Spectral trends in GC-EI-MS data obtained from the SWGDRUG mass spectral library and literature: A resource for the identification of unknown compounds

Abstract

Rapid identification of new or emerging psychoactive substances remains a critical challenge in forensic drug chemistry laboratories. Current analytical protocols are well-designed for confirmation of known substances yet struggle when new compounds are encountered. Many laboratories initially attempt to classify new compounds using gas chromatography-electron ionization-mass spectrometry (GC-EI-MS). Though there is a large body of research focused on the analysis of illicit substances with GC-EI-MS, there is little high-level discussion of mass spectral trends for different classes of drugs. This manuscript compiles literature information and performs simple exploratory analyses on evaluated GC-EI-MS data to investigate mass spectral trends for illicit substance classes. Additionally, this work offers other important aspects: brief discussions of how each class of drugs is used; illustrations of EI mass spectra with proposed structures of commonly observed ions; and summaries of mass spectral trends that can help an analyst classify new illicit compounds.

1. Introduction

In recent years, forensic laboratories have reported an influx of new illicit drugs [1-7]. Among the traditional substances reported, new/novel psychoactive substances (NPSs) or emerging synthetic drugs are becoming more prevalent because they (i) produce equivalent psychoactive responses by targeting similar receptors and reaction sites within the body, (ii) elude legal definitions of a controlled substance, and (iii) evade detection due to identification limitations [6-13]. These substances incorporate a variety of classes including cannabinoids, opioids, stimulants, and benzodiazepines [2,3]. These broad classes are comprised of various subcategories that share psychotropic effects, chemical structures, or legal definitions. For example, the opioids class consists of fentanyl-related compounds (FRC), opiates, utopioids, and the newly emerging nitazenes [10,14,15]. The emergence of these chemically diverse NPSs brings new analytical challenges for current instrumentation practices to identify and classify seized samples.

The Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG) identifies electron ionization-mass spectrometry (EI-MS) as a Category A technique due to its ability to provide structural information; it is commonly employed in conjunction with the Category B separation technique gas chromatography (GC) [16-21]. Briefly, in traditional EI, volatilized compounds are ionized via the bombardment of highly energetic electrons (70 eV) [22,23]. This process deposits excess energy to the compound, which may result in fragment and/or molecular ions that are observed in a mass spectrum as mass-to-charge (m/z) ratios. The abundance of these ions is related to the likelihood of occurrence and stability [22,23]. Laboratories benefit from a massive, evaluated mass spectral library of reproducible EI-MS fragmentation from which an unknown and known spectra can be compared to postulate a compound’s identification. With the introduction of NPSs, however, subtle alterations to the core chemical structure can result in nearly indistinguishable spectra and lead to misclassifications [8,10,24,25]. Alternatively, if emerging compounds are not present in the library, the mass spectrum may not provide a good match to any entry and thus, may lead chemists to attempt to identify based on a drug class’s mass spectral trends.

The goal of this work is to condense results of previous studies and to provide guidance into EI mass spectral trends across NPS classes. This paper also reviews literature of the most encountered classes [6] of NPSs and summarizes EI mass spectral trends for these substances through an exploratory analysis that identifies the most frequent ions within the SWGDRUG Mass Spectral Library [19]. The exploratory investigation performed on the merged database highlights mass spectral properties and identifies the most frequent ions within the classes. Finally, this work can serve as a reference guide for current observations and assist in narrowing potential classifications for unknown spectra.

2. Methods

2.1. Preprocessing to produce class database

To complete class-based analyses, class assignments were obtained from the Cayman Chemical product catalog and appended to the entries within the SWGDRUG Mass Spectral Library (version 3.10.) [19] using a custom R script described in the Supplemental Information. The resulting dataset, referred to as the merged database, yielded over 1,500 entries across seven broad classes and fifteen subclasses [6], the distribution of which is shown in Table 1. It is important to mention that this process is automated and thus, has occurrences where some compounds are members of multiple classes. For example, N-benzylpiperazine (BZP) and escaline are included in both stimulant and hallucinogen classes.

Table 1.

Summary of drug classes in the merged database discussed (n = 1,549) as well as the section number where they are discussed.

Broad Class	Subclass Breakdown
Barbiturates (n = 10) (3.1.)	N/A
Benzodiazepines (n = 56) (3.2.)	N/A
Anabolic Steroids (n = 50) (3.3.)	N/A
Cannabinoids (n = 448) (3.4.)	Phytocannabinoids (n = 24) (3.4.1.) Synthetic Cannabinoids (n = 424) (3.4.2.)
Opioids (n = 323) (3.5.)	Opiates (n = 36) (3.5.1.) Fentanyls (n = 237) (3.5.2.) Utopioids (n = 35) (3.5.3.) Nitazenes (n = 15) (3.5.4.)
Stimulants (n = 600) (3.6.)	Amphetamines (n = 174) (3.6.1.) Arylcyclohexylamines (n = 42) (3.6.2.) Cathinones (n = 229) (3.6.3.) Phenethylamines (n = 127) (3.6.4.) Piperazines (n = 28) (3.6.5.)
Hallucinogens (n = 62) (3.7.)	Tryptamines (n = 57) (3.7.1.) Lysergamides (n = 5) (3.7.2.)

To capture class similarities, mass spectral peaks were binned into one of five categories depending on their relative intensity/abundance (RA) – Base Peak (m/z values with a RA of 100%), High (m/z values with RAs of 50% to 99%), Medium (m/z values with RAs of 10% to 49%), Low (m/z values with RAs of 5% to 9%), and Ultra-Low (m/z values with RAs of less than 5%).

2.2. Data analysis

After verification of the merged database, a simple exploratory analysis was conducted to monitor the occurrence of m/z values within each RA category to illuminate potential trends and common fragments shared within each class. The Ultra-Low RA category is not discussed comprehensively in this manuscript as these ions are often not above critical thresholds/limits-of-detection in experimental data or may be noise. As a starting point, the top 10 reoccurring ions which are seen in at least two compounds are presented for each category. To provide further insight and understanding about the frequently occurring ions, the proposed structures of some fragment ions (found in relevant literature) have been included with a mass spectrum. The NIST MS Interpreter tool (version 3.4.4.) (available at chemdata.nist.gov) was utilized to elucidate structures when minimal literature was found. Additionally, neutral losses were investigated to highlight fragmentation commonalities that are not directly observable in a mass spectrum. Neutral losses were calculated by taking the difference between the molecular ion (M^+●) (or known molecular mass) and the observed fragment ions.

3. Results and discussion

The following sections present discussions and breakdowns of EI-MS trends for the frequently encountered drug classes [2,6,7] that have at least 10 compounds in the merged database, barring lysergamides, followed by an example of how this information can be used in practice. Each drug class section contains a description of their use, relevancy in casework, common core structure substitutions, a mass spectrum with select proposed structures, and a table showing top reoccurring ions and neutral losses from the merged database for Base Peak, High, and Medium RA categories. Also reported are the number of peaks across these bins, including the Low RA category, disaggregated by 40 Da m/z bins to demonstrate where peaks are commonly observed.

Each drug class section also contains a figure to summarize important information. Each figure depicts the core structure(s) of the class and, when appropriate, an example mass spectrum with potential annotations of major ions. Importantly, the proposed annotations are based, when possible, on EI-MS literature reports. When appropriate, literature reports leveraging electrospray ionization (ESI)-MS sources were used to guide mass spectral annotations. In cases where limited literature references were found to support the proposed annotated structures, the NIST MS Interpreter tool (version 3.4.4.) was employed to offer possible explanations. Three plots are also provided. The first plot provides a representation of the average distribution of peaks observed in the mass spectra for an NPS class – plotting the average number of peaks observed for Low, Medium, and High + Base Peak RA categories binned into 40 m/z increments across the mass scan range. The second plot is designed to show how informative the mass spectra for the class are by plotting the average number of Low, Medium, and High + Base Peaks RA ions. The y-axis of this plot is fixed in all figures to allow for class-to-class comparisons. The third plot provides information on the presence and abundance of molecular ions. As the purpose of this manuscript is to provide a resource for the classification of unknown drug compounds, the final section walks through how this information can be used with an example spectrum. Three additional examples of “unknown” spectra are provided in Supplemental Information.

