Raman spectroscopy in forensic analysis: identification of cocaine and other illegal drugs of abuse

Abstract

Current forensic methods for detecting and identifying cocaine and other drugs of abuse are destructive, so evidence cannot be re-analyzed. Raman spectroscopy, based on inelastic light scattering, allows for rapid, inexpensive and nondestructive analysis in forensic science. This review presents the state-of-the-art use of Raman spectroscopy as a confirmatory method for the identification of cocaine and other drugs of abuse in seized samples, including hidden compounds in legal materials such as beverages and clothes, among others, used for trafficking. Quantitative Raman spectroscopy is used to determine the actual drug concentrations in street cocaine and crack rocks and to identify possible adulterants in these samples for forensic toxicology and criminalistics. Finally, recent developments in Raman spectrometers (portable instruments and new excitation wavelengths) and advancements in data analysis offer exciting opportunities for new applications of Raman spectroscopy in the identification and quantification of drugs of abuse, including investigations conducted immediately at the scene of a crime. Copyright

Full text

Raman spectroscopy in forensic analysis:
identification of cocaine and other illegal
drugs of abuse
Ciro Augusto Fernandes de Oliveira Penido,
a
Marcos Tadeu Tavares Pacheco,
a,b
Igor K. Lednev
c
and Landulfo Silveira Jr.
a
*
Current forensic methods for detecting and identifying cocaine and other drugs of abuse are destructive, so evidence cannot be
re-analyzed. Raman spectroscopy, based on inelastic light scattering, allows for rapid, inexpensive and nondestructive analysis in
forensic science. This review presents the state-of-the-art use of Raman spectroscopy as a confirmatory method for the identifica-
tion of cocaine and other drugs of abuse in seized samples, including hidden compounds in legal materials such as beverages and
clothes, among others, used for trafficking. Quantitative Raman spectroscopy is used to determine the actual drug concentrations
in street cocaine and crack rocks and to identify possible adulterants in these samples for forensic toxicology and criminalistics.
Finally, recent developments in Raman spectrometers (portable instruments and new excitation wavelengths) and advancements
in data analysis offer exciting opportunities for new applications of Raman spectroscopy in the identification and quantification of
drugs of abuse, including investigations conducted immediatelyat the scene of a crime. Copyright © 2016 John Wiley & Sons, Ltd.
Additional supporting information may be found in the online version of this article at the publisher’s web site.
Keywords: Raman spectroscopy; drugs of abuse; cocaine; identification and quantification; forensic applications
Introduction
The widespread use of illicit drugs in contemporary society is an un-
precedented phenomenon in human history. Such drugs (e.g. co-
caine) are linked closely to increases in crime and violence. In
addition, the widespread use of illicit drugs constitutes a threat to
public health; the safety and welfare of mankind, particularly youth;
and the security and sovereignty of countries.
[1]
In a forensic context, an important aspect is the identification of
the composition of seized drugs, including fillers, so that materials
originating from different regions can be classified according to
their production sites. This information can be used to identify drug
trafficking patterns and distribution networks.
[2,3]
The characteriza-
tion of the impurities in drugs can also aid in the identification of
new laboratories manufacturing illicit drugs and new synthesis
routes and processes to obtain the drugs, which can, in turn, pro-
vide useful information to intelligence units or regulatory authori-
ties, for example by alerting authorities to monitor precursor
substances such as chemical reagents and adulterants.
[2,3]
These as-
pects are important, especially in the case of cocaine, because of
the ease of extraction and production of this drug in clandestine
laboratories and changes in the products and processes used in
extracting and refining this alkaloid.
[2,3]
Combined gas chromatography/mass spectrometry, high perfor-
mance liquid chromatography, ultraviolet spectrophotometry, Fou-
rier transform infrared spectroscopy (FTIR) and X-ray powder
diffraction are among the methods currently used in forensic toxi-
cology to detect illicit drugs, e.g. cocaine and amphetamines in-
cluding ecstasy, lysergic acid diethylamide and other synthetic
drugs. Despite their high sensitivity and molecular specificity, chro-
matographic techniques are destructive and time-consuming, and
they use reagents that are toxic to the operator and the environ-
ment. The use of optical techniques based on the absorption and
scattering of light for the rapid and precise evaluation of materials
in several fields has been reported recently.
[4–6]
Optical techniques, particularly Raman spectroscopy, have been
used for qualitative and quantitative analyses of illegal drugs and
adulterants.
[7]
The Raman technique is based on the inelastic scat-
tering of radiation by molecules and allows the evaluation of the
chemical composition of the sample. Raman spectroscopy is a rapid
and nondestructive way to perform sample characterization, does
not require chemical reagents and is not subject to interference
from water or moisture.
[8,9]
The principle advantage of this tech-
nique is its ability to be used rapidly at a crime scene, which can
prevent the arrest of innocent persons and record illegal activities
without destroying criminal evidence. Raman spectroscopy also
allows the custody chain to be maintained, given the possibility
of qualitative and quantitative evaluation of the sample to confirm
* Correspondence to: Landulfo Silveira Jr. Biomedical Engineering Institute,
Universidade Camilo Castelo Branco –UNICASTELO, Estr. Dr. Altino Bondesan,
500, São Jose dos Campos, SP, 12247-016, Brazil.
E-mail: l andulfo.silveira@gmail.com
aBiomedical Engineering Institute, Universidade Camilo Castelo Branco –
UNICASTELO, Estr. Dr. Altino Bondesan, 500, São Jose dos Campos, SP, 12247-
016, Brazil
bUniversidade Santa Cecília –UNISANTA, R. Dr. Osvaldo Cruz, 277, Santos, SP,
11045-907, Brazil
cDepartmentof Chemistry, University at Albany, State University ofNew York, 1400
Washington Avenue, Albany, NY, 12222, USA
28
J. Raman Spectrosc. 2016,47,28–38 Copyright © 2016 John Wiley & Sons, Ltd.
Review
Received: 18 September 2015 Revised: 19 November 2015 Accepted: 20 November 2015 Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/jrs.4864

its integrity from its seizure until the sample is placed in the
evidence file.
[10]
The objective of this paper is to present a literature review of the
applications of Raman spectroscopy in the forensic area, especially
in toxicology. Various types of Raman spectroscopic techniques
have been utilized for forensic purposes including portable Raman,
micro-Raman, fiber optic Raman for on-field applications, near-
infrared excitation for reducing fluorescence interference, surface-
enhanced Raman spectroscopy (SERS) to increase the sensitivity
and spatially offset Raman spectroscopy (SORS) for in-depth sample
analysis.
Fundamentals of Raman spectroscopy
Raman spectroscopy is based on the inelastic scattering of radiation
after it interacts with matter. The spectral information obtained by
Raman scattering is of a vibrational nature, through the interaction
of the incident radiation with the vibrating molecules of the mate-
rial. Inelastic light scattering was predicted theoretically by A.
Smekal in 1923
[11]
and discovered experimentally by Raman and
Krishnan in liquids
[12]
and by Landsberg and Mandelstam in
crystals.
[13]
When it is focused on a sample, incident radiation can undergo
reflection, absorption or scattering. Scattering is accompanied by
the polarization of the electron cloud of the molecule because of
the incident electromagnetic field.
[14]
If the induced dipole oscil-
lates at the frequency of the incident radiation, the emitted pho-
tons have the same wavelength as the incident radiation and
constitute elastic or Rayleigh scattering. Alternatively, the polariz-
ability of a molecular vibrational mode can change as a function
of the incident electromagnetic field, resulting in inelastic or Raman
scattering.
[8]
Therefore, Raman scattering provides information
about the molecular vibrational structure and can be used for iden-
tification purposes as a specific molecular fingerprint.
Figure 1 illustrates that the incidence of a photon may elevate
the molecule from its ground state S
0
to an exited energy state S
1
or S
2
, depending on its energy (wavelength). The excited state
can decay rapidly and nonradiatively to the lowest vibrational level
of the same excited state and then experience a decay to the
ground state S
0
, thus emitting a fluorescent photon. Alternatively,
the excited state may decay directly to any of the vibrational states
of the ground state, resulting in an inelastically scattered photon
(Raman scattering). If the energy of the incident photon is in the
near-infrared region, the molecule is elevated to a ‘virtual’energy
state that decays to any vibrational state of the ground level and
emits a photon without experiencing absorption.
[15–17]
The fre-
quencies of the vibrational energies observed in the Raman spec-
trum, known as Raman shifts, are determined by the difference
between the frequencies of the incident and scattered radiation.
According to Fig. 1, the use of near-infrared excitation (typically
between 750 and 1000 nm) provides a way to obtain Raman scat-
tering with little interference from molecular absorption. Absorp-
tion competes with scattering and may cause heating or
photodamage to the sample, as well as fluorescence interference
in the case of fluorescent molecules.
