GC–MS as a valuable tool for analysing cannabinoid-containing gummies and identifying the synthetic process used for their production
Forensic Chemistry, Volume 44, 2025, 100672: Graphical abstract
The goal of this study is to develop, optimize, and validate a GC–MS analytical protocol for identifying and quantifying both natural and semi-synthetic cannabinoids in complex edible matrices, specifically seized gummy sweets. Using cannabinoid-free commercial gummies spiked with known substances, the method was validated according to current guidelines.
This approach was then applied to six different types of seized recreational gummies to determine the cannabinoid composition and infer their origin—natural or synthetic—as well as possible synthetic routes. Beyond common cannabinoids, the study also identified less known compounds such as HHC-C9 and oleamide, providing deeper insight into the composition and potential pharmacological effects of these products.
The original article
GC–MS as a valuable tool for analysing cannabinoid-containing gummies and identifying the synthetic process used for their production
Arianna Bini, Elisa Roda, Stefano Protti, Luca Morini, Antonella Profumo, Marco Cavallo, Adolfo Gregori, Carlo Alessandro Locatelli, Daniele Merli
Forensic Chemistry, Volume 44, 2025, 100672
https://doi.org/10.1016/j.forc.2025.100672
licensed under CC-BY 4.0
Selected sections from the article follow. Formats and hyperlinks were adapted from the original.
Cannabis has been used for centuries for economical, medical, textile and recreational purposes, according to the ability of the single strain to produce specific bioactive compounds. The biological effects of Cannabis are due to the presence of a particular class of terpenolic compounds dubbed as cannabinoids. Such derivatives interact with different receptors, namely cannabinoid receptors CB1 and CB2 [1,2] as well as non-cannabinoids receptors such as 5-HT1A and α-2 receptors [3,4]. More than 100 cannabinoids have been discovered up to date, and the most famous are non-psychotropic cannabidiol (CBD) [5] and psychotropic Δ9-tetrahydrocannabinol (Δ9-THC) [6], thoroughly investigated for their effects on human health.
As the biologically active form of cannabinoids derives from the heat-induced decarboxylation of the acidic form present in the vegetable material, Cannabis has historically been consumed through smoking or vaporization [7,8]; however, due to the commercial availability of the decarboxylated cannabinoids, other preparations such as extracts and edibles (both solid and liquid) exist on the market and are lately gaining popularity [9,10]. It should be noticed that pharmacologically the term “cannabinoids containing edibles” refers to “cannabis-infused products meant for oral consumption, including oils, tinctures and oil-based capsules”, and thus includes also foods (e.g. chocolates, gummies, and hard candies) containing natural or semi-synthetic cannabinoids [11]. Edible – whether for medical or recreational uses – are often preferred to smoking for social and therapeutical reasons, since their use is more discreet while the psychoactive effects last longer [12].
Therefore, the market for cannabinoid-containing foods and supplements has grown steadily over the last few years [13], with CBD and Δ9-THC being the most widely available compounds; however, around 2021, hexahydrocannabinol (HHC) and Δ8-THC (that are currently neglected by several regulatory authorities) have been openly sold in a range of products in the US, and subsequently in Europe and elsewhere [14]. HHC is a semi-synthetic Δ8/Δ9-THC derivative, usually prepared via acid catalyzed cyclization of CBD – which produces Δ8-THC as the main product – followed by catalytic hydrogenation [15]. Although the human pharmacology and pharmacokinetics of HHC and Δ8-THC have not been thoroughly examined, anecdotal evidence suggests that their effects and active dose closely resemble those of the best known Δ9-THC [14,16,17].
Based on the marketing information, some of the preferred forms of consumption include gummies. The actual composition and cannabinoid content of these products is highly variable and, due to lack of quality control, marketed products may be contaminated either with extraction residues or synthetic byproducts [15].
