Journal of Chromatographic Science, Vol.

42, March 2004

Separation and Identification of Cannabis Components

by Different Planar Chromatography Techniques
(TLC, AMD, OPLC)

Downloaded from https://academic.oup.com/chromsci/article/42/3/130/282899 by guest on 23 August 2020

N. Galand1,*, D. Ernouf2, F. Montigny3, J. Dollet1, and J. Pothier1,*
1UFR des Sciences Pharmaceutiques, Laboratoire de Pharmacognosie, 31 Avenue Monge F-3720 Tours, France; 2UFR des Sciences
Pharmaceutiques, Laboratoire de Toxicologie, 31 Avenue Monge, F 37200 Tours, France; and 3UFR des Sciences Pharmaceutiques,
Plateau d’Analyse Chimique, 31 Avenue Monge, F-37200 Tours, France

Abstract metabolites cross reactivities.

Moreover, a wide variety of methodologies have been recom-
The use of cannabis is illicit in numerous countries, and the mended for the determination of marijuana samples or cannabis
increasing consumption has led to a multiplication of scientific plants: TLC (11), OPLC (12), HPLC (13), GC, and GC–MS (14,15),
studies. New methods of planar chromatography such as automated capillary electrochromatography (16), time-resolved fluoroim-
multiple development (AMD) and optimum performance laminar munological method (17), and immunoassay (18). Most of these
chromatography (OPLC) techniques can be used as a substitute for techniques require heavy and costly instruments and a lot of
the traditional thin-layer chromatography for the identification and time.
quantitation of the Indian hemp components. Each method offers its Planar chromatography is a suitable method to simultaneously
own advantage: high resolution with neither diffusion nor spot
screen numerous samples directly from plants. It has become a
stretching for AMD and speed, efficiency, and the possibility of
working in the semipreparative mode for OPLC.
modern technique with the commercialization of numerous
adsorbents and new appliances with automated development
chambers such as OPLC and AMD, which are interesting alterna-
tives to classical TLC.
OPLC opens up the possibility of analyzing an important
Introduction number of samples in a very short time. Moreover, this method
can be used in the semipreparative mode to purify products by
Because of the ever-increasing use of cannabis, it has become direct elution thanks to the fact that migration is linear in corre-
necessary to dispose of a whole range of efficient methods for the lation with time. hRf values are reproducible, and each com-
identification of its components and particularly for the charac- pound is eluted at a defined time.
terization of the “narcotic compound” ∆9-tetrahydrocannabinol AMD presents the best resolution without any spot diffusion
(∆9-THC). Analyses can be carried out from plants or biological and without oxidation because the microchamber is saturated
fluids. In urine, blood, and saliva samples, ∆9-THC and major with, for instance, methanol under a nitrogen atmosphere.
metabolites, such as 11-nor-∆9-THC-9-carboxylic acid, can be The aim of this study was to compare the performances of the
often required (1). Cannabinoids can be detected by numerous TLC, AMD, and OPLC techniques for the identification and quan-
and various analytical methods, including immunoassays (EMIT, titation of the cannabis components.
enzyme-linked immunosorbent assay, fluorescence polarization,
and radioimmunoassay) (2,3,4,5,6), planar chromatography tech-
niques [classical thin-layer chromatography (TLC), optimum
performance laminar chromatography (OPLC), and automated Experimental
multiple development (AMD)], gas chromatography–mass spec-
trometry (GC–MS) (7,8,9), and high-performance liquid chro- The standard solutions of ∆9-THC, ∆9-THC, and cannabinol
matography (HPLC)–MS (10). Generally, there is a good (CBN) were purchased from Sigma-Aldrich Chimie (Saint
quantitative correlation between these methods and few discrep- Quentin Fallavier, France). Because cannabidiol (CBD) is not
ancies even in the borderline region, especially if the cutoffs available, it was isolated from cannabis resin by OPLC in the
through immunoassay techniques are low, in spite of the different semipreparative mode (see the Semipreparative OPLC applied to
* Authors to whom correspondence should be addressed: emails [email protected] or
isolation of standard CBD section).
[email protected]. All the standard solutions were prepared in 0.5 mg/1 mL in

