Thermo-chemical conversion kinetics of cannabinoid acids in hemp (Cannabis sativa L.) using pressurized liquid extraction | Journal of Cannabis Research

On Aug 13, 2024

Decarboxylation studies

The concentrations of individual cannabinoids extracted from hemp were calculated over five different temperatures (80, 100, 120, 140, or 160 °C). For each temperature, a series of thermo-chemical conversions were performed over time points of 1, 3, 5, 10, 20, 40, 60, and 90 min. After experiments, the concentrations (mg/g) were plotted as a function of time (min) and temperature (°C) (Fig. 2). A relatively low conversion of all acidic cannabinoids into their respective neutral form was found at 80 °C, while an appreciably increase in conversion was shown in accordance with increasing temperature. However, at an elevated temperature of 160 °C the concentrations of all neutral forms started decreasing after 30 min of exposure likely resulting from degradation of the neutral cannabinoids. For instance, the CBGA to CBG and CBDA to CBD decarboxylation at 140 °C temperature, continued to show maximum concentration after 20 min. The CBDA to CBD decarboxylation was in good agreement with the earlier published report for this transformation (Olejar and Kinney, 2021). At a temperature of 140 °C CBD was maximum for over 20 min and almost remained constant throughout the 90 min, however at 160 °C the concentration of CBD started decreasing after 45 min. We predicted the reason for this to be the thermolability of cannabinoids resulting in unknown degradations products due to the increased temperature. Similar results were observed in the case of CBGA to CBG and CBCA to CBC. Interestingly, at elevated temperature (160 °C, or 140 °C in some) CBDA, CBGA and CBCA were completely decarboxylated in less than 1 min. This was attributed to the extraction cell requiring heating prior to the introduction of water (Fig. 2). At 140 °C, the maximum CBDV concentration resulted at 10 min. CBN accumulation was noted at higher temperatures, which is a well-known oxidation product of THC, but could not totally account for the observed losses of THC.

According to earlier literature the decrease in total cannabinoid concentration at elevated temperatures is a common observation. A loss of 60% in molar concentration of CBD_total at 120 °C, during hemp seed oil decarboxylation in open reactors was observed by Citti et al. (2018). Wang et al. has mentioned the unexplained decrease in THC, CBG and CBD forms in cannabis extracts when decarboxylation was carried out at 145 °C in a vacuum oven (Wang et al., 2016). Veress et al. (1990) also observed the same for CBD and THC at 122 and 145 °C attributing the losses to evaporation. According to a study of cannabis resin over several years under different storage conditions, the increase in total CBN concentration does not correspond to the decrease in total THCA + THC, explaining that degradation of THC can occur into other unknown compounds (Lindlost, 2010; Trofin et al., 2012). Another study on cannabis plant material under the effect of storage temperature disclosed that temperatures of 100 °C and above could lead to an accelerated THCA decarboxylation process followed by fast and rapid loss in THC. Likewise, Moreno et al., has also observed that the total molar concentration of CBD and CBG (sum of their acidic and neutral forms) plummeted by 90%, after 60 min. at 160 °C (Moreno et al., 2020). It was predicted to be the result of unidentified side products formation along with evaporative losses at higher temperature since the boiling point of CBG, THC and CBD lie in the range of 120—180 °C. During the kinetics analysis of thermo-chemical conversion in PLE, such losses in concentrations turned out to be insignificant. The sigmoidal curves for the molar concentrations of acid and neutral cannabinoid, came out equal and opposite as illustrated (Fig. 3). This fact was further confirmed when the combination of acid and neutral cannabinoid molar concentrations were plotted, and the result was almost a linear line through the time period, ruling out any possibility of a rise or drop in total concentration.

Kinetic Model

To understand the kinetics of the thermo-chemical conversion of cannabinoid acids to neutral cannabinoids a series of experiments examining cannabinoid contents following the conversion process were undertaken. Each series of experiments involved examining the concentration of the cannabinoid and its corresponding acid over a time range at a selected temperature. Utilizing this data, the reaction order was established from the graphical plots of this data. Considering the reaction matrix or simple model for CBDVA, CBDA, CBGA, and CBCA decarboxylation given in Scheme 1, where k’s are the rate constant for that cannabinoid decarboxylation.

The reaction matrix (Scheme 1) can be simplified into the Eqs. (1–4) given below.

$$\frac{d[CBDVA]}{dt}=k1[CBDV]$$

(1)

$$\frac{d[CBDA]}{dt}=k2[CBD]$$

(2)

$$\frac{d[CBGA]}{dt}=k3[CBG]$$

(3)

$$\frac{d[CBCA]}{dt}=k4[CBC]$$

(4)

Upon integration, the above Eqs. (1–4) can be converted into the generalized Eq. 5

$$ln\frac{[Co]}{[Ct]}=kt$$

(5)

where [C₀] and [C_t] stand for the acidic cannabinoid concentration at time 0 and t minutes, respectively. The concentrations following extraction on non-decarboxylated hemp exclusively are shown by the symbol [C₀] i.e. when decarboxylation time is zero. The rate order was established by plotting $ln\frac{[Co]}{[Ct]}$ vs time (Fig. 4). Each plot was examined for its linearity. A linear line was obtained from the plot of these values and from the resulting lines equations the value of k is determined. The graphical representation of the data is displayed (Fig. 4) and the extracted linear regressions and equations of best-fit line are expressed (Table 1). It should be noted from this data that at the higher temperatures 160 °C and occasionally 140 °C the reaction goes to completion almost immediately. In instances when a representative line, of at least three data points, is not available the temperature was excluded from further calculations. The reaction order was found to be a pseudo-first order reaction and the rate constant, k, was found to be equal to the slope of this line. A pseudo-first order reaction is defined as a reaction that appears to be first order: however, one reactant is typically found in gross excess so its change in concentration is negligible, or one reactant is a catalyst. Since thermo-chemical conversion utilizes water in excess compared to cannabinoid content therefore the reaction is considered a pseudo-first order reaction.

