Analytical and Bioanalytical Chemistry (2022) 414:5643–5656

https://doi.org/10.1007/s00216-022-04035-1

RESEARCH PAPER

Simultaneous evaluation of the enantiomeric and carbon isotopic

ratios of Cannabis sativa L. essential oils by multidimensional gas
chromatography
Lorenzo Cucinotta1,2 · Gemma De Grazia1 · Giuseppe Micalizzi1 · Luana Bontempo2 · Federica Camin2,3,4 ·
Luigi Mondello1,5,6 · Danilo Sciarrone1

Received: 19 December 2021 / Revised: 14 March 2022 / Accepted: 21 March 2022 / Published online: 7 April 2022
© Springer-Verlag GmbH Germany, part of Springer Nature 2022, corrected publication 2022

Abstract
Recent times have witnessed an upsurge of interest in hemp and hemp-derived products, as driven by the scientific findings
specific to the pharmacological properties of Cannabis sativa L. and its constituents. There has been evidence that the ter-
pene profile, along with the cannabinoid content, produces in humans the effects associated with different strains, beyond
fragrance perception. A great deal of effort has been put into developing analytical approaches to strengthen the scientific
knowledge on cannabis essential oil composition and provide effective tools for ascertaining the authenticity of commer-
cial cannabis samples. For this concern, enantio-selective-GC-C-IRMS has proven to be effective for assessing the ranges
characteristic of the genuine samples and detecting any fraudulent additions. This research aimed at providing for the first
time the enantiomeric and isotopic ratios of target terpenes in cannabis essential oils, obtained from microwave-assisted
hydro-distillation from the fresh and dried inflorescences of different cannabis varieties. Implementing multidimensional
gas chromatography separation was mandatory prior to detection, in order to obtain accurate δ13C values and enantiomeric
data from completely separated peaks. For this purpose, a heart-cut method was developed, based on the coupling of an
apolar first dimension column to a secondary chiral cyclodextrin-based stationary phase. Afterwards, the data gathered from
enantio-selective-MDGC-C-IRMS/qMS analysis of a set of genuine samples were used to evaluate the quality of nineteen
commercial cannabis essential oils purchased from local stores. Remarkably, the data in some cases evidenced enantiomeric
ratios and δ13C values outside the typical ranges of genuine oils. Such findings suggest the usefulness of the method developed
to ascertain the genuineness and quality of cannabis essential oils.

Keywords Multidimensional gas chromatography · Enantio-selective gas chromatography · Isotopic ratio mass
spectrometry · Cannabis sativa L. · Chiral terpenes

Published in the topical collection featuring Promising Early-

Career (Bio)Analytical Researchers with guest editors Antje
J. Baeumner, María C. Moreno-Bondi, Sabine Szunerits, and
Qiuquan Wang.

4
* Danilo Sciarrone Vienna International Centre, International Atomic Energy
dsciarrone@unime.it Agency, PO Box 100, 1400 Vienna, Austria
5
1 Chromaleont S.R.L., c/o Department of Chemical,
Department of Chemical, Biological, Pharmaceutical
Biological, Pharmaceutical and Environmental Sciences,
and Environmental Sciences, University of Messina,
University of Messina, 98168 Messina, Italy
98168 Messina, Italy
6
2 Department of Sciences and Technologies for Human
Traceability Unit, Research and Innovation Centre,
and Environment, University Campus Bio-Medico of Rome,
Fondazione Edmund Mach, San Michele All’Adige, via Mach
Rome, Italy
1, 38098 Trento, TN, Italy
3
Center Agriculture Food Environment (C3A), University
of Trento, San Michele All’Adige, Via Mach 1, 38010 Trento,
TN, Italy

13
Vol.:(0123456789)
5644 Cucinotta L. et al.

Introduction monodimensional and multidimensional GC [6–8]. Con-

versely, nowadays, key chiral components can be selected by
Cannabis is an herbal annual plant, for which research has natural sources having identical or similar enantiomeric
long been limited because of legal restrictions across the ratios with respect to the sample investigated [9], thereby
twentieth and the beginning of the twenty-first century, due reducing the possibility to detect any differences deriving
to its renowned psychotropic activity and the related illicit from adulterations. In such cases, additional analytical tech-
use. From a normative point of view, the cultivation has been niques are needed to highlight further typical traits of the
promoted in Europe only in the last decade, specifically for plant of origin. Gas chromatography coupled to isotope ratio
those varieties, e.g., Cannabis sativa L., with content in tet- mass spectrometry (GC-C-IRMS) is equally broadly recog-
rahydrocannabinol < 0.2% (https://​eur-​lex.​europa.​eu/​LexUr​ nized as a suitable technique to assess the origin and genu-
iServ/​LexUr ​iServ.​do?​uri=​OJ:L:​2013:​347:​0608:​0670:​IT:​ ineness of a sample. A GC-C-IRMS method can provide
PDF). As a consequence of the increased scientific interest, crucial information about the natural, synthetic, or biosyn-
nowadays, hemp market and the linked manufactures are thetic origin of specific compounds in a sample, allowing in
ever spreading. Among the products derived from Cannabis many cases to unveil fraudulent practices. Since the distribu-
sativa L., the essential oil obtained by the inflorescences is tion of carbon stable isotopes strictly depends on the photo-
one of the most distinctive, due to the harmonious balance synthetic carbon metabolism and the geographic origin of
between the main monoterpene and sesquiterpene compo- the plant, the evaluation of the δ13C values of key volatile
nents along with their oxygenated derivatives [1]. Cannabis compounds represents a powerful tool for genuineness
chemical composition may vary in relation to the inflores- assessment [10–12]. It is straightforward that a combined
cence being fresh or dry [1], its variety, and the extraction approach consisting of Es-GC-C-IRMS would allow for the
methodology used. In the last decades, steam distillation or simultaneous evaluation of both chiral and δ13C values, thus
hydro-distillation [2] has been the most used technique for enabling to detect fraudulent additions to genuine oil sam-
oil extraction, as well as the innovative microwave-assisted ples. The coupling of chiral separation and IRMS detection
hydro-distillation (MADH) apparatus [3, 4]. In MADH was first demonstrated by Mosandl et al. [13] for the differ-
extraction of the hemp inflorescences, the final yield can be entiation of biotechnological and synthetic γ-decalactone,
maximized by finely tuning some instrumental parameters, and later by few more authors [14–17]. Although represent-
i.e., microwave power and time program [4]. Alongside ing a powerful tool for authenticity assessment, this
genuine oils obtained from the inflorescences, also reconsti- approach suffers from important drawbacks, which have
tuted oils obtained by natural or synthetic sources are nowa- greatly limited its widespread diffusion. When dealing with
days more and more present on the market. Most of them are medium-to-high complex samples, incomplete separation of
declared to contain the typical hemp terpene compounds, the compounds of interest may occur, due to insufficient
such as α-pinene, myrcene, and (E)-β-caryophyllene, aiming chromatographic resolution. Within crowded zones across a
to emulate the flavor and other properties of natural cannabis chromatogram, co-elutions would in turn result in wrong
oils. Yet, no reference data are reported in the literature evaluation of the enantiomeric and isotopic ratios. Chiral
about the chemical composition of these oils. In this regard, compounds can be resolved into their enantiomers on a suit-
gas chromatography coupled to flame ionization detector able stationary phase, and in the presence of further co-
(GC-FID) and to mass spectrometry (GC–MS) are well- elutions, the enantiomeric ratios can be still calculated by
suited techniques for investigation of the qualitative and extracting specific m/z ions. On the other hand, this problem
quantitative composition in terms of volatile components. would negatively impact the δ13C value evaluation, due to
However, there is no chance to highlight the differences with the uneven distribution of carbon isotopes (12C and 13C)
respect to extracted genuine oils, by these means. The along the ­CO2 peak [18, 19]. A slightly different retention is
employment of more sophisticated analytical techniques is observable for the same molecules containing different
needed to tackle such a task. Enantio-selective gas chroma- amounts of the two isotopes. In fact, different van der Waals
tography (Es-GC) has historically played a key role in the dispersion forces during solute/stationary interaction cause
authenticity assessment of natural essential oils, proving to an enrichment in the heavier isotope of the initial part of the
be one of the most effective analytical approaches for this peak, while the lighter isotope will be more represented in
purpose [5]. In fact, chiral compounds show typical enantio- the final part. Thus, shifted isotopic values will be observed
meric ratios in natural samples, being the metabolites of with respect to those of a pure peak [18, 19], depending on
specific biochemical pathways of the plant of origin. Thus, which part of the peak is co-eluted. If the co-elution affects
the presence of unusual enantiomeric ratios for a given sam- the front of the peak, a more negative δ13C value will be
ple has often allowed unveiling fraudulent additions, by observed; on the contrary, a more positive δ13C value will
comparison to genuine reference values, using be measured if the tail of the peak of interest is affected by
co-elution. In the case of a complete co-elution, no MS

