Logan 1994
Logan 1994
Logan 1994
Ahtract
Liquid chromatography (LC) is used in forensic laboratories for numerous applications including examination of
drugs and poisons, plastics and polymers, oils, hydrocarbons, cosmetics, inks, dyes and pigments. LC with photodiode
array detection (PDA) is a hybrid technique which can provide complete UV-visible spectral information on a given
peak in a chromatogram, enabling determinations of peak purity to be made, and identification of unknown peaks to
be assigned by library searches of spectral information in combination with retention behavior. These are valuable
features normally associated with gas chromatography-mass spectrometry. The additional information available on
each peak makes LC-PDA a particularly attractive technique for the forensic laboratory where higher levels of
certainty are often demanded in test results. This paper reviews some of those applications for LC-PDA in the
forensic sciences, including drug screening, drug and pharmaceutical analysis, differentiation of inks, identification of
pesticides, fungi, quality control testing and profiling of cosmetics, street drugs and profiting of other complex
mixtures. The practical and technical limitations of the technique are explored and its place in the hierarchy of
methods available in forensic laboratories is evaluated.
Keywords: Liquid chromatography; UV-Visible spectrophotometry; Forensic analysis; Photodiode array detection;
Street drug analysis, Toxicology
Liquid chromatography (LC) has been used single wavelength, electrochemical activity at a
extensively in forensic laboratories for a number given potential, etc.). This has restricted the role
of years, with applications in drug testing, toxicol- of LC mostly to quantitative analysis of com-
ogy, and trace evidence examination including pounds whose identity was already known [2], and
dyes and intermediates, explosives, hydrocarbons, to two-dimensional profile comparisons for com-
lipids, optical brighteners, plastics and polymers, plex mixtures [3,4].
plasticizers and other additives [l]. Until recently The availability of low-cost diode array UV-
the detectors available for LC have given only visible spectrophotometric detection since the late
limited information about a chromatographic 1980s [5,6], now allows the collection of multiple
peak; namely its retention time, and its response data points on any chromatographic peak, and
to a single measuring probe (e.g., absorbance at a subsequent manipulation of the data, markedly
improving the discriminating power of LC for
qualitative as well as quantitative applications.
Correspondence to: Washington State Toxicology Laboratory,
A central tenet of chemical analysis in the
Department of Laboratory Medicine, University of Washing- forensic sciences is the confirmation of an analyt-
ton, 2203 Airport Way S., Seattle, WA 98134 (USA). ical result by an independent chemical means.
For example, drug testing laboratories certified such as in tests for cocaine, heroin, PCP, di-
by the National Institute of Drug Abuse (NIDA), azepam, marijuana, etc., but also in identification
typically rely on an initial drug screen using an of components in unknown mixtures such as illicit
immunoassay method, followed by a confirmatory drugs and pharmaceuticals, extracts from biologi-
quantitative analysis by gas chromatography-mass cal fluids, inks and dyes. This paper reviews some
spectrometry (W-MS) [7]. The initial identifka- of the applications in which LC-PDA has been
tion is based on biochemical recognition of part used routinely as both a complement to, and an
of the compound’s molecular structure, while the alternative to CC-MS in the forensic sciences.
confirmatory test is based on the volatility, polar-
ity, and predictable fragmentation behavior of the ZrWument considerations
compound in the source of the mass spectrome- Measuring UV-visible absorbance at a single
ter. The advent of photodiode array (PDA) spec- wavelength has been the favored detection
trophotometric detectors for LC provides the method for LC for many years, because of its
chromatographer with the means of identifying by sensitivity, general applicability and relatively low
retention time the nature of a peak in a chro- cost [1,9]. PDA detectors provide one means of
matogram, and confirming this identification by recording in real-time, constant flow mode, a full
compiling a ratiogram at two or more selective UV-visible spectrum of the contents of the de-
wavelengths, or by examining the complete UV- tector flow cell. The optics design for the PDA
visible spectrum of a given peak. The success of detector is the reverse of traditional LC UV-visi-
this technique depends on the compound in ques- ble detector as shown in Fig. 1. In a PDA detec-
tion having a spectrum with distinctive features tor, a collimator lens focuses white light through
which, while typically less specific than mass the flow cell, where absorbance across the com-
spectrometry [8], is often sufficient to provide plete spectrum takes place. The original light
confirmation with a high degree of confidence, signal minus the absorbed energy is then refo-
particularly when considered along with retention cused and split by a holographic grating. The
time information, the compounds behavior during resulting spectrum impinges on an array of pho-
a selective extraction, and any ancillary informa- todiodes, whose signals are amplified and fed to
tion from screening tests or other sources. the data system. Each photodiode therefore con-
The growing body of literature describing pho- tinually monitors a discrete section of the spec-
todiode array detection in LC analysis of forensic trum. The size of this section and the degree of
samples has borne this out, particularly where dispersion by the grating, determines the instru-
well-characterized compounds are being tested, ment’s spectral resolution. Figure 2 shows an
example of the three dimensional representation
of the spectral/ temporal data set known as a
spectrochromatogram. This particular example
shows a series of benzodiaxepines, separated
chromatographically and illustrates how their
identity can be further ascertained by comparison
of characteristic spectral features.
