Supercritical extraction strategies using CO2 and ethanol to obtain cannabinoid compounds from Cannabis hybrid flowers

Journal of CO₂ Utilization 28 (2018) 174–180

Contents lists available at ScienceDirect

Journal of CO2 Utilization journal homepage: www.elsevier.com/locate/jcou

Supercritical extraction strategies using CO2 and ethanol to obtain cannabinoid compounds from Cannabis hybrid flowers

T

Daniel Ribeiro Grijóa, Ignacio Alberto Vieitez Osoriob, Lúcio Cardozo-Filhoa,c,

⁎

a

Department of Chemical Engineering, Universidade Estadual de Maringá, Maringá, PR, Brazil Departmento de Ciencia y Tecnología de Alimentos, Universidad de la República, Montevideo, Uruguay c Center for Research, Centro Universitário da Fundação de Ensino Octávio Bastos (UNIFEOB), São João da Boa Vista, SP, Brazil b

ARTICLE INFO

ABSTRACT

Keywords: Cannabis Cannabinoid Supercritical extraction Decarboxylation Winterization

The genus Cannabis contains specific substances called cannabinoids that are found in high concentrations in the flowers of non-pollinated female plants. Cannabidiol (CBD) and Δ9-tetrahydrocannabinol (Δ9-THC) are the two main bioactive substances in this group with medicinal potential. However, these substances are found in low concentrations in fresh flowers. The decarboxylation technique can be applied to fresh flowers to promote an increase in the levels of these substances. Extracts with high levels of cannabinoids and without organic solvent residues have high medicinal and cosmetic potential. The purpose of this work was to present the results of cannabinoid extraction using pressurized fluids from two varieties of flowers of the genus Cannabis. The extractions were conducted using pure supercritical carbon dioxide (scCO2) and with ethanol as a co-solvent, comparing the use of decarboxylation and winterization techniques. The chemical profiles of cannabinoids (CBD, Δ9-THC and cannabinol (CBN)) and the essential oils in the extracts and in fresh flowers were analyzed using different chromatographic techniques. The decarboxylation technique employed maximized the levels of the cannabinoids of interest. The Sovová model used in the adjustment of the experimental kinetic extraction curves was adequate and satisfactory.

1. Introduction Cannabidiol (CBD) and Δ9-tetrahydrocannabinol (Δ9-THC) are the main cannabinoids with medicinal potential present in plants of the genus Cannabis [1–3]. However, the highest concentrations of cannabinoids in fresh flowers are cannabidiolic acid (CBDA) and Δ9-tetrahydrocannabinoic acid (Δ9-THCA). The transformation of these cannabinoid acids into their respective neutral cannabinoids, CBD and Δ9THC, is possible by a decarboxylation reaction [4–6]. This reaction is favored by several factors such as storage time [7], heating [8] and the use of alkaline conditions [7]. Controlled heating is the simplest technique used to promote decarboxylation and prevent the degradation of desirable cannabinoids [9]. Upon degradation, CBD can be converted to Δ9-THC [10] and/or cannabielsoin (CBE) [11] and Δ9-THC is converted to cannabinol (CBN) and/or Δ8-tetrahydrocannabinol (Δ8-THC) [12]. The CBD and Δ9-THC that form during decarboxylation are nonpolar and soluble in supercritical carbon dioxide (scCO2) [13]. However, the waxes present in the flowers are also extracted by scCO2. The

⁎

removal of these waxes through the “winterization” process can generate a desirable increase in the concentration of the cannabinoids in the extract. Syntactically, this process consists of suspending the extract in n-hexane and then decanting the waxes by severe cooling [14]. The purpose of this work is to present results of cannabinoid extraction using pressurized fluids from two varieties of flowers of the genus Cannabis. The extractions were conducted using pure scCO2 and with ethanol as a co-solvent, comparing the use of the decarboxylation and winterization techniques. The chemical profile of cannabinoids (CBD, Δ9-THC and CBN) and the essential oils in the extracts and in fresh flowers were analyzed using different chromatographic techniques. The Sovová model was used for the adjustment of the experimental kinetic curves of extraction. The decarboxylation technique was evaluated to maximize the pre-extraction contents of the cannabinoids of interest. The extraction conditions used in this work on Cannabis flowers with pressurized carbon dioxide were defined from the solubility data of CBD and Δ9-THC in scCO2, as described in the literature [14–16].

Corresponding author at: Department of Chemical Engineering, Universidade Estadual de Maringá, Maringá, PR, Brazil. E-mail addresses: [email protected], [email protected] (L. Cardozo-Filho).

https://doi.org/10.1016/j.jcou.2018.09.022 Received 22 May 2018; Received in revised form 21 August 2018; Accepted 27 September 2018 2212-9820/ © 2018 Elsevier Ltd. All rights reserved.