Currently, GC–MS and LC-MS are the gold standard methods for the identification and quantification of natural and semi-synthetic cannabinoids [18], especially when dealing with complex matrices. Different methods are presented in literature that can be adapted to the analysis of gummies. For example, ultrasound-assisted extraction (UAE) followed by solid-phase extraction (SPE) and HPLC/MS was proposed [19]; however, the sample preparation is tedious and HPLC can't efficiently separate several cannabinoids isomers (such as the THCs and isoTHCs), this last being a problem common to all the methods proposed that rely on this technique for the analysis step [20]. Therefore, the need for simple, alternative methods is highly desired [21]. GC–MS methods were proposed but not validated for this specific matrix [22] and no study reported in literature has attempted to establish the origin of the cannabinoid with which the edible was infused, or investigated the presence of a wide variety of olivetol-based cannabinoids as we do in the present paper. The last considerations are also valid in the case of e-cig liquid, for which analysis is much simpler as it requires only a dilution with the proper solvent [23]; in this way, HHC (1, Fig. 1) was found and quantified by adopting this strategy [24] but derivative HHC-C9 (2) was never found.
In view of the reports on intoxication cases due to the overconsumption of HHC and Δ8-THC, we describe herein a GC–MS protocol for the analysis of the different cannabinoids present in cannabinoid-containing gummy sweets after liquid extraction of the active ingredients.
In addition, we describe for the first time the identification of a new psychoactive substance (NPS), hexahydrocannabinonol (HHC-C9), in a gummy and an e-cig liquid, that, to the best of our knowledge, has not been previously described in the literature and has never been identified in recreational products. Oleamide was also identified for the first time in similar items.
The identification of the major and minor cannabinoids present, based on a lab-made GC–MS library consisting of more than 30 compounds, would provide a strategy to evaluate the potential of abuse of the analyzed items and the processes leading to their production.
2. Materials and methods
2.3. GC–MS conditions
GC–MS analyses have been performed with an Agilent Technologies 7890A single quadrupole GC/MS system (Agilent 5975C mass spectrometer – Agilent Technologies, Santa Clara, California, USA).
Chromatographic separation was performed on a HP-5MS capillary column (30 m length x 0.25 mm ID x 0.25 μm film thickness, Restek, Milan, Italy) with helium (>99.99 %) as carrier gas at a constant flowrate of 1.0 mL/min. An injection volume of 1 μL was employed. The injector temperature was set at 250 °C and operated in splitless mode. The oven temperature was programmed from 60 °C (isothermal for 4 min) to 300 °C (isothermal for 5 min) at the rate of 10 °C·min−1. Data acquisition started 5 min after injection. Mass transfer line temperature was set at 300 °C. All mass spectra were acquired with an electron ionization system (EI, Electron Impact mode) with ionization energy of 70 eV and source temperature of 250 °C, with spectral acquisition in Full Scan mode, positive polarity, over a mass range of 50–600 Da with a scan rate of 1460 amu·s−1. Standards were prepared in ethanol to obtain nine different concentrations (from 5.0 mg·mL−1 to 120 mg·mL−1), and olivetol (50.0 mg·mL−1) was added as an analogous internal standard. Non-deuterated IS was used, accordingly to other work dealing with the analysis of cannabinoids [[29], [30], [31]]; no loss of accuracy and repeatability is expected using non-deuterated IS, as several reports used an analogous internal standard with excellent results; furthermore, there are reports describing the disadvantages of deuterated internal standards [[32], [33], [34]]. TIC was used for the construction of calibration curves [32,35].
The data were analyzed by MSD 5975 VL data analysis software (Agilent Technology).
The open-source R-based software CAT (Chemometric Agile Tool) was used for PCA data treatment [36].
6. Analysis of real samples
6.1. Cannabinoids-containing gummies from NEWS project
Applying the extraction method III to 250 mg portions of cannabinoids-containing gummy sweets (3 portions, each extracted and analyzed in triplicate), a mixture of psychotropic cannabinoids was found, as shown in Table 2. The characteristics of the different gummies are reported in SI, Fig. S16 alongside the corresponding photos.
The cannabinoids described in SI paragraph 1.2 but not reported in Table 1 were below the LOD in all real samples. An example of chromatogram for each type of gummy is reported in SI, paragraph 2.2.