13
Simultaneous evaluation of the enantiomeric and carbon isotopic ratios of Cannabis sativa… 5645

“tricks” can be adopted, due to the conversion of the peak values. In order to calibrate the measured δ13C values to the
components to C ­ O2. In this case, evidence of such a situation VPDB scale, the ­CO2 reference gas was calibrated using
would be only obtained by checking the C ­ O2 ratio differen- four reference compounds, namely iodomethane (δ 13C
tiation [19]. Multidimensional gas chromatography exploited value − 54.59‰) and three alkanes from the Indiana mix
in the heart-cut mode (HC-MDGC) may represent the tech- A7: hexadecane (δ13C value: − 26.15‰), octadecane (δ13C
nique of choice for this task, thanks to the higher peak reso- value: − 32.70‰), and eicosane (δ13C value: − 40.91‰)
lution afforded by the coupling of two full-length columns (Indiana University, Bloomington, IN). Isotopic ratio meas-
with different selectivity [20]. In the literature, only few urements of the samples of interest were made by the fol-
works have employed an MDGC system before the IRMS lowing formula:
detection, and among these, only one deals with the detec-
Rsample − Rstandard
tion of enantiomers [16]. Whereas cannabinoids extracted 𝛿 13 C =
from marijuana samples have been investigated by means of Rstandard
compound-specific IRMS [21], no data are available relative where R represents the abundance ratio of the heavier
to the terpene fraction. In this research, an Es-MDGC system carbon isotope against the lighter one (13C /12C).
coupled to IRMS via a combustion chamber (C-IRMS), and All the samples were stored at + 4 °C.
to a quadrupole mass spectrometer (qMS), was developed.
An apolar column was used as the first dimension (1D) and Instrumental parameters
a chiral cyclodextrine-based stationary phase as the second
dimension (2D), to evaluate the enantiomeric and isotopic Microwave‑assisted‑hydro‑distillation (MAHD) conditions
ratios of well-separated target terpenes in cannabis oils, for
the first time. A MAHD method, earlier described by Mical- Six-hundred milliliters of ultrapure water was added to 200 g
izzi et al. [4], was adopted to obtain in-house genuine oils of inflorescence, and the biomass was uniformly mixed. The
from the fresh and dried inflorescences of different cannabis resulting mixture was placed inside a 2-L ETHOS-X glass
varieties. In parallel, commercial oils declared as reconsti- reactor and then into a Milestone “Ethos X” extractor (Mile-
tuted or natural were analyzed, and the results were com- stones, Sorisole, Italy). The extraction was carried out under
pared to those obtained for the genuine samples extracted previously optimized conditions [4], and the essential oils
in-house. To the best of our knowledge, this is the first time were collected from the distillation system.
that δ13C values are provided for the terpene fraction of Can-
nabis sativa L. essential oil samples. The data gathered from Monodimensional GC‑C‑IRMS/qMS conditions
analysis of the different sample varieties were used to estab-
lish the characteristic enantiomeric and isotopic ranges of The system consisted of a GC2010 Plus gas chromatogra-
Cannabis sativa L. essential oil. pher equipped with an AOC-20i autosampler (Shimadzu
Europa, Duisburg, Germany). The instrument was directly
connected via a zero dead-volume tee-union to a QP2010
Experimental section Ultra quadrupole mass spectrometer (Shimadzu Europa) and
to a VisION IRMS system, preceded by a GC V furnace
Sample preparation system (Elementar Analysensysteme GmbH, Langenselbold,
Germany) operated at 850 °C. A split/splitless injector was
Fifteen samples of dry hemp inflorescences belonging to maintained at 280 °C, at a split ratio 10:1. The same GC-C-
Futura 75, Kompolti, Felina 32, Tisza, and CS (Carmagnola IRMS/qMS system configuration was used exploiting two
Selezionata) varieties and one sample of fresh hemp inflo- different stationary phases, alternatively:
rescences of Futura 75, registered in the EU Plant variety
database (https://e​ c.e​ uropa.e​ u/f​ ood/p​ lant/p​ lant_p​ ropag​ ation_​ – A MEGA-DEX ASX 1 column, 25 m × 0.25 mm
mater​ial/​plant_​varie​ty_​catal​ogues_​datab​ases/​search/​pub- I.D. × 0.25 μm df (MEGA, Milano, Italy), was used as
lic/​index.​cfm?​event=​Searc​hVari​ety&​ctl_​type=​A&​speci​ chiral stationary phase, ramped from 50 to 220 °C at
es_​id=​240&​varie​ty_​name=​&​listed_​in=​0&​show_​curre​nt=​ 2 °C/min.
on&​show_​delet​ed), were provided by the Canapar group – A capillary SLB-5 ms column, 30 m × 0.25 mm
(Ragusa, Italy). Nineteen commercial hemp oil samples, I.D. × 0.25 μm df (Merck Life Science, Darmstadt, Ger-
16 reconstituted and 3 declared as naturals, were purchased many), was used as apolar stationary phase, ramped from
from local stores. All the samples were diluted 1:10 in 50 to 280 °C at 3 °C/min.
n-hexane before GC analysis. A ­C7-C30 n-alkane mix, kindly
provided by Merck Life Science (Darmstadt, Germany), was
used for the calculation of Linear Retention Index (LRI)

13
5646 Cucinotta L. et al.

The columns were operated at a constant flow rate of system by means of a GC V furnace system (Elementar
carrier gas (helium) of 1 mL/min. A pressure program was Analysensysteme GmbH) maintained at 850 °C. The split/
applied to the injector during the analyses, on the basis splitless injector was maintained at 280 °C, at a split ratio
of the total resistance (column + retention gaps). For the 10:1. A constant helium flow of 1.0 mL/min was delivered to
chiral column, the program pressure started from 74 kPa, the 1D column, an SLB-5 ms 30 m × 0.25 mm I.D. × 0.25 μm
at 1.63 kPa/min, to 126 kPa. For the apolar column, the df (Merck Life Science, Darmstadt, Germany). A pressure
pressure was ramped from 85 to 165 kPa at 1.04 kPa/min. program was used, from 185 kPa (7 min) to 247 kPa (5 min)
The column effluent was diverted to the IRMS system via at 1.89 kPa/min, to 300 kPa at 1.89 kPa/min, and finally to
a 0.85 m × 0.32 mm I.D. uncoated column, located inside a 330 kPa at 9.48 kPa/min. The GC1 oven was ramped as fol-
combustion chamber, and in parallel to the qMS system via lows: 50 °C (7 min) to 227 °C at 3 °C/min with an isotherm
a 2 m × 0.1 mm I.D. uncoated column. The qMS ion source at 150 °C (5 min), finally to 280 °C at 15 °C/min. The FID
and interface temperature was maintained at 200 °C; a mass was connected to the DS device via a 0.25 m × 0.18 mm i.d.
range 40–400 m/z was monitored at an acquisition speed of stainless steel uncoated column and used to monitor the 1D
10 Hz. GCMS data were acquired by the GCMS solution eluent. FID conditions were as follows: 330 °C; ­H2 flow,
software ver. 4 (Shimadzu Europa). Compound identification 40.0 mL/min; air flow rate, 400 mL/min; sampling rate,
was carried by using the FFNSC 4.0 mass spectral library 80 ms equal to 12.5 Hz. GC2 was equipped with a MEGA-
database (Shimadzu Europa), exploiting a double-filter DEX ASX 1 chiral column 25 m × 0.25 mm I.D. × 0.25 μm
approach based on spectral similarities and Linear Reten- df (MEGA, Milano, Italy), and the temperature was ramped
tion Index (LRI) values. The VisION IRMS was a bench- as follows: 40 °C (22 min) to 76 °C (5 min) at 2 °C/min, to
top 5-kV system equipped with an integrated gas delivery 145 °C at 3 °C/min, and finally to 195 °C at 8 °C/min. The
2
monitoring system. The combustion chamber was equipped D column was connected at one side to the DS device and
with a high-performing silicon carbide tube furnace for the at the other side to a zero dead-volume tee-union (Valco).
quantitative, fractionation-free conversion of the delivered A pressure program was applied to the APC device in order
compounds to pure gases (­ CO2 and ­H2O). The C ­ O2 pro- to maintain a constant carrier flow also in the 2D column
duced by the combustion of each compound was transferred (≈1 mL/min): 140 kPa (22 min) to 165 kPa (5 min), to a
to the IRMS, while the H ­ 2O produced was removed through final pressure of 210 kPa at 1.39 kPa/min. An auxiliary He
a nafion membrane. The system was designed with reduced line (sample line He), automatically controlled through a
dead volumes to maintain the chromatographic resolution second channel of the APC unit, was used in the furnace to
at the IRMS. The following settings were applied to the allow a proper control over the open split conditions for the
VisION system: acceleration voltage, 3795 V; trap current, IRMS. Also in this case, the APC was operated in constant
600 mA; magnet current, 3700 mA. An electron-impact flow mode, to maintain the open split in a steady state. MS
ionization (EI) gas source, a variable field, stigmatically and IRMS conditions were the same used in the monodimen-
focused electromagnet for beam separation and multi-chan- sional approach. All the MDGC analyses were carried out
nel Faraday collectors for beam detection were used. IRMS at least in triplicate and the standard deviations for IRMS
data were collected by IonOS stable isotope data processing measurements were found to be < 0.5 ‰.
software ver. 4.5 (Elementar Analysensysteme GmbH). The
apex track integration method was exploited to automatically
find the correct starting and finishing points of the peaks. Results and discussion