There are several different experimental methods for acquiring
Raman scattered photons and obtaining Raman spectra. Most of
the widely used methods are briefly described here.
Dispersive and Fourier transform Raman spectroscopy
To obtain the Raman spectrum, two types of instruments are typi-
cally used: dispersive and Fourier transform Raman (FT-Raman)
spectrometers. The dispersive spectrometer uses an imaging spec-
trograph in its collection system, with a diffraction grating and a
multichannel detector, usually a charge-coupled device (CCD) sili-
con camera. Raman scattering is obtained by directing an excitation
laser onto the sample. The scattered radiation is collected by an op-
tical system coupled to the spectrograph through a slit aperture,
dispersed by a diffraction grating and focused on the CCD to be
converted into an electrical signal for computer reading and stor-
age. The inelastic scattering occurs within the sensitivity range of
the CCD, which is usually from the visible region to 950 nm. In a dis-
persive spectrometer, the spectrum is collected in a single exposure
of the CCD camera, and the spectral resolution depends on the
number of grooves of the grating and the width of the spectro-
graph entrance slit.
[8]
An FT-Raman spectrometer usually uses a laser in the infrared re-
gion (1064 nm) and an interferometer, which records an interfer-
ence pattern after the radiation from the sample is reflected by a
moving and a fixed mirror and combined in an interferometer. This
interferogram is detected by a germanium single channel detector,
and software transforms the interference pattern into a spectrum
through the mathematical analysis called Fourier transformation.
[16]
The main advantage of FT-Raman is the use of a wavelength that is
poorly absorbed by fluorescent molecules, but the spectral resolu-
tion is dependent on sample timing.
[18,19]
The use of dispersive Raman spectrographs with excitation
wavelengths up to 830 nm provides higher efficiency compared
with FT-Raman instruments with 1064 nm excitation because of
the dependence of the scattering intensity on the light wavelength
(Iα1/λ
4
). The dispersive Raman approach also has better spatial res-
olution in the case of microscopy and faster data acquisition be-
cause the spectrum is recorded during a single reading of the
CCD camera. However, for highly fluorescent samples, FT-Raman
technique serves as a viable solution to the noise generated by
the fluorescence. Currently, dispersive Raman systems using multi-
channel detectors and infrared excitation (1064 nm) are presented,
with the possibility ofobtaining spectra with a good signal-to-noise
ratio in fluorescent samples quickly and with good resolution.
[20]
Figure 1. Simplified Jablonski diagram for the processes of absorption,
emission and Raman scattering. From Hanlon et al.,
[5]
with permission
from IOP Publishing. All rights reserved.
29
Raman spectroscopy in forensic toxicology
J. Raman Spectrosc. 2016,47,28–38 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

Surface-enhanced Raman spectroscopy
The Raman scattered signal can be significantly increased by the
plasmonic effect at metal surfaces, called SERS. The increase in
the Raman signal because of the SERS effect was discovered by
assessing pyridine on a rough silver surface.
[21]
The SERS effect oc-
curs when plasmonic materials, which are traditionally noble and
coinage metals (e.g. silver, gold and copper) with nanoscale fea-
tures, are illuminated with light. The enhancement results from
the amplification of the electromagnetic field by the excitation of
localized surface plasmon resonances.
[22]
The electromagnetic field
enhancement occurs preferentially in the gaps, crevices and sharp
features of the material surface. Depending on the structure of
the supporting plasmonic material, the enhancement factor can
be approximated by the increase in the amplitude of the electro-
magnetic field by 10
4
–10
10
.
[23]
Other mechanisms involved in Ra-
man signal enhancement are related to the metal–molecule
charge transfer, where the excitation wavelength is resonant with
the charge transfer electronic transition.
[24]
This mechanism, called
chemical enhancement, is responsible for Raman signal enhance-
ments on the order of 10
1
–10
3
.
Spatially offset Raman spectroscopy
The SORS can beused to obtain Raman scattering from compounds
located under opaque surfaces.
[25]
In SORS, the radiation scattered
by the material is collected in backscattering geometry from points
that are spatially (laterally) displaced from the excitation point. SOR
spectra consist of contributions from various layers of the sample
located at different depths. Most importantly, the surface signal be-
comes less prominent, and the collected SOR spectra contain a
greater contribution from subsurface components when the dis-
tance between the illumination and collection points increases. In
addition, the Raman spectrum collected from the surface without
excitation–collection offset could be quantitatively subtracted from
the SOR spectra to further increase the desirable contribution from
the deep layers of the sample.
[25,26]
The discrimination of the spec-
tra collected from deeper layers from the spectra collected from
surface can be achieved using a variety of multivariate analysis
methods, being PCA the most used.
[25,27]
The SORS technique al-
lows chemical information about a sample inside a sealed container
to be obtained, for example plastic containers, including bottles
using suitable probe and sample geometry.
[25–27]
Raman spectroscopy in forensic toxicology
Raman spectroscopy and its main variants, including resonance Ra-
man spectroscopy, SERS and SORS, have been proposed as tech-
niques to be used in routine analysis in forensic toxicology, with
the main advantage based on the possibility of a rapid, accurate,
nondestructive and reagent-free analysis. One of the first attempts
to use Raman spectroscopy for evaluating drugs, including cocaine,
was made by Hodges et al.
[28]
Since then, the selection of Raman as
the technique of choice in drug analysis has been increasing.
The advances in the detection and identification of drugs of
abuse for forensic purposes using Raman spectroscopy have been
reviewed,
[29–31]
emphasizing the desirable characteristics of the
technique of choice for forensic applications, such as preserving
the chain of custody, performing nondestructive analysis and offer-
ing the possibility of in situ rapid determination and quantification
of the purity of a drug. These applications include identification
and quantification of drugs of abuse in different types of forensic
evidence, including bulk street drugs and traces found in drinks,
fibers/clothing, fingerprints, fingernails, banknotes and body fluids.
Discriminating cocaine from other illicit drugs
The determination of the main spectral features of several types of
drugs of abuse and the discrimination of these drugs according to
their spectral features using a suitable discrimination technique
have been demonstrated. Ryder
[32]
applied PCA to the near-
infrared (785 nm) dispersive Raman spectra to discriminate be-
tween cocaine, heroin and methamphetamine (MET) diluted with
several ‘cutting’materials, even when the spectra were visually sim-
ilar. Leger and Ryder
[33]
conducted a study showing that the pre-
processing technique could affect the quality of the quantification
of drugs (cocaine, heroin and MET) in mixed samples. They success-
fully utilized the method of the first derivative of the Raman spectra
and background subtraction by a ‘modified’polynomial baseline
fitting.
Identifying the type of cocaine present
Cocaine typically occurs in freebase and hydrochloride (HCl) chem-
ical forms, with the physical appearance of crack rocks (freebase),
paste (freebase) or powder (freebase or hydrochloride). While free-
base rocks and paste are usually smoked, freebase and hydrochlo-
ride powders are typically inhaled. The HCl form is also used for
arterial/venous injection as it is soluble in water. Therefore, the
identification of the type of cocaine is important for determining
the potential for addiction, as it influences the dose consumed in
an overdose. Additionally, the route of traffic may be indicated by
the form of the drug at the time of seizure.
The differences in the Raman spectra of freebase cocaine and co-
caine HCl were demonstrated for the first time by Carter, Brewer
and Angel.
[10]
Additionally, these authors demonstrated that the
spectra of commonly used cutting agents, such as benzocaine
and lidocaine, are different from the spectra of cocaine. Penido
et al.
[34]
also showed the differences in the Raman spectra of seized
samples of freebase cocaine, freebase cocaine paste, cocaine HCl
and crack rocks, as well as common cutting agents such as lido-
caine, benzocaine, caffeine, sodium carbonate and aluminum sul-
fate. They also showed spectral differences in the Raman spectra
of freebase paste and crack that are compatible with degradation
products as benzoic acid and benzoylecgonine.
Figure 2 shows the spectra of cocaine in different forms, as
shown in Penido et al.
[34]
Table 1 presents a tentative assignment
of Raman bands based on the recent literature. Common Raman
bands for cocaine (freebase, HCl and crack) appear at 848, 874
and 898 cm
1
(C–Cstretching–tropane ring); 1004 cm
1
(symmet-
ric stretching–aromatic ring); 1279 cm
1
(C–Nstretching);
1453 cm
1
(asymmetric CH
3
deformation); and 1605 and
1712 cm
1
(C=C and C=O stretching). The cocaine HCl exhibits Ra-
man bands with similar peak positions but lower relative intensity
of the 848 and 898 cm
1
(tropane ring) bands. The bands at 1026
and 1207 cm
1
in the spectrum of the HCl form correspond to
the 1036 and 1183 cm
1
bands, respectively, in the freebase co-
caine spectrum. The protonated nitrogen due to the presence of
hydrochloric acid in hydrochloridric cocaine changes the polariz-
ability of the tropanic ring and causes spectral differences between
freebase and HCl cocaines. Specifically, peaks at 1183, 1319 and
1735 cm
1
, which are characteristic of freebase cocaine, are not
present in the spectrum of the HCl form.