Journal of CO₂ Utilization 28 (2018) 174–180

D.R. Grijó et al.

2. Materials and methods

higher than 0.9999.

2.1. Chemicals

2.2.4.2. Decarboxylation. Samples of the flowers were heated at temperatures of 90, 110 and 140 °C for a period of three hours. The cannabinoid content of each sample was determined at the intervals of 30, 60, 120 and 180 min.

Table 1 presents some information on the chemicals used in this work. All chemicals were used without further treatment. 2.2. Plant material

2.3. Organic solvent extraction

The flowers used were of two hybrid varieties. The “GSC” variety called “Girl Scout Cookies” (The Cali Connection) consists of 60% C. sativa and 40% C. indica with an estimated chemical composition of approximately 25% Δ9-THC. The “DMII” variety called “Durga Mata II CBD” (Paradise Seeds) consists of 10% C. sativa and 90% C. indica with a chemical composition estimated at 6.5% CBD and 7.5% Δ9-THC.

A 3 mL volume of a methanol: chloroform solution (9:1 v/v) was added to 150 mg of ground flower samples. The mixture was homogenized using an ultrasonic agitator (UltraCleaner, 1400 A) for 15 min at 37 Hz and 40 °C. Then, the mixture was centrifuged (CentriBio) for 10 min at 3000 rpm and the supernatant was collected. To determine the oil content in the sample without heating, the supernatant was dried in an oven (New Ethics, 400 /4ND) with air circulation at 35 °C for approximately 20 h until the sample weight was constant. For analysis of the cannabinoid composition, 200 μL of the supernatant was diluted in methanol to obtain a concentration of 1 mg of the solid sample used in 1 mL of solution [18].

2.2.1. Milling The flower samples were milled using a knife mill (Solab, SL-30). Samples were classified using sieves (Bertel) of 10, 14, 20, 28, 35, 48, 65 and 100 mesh. The mean particle diameter of the samples was calculated using the Sauter equation (Eq. (1)) [17]:

1

DSauter = n i=1

¯

2.4. Supercritical extraction

n

(1)

Dn

In order to potentiate the cannabinoids of interest (CBD and Δ9THC) two extraction strategies using supercritical carbon dioxide (scCO2) pure and cosolvent were studied. In the first technique, the samples were initially subjected to the decarboxylation process (heating) and in sequence subjected to the extraction process with scCO2. In the second technique, the samples are extracted using scCO2 and ethanol as a co-solvent. Synthetically, the equipment used in the scCO2 extractions consisted of a syringe pump (ISCO), a stainless steel extractor with a capacity of 58 mL (19.40 cm in height and 1.95 cm in diameter) and micrometric valves for depressurizing. The experimental extraction apparatus is described in several studies by our research group available in the literature [17]. The extraction conditions were defined from studies on the solubility of cannabinoids in scCO2 [15,16].

¯

where Δφ is the sample fraction retained in the sieve and D is the average of the sieve diameters. 2.2.2. Moisture The relative humidity of the samples (3 g) were determined in triplicate using a furnace (New Ethics, 400/4ND) with air circulation at 35 °C for approximately 20 h until the sample weight was constant. 2.2.3. Density The density of the flower samples was determined in triplicate using a pycnometer (Quantachrome Instruments, 440-C Stainless Steel) of Helium gas at 20 psi. 2.2.4. Potential of cannabinoids 2.2.4.1. Calibration curves. The calibration curves for the concentrations of the cannabinoids CBD, Δ9-THC and CBN were performed using six dilutions (0.001, 0.003, 0.005, 0.010, 0.050 and 0.100 mg mL−1) in methanol using a Shimadzu HPLC 20 A with diode array detector at 220 nm and RP-8 column (SUPELCOSIL (TM): 250 x 4.5 mm, 5 μm) at 35 °C. The mobile phase used was a solution of acetonitrile and water (8:2 v/v) under isocratic conditions with a flow rate of 1 mL/min for 10 min [18]. The curves showed linearity (R²)

2.4.1. With decarboxylation The extraction experiments were conducted using 2 g of sample in each analysis at temperatures of 50, 60 and 70 °C and flow rate of 2.5 mL min−1. The pressures used were 16.5, 20.7 and 24.9 MPa for the “GSC” variety and 12.8, 18.4 and 24.0 MPa for the “DMII” variety. The samples were submitted to the decarboxylation process at 140 °C for 30 min before extraction.