MDGC‑C‑IRMS/qMS conditions Extraction of the cannabis oil samples by MAHD

The MDGC-C-IRMS/qMS prototype consisted of an AOC- Extraction by MAHD was applied to sixteen genuine sam-
20i autosampler and two GC-2010 Plus gas chromatog- ples consisting of fresh and dry hemp inflorescences. By
raphers (defined as GC1 and GC2), connected by means the employment of a microwave-assisted system, heat was
of a heated transfer line (Shimadzu Europa). GC1 was applied to the soaked biomass to facilitate the break of the
equipped with a split/splitless injector, a flame ionization oileferous glands and the release of essential oil. By this
detector (FID), and a Deans-switch (DS) transfer device. method, the first drops of essential oil fell after about 5 min
GC1 was connected to an advanced pressure control unit of distillation at 1200 W, at a temperature around 94 °C. The
(APC), which supplied the same carrier gas (He) (Shimadzu samples selected for extraction are listed in Table 1. One
Europa) allowing to divert the first column eluent to the sample was obtained by fresh hemps of Futura 75 (sample
FID or to the second column in the GC2. The latter was 1), while the dry samples belonged to different varieties: 9 of
hyphenated in parallel to a QP2010 Ultra quadrupole mass Futura 75 (samples 2–10), 2 of Felina 32 (samples 11–12), 1
spectrometer (Shimadzu Europa) and to a VisION IRMS of Tisza (sample 13), 2 of Kompolti (samples 14–15), and 1

13
Simultaneous evaluation of the enantiomeric and carbon isotopic ratios of Cannabis sativa… 5647

Table 1  Cannabis varieties selected for essential oil extraction from results, on the differences among the distinct Cannabis
hemp inflorescences sativa varieties. Nonetheless, the data gathered from analy-
ID Varieties Fresh/dry Origin sis of the essential oil samples were used to establish the
characteristic ranges of the enantiomeric and isotopic ratios
1 Futura 75 Fresh Italy
of Cannabis sativa L.
2 Futura 75 Dry Italy
3 Futura 75 Dry Italy
Monodimensional Es‑GC‑C‑IRMS/qMS analysis
4 Futura 75 Dry Italy
5 Futura 75 Dry Italy
Es-GC-C-IRMS/qMS analyses were initially carried out to
6 Futura 75 Dry Italy
determine both the enantiomeric and isotopic ratios of the
7 Futura 75 Dry Italy
genuine cannabis oils. To this purpose, a chiral MEGA-DEX
8 Futura 75 Dry Italy
ASX-1 stationary phase proved to be effective for the separa-
9 Futura 75 Dry Italy
tion of the target enantiomers before the IRMS determina-
10 Futura 75 Dry Croatia
tion. Still a number of issues arose in the monodimensional
11 Felina 32 Dry Italy
GC analysis of some genuine cannabis oils, for the separa-
12 Felina 32 Dry Italy
tion of the main chiral terpenes. This is shown in Fig. 1A for
13 Tisza Dry Italy
a Futura 75 essential oil. As can be easily noticed, although
14 Kompolti Dry Italy
α-pinene and β-pinene enantiomers were efficiently sepa-
15 Kompolti Dry Italy
rated by Es-GC-C-IRMS/qMS, co-elutions occurred with
16 CS (Carmagnola Dry Italy
other achiral components. In the case of ( +)-α-pinene,
selezionata)
present in higher amount with respect to heptanal, the par-
tial peak overlapping might have a limited influence on the
peak area evaluation. Thus, it would not affect the enantio-
of CS (Carmagnola Selezionata) (sample 16). Unfortunately, meric ratio assessment and the δ13C value evaluation [19].
the high cost and the difficulties in retrieving genuine plant Otherwise, the co-elution occurring between ( +)-β-pinene
material have limited the number of oil samples included and another major compound (e.g., myrcene) introduced a
in this research. This represents a common issue for stud- critical issue. The incomplete separation in fact hindered
ies focused on ascertaining the genuineness ranges, since accurate evaluation of both the enantiomeric and isotopic
natural samples are required to assess the characteristic δ13C ratios. Concerning the enantiomeric ratio evaluation, the use
values and enantiomeric ratios. The limited size of samples of simultaneous qMS detection would in principle allow to
available prevented from obtaining statistically significant resolve the chromatographic co-elution by monitoring the

Fig. 1  Es-GC-C-IRMS/qMS analysis showing the co-elutions of ( +)-α-pinene and ( +)-β-pinene peaks (A) and (-)-limonene peak (B) in a
Futura 75 essential oil, and the enantiomers of selina diene isomers in a Kompolti essential oil (C)

13
5648 Cucinotta L. et al.