Laussmann et al.
[39]
presented an interesting case of the identifi-
cation of cocaine in a suspicious black powder confiscated by
30
C. A. F. de Oliveira Penido et al.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2016,47,28–38

German customs. Cocaine, mixed with copper, iron, thiocyanate
and graphite, was identified by the use of several analytical tech-
niques, including Raman spectroscopy. The authors suggested that
the graphite may have beenused to mask the infrared spectrum of
cocaine, but Raman spectroscopy was able to detect and differenti-
ate these compounds.
Identifying cocaine in the ‘real world’in a forensic context
Raman spectroscopy has been evaluated as a possible tool for the
identification of cocaine in various evidentiary samples that could
be present at a crime scene, such as fingerprints and nails, as well
as materials that are commonly used to hide and transport the
drug, such as soaked textile fibers. The main advantages of optical
techniques such as Raman spectroscopy are that little or no sample
preparation is needed, no visible alteration of the evidentiary mate-
rial occurs and the chain of custody is maintained. The main disad-
vantage of using Raman spectroscopy is related to the time spent
to locate a drug crystal on the surface of the suspected material.
This process could be automated by using scanning Raman
microspectroscopy.
The possibility of detecting illegal drugs, including cocaine, in la-
tent fingerprints at a crime scene has been investigated. Day et al.
[40]
reported on the detection of cocaine HCl, codeine phosphate,
Figure 2. Raman spectra of cocaine samples in different forms: (A) yellow
freebase paste, (B) white freebase paste, (C) crack rock, (D) hydrochloride
cocaine powder, and (E) freebase cocaine powder, with their characteristic
peaks labeled according to Table 1. The presence of the peak at
1639 cm
1
in samples B and C suggests the occurrence of benzoic acid, a
degradation product of cocaine. Additionally, the differences in the
intensities of the peaks a t 1605 and 1712 cm
1
indicate degradation into
benzoylecgonine. Adapted from Penido et al.,
[34]
with permission from
John Wiley & Sons. All rights reserved.
Table 1. Peak positions of the main Raman and FTIR bands of different forms of cocaine and adulterants with their tentative vibrational assignments
based on the recent literature and the chemical structure
Compound Raman peak
position (cm
1
)
Assignment
[10,18,35–38]
Chemical structure
a
Cocaine: freebase, HCl,
paste and crack
848, 874, 898 C–C stretching (tropane ring) Freebase cocaine
1004 Symmetric stretching –aromatic ring breathing
1036 Asymmetric stretching –aromatic ring
1165 C–Nstretching
1183 C–Nstretching
1279 C–Nstretching
1319 C–H twisting
1453 Asymmetric CH
3
deformation
1605 C=C stretching –aromatic ring
1712 C=O symmetric stretching –carbonyl
1735 C=O asymmetric stretching –carbonyl
Peaks exclusive to
cocaine HCl
848, 898
b
C–C stretching (tropane), change the
polarizability by protonated nitrogen
Cocaine HCl
1026
b
Asymmetric stretching –a romatic r ing (freebase at 1036 cm
1
)
1207
b
C–N stretching shifted by the protonated
nitrogen (freebase at 1183 cm
1
)
1462
b
Asymmetric CH
3
deformation (freebase at 1453 cm
1
)
1596
b
C=C stretching –aromatic ri ng (freebase at 1605 cm
1
)
1601
b
C=C stretching –aromatic ri ng (freebase at 1605 cm
1
)
1716
b
C=O symmetric stretching –carbonyl
(freebase at 1712 cm
1
)
Peaks exclusive to paste cocaine 1069
c
O–C–O asymmetric stretching of sodium carbonate
(contaminant or adulterant)
1639
c
C=O stretching –benzoic acid
Peaks exclusive to crack cocaine 1639
c
C=O stretching –benzoic acid
Adapted from Penido et al.
[34]
a
Available from http://pubchem.ncbi.nlm.nih.gov/compound/ by searching ‘cocaine’and ‘cocaine HCl’.
b
Indicates Raman peaks with differences in the position and intensity, caused by protonated nitrogen because of the presence of hydrochloric acid in
cocaine HCl compared with freebase cocaine.
c
Indicates Raman peaks of adulterants or degradation products in the samples of freebase cocaine.
31
Raman spectroscopy in forensic toxicology
J. Raman Spectrosc. 2016,47,28–38 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

amphetamine sulfate, barbital and nitrazepam, and other noncon-
trolled substances used to adulterate the drugs in contaminated
fingerprints on a steel slide. The same authors applied cyanoacry-
late fuming (‘super glue’) as an enhancer agent to the contami-
nated slide.
[41]
According to the authors, spectral features from
contaminants could be detected easily even with the cyanoacrylate
polymer. However, the most difficult aspect ofevaluating latent fin-
gerprints was to locate the suspect substance visually in the finger-
print to obtain a Raman spectrum. West and Went
[42]
evaluated the
possibility of using Raman spectroscopy in the detection of drugs of
abuse such as cocaine, ecstasy, ketamine and amphetamine in la-
tent fingerprints that had been treated with powders and also sub-
sequently lifted with adhesive tape. The authors verified that the
application of aluminum-based or iron-based powders to the con-
taminated fingerprints did not interfere with the Raman spectra ob-
tained for the contaminants. Additionally, contaminated
fingerprints developed with powders and lifted with tapes
contained bands from the tape, but these bands did not interfere
with the detection process.
Molina Moreno et al.
[43]
evaluated the possibility of the detection
of drug particles collected with the aid of transparent and colored
adhesive tapes. They recorded spectra of drugs, including cocaine,
metabolites and cutting agents, for a standard database, and the
spectra of some compounds using a fingerprint lifting tape and
two colored (white and green) packaging tapes. The confocal Ra-
man spectra of the particles could easily identify the drug, although
the presence of contaminants such as sugar and textile fibers in the
collection area interfered with the measurements and increased
the analysis time.
Human nails can be contaminated after handling or abusing ille-
gal substances. In this context, the detection of residues from drugs
and adulterants in a forensic context has been proposed by Ali
et al.
[36]
The authors evaluated the possibility of identifying crystals
(5–20 μm in size) of pure cocaine HCl, seized cocaine HCl (77% pu-
rity) and paracetamol on the nail surface using a confocal micro-
Raman spectrometer. Drug crystals were detected even under a
layer of a red nail varnish applied after the contamination of the nail
with the drug. Figure S1 (Supporting Information) shows the spec-
tra of a crystal of pure cocaine HCl into the surface of the nail and a
microphotography of the crystal into the nail surface (from Ali
et al.
[36]
).
The possibility of detecting drugs in clothes or in textile fibers
is of extreme interest because these materials can hide drugs
within their fibers, including textile fibers lifted by the adhesive
tape at the scene of the crime and stored in evidence files. Ali
et al.
[44]
studied the possibility of detecting and identifying
drugs of abuse (cocaine HCl and MET HCl) in situ on undyed nat-
ural synthetic fibers and colored textile specimens using confo-
cal micro-Raman. Despite the presence of spectral Raman
bands and background fluorescence arising from the natural
and synthetic polymers and dyed textiles, the drugs could be
identified by their characteristic Raman bands in less than
3 min, even when they were trapped between the fibers. West
and Went
[45]
evaluated the detection of drugs in textile fibers
contaminated with ecstasy, cocaine, ketamine and amphet-
amine after being lifted with adhesive tapes and evaluated by
Raman spectroscopy. This study showed that the drugs could
be identified in the adhesive tape even when the spectra were
recorded through the wall of the evidence bag. Ali et al.
[46]
de-
scribed the application of benchtop and portable Raman spec-
trometers for the in situ detectionofcocaineHClinnaturaland
synthetic fibers and dyed textiles impregnated with the drug.
They found that high-quality spectra of the drug could be ac-
quired in situ within seconds and without any sample prepara-
tion or alteration of the evidential material. Ali et al.
[47]
used
confocal micro-Raman to detect trace amounts of street samples
of cocaine HCl and MET trapped between the fibers of natural
and synthetic textiles without significant interference from the
bands of the fiber substrates. Ali and Edwards
[48]
evaluated the
application of a portable Raman spectrometer for the analysis
of natural and artificial fiber items of clothing impregnated with
drugs of abuse. They focused on the growing problem of the im-
pregnation of items of clothing with drugs of abuse for smug-
gling through airports and other ports of entry. Textile pieces
were soaked with cocaine HCl, MET HCl and amphetamine sul-
fate diluted in water/ethanol solutions. With the use of PCA,
the feasibility of automatic spectral recognition was shown.