3.1. Barbiturates (n = 10)

Barbiturates, which originally stemmed from barbituric acid, are classified as tranquilizers and are utilized for sedation. The core structure (Fig. 1) may contain substitutions at the R₁ and R₂ positions [24,26-28]. These compounds are not frequently observed in forensic casework – there are no barbiturates in the top 25 drugs in the NFLIS 2020 Annual report [6] or in the top reported tranquilizers.

Fig. 1.

Provided in this figure is (A.) the core skeletal structure of barbiturates (outlined in red) and mass spectrum of pentobarbital with select proposed structural annotations [29]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the barbiturates class (n = 10). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Looking at the reoccurring ions for the 10 barbiturates in the merged database (Table 2), the ions at m/z 156, m/z 143, and m/z 168 in the Base Peak RA category, as well as m/z 41, m/z 43, and m/z 55 in the Medium RA category, correspond to the loss of the alkyl chain in the R₂ position. For the High RA category, the observed m/z values of m/z 141 and m/z 167 are the result of the intact ring structure but with varying N-alkyl chains at the R₁ and R₂ positions [29,30]. Other Medium RA peaks such as m/z 98 (Fig. 1A) may correspond to a ring-closing after the fragmentation of main ring structure. For the entire spectrum, it is unlikely to observe significant peaks above m/z 200, and the low mass range (m/z 40 to m/z 120) is often dominated by Low and Medium RA peaks.

Table 2.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the barbiturates class (n = 10). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
156	4 (40)	141	4 (40)	41	5 (50)	171	6 (60)
143	2 (20)	167	2 (20)	43	5 (50)	71	5 (50)
168	2 (20)			55	5 (50)	141	5 (50)
				98	4 (40)	114	4 (40)
				112	3 (30)	143	4 (40)
				157	3 (30)	183	4 (40)
				42	2 (20)	28	3 (30)
				51	2 (20)	43	3 (30)
				53	2 (20)	70	3 (30)
				56	2 (20)	85	3 (30)

From the merged database and literature reports of this limited class, potential criteria for identifying common barbiturate-related compounds include (i) lack of a molecular ion above 5% RA, and (ii) presence of a peak corresponding to the fragment of the R₁ and R₂ groups which yields ions at m/z 41 and m/z 43. Neutral losses of 171 Da, 71 Da, and 141 Da were commonly observed. Because of the small number of barbiturates in the merged database, no further conclusions were drawn.

3.2. Benzodiazepines (n = 56)

In recent years, benzodiazepines have seen increased prevalence in casework – driven by both prescription and illicit designer compounds. The NFLIS 2020 Annual Report [6] contained four benzodiazepines in the top 25 drugs and the list of top tranquilizers is dominated by these compounds [24,31,32]. More timely resources, like NPS Discovery [7], have recently reported new benzodiazepines, such as fluclotizolam and deschloroetizolam. The general backbone of these compounds consists of a benzene ring fused to a diazepine ring along with various side chains of substituents (Fig. 2A). Most of the substances in this class possess halogens (Cl or Br) with characteristic isotopic distributions (Fig. 2A).

Fig. 2.

Provided in this figure is (A.) the core skeletal structure of benzodiazepines (outlined in red) and mass spectrum of diazepam with select proposed structural annotations [33-36]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the benzodiazepines class (n = 56). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

For the Base Peak and High RA categories in this class, most of the m/z values are greater than m/z 200 and correspond to molecular ions (Fig. 2A) remaining largely intact; thus, these specific ions are not diagnostic for class identification. Most of these structures cleave the smaller functional groups [36,37] (i.e., alcohols, carbonyls, methyl groups, and halogens) resulting in larger number of high mass ions versus other classes (Fig. 2B). Ions at m/z 75 and m/z 77 are observed across the RA categories bins (Table 3) and can be explained by fragmentation at the phenyl ring and the R₄ group [32,38,39].

Table 3.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the benzodiazepines class (n = 56). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
205	3 (5)	75	6 (11)	75	33 (59)	29	45 (80)
286	3 (5)	77	5 (9)	102	29 (52)	27	36 (64)
75	2 (4)	288	5 (9)	177	28 (50)	28	28 (50)
204	2 (4)	269	4 (7)	51	26 (46)	63	19 (34)
222	2 (4)	286	4 (7)	163	26 (46)	26	18 (32)
223	2 (4)	287	4 (7)	76	26 (46)	35	16 (29)
259	2 (4)	294	4 (7)	89	25 (45)	104	13 (23)
267	2 (4)	315	4 (7)	63	24 (43)	119	13 (23)
275	2 (4)	203	3 (5)	77	24 (43)	43	12 (21)
283	2 (4)	239	3 (5)	151	23 (41)	189	12 (21)

The Medium RA category consists of ions commonly found in numerous reports [34,36,37,40-42]. As suggested by both the NIST MS Interpreter tool and literature reports [36,37,43,44], ions at m/z 102, m/z 151, m/z 163, and m/z 177 correspond to fragmentation across the diazepine ring of various core structures.

From the merged database investigation and literature reports, potential criteria for identification of novel benzodiazepines include (i) a molecular ion with at least Medium RA, (ii) ions at m/z 75 and/or m/z 77, (iii) mid-range RA ions indicative of the diazepine ring structure, (iv) a high number of Low and Medium RA peaks spread across the mass range, and (v) few to no High RA peaks below m/z 200. Neutral losses of 29 Da, 27 Da, and 28 Da are common and are the result of cleaving smaller functional groups.

3.3. Anabolic steroids (n = 50)

Anabolic steroids are the most frequently abused substances by participants in sporting events and athletic competitions [45-47]. These compounds are not frequently observed in forensic casework (accounting for less than 1% of drug reports in the NFLIS 2020 Annual Report [6]) but a large range of designer steroids are observed [48]. Newly emerging steroids that elude GC–MS identification are often modified at the 17α-position (R₂) of the core structure (Fig. 3A) [46]. These substances often possess an alcohol or ketone in the R₂ position and, when ionized, are cleaved at the multiple ring sites.

Fig. 3.

Provided in this figure is (A.) the core skeletal structure of anabolic steroids (outlined in red) and mass spectrum of testosterone with select proposed structural annotations [49,50]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the anabolic steroids class (n = 50). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Fragmentation of these molecules can occur across the entire core structure. The Base Peak RA category ions at m/z 122 and m/z 124 (Table 4) appear to be the product ions from the fragmentation across the B-ring fragmentation (Fig. 3A) [49-52]. Out of the 50 compounds in the merged database, two compounds (Methyltestosterone and 7-keto Dehydroepiandrosterone) possessed a molecular ion of m/z 302 suggesting no degradation/cleavage across the B-/C-ring. Other Base Peak and High RA categories ions involve m/z 43, m/z 91, and m/z 105 which are attributed to the C-/D-ring fragmentation [53]. For the structure of m/z 79, it has been suggested, via ESI fragmentation, to represent the cation of 5-methyl-cyclopenta-1,3-diene [54].