130 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Journal of Chromatographic Science, Vol. 42, March 2004

methanol. Cannabis resin (0.1 g) was extracted by shaking at scanned from m/z 35–450. Chromatographic data were acquired
room temperature for 20 min with 10 mL of hexane. The filtrate using HP Chemstation software (Hewlett-Packard).
was evaporated to dryness and the residue dissolved in 1 mL of
toluene. A hemp sample (0.5 g) was extracted for 10 min with 20
mL of hexane. After filtration, the extract was evaporated under
vacuum and the residue dissolved in toluene. All solvents were Results and Discussion
purchased from Carlo Erba Reactifs (Val de Reuil, France) and
then distilled. A Linomat IV (Camag, Muttenz, Switzerland) was TLC
used for sample applications. A TLC-MAT Desaga (Bionisis, Le Comparison of various eluents used in TLC for the
Plessis-Robinson, France), OPLC 50 (Bionisis), and AMD (Camag) separation of cannabinoids
were used for the chromatographic studies. TLC and AMD were TLC is a suitable method for screening different samples.
performed on 10- × 20-cm plates (precoated silicagel HPTLC F254) In the literature, the eluents that are mostly used are eluent A,
(Merck Art. 11764) (VWR International SAS, Fonterlaysous Bois, isooctane–ethyl acetate–acetic acid (30:10:1, v/v/v) (20); eluent B,

Downloaded from https://academic.oup.com/chromsci/article/42/3/130/282899 by guest on 23 August 2020

France). OPLC was performed on an HTSorb BSLA 011 and HT petroleum ether–diethyl ether (90:10) (21); eluent C, ace-
Sorb BSLA 003 (Bionisis). The chromatograms were derivated by tone–methylene chloride–diisopropyl ether–hexane (1:1:3:20,
spraying with Fast Blue salt B reagent (19). v/v/v/v) (22); eluent D, toluene–chloroform–methanol (100:10:1,
For classical TLC and TLC-MAT, the eluent used was v/v/v) (23); eluent E, hexane–dioxane (90:10, v/v) double migra-
tion (24); and eluent F, hexane–diethyl ether (80:20, v/v) (19).
The eluents A and B result in a clean separation between ∆9-
hexane–diethyl ether (80:20, v/v). For AMD, the elution gradient
THC and CBN but not between ∆9-THC and CBD.
was acetone (100) (bottle 1), diisopropylether (100) (bottle 2),
hexane (100) (bottles 3–6), and migration during 20 steps. For
OPLC, the eluent was isooctane–diethyl ether (90:10, v/v). The With eluent C, the main cannabinoids are separated, but the
external pressure was 50 bars, flow rate 100 µL/min, flash volume spots are stretched.
The best results were obtained with eluent F, hexane–diethyl
ether (80:20, v/v) (Table I), which allowed a clean separation of ∆8-
75 µL, eluent volume 1000 µL, and the migration time 607 s.
THC, ∆9-THC, CBN, and CBD (Figure 1).
The GC–MS instrumentation used consisted of a Hewlett-
The separation of cannabinoids ∆9-THC, CBN, and CBD by clas-
Packard system (HP 5890 series II GC with an HP5989A
quadrupole MS) (Palo Alto, CA). An HP-5MS 15-m × 0.25-mm ×
0.25-µm capillary column and a helium (99.99%) carrier gas at a sical TLC is not easy because these derivatives possess chemical
structures with very close substitutes. Besides, the molecular
weights of ∆9-THC and CBD are the same (314.47), and the
flow rate of 1.3 mL min–1 were used. The injector temperature
was maintained at 250ºC, and all injections were made in the
splitless mode. The GC oven temperature was held at 50ºC for molecular weight of CBN is very close (310.44).
1 min and then programmed to 250ºC at 10ºC min–1 and main-
tained for 10 min. The GC–MS transfer line was maintained at
280ºC, electron ionization at 70eV, and the mass spectrum