Figure S2 (Supporting Information) shows the scanning electron
micrograph of a piece of wool impregnated with cocaine HCl
and the spectra of drug-impregnated wool taken with different
Raman instruments (from Ali and Edwards
[48]
), showing distinct
peaks that can be used to identify and discriminate the drugs.
The potential of SORS to detect cocaine concealed inside
transparent glass bottles containing alcoholic beverages has
been demonstrated by Eliasson et al.
[37]
The detection of cocaine
in a brown bottle of rum required only 1 s. This application holds
promise as a method for the fast, noninvasive analysis of
suspected beverages, permitting drug trafficking to be com-
bated more effectively.
Quantifying cocaine in seized and adulterated samples
Raman spectroscopy has been proposed for quantifying drugs in
seized and ‘cut’samples. Modern drug laws require that a seized
sample be characterized for both the illegal substances present
and the quantity of each of those substances. The correct quan-
tification of the drug is important for the chain of custody, given
the possibility of quantitative evaluation of the sample from sei-
zure to the laboratory without sample destruction, proving its
integrity.
Hodges et al.
[28]
employed FT-Raman spectroscopy to identify
cocaine HCl, heroin and amphetamine sulfate in samples diluted
(‘cut’) with several ‘cutting’agents commonly encountered in
seizures. Some ‘cutting’agents exhibited fluorescence, compli-
cating the identification of the drug. Raman features of cocaine
could be seen in concentrations as low as 5% when the cocaine
was ‘cut’with sorbitol. In the same context, Ryder et al.
[49]
employed micro-Raman spectroscopy with 785 nm excitation
to quantify binary mixtures of narcotics (cocaine, heroin and
MET) with solid diluents (i.e. flour, baby milk formula, glucose,
maltose, talc and sodium bicarbonate) at different concentra-
tions. An accurate prediction model was generated using a mul-
tivariate algorithm based on partial least squares (PLS) that can
predict the concentration of cocaine in glucose from a single Ra-
man spectrum with a prediction error of 2.3%. In 2000, the same
authors obtained Raman spectra from a series of 33 solid mix-
tures containing cocaine, caffeine and glucose (9.8–80.6% by
weight of cocaine) and used the PCA to verify that the first three
PCA factors have the spectral information related to the concen-
tration of cocaine.
[50]
A PLS model was developed to predict the
concentration of cocaine in the mixtures of caffeine and glucose,
with prediction error as low as 4.1% for cocaine.
Aiming to achieve the quantification of cocaine mixed with com-
mon ‘cutting’agents, Penido et al.
[51]
used dispersive Raman
32
C. A. F. de Oliveira Penido et al.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2016,47,28–38

spectroscopy (830 nm excitation) and FTIR to estimate the concen-
tration of crack freebase in binary mixtures made with adulterants
such as sodium carbonate, caffeine, benzocaine and lidocaine
(20%, 40%, 60% and 80% by weight of crack and adulterant). A prin-
cipal components regression (PCR) model was built based on the
PCA scores. The authors found that Raman spectroscopy allowed
for as low as 7% prediction error for the crack–caffeine mixture,
which was significantly lower than that of the FTIR. Figure 3 pre-
sents the Raman spectra of binary mixtures showing differences
in the intensities of the main peaks of cocaine and adulterant,
and Fig. 4 presents the estimated percentage of each binary mix-
ture based on the PCR model.
[51]
Identification and quantification of other alkaloids
The possibility of using the Raman technique for the identification
and quantification of drugs and controlled substances is not re-
stricted to cocaine. Although not yet used routinely, Raman spec-
troscopy has been proposed to identify and quantify alkaloids
with similar structures (heroin, morphine and codeine). Morphine
is the main alkaloid from Papaver somniferum and directly affects
the central nervous system. Morphine is used as a potent analgesic
in oncologic patients and as a narcotic.
Zhao et al.
[52]
presented a study in which a micro-Raman spec-
trum was measured with 514 nm excitation in trace contraband
heroin, using a slick Al slice as a gasket for the drug. Rana et al.
[53]
showed that morphine, codeine and hydrocodone exhibit signifi-
cant differences in their Raman spectra despite the similarities in
their structures. Additionally, they showed that the dispersive spec-
trum obtained at 785 nm excitation showed a poor signal-to-noise
ratio compared with SERS spectra because of strong fluorescence
interference. Figure S3 (Supporting Information) presents the Raman
and SERS spectra of morphine and codeine, indicating spectral
differences suitable for discrimination in both normal Raman and
SERS. Baranska and Kaczor
[54]
obtained Raman spectra of morphine
and its metabolite morphine-3-glucuronide (which exhibits an
antagonistic biological effect), presenting an interpretation of the
vibrational spectra based on computational models, thus allowing
an ‘in situ’identification of its active form.
Ivanova and Spiteller
[55]
reported the Raman spectra of cytosine
derivatives, which are of interest for forensic chemistry. These alka-
loids have an established biochemical profile related to the binding
affinity of the nicotinic acetylcholine receptors (anti-cancer activity),
and the authors reported a case of fatal intoxication with cytosine.
Identification of amphetamines
Beyond cocaine identification, one of the most prominent applica-
tions of Raman spectroscopy is the detection of amphetamines.
The most common amphetamines, phenethylamine, ephedrine
and 3,4-methylenedioxy-methamphetamine (MET, known as ec-
stasy) are stimulants of the central nervous system used to treat
obesity that also cause cerebral necrosis vasculitis and cardiovascu-
lar problems.
[56]
MET has become an epidemic in the United States,
with a number of clandestine laboratories producing the drug.
[18]
Amphetamines can be synthesized from the precursors ephedrine
and pseudoephedrine. Most studies with the objective of the
detection of amphetamines are based on SERS, which will be
described in the Section on SERS for the Identification and Quanti-
fication of Cocaine and Other Drugs of Abuse.
Frederick et al.
[57]
evaluated the feasibility of detecting single
drug crystals, including amphetamine, onpaper currency by means
of Raman spectroscopy. Paper bills ($) were contaminated with two
nonpharmacologically active drug surrogates, isoxsuprine and
norephedrine, and two common excipients, benzocaine and lido-
caine. Despite a high background fluorescence coming mostly from
the light green and white areas, the spectra of individual crystals of
drugs and excipients are of sufficient quality for drug identification.
Following the same approach, Noonan et al.
[38]
reported a method
to reduce the fluorescence background from the paper currency by
photobleaching. They also determined the percent composition of
individual drug components in mixtures by systematically sampling
the surface of the dollar bill.
Kataien et al.
[58]
proposed a method to quantify the amphet-
amine concentration in seized street samples by dissolving the
Figure 3. Raman spectra of binary mixtures of crack with the following: A,
sodium carbonate; B, caffeine; C, benzocaine; andD, lidocaine. Adapted from
Penido et al.,
[51]
with permission from Taylor and Francis. All rights reserved.
33
Raman spectroscopy in forensic toxicology
J. Raman Spectrosc. 2016,47,28–38 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

sample in an acidic solution of sodium dihydrogen phosphate,
which was used as internal standard. A model based on PLS was de-
veloped based on the second derivative spectrum of the samples
using the liquid chromatography quantification as a standard.
Weston
[59]
evaluated mixtures of ‘crystal meth’(methylsulfone
and methamphetamine) by dissolving the mixtures in water and
collecting the spectra of the aqueous solution. By comparing these
spectra with spectra of methylsulfone and methamphetamine
mixtures of known composition, the composition of the sample
was obtained in minutes, demonstrating the suitability of the
method for a practical semi-quantitative analysis.
Berg et al.
[60]
evaluated differences in the normal Raman and
SERS spectra of amphetamine and amphetamine-H+ and between
different conformers. The vibrational bands from salt anions with
internal bonds (sulfates, hydrogen phosphates etc.) need to be con-
sidered when employing these spectra for identification purposes.
Furthermore, discrimination between the free and protonated
forms of amphetamine salts can be achieved using Raman spec-
troscopy, which can be utilized by the forensic community for drug
profiling studies. Fenton et al.
[61]
developed a PLS regression model
to quantify the concentrations of simulated street-drug samples by
‘cutting’samples of drug surrogates, including isoxsuprine,
norephedrine, benzocaine and lidocaine, with variable concentra-
tions of ‘cutting’agents such as mannitol, lactose, methylsulfonyl
methane, procaine HCl, baking soda, baking powder and corn
starch. The PLS model yielded prediction errors in the 4% range
using two latent variables.
Taplin et al.
[62]
measured the Raman spectra of phenethylamine,
ephedrine and MET and made assignments by calculating the
harmonic vibrational frequencies using density functional theory.
In this study, the authors also presented a set ofdiscriminant bands,
useful for distinguishing the three compounds, despite their struc-
tural similarities. Triplett et al.