The issue of resolution is crucial in the selec-
tion of a PDA instrument for forensic applica-
tions. Resolution is the instrument’s ability to
discriminate between two closely adjacent wave-
Fig. 1. Optics of conventional and photodiode array detectors. lengths. Two wavelengths are said to be resolved
(a) Conventional variable-wavelength UV-visible detector for if the trough between the two peaks is lower than
LC. (1) Light source, (2) grating, (3) monochromator, (4) flow
cell, (S) single photodiode. (b) Reversed optical bench for
80% of the maximum. The ratio of the spectral
photodiode array (PDA) detection for LC. (1) Light source, bandwidth, i.e. the bandwidth of the source light
(2) lens system, (3) flow cell, (4) grating, (5) photodiode array. at half the intensity, to the natural bandwidth, i.e.
B.K Logan/Anal. Chim. Acta 288 (1994) 111-122 113
the width of the absorption bandwidth at half the tion is usually unnecessary and poorer resolution
maximum absorbance, must be 0.1 or less other- compromises the value of the spectroscopic data.
wise inaccuracies in absorbance measurement for A competing technology to PDA which also
that band will result [&lo]. Most organic com- allows the collection of full W-visible spectra is
pounds in an aqueous environment have absorp- the rapid mechanical scanning UV-visible detec-
tion bands with natural bandwidths greater than tor. In this instrument, the light source is split
20 nm, and therefore adequate resolution can be prior to passing through the flow cell, as with the
achieved on an instrument with a spectral band- standard optical bench. In rapid mechanical scan-
width of 2 nm [lO,ll]. However, for some applica- ning instruments however, the grating is mounted
tions e.g., the fine absorbance detail of the five on a transducer platform, the frequency of which
‘fingers’ of benzene, 2 nm resolution would be is set at several cycles per second to ensure that,
insufficient. In practical terms, resolution is pri- for typical chromatographic peaks, several full
marily a function of the number of diodes on the spectral scans take place during the passage of
array, and increased resolution is achieved only at each component through the flow cell. This coun-
the cost of sensitivity, or spectral range. In con- ters risks of obtaining skewed spectral data on
sidering the use of LC-PDA, for most applica- peak upslope or downslope. This approach may
tions including forensic science, an instrument impart improved sensitivity and can offer resolu-
with 2 nm resolution is optimum, greater resolu- tion down to 1 nm, allowing the measurement of
Ih 6
Fig. 2. Spectrochromatogram showing spectral and chromatographic data for the separation of six benzodiazepines. (1) Midazolam,
(2) flurazepam, (3) oxazepam, (4) nitrazepam, (5) alprazolam, (6) clonazepam. Separation was accomplished on a 25cm Lichrosorb
RP-8 column (Merck), using 35% acetonitrile in 0.05 M KH,PO,-H,PO, buffer (pH 3). Flow-rate was 1.5 ml min-’ (Gilson 307)
and all spectra were collected between 190 and 400 nm (Hewlett Packard, 1040 M).
114 B.K. L,ogan/And Chim. Acta 288 (1994) 111-122
spectroscopic features with natural bandwidths of must be injected individually under each set of
10 nm [10,12]. Both approaches achieve the goal elution conditions to determine peak elution or-
of obtaining full spectral information of chro- der. In cases where peaks co-elute under one set
matographic peaks, and additional features such of chromatographic conditions, having spectral
as library matching, peak purity determinations, information about each component in the mix-
and derivative spectra are functions of the avail- ture can indicate co-elution [13,14]. Where ma-
able software. The relative advantages of each nipulation of chromatographic conditions will not
configuration can be debated, but the capabilities allow the separation of two or more compounds,
of the software system, the required sensitivity for spectral deconvolution and spectral suppression
a given assay, and the choice of chromatographic software is available which allows independent
conditions are likely to be more important con- quantitative measurement of both components in
siderations for the typical user 193. that peak, if their spectra are sufficiently distinct
ml.