Table 1 Chemicals employed in this work. Chemical

IUPAC Nomenclature

Molecular formula

Molar mass (g∙mol−1)

Supplier (country)

Minimum puritya (%)

Carbon dioxide Ethanol Acetonitrile

Carbon dioxide Ethanol Acetonitrile

CO2 C2H6O C2H6N

44.01 46.07 41.05

99.99 99.8 99.5

Water Helium Chloroform

Water Helium Chloroform

H2O He CHCl3

18.02 4.00 119.37

Methanol Hexane CBD Δ9-THC CBN

Methanol Hexane Cannabidiol Δ9-tetrahydrocannabinol Cannabidiol

CH4O C6H14 C21H30O2 C21H30O2 C21H26O2

32.04 86.18 314.47 314.47 310.44

Linde (Brazil) PanReac (Brazil) Sigma (Uruguay) Sartorius (Uruguay) Linde (Brazil) Nuclear (Brazil) PanReac (Brazil) PanReac (Brazil) Grace Davison Discovery Science (United States)

a

Purities were provided by the manufacturers. 175

Ultra-pure 99.999 99.8 99.9 99.0 99 91 97.9

Journal of CO₂ Utilization 28 (2018) 174–180

D.R. Grijó et al.

2.4.2. With co-solvent Extraction of the “GSC” variety was conducted at 50 °C and 16.5 MPa and the “DMII” variety at 50 °C and 24.0 MPa. The mass used in each extraction was approximately 6 g of sample. The ethanol content in the total flow rate (2.5 mL min−1) was 6% (v/v).

was separated and n-hexane was removed on a rotary evaporator (Fisatom, model 802) at 40 °C. 2.4.5. Composition of cannabinoids in the extracts To carry out the analysis of the cannabinoid composition, a sample preparation step was necessary. For this, 20 mL of methanol was added to 30 mg of each extract. The solution was then subjected to ultrasonic agitation at 37 Hz and 40 °C for 30 min. An aliquot of the homogeneous sample was diluted in methanol in the proportion of 1:10. This aliquot was subjected to 20 min of ultrasonic agitation at 37 Hz and 40 °C before being injected into HPLC. The composition of the cannabinoids in the extracts was determined before and after the “winterization” process.

2.4.3. Mathematical modelling The modelling of the experimental data of the oil extractions was carried out using the Sovová model [19]. This model considers that the oil is divided into a fraction of easy access and a fraction of harder access. The easily accessible fraction of oil is related to the convective mass transfer mechanism and is therefore obtained at a constant extraction rate (CER). The mass extracted as a function of time m (t ) for t < tCER (Equation 2) is related to the dimensionless factor Z (Eq. 3):

Z=

KF • mS • F mF •(1 BED )•

2.5. Analyses of essential oils

(2)

m (t ) = mF •YS•t•[1 exp( Z )]

Analyses of essential oils were carried out on fresh flowers and extracts obtained with ethanol as a co-solvent. Samples were separated in 20 mg quantities and packed in 2 mL amber vials. The vials were wrapped with foil and kept in a heating block (Tecnal TE-021 Dry Block) at 120 °C for 1 h. Then, 2.5 mL of the vapour were collected manually with a glass syringe and injected into a gas chromatograph (Thermo Fisher Scientific, FOCUS GC) with injector at 280 °C. The drag gas used was helium with a flow rate of 1 mL min−1 and split ratio of 1/ 10. The column used was DB5-MS (30 m x 0.25 mm x 0.25 μm). The furnace heating ramp was 50 °C (1 min), 250 °C at 8 °C min−1, 300 °C at 30 °C min−1 and held for 3 min. The transfer temperature for the quadrupole mass detector (Thermo Fisher Scientific, DSQ II) was 280 °C and the data were recorded after 4 min. The identification of the essential oils was performed from the mass-to-charge ratios [20] and by linear temperature programmed retention index (LTPRI) (Eq. (10)) [21]. Mass-to-charge ratios of m/z 45–450 were scanned [20]. The LTPRI index was identified using C9 to C16 alkane standards and previous work described in literature [22,23].

(3)

S

where mF is the mass flow rate of the solvent, YS is the solubility of the oil in the solvent, mS is the sample mass on an oil-free basis, q0 is the ratio of the mass of available oil to the sample mass on an oil-free basis, F and S are the fluid and solid densities respectively, KF is the coefficient of mass transfer by convection and BED is the porosity of the bed. The extraction of the oil fraction of harder access is controlled by the diffusion mechanism in the solid and causes a fall in extraction rate (FER). The mass extracted as a function of time m (t ) for tCER ≤ t < tFER (Eq. (4)) is related to the dimensionless factors Z and W (Eq. (5)): m (t ) = mF •YS• t tCER•exp

Z•YS 1 W •mF •ln exp •(t tCER) r W •q0 (1 r ) mS

Z

(4)

K S•mS W= mF •(1 BED )

(5)

where K S is the diffusion mass transfer coefficient and r corresponds to the fraction of available oil in harder access. tFER (Eq. (6)) The mass extracted as a function of time m (t ) for t will occur only by diffusion. In this case there is a low extraction rate (LER).

m (t ) = mS q0

W •q0 YS •ln 1 + exp W YS

1 •exp

LTPRI = 100•n + 100•

3.1. Characterisation of the samples

W •mF •(tCER t ) •r mS

The relative humidity of the flowers of the “GSC” and “DMII” varieties was 2.76 ± 0.13 and 2.30 ± 0.03%, respectively. The density of the flowers of the “GSC” and “DMII” varieties was 1.30 ± 0.03 and 1.50 ± 0.10 g mL−1, respectively. The mean Sauter diameter of the “GSC” and “DMII” varieties was 0.24 and 0.32 mm, respectively.

