General approaches for residue
analytical method development
and validation

Thomas J. Class and Reiner Bacher

PTRL Europe GmbH, Ulm, Germany

1 Introduction

Analytical chemistry is an important field in the life sciences whether the main focus
is health (pharmaceutical chemistry), nutrition (food chemistry), food supply (pes-
ticide chemistry), environment (water chemistry, waste minimization, disposal or
treatment) or lifestyle (textiles, mobility, cosmetics). Thus chemists (and other scien-
tists) working analytically, whether they are trained originally as analytical chemists
or whether they come from a different field and use analytical chemistry as support
for their research area, play an important role in supporting the progress in the life
sciences.

Each chemist working analytically uses (sometimes without any awareness) the

analytical process, a scheme (see Figure 1) by which most analytical problems are
assessed. The analytical process is a multi-step approach to solving questions by
analytical chemistry and includes the following steps:

r

Define the problem and the question(s) to be answered. These may originate in any
field of the life sciences, or in any technical or scientific area, or even in politics or
society.

r

Define the analytical approach, such as the material and the analytes to be looked
for so as to (possibly) answer the questions asked and to solve the problems.

r

Select an appropriate analytical method, with definition of its purpose and utility.
If none of the available methods fits the analytical purpose, try to deduce method
approach(es) from existing methods for structurally related compounds or materials
by introducing carefully selected modifications and adaptations.

r

Other considerations could include availability of reagent(s) or equipment, method
for routine analyses vs limited samples, and confirmatory method vs multi-residues.

r

Plan for method validation and/or analytical quality control.

r

Define the specimen(s) and the sampling procedure(s) to obtain a representative
sub-sample of the materials to be examined.

Handbook of Residue Analytical Methods for Agrochemicals.

C

2003 John Wiley & Sons Ltd.

General approaches for residue analytical method development and validation

51

Problem and Question(s) of User

Analytical Problem

Object(s) and Analyte(s)

Analytical Approach

Sampling

Interpretation

and Answers

Results and Report

Sample (Pre-) Processing,

Extract Preparation,

Derivatization

Determination:

Identity, Quantity,

Concentration

Figure 1

The analytical process

r

After sampling, homogenization, extraction, cleanup, concentration and possible
derivatization, use a suitable determination method which provides sufficient se-
lectivity or specificity and sufficient sensitivity.

r

Ensure that the analytical methodology gives reliable results in terms of identity
(absence of false-positive findings) and of absence (no false-negative findings) of
the analyte(s). This requires processing of concurrent analytical quality control
samples.

r

Reliable identification may require confirmation by a method with different selec-
tivity or employing a different analytical principle.

r

Ensure that quantitation yields accurate and precise results by monitoring the back-
ground, recoveries and standard deviations.

r

Report the results in a comprehensive manner to allow interpretation of the findings
and the drawing of conclusions.

r

Answer the questions and solve the problem posed at the beginning of the analytical
process.

The analytical chemist is not involved in the entire analytical process in all cases. It

is always preferable, however, not only to focus on the analytical method, but also to
consider the background of the analytical task and the consequences of the analytical
results.

2 Approaches to analytical method development

2.1 Properties of the analyte(s)

The structure of the compounds to be analyzed is usually known, but there are
cases where unknown metabolites, degradation products, conjugates or species (e.g.

52

Regulatory and scientific consideration for residue analytical methods

metal complexes, adducts) need to be searched for. This is the case, for example,
in metabolism or degradation studies, or when (unwanted or positive) effects are
presumably caused by compounds of unknown identity.