[63]
used Raman spectroscopy in a
screening test for amphetamines from clandestine laboratory liq-
uids. Solutions of MET HCl in ethanol, diethyl ether and Coleman
fuel were prepared in concentrations ranging from 0.5% to 10%
w/v, and a peak at 1003 cm
1
was correlated with the
concentration of MET. Figure S4 (Supporting Information) shows
the Raman spectra of MET HCl and the solvents as presented in
the study of Triplett et al.,
[63]
showing the distinct peaks of MET
HCl compared with the solvents, particularly the one at 1003 cm
1
.
Identification of other substances of abuse: ‘legal highs’
‘Legal highs’are substances that are not covered by current drug
laws but present psychoactive effects and are used recreationally
as substitutes for illegal drugs. The identification of these ‘legal
highs’is challenging, as they often do not match their label claim
and contain a wide range of impurities and/or adulterants. There-
fore, the identification of these legal substances used in cases of
abuse, including newly developed drugs, is an area of toxicology
that can benefit from Raman spectroscopy.
The work of Stewart et al.
[64]
showed Raman spectra of a repre-
sentative range of ketophenethylamines, a rapidly growing family
of cathinone-related ‘legal highs’. These spectra showed changes
that were associated with the pattern of substitution on the aro-
matic rings and on the ethylamine substituent, thus providing a
tool for rapid identification of this class of drugs in seized samples
and the rapid detection of minor constituents, giving a composition
profile that can be used for drug intelligence work. Christie et al.
[65]
evaluated the discrimination of cathinone regioisomers from
mephedrone and flephedrone, identified in ‘legal highs’products
in Irish head shops, through the Raman spectra of the respective
isomers (methylenedioxypyrovalerone and naphyrone).
Montalvo et al.
[66]
employed confocal Raman microscopy to
obtain a spectral signature of benzodiazepine products, which are
widely used in drug-facilitated crimes. PCA showed spectral vari-
ability among batches of the same benzodiazepine drug, mainly
attributed to the heterogeneity of the drug products. Spectral
differences as a function of the dose, type of active ingredient used
and manufacturer were found. For low doses, the differentiation
was achieved by excipients because the spectral contribution of
the active ingredient was almost absent.
Figure 4. Plot of the estimation of each binary mixture of crack and adulterant (sodium carbonate, caffeine, benzocaine and lidocaine, as indicated in each
plot) with mathematical concentrations (lines) and the expected concentrations of real samples (squares) using the Raman spectra. Adapted from Penido
et al.,
[51]
with permission from Taylor and Francis. All rights reserved.
34
C. A. F. de Oliveira Penido et al.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2016,47,28–38

In a recent study, Assi et al.
[67]
constructed spectral libraries (near-
infrared, Raman and attenuated total reflectance spectroscopies) of
three substances found in ‘legal highs’, including dextromethor-
phan, 2-aminoindane and lidocaine and their mixtures with caf-
feine. A model dilution of these substances with caffeine (5–95%)
was employed to evaluate the concentrations of the ‘legal highs’
found in Internet products. The ‘legal high’constituents in most
of the model mixtures were identified within a minimum range of
20–75% by Raman spectroscopy, demonstrating that simple library
mixtures could be used to identify test substances. Some samples
presented higher fluorescence, making the quantification more
difficult.
Identification of drugs of abuse and metabolites in biological
tissues and fluids
The molecular information provided by the Raman spectroscopic
fingerprint is of interest in forensic toxicology and pathology be-
cause residues of drugs and medicines may be assessed in situ,non-
destructively, by biopsy evaluation. The possibility of detecting
drugs of abuse behind bulk tissues has been demonstratedby Ozer
et al.
[68]
Raman spectra of morphine HCl were collected through the
wall of a lamb stomach with 780 nm excitation (optical window),
and the spectral library of the dispersive micro-Raman spectrome-
ter matched the spectrum of morphine HCl as the spectrum of
codeine (methyl morphine) in the spectrum of the tissue with
morphine HCl behind the tissue.
Urine samples contaminated with antibiotics, pain relievers and
barbiturates have been evaluated by Farquharson et al.
[69]
by using
a developed silver-doped sol-gel active medium for trace drug anal-
ysis in urine using SERS measured on a prototype FT-Raman spec-
trometer. Inscore et al.
[70]
used gold-doped and silver-doped sol-
gels immobilized in glass capillaries to detect drugs of abuse and
metabolites in saliva. The method could detect the presence of
the drugs amphetamine, diazepam and methadone in contami-
nated saliva, with a consistent detection of 50 ppb cocaine.
Very recently, D’Elia et al.
[71]
reviewed the application of vibra-
tional techniques (infrared, Raman and SERS) in detecting and
quantifying drugs of abuse in oral fluid (saliva). The researchers
aim to lower the limit of detection to detect quantities of drugs
smaller than the cut-off limits established by law for drug controls.
The authors summarize that SERS has been shown to be the most
sensitive technique for the detection of illicit drugs in oral fluid.
SERS for the identification and quantification of cocaine and
other drugs of abuse
Because of a significant increase in Raman scattering efficiency by
the surface plasmons on nanostructured metal surfaces and metal
nanoparticles (colloids), SERS has been attracting interest in the tox-
icology field. A number of papers have proposed new SERS-active
media for the detection of drug traces. Horvath et al.
[72]
evaluated
the SERS detection of heroin, codeine and cocaine samples after
separation on a thin layer chromatographic gel and subsequent
coverage with silver colloids. A detection limit of 0.2μg/mm
2
was
achieved. The authors observed a weak matrix effect from the
silicate-based stationary phase, which allowed SERS to be used as
an identification method. To enhance the reproducibility of the
manufacture of active SERS substrates, Lee et al.
[73]
proposed
gold-doped and silver-doped sol-gels that provide a unique envi-
ronment for stabilizing SERS-active metal particles in the porous
silica network. Sulk et al.
[74]
demonstrated the use of SERS for the
detection of amphetamine and methamphetamine using a cou-
pling reaction of the amines with 2-mercaptonicotinic acid prior
to the surface derivatization. Detection limits of 19 and 17 ppm
were found for amphetamine and methamphetamine, respectively.
Carter et al.
[10]
reported a SERS spectrum of freebase cocaine in a
colloidal silver solution, as shown in Fig. 5. Rana et al.
[53]
detected
spectral differences of morphine, codeine and hydrocodone by
measuring SERS directly in drug solutions. Yang et al.
[75]
reported
on the ultrasensitive SERS of cocaine and an organophosphate
pesticide by using metastable silver nanoparticles.
Sagmuller et al.
[76]
proposed to combine liquid chromatography
and SERS for drug detection and identification. They separated
individual elements using an acetonitrile-free eluent and pipetted
fractions onto a microliter plate containing a matrix-stabilized silver
halide dispersion for SERS analysis. The detection limit was as low as
1μg per well of the plate for cocaine, heroin and amphetamine.
Faulds et al.
[77]
compared methods based on SERS using colloid
suspensions and vapor-deposited films of silver and gold to detect
amphetamines. The authors found that gold colloid and vapor films
presented lower detection limits, as low as 10 μmol/dm
3
, compared
with silver. Ackermann et al.
[78]
proposed a microfluidic device with
a liquid/liquid two-phase-segmented flow system and SERS detec-
tion for the online monitoring of the drug promethazine (used as a
legal high).
Chen et al.
[79]
reported a proof of principle of a reagentless
aptameric sensor based on SERS that can determine cocaine at a
concentration of 1 μM under optimized assay conditions. Sanles-
Sobrido et al.
[80]
detected the cocaine metabolite benzoylecgonine
by employing a biointerface (a monoclonal antibody) supported on
silver-coated carbon nanotubes. Farquharson et al.
[81]
evaluated the
ability of SERS based on fused gold colloids trapped within a porous
glass matrix to detect drugs in saliva in ng/ml concentrations in less
than 10min. The method was designed to detect drug-related
problems resulting in visits to emergency rooms rapidly. Detection
of 50 ng/ml of cocaine in saliva could be achieved.
Fox et al.
[82]
used gold-coated films with electrostatic lifting as a
method of residue extraction for SERS that provides ultra-trace
analysis of drug residues and other chemicals of interest. Yu and
Figure 5. Raman spectra of (a) cocaine · HCl and (b) freebase cocaine taken
with a silver colloid solution. The spectra show significant differences
between the two forms. Adapted from Carter et al.,
[10]
with kind
permissions of the authors and of the Society for Applied Spectroscopy
conveyed through Copyright Clearance Center. All rights reserved.
35
Raman spectroscopy in forensic toxicology
J. Raman Spectrosc. 2016,47,28–38 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

White
[83]
demonstrated a low-cost paper-based SERS substrate
based on a gold colloid inkjet printed on a paper-based surface
swab, mounted with a lateral-flow dipstick to collect analyte mole-
cules directly from a large-area surface. A detection limit as low as
9 ng of heroin and 15 ng of cocaine was achieved. Most recently,
Massarini et al.