The reference or library spectra should be
Method development applications acquired in the mobile phase to be used for the
The method development stage of any LC particular application. When using gradient elu-
assay is as critical in forensic science as in any tion LC methods, the spectrum may change de-
other field and is made considerably more pending on the content of the mobile phase at
straightforward when a PDA detector is avail- the time of elution from the column, particularly
able. It aids in selecting the optimum wavelength if a pH or ionic strength gradient is being run.
for maximum sensitivity and eliminating interfer- Because UV-visible spectra are sensitive to the
ence from other components in the sample or in conditions of pH, ionic environment, and solvent
biological extracts-a frequent problem in deal- system, generic UV-visible spectral libraries can
ing with forensic samples. be of limited use, and comparative spectra should
In an LC-PDA method for colchicine in post- always be acquired under local conditions. As
mortem bile [91, a wavelength of 390 nm was proper identification should be based on the be-
selected after examining the spectrochro- havior of the compound in the time domain as
matogram for extracts from drug-free bile sam- well as the spectral domain, the generally variable
ples, resulting in an interference-free chro- interlaboratory reproducibility of LC assays lends
matogram and a sensitive assay with minimal further arguments for users to construct their
sample preparation. Other peaks with similar own spectral/ chromatographic libraries.
spectral features to colchicine were noted and In toxicological applications, when method de-
assumed to be metabolites or conjugates. velopment reaches the stage of testing biological
During method development, knowing the extracts, PDA detection provides the ability to
UV-visible absorption characteristics of the com- examine through software, ratiograms at two or
ponents of a mixture allows their relative elution more wavelengths to ensure separation of the
order to be determined through a single injection analyte of interest from any co-extracted material
of the mixture, providing they have sufficiently -a frequent problem when dealing with putri-
different spectra. fied post mortem material routinely encountered
Once peaks have been identified, elution con- in post mortem forensic toxicology. In a homoge-
ditions can be changed on successive runs, and a neous peak, the ratio at two carefully chosen
mixture of all components can be injected simul- characteristic wavelengths, should be constant
taneously. Any changes in peak elution order throughout the peak, independent of concentra-
which might otherwise have gone undetected, can tion, and thus identical from run to run [16].
be determined spectrophotometrically, as can the Performing a plot of chromatograms at two wave-
identity of any other drug or artifact which could lengths results in a ratiogram which will show the
potentially interfere with the analyte of interest. presence of non-homogeneous peaks. These may
In traditional method development, compounds appear with non-linear apices (Fig. 3) if the peaks
B.K Logan /Anal. Chim. Acta 288 (1994) 111-122 115
only the best match, but the top three or four exploited in the identification of unknown spec-
matches. Library comparison is undoubtedly one tra. Inspection of any collection of W-visible
of the most useful and commonly used features of spectra will show that many compounds within a
PDA detectors in the forensic sciences, for the class will share the same spectral features, for
purposes of suggesting possible identities of un- example [201 benzodiazepines, opiates and phe-
known peaks, for checking peak purity, and for nothiazines, and some of their metabolites each
confirming the identities of compounds identified possess features which would indicate to which
by other methods. class they belong. In contrast to fragmentation
patterns in mass spectrometry which can change
appreciably following relatively minor structural-
Automated method deveibpment
modifications, W-visible spectra display greater
As noted above, determining optimum chro-
similarity within a class. PDA detection can
matographic conditions for a new assay can be a
therefore be a useful tool in identifying, for ex-
time consuming and subjective process. A variety
ample, an unknown drug according to its class
of method development and optimization soft-
even when its absolute identity camrot be imme-
ware packages are available commercially which
diately determined. There are significant excep-
exploit the multiwavelength data set generated by
tions to this general rule however, since some
the PDA. Several of these systems are entirely
metabolic modifications, particularly those involv-
automated, selecting the chromatographic condi-
ing changes in saturation, conjugation, and hy-
tions, performing the analysis and collecting and
droxylation of delocalized systems, and major
evaluating the data before selecting the solvent
conformational changes such as ring opening, can
conditions for a subsequent run. These programs
also cause considerable changes in absorbance
make use of mixture design statistical techniques
behavior [21].