Table 4.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the anabolic steroids class (n = 50). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
124	8 (16)	91	22 (44)	119	41 (82)	181	31 (62)
43	7 (14)	79	20 (40)	41	39 (78)	207	31 (62)
122	6 (12)	43	14 (28)	133	39 (78)	195	29 (58)
91	4 (8)	93	13 (26)	131	38 (76)	183	27 (54)
105	2 (4)	55	12 (24)	145	37 (78)	193	27 (54)
302	2 (4)	81	9 (18)	77	37 (74)	209	27 (54)
		147	9 (18)	121	36 (72)	171	26 (52)
		105	8 (16)	107	36 (72)	221	26 (52)
		107	8 (16)	135	36 (72)	167	25 (50)
		77	7 (14)	105	35 (71)	157	24 (48)

For the Medium RA category in Table 4, the top observed ions, barring m/z 41 and m/z 77, were above 100 Da and are the remaining fragmentation products of the B-/C-ring structures [52]. It is important to note that ions have been previously reported in ESI-MS literature as “characteristic” of anabolic steroids (ie., m/z 97, m/z 109, m/z 143, m/z 169, and m/z 183) [51,53-60] did not appear among the top reoccurring EI-MS generated ions observed here.

From the merged database investigation and literature reports, potential criteria for identifying common anabolic steroids with EI-MS include (i) the presence of peaks at m/z 91, m/z 105, m/z 119, m/z 122, m/z 124, m/z 131, and/or m/z 133 at Medium RA or greater (ii) moderate number of peaks of Low and Medium RA that span the mass range, and (iii) presence of a molecular ion, though not at a uniform intensity. There are a number of neutral losses that are common to this class including 181 Da, 207 Da, and 195 Da.

3.4. Cannabinoids (n = 448)

Cannabinoids were first discovered from the derivation of the hemp plant Cannabis sativa and have been utilized for both recreational and medical purposes [61]. Cannabinoids are a structurally diverse family of compounds and are one of the most widely abused substances [6,7,61-63]. Depending on the source (synthetic, naturally occurring, metabolism-dependent), cannabinoids can be classified into three groups: phytocannabinoids, endocannabinoids, and synthetic cannabinoids. For this study, only phytocannabinoids and synthetic cannabinoids are discussed.

3.4.1. Phytocannabinoids (n = 24)

Phytocannabinoids are naturally occurring, plant-derived products and include Δ⁹-tetrahydrocannabinol (Δ⁹-THC), Δ⁹-tetrahydrocannabivarin (Δ⁹-THCV), cannabinol (CBN), cannabidiol (CBD), cannabidivarin (CBDV), cannabigerol (CBG), and cannabichromene (CBC) [64-66]. Δ⁹-THC was the second most frequently reported compound in the NFLIS 2020 Annual Report [6], accounting for approximately 15% of all identifications. The core skeletal structure of these compounds is represented in Fig. 4A along with the common substitution sites – the isopropenyl residue at the terpenoids moiety (1), the resorcinol nucleus (2) and the alkyl side chain (3) [65].

Fig. 4.

Provided in this figure is (A.) the core structure of the phytocannabinoids (outlined in red) highlighting the terpenoids moiety (red box 1), the resorcinol nucleus (blue box 2) and the alkyl side chain (black box 3) and mass spectrum of Δ⁹-THC with select proposed structural annotations [67-70]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the phytocannabinoids class (n = 24). Mass spectrum obtained from SWGDRUG Mass Spectral Libraiy (version 3.10.).

Roughly 29% of the compounds in this class displayed a Base Peak RA ion at m/z 231 (Table 5) which can be attributed to structural similarities at the resorcinol nucleus (B) and the 5-carbon alkyl side (C) [69,71]. The remaining ions in the Base Peak RA category are caused by fragmentation occurring at the terpenoids moiety or along the alkyl side chain. This observation remains consistent for the ions in the High RA category as the substitutions become bulkier [72].

Table 5.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the phytocannabinoids class (n = 24). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
231	7 (29)	299	3 (13)	174	16 (67)	83	16 (67)
137	2 (8)	314	3 (13)	41	14 (58)	68	9 (38)
203	2 (8)	271	2 (8)	43	9 (38)	43	8 (33)
313	2 (8)			232	9 (38)	15	7 (29)
				91	9 (38)	85	6 (25)
				77	8 (33)	121	6 (25)
				69	7 (29)	140	5 (21)
				175	6 (25)	112	4 (17)
				79	6 (25)	127	4 (17)
				260	6 (25)	141	4 (17)

With the Medium RA category, the m/z 174 ion is the most frequently observed and corresponds to the fragmentation of the alkyl side chain and across the terpenoids moiety [70,71]. Other common ions such as m/z 41, m/z 43, m/z 232, and m/z 91 correspond to fragments of the core structure containing the aromatic ring and the alkyl chain. The observations of these ions are also documented frequently throughout various literature reports [61,73,74].

From the merged database investigation and literature reports, potential criteria for identifying common phytocannabinoids include (i) the presence of ions at m/z 231/232, m/z 313/314, and m/z 174, (ii) a number of Low and Medium RA peaks across the m/z range, and (iii) a limited number of High RA peaks. Neutral losses of 83 Da and 68 Da are common. Because of the fragmentation nature of these substances, the presence and RA of a molecular ion is inconsistent.

3.4.2. Synthetic cannabinoids (n = 424)

Synthetic cannabinoids (SCs) such as “Spice”, “K2”, and “Joker”, have often eluded efforts to be regulated due to slight structural modifications [75]. According the NFLIS 2020 Annual Report [6], synthetic cannabinoids account for approximately 1% of all identifications and are spread across many different compounds and core structures (Fig. 5A). New synthetic cannabinoids are frequently reported by groups such as NPS Discovery, who reported 10 new compounds in 2021 alone [7].

Fig. 5.

Provided in this figure is (A.) the reported core structures of synthetic cannabinoids [76]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the synthetic cannabinoids class (n = 424).

Within this diverse subclass, there are a few notable trends between compounds across the RA categories (Table 6). For instance, the Base Peak RA ions at m/z 214 and m/z 232 correlate to core structures that are non-fluorinated and fluorinated, respectively, and are formed by fragmentation between the carbonyl carbon and benzyl carbon (i.e., the á-carbon) with one electron going into the indole-carbonyl system and the other electron forming the methoxybenzyl radical leaving group [77,78]. For the ions at m/z 215 and m/z 233, most of the core structures resemble the alkylindoles, indole, and carboxamides, barring substances similar to 3-epi CP 47,497 which has a cyclohexylphenol core structure. For compounds with a naphthoylindole, benzoylindole, or phenylacetylindole core structure, a common fragment ion of m/z 144 may form via a 1,3-hydride transfer from the alkyl carbon between the nitrogen and the alkyl group (R₃), forming a π-bond on the pentene leaving group [77].

Table 6.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the synthetic cannabinoids class (n = 424). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
232	47 (11)	127	27 (11)	144	187 (44)	57	87 (21)
214	44 (10)	145	21 (3)	116	112 (26)	17	84 (20)
233	31 (7)	284	20 (3)	41	106 (25)	71	67 (16)
109	29 (7)	144	18 (4)	254	93 (22)	127	63 (15)
215	21 (5)	155	18 (4)	128	88 (21)	15	61 (14)
127	14 (3)	298	14 (3)	43	83 (16)	186	58 (14)
284	11 (3)	253	14 (3)	241	75 (16)	214	56 (13)
100	10 (3)	214	13 (3)	115	72 (16)	85	47 (11)
241	8 (2)	232	11 (3)	145	70 (16)	144	47 (11)
355	8 (2)	341	10 (3)	270	70 (11)	115	44 (10)

Additionally, the m/z 109 ion has been recorded to be characteristic of specific fluorine-containing synthetic cannabinoids [79]. The structure of this ion is attributed to a benzyl moiety possessing a fluorine atom. Some structures possessing this Base Peak RA and High RA category ion include AB-FUBICA and FUB-AEB which have a relatively simplistic mass spectra consisting of a few notable peaks. Within the merged database, the Base Peak RA category ion of m/z 109 may be associated with other ions at m/z 252/253 which have varying intensities ranging from Low to High RA. These observed fluorinated compounds share indole or indazole carboxamide core structures which cleaves at the amino group near the R₁ group, as suggested by both the NIST MS Interpreter tool and literature reports [80,81].