[84]
systematically evaluated commercial SERS sub-
strates in terms of limit of detection, limit of identification and dy-
namic range for common narcotic drug analytes, including
cocaine, amphetamine and morphine.
SORS for identification of drugs inside containers and
packages
The emerging SORS technique has been described as a relevant
technique for forensic applications, including screening of illicit
drugs hidden under and inside packages or concealed inside food-
stuffs and beverages. SORS was used by Eliasson et al.
[37]
for detect-
ing dissolved cocaine inside a transparent glass bottle containing
an alcoholic beverage (rum) in less than 1 s, with a detection limit
estimated to be of the order of 9 g of pure cocaine per 0.7l of
rum. Figure S5 (Supporting Information) presents the Raman spec-
tra of rum and a mixture of rum and 70% purity cocaine (430 g/L)
taken through the brown glass bottle with the laser positioned
45° with respect to the spectrometer optical input, as presented
by Eliasson et al.
[85]
The peaks of cocaine are clearly distinguishable
from the spectra of ethanol of rum.
Olds et al.
[86]
demonstrated the SORS technique in several sce-
narios that are relevant to drug detection and forensics applica-
tions. The examples include analysis of a multilayered postal
package to identify a concealed substance; identification of a med-
icine inside its plastic blister pack; analysis of an envelope contain-
ing a powder; and identification of a drug dissolved in a clear
solvent, contained in a nontransparent plastic bottle. In a subse-
quent study, Olds et al.
[87]
demonstrated the noninvasive quantifi-
cation of ternary drug mixtures inside an opaque plastic container
employing multivariate statistics applied to Raman spectra, achiev-
ing error of prediction as low as 3.8%. Cletus et al.
[88]
described a
modified inverse SORS sensor capable of real-time detection of
concealed substances under real-life environments and allowing
noncontact examination of the tested exhibits under background
light. Although no measurements were reported for drugs of abuse,
the authors discussed the potential application of the SORS for fo-
rensic and national security purposes, including real-time tests at
airports, border control terminals, mail distribution authorities and
customs checkpoints.
Advances in the Raman instrumentation and
data analysis: perspectives of the technique in
the detection of cocaine and other illegal
drugs of abuse in a forensic context and in low
content
Recent advances in Raman instrumentation for rapid analysis of
samples in real-timeopen new opportunities for on-field drug anal-
ysis. Izake
[7]
discussed the technological achievements in the opti-
cal methods based on the Raman effect for the stand-off and
noncontact analysis and identification of chemical and biological
hazards within the forensic and homeland contexts. Vandenabeele
et al.
[89]
described the ability of Raman spectroscopy to record data
nondestructively, without sample preparation, and the subsequent
transfer of samples to the analytical laboratory. As a first-pass
screening probe for forensic crime scenes, Raman spectroscopy
has proven to be a valuable tool for the early detection of danger-
ous and prohibited materials such as drugs of abuse and their
chemical precursors. There is still a need for new developments in
the area of miniaturized instrumentation, which extends the con-
cept and breadth of the analytical requirement to facilitate data col-
lection ‘on field’.
Portable Raman instruments have been proposed for use in air-
ports and at crime scenes for the rapid analysis of illicit drugs.
Valussi and Underhill
[90]
presented a proof-of-concept handheld
Raman spectrometer for the rapid analysis of illicit drugs from
seized samples of amphetamine, cocaine, heroin and a set of
adulterants/contaminants and solvents. Solid samples were ana-
lyzed in plastic bags (‘polybags’and common white plastic bags),
while liquid samples were analyzed inside colorless and brown
glass and in plastic bottles. The authors noted that virtually all
seized samples exhibited some fluorescence, and identification
was difficult where the fluorescence background was strong (e.g.
brown heroin). Hargreaves et al.
[91]
proposed using portable Raman
instruments with 785 nm excitation (an ‘industrial standard’wave-
length) to analyze suspect powders, including cocaine HCl, MET
and amphetamine sulfate with unknown constituents. The mea-
surements and analysis could be carried out rapidly (in less than a
minute) and with a high degree of discrimination in a real airport
environment by comparing a measured spectrum with library stan-
dards. Weyermann et al.
[92]
tested a transportable Raman spectrom-
eter for detecting drugs seized during border control. The analysis
was optimized using reference standards of heroin, cocaine and
amphetamine for real conditions and parameters such as interfer-
ence from room/ambient light, focalization distance, integration
time, repeatability and limit of detection.
The possibility of using different excitation wavelengths to ob-
tain Raman spectra in dispersive Raman spectrometers has been
evaluated by several authors. A new dispersive Raman spectrome-
ter with excitation at 1064 nm was proposed as an alternative to
785/830nm spectrometers in the analysis of highly fluorescent
samples. Hargreaves et al.
[93]
compared the effect of 785 and
1064 nm excitation wavelengths on the Raman spectra of cocaine
HCl, cocaine freebase, MET, amphetamine, heroin and cannabis.
With the use of 1064 nm excitation, Raman bands from heroin
and cannabis could be clearly identified. This identification was
not possible with the 785nm wavelength because of strong fluo-
rescence, demonstrating the major ad vantage of the 1064 nm exci-
tation for the analysis of these types of drugs despite the longer
integration time required to obtain the spectra. Yang et al.
[94]
used
a newly developed three-wavelength (532, 785 and 1064 nm) exci-
tation micro-Raman to observe a variety of samples. Because
fluorescence significantly reduces the quality of Raman spectra in
real-world measurements, the 1064 nm excitation was preferable
for the detection of explosives and street drugs. Vitek et al.
[95]
also
stated the b enefits of using 1064 nm excitation compared with
the standard 785 nm in selected samples of geological, geobio-
logical and forensic areas, expanding the types of samples that
can be measured by a miniaturized Raman spectrometer without
interference from fluorescence background emissions. In a recent
study, Ali et al.
[96]
used a novel 1064 nm dispersive Raman system
with a detector based on a newly developed transfer electron
InGaAs photocathode and electron bombardment CCD technology
to detect drugs and explosives, both neat and in plastic and glass
containers. Raman spectra of street samples of cocaine hydrochlo-
ride, MET, amphetamine and heroinmeasuredat1064nmexcitation
36
C. A. F. de Oliveira Penido et al.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2016,47,28–38

showed lower background fluorescence when compared with the
785 nm excitation using the same integration time.
The development of Raman instrumentation is accompanied by
significant progress in data processing and analysis. In this context,
Ng et al.
[97]
tested spectral library searching algorithms for identify-
ing illegal substances deposited in fingerprints based on the infra-
red absorption features that can easily be applied to Raman
features. Algorithms included conventional Euclidean distance
searching, spectral angle mapping (also called cosine correlation
analysis) and correlation algorithms, which give better results when
used with second derivatives and reference spectra. Noonan
et al.
[98]
implemented several preprocessing techniques such as
truncating, Savitzky–Golay smoothing, normalization, differentiat-
ing and mean centering to implement a discriminant algorithm
using PCA in the spectra of drugs (isoxsuprine and norephedrine)
mixed with adulterants (benzocaine and lidocaine) in the range
from 10% to 100%. Authors found that it was possible to resolve
the spectral differences between these samples and correctly clas-
sify them 100% of the time.
At last, but not less important, is the development of calibration
methods based on multivariate statistics such as PCR and PLS
aiming quantitative analysis of mixtures, and multivariate methods
with potential for classification of illegal drug samples.
[99]
PCR
(based on PCA factors) and PLS are multivariate methods suitable
for quantitative analysis where all the useful spectral information
is used in the model, by correlating the concentrations of the ana-
lyte of interest with the spectral features in a regression
model.
[99–101]
The PLS differs from PCA in the way the factors are
calculated. PLS considers the information about the concentration
of the analytes in the decomposition of the scores and vectors,
resulting in higher weights for the factors related to higher analyte
concentrations.
[102]
This may result in models with better correla-
tions than the PCR, mainly when the analyte is in the limit of detec-
tion. These multivariate techniques have been used in the
quantitative analysis of several drugs of abuse mixed with
adulterants/contaminants using Raman spectra.
[7,31,48–51,58,61,87]
In
a recent study, Li et al.
[103]
developed a robust and accurate analyt-
ical methodology to quantify low-content (<0.1%) solid mixtures
based on PLS, being applicable to systems that cannot be easily an-
alyzed chromatographically, such as hydrate, polymorph or solvate
contamination. The method involves five steps to produce an over-
all accurate sample concentration, being the detection limit compa-
rable with liquid chromatography.
PCA and PLS are also used for discrimination and classification
of samples in groups.
[99]
Because the scores of both techniques
are calculated in a way they represent the intensities of the
spectral variations among samples, the distance between sam-
ples to a group can be calculated by a suitable discrimination
technique.