extracted m/z ions of the compounds of interest. Unfortu- Monodimensional GC‑C‑IRMS/qMS analysis
nately, the EI spectra of monoterpene compounds will be
characterized by the same fragmentation pattern and show Inaccurate results were also obtained by GC-C-IRMS/qMS
nearly identical fragment ions, thus precluding the use of analysis performed on the apolar column. As can be seen in
this approach. From the IRMS standpoint, a similar approach Fig. 2, co-elutions affected the target oil compounds elut-
cannot be envisaged. Actually, all the organic matter must ing in the monoterpene and sesquiterpene zone. Figure 2A
be converted to C­ O2 before the δ13C value determination, shows a partial co-elution involving limonene and eucalyptol
with a consequent loss of any analyte structural informa- in a Futura 75 sample. Due to their similar LRIs on an apo-
tion [18]. As a consequence, the occurrence of a co-elution lar stationary phase (1030 and 1032 respectively), a partial
involving the right part (tail) of the peak, as for the case of overlap occurred, again generating unreliable results for the
( +)-β-pinene, is expected to cause a relevant positive shift δ13C values.
of the δ13C value. The latter is due to the depletion of the Figure 2B shows the critical case of selina-4(15),7(11)-
44
CO2 fragment, related to a lower amount of β-pinene mol- diene, identified by exploiting double-filter search based
ecules containing 12C. Figure 1B shows a further co-elu- on mass spectral similarity (> 90%) and LRI (± 5 units
tion occurred in a Futura 75 oil, involving the (-)-limonene window). Due to its Gaussian shape, it would have been
enantiomer and (E)-β-ocimene. Being the latter a higher regarded as a pure peak, at first sight. Yet, a deeper insight
abundant compound, a wrong estimation of the δ13C value into the MS spectrum along the peak highlighted the pres-
should be expected, unlike the case of ( +)-α-pinene dis- ence of co-eluted (E)-α-bisabolene in the peak tail. Actu-
cussed above. Moving forward in the chromatogram, further ally, these components are characterized by the same LRI
critical cases may be observed in the sesquiterpene elution value of 1540 on the column employed, as reported in the
zone of a Kompolti sample (Fig. 1C). Selina-3,7(11)-diene FFNSC 4.0 database. Figure 2C shows how this co-elution
and selina-4(15),7(11)-diene compounds were successfully can be resolved by monitoring the extracted ions from the
separated into their enantiomers; nevertheless, the late elut- MS total ion current (TIC) (lower trace) and by 45/44CO2
ing enantiomer of selina-4(15),7(11)-diene and the first elut- ratio differentiation at the IRMS (upper trace). In detail,
ing enantiomer of selina-3,7(11)-diene overlap with germa- m/z 93 and m/z 105 were extracted from the TIC, being the
crene B peak. Likewise, unreliable results may be predicted base peaks for (E)-α-bisabolene and for selina-4(15),7(11)-
in this case, notably a more positive δ13C value for selina- diene, respectively. The higher amount of the m/z 93 ion in
4(15),7(11)-diene and a more negative δ13C value for selina- the tail of the peak clearly suggested the presence of (E)-α-
3,7(11)-diene, with respect to the true values. bisabolene, as further confirmed by the unusual 45/44CO2

Fig. 2  GC-C-IRMS/qMS analysis showing the co-elutions of differentiation (IRMS: upper trace) and extracted ion chromatogram
limonene and eucalyptol peaks (A) and selina-4(15), 7(11)-diene, and (qMS: lower traces) of selina-4(15),7(11)-diene peak (C)
(E)-α-bisabolene peaks (B) in a Futura 75 essential oil. 45/44CO2 ratio

13
Simultaneous evaluation of the enantiomeric and carbon isotopic ratios of Cannabis sativa… 5649

ratio differentiation reported in the upper profile. On this column and a chiral MEGA-DEX ASX-1 as 2D. The first
basis, the need to implement a multidimensional approach step of any HC-MDGC analysis involves the identification
emerged clearly, to increase the separation capability. of the peaks of interest after a 1D analysis performed in the
stand-by mode (no heart-cuts selected). Since no qualita-
Es‑MDGC‑C‑IRMS/qMS approach tive information was available after the 1D detector (FID),
the identity of each peak of interest was confirmed on the
System configuration basis of the elution order in monodimensional GC–MS on
the same stationary phase. The experimental LRIs of the
Different MDGC approaches are described in the literature peaks calculated for a 1D stand-by analysis were matched
exploiting both single- and double-oven configurations. to the theoretical values, allowing for a ± 10 units toler-
Apart from the economical aspect, the technical advantage ance (see Electronic Supplementary Material, Fig. S3 and
of a double-oven configuration relies on the capability for Table S1) [22]. Cannabis oil samples and a C ­ 7–C30 alkane
separate optimization of column temperature programs. homologous series were analyzed under the same conditions
Such configuration requires the coupling of a GC–MS to a and the experimental LRIs were calculated by the following
GC-FID system via a heated transfer line. Since in GC–MS formula:
the spectrometer is located on the left side of the GC, the ( )
transfer line connecting to the (second) GC-FID instrument 100n + 100 tR x − tR n
LRIx = (1)
is usually located on the right side of the instrument (see
( )
tR n + 1 − tR n
Electronic Supplementary Material, Fig. S1). In such con-
figuration, the GC-FID system is regarded as the first chro- where n and tR n are the carbon number and retention time
matographic dimension (1D) and thus the GC–MS system of the alkane eluted before the peak of interest, and tR n + 1
as the second one (2D). Hereby, given the need for splitting is the retention time of the alkane eluted after the peak of
the second column eluate between the qMS and IRMS sys- interest.
tems, an inverted configuration was chosen (see Electronic
Supplementary Material, Fig. S2). Hence, the primary col- Es‑MDGC‑C‑IRMS/qMS analysis
umn (1D) and the DS device were located by the GC–MS
side while the secondary column, connected to the DS, was The oil samples were separated by MDGC prior to meas-
passed through the heated transfer line and located by the urement of the δ13C values and enantiomeric ratios of the
GC-FID system. The latter was provided with a T-union peaks of interest. Figure 3 shows the 1D FID chromatogram
located in the oven, connected at one side to the (2D) outlet of a cannabis essential oil acquired in the stand-by mode
and at the other side to two retention gaps. The latter were on the apolar (5%) column (black trace). Among the main
different in length and I.D. and connected to the two spec- chiral terpenes, a total of nine compounds were selected
trometers. The configuration here described was chosen on for the heart-cut: α-pinene, β-pinene, limonene, linalool,
the basis of the required split ratio of about 1:10 to the qMS α-terpineol, (E)-β-caryophyllene, selina-4(15),7(11)-diene
and IRMS, respectively. The low natural abundance of the and selina-3,7(11)-diene, and β-caryophyllene oxide. Minor
13
C isotope (about 1% of the 12C isotope) actually leads to chiral components, namely camphene, borneol, fenchyl alco-
a proportional lower sensitivity of IRMS detection. Thus, a hol, and (E)-nerolidol, were not taken into consideration
higher resistance was applied to the qMS side to compensate due to their low amount in the oils. As shown in Table 2,
for this difference, obtained via a 2 m × 0.1 mm I.D. reten- the experimental LRI values ­(LRIexp), calculated on the 1D
tion gap. At the IRMS side, a 0.8 m × 0.32 mm I.D. reten- apolar column, ranged within ± 5 units with respect to the
tion gap was used. This corresponded to the shortest reten- theoretical data ­(LRItheor).
tion length needed for the desired split ratio, with minimum A heart-cut window was then selected, for each com-
dead volumes to preserve the chromatographic resolution. pound to be transferred to the second chiral column in the
Finally, such inverted configuration allowed for positioning “cut” mode (pink trace in Fig. 3).
the T-union as close as possible to the IRMS system. The combination of two chromatographic separation
mechanisms provided substantial benefits over the monodi-
Selection of the peaks of interest after 1D stand‑by analysis mensional approach. First, a satisfactory separation was
achieved for all the enantiomeric compounds transferred
An MDGC method was implemented as front-end separa- from 1D. Furthermore, all the co-elutions occurring with
tion, before simultaneous qMS and IRMS detection. This, other (achiral) sample components were prevented. The gain
in order to overcome the separation issues discussed in the in separation attained by MDGC can be appreciated in Fig. 4,
“Monodimensional GC-C-IRMS/qMS analysis” section. In showing different zoomed zones of the chromatogram. In
detail, an apolar SLB-5 ms column was employed as 1D detail, α-pinene, β-pinene, and limonene enantiomers show

13
5650 Cucinotta L. et al.