For structures that are similar in structure to indazole carboxamide cannabinoids (i.e., AB-PINACA and AB-CHMINACA), the common fragment ions of m/z 145 and m/z 241 are observed which can be attributed to the main core structure [79,82-85]. Minimally substituted cyclohexylphenols structures (i.e., CP-47497) exhibit similar fragments like m/z 215 and m/z 233 [86].

With the merged database investigation and literature reports [81,87-89], potential criteria for identification of synthetic cannabinoids include (i) presence of m/z 109, m/z 116, m/z 144/145, m/z 214/215, and m/z 232/233, (ii) Low and Medium RA peaks, a majority of which are greater than m/z 240, and (iii) few to no High RA peaks below m/z 200. The presence and RA of a molecular ion is inconsistent across this class due to the wide variety of structures. Some common neutral loss ions include 57 Da, 17 Da, and 71 Da.

3.5. Opiates and opioids (n = 323)

With the initial extraction from the poppy seed, researchers identified the main alkaloids responsible for its pharmacologic effects [90]. These compounds are mainly used to mitigate pain by activating receptors in the brain to release endorphins and thus, depressing the central nervous system [6,90-92]. The first iteration of these substances was termed opiates, which contain both natural and semi-synthetic compounds. However, the use of natural opiates was quickly overshadowed by the development of heroin, a semi-synthetic opiate, which exhibits a similar core structure to morphine but with the addition of two acetyl groups. Since then, heroin has also been overshadowed with the reemergence of synthetic opioids (e.g., fentanyls, utopioids, and nitazenes) which are not structurally related to morphine but still target the μ-opioid receptors and have seen increasing prevalence in casework.

3.5.1. Opiates (n = 36)

First derived from the Papaver somniferum plant, opiates are naturally occurring compounds which are comprised of substances such as morphine, codeine, thebaine, papaverine, and noscapine [90]. According to the NFLIS 2020 Annual Report [6], substances like heroin, buprenorphine, and oxycodone represent the fifth, seventh, and eighth frequently observed opiates in casework accounting for approximately 10% of all drug identifications.

Most of the opiates within this class have similar core structures (Fig. 6A) with substitutions commonly found at the R₁ and R₃ groups. Looking at the merged database information, the Base Peak and High RA categories contain multiple ions which correspond to molecular ions (Table 7). From these two categories, m/z 285 and m/z 42 are ions observed at similar frequencies which can be attributed to the fused-ring core structure and fragments of smaller functional groups, respectively. Within the Medium RA category, ions at m/z 115 and m/z 128 occur in over 80% of the compounds. Interestingly, there have been few reports (e.g., ESI fragmentation) suggesting the structures of these ions are attributed to the losses from larger neutral fragments [93-95].

Fig. 6.

Provided in this figure is (A.) the core structure of opiates (outlined in red) and mass spectrum of heroin with select proposed structural annotations [93,94]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the opiates class (n = 36). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Table 7.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the opiates class (n = 36). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
285	5 (14)	42	5 (14)	115	31 (86)	57	16 (44)
301	4 (11)	55	3 (8)	128	30 (83)	85	14 (39)
303	3 (8)	70	3 (8)	77	24 (67)	259	14 (39)
317	3 (8)	162	3 (8)	127	22 (61)	87	13 (36)
287	2 (6)	216	3 (8)	44	22 (61)	101	13 (36)
299	2 (6)	327	3 (8)	91	21 (58)	128	13 (36)
343	2 (6)	43	2 (6)	42	20 (55)	100	12 (33)
		150	2 (6)	70	18 (50)	116	12 (33)
		157	2 (6)	200	17 (47)	245	12 (33)
		256	2 (6)	131	17 (47)	17	11 (31)

From the merged database investigation, potential criteria for identifying common opiates include (i) a molecular ion with at least a High RA, (ii) ions at m/z 115 and/or m/z 128, and (iii) a high number of Low and Medium RA ions spread across the m/z range. Neutral losses of 57 Da, 85 Da, and 259 Da are common.

3.5.2. Fentanyls (n = 237)

The rise, and continued prevalence, of fentanyl in casework has been well documented [6]. Fentanyl is the fourth most frequently reported drug in the NFLIS 2020 Annual Report [6], accounting for 9% of all identifications. While the number of fentanyl analogs reported has decrease, new analogs, such as bromofentanyl, chlorofentanyl, and fluorofentanyl are still prevalent and other analogs may continue to emerge [7]. The core structure of fentanyl is illustrated in Fig. 7A to which common substitutions occur at the amide (R₁), aniline (R₂), piperidine (R₃), and phenethyl chain (R₄) groups [96].

Fig. 7.

Provided in this figure is (A.) the core skeletal structure of fentanyl (outlined in red) and mass spectrum of fentanyl with select proposed structural annotations [96-102]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the fentanyl class (n = 237). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

This subclass posseses a few notable trends between compounds across the RA categories. For instance, the frequently observed Base Peak RA ions at m/z 245 and m/z 259 differ by 14 Da which, structurally, correspond to the presence of a methyl group on the fragment ion caused by breaking the phenethyl chain and piperidine. The compounds found with ions m/z 245 and m/z 259 share a core similarity to fentanyl. Conversely, structures that generate a Base Peak RA m/z 91 ion substitute the core structure with a N-benzyl or N-methyl moiety [101-103]. Compounds that produce an ion at m/z 95 (Table 8) undergo cleavage along the amide C─N bond correlating to resonance-stabilization of the R₁ functional group (i.e., 2-Furanyl fentanyl) [96].

Table 8.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the fentanyl class (n = 237). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
245	36 (15)	146	31 (13)	42	173 (73)	91	140 (59)
259	25 (11)	160	19 (8)	96	152 (64)	134	74 (31)
91	16 (7)	82	13 (6)	91	148 (62)	259	45 (19)
95	15 (6)	189	12 (5)	105	137 (58)	204	43 (18)
277	15 (6)	164	10 (4)	77	107 (45)	293	43 (18)
273	10 (4)	203	9 (4)	57	76 (32)	279	38 (16)
231	8 (3)	69	8 (3)	146	71 (30)	245	37 (16)
275	8 (3)	96	7 (3)	132	58 (24)	190	32 (14)
83	7 (3)	105	7 (3)	189	54 (15)	297	32 (14)
271	7 (3)	176	5 (2)	41	45 (15)	307	32 (14)

For the High RA category, the ions at m/z 146 and m/z 189 reveal compounds with no substitution of the aniline (R₂) or piperidine (R₃) regions. These ions also appear in the Medium RA category (Table 8) along with m/z 42, m/z 91, m/z 96, and m/z 105 which are also reported in literature [15,96,97,104,105]. An ion at m/z 96 can occur for a variety of reasons including the isotope of a Base Peak RA ion of m/z 95 (i.e., furanyl fentanyl and its analogs) [106] or as the fragmentation of the piperidine ring (suggested by NIST MS Interpreter tool). For the ions at m/z 42, m/z 91, and m/z 105, there are a variety of reports that depict the formation of these ions as fragmentation of the pyridine ring, the tropylium ion, and a methyl-substituted tropylium ion, respectively [91,96,97,104,107,108]. From the merged database investigation and literature reports, potential criteria for identifying common fentanyl-related compounds include (i) the presence of ions at m/z 245 or m/z 259, m/z 91, and m/z 105, (ii) lack of an appreciable (>5% RA) molecular ion, and (iii) spectra dominated by Low and Medium RA peaks below m/z 160. Neutral losses of 91 Da and 134 Da are common. It should be noted that tools and statistical strategies exist for the determination potential structures of novel fentanyls with single modifications [11,15,96].