The periods delimiting the extraction modelling are described in Eqs. (7) and (8).

(1 r )•mS •q0 YS•Z•mF

tFER = tCER +

W •q0 mS •ln r + (1 r )•exp W •mF YS

(7)

3.2. Supercritical extracts obtained with decarboxylation and without a cosolvent

(8)

3.2.1. Decarboxylation Fig. 1 shows the increase in the percentage of the cannabinoids of interest (CBD and Δ9-THC) in the varieties processed at different temperatures by the decarboxylation process. Fig. 1 (a) shows that the initial Δ9-THC composition of approximately 1.5% was increased to 15% and that CBD levels were not significant for the “GSC” variety. Fig. 1(b) and (c) shows that the initial content of 1% of both cannabinoids was increased to approximately 6% CBD and 5% Δ9-THC in the “DMII” variety. The decarboxylation stage of both varieties at 140 °C for 30 min increased the concentration of CBD and Δ9-THC.

The parameters r , Z and W were calculated by minimising the objective function given by Eq. (9): n exp

N

OF = i=1 j =1

2 (micalc miexp ,j ,j )

(10)

3. Results and discussion

(6)

tCER =

tRi tRn tR (n+ 1) tRn

(9)

where n exp is the number of experiments performed, N is the number and miexp of experimental data of each experiment, micalc ,j , j are the calculated and experimental masses, respectively. 2.4.4. Winterization process To each gram of crude extract obtained from scCO2 extraction was added to 10 mL of hexane. The mixture was then cooled to -80 °C in an Ultra-Freezer (ColdLab, model CL580-86 V) for 20 h. The supernatant

3.2.2. Extraction kinetics Fig. 2 shows the experimental kinetic curves obtained and 176

Journal of CO₂ Utilization 28 (2018) 174–180

D.R. Grijó et al.

Fig. 1. Variation of the composition of the cannabinoids of interest at different temperatures ( 140 °C; 110 °C and 90 °C): (a) Δ9-THC in the “GSC” variety, (b) Δ9-THC in the “DMII” variety and (c) CBD in the “DMII” variety.

Fig. 2. Kinetic curves of extraction with decarboxylation for (a) the “GSC” variety and (b) the “DMII” variety: ▲ (50 °C; 16.5 MPa for “GSC” and 12.8 MPa for “DMII”), □ (50 °C; 24.9 MPa for “GSC” and 24.0 MPa for “DMII”), ○ (70 °C; 16.5 MPa for “GSC” and 12.8 MPa for “DMII”), • (70 °C; 24.9 MPa for “GSC” and 24.0 MPa for “DMII”), △ (60 °C; 20.7 MPa for “GSC” and 18.4 MPa for “DMII”), models.

Table 2 Sovová parameters of extraction with decarboxylation of the “GSC” variety. Run

▲

T (ºC) P (MPa) −1 F (g mL ) Yield (%) mF (g min−1) −1 BED (g mL ) r q0 Z W YS (g g−1) tCER (min) tFER (min) KF (min−1) K S (min−1) AARD (%)

50 16.5 0.732 20.5 2.365 0.056 5.8·10−1 3.3·10−1 6.0 5.3·10−2 4.2·10−3 3.9 4.6·101 6.8·10−1 3.4·10−3 3.8

Table 3 Sovová parameters of extraction with decarboxylation of the “DMII” variety.

□

○

•

△

Run

▲

24.9 0.833 24.5 2.816 0.058

70 16.5 0.566 14.9 1.884 0.058

24.9 0.736 20.4 2.349 0.061

60 20.7 0.734 22.2 2.412 0.058

2.395 0.065

6.2 7.0·10−2 7.8·10−3 1.6 1.9·101 7.7·10−1 5.6·10−3 2.1

4.5 9.2·10−3 4.1·10−3 6.0 3.9·101 5.8·10−1 5.2·10−4 3.5

1.1·101 2.5·10−2 8.2·10−3 1.1 1.6·101 1.3 1.7·10−3 1.3

1.1·101 3.1·10−2 7.7·10−3 1.1 1.7·101 1.3 2.1·10−3 6.4

8.9 6.0·10−2 6.2·10−3 1.8 2.9·101 1.1 4.1·10−3 4.7

T (ºC) P (MPa) −1 F (g mL ) Yield (%) mF (g min−1) −1 BED (g mL ) r q0 Z W YS (g g−1) tCER (min) tFER (min) KF (min−1) K S (min−1) AARD (%)