Table I. hRf and Colors with Fast Blue Salt Reagent of

Cannabinoides

hRf
1 2 3 4 5 6
Colors with
Cannabinoids TLC-MAT AMD OPLC Fast Blue salt B
Figure 2. AMD of cannabinoids: (1) 7 µL cannabis extract, (2) 1 µL CBD, (3)
CBN 59 73 23 violet 1 µL CBN, (4) 2 µL cannabis resin, (5) 3 µL ∆9 THC, and (6) 5 µL cannabis
∆9-THC 66 76 28 purple extract.
CBD 73 79 34 orange-red

1 2 3 4 5 6 7

Figure 1. TLC-MAT of cannabinoids: (1) 1 µL CBD, (2) 7 µL cannabis extract, 1 2 3 4

(3) 2 µL ∆8-THC, (4) 1 µL CBN, (5) 2 µL cannabis resin, (6) 3 µL ∆9-THC, and Figure 3. OPLC of cannabinoids: (1) 1 µL CBN, (2) 2 µL cannabis resin, (3) 3
(7) 7 µL cannabis extract. µL ∆9-THC, and (4) 1 µL CBD.

131
 Journal of Chromatographic Science, Vol. 42, March 2004

The analysis of the chromatograms revealed two different “zigzag” shape because the viscosity of the eluents was too low
groups of cannabinoids, a first group, the least polar, composed of and the silicagel plates were not homogeneously permeated deep
CBD, CBN, and ∆9-THC (upper hRf values) and a second one, inside their structure. To solve this problem, the viscosity was
which consisted in many compounds with lower hRf. increased by replacing hexane with a higher homologous such as
The detection limit with the Fast Blue salt reagent was 0.25 µg isooctane, which does not change the elution power of the eluent
for ∆9-THC, CBD, and CBN. but increases inner pressure, leading to higher hRf values, thus
Different eluents were tested: isooctane, heptane, hexane, and improving considerably the shape of the stripe—the best results
pentane–diethyl ether (90:10, v/v). The comparison between were obtained with isooctane–diethyl ether (90:10, v/v) as eluent
these four alkanes showed that the separation capability (Table I) (Figure 3). Moreover, hexane–diethylether (80:20, v/v)
decreased when the carbon-bearing chain lengthened. used in the semipreparative mode offers the advantage of evapo-
After this traditional TLC study, these compounds were studied rating easily because of its low viscosity.
with modern planar chromatographic methods such as AMD and
OPLC, with the aim of optimizing their separation and identifica- Semipreparative OPLC applied to isolation of standard CBD
CBN and ∆9-THC were obtained from Sigma-Aldrich. Because

Downloaded from https://academic.oup.com/chromsci/article/42/3/130/282899 by guest on 23 August 2020