3.5.3. Utopioids (n = 35)

In conjunction with fentanyl-related cases, other new synthetic opioids have emerged within the past decade [4,5,109]. Many of these initial cases reported U-47700, however, an influx of other U-compounds (U-51754, U-49900, U-48800, etc.) started to appear to elude legislature once U-47700 was scheduled [110]. While the U-series compounds are not among the most frequently reported drugs, new U-compounds (such as trifluoiOmethyl-U-47700 and naphthyl-U-47700) have been reported recently [6,7]. The core structure of this subclass is illustrated in Fig. 8A and is generally substituted on the amine group of the cyclohexane ring (R₁ and R₂), the amine group (R₃), or the phenyl ring (R₄) [109].

Fig. 8.

Provided in this figure is (A.) the core skeletal structure of utopioids (outlined in red) and mass spectrum of U-47700 with select proposed structural annotations [91,100,111,112]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the utopioids class (n = 35). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Although the merged database contains only 35 compounds, there are still notable trends for these compounds. Across the Base Peak and High RA categories (Table 9), ions at m/z 84 and m/z 125 (present in 77% and 69% of compounds, respectively) relate to the N, N-dime-thylcyclohexanamine fragment (Fig. 8A) [113]. The m/z 125 ion is caused by cleaving the bond between the amide moiety and the cyclohexane [100,111,114,115]. The fragmentation pathway for the generation of the m/z 84 ion has not been proposed. Using the NIST MS Interpreter tool, one potential structure for the m/z 84 ion (Fig. 8A) involves fragmentation across the cyclohexane ring while maintaining the amide group. The remaining m/z values in this category correspond to compounds with varying methyl groups that still fragment at the bond between the amide moiety and the cyclohexane.

Table 9.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the utopioids class (n = 35). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
84	19 (54)	125	18 (51)	71	26 (74)	183	20 (57)
125	6 (17)	84	8 (23)	58	25 (71)	218	10 (29)
111	3 (9)	70	2 (6)	124	24 (69)	244	10 (29)
112	2 (6)	141	2 (6)	110	20 (57)	155	9 (26)
				42	19 (54)	181	8 (23)
				173	10 (23)	186	7 (20)
				145	10 (29)	211	7 (20)
				147	9 (26)	240	7 (20)
				175	8 (23)	169	6 (17)
				126	8 (23)	203	6 (17)

For the Medium RA category, ions at m/z 58 and m/z 71 occur in over 70% of the compounds and suggest the cleavage from the cyclohexane and a 1,2 dissociation of the cyclohexane ring. Interestingly, many of these compounds possess halogens on the aromatic ring (R₄) yielding characteristic isotopic distributions. There are several compounds in this class that have two chlorine atoms present, creating a unique isotopic pattern consisting of three peaks, each separated by 2 Da with the ratio of 100:63.9:10.2 [22]. Two examples of this pattern are observed in the U-47700 mass spectrum (i.e., m/z 145/147/149 and m/z 173/175/177) which corresponds to the benzene ring containing two chlorine atoms with and without a carbonyl functional group [112].

From the merged database analysis and literature reports, potential criteria for identifying traditional U-series compounds include (i) ions at m/z 84, m/z 125, m/z 58, and/or m/z 71, (ii) lack of an appreciable (>5% RA) molecular ion, and (iii) a number of Low and Medium RA peaks below m/z 160. For the U-series substances that did not follow this trend, the NIST MS Interpreter tool was used to understand the fragmentation. Within the High RA category, the m/z 70 ion is suggested (via the NIST MS Interpreter tool) to result from fragmentation across the cyclohexane for both 3,4-Difluoro-N-desmethyl U-47700 and N-Desmethyl U-47700. The m/z 141 ion appears (suggested by the NIST MS Interpreter tool) to be generated by fragmenting the amine and carbonyl groups on 3,4-Difluoro-N-desmethyl U-47700 and 3,4-Difluoro Isopropyl U-47700 similarly to the other utopioids [100]. Even with substitutions, the substances included in the merged database exhibit similarly distributed mass spectral characteristics. Common neutral losses for this class includes 183 Da, 218 Da, and 244 Da.

3.5.4. Nitazenes (n = 15)

Contributing to the surge of synthetic opioids is a new class of compounds commonly referred to as “nitazenes”. Like utopioids, nitazenes are not among the most frequently observed compounds but continued detection of new compounds in casework in recent months (such as N-piperidinyl etonitazene and metonitazene) has been reported [116]. The core structure of these compounds has three substitution regions (Fig. 9A) located on the aminoethyl chain (R₁), the benzimidazole core (R₂), and the connected benzyl moiety (R₃).

Fig. 9.

Provided in this figure is (A.) core skeletal structure of the nitazenes compounds (outlined in red) and mass spectrum of metonitazene with select proposed structural annotations [14,116,117]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the nitazenes class (n = 15). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

The mass spectra of a nitazene may possess several noteworthy ions (Table 10). An ion at m/z 86 occurs in the Base Peak RA category in 80% of the substances, corresponding to the fragmentation (Fig. 9A) of diethyl aminoethyl moiety (R₁). The three compounds that do not possess this fragment ion (N-Desethyl Isotonitazene, N-Pyrrolidino Etonitazene, N-Desethyl Etonitazene) have Base Peak RA ions at m/z 149, m/z 84, and m/z 135, respectively, which can be attributed to varying substitutions at the R₁ moiety.

Table 10.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the nitazenes class (n = 15). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
86	12 (80)	107	5 (33)	87	12 (80)	261	5 (33)
		207	3 (20)	107	6 (40)	310	4 (27)
		121	2 (13)	58	5 (33)	324	3 (20)
		235	2 (13)	77	3 (20)	57	2 (13)
						130	2 (13)
						233	2 (13)

The second frequently encountered fragment, m/z 107, corresponds to the benzyl moiety (R₃) with an ether group while the ions at m/z 207 and m/z 235 correspond to the resulting benzimidazole core (R₂) with the difference of an ethyl substitution [117]. Other frequently observed ions at (i) m/z 87, (ii) m/z 58, and (iii) m/z 77 are formed from (i) the natural isotope of m/z 86, (ii) fragmentation of the diethyl aminoethyl fragment (R₁), and (iii) the resulting phenyl ring fragment of the benzyl moiety (R₃).

From the limited research [14,116,118,119] along with this merged database investigation, potential characteristic ions for nitazenes include (i) ions at m/z 86 and m/z 107, (ii) a mass spectrum containing few peaks, (iii) lack of an appreciable (>5% RA) molecular ion, and few, if any, ions above m/z 120. Common neutral losses observed are 261 Da and 310 Da.

3.6. Stimulants (n = 600)

Stimulants represent a wide range of compounds that are used to create feelings of euphoria. These compounds are amongst the most frequently encountered in forensic casework [1] and recent reports have indicated that an increase in overdoses involving stimulant use in combination with fentanyl-related compounds may be coming [120,121].

3.6.1. Synthetic cathinones (n = 229)

Among this new wave of stimulants are synthetic cathinones which are often labelled as “bath salts”, “plant food”, or “insect repellant” to elude law enforcement [122-124]. Eutylone is the only cathinone listed on the NFLIS 2020 Annual Report [6] as a top 25 reported drug (ninth on the list, compromising 1% of all identifications) but, like synthetic cannabinoids, there are a wide range of these novel compounds that have been identified in casework [6]. Generally, synthetic cathinones have several substitution locations on the core structure (Fig. 10A) at the benzyl moiety (R₁), the α-alkyl carbon (R₂), and the amino group (R₃ and R₄). The EI mass spectra of synthetic cathinones are dominated by the formation of iminium and acylium ions and readily dissociate at the α;- and β-carbon bond [125-129].