such as overlapping resolution mapping and sim-
A further limitation which applies to LC-PDA
plex search algorithms to produce a user-weighted
is a result of the W-visible absorbance proper-
optimized separation [19]. Most of these systems
ties of many of the solvents typically used in
use PDA data to construct ratiograms or use full
normal phase LC, such as chlorinated hydrocar-
spectral comparisons to detect peak crossover or
bons, n-alkanes, unsaturated hydrocarbons, and
coelution during the optimization procedure. Un-
cyclic hydrocarbons, ethers, amides and esters. As
detected peak crossover between iterations can
these compounds can have very high absorbances
invalidate the optimization.
in the low W (190450 nm), and many analytes
have their only significant absorbance in this re-
Pradtical considerations gion, signal-to-noise ratios tend to be very poor
As/a result of the amount of data generated, and spectral comparisons can be compromised.
PDA instruments require a computer with a disk This also prevents the use of some mobile phase
drive. Disk space can be used most economically additives and LC-PDA is often limited to use
by collecting full spectral data only when a peak with reversed phase LC using solvent systems
elutes (threshold values can be set on most in- composed of buffer, water, short chain alcohols,
struments, as with computing integrators). For nitriles and organic acids. Constraints on design
peak purity determinations, complete chro- forced by the optical layout, together with the
matograms at several wavelengths must also be electronics noise limitations of the instruments,
collected throughout the run. typically result in a sensitivity of 2-5 times less
Photodiode array detection has many of the than state-of-the-art single variable wavelength
same limitations as conventional W-visible W-visible detectors for LC [&lo]. The greater
spectroscopy, specifically, its less specific nature potential for stray light in reversed optics systems
compared to mass spectrometry, and particulary may also limit the linear range of the instrument,
for compounds which share major structural fea- although in practice this is not a widely reported
tures. While this can be a limitation, it can be problem.
B.L Logan /Ad. Chim. Acta 288 (1994) Ill-122 117
Forensic science applications as this permits the computer library match system
LC with PDA detection has two major to tentatively identify unknown peaks according
strengths which make it an indispensable tech- to chromophore class, which can aid with identifi-
nique for forensic laboratories. The first is the cation of compounds unknown to the library. In
ability to provide reliable confirmation of peak addition, the authors urge that laboratories
identity and purity, equivalent to selected ion adopting this method should acquire their own
monitoring in mass spectrometry. The second spectral library to ensure accurate criteria for
valuable feature is in the wealth of spectral infor- identification under local conditions. The mobile
mation which is available for each peak in a phase composition which ran from 10 to 50%
chromatogram. This makes LC-PDA suitable for acetonitrile in 0.05 M phosphate buffer (pH 31,
constructing three dimensional spectrochro- allowed elution of compounds of extremely dis-
matograms for profiling of complex mixtures, such parate polarity in a single run, for example co-
as in the comparison of inks, dyes, copier toners, caine, cocaethylene and benxoylecgonine were all
cosmetics, explosives, oils, hydrocarbons and pig- detected, and morphine and thioridaxine have
ments, all of which have been shown to have been detected in the same extract. Classes of
applications in the forensic sciences [ll. To date compounds were found to elute within discrete
this feature has been exploited only to a limited blocks with respect to retention time, as com-
extent 19,121. pounds within a class are typically close in polar-
ity. This extraction procedure has provided ex-
Drug testing in biological jluih and times tracts of excellent quality from urine samples and
The major forensic application for LC-PDA is has been used successfuIly with cerebrospinal
in drug analysis-both of street drugs and in fluid and vitreous humor samples. Untreated
detecting and identifying drugs and poisons in postmortem blood samples however were not
extracts from biological fluids and tissues. Be- compatible with the SPE procedure 1221, and
cause of the increased certainty of a peak identi- samples with high salt concentration can reduce
fication many LGPDA methods have been re- recoveries for very polar compounds. The LC-
ported which attempt to provide a systematic PDA method was validated by comparison against
toxicological screen and confirmation for drugs in an enzyme multiplied immunoassay technique
serum or urine samples. This application has (EMIT)-GC/GC-MS procedure with excellent
previously been the sole domain of GC-MS. agreement.