tion.
obtaining CBD was not possible, this compound was isolated from
AMD cannabis resin by OPLC in the semipreparative mode. In the lit-
The “universal gradient” nº1 with methanol, methylene chlo- erature, few works have reported this technique. First of all, this
ride, and hexane is far too polar. Therefore, it was necessary to method had been tested on opium (25) and xanthines from tea
decrease the polarity by replacing methanol with methylene chlo- leaves extracts (26). In the case of cannabis extract, two series of
ride. compounds are shown. The first one has an hRf above 50, which
First, two gradients were tested. Elution gradient 1A was is easily and quickly carried out, and the second one has an hRf
methylene chloride (100), methylene chloride (100), methylene below 50. For the latter, it was necessary to increase the eluent
chloride–hexane (50:5), hexane (100), hexane (100), and hexane polarity. The aim of this work was to obtain pure compound from
(100) during 25 steps. the resin of cannabis by coupling the chromatograph with a frac-
Elution gradient 1B was diethyl ether (100)–hexane (50:50), tion collector. The extract was applied inline with Linomat IV and
hexane (100), hexane (100), and hexane (100). eluted with hexane–diethyl ether (90:10, v/v). The migration of
These two eluents are interesting for revealing most of the con- the eluent was performed during the time required to begin the
stituents of cannabis but do not separate ∆9-THC with CBN and elution process. Because OPLC allows for linear migration in cor-
CBD very clearly. relation with time, it was able to determine the instant when CBD
In AMD, the best separation of the three interesting compounds was collected. Thanks to OPLC, it was possible to obtain pure
was realized in high-performance TLC with the elution gradient CBD.
1C acetone (100), diisopropylether (100), hexane (100), hexane The advantage of OPLC compared with TLC in the semiprepar-
(100), and hexane (100) during 20 steps (Table I) (Figure 2). ative mode is that no scraping and eluting of bands are necessary.
The visualization of the chemical constituents was accom- In OPLC, the components can be eluted from the plate and
plished by spraying Fast Blue salt B reagent (19). obtained pure by coupling with a fraction collector.
The different cannabinoids were identified by their hRf and the Every elution fraction was evaluated by analytical TLC with
color of the spots (purple for ∆9-THC, orange-red for CBD, and hexane–diethyl ether (80:20, v/v) as eluent. After derivation by
violet for CBN). Fast Blue salt B reagent (19), four fractions were obtained, giving
only one spot in TLC. The control of the structural study per-
OPLC formed by GS–MS analysis allowed the identification of a com-
Analytical OPLC pound present in the sample as being CBD (Figure 4).
Applying the classical TLC eluent hexane–diethyl ether (80:20,
v/v), OPLC gave clear separation, but the different stripes took a Structural analysis of CBD
From chromatographic data obtained by
GC–MS, the organic compounds were identified
by computer. Standard reference mass fragmen-
tograms listed in the National Institute of
Standards and Technology (NIST) library were
compared with the specific results obtained here.
The total ion chromatogram obtained showed
one major compound. A search in the database
spectral library indicated that this substance
might very likely be “2-[3-methyl-6-(1-methyl-
ethenyl)-cyclohex-2-en-1-yl]-5-pentyl benzene-
1,3-diol (Figure 1). The identified compound is
also known under the synonym CBD. This
Figure 4. (Top) MS–GC of CBD obtained from cannabis hemp by semipreparative mode OPLC and
(Bottom) MS of CBD from NIST library. assumption is in full agreement with the mass
spectrum of CBD investigated by Baptista (27).

132
Journal of Chromatographic Science, Vol. 42, March 2004

Furthermore, the chromatograms showed other substances with drocannabinol, and 11-Nor-∆9-tetrahydrocannabinol-9-carboxylic
acid. J. Anal. Toxicol. 25: 538–49 (2001).
low concentrations, which can be assigned as aliphatic and
2. D. Altunkaya, A.J. Clatworthy, R.N. Smith, and I.J. Start. Urinary
ethylenic hydrocarbons. Unfortunately, these compounds cannot cannabinoid analysis: comparison of four immunoassays with gas
be identified with certainty because of their poor mass spectra chromatography-mass spectrometry. Forensic Sci. Int. 50: 15–22,
resolution. (1991).
3. S. Kerrigan and W.H. Philips, Jr. Comparison of ELISAs for opiates
methamphetamine-cocaine metabolite, benzodiazepines, phency-
clidine, and cannabinoids in whole blood and urine. Clin. Chem. 47:
540–47 (2001).
Conclusion 4. T. Korte, J. Pykalainen, P. Lillsunde, and T. Seppala. Comparison of
RapiTest with Emit d.a.u. and GC–MS for the analysis of drugs in
The modern planar chromatography techniques are reliable urine. J. Anal. Toxicol. 21: 49–53 (1997).
5. A.D. Fraser and D. Worth. Monitoring urinary excretion of cannabi-
because they are automated and inexpensive, they allow a better noids by fluorescence-polarization immunoassay: a cannabinoid-to-
resolution than classical TLC, and can potentially replace slower creatinine ratio study. Ther. Drug. Monit. 24: 746–50 (2002).