The following physico-chemical properties of the analyte(s) are important in

method development considerations: vapor pressure, ultraviolet (UV) absorption
spectrum, solubility in water and in solvents, dissociation constant(s), n-octanol/water
partition coefficient, stability vs hydrolysis and possible thermal, photo- or chemical
degradation. These valuable data enable the analytical chemist to develop the most
promising analytical approach, drawing from the literature and from his or her expe-
rience with related analytical problems, as exemplified below. Gas chromatography
(GC) methods, for example, require a measurable vapor pressure and a certain ther-
mal stability as the analytes move as vaporized molecules within the mobile phase.
On the other hand, compounds that have a high vapor pressure will require careful
extract concentration by evaporation of volatile solvents.

A UV spectrum with a pronounced absorption above 210 nm allows UV detection

after liquid chromatography (LC), but an absorption maximum in the range of visible
light may also decompose during cleanup procedures and require the elimination of
light when handling extracts.

Water solubility, dissociation constant(s) and n-octanol/water partition coefficients

allow one to predict how an analyte may behave on normal-phase (NP), reversed-
phase (RP), or ion-exchange solid-phase extraction (SPE) for sample enrichment and
cleanup.

2.2 Functional groups of the analyte(s)

The presence of heteroatoms usually provides a convenient feature for improving
selectivity by employing selective detection mechanisms. GC may then use: flame
photometric detection (FPD) for S and P atoms and to a certain extent for N, Se, Si
etc.; thermoselective detection (TSD) and nitrogen–phosphorus detection (NPD) for
N and P atoms; electron capture detection (ECD) for halogen atoms (F, Cl, Br, and
I) and for systems with conjugated double bonds and electron-drawing groups; or
atomic emission detection (AED) for many heteroatoms.

The isotopic patterns of

35

/37

Cl

n

or

79

/81

Br

m

atoms present in the analyte molecule

are very helpful when mass spectrometry (MS) is used for specific detection and iden-
tification after GC or LC. The presence of several halogens in a molecule, however,
may decrease the sensitivity in normal resolution (quadrupole or ion trap) MS and
tandem mass spectrometry (MS/MS) due to the presence of an isotopic ion pattern,
but when high-resolution mass spectrometry (HRMS) (magnetic field HRMS) detec-
tion is employed (using the sum of the mass defects of the halogen atoms present
in a fragment ion), the selectivity and consequently sensitivity are improved tremen-
dously. Multiple fragmentation ions are formed in the electron ionization (EI) source
when labile bonds break, whereas chemical ionization (CI) or electrospray ionization
(ESI) sources result in softer ionization and thus less bond breakage and a limited
number of ions with less fragmentation.

On the other hand, electronegative substituents such as F, Cl and Br atoms, but

also NO

2

and COOH moieties, are extremely sensitive when negative ions and

General approaches for residue analytical method development and validation

53

fragments are monitored by gas chromatography/mass spectrometry (GC/MS) or
liquid chromatography/mass spectrometry (LC/MS).

Functional groups and reactive moieties may cause losses by adsorption on the

matrix and surfaces. On the other hand, functional groups may be used to increase the
selectivity and sensitivity of the methods by derivatization, forming distinct deriva-
tives with different properties during cleanup and chromatographic separation and
detection. Examples are the esterification of carboxylic acid moieties (using, e.g.,
diazomethane, diazoethane, trimethylsilyldiazomethane, sulfuric acid–methanol or
butanol), silylation of carboxylic acid, hydroxyl or amino moieties, benzylation of
OH moieties, cyclization of two neighboring moieties (using, e.g., acetylacetone or
aceticanhydride), etc. Precolumn derivatization of functional groups is mandatory if
GC is employed for polar or thermolabile analytes. Postcolumn derivatization after
high-performance liquid chromatography (HPLC) may drastically improve selectivity
as UV detection can be substituted by fluorescence detection.

2.3 Properties of the sample material

The composition, properties and size (weight, volume) of the sample material to be
analyzed are important aspects for analytical method development and for analyte
enrichment vs depletion of sample matrix.