Fig. 3  1D stand-by (black

chromatogram) and cut (pink
chromatogram) FID analysis of
a Cannabis sativa L. essential
oil. Cut windows: I: α-pinene,
II: β-pinene, III: limonene,
IV: linalool, V: α-terpineol,
VI: (E)-β-caryophyllene, VII:
selina-4(15),7(11)-diene and
selina-3,7(11)-diene, VIII:
β-caryophyllene oxide

Table 2  The theoretical and experimental LRI values calculated for up as separated peaks in Fig. 4A, whereas in the correspond-
a 1D stand-by analysis of a Kompolti oil (apolar 1D column). ­LRItheor ing Es-GC-C-IRMS/qMS analysis important co-elutions
are reported in the FFNSC 4.0 mass spectral database for the same
occurred with heptanal, myrcene, (E)-β-ocimene, and euca-
stationary phase (SLB-5 ms)
lyptol. Figure 4B shows the separation obtained for linalool
Peak ID Compound LRItheor LRIexp and α-terpineol enantiomers. Figure 4C shows the (E)-β-
1 α-Pinene 933 935 caryophyllene peak separated from γ-elemene. Noticeably,
2 β-Pinene 978 980 these compounds fully co-eluted on the apolar 1D column
3 Limonene 1030 1032 due to the very close LRI values (1424 vs 1432) and to the
4 Linalool 1101 1103 high amount of caryophyllene in the sample. Moreover, the
5 α-Terpineol 1195 1200 separation of the four enantiomers of selina diene isomers
6 (E)-β-Caryophyllene 1424 1427 is evident in the same chromatogram, free from overlap-
7 Selina-4(15),7(11)-Diene 1540 1544 ping with germacrene B and (E)-α-bisabolene peaks. Sepa-
8 Selina-3,7(11)-diene 1546 1550 ration on a single chiral column had resulted in co-elution of
9 β-Caryophyllene oxide 1587 1591 selina-4(15),7(11)-diene with germacrene B, as illustrated
in Fig. 1. Likewise, monodimensional analysis on the apolar

Fig. 4  MDGC separation of the target monoterpenes in a Futura 75 (A) and a Felina 32 (B) essential oil sample and of the target sesquiterpenes
in a Futura 75 essential oil sample (C)

13
Simultaneous evaluation of the enantiomeric and carbon isotopic ratios of Cannabis sativa… 5651

column has led to the co-elution with (E)-α-bisabolene, as

illustrated in Fig. 2.
After MDGC separation, only a minor co-elution was

− 24.3
− 24.6
− 25.9
− 30.9
− 24.9
− 28.6
− 27.7
− 27.2
− 24.3
δ13C
observed between the second eluting selina-4(15),7(11)-

nd
nd
nd
nd
nd
nd
nd
nd
nd
diene peak and the first eluting peak of selina-3,7(11)-diene

6 Futura 75
enantiomers. Unfortunately, no data are available about the

ER %
elution order of these enantiomers on the chiral stationary

91.7
39.8
60.2
40.6
59.4
17.0
83.0
13.0
37.0
63.0
96.0
87.0

100

8.3
4.0

nd
nd
nd
phase used. Finally, the separation of (E)-β-caryophyllene
oxide enantiomers could be appreciated only in the dried

− 23.6
− 25.0
− 25.1
− 24.3
− 29.0
− 27.0
− 27.7
− 24.3
hemp inflorescence samples. This compound is in fact gen-

δ C
13

nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
erated as a result of the oxidation occurring upon drying
of the plant material [4]. The enantiomeric ratios and δ13C

5 Futura 75
values obtained within a single analysis for all the sample

ER %

93.0
38.9
61.1
40.4
59.6
13.0
42.5
57.5
96.3
87.0

100
components are reported in Table 3.

7.0
3.7

nd
nd
nd
nd
nd
Differences were found in the qualitative and quantitative
composition in monoterpenes and sesquiterpene, accord-

− 24.1
− 24.4
− 24.7
− 24.4
− 28.6
− 25.3
− 26.7
− 27.3
ing to the type of inflorescence and the variety itself [1,

δ C
13

nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
4]. Moreover, characteristic enantiomeric ratios and δ13C

4 Futura 75
values were observed for the target compounds in all the
sixteen genuine oils. The enantiomeric ratios of all the natu-

ER %

96.7
40.6
59.4
41.8
58.2
14.2
28.7
71.3
96.5
85.8

100

3.3
3.5

nd
nd
nd
nd
nd
ral samples were in good agreement with the literature data
[3, 4], although in this research a higher number of samples
were investigated. A marked preponderance was observed

− 24.4
− 24.8
− 25.1
− 24.8
− 29.1
− 25.7
− 28.8
− 28.1
δ C
in all the hemp varieties of ( +)-α-pinene (93.4–97.1%),
13

nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
( +)-β-pinene (78.6–88.0%), and (-)-β-caryophyllene oxide
3 Futura 75

(91.7–98.9%). Similarly but to a lesser extent, ( +)-lin-

ER %

alool was predominant in all the varieties investigated

95.9
43.4
56.6
41.4
58.6
12.1
54.1
45.9
95.6
87.9

100

4.1
4.0

nd
nd
nd
nd
nd
(57.8–74.9%). In contrast, a similar trend could not be
Table 3  Enantiomeric ratios (ER %) and δ13C values data of genuine Cannabis sativa L. oils

identified for limonene and α-terpineol enantiomers, whose

enantiomeric ratios greatly varied in the samples. A racemic

− 26.4
− 25.9
− 27.0
− 25.7
− 30.3
− 30.2
− 28.4
− 29.8
− 29.6
− 29.3
δ C
13

behavior was observed for selina diene isomers, with only a

nd
nd
nd
nd
nd
nd
nd
nd

slight predominance of the late eluted enantiomers for both

2 Futura 75

the 4(15) and 3(11) isomers. In this case, however, the elu-
ER %

97.9
40.4
59.6
39.5
60.5
41.7
58.3
29.8
70.2
38.1
61.9
13.9
97.1
86.1

tion order of the two enantiomers is unknown. Concerning

100

2.1
2.9

nd

the (E)-β-caryophyllene enantiomers, the occurrence of the

levorotatory form at nearly 100% is widely reported, and the
− 28.6
− 27.3
− 27.0
− 30.2
− 25.4
− 29.5
− 30.7
− 30.1
− 28.6
− 28.4
− 28.9
− 28.8

data obtained hereby were in accordance with this literature

δ C
13

nd
nd
nd
nd
nd
nd

finding [3, 4]. Dealing with isotopic distribution, differences

were observed in terms of δ13C values among the specific
1 Futura 75

varieties, as reported in Table 3. The δ13C values of pure

ER %

97.1
40.5
59.5
41.4
58.6
68.7
30.6
69.4
31.3
14.6
58.7
41.3
97.1
85.4

100

enantiomer peaks could be determined, after the 2D separa-

2.9
2.9

nd

tion obtained on the chiral stationary phase. However, for

low concentrated components and depending on the enan-
(-/ +)
(-/ +)
(-/ +)
(-/ +)

( +)
( +)
( +)
( +)
( +)
( +)
( +)

(-)
(-)
(-)
(-)
(-)
(-)
(-)

tiomeric ratio, in few cases, only one enantiomer could be

measured. A linearity test automatically performed by the
system showed consistent δ13C values in the response inter-
val above 0.5 nA; thus, signals below this threshold were
discarded. According to the few data available in the litera-
Selina-4(15),7(11)-diene

ture, no significant δ13C value differences are to be expected

β-Caryophyllene oxide
(E)-β-Caryophyllene

Selina-3,7(11)-diene

within enantiomers of the same molecule [13–17]. In gen-

eral, more negative δ13C values were found for samples
α-Terpineol
Limonene

13–16 (Tisza, Kompolti, and Carmagnola Selezionata) with

α-Pinene

β-Pinene

Linalool

respect to samples 1–10 (Futura 75) and 11–12 (Felina 32).