Table 1 Cannabinoid reaction order determination, providing the coefficient of correlation and equation of the best-fit line

This result was not unexpected as initially thermal decarboxylation was considered first order (Veress et al., 1990). Later it was discovered that this reaction was catalyzed by formic acid and it is now considered a pseudo-first order reaction (Perrotin-Brunel et al., 2011). However, in both these instances there is a large degree of loss, which results from evaporation of the cannabinoids and creates a complex kinetics model where degradation of the neutral cannabinoids must also be factored in with the evaporation of the cannabinoids. Furthermore, the value of 1/T, where T is the temperature expressed in °K, and ln(k) can be established. Once established these values may be plotted as ln(k) versus (1/T) following the Arrhenius equation to calculate the activation energy, E_a, and the frequency or pre-exponential factor, A (Fig. 5). The value of the frequency constant, A, was established by obtaining the value of y when x = 0 using Eq. 6 and Eq. 7.

where y = ln(k), m is the slope, x = (1/T), and c is the x-intercept.

$$\text{ln}k=\text{ln}A-\frac{{E}_{a}}{RT}$$

(7)

where k is the reaction rate constant, A is the frequency factor, E_a is the activation energy, T is temperature in °K, and R is the universal gas constant (8.3144 J K⁻¹ mol⁻¹). The obtained value of the x-intercept is the ln(A) and therefore Eq. 8 must be applied.

The slope of the line is equal to ($\frac{-{E}_{a}}{R}$), where R is the gas constant and therefore must be converted to obtain the activation energy E_a. The calculated rate constants, activation energies and frequency constant are included (Table 2).

Table 2 Rate constants, activation energies and pre-exponential factors for all acidic cannabinoids studied

Expanding beyond this it is possible to calculate a mass balance corrected model Scheme 2).

Where Xⁱ is an unknown precursor, Yⁱ is an unknown degradation product, and k are reaction rate constants.

Once the k, E_a and A values are established, using Scheme 1, a simple model of thermo-chemical conversion was generated. Unfortunately, this model does not consider other factors that contribute to the formation of neutral cannabinoids or degradation.

As such, these equations must be applied to obtain values of k for Scheme 2. From Scheme 2, a more complete model of thermo-chemical conversion was examined. Through this process k_2a and k_2b were found to negligible in the cannabinoids studied at the temperatures and times where maximum extraction was expected to occur. Consequently, the simple model was used for the predictive modeling of the cannabinoid thermo-chemical conversion and extraction.

The simple generated model further determined the favorable conditions leading to the maximum concentration of the desirable cannabinoid. The maximum concentration of CBDV, CBD, CBG, and CBC according to the model can be attained at each temperature, along with the predicted time (Fig. 6). The prediction graphs suggest that the thermo-chemical conversions using PLE are faster and take less time as compared to oven decarboxylation as reported by Moreno and others (Moreno et al., 2020). For instance, in the case of CBDA at 140 °C, the maximum concentration can be achieved in 6.3 min through thermo-chemical conversion, while thermal decarboxylation required 27 min at same temperature, based on the mass balance model of Moreno et al. (2020). According to the mass balance model of Moreno, the optimum conditions to get the maximum CBD is preferably at lower temperature (80 °C) with a longer time period (25 h), since at lower temperature the decomposition reaction is minimized for the associated higher activation energy of CBDA-CBD decarboxylation. However, in industry decarboxylation and isolation time are critical not only for being able to produce product but also to minimize operating expenses, such as power and labor. Consequently, thermo-chemical decarboxylation can be far superior.

Prediction Model Verification

After the model was generated, it was verified for CBD and CBG maximum extraction concentration at a given temperature in two cannabis varieties. These two cannabinoids were chosen due to existing research and economic interests. Because the bench-top PLE does not perform fractions of a minute, times were rounded to the closest whole minute. The selected conditions for CBD were thermo-chemical conversion at 130 °C for 9 min using water as the solvent, followed by extraction with ethanol at 120 °C, for 2 static cycles of 3 min each. Similarly, the conditions for CBG were thermo-chemical conversion at 130 °C for 14 min followed by ethanol extraction at 120 °C for 2 static cycles of 3 min each.

The comparison of the means of triplicate processes of the control conditions and the model-generated conditions were tabulated (Table 3). It should be noted that while the model performs well for cannabinoids that are in larger quantity, it begins to falter when these compounds are in trace amounts. This is evidenced by the low predictions for the CBG in CBD rich hemp and the negative concentration of the CBD in CBG rich hemp. The reason for both errors is that the model’s rate constants for the specified temperature and reaction time are higher than the cannabinoids concentrations present in the hemp. As a result, in both cases (CBD and CBG rich hemp), the extracted cannabinoid from the control and the model conditions, exceed the prediction.

Table 3 CBD and CBG model verification using CBD and CBG rich hemp varieties

The models performed as expected for the major cannabinoids in each variety. The CBD model predicts the CBD values for the control, which the model was based on, and the model conditions. While the CBG model predicts the values for the control well, the model conditions outperform the prediction. This discrepancy may be due to the model being generated from a variety of hemp that is not high in CBG or there may be an unknown precursor forming CBG under the predicted conditions. The latter was not expected, as there was no evidence of a precursor during the model development. Furthermore, it is not expected that the hemp variety used will affect the kinetics of the process; however the kinetics of the cannabinoids in the associated cannabinoid-rich matrix may differ based on concentration. Consequently, further studies should be done to elucidate the relationship of cannabinoid concentration to k values.

Source link