Conversion Characteristics of Some Major Cannabinoids from Hemp (Cannabis sativa L.) Raw Materials by New Rapid Simultaneous Analysis Method

This study was carried out to develop a high-performance liquid chromatography method for short-time analysis of the main cannabinoids in the inflorescence of hemp (Cannabis sativa L.). We also performed decarboxylation of the raw material using our advanced analysis technique. In this study, the UV spectrum was considered to analyze each of the four common cannabinoids, solvents, and samples, where the uniform elution of acidic cannabinoids without peak tailing and acids was tested. Optimal results were obtained when readings were taken at a wavelength of 220 nm using water and methanol containing trifluoroacetic acid as mobile phases in a solvent gradient system. The established conditions were further validated by system suitability, linearity, precision, detection limit, and quantitation limit tests. The decarboxylation index (DT50) confirmed that Δ9-THCA decarboxylated faster than CBDA, and both maintained a linear relationship with time and temperature. In addition, the loss of cannabidiol was better prevented during the decarboxylation process in the natural state than in the extracted state.


Introduction
Linnaeus classified hemp in Europe as a single species (Cannabis sativa L.) [1], and de Lamarck separately classified Indian varieties as Cannabis indica Lam. [2]. Janischevsky classified Russian-born hemp as Cannabis ruderalis Jansich. without any characteristics [3]. Even today, it remains controversial whether all cannabis should be viewed as a single species of C. sativa or as multiple species divided into C. sativa, C. indica, and C. ruderalis [4]. For convenience, it is indicated as 'Cannabis sativa L.'.
Hemp (Cannabis sativa L.) is an annual herb belonging to the Cannabaceae family, first recorded in China for the purpose of obtaining fiber [5]. Since then, it has been exploited for medical purposes for more than 10,000 years [6]. As many countries have approved hemp for medical applications, there has been a surge in the use of hemp drug therapy for pain relief [7]. Hemp contains more than 525 compounds and is known to express approximately 109 cannabinoids [8]. Cannabinoid refers to the C21 terpenophenol compound uniquely found in hemp [9]. Although expressed in small amounts in leaves, cannabinoid is mostly concentrated in the trichome part distributed in the bracts of the female flower of the plant. Its concentration is higher in the stalked-type trichome among the two types, 'sessile' and 'stalked' [10]. They are mostly present in the resin secreted from the trichomes of female plants, while male C. sativa have few glandular trichomes that can In this study, high-performance liquid chromatography (HPLC) was used to efficiently and accurately quantify four major cannabinoids (∆9-THC, ∆9-THCA, CBD, and CBDA) that can be naturally and artificially converted. These compounds exhibit physiological activities. Simultaneous analysis conditions were developed, established, and employed to study the decarboxylation of cannabis inflorescence.

Standard and Sample Preparation
A sample of the inflorescence part of cannabis (Cannabis sativa L.) was received from the Korea Cannabis Genetic Resource Center (Chuncheon, Korea). An inflorescence sample of a plant grown for 105 days (50 days after flowering treatment) in a nutrient solution cultivation system after sowing was used. The standard products cannabidiol (CBD) and Δ9-tetrahydrocannabinol (∆9-THC) were obtained from CAYMAN (Ann Arbor, MI, USA), the cannabidiolic acid (CBDA) standard product was from Cerilliant (Round Rock, TX, USA), and the Δ9-tetrahydrocannabinolic acid (∆9-THCA) standard product was manufactured by Sigma-Aldrich (St. Louis, MO, USA).
The inflorescence of freeze-dried hemp was pulverized using a chopper (YPT-1402, Young Polymer Tech Co., Incheon, Korea) certified by Korea Conformity Laboratories and filtered through a 12-mesh test sieve. The furnace particles were found to be homogeneous. In brief, 0.1 g of hemp sample was sonicated (JEIO TECH Co., Ltd., Daejeon, Korea) for 30 min with 30 mL extra pure grade MeOH (Daejung Chemicals and Metals Co., Ltd., Siheung, Korea) as an extraction solvent. After ultrasonic extraction at room temperature (23 ± 2 °C), the supernatant was sampled, filtered using a 0.45-micrometer syringe filter (Advantec, Tokyo, Japan), and used for HPLC analysis.
All solutions except CBDA in stock state were precisely weighed and then diluted using MeOH. The standard products and samples used in this study were received and In this study, high-performance liquid chromatography (HPLC) was used to efficiently and accurately quantify four major cannabinoids (∆9-THC, ∆9-THCA, CBD, and CBDA) that can be naturally and artificially converted. These compounds exhibit physiological activities. Simultaneous analysis conditions were developed, established, and employed to study the decarboxylation of cannabis inflorescence.