50 12.8 0.627 13.6 1.975 0.062 2.3·10−1 5.1·10−1 1.3·101 2.4·10−2 4.1·10−3 1.1 2.4·101 1.5 1.2·10−3 5.3

calculated for the flower varieties studied using the Sovová model as a function of the percentage mass yield and the mass ratio between the scCO2 mass used and the initial sample mass. The percent mass yields obtained are compatible with the values available in the literature [14]. The mathematical model used satisfactorily described the experimental kinetic curves. The percentage mass yield obtained with scCO2 for each variety was lower than the percent yields obtained with organic solvents. The weight percent yield obtained using an organic solvent for the extraction of the oil of the flowers of the “GSC” and “DMII” varieties was 37.32 ± 1.44 and 30.78 ± 0.22%, respectively. The highest yields were obtained with the highest solvent densities for the “GSC” variety. For the “DMII” variety, cross-over behavior was observed [24], because in some extractions, the highest yields were not obtained with the highest densities. This was possibly due to the high concentration of both cannabinoids (CBD and Δ9-THC).

□

○

•

△

24.0 0.826 17.0 2.677 0.062

70 12.8 0.390 6.1 1.182 0.062

24.0 0.724 18.2 2.422 0.062

60 18.4 0.695 17.4 2.283 0.062

3.278 0.062

5.5 4.4·10−2 7.2·10−3 9.1·10−1 8.5 8.4·10−1 3.7·10−3 1.4

2.3 3.2·10−7 1.3·10−3 5.0·101 1.7·102 2.7·10−1 8.7·10−9 7.8

4.1·101 8.3·10−2 1.0·10−2 1.3·10−1 9.1 4.8 4.7·10−3 2.5

4.0·101 7.2·10−2 7.3·10−3 1.8·10−1 1.3·101 5.0 4.1·10−3 2.0

4.0·101 5.4·10−2 8.4·10−3 1.6·10−1 1.0·101 4.8 3.0·10−3 1.4

Tables 2 and 3 contain all the relevant information on the operating conditions of extraction and the calculated parameters of the Sovová model [19] for each variety studied. From the data in Tables 2 and 3, it can be observed that, for both varieties, the solvent density ( F ) had the most relevant effect on solute solubility values (YS ). Consequently, the same behavior was observed for the values of the mass transfer coefficients, KF and K S . These behaviors are common in extracting from a plant matrix using scCO2. For the “DMII” variety, cross-over behavior was observed when the extraction pressure was set. This is because, at a specific pressure, an increase in solubility occurs due to an increase in temperature [24]. The values of the calculated parameters using the Sovová model are compatible with the values available in the literature and can help in the scheduling of a pilot unit for supercritical extraction. The ease of scalability using supercritical technology can be observed by the similar

177

Journal of CO₂ Utilization 28 (2018) 174–180

D.R. Grijó et al.

Table 4 Composition of the extracts with decarboxylation of the variety “GSC”. Extract

Cannabinoid

50 °C 16.5 MPa

50 °C 24.9 MPa

70 °C 16.5 MPa

70 °C 24.9 MPa

60 °C 20.7 MPa

Crude

CBD Δ9-THC CBN

5.08 ± 1.01 87.91 ± 1.10 0.53 ± 0.59

1.21 ± 0.17 80.80 ± 0.85 0.62 ± 0.04

1.35 ± 0.04 77.00 ± 0.22 0.50 ± 0.07

1.81 ± 0.31 88.29 ± 1.94 1.18 ± 0.41

2.40 ± 0.15 80.82 ± 5.29 0.74 ± 0.02

Winterized

CBD Δ9-THC CBN

2.22 ± 0.74 85.41 ± 4.26 0.96 ± 0.09

1.38 ± 0.14 84.12 ± 6.45 3.56 ± 0.28

1.94 ± 0.39 80.59 ± 0.93 2.18 ± 0.03

4.09 ± 0.70 88.51 ± 1.54 3.71 ± 0.66

2.24 ± 0.10 76.20 ± 3.82 2.54 ± 0.12

Table 5 Composition of the extracts with decarboxylation of the variety “DMII”. Extract