Sampling of air to determine worker exposure or for environmental purposes usu-

ally includes the easy task of eliminating first the air (N

2

, O

2

, trace gases) and then

the humidity present in the air (water which may condense and saturate adsorption
columns). Particles (e.g. salts, soil, soot) present in the air and trapped during air
sampling may also contain active species or adsorptive surfaces and thus cause losses
of the analytes. Large volumes (liters to cubic meters, high-volume sampling up to
1000 m

3

) of air, however, are usually sampled by eliminating air and water without

losing the analytes, thus reducing the sample size by a large factor. These procedures
depend on the vapor pressure and the adsorption or absorption mechanisms that retain
the analyte.

Sampling of water for monitoring purposes allows sample sizes of 1–3 L or less.

Eliminating water and retaining the analyte are relatively easy for organic molecules
with a high n-octanol/water partition coefficient, but become more difficult for ana-
lytes with ionic properties and high solubility in water. If the water sample contains
a high load of salts, silt or organic matter such as surface (river, pond, sea) water,
possible adsorption on filtered matter needs to be considered.

Plant material water contents range from high (

>90%, e.g. vegetables) to low

(

<10%, e.g. straw, herbs, tea, hops, etc.). Thus the ratio between the analytes (residues)

and the organic matter potentially interfering with the analysis is very different for,
e.g., cucumber and camomile tea. Other ingredients in plant materials such as acids,
oil, sugars, starch or substances typically for the taste and effect of plant materials
may have properties similar to those of the analytes and thus interfere in or influence
the cleanup procedures.

Materials of animal origin such as tissue, fat, milk, egg or blood contains usually

relatively large amounts of fat, proteins and carbohydrates that need to be reduced
during cleanup to allow enrichment of the analytes to be searched for.

54

Regulatory and scientific consideration for residue analytical methods

Sample size may influence the analytical approach, e.g. 0.1 L of milk is easily ob-

tained and extracted, but 10 mL of blood should be sufficient for monitoring purposes,
and 5 g of fat is already the upper limit for an efficient fat cleanup by partition or gel
permeation chromatography (GPC).

2.4 Availability and practicality of analytical instrumentation

Analytical instrumentation ranges from that for generally available techniques such as
liquid chromatography/ultraviolet detection (LC/UV) or gas chromatography/flame
ionization detection (GC/FID) with their limitations with regard to selectivity and sen-
sitivity on the one hand, to very sophisticated techniques such as GC, LC or capillary
electrophoresis (CE) coupled to triple-quadrupole tandem mass spectrometry
(MS/MS) in space, ion trap MS

n

in time, time-of-flight (TOF), or high resolution

(HRMS) mass spectrometers on the other. The availability of the very sophisti-
cated (and expensive) instrumentation (such as GC with HRMS, or LC with triple-
quadrupole MS/MS) may lead to the over-use of this instrumentation with very little
attention given to cleanup and chemistry, whereas the less sensitive/selective tech-
niques require one to focus more on the laboratory procedures of the method.

2.5 Consideration of time, throughput, ruggedness and quality

In the development of analytical methods one has to consider also cases where a
fast response is required, e.g. clinical and forensic chemists or toxicologists need
methods which yield results in a few minutes or hours to allow a fast response in cases
of poisoning. In this event, accurate quantitative results may be of less importance,
but the time from sampling to result may be lifesaving, whereas the throughput (i.e.
number of analyses per day) is not so much of concern.

Residue analysts working in enforcement laboratories are required to analyze spec-

imens collected in the market or obtained from import/export facilities in time to ex-
clude any goods with unacceptable residues being sold or imported. Their analytical
problem is focused mainly on the presence or absence of regulated residues where
they need to avoid any false-positive or false-negative results. Hence for these analysts
the analytical method needs to give them reliable results in a day or two or before the
foodstuff of plant or animal origin is sold, consumed or spoiled. On the other hand,
they are required to collect and analyze large numbers of representative samples as
tolerances or maximum residue levels (MRLs) need to be surveyed and enforced.