13
 Table 3  (continued)
5652

7 Futura 75 8 Futura 75 9 Futura 75 10 Futura 75 11 Felina 32 12 Felina 32

ER % δ13C ER % δ13C ER % δ13C ER % δ13C ER % δ13C ER % δ13C

13
α-Pinene (-) 3.4 − 26.1 3.7 − 27.8 3.4 − 30.5 3.9 − 32.5 4.0 nd 3.9 − 26.4
( +) 96.6 − 25.9 96.3 − 27.3 96.6 − 29.9 96.1 − 32.3 96.0 − 30.8 96.1 − 26.4
β-Pinene ( +) 87.6 − 23.0 86.7 − 24.9 88.0 − 27.5 83.3 − 29.7 86.5 − 28.8 87.0 − 23.1
(-) 12.4 nd 13.3 nd 12.0 nd 16.7 nd 13.5 nd 13.0 nd
Limonene (-) 29.8 nd 41.0 − 28.0 36.8 − 30.2 59.5 nd 64.4 − 31.1 39.9 nd
( +) 70.2 − 27.5 59.0 − 28.5 63.2 − 30.8 40.5 − 33.8 35.6 − 31.3 60.1 − 27.8
Linalool (-) nd nd nd nd nd nd nd nd 42.2 − 30.3 nd nd
( +) nd nd nd nd nd nd nd nd 57.8 − 30.6 nd nd
α-Terpineol (-) nd nd nd nd nd nd nd nd 50.8 nd nd nd
( +) nd nd nd nd nd nd nd nd 49.2 nd nd nd
(E)-β-Caryophyllene (-) 100 − 24.1 100 − 23.9 100 − 26.4 100 − 28.0 100 − 27.7 100 − 24.2
( +) nd nd nd nd nd nd nd nd nd nd nd nd
Selina-4(15),7(11)-diene (-/ +) 38.8 − 24.5 39.5 − 24.6 42.8 − 27.5 43.7 − 29.3 41.4 − 27.8 40.5 − 23.8
(-/ +) 61.2 nd 60.5 nd 57.2 nd 56.3 nd 58.6 nd 59.5 nd
Selina-3,7(11)-diene (-/ +) 41.8 nd 40.6 nd 44.0 nd 43.88 nd 36.6 nd 41.3 nd
(-/ +) 58.2 − 23.0 59.4 − 24.7 56.0 − 27.2 56.12 − 29.4 63.4 − 28.9 58.7 − 24.2
β-Caryophyllene oxide (-) 92.7 − 23.2 93.6 − 25.0 97.4 − 27.7 95.01 − 28.9 96.2 − 28.9 93.9 − 23.5
( +) 7.3 nd 6.4 nd 2.6 nd 4.99 nd 3.8 nd 6.1 nd

13 Tisza 14 Kompolti 15 Kompolti 16 Carmagnola

ER % δ13C ER % δ13C ER % δ13C ER % δ13C
α-Pinene (-) 5.5 − 33.8 4.6 nd 2.9 nd 6.6 nd
( +) 94.5 − 32.9 95.4 − 30.8 97.1 − 32.5 93.4 − 32.2

β-Pinene ( +) 83.4 − 30.5 83.6 − 29.3 86.2 − 30.3 78.6 − 30.2

(-) 16.6 nd 16.4 nd 13.8 nd 21.4 nd
Limonene (-) 70.0 − 32.5 75.3 − 30.8 63.1 nd 79.0 − 33.6
( +) 30.0 − 32.9 24.7 − 30.2 36.9 nd 21.0 − 33.2
Linalool (-) nd nd nd nd nd nd 25.1 − 32.1
( +) nd nd nd nd nd nd 74.9 − 31.9
α-Terpineol (-) 53.1 nd nd nd nd nd 55.5 − 36.0
( +) 46.9 nd nd nd nd nd 44.5 nd
(E)-β-Caryophyllene (-) 100 − 28.9 100 − 26.8 100 − 27.1 100 − 27.9
( +) nd nd nd nd nd nd nd nd
Selina-4(15),7(11)-diene (-/ +) 44.8 − 29.7 40.8 − 29.1 41.0 − 28.9 42.8 − 28.7
(-/ +) 55.2 nd 59.2 nd 59.0 nd 57.2 nd
Selina-3,7(11)-diene (-/ +) 44.7 nd 35.1 nd 37.2 nd 40.8 nd
(-/ +) 55.3 − 31.2 64.9 − 29.4 62.8 − 30.3 59.2 − 29.0
β-Caryophyllene oxide (-) 98.9 − 29.7 96.8 − 29.6 97.1 − 29.5 98.5 − 29.9
( +) 1.1 nd 3.2 nd 2.9 nd 1.5 nd
Cucinotta L. et al.
Simultaneous evaluation of the enantiomeric and carbon isotopic ratios of Cannabis sativa… 5653

δ13C values, according to Mosandl et al. [13]. Conversely,

a more positive δ13C value was measured for the second
eluted enantiomer with respect to the first eluted enantiomer
of selina diene 4(15),7(11), viz., − 27.2‰ vs − 29.3‰. The
same discrepancy, but with an opposite trend, was observed
for the first eluted selina diene 3,7(11) enantiomer, with a
more negative δ13C value with respect to the second eluted
enantiomer, viz., − 32.0‰ vs − 29.4‰. These findings sug-
gest the occurrence of isotopic fractionation (as discussed
earlier), with opposite trends for co-elutions affecting the
initial (front) or final (tail) part of a peak. For this reason,
the δ13C values of these two compounds are not reported
[10, 18, 19].

Analysis of commercial samples

Nineteen commercial cannabis essential oils purchased from

local stores were analyzed. Among these, three oils were
Fig. 5  (A) Comparison between the δ13C values range of Italian declared as genuine, while the remaining were labelled as
Futura 75 (samples 1–9) and Croatian Futura 75 (sample 10). (B) reconstituted. The enantiomeric and δ13C values determined
Results obtained by the use of (-)-(E)-β-caryophyllene reference com- for these samples are shown in Table 4. Among the three
pound (i-ISTD)
commercial oils labelled as genuine, all the samples showed
chiral and isotopic behavior similar to those of a genuine
sample.
Among Futura 75 samples, different values were observed Most chiral and isotopic ratios determined for the recon-
between oils of different origin, namely from Italy (sam- stituted oil samples were outside the ranges characteristic of
ples 1–9) and Croatia (sample 10), as can be appreciated in the genuine samples. In some oils, the terpene fraction was
Fig. 5A. Specifically, sample 10 showed more negative δ13C represented mainly by α-pinene and (E)-β-caryophyllene, as
values with respect to samples 1–9. Given the genuineness in the case of samples 1–5, while other terpenes were found
of the Croatian oil, one of the terpene components was used in very low concentrations or were absent. The enantiomeric
as reference, to avoid the isotopic discrimination due to the ratios in all the reconstituted samples fell outside the range
geographical origin [15]. (-)-(E)-β-Caryophyllene was cho- of genuineness: (-)-α-pinene enantiomer was > 70% in sam-
sen as isotopic internal standard (i-ISTD), being one of the ples 1–8, and close to racemic values for samples 9–16, i.e.,
characteristic compound in cannabis essential oil and bio- from 39.8 to 47.7% (while in the genuine samples the range
genetically related to the other compounds investigated. The was from 2.9 to 6.6%). The same inverted ratio was observed
i-ISTD approach allowed us to ascertain the genuineness of for β-pinene enantiomers, with values for the dextrorotatory
all the Futura samples, regardless of their origin (Fig. 5B). enantiomers ranging from 2.0 to 11.6%, greatly below those
In the case of the two selina diene isomers, the small of the genuine samples (78.6 to 88.0%). Similarly, (-)-lin-
co-elution occurring between the second eluted enantiomer alool enantiomer was found at higher amount (> 66.0%)
of the 4(15),7(11) isomer and the first eluted enantiomer in all the commercial samples with respect to the genuine
of the 3,7(11) isomer was negligible to the aim of enantio- ones (< 42.2%). The enantiomeric ratio of (-)-limonene was
meric ratio evaluation. Yet, such a small lack of separation similar for samples 1–5 (> 80.0%), while in the other cases
significantly affected the evaluation of δ13C values in all the very low ratios were observed (< 3.3%). Those values were
samples analyzed. A representative example is illustrated in all cases outside the characteristic ranges of the genuine
by the δ13C values measured for Futura 75 oil (sample samples. Some of the reconstituted samples did not present
10). Among the four peaks corresponding to the distinct α-terpineol, apart for samples 6, 10–14. and 16 where it was
enantiomers of selina diene isomers, baseline separation present almost in a racemic form. Selina-4(15),7(11) and
was achieved only for the first (4(15),7(11)) and the fourth selina-3,7(11) diene were not detected in all the reconstituted
(3,7(11)) peaks. Thus, it could be assumed that accurate δ13C samples, while the enantiomeric ratio of β-caryophyllene
values were determined for these two compounds. On the oxide was in agreement with the genuine oil range. Finally,
other hand, being the second and third peaks co-eluted to (E)-β-caryophyllene was determined only in the levoro-
a very small extent, no significant δ13C value differences tatory form. Concerning the δ13C values, the values for
should be expected with respect to their relative enantiomer most components agreed with the isotopic data obtained

13
 Table 4  Enantiomeric ratios (ER%) and δ13C values obtained for the commercial Cannabis sativa L. oils, and their origin as declared on the label
5654