A method for the extraction of basic drugs A standardized system for drug screening of
from urine using a cation-exchange solid phase 225 street drug and pharmaceutical samples [23]
extraction (SPE) procedure followed by LC with enhances identification based on retention behav-
PDA detection [20] used a badient elution re- ior by relating it to the 1-nitroalkane scale 1241,
versed-phase system with a C-8 column (Li- and includes a daily analysis of acid/neutral and
chrospher 100 CH-8, Merck, Darmstadt) to screen basic drug standards to provide correction fac-
for over 100 drugs and metabolites in a l-ml tors. This approach effectively compensates for
urine sample. The authors recommended that changes in retention time with column deteriora-
although retention times were reproducible tion and may improve interlaboratory compar-
within-day and over extended periods of time, it isons, provided that identical chromatographic
was important that standards be run frequently, conditions are used. The system has also been
to prevent peaks falling outside the time domain used with simple ethyl acetate extracts from post
identification window of the library search rou- mortem blood, although the extract quality ap-
tine. This could lead to misidentified peaks when pears poor and unsuitable for routine screening
the spectrum was actually available. Our subse- purposes with typical forensic specimens.
quent experience with this method has been that Microbore LC has achieved popularity with
the setting of a wide window (20% or 3-5 min some users of LC, and has several advantages
absolute), is the optimum for a screening method, over normal-bore LC, particularly when PDA de-
118 B.K Logan /Anal. Chim.Acta 288 (1994) III-122
tection is being used. Along with the usual reduc- [281. This procedure uses a dual isocratic chro-
tion in mobile phase volumes, the elution volume matographic system approach with two different
for the peak is lower, therefore the concentration columns, and a multiple step fractionated manual
of the eluted drugs is consequently higher, im- liquid-liquid extraction. The authors also address
proving sensitivity by a factor of 2-3. This offsets the issue of co-elution of peaks and the use of
to some extent the lower sensitivity typically peak purity determinations to flag this.
available from diode array detectors [25]. Similar In consideration of the above methods, it
effects can be obtained through the use of low should be noted that while urine is certainly a
dispersion chromatographic techniques [26]. In a relevant forensic specimen, serum is considerably
method for drug screening in serum samples, less so and some of these methods which specifi-
microbore chromatography has been used effec- cally discuss serum would require considerable
tively in this way [21]. The authors describe reten- changes in extraction procedures to make them
tion time information for over 350 drugs, metabo- suitable for post mortem whole blood, tissue ho-
lites and artifacts. Also noted is the utility of mogenates or other typical forensic specimens.
using UV-visible spectral information to classify In-line solid phase extraction by on-column injec-
drugs according to their chromophore class which, tion with column switching has been described
with some notable exceptions, can be very helpful for post mortem vitreous humor and cere-
in identifying unknown peaks. This method was brospinal fluid sample, specifically for the identi-
used successfully with a simple dichloromethane fication of barbiturates [291 and cocaine and ben-
extraction procedure from clinical serum samples zoylecgonine [30]. Diode array detection was used
for the diagnosis of drug poisoning, but was not to confirm the identity and purity of the drug
used with autopsy material. peaks in the samples, and aided in the identifica-
An isocratic LC method with PDA detection tion of other peaks as possible metabolites. A
[27] was characterized for over 30 drugs and rapid scanning UV detector has also been used
metabolites including antidepressants, antihis- for the measurement of cocaine and metabolites
tamines, phenothiazines and analgesics in serum. in post mortem brain samples after enzymatic
The authors make the point that the two tech- digestion and solid phase extraction 1311, again
nique approach-identification and quantitation providing the added assurance of spectral infor-
chromatographically, and confirmation spec- mation in determining peak identity and peak
trophotometrically-can be sufficient to ensure purity prior to quantitation.
