Since their appearance in the illicit drug market, the number of new psychoactive substances (NPS) is growing at an alarming fast rate; as a consequence, the task of analysis and identification of NPS is posing a big challenge for forensic scientists on one side, and regulatory bodies, for the design and delivery of effective evidence-based responses to drug problems (Zuba, 2014; Lee et al., 2019).

In its latest report, the European Monitoring Center for Drugs and Drug Addiction (EMCDDA) revealed a market that is both resilient and reflective of developments taking place at the global level; the value of the NPS market is unknown actually, however 55 new substances were reported to the European Union Early Warning System (EWS) in 2018, bringing the total number of NPS monitored to 731 (EMCDDA, 2019). Undoubtedly, the shaping and implementation of policy responses and law enforcement activity have contributed to slow-down in appearance of NPS, with respect to the previous decade. However, NPS continue to pose serious cross-border threats to health, with potent synthetic opioids (mainly fentanyls), synthetic cannabinoids and designer benzodiazepines appearing on the market, associated with reports of health emergencies and deaths. Moreover, drug overdoses are commonly associated with the intake (deliberate or not) of multiple substances and thus, health threats and diagnosis may be overlooked without the disposal of adequate forensic and toxicological data. As a consequence, introducing comprehensive screening and increasing the reliability of testing is a central focus for many countries, who have made significant investments in this area. Unfortunately, drug designers are working incessantly to synthesize non-controlled analogs of the drugs of abuse, aiming to get around the existing anti-drug laws, by introducing slight modifications to the chemical structures (UNODC, 2018; Kraenenburg et al., 2019). The constant introduction of new drugs in turn creates a need for reference material to confirm structural elucidation of uncommon or newly encountered substances (Brandt et al., 2014).

In this context, researchers have put considerable efforts in developing advanced chromatographic instrumentation and techniques, aiming to achieve reliable identification of organic compounds in complex mixtures. Approaches based on high-resolution gas chromatography coupled to mass spectrometry (HRGC-MS) are the workhorse analytical tool employed in forensics laboratories, affording the selectivity and sensitivity required for most analytes commonly encountered in seized drugs; however, these hyphenated techniques presents inherent weak points which relate to both the GC and MS counterparts (International, 2015, 2017; Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG), 2019). The data afforded by GC-MS are in fact affected by measurement uncertainties, of different magnitude and sources, of which analysts must be aware: first the uncertainty of measurement for GC retention time, expressed as absolute or relative time units (compared to a known reference standard). Second, the uncertainty of measurement of relative abundances of MS ions obtained by electron impact (EI) ionization; in both cases, specific acceptance criteria recommended by different governing bodies are not uniform (Davidson et al., 2018; Kelly and Bell, 2018).

Whereas, generally accepted as the gold standard of forensic drug analysis, yet GC-MS in some cases suffers from clear limitations, for the identification of co-eluting regioisomeric forms of synthetic drugs of identical elemental composition and yielding identical fragmentation patterns. In some circumstances, positional isomers, and diastereomers may be separated chromatographically, but identification cannot be attained, univocally, on the sole basis of the retention behavior. Whenever structural assignment is mandatory, a further analytical step may be required, consisting of compound isolation or targeted organic synthesis (followed by purification/concentration, eventually) prior to further characterization (Abiedalla et al., 2019; Kraenenburg et al., 2019).

To this concern, the combination of high-resolution gas chromatography (HRGC) to Fourier Transform Infrared Spectroscopy (FTIR) provides a unique tool, through the combination of high efficient separation, and highly specific identification. GC-FTIR allows to quick identifying functional groups in unknown substances, based on the retention behavior of the analytes and the IR absorption bands. Relying on distinct chemical properties, IR may well complement the information afforded by mass spectrometry (MS), in achieving structural identification of volatile and semi-volatile molecules. Moreover, by measuring small energy differences based on rotational and vibrational amplitudes between individual molecular bonds, FTIR spectroscopy enables to overcome one limitation of MS detection, in discriminating regioisomeric compounds (Kempfert, 1988; Almalki et al., 2019).