Downloaded from https://academic.oup.com/chromsci/article/42/3/130/282899 by guest on 23 August 2020

and more costly methods (GC–MS), thus increasing the produc- 6. D. Perez-Bendito, A. Gomez-Hens, and A. Gaikwad. Direct
tivity of the laboratory thanks to their ability to analyze several stopped-flow fluorescence polarization immunoassay of abused
samples at the same time (up to 20). drugs and their metabolites in urine. Clin. Chem. 40: 1489–93
(1994).
Therefore, with traditional TLC, it was possible to separate D9- 7. R.L. Foltz and I. Sunshine. Comparison of a TLC method with EMIT®
THC, CBD, and CBN from cannabis resin and Indian hemp herb, and GC–MS for detection of cannabis in urine. J. Anal. Toxicol. 14:
but this method did not offer a clean separation of the most polar 375–78 (1990).
compounds; four spots for cannabis extracts with classical TLC 8. M.L. Weaver, B.K. Gan, E. Allen, L.D. Baugh, F.-Y. Liao, R.H. Liu,
and eight spots for the resin could be obtained with AMD. J.G. Langner, A.S. Walia, and L.F. Cook. Correlations on radioim-
munoassay, fluorescence polarization immunoassay, and enzyme
AMD offers high resolution without any stretching of spots, the immunoassay of Cannabis metabolites with gas chromatography/
focalization of which gives the possibility of making dosage by mass spectrometry analysis of 11-Nor-∆9-tetrahydrocannabinol-9-
scanner densitometry. carboxylic acid in urine specimen. Forensic Sci. Int. 49: 43–56
These two modern techniques, OPLC and AMD, are repro- (1991).
ducible because they are completely automated. They can provide 9. J. Teske, K. Putzbach, W. Engewald, and R.K. Müller. Determination
of cannabinoids by gas chromatography-mass spectrometry and
interesting information about the composition of different sam- large-volume programmed-temperature vaporiser injection using
ples of Indian hemp and open up the possibility to determine the 25µl of biological fluid. J. Chromatogr. B 772: 299–306 (2002).
geographical origin of different samples. 10. W. Weinmann, S. Vogt, R. Goerke, C. Muller, and A. Bromberger.
The benefits of OPLC compared with TLC are numerous, Simultaneous determination of THC-COOH and THC-COOH-glu-
namely efficiency, reproducibility, small consumption of devel- curonide in urine samples by LC/MS/MS. Forensic Sci. Int. 113:
381–87 (2000).
oping eluent, and shorter analysis delay. Consequently, the spots 11. D. Debruyne, F. Albessard, M.C. Bigot, and M. Moulin. Comparison
have a more regular shape and diffusion and stretching is not as of three advanced chromatographic techniques for Cannabis identi-
pronounced as in TLC. Another major advantage of OPLC com- fication. Forensic Sci. Int. 106: 135–46 (1999).
pared with TLC is the possibility to extend this method to 12. P. Oroszlán, G. Verzár-Petri, E. Mincsovics, and T. Székely.
semipreparative chromatography, in which no scraping and Separation, quantitation and isolation of cannabinoids from
Cannabis sativa L. by overpressured layer chromatography.
eluting of bands are necessary because the components can be J. Chromatogr. 388: 217–24 (1987).
drained from the plate and obtained pure by coupling OPLC to a 13. O. Zoller, P. Rhyn, and B. Zimmerli. High-performance liquid chro-
collector. matographic determination of delta 9-tetrahydrocannabinol and the
An additional benefit of planar chromatographic techniques corresponding acid in hemp containing foods with special regard to
versus HPLC and GC lies in the fact that it is possible to detect in the fluorescence properties of delta 9-tetrahydrocannabinol.
J. Chomatogr. A 872: 101–10 (2000).
the cannabis samples other addictive products belonging to dif- 14. A.J. Poortman-Van Der Meer and H. Huizer. A contribution to the
ferent chemical classes (e.g., alkaloids, like opiates and deriva- improvement of accuracy in the quantitation of THC. Forensic Sci.
tives; cocaine; and nicotine) mixed in a single cannabis sample by Int. 101: 1–8 (1999).
using specific reagents (e.g., iodoplatinate or Dragendorff, in the 15. S.A. Ross, Z. Mehmedic, T.P. Murphy, and M.A. Elsohly. GC–MS
case of alkaloids). analysis of the total delta 9-THC content of both drug and fiber-type
cannabis seeds. J. Anal. Toxicol. 4: 715–17 (2000).
TLC, OPLC, and AMD can also supply interesting information 16. I.S. Lurie, R.P. Meyers, and T.S. Conver. Capillary electrochromatog-
in regards to the composition of various samples of hemp and raphy of cannabinoids. Anal. Chem. 70: 3255–60 (1998).
offer the possibility of determining its origin. 17. M.A. Bacigalupo, A. Ius, G. Meroni, G. Grassi, and A. Moschella.
Time-resolved fluoroimmunoassay for delta(9)-tetrahydrocannabinol
as applied to early discrimination of Cannabis sativa plant. J. Agric.
Food Chem. 47: 2743–45 (1999).
18. H. Tanaka and Y. Shoyama. Monoclonal antibody against tetra-
References hydrocannabinolic acid distinguishes Cannabis sativa samples from
different plant species. Forensic Sci. Int. 106: 135–46 (1999).
1. J.E. Manno, B.R. Manno, P.M. Kemp, D.D. Alford, I.K. Abukhalaf, 19. H. Wagner and S. Bladt. Plant Drug Analysis—A Thin Layer
M.E. Mc Williams, F.N. Hagaman, and M.J. Fitzgerald. Temporal Chromatography Atlas, 2nd ed. Spinger Verlag, Berlin, Germany,
indication of marijuana use can be estimated from plasma and urine 1996, p. 260–61.
concentrations of ∆9-tetrahydrocannabinol, 11-hydroxy-∆9-tetrahy- 20. G. Alemany, A. Gamundi, M.C. Nicolau, and D. Serro. A simple