Throughput may also be an issue when monitoring programs for groundwater

or for characteristic consumer/market baskets yield very large numbers of samples
for analysis. Such monitoring programs are expected to yield reliable results and
therefore require special care in terms of accuracy and precision of the results. This
is often ensured by frequent and rigid quality assurance measures such as intra- and
inter-laboratory comparison tests and the use of certified reference materials.

An analytical method can be considered rugged when it can be transferred from one

laboratory to another with comparable experience without much effort in adaptation to
the different technical personnel and to the different equipment and instrumentation.

General approaches for residue analytical method development and validation

55

The above leads to the concept of the quality of an analytical method (not to

be confused with the quality of the results). A method should be as simple or as
sophisticated as necessary to serve its purpose in yielding reliable results and in
answering the questions posed at the beginning of the analytical process.

3 Practical examples

3.1 Extending the scope of the multi-residue method DFG S19

The multi-residue method DFG S19

1–3

was intended to be used in state enforcement

laboratories or in private contract or food industry laboratories. It was aimed initially
only at plant materials and water and included a relatively large number of pesticides
which are amenable to GC.

Its principles include polar extraction with acetone–water (2 : 1, v/v), homoge-

neous partitioning of the target molecules into an organic solvent, GPC cleanup on
Bio-Beads, fractionation by adsorption column chromatography on silica gel (SiO

2

)

deactivated with 1.5% water and finally GC with various selective detection methods
(NPD, ECD, FPD).

Confirmation of suspected residue findings relies on the various chromatographic

principles of cleanup and determination (GPC, NP-LC, GC), and is further supported
by re-analysis of the final extract(s) on a GC stationary phase of different polarity,
providing modified selectivity, or by the use of GC with specific mass spectrometric
detection [GC/MS or gas chromatography/tandem mass spectrometry (GC/MS/MS)].

The practicality of the method was further improved by introducing a one-beaker

extraction and partition step, dichloromethane being replaced with ethyl acetate–
cyclohexane as the organic phase during homogeneous partitioning. These solvents
together with the acetone portion of the extraction form the upper organic phase,
whereas the hydrophilic matrix constituents remain in the aqueous phase saturated
with sodium chloride.

The method procedures are very time-efficient by always using a well defined

portion of the extracts, thus avoiding multiple extractions, time-consuming rinses and
overloading of the chromatographic cleanup systems with co-extracted sample matrix.

For oily crops such as nuts and oilseeds, a slightly different extraction procedure

with 10% acetone in acetonitrile is used.

The scope of the multi-residue method is extended permanently by testing and then

including further active substances that can be determined by GC. Acidic analytes
(such as phenoxyacetic acids or RCOOH metabolites) are included into the homo-
geneous partitioning by acidifying the raw extracts to a pH below the pK

s

value of

the carboxylic acids. To include these analytes in the GC determination scheme they
have to be derivatized with diazomethane, diazoethane, trimethylsilyldiazomethane,
acidic esterification or benzylation, or by silanizing the COOH moiety.

Another extension of the DFG S19 method was achieved by applying it successfully

to foodstuffs of animal origin such as whole milk and egg, muscle meat, offal, fat
and honey. Depending on water and fat content, either water–acetone (e.g. for milk,
meat, possibly egg and honey) or acetone–acetonitrile (e.g. for offal, egg, fat) solvent
extraction is preferable. When high fat or oil contents in the raw extract are expected,

56

Regulatory and scientific consideration for residue analytical methods

an additional hexane–acetonitrile partitioning/wash step prior to GPC reduces the
load on the size-exclusion GPC column and thus avoids overloading the pores, which
can cause a shift of the analyte retention times to earlier elution.