Reconstituted 1 Reconstituted 2 Reconstituted 3 Reconstituted 4 Reconstituted 5 Reconstituted 6 Reconstituted 7

13 13 13 13 13 13
ER % δ C ER % δ C ER % δ C ER % δ C ER % δ C ER % δ C ER % δ13C

13
α-Pinene (-) 82.3 − 30.4 82.4 − 30.5 80.9 − 29.0 82.8 nd 81.3 − 28.6 71.3 − 29.9 75.9 − 28.4
( +) 17.7 − 31.5 17.6 − 31.5 19.1 − 30.3 17.2 nd 18.7 − 30.0 28.7 − 31.7 24.1 − 30.0
β-Pinene ( +) nd nd nd nd nd nd nd nd 11.6 nd 2.8 nd 2.3 − 31.3
(-) nd nd nd nd nd nd nd nd 88.4 18.6 97.2 − 32.0 97.7 − 30.5
Limonene (-) 91.1 − 32.1 95.1 − 29.5 80.8 − 31.1 nd nd 90.1 − 31.0 2.8 nd 1.6 nd
( +) 8.9 nd 4.9 nd 19.2 nd nd nd 9.9 nd 97.2 − 30.7 98.4 − 29.3
Linalool (-) 98.2 − 26.9 98.1 − 24.3 nd nd 95.5 nd 97.0 − 25.2 74.8 − 30.8 71.1 − 31.0
( +) 1.8 nd 1.9 nd nd nd 4.5 nd 3.0 nd 25.2 nd 28.9 − 29.7
α-Terpineol (-) nd nd nd nd nd nd nd nd nd nd 50.0 − 29.7 nd nd
( +) nd nd nd nd nd nd nd nd nd nd 50.0 − 30.2 nd nd
(E)-β-Caryophyllene (-) 100 − 35.8 100 − 33.0 100 − 32.9 100 nd 100 − 32.3 100 − 33.4 100 − 35.4
( +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
Selina-4(15),7(11)-diene (-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
(-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
Selina-3,7(11)-diene (-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
(-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
β-Caryophyllene oxide (-) 100 − 39.3 100 − 36.6 100 − 36.8 nd nd nd nd 98.5 − 37.0 98.0 − 39.4
( +) nd nd nd nd nd nd nd nd nd nd 1.5 nd 2.0 nd

Reconstituted 8 Reconstituted 9 Reconstituted 10 Reconstituted 11 Reconstituted 12 Reconstituted 13 Reconstituted 14

ER % δ13C ER % δ13C ER % δ13C ER % δ13C ER % δ13C ER % δ13C ER % δ13C
α-Pinene (-) 74.6 − 28.9 39.8 − 28.0 47.7 − 29.1 45.1 − 28.9 41.3 − 29.4 46.7 − 29.6 43.2 − 29.1
( +) 25.4 − 30.3 60.2 − 28.3 52.3 − 28.9 54.9 − 28.7 58.7 − 28.8 53.3 − 29.6 56.8 − 28.7
β-Pinene ( +) 2.3 − 31.9 2.5 nd 2.8 nd 2.6 − 30.3 2.0 − 29.8 2.5 nd 2.4 nd
(-) 97.7 − 31.2 97.5 − 31.4 97.2 − 30.9 97.4 − 29.9 98.0 − 30.5 97.5 − 30.6 97.6 − 30.4
Limonene (-) 1.7 nd 3.3 − 28.3 2.0 nd 0.9 − 27.2 1.3 nd 1.2 nd 1.3 nd
( +) 98.3 − 28.7 96.7 − 27.8 98.0 − 28.7 99.1 − 29.5 98.7 − 29.0 98.8 − 28.1 98.7 − 28.5
Linalool (-) 68.6 − 31.5 73.3 − 30.9 69.8 − 31.3 69.9 − 30.4 71.6 − 31.3 nd nd 73.5 − 31.4
( +) 31.4 − 29.9 26.7 − 29.2 30.2 − 29.8 30.1 − 28.7 28.4 − 29.3 nd nd 26.5 − 29.8
α-Terpineol (-) nd nd nd nd 46.7 − 30.3 49.3 − 29.7 47.9 − 30.2 48.8 − 29.7 43.2 − 30.8
( +) nd nd nd nd 53.3 − 30.1 50.7 − 30.5 52.1 − 30.6 51.2 − 30.0 56.8 − 30.2
(E)-β-Caryophyllene (-) 100 − 34.5 100 − 33.3 100 − 33.6 100 − 32.1 100 − 33.6 100 − 34.1 100 − 34.2
( +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
Selina-4(15),7(11)-diene (-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
(-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
Selina-3,7(11)-diene (-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
(-/ +) nd nd nd nd nd nd nd nd nd nd nd nd nd nd
β-Caryophyllene oxide (-) 98.3 − 38.0 98.0 nd 98.5 − 38.1 98.8 − 37.1 97.9 − 38.0 98.4 − 37.9 97.8 − 38.1
( +) 1.7 nd 2.0 nd 1.5 nd 1.2 nd 2.1 nd 1.6 nd 2.2 nd
Cucinotta L. et al.
 Simultaneous evaluation of the enantiomeric and carbon isotopic ratios of Cannabis sativa… 5655

for the genuine samples, apart for (E)-β-caryophyllene

and β-caryophyllene oxide, which were both more nega-
− 31.1
− 30.1

− 29.6
− 30.2
− 30.5
− 32.1

− 27.5

− 28.5

− 28.3
− 28.5
tive. The latter evidence suggests the possible addition of
δ13C

nd
nd
nd
nd

nd

nd
nd

nd
exogenous (E)-β-caryophyllene to the commercial samples,
as one of the major oil terpenes. Such an hypothesis also
Natural 3

gives a rationale for the more negative values determined for

ER %

97.2

82.0
18.0
42.8
57.2

39.4
60.6
40.5
59.5
98.4
100
2.8

1.6
β-caryophyllene oxide, being it produced upon oxidation of

nd
nd
nd
nd

nd
(E)-β-caryophyllene.
− 32.1
− 30.7

− 30.3
− 30.9
− 31.6
− 32.3

− 28.1

− 29.1

− 28.8
− 28.9
δ13C

nd
nd
nd
nd

nd

nd
nd

nd
Conclusions
Natural 2
ER %

In this research, the enantiomeric ratios and δ13C values of

95.0

80.5
19.5
51.3
48.7

40.5
59.5
42.5
57.5
98.6
100
5.0

1.4
nd
nd
nd
nd

nd
main terpene constituents of different varieties of Cannabis
sativa L. oils were determined, for the first time. An MDGC
− 31.5
− 30.4

− 28.5

− 30.9
− 29.2

− 28.0

− 28.8
instrumentation and method were implemented, exploiting a
δ13C

nd

nd
nd
nd
nd

nd
nd
nd
nd
nd

nd
combination of apolar (1D) and chiral (2D) stationary phases,
with parallel MS and IRMS detection. The coupling of two
Natural 1

separation mechanisms allowed, at first instance, for resolv-

ER %

97.4

88.1
11.9
54.7
45.3

96.5
100
2.6

3.5
nd
nd
nd
nd

nd
nd
nd
nd
nd

ing most of the co-elutions observed in a monodimensional

analysis performed on any of the two columns. A more in-
depth analysis of the results revealed also the higher accu-
− 28.5
− 28.4

− 31.5
− 27.9
− 28.1

− 33.9
δ13C

racy of the data detected/measured by the multidimensional

nd

nd
nd
nd
nd

nd
nd
nd
nd
nd
nd
nd
Reconstituted 16

approach, as a consequence of increased peak separation.

The analysis of a set of commercial oil samples in some
ER %
42.3
57.7

97.6

98.4
66.1
33.9
49.1
50.9

98.0

cases evidenced enantiomeric ratios and δ13C values out-

100
2.4

1.6

2.0
nd
nd
nd
nd
nd

side the typical ranges of genuine oils. Such findings suggest

the usefulness of the Es-MDGC-C-IRMS/qMS method, as
− 29.1
− 28.5

− 30.8

− 28.4

− 34.2

− 38.2
δ13C

an effective tool to ascertain the genuineness and quality

nd

nd

nd
nd
nd
nd

nd
nd
nd
nd
nd

nd
Reconstituted 15

of these samples. Current efforts are devoted to expand the

dataset of oil samples investigated, so to increase the statisti-
cal confidence of the genuineness ranges established for the
ER %
42.0
58.0

97.7

97.9

100

100
2.3

2.1

nd
nd
nd
nd

nd
nd
nd
nd
nd

nd

key oil components.