absolute identification to the degree of certainty By making the chromatographic conditions
required for most clinical purposes. This makes more selective for a given drug or group of struc-
LC-PDA an indispensable technique for clinical turally related compounds, the specificity of an
laboratories. The procedure described was partic- LC-PDA method can be further enhanced. A
ularly successful in identifying the class of tri- number of methods developed for drug assays in
cyclic antidepressants, and discusses the common clinical samples have been described. The experi-
problems of interference and false positives with ence of these workers validates the use of LC-
traditional single wavelength instrumentation. A PDA, and is certainly relevant to forensic drug
reversed phase ion-pair LC-DA procedure has testing. Among the most common applications is
also been described for drug screening of serum the analysis of benzodiazepines and their metabo-
samples in an emergency toxicology setting [181. lites in both serum and urine. One method [32]
This report also evaluated the discriminatory effi- allows for the specific identification of more than
ciency of five different similarity tests in library 11 benzodiazepines in serum, and utilized library
searches and concluded that a multicomponent comparison and peak purity features of the avail-
analysis provided the most reliable results. The able software. Another report validated the LC-
application of LC-PDA to the identification of PDA method against an EMIT benzodiazepine
drugs in post mortem blood samples has been assay [33]. The sample preparation includes sam-
described, using an integrated LC-PDA system ple hydrolysis resulting in conversion of the ben-
B.K.Logan/Anal. Chti Acta 288 (1994)Ill-122 119
zodiazepines and their metabolites to the corre- methods are eminently suitable for this applica-
sponding benzophenone (13 in all are character- tion, as the number of likely analytes is relatively
ized here, corresponding to 15 of the most com- low and they are well characterized and easily
monly encountered benzodiazepines and their recognizable in terms of their spectral properties.
metabolites). This conversion limits identification
of the specific benzodiazepine present, but im- Drug testing in sports
proves the sensitivity of the assay. These workers Drug testing in sports is another area of foren-
reported concordant results between LC-PDA sic science where LC-PDA is being used exten-
and EMIT in greater than 80% of cases. LC-PDA sively. Diuretics have been used by athletes to
was also used effectively for the detection, identi- reduce body weight prior to weigh-in to allow
fication and quantitation of chlordiazepoxide and them to qualify in lower weight categories. They
its metabolites in serum [34], while another have also been used to enhance the excretion of
method reported a sensitive and specific gradient other banned performance enhancing drugs. The
elution LC method for the separation and identi- identification of diuretics in the urine of athletes
fication of 19 benzodiazepines in biological fluids has therefore been an important determination in
using a rapid scanning W-visible detector [35]. sports toxicology since the International Olympic
Reported detection limits were in the 3-5 ng Committee (IOC) banned their use in 1988. LC-
ml-’ range. The method uses an efficient solid PDA has been described as a screening method
phase extraction procedure, which appears to for these compounds [40,41], and although the
work well for clinical specimens, but would prob- UV-visible spectra are probably sufficiently dis-
ably make the method unsuitable for untreated tinct on their own to allow confirmation of iden-
post mortem blood. tity, the protocol for IOC drug testing currently
requires GC-MS confirmation. The measure-
Analysis of street drug material ment of diuretics however illustrates the advan-
An LC-PDA method for the detection of man- tage that LC has over GC for screening for cer-
ufacturing by-products and impurities in illicitly tain compounds. The polar structural features of
produced cocaine has been described in which the drugs in this class make them unsuitable for
dual UV detection at 215 and 277 run was used GC without prior derivatization, making screen-
[36]. The procedure identified benzoic acid, cin- ing for their presence by GC or GC-MS a more
namic acid (cis and tram) and several isomers of tedious procedure. The xanthine stimulants such
truxillic and truxinic acid, benzoylecgonine and as caffeine, theophylline and theobromine are
cinnamoylcocaine (cis and tram). The use of banned in certain sports also. An LC-PDA
complete spectral information greatly facilitated method for the gradient elution separation and
compound identification. A9-tetrahydrocan- identification of eight of these compounds and
nabinol (THC), the active component of mari- some of their metabolites has been described,
juana was identified in extracts from THC-con- where identification is based on retention time
taining gelatin capsules, using LC-PDA to evalu- and peak identity confirmed by UV-visible data
ate peak homogeneity during the chromato- WI.