Fig. 10.

Provided in this figure is (A.) the core skeletal structure of synthetic cathinones (outlined in red) and mass spectrum of butylone with select proposed structural annotations [3,128,129]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the cathinones class (n = 229). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

For synthetic cathinones, it has been reported that commonly identified Base Peak RA ions of m/z 58, m/z 72, m/z 86, or m/z 100 (Table 11) represent fragment ions corresponding to varying alkyl chain lengths in the R₃ substitution [127]. This observation is consistent with what was observed in the merged database.

Table 11.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the cathinones class (n = 229). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
72	36 (16)	134	4 (2)	44	41 (18)	149	49 (21)
58	30 (13)			77	36 (16)	105	36 (16)
86	25 (11)			95	22 (10)	119	35 (15)
100	23 (10)			65	20 (9)	123	22 (10)
126	23 (10)			56	20 (9)	128	22 (10)
140	14 (6)			91	20 (9)	100	21 (9)
128	13 (6)			58	19 (8)	133	21 (9)
112	11 (5)			149	18 (8)	114	19 (8)
98	10 (4)			127	16 (7)	161	17 (7)
114	8 (4)			75	14 (6)	139	16 (7)

The mass spectra for the cathinones consist of one Base Peak RA ion followed by a few Medium RA peaks. In addition to the previously mentioned m/z values, there are trends with regards to ions at m/z 98, m/z 112, m/z 126, and m/z 140. For the compounds that exhibit these m/z values, the amino fragment is instead substituted by a pyrrolidine moiety. Within the Medium RA category, m/z values of m/z 44 and m/z 77 correspond to minimal or no substitutions on the amino group and the aromatic rings, respectively.

From some published literature [127,130] along with this merged database investigation, potential criteria for the identification of cathinones include (i) Base Peak and High RA ions at m/z 58, m/z 72, m/z 86, and/or m/z 100, (ii) Medium RA ions of m/z 44 and m/z 77, (iii) spectra with only a few appreciable ions, and (iv) lack of an appreciable (>5% RA) molecular ion. Spectra of these compounds often have few peaks of appreciable intensity and those that are present are typically below m/z 120. Common the neutral losses include 149 Da, 105 Da, and 119 Da.

3.6.2. Amphetamines (n = 174)

One of the most abused compound classes to date is the amphetamine derivatives class which had a global seizure of 144,000 kg (144 metric tons) between 2003 and 2012 [131]. Methamphetamine was the most frequently reported drug in the NFLIS 2020 Annual Report [6] accounting for approximately 29% of all identifications. Amphetamine and 3,4-methylenedioxy-N-methylamphetamine (MDMA) are also frequently encountered compounds [132]. The core structure of amphetamines is derived from phenethylamine, represented in Fig. 11A, and can be substituted on the phenyl ring (R₁), the amino group, and the alkyl chain.

Fig. 11.

Provided in this figure is (A.) the core skeletal structure derived from phenethylamine (outlined in red) and mass spectrum of methamphetamine with select proposed structural annotations [133]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the amphetamines class (n = 174). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Under EI conditions, amphetamines readily dissociate at the α- and β-carbon bond, resulting in simple mass spectra of very few peaks. The mass spectra for these compounds often exhibit a Base Peak RA ion with up to one additional High RA peak and a small number of Medium and Low RA peaks.

Some frequent ions identified by the merged database investigation and literature values are m/z 44, m/z 58, and m/z 72 which are comprised of the fragmentation of the α- and β-carbon bond and the resulting in the amino groups (Fig. 11A) [131,134,135]. Out of the 174 compounds in the merged database, these three ions account for approximately 69% of all Base Peak RA ions observed (Table 12). Aside from these ions, as well as m/z 77 and m/z 91, the remaining ions are largely specific to unique subsets of compounds.

Table 12.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the amphetamines class (n = 174). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
44	58 (33)	44	7 (4)	77	32 (18)	43	43 (25)
58	43 (25)	91	6 (3)	91	29 (17)	58	29 (17)
72	19 (11)	43	3 (2)	44	21 (12)	44	24 (14)
86	5 (3)	166	3 (2)	65	16 (9)	91	22 (13)
91	5 (3)	102	2 (1)	89	16 (9)	72	21 (12)
84	4 (2)	130	2 (1)	131	15 (9)	131	16 (9)
100	2 (1)	135	2 (1)	151	13 (7)	135	15 (9)
116	2 (1)	149	2 (1)	135	10 (6)	151	15 (9)
118	2 (1)	150	2 (1)	90	9 (5)	100	12 (12)
121	2 (1)	164	2 (1)	63	9 (5)	86	11 (6)

From the literature [132,134,135] and merged database investigation, potential criteria for identification of amphetamines include (i) ions at m/z 44, m/z 58, m/z 72, and/or m/z 91, (ii) a mass spectrum with a Base Peak RA, possible one additional High RA peak, and relative few Medium or Low RA peaks, (iii) spectra where nearly all of the peaks are below m/z 160, and (iv) lack of an appreciable (>5% RA) molecular ion. Common neutral losses are 43 Da and 58 Da.

3.6.3. Phenethylamines (n = 127)

Phenethylamines are another class of compounds increasingly being observed in casework that, like synthetic cannabinoids, have a wide range of core structures. These core structures include amphetamine-like, “2C”-series, “D”-series, “NBOMe”-series, “FLY”- series, and “DRAGONFLY”-series compounds [136-140]. These sub-groups are based on the different substitutions found on the aromatic ring (R₁), on the α- and β-carbon alkyl chain (R_α and R_β), or on the nitrogen atom (R_N) (Fig. 12A).

Fig. 12.

Provided in this figure is (A.) common derivatives of the phenethylamine core structure [144]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the phenethylamines class (n = 127).

The core phenethylamine structure is primarily preserved in the amphetamine-like compounds with simple substitutions on the R_α or on the phenyl group (observed with MDMA). More complex structures were added to the phenyl group with the introduction of the “2C”-series where there are two methoxy groups on the 2- and 5- positions [141]. The “D”-series consists of two methoxy groups on the phenyl ring, and a single substituent on the 4- position of the phenyl ring. The NBO-series, and similarly the NBOMe-series (Fig. 12A), stem from the “2C”-series and have methoxy groups at the 2- and 5- positions of the phenyl ring, a substitution on the 4- position of the phenyl ring, and a methoxy or other substitutions at the ortho-, meta-, or para- positions of the N-benzyl ring [142]. The FLY-series and DRAGONFLY-series consists of tetrahydrobenzodifuran and benzodifuran located on the phenyl group where the only difference is the DRAGONFLY-series is fully substituted on this ring [143].

Compounds in this class that consist of amphetamine-like structures show similarities with the amphetamines class discussed above, dominated by a formation of ions at m/z 91 and m/z 44. Within the 2C class, there appear to be no common peaks. However, there is a neutral loss of 15 Da which can be explained by the loss of the methyl radical [133]. The d-series has very simplistic mass spectra with the primary Base Peak RA category ion of m/z 44 followed by Medium RA peaks [145] (Table 13).

Table 13.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the phenethylamines class (n = 127). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
121	24 (19)	150	5 (4)	91	45 (35)	29	28 (22)
245	12 (9)	146	4 (3)	77	28 (22)	44	26 (20)
259	12 (9)	91	3 (2)	146	19 (15)	28	22 (17)
44	11 (9)	164	3 (2)	150	17 (13)	30	21 (17)
84	8 (6)	183	3 (2)	246	14 (11)	43	15 (12)
58	6 (5)	44	2 (2)	260	14 (11)	149	15 (12)
72	3 (2)	138	2 (2)	189	13 (10)	58	14 (11)
86	3 (2)	151	2 (2)	56	12 (9)	72	13 (10)
167	3 (2)	165	2 (2)	110	12 (9)	86	12 (9)
91	2 (2)	167	2 (2)	96	11 (9)	60	10 (8)

The NBOMe-phenethylamines have distinct spectra with characteristic fragments of the tropylium cation (m/z 91), the methoxy-substituted tropylium ion (m/z 121 (Base Peak)), and the methoxy-substituted tropylium ion with a secondary amine chain (m/z 150) [137] which can help differentiate between positional isomers [146]. Despite substitutions on the aromatic ring in the NBOMe-phenethylamines, these three ions dominate the spectra and consequently can be considered characteristic ions for this subclass [137].