With the use of ion trap GC/MS for determination, the method became more

universal when the ion trap mass spectrometer was operated in the full-scan mode,
providing for many analytes sufficient sensitivity and a high specificity by providing
full mass spectra. With quadrupole GC/MS operated in the selected ion monitoring
(SIM) mode, or with ion trap GC/MS operated in the selected ion storage (SIS) mode,
the sensitivity of the method was further improved, but the universal multi-residue
character of the DFG S19 method, however, was reduced to a target method still
very useful for the confirmation of suspect residue findings. A further improvement
was achieved by using ion trap or quadrupole GC/MS/MS for determination of target
analytes in the final extracts.

As the sensitivity and selectivity of the above GC/MS methods are for many analytes

around 1 pg µL

−1

injected into the GC system, cleanup by SiO

2

fractionation can

be omitted when larger sample sizes (25–100 g) are possible. For difficult dry (e.g.
hops, pharmaceutical herbs) or oily (e.g. rape seed, fat, liver) materials which start
with smaller sample sizes (5–10 g) and tend to overload the chromatographic cleanup
systems, however, cleanup is still an important requirement as the GC injection system
is vulnerable when the ratio of co-extracted material to analyte is too high.

A further extension of the DFG S19 method was achieved when polar analytes

and those unsuitable for GC were determined by LC/MS or more preferably by
liquid chromatography/tandem mass spectrometry (LC/MS/MS). Triple-quadrupole
MS/MS and ion trap MS

n

have become more affordable and acceptable in the recent

past. These techniques provide multiple analyte methods by employing modes such
as time segments, scan events or multiple injections. By improving the selectivity and
sensitivity of detection after HPLC separation, the DFG S19 extraction and cleanup
scheme can be applied to polar or high molecular weight analytes, and cleanup steps
such as SiO

2

fractionation or even GPC become unnecessary.

What can be achieved by the fully extended DFG S19 approach is impressively

demonstrated by the residue analysis of many pesticides during minor crop registra-
tion. Considering an analyte such as azoxystrobin, which is tested for application in
minor crop cultures or pharmaceutical herbs such as artichokes, peppermint, camomile
or St. John’s Wort, GC is expected to result in problems of thermal degradation or
unacceptable tailing of its large and polar molecule, and LC/MS/MS becomes the
method of choice. Further, as pharmaceutical herbs or teas are very dry materials,
only a small sample size, e.g. 5–10 g, can be extracted. This small size of the analyti-
cal samples is still acceptable and considered sufficiently representative if taken from
a larger quantity previously homogenized frozen with dry-ice. To circumvent possible
matrix effects (e.g. precipitation) in the final extracts for LC injection, atmospheric
pressure chemical ionization (APCI) and ion trap MS

n

detection of positive ions, the

raw extract obtained by homogeneous partition is fractionated by GPC, which in this
case reduces the presence of larger biomolecules. A further improvement in speci-
ficity without loss in sensitivity is achieved by employing an ion trap MS

3

method.

Once this target method has been established for a few different plant materials, it
can be easily extended to related analytes such as pyraclostrobin or most other plant
materials.

General approaches for residue analytical method development and validation

57

Another successful adaptation of the fully extended DFG S19 approach is the de-

termination of, e.g., fenpyroximate in all type of berries by LC/MS/MS with APCI
monitoring of positive ions directly in the S19 raw extract, and further the deter-
mination of trifluralin by LC/MS/MS with APCI monitoring of negative ions after
performing a short SPE cleanup on an ion-exchange material. Similar approaches have
used GC/MS/MS for, e.g., fenpropimorph and kresoxim methyl in St. John’s Wort
and peppermint.

Once several target methods employing, e.g., LC/MS/MS techniques have been

combined, a multi-residue method will evolve which includes the DFG S19 extraction
procedures in combination with the generally applicable GPC cleanup and requires
automatic multiple injections to circumvent the limitations of the limited HPLC peak
capacity and the target-specific MS/MS methods.

3.2 What can go wrong?

The above examples can be extended to the majority of older and newer active sub-
stances described in, e.g., ‘The Pesticide Manual’

4

and to numerous relevant metabo-

lites featuring hydroxyl or carboxyl moieties or even for conjugates; however, there
remain various active substances and metabolites that still require careful and exten-
sive method development.