graphic optimization process [37]. Sub- and su- Alsoof concern in athletic drug testing is the
percritical fluid chromatography were also suc- use of corticosteroids as performance enhancing
cessfully attempted with a PDA detector, using drugs. LC-PDA has been shown to be capable of
carbon dioxide as the primary mobile phase com- resolving and identifying spectrally 10 commonly
ponent. This was then applied to the separation, used corticosteroids [43]. Because of the struc-
identification and purity determination of opiate tural features shared by many of these com-
alkaloids from a poppy straw extract [38]. LC-PDA pounds, first derivative spectroscopy was found to
has also been used for the identification of com- be more useful in discriminating between them
ponents, cutting agents, impurities and reaction than the spectra themselves. Methods have also
by-products in illicit heroin samples [39]. LC-PDA been described using LC-PDA for the identifica-
120 B.R Logan/Anal. Chitn. Acta 288 (1994) Ill-122
tion of more than 60 anabolic and androgenic used in cosmetics [57]. A method using multiple
steroids in illicit drug preparations [44-441. These wavelength detection has been described for the
compounds are used by athletes and also used differentiation of over 100 non-ball pen inks,
illegally in animal racing and in cattle breeding using a rapid scanning UV-visible detector [12].
and rearing, which may constitute a forensic set- This work illustrates the utility of three dimen-
ting. A method has been described for the analy- sional information sets in discriminating between
sis of these steroids in tissue from slaughtered complex mixtures of very similar composition. In
cattle using LC-PDA [47]. addition, the use of first and second derivative
The general experience of those using LC-PDA spectra, was used effectively to distinguish be-
in drug testing applications which have relevance tween ink components with similar W-visible
in forensic toxicology, represents an enthusiastic spectra.
endorsement of the technique. Features most Identification of pesticides also occasionally
commonly used include multiple wavelength mon- forms part of forensic cases, and the chromato-
itoring, library comparisons, first and second graphic properties of 51 of the common pesti-
derivative spectra, peak purity determinations and cides have been compiled [58], including UV-
peak integration features for quantitation. visible absorption maxima, and retention times.
2 R. Gill, T.S. Bal and A.C. Moffat, J. Forensic Sci. Sot., 22 33 E.I. Minder, R. Schaubhut and F. Simmler, Tozicol. Lett.,
(1982) 165. 45 (1989) 93.
3 W.A. Saner, G.E. Fitzgerald and J.P. Welsh, Anal. Chem., 34 E.I. Minder, J. Tozicol. Clin. Tozicol., 27 (1989) 117.
48 (1976) 1747. 35 P. Mura, A. Piriou, P. Fraillon, Y. Papet and D. Reiss, J.
4 L.F. Colwell and B.L. Karger, J. Assoc. Off. Anal. Chem., Chromatogr., 416 (1987) 303.
60 (1977) 613. 36 I.S. Lurie, J.M. Moore, D.A. Cooper and T.C. Kram, J.
5 A.J. Owen, The Diode Array Advantage, Hewlett Packard, Chromatogr., 405 (1987) 273.
Waldbronn, 1988. 37 G. Ray, M. Crook, N. West, M. Kwoka, G. Rehagen, J.
6 M.K. Tobias, Am. Lab., March (1989) 102. Cox, E. Murrill and K. Flora, J. Chromatogr., 317 (1984)
7 K. Aziz, Clinics in Laboratory Medicine, 10 (1990) 493. 455.
8 T. Alfredson and T. Sheehan, J. Chromatogr. Sci., 24 38 J.L. Janicot, M. Caude and R. Rosset, J. Chromatogr., 437
(1986) 473. (1988) 351.
9 B.K. Logan, H.S. Nichols, G.S. Femandez and D.T. 39 H.A. Billiet, R. Walters, G.L. De and H. Huizer, J.
Stafford, Crime Laboratory Digest, 17 (1990) 5. Chromatogr., 368 (1986) 351.
10 W.H. Hopkins, R.G. Nordman and B.G. Willis, Proc. Sot. 40 S.J. Park, H.S. Pyo, Y.J. Kim, M.S. Kim and J. Park, J.
Photo-optical Instrumentation Engineers, 191 (1979) 2155. Anal. Tozicol., 14 (1990) 84.
11 A. Knowles and C. Burgess, Practical Absorption Spec- 41 S.F. Cooper, R. Masse and R. Dugal, J. Chromatogr., 489
trometry, Chapman and Hall, London, 1984. (1989) 65.
12 I.R. Tebbet, C. Chen, M. Fitzgerald and L. Olson, J. 42 T.E. Leakey, J. Chromatogr., 507 (1990) 199.
Forensic Sci., 37 (1992) 1149. 43 S.J. Park, Y.J. Kim, H.S. Pyo and J. Park, J. Anal. Tozicol.,
13 N.H. Anderson, M.R. Gray and C.J. Hinds, J. Pharm. 14 (1990) 102.
Biomed. Anal., 8 (1990) 853. 44 J.O. De Beer, J. Chromatogr., 489 (1989) 139.