The merged database consists of four compounds within the FLY and DragonFLY categories – 2C-B-FLY, N-(2C-B-fly) fentanyl, N-(3C-B-fly) fentanyl, and Bromo-Dragon-FLY. The two fentanyl analogs exhibit similar mass spectral characteristics to the fentanyl fragmentation m/z 146, m/z 189, and m/z 245. For the 2C-B-FLY and Bromo-Dragon-FLY compounds, both show isotopic distributions of Br and fragmentation of the amino group on the tail. No further inferences were made for these compounds due to lack of compounds representing this specific structure.

Based on the merged database information and reported literature, potential diagnostic characteristics of phenethylamines include (i) spectra with a Base Peak RA ion and up to one additional High RA peak, (ii) spectra with a slightly higher number of Medium and Low RA peaks than other stimulant subclasses, (iii) spectra dominated with peaks below m/z 160, and (iv) spectra with either no molecular ion or a molecular ion with Medium RA. Common neutral losses are 29 Da and 44 Da.

3.6.4. Arylcyclohexylamines (n = 42)

The arylcyclohexylamine subclass consists of commonly observed compounds, such as phencyclidine (PCP) which is one of the top 25 reported compounds [6], as well as novel emerging compounds such as recently reported F-PCP, Cl-PCP, POXP, and PTHP. Compounds in this class consist of a core structure with substitutions occurring at the aromatic ring (R₁), at the cyclohexane ring (R₂), and at the amine moiety (R₃ and R₄) (Fig. 13A).

Fig. 13.

Provided in this figure is (A.) the core skeletal structure of arylcyclohexylamines (outlined in red) and mass spectrum of PCP with select proposed structural annotations [147,148]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the arylcyclohexylamines class (n = 42). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Investigating the mass spectra of the 42 compounds in the merged database, there are no obvious trends within the Base Peak and High RA categories but there is commonality in the fragmentation pathways. For instance, compounds that are structurally similar to PCP exhibit neutral losses of 42/43 Da. Substances that have a neutral loss of 57 Da are indicative of a ketone substitution on the cyclohexane ring (Fig. 13A) which is found in ketamine-like compounds. Using the NIST MS Interpreter tool as a guide, other common neutral losses of 28 Da and 85 Da are caused by various fragmentation across the cyclohexane ring (Table 14).

Table 14.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the arylcyclohexylamines class (n = 42). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
164	3 (7)	91	2 (5)	115	16 (38)	57	33 (79)
176	3 (7)	105	2 (5)	91	16 (38)	85	22 (52)
190	3 (7)	150	2 (5)	77	11 (26)	28	20 (48)
204	3 (7)	157	2 (5)	117	10 (24)	56	20 (48)
160	2 (5)	256	2 (5)	84	9 (21)	43	18 (43)
166	2 (5)			104	8 (19)	99	15 (36)
200	2 (5)			134	8 (19)	42	14 (33)
216	2 (5)			102	7 (17)	71	13 (31)
				166	7 (17)	113	13 (31)
				138	7 (17)	126	13 (31)

For the ions in the Medium RA category, m/z 91 and m/z 115 each appear in 38% of the compounds, suggesting fragmentation and formation of the phenyl ring with a variety of functional groups [94].

Based on the merged database information and reported literature, potential diagnostic characteristics of arylcyclohexylamines include (i) ions at m/z 91 and m/z 115, numerous Low and Medium RA ions below m/z 200, and (iii) one to two High RA peaks in the m/z 120 to m/z 240 range. The presences, and RA, of a molecular ion is inconsistent. Common neutral losses for this class are 57 Da, 85 Da, 28 Da, 56 Da, and 43 Da.

3.6.5. Piperazines (n = 28)

Piperazines are compounds often found at raves and other nightclubs and are sold as “ecstasy” pills or under names such as “Frenzy”, “-Bliss”, “Charge”, “Herbal ecstasy”, “A2”, “Legal X” and “Legal E” [149]. According to the NFLIS 2020 Annual Report [6], these compounds are not frequently encountered in casework but are still of interest to the community. The core structure of these substances is piperazine which, like other core structures, is amenable to substitution. The piperazine sub-class can be further split into three categories (Fig. 14A) – benzylpiperazines (i.e., N-benzylpiperazine (BZP)), phenylpiperazines (i.e., 1-(3-chlorophenyl)piperazine (mCPP)), and thienylmethylpiperazines (i.e., 1-[(5-Ethylthien-2-yl)methyl]piperazine) [149].

Fig. 14.

Provided in this figure is (A.) the core skeletal structure of a piperazine ring (centered) with observed piperazine designer drugs (outlined in red) of the benzylpiperazine (left) and phenylpiperazines (right) structures. Also provided is a mass spectrum of BZP with select proposed structural annotations [150,151]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the piperazines class (n = 28). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

Investigating the mass spectra of the 28 compounds in the merged database, there are some commonalities of the ions spanning the RA categories and even amongst the neutral losses. For instance, neutral losses of 42 Da and 85 Da are common across the subclass (Table 15). Using the NIST MS Interpreter tool as a guide, each mass spectrum and their structures were combed for commonalities.

Table 15.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the piperazines class (n = 28). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
91	4 (14)	56	2 (7)	56	18 (64)	42	22 (79)
188	4 (14)			85	6 (21)	85	20 (71)
138	3 (11)			230	5 (18)	57	14 (50)
105	2 (7)			65	4 (14)	58	14 (50)
135	2 (7)			172	4 (14)	40	7 (25)
150	2 (7)			122	4 (14)	56	7 (25)
154	2 (7)			156	4 (14)	41	5 (18)
172	2 (7)			173	4 (14)	91	4 (14)
				77	4 (14)	134	4 (14)
				91	4 (14)	84	3 (11)

Substances exhibiting a loss of 85 Da were found to be benzylpiperazines with substitutions on the aromatic ring. The traditional phenylpiperazine-like compounds, possessing substitutions on the aromatic ring, have Base Peak RA ions with neutral losses of 42 Da [152,153]. Neutral losses of 58 Da are also attributed to phenylpiperazine-like compounds where the nitrogen atoms are not connected to the phenyl ring [152,153]. There are few instances where a neutral loss 72 Da is observed (i.e., 1-(4-Methoxyphenyl)piperazine) which indicates a phenylpiperazine-like compound with substitutions on both the aromatic and piperidine rings. Other instances occur with mass spectral differences of 91 Da where those structures share the core benzylpiperazine-like moiety but with bulkier side chains or different substitutions on the piperidine ring. Interestingly, these compounds always had a molecular ion present between 10% and 50% RA.

Based on the merged database information and reported literature, potential diagnostic characteristics of piperazines include (i) presence of an ion at m/z 56 with Medium or High RA, (ii) only one peak of High RA along with several Low and Medium RA peaks below m/z 200, and (iii) presence of a molecular ion with Low or Medium RA. Common neutral losses include 42 Da, 85 Da, 57 Da, and 58 Da.