To better visualize the possible similarities among the samples, the dataset regarding the analysis of the real samples was submitted to PCA (as previously described in paragraph 4.8.). The first and second components contributed to 60.2 % and 27.3 %, respectively, of the explained variance; therefore, we based our considerations on these first two components only. Fig. 2 helpfully displays the relationships between the real samples (objects) and the compounds found during the analysis (variables), allowing for the identification of different populations present in the samples.
Forensic Chemistry, Volume 44, 2025, 100672: Fig. 2. Biplot of the PCA model on the first two principal components, with red arrows representing the vectors of the loading values of the variables (compounds reported in Table 2 minus Δ9-THCV, CBD and CBE). The colored dots are the scores of the gummies: the blue dots are the three replicates relative to the analysis of sample C1, red dots are the three replicates relative to the analysis of sample C2, and so on. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
As can be observed, a clear distinction occurs among the samples, and four classes can be identified: one comprising C1, one comprising gummies from C2 to C4, one for C5 and the last for C6. Thus, as described in paragraph 4.8, the data set limited to samples from C2 to C4 was independently submitted to PCA to define the statistical boundaries of this population; the other samples were then projected as an external set and properly recognized as independent of C2-C4 set.
This step has been performed according to a class modeling approach. [49] T2 outliers are samples presenting extreme values for the population taken into consideration (in parallel with univariate statistics), while Q outliers are samples in which the correlation between variables doesn't align with what can be observed in the reference group. In our case, taking the gummies from C2 to C4 as the population, samples C1 and C6 are outliers at 99.9 % according to both statistics, while C5 is an outlier at 99.9 % in T2 and 95 % in Q. In conclusion, C5 presents extreme values (identifiable as a different composition of the active principle) with a more similar correlation between compounds. The graphs supporting these conclusions have been reported in SI (paragraph 3, figure from S23 to S26) for the sake of completeness.
Following this statistical analysis and the identification of the different populations, we grouped the C2 to C4 gummies population together.
7. Conclusions
The quantification of the main compounds found in recreational gummy sweets, obtained from PCC within the NEWS Project, was achieved using a simple, low-cost extraction method that minimizes the sample preparation procedure, yet efficiently defines the sample's composition, which in turn can be used to argue the synthetic method used to obtain the psychotropic cannabinoids. Given the growing interest in improving the routinely available GC–MS and LC-MS techniques for the identification and quantification of newly synthesized cannabinoids, the development of this straightforward yet effective method for the extraction of cannabinoids from such a complex matrix is valuable in simplifying food analyses, making it faster and more accurate.
Two types of gummies (C1 and C2–4) have been found to contain a number of cannabinoids that suggests that a synthetic procedure was used to obtain the psychoactive mixture: in both cases, the presence of synthetic by-products (such as CBL, CBE, Δ7-THC, Δ8-isoTHC and Δ4(8)-isoTHC) indicates that natural CBD underwent acid-catalyzed cyclization of to form THC isomers. In the case of gummy C1, this procedure is followed by an incomplete catalytic hydrogenation to give HHC. Both the third and fourth types of gummies, namely C2–4 and C5, contain comparable amounts of oleamide, an endocannabinoid and CB1 modulator, the presence of which enhances the effect of cannabinoids on the human body; nonetheless, due to their significantly different cannabinoid content, these gummies cannot be grouped together for comparison. The C5 gummy was found to contain only Δ9-THC, with all other cannabinoids, including synthetic by-products, to be below the LOD. This suggests that natural Δ9-THC may have been used in the preparation. Similarly, most natural and semi-synthetic cannabinoids (including synthetic by-products) were found to be below the LOD in the C6 gummy, while the active principle identified as HHC-C9, an unprecedented synthetic HHC analogue.
All the analyzed items have the potential for abuse, considering that an active dose of 5 mg–10 mg of THC or THC-like compounds are considered psychotropic when assumed orally.
It should be noticed that the use of certified standards and procedures is rarely feasible in the analysis of NPS, as the innovation in this area precludes the possibility of obtaining certified reference materials and standards. For this reason, an approach based on the synthesis of the suspected compounds or extracting them from sized items and the availability of an in-house MS library, seems to be the most feasibly approach, being a common and sometime obligatory practice in forensic science.