Supplementary Information The online version contains supplemen-

tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00216-0​ 22-0​ 4035-1.

Declarations

Ethics approval Not applicable.

Source of biological material Not applicable.

(-/ +)
(-/ +)
(-/ +)
(-/ +)
( +)

( +)

( +)

( +)

( +)

( +)

( +)
(-)

(-)
(-)

(-)

(-)

(-)

(-)

Statement on animal welfare Not applicable.

Conflict of interest The authors declare no competing interests.

References
Table 4  (continued)

Selina-4(15),7(11)-diene

β-Caryophyllene oxide

1. Sommano SR, Chittasupho C, Ruksiriwanich W, Jantrawut P.

(E)-β-Caryophyllene

Selina-3,7(11)-diene

The cannabis terpenes. Molecules. 2020. https://​doi.​org/​10.​3390/​

molec​ules2​52457​92
α-Terpineol
Limonene

2. Ternelli M, Brighenti V, Anceschi L, Poto M, Bertelli D, Licata

α-Pinene

β-Pinene

Linalool

M, Pellati F. Innovative methods for the preparation of medical

13
5656 Cucinotta L. et al.

Cannabis oils with a high content of both cannabinoids and terpe- mass spectrometry. Anal Chem. 2019. https://d​ oi.o​ rg/1​ 0.1​ 021/a​ cs.​
nes. J Pharm Biomed Anal. 2020. https://​doi.​org/​10.​1016/j.​jpba.​ analc​hem.​8b048​38
2020.​113296 13. Mosandl A, Hener U, Georg Schmarr H, Rautenschlein M. Chiro-
3. Fiorini D, Scortichini S, Bonacucina G, Greco NG, Mazzara E, specific flavor analysis by means of enantioselective gas chroma-
Petrelli R, Torresi J, Maggi F, Cespi M. Cannabidiol-enriched tography, coupled on-line with isotope ratio mass spectrometry.
hemp essential oil obtained by an optimized microwave-assisted J High Resolut Chromatogr. 1990. https://​doi.​org/​10.​1002/​jhrc.​
extraction using a central composite design. Ind Crops Prod. 2020. 12401​30722
https://​doi.​org/​10.​1016/j.​indcr​op.​2020.​112688 14. Silfer JA, Engel MH, Macko SA, Jumeau EJ. Stable carbon isotope
4. Micalizzi G, Alibrando F, Vento F, Trovato E, Zoccali M, Guar- analysis of amino acid enantiomers by conventional isotope ratio
naccia P, Dugo P, Mondello L. Development of a novel micro- mass spectrometry and combined gas chromatography/isotope
wave distillation technique for the isolation of Cannabis sativa L. ratio mass spectrometry. Anal Chem. 1991. https://​doi.​org/​10.​
essential oil and gas chromatography analyses for the comprehen- 1021/​ac000​04a014
sive characterization of terpenes and terpenoids, including their 15. Mosandl A. Enantioselective capillary gas chromatography and
enantio-distribution. Molecules. 2021. https://​doi.​org/​10.​3390/​ stable isotope ratio mass spectrometry in the authenticity control
molec​ules2​60615​88 of flavors and essential oils. Food Rev Int. 1995. https://​doi.​org/​
5. Mondello L, Costa R, Sciarrone D, Dugo G. The chiral compound 10.​1080/​87559​12950​95410​63
of citrus oils. In: Citrus oils. Composition, advanced analytical 16. Badea SL, Danet AF. Enantioselective stable isotope analysis
techniques, contaminants, and biological activity; Dugo, G, Mon- (ESIA) — a new concept to evaluate the environmental fate of
dello, L, editors; CRC Press Taylor & Francis Group: Boca Raton, chiral organic contaminants. Sci Total Environ. 2015. https://​doi.​
FL, USA, 2010. pp. 349–403 org/​10.​1016/j.​scito​tenv.​2015.​01.​082
6. Dugo P, Ragonese C, Russo M, Sciarrone D, Santi L, Cotroneo 17. Reichert S, Fischer D, Asche S, Mosandl A. Stable isotope label-
A, Mondello L. Sicilian lemon oil: composition of volatile and ling in biosynthetic studies of dill ether, using enantioselective
oxygen heterocyclic fractions and enantiomeric distribution of multidimensional gas chromatography, online coupled with iso-
volatile components. J Sep Sci. 2010. https://d​ oi.o​ rg/1​ 0.1​ 002/j​ ssc.​ tope ratio mass spectrometry. Flavour Fragr J. 2000;15:300–8.
20100​0578 18. Matucha M, Jockisch W, Verner P, Anders G. Isotope effect in
7. Mondello L, Casilli A, Tranchida PQ, Sciarrone D, Dugo P, gas—liquid chromatography of labelled compounds. J Chroma-
Dugo G. Analysis of allergens in fragrances using multiple heart- togr A. 1991. https://​doi.​org/​10.​1016/​0021-​9673(91)​85030-J
cut multidimensional gas chromatography-mass Spectrometry. 19. Cucinotta L, De Grazia G, Salerno TMG, Donnarumma D, Donato
LC-GC Europe. 2008;21(3):130–7. P, Sciarrone D, Mondello L. Overcoming the lack of reliability
8. Bonaccorsi I, Sciarrone D, Cotroneo A, Mondello L, Dugo P, associated to monodimensional gas chromatography coupled to
Dugo G. Enantiomeric distribution of key volatile components in isotopic ratio mass spectrometry data by heart-cut two-dimen-
Citrus essential oils. Rev Bras Farmacogn. 2011. https://​doi.​org/​ sional gas chromatography. J Chromatogr A. 2021. https://​doi.​
10.​1590/​S0102-​695X2​01100​50001​23 org/​10.​1016/j.​chroma.​2021.​462473
9. Do TKT, Hadji-Minaglou F, Antoniotti S, Fernandez X. Authen- 20. Tranchida PQ, Sciarrone D, Dugo P, Mondello L. Heart-cutting
ticity of essential oils. Trends Anal Chem. 2015. https://​doi.​org/​ multidimensional gas chromatography: a review of recent evolu-
10.​1016/j.​trac.​2014.​10.​007 tion, applications, and future prospects. Anal Chim Acta. 2012.
10. Sciarrone D, Schepis A, Zoccali M, Donato P, Vita F, Creti D, https://​doi.​org/​10.​1016/j.​aca.​2011.​12.​015
Alpi A, Mondello L. Multidimensional Gas chromatography 21. Muccio Z, Wöckel C, An Y, Jackson G.P. Comparison of bulk and
coupled to combustion-isotope ratio mass spectrometry/quadru- compound-specific δ13C isotope ratio analyses for the discrimina-
pole MS with a low-bleed ionic liquid secondary column for the tion between cannabis samples. J. Forensic Sci 2012. https://​doi.​
authentication of truffles and products containing truffle. Anal org/​10.​1111/j.​1556-​4029.​2011.​02046.x
Chem. 2018. https://​doi.​org/​10.​1021/​acs.​analc​hem.​8b003​86 22. Bicchi C, Binello A, D’Amato A, Rubiolo P. Reliability of Van den
11. Zhang L, Kujawinski DM, Federherr E, Schmidt TC, Jochmann Dool retention indices in the analysis of essential oils. J Chromat
MA. Caffeine in your drink: natural or synthetic? Anal Chem. Sci. 1999. https://​doi.​org/​10.​1093/​chrom​sci/​37.8.​288
2012. https://​doi.​org/​10.​1021/​ac203​197d
12. Reay MK, Knowles TDJ, Jones DL, Evershed RP. Development Publisher's note Springer Nature remains neutral with regard to
of alditol acetate derivatives for the determination of 15N-enriched jurisdictional claims in published maps and institutional affiliations.
amino sugars by gas chromatography–combustion–isotope ratio

13