3.7. Hallucinogens (n = 62)

Hallucinogenic compounds are often found in plants and fungi [154] or are synthetic compounds based on these naturally occurring compounds [155]. Many of the phenethylamine substances, such as PCP and ketamine, can be categorized as both a stimulant and hallucinogen depending on dosage and target sites in the CNS [156-158]. Amongst the hallucinogens, the psychedelics are the most studied and reported substances and are typically comprised of phenethylamines (previously discussed, Section 3.6.3) and tryptamines [121,155,159,160].

3.7.1. Tryptamines (n = 57)

Some well-known tryptamines like dimethyltryptamine (DMT), lysergic acid diethylamide (LSD), and psilocybin are either natural or semisynthetic products originating from various fungi or plants, and even animals [159,161]. These compounds are frequently reported in casework, with both LSD and psilocin being amongst the top 25 most frequently reported drugs in the NFLIS 2020 Annual Report [6]. NPS Discovery has also reported the observation of new, designer tryptamines, such as 4-acetoxy EPT, in recent months [7]. Tryptamines can be categorized into two subclasses based on structure, ergolines (synthesized from the ergot fungus) [162] and simple tryptamines [163].

The scaffolding of tryptamines varies between subclasses but is centered around the indole ring structure [162-165]. For simple tryptamines (Fig. 15A), the indole ring attaches to an amino group via a two-carbon sidechain (monoamine chain) and can have various substitutions at the nitrogen (R_N1 and R_N2) and at the α-carbon (R_α). Like the amphetamines class, the simple tryptamines readily dissociate at the α- and β-carbon bond and have relatively simplistic mass spectra [166].

Fig. 15.

Provided in this figure is (A.) the core skeletal structure of tryptamines (outlined in red) and mass spectrum of DMT with select proposed structural annotations [139,167]. Supplemental histograms depict (B.) average distribution of Low (light blue), Medium (blue), and High + Base Peak (dark blue) ions across the m/z range, (C.) average number of peaks for each RA category (excluding Ultra-Low), and (D.) frequency of the presence of a molecular ion across the RA categories for the tryptamines class (n = 57). Mass spectrum obtained from SWGDRUG Mass Spectral Library (version 3.10.).

For these substances, this fragmentation often involves the monoamine group and indole core. For the 57 simple tryptamine compounds, there are slight trends found amongst the Base Peak and High RA categories (Table 16). Fragmentation of the α- and β-carbon bond, resulting in amino groups with vaiying alkyl chains produces ions at m/z 58, m/z 72, m/z 86, m/z 100, m/z 114, and m/z 142. The m/z 84 ion is the product of the methyl-N-allyltryptamine (MALT) substances which has a double bond on the amino group. The remaining fragment ions in the Base Peak and High RA categories (m/z 130, m/z 131, m/z 142, and m/z 161) correspond to a resulting α cleavage and an indole core structure with varying substitutions [166].

Table 16.

Most frequently observed ions, neutral losses, and the total percentage of these values found in the tryptamines class (n = 57). The neutral losses represent those from Base Peak, High, and Medium RA ions, combined.

Base Peak RA		High RA		Medium RA		Neutral Loss
m/z	Count (%)	m/z	Count (%)	m/z	Count (%)	Da	Count (%)
114	12 (21)	44	5 (9)	146	9 (16)	160	17 (30)
58	11 (19)	130	5 (9)	160	9 (16)	130	12 (21)
86	11 (19)	58	2 (4)	44	8 (14)	188	12 (21)
72	3 (5)	131	2 (4)	72	8 (14)	146	10 (18)
100	3(5)			77	6 (11)	202	9 (16)
131	3 (5)			117	6 (11)	58	7 (12)
84	2 (4)			130	6 (11)	43	6 (11)
142	2 (4)			115	5 (9)	44	6 (11)
161	2 (4)			103	4 (7)	114	5 (9)
				143	4 (7)	29	4 (7)

Based on the merged database information and reported literature, potential diagnostic characteristics of tryptamines include (i) an ion at m/z 58, m/z 72, m/z 86, m/z 100, and/or m/z 114, (ii) mass spectra exhibiting one or two High RA peak along with a higher number of Medium RA peaks, (iii) spectra with peaks below m/z 160, and (iv) either no or a Medium RA molecular ion. Neutral losses of 130 Da, 146 Da, 160 Da, and 188 Da are common and correspond to the remaining indole moiety with varying substitutions.

3.7.2. Lysergamides (n = 5)

For the five lysergamides, molecular ions are in the High RA category which is expected for these fused ring structures. Common ions reported within literature and the merged database investigation for these ergoline compounds are m/z 181, m/z 207, and m/z 221. Additionally, the neutral loss of 102 Da is common from the molecular ion to the next highest peak. This process has been proposed as a sequential loss of 101 Da from N,N-diethyl acrylamide followed by the loss of a radical H• [168,169] resulting in the ergoline base structure with a methyl group. These conclusions from the merged database investigation are conservative as there were only five lysergamide entries.

3.8. Example interpretation of an unknown mass spectrum

As discussed in this work, there are mass spectral trends that can be used to help analysts classify unknown compounds. An example of how this is demonstrated in Fig. 16 using the synthetic cathinone “iProne” as an example. Several additional examples are also provided in the Supplemental Information.

Fig. 16.

Provided in this figure is an example mass spectrum with supplemental observations of the major ions. The occurrences of these ions correspond to the RA bins that were implemented in the code.

The “unknown” spectrum provided in Fig. 16 has no notable m/z values above m/z 200 but does have a number of ions with appreciable intensity below m/z 200. Additionally, the spectrum has a Base Peak RA ion at m/z 86 and does not appear to have a molecular ion of notable abundance. Using these characteristics, one can narrow down the list of likely classes to nitazenes, cathinones, or amphetamines. Nitazenes (Fig. 9) were found to have, on average, only one Medium RA ion along with a Base Peak RA ion which is not consistent with this “unknown” mass spectrum where a number of Medium RA ions (m/z 149, m/z 121, m/z 91, m/z 65, and m/z 44) are observed. Other spectral characteristics, namely the presence of the m/z 44 ion with Medium RA and a presumed neutral loss of 105 Da provides additional clues this “unknown” is more likely a cathinone rather than an amphetamine. Given these spectral characteristics, this compound is likely a cathinone with a benzodioxole core structure (similar to Fig. 10A).

4. Conclusions

The influx of new/novel illicit psychoactive substances continues to strain limited resources and increase backlogs of forensic laboratories. For most drug chemists, GC-EI-MS remains the primary technique for identifying and characterizing NPSs. While this resource can help classify an unknown compound, definitive identification, especially when a standard is not available, will require the use of other analytical techniques or statistical approaches. Other analytical techniques that can provide additional information for structural elucidation include LC-MS/MS [170,171], ¹H NMR [78,115], spectroscopic methods [106,172,173], and DART-MS/MS [12,18,174,175]. An orthogonal approach is to leverage multivariate statistics or machine learning algorithms to identify important features contributing to compound characterization within mass spectra [11,15,176,177]. Use of this approach has been successful in differentiating illicit substances and positional isomers based upon mass spectra to explain the influence of slight structure modifications [51,77,93,125,137].

It should be noted that this work only discusses the mass spectral component of the GC-EI-MS and does not discuss chromatographic retention times, which can provide another critical component for unknown compound classification. However, because of a lack of method standardization in the field, no definitive statements regarding retention characteristics can be made. Individual laboratories can monitor retention times of known compounds and use them to better narrow down the potential class(es) of unknown compounds. Laboratories may also want to consider locked retention times or retention indices to allow for comparison of retention data across different methods or instruments. For researchers interested in this field, additional high-resolution MS and/or labelling experiments would greatly assist in the understanding of more complex fragmentation mechanisms in order to fill gaps in the literature. Ultimately, the interest of the forensic community to increase and understand common trends in drug analysis is crucial as it can potentially lead to quicker characterizations and decrease backlogs.

Data availability

Data will be made available on request.