[102]
Because the nature of the PLS algorithm is related
to the canonical correlation analysis (CCA), the correlation be-
tween two sets of variables and the CCA is related to a linear dis-
criminant analysis.
[104]
It is expected that a discrimination by PLS
may achieve more reliable results than PCA because of the intra-
group variability, instead of the total variability of PCA; therefore,
discrimination by PLS (PLS/DA) can be successfully used when the
variability within groups is greater than the variability among
groups,
[104]
which may be the case of multicomponent mixtures
of drugs and adulterants. PLS/DA presented better results than
PLS in terms of discrimination of Raman spectra of skin
cancer,
[105]
which may be of interest in the forensic discrimination
of drugs. PCA has been employed to discriminate drugs in mix-
tures with high selectivity.
[32,48,98]
Final remarks
The Raman technique is a nondestructive method that can perform
rapid analysis on drugs of abuse to preserve the evidence recov-
ered at crime scenes and maintain the chain of custody. The Raman
technique is able to identify and quantify all sorts of illicit drugs, in-
cluding cocaine, morphine and amphetamines, even in the pres-
ence of contaminants and adulterants. The measurements can be
carried out from the outside of containers by using modern porta-
ble spectrometers that do not need sample preparation or reagents
in most applications. The technique allows for the identification of
different presentations of cocaine and the quantification of adulter-
ants in binary mixtures. It is important to emphasize that Raman
spectroscopy does not generate any chemical waste, thus
preventing contamination of the environment and eliminating haz-
ardous conditions for the investigator. Raman spectroscopy is an
economical method, dispensing resources used in managing
chemical liabilities. It is a fast, accurate and reliable methodology
with great potential to further improve the combat against drug
contraband in airports, post offices and other public areas.
Acknowledgements
This project was supported in part by Award No. 2014-DN-BX-K016
awarded by the National Institute of Justice, Office of Justice Pro-
grams, U.S. Department of Justice (I. K. Lednev), and in part by Grant
No. 2009/01788-5 from São Paulo Research Foundation –FAPESP
(L. Silveira Jr.). The opinions, findings, and conclusions or recommen-
dations expressed in this publication are those of the authors and do
not necessarily reflect those of the U.S. Department of Justice.
References
[1] United Nations Office on Drugs and Crime, World Drug Report 2014,
United Nations, New York, 2014.
[2] I. Bayer, H. Ghodse, B. Narcotics 1999,51,1.
[3] L. A. King, I. Ujvary, S. D. Brandt, Drug Test. Anal. 2014,6, 879.
[4] R. Karoui, C. Blecker, Food Bioprocess. Technol. 2011,4, 364.
[5] E. B. Hanlon, R. Manoharan, T. W. Koo, K. E. Shafer, J. T. Motz,
M. Fitzmaurice, J. R. Kramer, I. Itzkan, R. R. Dasari, M. S. Feld, Phys.
Med. Biol. 2000,45,R1.
[6] J. Torrent, V. Barron, Diffuse Reflectance Spectroscopy (Methods of Soil
Analysis. Part 5. Mineralogical Methods), Soil Science Society of
America, Madison, 2008.
[7] E. L. Izake, Forensic Sci. Int. 2010,202,1.
[8] S. Ewen, G. Dent, Modern Raman Spectroscopy: A Practical Approach,
John Wiley & Sons, New York, 2005.
[9] R. L. McCreery, Raman Spectroscopy for Chemical Analysis,Wiley-
Interscience, New York, 2000.
[10] J. C. Carter, W. E. Brewer, S. M. Angel, Appl. Spectrosc. 2000,54, 1876.
[11] A. Smekal, Naturwissenschaften 1923,11, 873.
[12] C. V. Raman, K. S. Krishnan, Nature 1928,121, 501.
[13] G. S. Landsberg, L. I. Mandelstam, Naturwissenschaften 1928,16, 557.
[14] N. Colthup, L. Daly, S. Wiberley, Introduction to Infrared and Raman
Spectroscopy, Academic Press, New York, 2012.
[15] P. Larkin, Infrared and Raman Spectroscopy: Principles and Spectral
Interpretation, Waltham, Elsevier, 2011.
[16] I. R. Lewis, H. G. M. Edwards, Handbook of Raman Spectroscopy: From the
Research Laboratory to the Process Line, Marcel Dekker, New York, 2001.
[17] J. M. Hollas, Modern Spectroscopy,JohnWiley&Sons,WestSussex,2004.
[18] F. Smith, Handbook of Forensic Drug Analysis, Elsevier Academic Press,
Burlington, 2004.
[19] A. Negrusz, G. Cooper, Clarke’s Analytical Forensic Toxicology,
Pharmaceutical Press, London, 2013.
[20] K. L. Wise, J. B. Cooper, C. L. Schoen, Dispersive Near-IR Raman
Spectrometer, US Patent no. US 6373567 B1. Year Published: 2002.
[21] M. Fleischmann, P. J. Hendra, A. J.McQuillanet, Chem. Phys. Lett. 1974,
26, 163.
37
Raman spectroscopy in forensic toxicology
J. Raman Spectrosc. 2016,47,28–38 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

[22] P. L. Stiles, J. A. Dieringer, N. C. Shah, R. P. Van Duyne, Annu. Rev. Anal.
Chem. 2008,1, 601.
[23] B.Sharma,R.R.Frontiera,A.I.Henry,E.Ringe,R.P.VanDuyne,Mat.
Today 2012,15, 16.
[24] L. Jensen, C. M. Aikens, G. C. Schatz, Chem. Soc. Rev. 2008,37, 1061.
[25] P. Matousek, I. P. Clark, E. R. Draper, M. D. Morris, A. E. Goodship,
N. Everall, M. Towrie, W. F. Finney, A. W. Parker, Appl. Spectrosc.
2005,59, 393.
[26] C. Eliasson, N. A. Macleod, P. Matousek, Anal. Chem. 2007,79, 8185.
[27] P. Matousek, Chem. Soc. Rev. 2007,36, 1292.
[28] C. M. Hodges, P. J. Hendra, H. A. Willis, T. Farley, J. Raman Spectrosc.
1989,20, 745.
[29] S. C. Pinzaru, I. Pavel, N. Leopold, W. Kiefer, J. Raman Spectrosc. 2004,
35, 338.
[30] N. A. Macleod, P. Matousek, Pharm. Res. 2008,25, 2205.
[31] M. J. West, M. J. Went, Drug Test. Anal. 2011,3, 532.
[32] A. G. Ryder, J. Forensic Sci. 2002,47, 275.
[33] M. N. Leger, A. G. Ryder, Appl. Spectrosc. 2006,60, 182.
[34] C. A. F. O. Penido, M. T. T. Pacheco, R. A. Zangaro, L. Silveira, J. Forensic
Sci. 2015,60, 171.
[35] A. P. Gamot, G. Vergoten, G. Fleury, Talanta 1985,32, 363.
[36] E. M. A. Ali, H. G. M. Edwards, M. D. Hargreaves, I. J. Scowen, Anal.
Bioanal. Chem. 2008,390, 1159.
[37] C. Eliasson, N. A. Macleod, P. Matousek, Anal. Chim. Acta 2008,607,50.
[38] K. Y. Noonan, M. Beshire, J. Darnell, K. A. Frederick, Appl. Spectrosc.
2005,59, 1493.
[39] T. Laussmann, I. Grzesiak, A. Krest, K. Stirnat, S. Meier-Giebing,
U. Ruschewitz, A. Klein, Drug Test. Anal. 2015,7, 56.
[40] J. S. Day, H. G. M. Edwards, S. A. Dobrowski, A. M. Voice, Spectrochim.
Acta A 2004a,60, 563.
[41] J. S. Day, H. G. M. Edwards, S. A. Dobrowski, A. M. Voice, Spectrochim.
Acta A 2004b,60, 1725.
[42] M. J. West, M. J. Went, Spectrochim. Acta A 2009,71,1984.
[43] V. Molina Moreno, M. Lopez-Lopez, J. C. Atoche, C. Garcia-Ruiz, Sci.
Justice 2014,54,164.
[44] E. M. A. Ali, H. G. M. Edwards, M. D. Hargreaves, I. J. Scowen, Anal. Chim.
Acta 2008,615, 63.
[45] M. J. West, M. J. Went, Forensic Sci. Int. 2009,189, 100.
[46] E. M. A. Ali, H. G. M. Edwards, M. D. Hargreaves, I. J. Scowen, J. Raman
Spectrosc. 2010,41, 938.
[47] E. M. A. Ali, H. G. M. Edwards, I. J. Scowen, Spectrochim. Acta A 2011,
80,SI2.
[48] E. M. A. Ali, H. G. M. Edwards, J. Raman Spectrosc. 2014,45, 253.
[49] A. G. Ryder, G. M. O’Connor, T. J. Glynn, J. Forensic Sci. 1999,44, 1013.