Cannabinoid

50 °C 12.8 MPa

50 °C 24.0 MPa

70 °C 12.8 MPa

70 °C 24.0 MPa

60 °C 18.4 MPa

Crude

CBD Δ9-THC CBN

35.23 ± 6.69 31.10 ± 4.17 0.31 ± 0.02

33.81 ± 0.43 27.96 ± 0.57 0.23 ± 0.13

43.00 ± 3.54 30.45 ± 3.44 0.50 ± 0.29

33.42 ± 2.48 29.42 ± 1.51 0.22 ± 0.08

38.66 ± 3.59 38.72 ± 1.14 0.39 ± 0.01

Winterized

CBD Δ9-THC CBN

29.18 ± 8.66 34.43 ± 2.89 0.33 ± 0.02

29.48 ± 0.94 35.13 ± 2.65 0.32 ± 0.03

43.08 ± 2.79 33.88 ± 3.07 0.55 ± 0.05

32.69 ± 3.12 31.75 ± 3.15 0.48 ± 0.12

33.60 ± 0.98 37.76 ± 1.87 0.62 ± 0.10

Fig. 3. Comparison between the kinetic curves of extractions using different techniques (▲ scCO2 with decarboxylation, ◊ scCO2 with a co-solvent, maximum yield with an organic solvent): (a) 50 °C and 16.5 MPa for the “GSC” variety, (b) 50 °C and 24.0 MPa for the “DMII” variety, models.

Table 6 Sovová parameters of the extractions with ethanol as a co-solvent. Variety

“GSC”

“DMII”

T (ºC) P (MPa) −1 F (g mL ) mF (g min−1) −1 BED (g mL ) r q0 Z W YS (g g−1) tCER (min) tFER (min) KF (min−1) K S (min−1) AARD (%)

50 16.5 0.763 2.198 0.142 8.3·10−1 4.1·10−1 9.6·10−2 7.4·10−2 4.6·10−2 3.1·101 3.5·101 9.0·10−3 8.4·10−3 9.0

24.0 0.846 2.484 0.130 10.0·10−1 3.6 ·10−1 8.1·10−3 2.7·10−1 1.4·10−2 6.1 1.0·101 1.0·10−3 1.2·10−2 1.5

Table 7 Comparative composition of supercritical extracts obtained by different techniques. Cannabinoid

CBD (%) Δ9-THC (%) CBN (%)

Variety: “GSC”

Variety: “DMII”

with ethanol

decarboxylation

with ethanol

decarboxylation

0.00 12.65 0.09

5.08 ± 2.01 87.91 ± 1.10 0.53 ± 0.59

5.57 9.20 0.22

33.81 ± 0.43 27.96 ± 0.57 0.23 ± 0.13

of the flowers of the varieties “GSC” and “DMII” are shown in Tables 4 and 5, respectively. For both varieties, it was observed that the winterization procedure did not promote homogeneous changes in the concentrations of the evaluated cannabinoids. Extracts with high concentrations of Δ9-THC, between 75 to 90%, were obtained from the variety “GSC”. For the “DMII” variety, the obtained extracts had a significantly higher CBD concentration than that of Δ9-THC. This result is highly significant for the treatment of diseases such as epilepsy [25]. Extracts containing similar concentrations of CBD and Δ9-THC were also obtained, which are indicated for others medicinal applications, such as for multiple sclerosis [26,27].

values of the percentage mass yields obtained in the present study and in the work by Rovetto and Aieta [14], even for a flower mass ratio of approximately 1:250. 3.2.3. Composition of the extracts The cannabinoid composition of the crude and winterized extracts 178

Journal of CO₂ Utilization 28 (2018) 174–180

D.R. Grijó et al.

Table 8 Comparative composition of essential oils in the extracts and fresh flowers. Essential oil; retention index m/z reference; m/z experimental

α-Pinene; 937 93;91(44);92(42) 93.02;91.00(45.30);92.04(41.62) β-Pinene; 983 93;41(61);69(36) 93.03;41.03(18.22);69.03(30.28) Mircene; 992 41;93(87);69(80) 40.98(54.02);93.01(100);68.95(81.92) D-Limonene; 1033 68;93(60);67(45) 68.04;93.02(76.53);67.06(78.93) Cariofilene; 1431 93;133(91);91(86) 93.03;133.02(93.99);90.99(77.93) α-Humulene; 1467 93;80(31);121(30) 93.03;80.04(23.62);121.06(30.48) β-Panasinsene; 1549 161;105(60);133(52) 161.04;104.98(50.97);133.05(51.93) Selina-3,7(11)-diene; 1553 161;122(62);107(61) 161,06;122,07(62,71);107,05(53,30)

% “GSC”