Standard and Sample Preparation
A sample of the inflorescence part of cannabis (Cannabis sativa L.) was received from the Korea Cannabis Genetic Resource Center (Chuncheon, Korea). An inflorescence sample of a plant grown for 105 days (50 days after flowering treatment) in a nutrient solution cultivation system after sowing was used. The standard products cannabidiol (CBD) and ∆9-tetrahydrocannabinol (∆9-THC) were obtained from CAYMAN (Ann Arbor, MI, USA), the cannabidiolic acid (CBDA) standard product was from Cerilliant (Round Rock, TX, USA), and the ∆9-tetrahydrocannabinolic acid (∆9-THCA) standard product was manufactured by Sigma-Aldrich (St. Louis, MO, USA).
The inflorescence of freeze-dried hemp was pulverized using a chopper (YPT-1402, Young Polymer Tech Co., Incheon, Korea) certified by Korea Conformity Laboratories and filtered through a 12-mesh test sieve. The furnace particles were found to be homogeneous. In brief, 0.1 g of hemp sample was sonicated (JEIO TECH Co., Ltd., Daejeon, Korea) for 30 min with 30 mL extra pure grade MeOH (Daejung Chemicals and Metals Co., Ltd., Siheung, Korea) as an extraction solvent. After ultrasonic extraction at room temperature (23 ± 2 • C), the supernatant was sampled, filtered using a 0.45-micrometer syringe filter (Advantec, Tokyo, Japan), and used for HPLC analysis.
All solutions except CBDA in stock state were precisely weighed and then diluted using MeOH. The standard products and samples used in this study were received and tested with the approval of the Seoul Regional Food and Drug Administration (approval number: 1806).
To determine the absorption region of each component, the absorbance spectrum was measured in the UV region of 190-400 nm. The spectra of methanol and acetonitrile were also measured in the same region to avoid the cut-off section of the solvent.

Validation for Establishing Analysis Conditions
To determine the suitability of the developed analysis conditions, tests of the system suitability, specificity, linearity, range, accuracy, precision, detection limit, and quantitation limit were conducted with reference to the ICH guideline Q2 (R1) [22]. The United States Pharmacopeia (USP) was used to determine the suitability of the system.

Investigation of the Correlation between Hemp Inflorescence and Decarboxylation
Samples ground in the same manner as previously described were placed in a preheated, closed, dry oven (Sungchan Science Co., Pocheon, Korea) and sampled at 5-min intervals for 60 min. The temperature conditions were 90, 105, 120, and 135 • C.
To compare the decarboxylation of the extract and the natural product, two different types of cannabis extract were additionally prepared, extracted by applying the above extraction method, and concentrated at 30 • C using a rotary evaporator (Tokyo Rikakikai). The moisture was removed using a freeze-dryer (FD8512, IlShinBioBase Co., Ltd., Dongducheon, Korea). After decarboxylation at 135 • C for 30 min in the absence of any solvent, the product was recovered with MeOH and used for analysis.