133
 Journal of Chromatographic Science, Vol. 42, March 2004

method for plasma cannabinoid separation and quantification. 25. J. Pothier, N. Galand, S. Fouchécourt, and C. Viel. Isolation of alka-
Biomed. Chromatograph. 7: 273–74 (1993). loids by over thin-layer chromatography. Fitoterapia 68: 42–44
21. K D. Parker, J.A. Wright, A.F. Halpern, and C.H. Hine. Preliminary (1997).
report on the separation and quantitative determination of cannabis 26. E. Mincsovics, M. Garami, L. Kecskès, and B. Tapa. Personal over-
constituents present in plant material, and when added to urine, by pressured-layer chromatography (OPLC) basic system flexible tool in
thin-layer and gas chromatography. Bull. Marc. 20: 9 (1968). analytical and semipreparative work. J. AOAC Int. 82: 587–98
22. B. Szabady, Z. Fater, and S. Nyiredy. Comparative study of auto- (1999).
mated development chambers. J. Planar Chromatogr. 12: 82–85 27. M.J. Baptista, P. Venâncio Monsanto, E. Gouveia Pinho Marques,
(1999). A. Bermejo, S. Ávila, A. Martelo Castanheira, C. Margalho, M.
23. A. Schurr and B.M. Rigor. Cannabis extract but not ∆1-tetrahydro- Barroso, and D. Nuno Vieira. Hair analysis for ∆9-THC-COOH, CBN
cannabinol, inhibits human brain and liver monoamine oxidase: the and CBD, by GC/MS-EI. Comparison with GC/MS-NCI for ∆9-THC-
vascular system. Gen. Pharm. 15: 171–74 (1984). COOH. Forensic Sci. Int. 128: 66–78 (2002).
24. E. Stahl. Analyse Chromatographique et Microscopique des
Drogues. Tec. et Doc., Paris, France, 1975, pp. 137–41. Manuscript accepted November 26, 2003

Downloaded from https://academic.oup.com/chromsci/article/42/3/130/282899 by guest on 23 August 2020

134