“DMII”

in natura

extract

in natura

extract

22.18 ± 2.76

4.79 ± 0.45

27.47 ± 0.08

17.99 ± 1.32

8.15 ± 1.38

2.20 ± 0.08

14.24 ± 0.22

6.08 ± 0.86

30.57 ± 2.21

15.35 ± 2.83

17.99 ± 0.41

12.34 ± 0.50

1.42 ± 0.23

0.81 0.13

9.00 ± 0.15

7.45 ± 0.44

15.18 ± 2.36

31.67 ± 0.28

12.80 ± 0.59

20.97 ± 1.78

4.85 ± 0.80

11.09 ± 0.63

4.35 ± 0.19

8.75 ± 0.26

4.92 ± 1.47

12.62 ± 1.57

2.44 ± 0.38

9.15 ± 1.96

6.69 ± 2.15

16.49 ± 2.10

3.18 ± 0.37

11.17 ± 2.15

3.3. Supercritical extracts obtained with a co-solvent and without decarboxylation

4. Conclusions Supercritical extraction with pure scCO2 and using ethanol as a cosolvent were more selective for the extraction of CBD and Δ9-THC than extraction with an organic solvent mixture. The application of the decarboxylation technique favored the extraction of CBD and Δ9-THC. The extracts obtained in this work presented levels of CBD and Δ9-THC higher than those described in the literature [14,28]. The use of the “winterization” technique did not significantly alter the cannabinoid contents of the extracts. The different chromatographic techniques used were adequate and efficient. The thermal analysis of the flowers was efficient and was able to define the optimal decarboxylation conditions to obtain high levels of the cannabinoids of interest. The Sovová model applied to represent the experimental kinetic curves was adequate and satisfactory.

3.3.1. Extraction kinetics Fig. 3 shows the kinetic curves obtained from the supercritical extractions using ethanol as a co-solvent and by the decarboxylation process under the same conditions of temperature and pressure. The value of the mass yield of each variety obtained using the organic solvent mixture is also provided as a reference. It can be seen that the highest supercritical yields were obtained using ethanol as a co-solvent. The high content of cannabinoid acids and essential oils present in the extracts obtained with a co-solvent may have favored the yield values [9,14]. Table 6 shows the operating conditions of extraction and the parameters calculated using the Sovová model [19]. The solubility values (YS ) of the extracts using ethanol as a co-solvent were significantly higher than for the extracts using the decarboxylation technique. The low values of KF in Table 6 compared to the values in Tables 2 and 3 may be due to the high concentrations of essential oils in the non-decarboxylated samples.

References [1] R.G. Pertwee, Pharmacological and therapeutic targets for Δ9-tetrahydrocannabinol and cannabidiol, Euphytica 140 (2004) 73–82. [2] J.D. Wilkinson, E.M. Williamson, Cannabinoids inhibit human keratinocyte proliferation through a non-CB1/CB2 mechanism and have a potential therapeutic value in the treatment of psoriasis, J. Dermatol. Sci. 45 (2) (2007) 87–92. [3] M. Pucci, C. Rapino, A. Di Francesco, E. Dainese, C. D’Addaria, M. Maccarrone, Epigenetic control of skin differentiation genes by phytocannabinoids, Br. J. Pharmacol. 170 (3) (2013) 581–591. [4] H. Perrotin-Brunel, W. Bujis, J. van Spronsen, M.J.E. van Roosmalen, C.J. Peters, R. Verpoorte, G.J. Witkamp, Decarboxylation of Δ 9-tetrahydrocannabinol: kinetics and molecular modeling, J. Mol. Struct. 987 (2011) 67–73. [5] M. Taschwer, M.G. Schmid, Determination of the relative percentage distribution of THCA and Δ9-THC in herbal Cannabis seized in Austria – impact of different storage temperatures on stability, Forensic Sci. Int. 254 (2015) 167–171. [6] M. Wang, Y.H. Wang, B. Avula, M. Radwan, A.S. Wanas, J. van Antwerp, J.F. Parcher, M.A. ElSohly, I.A. Khan, Decarboxylation Study of Acidic Cannabinoids: A Novel Approach Using Ultra-High-Performance Supercritical Fluid Chromatography / Photodiode Array-Mass Spectrometry, Cannabis, Cannabinoid Res. 1 (1) (2016) 262–271. [7] A.N. Masoud, N.J. Doorenbos, Mississippi grown Cannabis sativa L. III. Cannabinoid and cannabinoid acid content, J. Pharm. Sci. 62 (1973) 313–315. [8] M. Kimura, K. Okamoto, Distribution of tetrahydrocannabinolic acid in fresh wild Cannabis, Experientia 26 (1970) 819–820. [9] L. Romano, A. Hazekamp, Cannabis Oil: chemical evaluation of an upcoming

3.3.2. Composition of the extracts Table 7 shows a comparison of the cannabinoid composition of the winterized extracts obtained with scCO2 for the different techniques used. It can be observed that the extracts with the highest concentrations of CBD and Δ9-THC were obtained with the application of the decarboxylation technique. The extracts obtained with ethanol as a cosolvent had low concentrations of CBD and Δ9-THC [12]. The percentages of the essential oils in fresh flowers and in the extracts obtained with ethanol as a co-solvent are shown in Table 8. There were slight differences in the contents of essential oils in the fresh flowers in both varieties. Higher levels of volatile essential oils were obtained with higher extraction pressures.

