple* is used for programmer control.
This placement is non-critical.
The programmer contains an initial control circuit which is used mainly
in isothermal control of the oven and a final control circuit activated
for programming. The programmer circuitry may be used to raise the
temperature at a specific rate and automatically return to a set
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 problem 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
- 15 -
-------
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 (Ni ) 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 an H (tritium) detection 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 ^egr 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.
- 16 -
-------
(2) Check 1 percent resistors on amplifier board for continuity.
(3) Check 26N699 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.
9.1. 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 base line (may be observed on peaks): check for dirty
recorder slide wire, line voltage changes, recorder drive, recorder
gain control.
- 17 -
-------
(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.
- 18 -
-------
Troubleshooting:
(1) Leave column cold. Eliminates (2) and (8) in A.
(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.
- 19 -
-------
(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,.
X4) Check 0-110 VAC output from heat control.
(c) Inlet block will not heat.
Symptom: no temperature increase on oyrometer.
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.
- 20 -
-------
(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 recorder.
Caution: use only volt-ohmeter. Do not use vacuum volt meter.
- 21 -
-------
(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..
- 22 -
-------
(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.
Symptom: 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.
- 23 -
-------
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.
Troubleshooting procedure:
(1) Check standard.
(2) Check range and gain setting.
(3) Replate cell.
(4) Replace pyrolysis tube.
(5) Test flow system for leaks.
- 24 -
-------
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.
- 25 -
-------
Troubleshooting procedure:
(I) 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.
- 26 -
-------
(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.
2/
(4) Check variable 1-110 VAC output of heater control with volt
meter.
Caution: use only a volt-ohmeter. Do not use a vacuum tube volt meter.
- 27 -
-------
(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.
2/
(4) Check variable 0-110 VAC output of heater control with volt
meter.
9.M. TROUBLESHOOTING THE FLAME PHOTOMETRIC DETECTOR (FPD)
(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.
2/
See footenote on page 27.
In any case of noisy baseline, make certain the recorder gain is properly
adjusted and the slide wire is clean.
- 28 -
-------
(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.
- 29 -
-------
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.
- 30 -
-------
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.
-------
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.
- 2 -
-------
TABLE 10-1
RELATIVE PERFORMANCE RANKINGS
CHECK SAMPLE NO. 26, MIXTURE IN SOLVENT
Lab . Code
Number
l61.
/ 137.
S 135.
162.
*" 87.
" 113A.
* 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.
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
- 3 -
-------
-------
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)
U.S. GOVERNMENT PRINTING OFFICE: 1978740-261/332 Region No. 4
-------
TECHNICAL REPORT DATA
[Please read Inunctions on the reverse before completing)
1. REPORT NO.
FPA-600/1 -76-017
4. TITLE AND SUBTITLE
MANUAL OF ANALYTICAL OUALIT
AND RELATED COMPOUNDS In Hu
Samples
2.
Y CONTROL FOR PESTICIDES
man and Environmental
7 AUTHOR(S)
>loseoh Sherma
9. PERFORMING ORGANIZATION NAME Als
Department of Chemistry
Lafayette Colleae
Eastern, Pennsylvania
12. SPONSORING AGENCY NAME AND ADC
Health Effects Research Lab
Office of Research and Deve
U.S. Environmental Protecti
Research Triannle Park, N.C
D ADDRESS
RESS
oratory
Torment
on Aqency
. 27711
3 RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
February 1976
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1EA615
11. CONTRACT/GRANT NO.
68-02-1727
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This manual primarily provides the pesticide chemist with a systematic protocol
for the prevention and control of analytical procedures which arise in the
analysis of human or environmental media. The sections dealing with inter- and
intra-laboratory quality control, the evaluation and standardization 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 pesti-
cide analytical chemist. Section 7 discusses many aspects of the problem areas
involved in extraction and isolation techniques for pesticides in various types
of samples. Techniques for confirming the presence or absence of pesticides in
sample materials are treated, at some length. This highly important area provides
validation of the 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.
a. DESCRIPTORS
Quality control
Chemical analysis
Chemical tests
Pesticides
Bioassay
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
KEY WORDS AND DOCUMENT ANALYSIS
b. IDENTIFIERS/OPEN ENDED TERMS
Environmental samples
19 SECURITY CLASS (n,n Report!
UNCLASSIFIED
20 SECURITY CLASS (This page/
UNCLASSIFIED
c. COSATI f ield/Group
07, C
07, B
21. NO. OF PAGfcS
28^
22. PRICE
EPA Form 2220-1 (9-73)
276
------- |