First attempts to interface a gas chromatograph to IR spectroscopy date back to four decades ago (Griffiths et al., 2008); however the real milestone came in the late 1960s, with the replacement of conventional gratings or dispersive elements with interferometers and FT mathematics (Low and Freeman, 1967). Though exploiting the clear advantage of speed of analysis, yet those hyphenated instruments used high-capacity GC columns and were operated in the stopped-flow mode (Low, 1971; Shaps and Varano, 1977). A flow-through gas cell, known as the light pipe (LP), was used as an interface to deliver vapor-phase IR spectra of eluting solutes (Visser, 2002), with the addition of a flow of make-up gas to compensate for the greater internal diameter of the LP device over that of the GC column (typically, 1.5 vs. 0.32 mm i.d.). In such a way, the chromatographic resolution was kept, even at the cost of longer residence time of the analytes in the interface and, thus, a decrease in the sensitivity. The latter was further impaired by the higher temperature required in the interface for the less volatile GC components, creating background noise in the spectrum (Brown et al., 1985). Indeed, the coupling of GC and FTIR has always posed the need to compromise between sensitivity, and speed. The narrow bandwidths of GC peaks often did not allow for adequate detector sampling, to record a useful spectrum; on the other hand, the amount of analyte required in most cases overwhelmed the GC column capacity. Later on, sample trapping techniques have been developed, aiming to achieve lower detection limits than those afforded by LPs: the matrix isolation (MI) interface (Reedy et al., 1985), and the direct deposition interface (DD) (Fuoco et al., 1986). Both consisted of cryogenic devices allowing for mobile-phase elimination by trapping the GC-separated compounds eluting from the sub-ambient temperature to 11 K (MI) or 77 K (DD). All cold trapping interfaces rely on the use of high vacuum (needed to prevent interferences from environmental water and carbon dioxide), and leak-tight interface housing. In MI, also a cage of 1–2% of argon is added to the carrier gas for freezing the analytes. A critical comparison between light-pipe and sample trapping interfacing methods has been made by Schneider and Demirgian (1986), also discussing pros and cons in terms of applicability to sample analysis. The light-pipe interface provided a relatively quick and inexpensive way to obtain library-searchable vapor-phase spectra, in real time and in a non-destructive way, from GC eluted components. Whereas, MI interfaces implied a two-step process, since collection of IR spectra from the trapped analytes occurred after the collection was completed. In DD interface, the immobilized spots pass through an external IR beam a few seconds after the deposition, and multiple scans can be taken, as long as the cryogenic conditions are maintained. Both deposition techniques afford at least two orders of magnitude more sensitivity than the LP devices, with the resolution increased from 8 to 4 cm⁻¹. As a result of the sharper absorption bands compared with those obtained from free rotating molecules in the gas-phase, also the specificity was increased, in terms of differentiation between closely related molecules. Such features pave the way for the use of solid deposition GC-FTIR as a viable alternative to GC-MS approaches in forensic laboratories, for the reliable identification of NPS such as synthetic cathinones, cannabinoids, phencyclidine analogs, etc.

Originally developed for achieving the desired selectivity toward the cannabinoid receptors CB1 and CB2, synthetic cannabinoids or cannabimimetics were used in the clinical practice to deliver high therapeutic activity (anti-inflammatory and analgesic properties) from unwanted side effects (psychotropic activity). Unfortunately, the information generated by the scientific community has been promptly misused by clandestine laboratories, and these compounds have quickly found their way into the recreational drug market. The detailed pharmacological activities of these analogs are not known, which makes easy access and use of these drugs very dangerous to human health; moreover, synthetic cannabinoids are typically full agonists on the CB1 receptor, thus leading to maximum activation, even at significantly lower doses. Besides their higher potency with respect to the conventional drugs, long half-lives, and formation of active metabolites represent additional hazards deriving from illegal use of cannabimimetics (Bretteville-Jensen et al., 2013; ElSohly et al., 2014). The latter encompasses a wide range of chemical structures, and new analogs are constantly introduced on the market after the preceding drug comes under regulation; this poses additional challenges to drug law enforcement and forensic scientists. Analytical techniques developed so far for the detection and/or quantification of synthetic cannabinoids include colorimetric, immunochemical, and chromatographic methods (Namera et al., 2015).

In this research, the feasibility of using GC-FTIR with solid deposition of the analyte is shown, to achieve unambiguous structure elucidation of 1-pentyl-3-(1-naphthoyl)indole, commonly known as JWH-018, a synthetic cannabinoid identified as component of illegal “incense blends.” Furthermore, the results obtained from the analysis of a seized sample containing several synthetic cannabinoids were compared to those afforded by GC-EI-MS, in terms of identification of unknown components. Detection and identification of unknown NPS in real samples is a major concern when legal or regulatory issues are involved; quantification of targeted analytes may be required, eventually. In a similar way as for MS-based identification, IR spectral data are searched into commercial or custom-made libraries, containing hundreds to thousands IR spectra of standard compounds. When reference materials are measured one at the time, in pure form or constant matrix, then LOD would be sufficient to describe the performance of the measuring system. In this study, another validation parameter was investigated, i.e., the limit of identification (LOI), defined as the lowest analyte concentration that yields a library searchable IR spectrum. In this concern, LOI is related to LOD in that a detectable signal is entailed, but this must also allow for a correct identification to be made, from a defined database. Most often overlooked in similar studies, LOI is a key parameter in determining the possibility for reliable identification of an unknown compound, contained at a certain amount in a given sample, and often in the presence of a noisy background. In a very straightforward way, LOI is sometimes estimated by simply doubling the LOD value (Lanzarotta et al., 2017).