The following are some of the main pitfalls one can expect:

r

Ionic or amphotheric character of the analytes, e.g. asulam and its metabolites
acetylasulam, sulfanilamide, acetylsulfanilamide:

Asulam

S

N

H

O

O

O

O

H

2

N

Sulfanilamide

S

NH

2

O

O

H

2

N

The carbamate –NH– moiety present in asulam has acidic properties (e.g. the pK

a

value for asulam is 4.82). On the other hand, the –NH

2

moiety present in sulfanil-

amide has a slightly alkaline character. Considering these properties, the partition
of these analytes into an organic solvent should depend strongly on the pH value
in the aqueous phase.

Such analytes require carefully chosen extraction conditions in terms of pH,

solvent composition and technique. Also, these analytes tend to become lost by
adsorption on (glass) surfaces or undergo conjugation so that a chemical or en-
zymatic deconjugation step may be required. Often only the use of radiotracers

58

Regulatory and scientific consideration for residue analytical methods

allows a time- and cost-efficient development of such extraction and the required
cleanup methods.

r

Volatile analytes. As residue analysis is also trace analysis in the lower ppm
(mg kg

−1

) to ppb (µg kg

−1

) range, concentration steps usually involve evaporation

of solvents (sometimes with traces of water present) to near dryness. The volatility
of analytes can be deduced from their elution temperatures in GC, and thus when-
ever an analyte elutes from a nonpolar GC phase of film thickness

≤0.25 µm below

approximately 150

◦

C, losses due to co-evaporation during concentration by the

rotary evaporator or by a stream of nitrogen need to be avoided.

r

Labile analytes. Labile analytes may degrade during extraction and cleanup when
stress in terms of temperature or pH is applied. They also tend to degrade on GC
injection and may even undergo extensive fragmentation during MS ionization
even with soft techniques such as ESI or CI. Therefore, one needs to consider
derivatization at an early stage of the analytical method, thus enhancing stability
and possibly detectability.

3.3 Beyond the limits

Residue analytical chemistry has extended its scope in recent decades from the ‘sim-
ple’ analysis of chlorinated, lipophilic, nonpolar, persistent insecticides – analyzed
in the first SiO

2

fraction after the all-destroying sulfuric acid cleanup by a gas chro-

matography/electron capture detection (GC/ECD) method that was sometimes too
sensitive to provide linearity beyond the required final concentration – to the moni-
toring of polar, even ionic, hydrophilic pesticides with structures giving the chemist no
useful feature other than the molecule itself, hopefully to be ionized and fragmented
for MS or MS

n

detection.

The required limit of quantitation (LOQ) and limit of detection (LOD) have been

extended to the parts per billion range as the European Community (EC) ‘baby food’-
related guideline and the US ‘consumer basket’ requirements became effective.

Modern analytical techniques in combination with conventional analytical experi-

ence and thinking thus try to meet these new requirements by pushing residue analysis
to extended limits.

References

1. H. P. Thier and H. Zeumer (eds.), ‘DFG Manual of Pesticide Residue Analysis,’ VCH, Weinheim,

1987, pp. 71–74, 75–78 and 383–400.

2. W. Specht, S. Pelz, and W. Gilsbach, Fresenius’ J. Anal. Chem., 353, 183 (1995).
3. Bundesinstitut f¨ur Gesundheitlichen Verbraucherschutz und Veterin¨armedizin, ‘Official Collec-

tion of Test Methods According to

§35 LMBG (Law of Food and Commodities),’ edited Beuth

Verlag, Berlin, 1999, Modular Multi Method L 00.00-34.

4. C. D. S. Tomlin (ed.), ‘The Pesticide Manual,’ twelth edition, Version 2.1, British Crop Pro-

tection Council, Farnham, 2001–2.