179

Journal of CO₂ Utilization 28 (2018) 174–180

D.R. Grijó et al. Cannabis based medicine, Cannabinoids 1 (1) (2013) 1–11. [10] R. Adams, D.C. Pease, C.K. Cain, J.H. Clark, Isomerization of cannabidiol to tetrahydrocannabinol, a physiologically active product. Conversion of cannabidiol to cannabinol, J. Am. Soc. Brew. Chem. 62 (9) (1940) 2402–2405. [11] A. Shani, R. Mechoulam, Cannabielsoic acids. Isolation and synthesis by a novel oxidative cyclization, Tetraedron 30 (15) (1974) 2437–2446. [12] E.R. Garret, J. Tsau, Stability of tetrahydrocannabinols I, J. Pharm. Sci. 63 (10) (1974) 1563–1574. [13] L. Eory, V. Szalay, T. Veress, Supercritical fluid extraction of tetrahydrocannabinol and tetrahydrocannabinol acid, Probl. Forensic Sci. XLVII (2001) 328–332. [14] L.J. Rovetto, N.V. Aieta, Supercritical carbon dioxide extraction of cannabinoids from Cannabis sativa L, J. Supercrit. Fluids 129 (2017) 16–27. [15] H. Perrotin-Brunel, P.C. Perez, M.J.E. van Roosmalen, J. van Spronsen, G.J. Witkamp, C.J. Peters, Solubility of Δ 9-tetrahydrocannabinol in supercritical carbon dioxide: Experiments and modeling, J. Supercrit. Fluids 52 (2010) 6–10. [16] H. Perrotin-Brunel, M.C. Kroon, M.J.E. van Roosmalen, J. van Spronsen, C.J. Peters, G.J. Witkamp, Solubility of non-psychoactive cannabinoids in supercritical carbon dioxide and comparison with psychoactive cannabinoids, J. Supercrit. Fluids 55 (2010) 603–608. [17] N.P. Povh, M.O.M. Marques, M.A.A. Meireles, Supercritical CO2 extraction of essential oil and oleoresin from chamomile (Chamomilla recutita [L.] Rauschert), J. Supercrit. Fluids 21 (2001) 245–256. [18] UNODC, United Nations Office on Drugs and Crime. Recommended Methods for the Identification and Analysis of Cannabis and Cannabis products, (2009). [19] H. Sovová, Rate of the vegetable oil extraction with supercritical CO2–I modeling of extraction curves, Chem. Eng. Sci. 49 (1994) 409–414.

[20] P.J. Linstrom, W.G. Mallard, The NIST Chemistry WebBook: A Chemical Data Resource on the Internet, J. Chem. Eng. Data 46 (5) (2001) 1059–1063. [21] H.D. Dool, P.D. Kratz, A generalization of the retention index system including linear temperature programmed gas—liquid partition chromatography, J. Chromatogr. A 11 (1963) 463–471. [22] A. Bertoli, S. Tozzi, L. Pistelli, L.G. Angelini, Fibre hemp inflorescences: From cropresidues to essential oil production, Ind. Crops Prod. 32 (2010) 329–337. [23] C. Da Porto, D. Decorti, A. Natolino, Separation of aroma compounds from industrial hemp inflorescences (Cannabis sativa L.) by supercritical CO2 extraction and on-line fractionation, Ind. Crops Prod. 58 (2014) 99–103. [24] M.B. King, T.R. Bott, Extraction of Natural Products Using Near-Critical Solvents, first edition, Academic & Professional, Blackie, 1993, p. 21. [25] M. Tzadok, S. Uliel-Siboni, I. Linder, U. Kramer, O. Epstein, S. Menascu, A. Nissenkorn, O.B. Yosef, E. Hyman, D. Graot, M. Dor, T. Lerman-Sagie, B. BenZeev, CBD-enriched medical cannabis for intractable pediatric epilepsy. The current Israeli experience, Seizure 35 (2016) 41–44. [26] R.J.M. Niesink, M.W. van Laar, Does cannabidiol protect against adverse psychological effects of THC? Psiquiatria frontal 4 (2013) 130. [27] A. Molnar, S. Fu, J. Lewis, D.J. Aallsop, J. Copeland, The detection of THC, CBD and CBN in the oral fluid of Sativex® patients using two on-site screening tests and LC–MS/MS, Forensic Sci. Int. 238 (2014) 113–119. [28] J. Omar, M. Olivares, M. Alzaga, N. Etxebarria, Optimisation and characterisation of marihuana extracts obtained by supercritical fluid extraction and focused ultrasound extraction and retention time locking GC-MS, J. Sep. Sci. 36 (8) (2013) 1397–1404.

180