14 J.B. Castledine, A.F. Fell, R. Modin and B. Sellberg, J. 45 M. Mulholland and D. Knight, Analyst, 113 (1988) 165.
Pharm. Biomed. Anal., 9 (1991) 619. 46 M.J. Walters, R.J. Ayers and D.J. Brown, J. Assoc. Off.
15 E. Qwino, B.J. Clark and A.F. Fell, J. Chromatogr. Sci., 29 Anal. Chem., 73 (1990) 904.
(1991) 450. 47 E.H. Jansen, B.H. van Blitterswijk and R.W. Stephany,
16 E.M. Kirk and A.F. Fell, Clin. Chem., 35 (1989) 1288. Vet. Q, 6 (1984) 60.
17 A.F. Fell, B.J. Clark and H.P. Scott, J. Chromatogr., 316 48 D.N. Buchanan and J.G. Thoene, J. Biomed. Chromatogr.,
(1984) 423. 1 (1986) 38.
18 E.I. Minder, R. Schaubhut, C.E. Minder and D.J. Vonder- 49 D.N. Buchanan and J.G. Thoene, Clin. Chem., 32 (1986)
Schmitt, J. Chromatogr., 419 (1987) 135. 169.
19 G. D’Agostino and L. Castagnetta, J. Chromatogr., 338 50 D.N. Buchanan and J.G. Thoene, J. Chromatogr., 344
(1985) 11. (1985) 23.
20 B.K. Logan, D.T. Stafford, I.R. Tebbett and C.M. Moore, 51 E. Jellum, J. Chromatogr., 452 (1988) 435.
J. Anal. Tozicol., 14 (1990) 154. 52 M. Sharman, A.L. Patey and J. Gilbert, Food Addit.
21 A. Turcant, C.A. Premel, A. Cailleuz and P. Allain, Clin. Contam., 7 (1990) 649.
Chem., 37 (1991) 1210. 53 J.C. Frisvad and U. Thrane, J. Chromatogr., 404 (1987)
22 B.K. Logan and D.T. Stafford, J. Forensic Sci., 34 (1989) 195.
553. 54 J.C. Frisvad, 0. Filtenborg and U. Thrane, Arch. Environ.
23 M. Bogusz and M. Wu, J. Anal. Toxicol., 15 (1991) 188. Contam. Tozicol., 18 (1989) 331.
24 M. Bogusz, G. Neidl Fisher and R. Aderjan, J. Anal. 55 S. Bomer and R. Brenneisen, J. Chromatogr., 408 (1987)
Tozic~l., 12 (1988) 325. 402.
25 H.S. Lu and P.H. Lai, J. Chromatogr., 368 (1986) 215. 56 R. Matissek, Z. Lebensm. Unters. Forsch., 183 (1986) 273.
26 M. Mulholland, Analyst, 111 (19861601. 57 H. Schulz and G. Albroscheit, J. Chromatogr., 442 (1988)
27 P.R. Puopolo, S.A. Volpicelli, D.M. Johnson and J.G. 353.
Flood, Clin. Chem., 37 (1991) 2124. 58 M.D. Osselton and R.D. Snelling, J. Chromatogr., 368
28 E.M. Koves and J. Wells, J. Forensic Sci., 37 (1992) 42. (1986) 265.
29 B.K. Logan and D.T. Stafford, Forensic Sci. Int., 41 (1989) 59 A. Curry, Poison Detection in Human Organs, Thomas,
125. Springfield, 2nd edn., 1969.
30 B.K. Logan and D.T. Stafford, J. Forensic Sci., 35 (1990) 60 A.C. Moffat, Clarke’s Isolation and Identification of Drugs,
1303. Vol. 2, The Pharmaceutical Press, London, 1986.
31 S.P. Browne, C.M. Moore, J. Scheurer, I.R. Tebbett and 61 T. Mills and C.J. Roberson, Instrumental Analysis for
B.K. Logan, J. Forensic Sci., 36 (1991) 1662. Drug Analysis, Elsevier, Amsterdam, 2nd edn., 1987.
32 P.R. Puopolo, M.E. Pothier, S.A. Volpicelli and J.G. Flood,
Clin. Chem., 37 (1991) 701.