[50] A. G. Ryder, G. M. O’Connor,T.J.Glynn,J. Raman Spectrosc. 2000,31, 221.
[51] C. A. F. O. Penido, L. Silveira, M. T. T. Pacheco, Instrum. Sci. Technol.
2012,40, 441.
[52] J. T. Zhao, C. Y. Xu, Y. B. Duan, P. X. Zhang, M. Z. Si, Q. Yang, Spectrosc.
Spectral Anal. 2002,22, 588.
[53] V. Rana, M. V. Canamares, T. Kubic, M. Leona, J. R.Lombardi, J. Forensic
Sci. 2011,56, 200.
[54] M. Baranska, A. Kaczor, J. Raman Spectrosc. 2012,43, 102.
[55] B. Ivanova, M. Spiteller, J. Mol. Struct. 2013,1034, 173.
[56] S. B. Karch, Drug Abuse Handbook, CRC Press, Boca Raton, 2006.
[57] K. A. Frederick, R. Pertaub, N. W. S. Kam, Spectrosc. Lett. 2004,37, 301.
[58] E. Katainen, M. Elomaa, U. M. Laakkonen, E. Sippola, P. Niemela,
J. Suhonen, K. Jarvinen, J. Forensic Sci. 2007,52, 88.
[59] R. G. Weston, J. Forensic Sci. 2010,55, 1068.
[60] R. W. Berg, T. Norbygaard, P. C. White, S. Abdali, Appl. Spectrosc. Rev.
2011,46, 107.
[61] O. S. Fenton, L. A. Tonge, T. H. Moot, K. A. Frederick, Spectrosc. Lett.
2011,44, 229.
[62] F. Taplin, D. O’Donnell, T. Kubic, M. Leona, J. Lombardi, Appl. Spectrosc.
2013,67, 1150.
[63] J. S. Triplett, J. A. Hatfield, T. L. Kaeff, C. R. Ramsey, S. D. Robinson,
A. F. Standifer, J. Forensic Sci. 2013,58, 1607.
[64] S. P. Stewart, S. E. J. Bell, N. C. Fletcher, S. Bouazzaoui, Y. C. Ho,
S. J. Speers, K. Laota Peters, Anal. Chim. Acta 2012,711,1.
[65] R. Christie, E. Horan, J. Fox, C. O’Donnell, H. J. Byrne, S. McDermott,
J. Power, P. Kavanagh, Drug Test. Anal. 2014,6, 651.
[66] G. Montalvo, L. Lopez-Melero, F. Ortega-Ojeda, M. A. Pena,
C. Garcia-Ruiz, Anal. Method 2014,6, 9536.
[67] S. Assi, A. Guirguis, S. Halsey, S. Fergus, J. L. Stair, Anal. Method 2015,7,
736.
[68] A.G.Ozer,H.O.Tabakoglu,S.Cengiz,J. Biomed. Opt. 2014,19,
057006.
[69] S. Farquharson, Y. H. Lee, R. A. Lieberman, Proc. SPIE 2000,4200, 89.
[70] F. Inscore, C. Shende, A. Sengupta, H. Huang, S. Farquharson, Appl.
Spectrosc. 2011,65, 1004.
[71] V. D’Elia, G. M. García, C.García Ruiz, Appl. Spectrosc.Rev. 2015,50, 775.
[72] E. Horvath, J. Mink, J. Kristof, Mikrochim. Acta 1997,14, 745.
[73] Y. H. Lee, S. Farquharson, H. Kwon, P. Rainey, M. Fallahi,
R. J. Nordstrom, T. R. Todd, Proc. SPIE 1999,3537, 252.
[74] R. A. Sulk, R. C. Corcoran, K. T. Carron, Appl. Spectrosc. 1999,53, 954.
[75] L. Yang, H. Liu, J. Wang, F. Zhou, Z. Tian, J. Liu, Chem. Commun. 2011,
47, 3583.
[76] B. Sagmuller, B. Schwarze, G. Brehm, G. Trachta, S. Schneider, J. Mol.
Struct. 2003,661, 279.
[77] K. Faulds, W. E. Smith, D. Graham, R. J. Lacey, Analyst 2002,127, 282.
[78] K. R. Ackermann, T. Henkel, J. Popp, ChemPhysChem 2007,8, 2665.
[79] J. Chen, J. Jiang, X. Gao, G. Liu, G. Shen, R. Yu, Chem.-Eur. J. 2008,14,
8374.
[80] M. Sanles-Sobrido, L. Rodriguez-Lorenzo, S. Lorenzo-Abalde,
A. Gonzalez-Fernandez, M. A. Correa-Duarte, R. A. Alvarez-Puebla,
L. M. Liz-Marzan, Nanoscale 2009,1, 153.
[81] S. Farquharson, C. Shende, A. Sengupta, H. Huang, F. Inscore,
Pharmaceutics 2011,3, 425.
[82] J. D. Fox, K. N. Waverka, G. F. Verbeck, Forensic Sci. Int. 2012,216, 141.
[83] W. W. Yu, I. M. White, Analyst 2013,138, 1020.
[84] E. Massarini, P. Wasterby, L. Landstrom, C. Lejon, O. Beck,
P. O. Andersson, Sensor Actuat. B - Chem. 2015,207, 437.
[85] C.Eliasson,N.A.Macleod,P.Matousek,Central Laser Facility Annual
Report 2007/2008, Science and Technology Facilities Council,
Rutherford Apleton Laboratory, Central Laser Facility, 2008; 254.
http://www.clf.stfc.ac.uk/resources/PDF/ar07-08_s7_non-invasive_
detection.pdf.
[86] W. J. Olds, E. Jaatinen, P. Fredericks, B. Cletus, H. Panayiotou, E. L. Izake,
Forensic Sci. Int. 2011,212, 69.
[87] W. J. Olds, S. Sundarajoo, M. Selby, B. Cletus, P. M. Fredericks,
E. L. Izake, Appl. Spectrosc. 2012,66, 530.
[88] B. Cletus, W. Olds, P. M. Fredericks, E. Jaatinen,E. L. Izake, J. ForensicSci.
2013,58, 1008.
[89] P. Vandenabeele, H. G. M. Edwards, J. Jehlicka, Chem. Soc. Rev. 2014,
43, 2628.
[90] S. Valussi, M. Underhill, Proc. SPIE 2006,6402, 64020I.
[91] M. D. Hargreaves, K. Page, T. Munshi, R. Tomsett, G. Lynch,
H. G. M. Edwards, J. Raman Spectrosc. 2008,39, 873.
[92] C. Weyermann, Y. Mimoune, F. Anglada, G. Massonnet, P. Esseiva,
P. Buzzini, Forensic Sci. Int. 2011,209, 21.
[93] M. D. Hargreaves, A. D. Burnett, T. Munshi, J. E. Cunningham,
E. H. Linfield, A. G. Davies, H. G. M. Edwards, J. Raman Spectrosc.
2009,40, 1974.
[94] W. Yang, H. Wu, J. Qian, L. Chandler, C. Lieber, C. Dentinger, Proc. SPIE
2012,8546, 854603.
[95] P. Vitek, E. M. A. Ali, H. G. M. Edwards, J. Jehlicka, R. Cox, K. Page,
Spectrochim. Acta A 2012,86, 320.
[96] E. M. A. Ali, H. G. M. Edwards, R. Cox, J. Raman Spectrosc. 2015,46, 322.
[97] P. H. R. Ng, S. Walker, M. Tahtouh, B.Reedy, Anal. Bioanal. Chem. 2009,
394, 2039.
[98] K. Y. Noonan, L. A. Tonge, O. S. Fenton, D. B. Damiano, K. A. Frederick,
Appl. Spectrosc. 2009,63, 742.
[99] Y. Roggo, P. Chalus, L. Maurer, C. Lema-Martinez, A. Edmond, N. Jent,
J. Pharm. Biomed. Anal. 2007,44, 683.
[100] E. V. Thomas, Anal. Chem. 1994,66, 795A.
[101] M. L. O’Connell, A. G. Ryder, M. N. Leger, T. Howley, Appl. Spectrosc.
2010,64, 1109.
[102] P. Geladi, B. R. Kowalski, Anal. Chim. Acta 1986,185,1.
[103] B. Li, A. Calvet, Y. Casamayou-Boucau, C. Morris, A. G. Ryder, Anal.
Chem. 2015,87, 3419.
[104] M. Barker, W. Rayens, J. Chemometrics 2003,17, 166.
[105] F. L. Silveira, M. T. T. Pacheco, B. Bodanese, C. A. Pasqualucci,
R. A. Zângaro, L. Silveira, Lasers Surg. Med. 2015,47,6.
Supporting Information
Additional supporting information may be found in the online
version of this article at the publisher’swebsite.
38
C. A. F. de Oliveira Penido et al.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2016,47,28–38