HPLC Condition Setting
The absorbance spectrum of the standard solution in the UV region was measured and compared with the mAU value. As a result, ∆9-THC, ∆9-THCA, CBD, and CBDA had their maximum absorbance at 210.50, 221.24, 209.09, and 221.51 nm, respectively ( Figure 2). Cannabinoids in the form of acidic precursors had relatively low absorbance in the UV region. The cut-off section of the mobile phase allowed us to judge whether it was easy to analyze at the selected wavelength ( Figure 3). Acetonitrile and HPLC water showed an absorbance lower than or similar to that of the baseline at 220 nm, and methanol had a slightly higher absorbance than that of the baseline. This observation was consistent with that reported by Welch et al. [23]. We chose 220 nm as the optimal wavelength and acetonitrile as the organic mobile phase to facilitate sensitive peak detection, solvent absorption, and simultaneous analysis.
Peak tailing tends to lower the peak height and increase the limit of detection (LOD) and limit of quantitation (LOQ). As the area of the peak cannot be accurately determined, the reliability of quantification is automatically reduced. In the case of ∆9-THCA, which was eluted at the end, tailing occurred due to diffusion in the column and unstable kurtosis; therefore, the flow rate was set to 1.5 mL/min.
The organic acids 0.05% TFA, 0.10% PA, and 0.10% FA were used under gradient conditions, and the gradient eluent set to simulation was used as the elution method to increase the elution rate (initiation, 70% ACN; 2 min, 70% ACN; 8 min, 85% ACN; 9 min, 70% ACN; 12 min, 70% ACN). FA and TFA showed a raised baseline ( Figure 4). However, TFA was selected as the optimal organic acid because PA is not very compatible with mass spectrometry (MS), such as liquid-chromatography-mass-spectrometry (LC-MS) or liquid chromatography quadrupole time-of-flight (LC-Q-ToF).
The chromatography conditions were as follows: mobile phase A was water containing 0.05% TFA and mobile phase B was ACN containing 0.05% TFA. The gradient conditions were as follows: 70% B initially, 70% B for 2 min, 85% B for 8 min, 70% B for 9 min, and 70% B for 12 min. All four cannabinoids were measured at 220 nm using a UV-Vis detector. The flow rate was 1.5 mL/min. The injection volume was 10 µL, and the column oven temperature was set to 35 • C to prevent the destruction of acid-type cannabinoids and to reduce the retention time.     Peak tailing tends to lower the peak height and increase the limit of detection (L and limit of quantitation (LOQ). As the area of the peak cannot be accurately determ the reliability of quantification is automatically reduced. In the case of Δ9-THCA, w Molecules 2021, 26, x FOR PEER REVIEW was eluted at the end, tailing occurred due to diffusion in the column and unstab tosis; therefore, the flow rate was set to 1.5 mL/min.
The organic acids 0.05% TFA, 0.10% PA, and 0.10% FA were used under g conditions, and the gradient eluent set to simulation was used as the elution me increase the elution rate (initiation, 70% ACN; 2 min, 70% ACN; 8 min, 85% ACN 70% ACN; 12 min, 70% ACN). FA and TFA showed a raised baseline (Figure 4). Ho TFA was selected as the optimal organic acid because PA is not very compatible wit spectrometry (MS), such as liquid-chromatography-mass-spectrometry (LC-MS) o chromatography quadrupole time-of-flight (LC-Q-ToF). The chromatography conditions were as follows: mobile phase A was water c ing 0.05% TFA and mobile phase B was ACN containing 0.05% TFA. The gradient tions were as follows: 70% B initially, 70% B for 2 min, 85% B for 8 min, 70% B for and 70% B for 12 min. All four cannabinoids were measured at 220 nm using a U detector. The flow rate was 1.5 mL/min. The injection volume was 10 μL, and the c oven temperature was set to 35 °C to prevent the destruction of acid-type cannab and to reduce the retention time.

System Suitability Test
A standard stock solution was prepared, and four types of standard products were mixed and measured nine times at the same concentration. The retention time did not exceed 2.00% of the relative standard deviation in % (RSD%) values for all four cannabinoids, Molecules 2021, 26, 4113 6 of 11 and the peak area was less than 1.00%. The tailing factor was less than 2.0, and all theoretical plates had values of 1000 or more, thus meeting all USP standards. For resolution, CBDA, CBD, ∆9-THC, and ∆9-THCA with similar retention times were compared, thus showing excellent resolution (Table 1).

Linearity
To test linearity, standard products were prepared at concentrations of 20, 40, 80, 160, 320, and 640 µg/mL, and calibration curves were measured using newly developed analysis methods. The coefficient of determination (R 2 ) of the evaluated calibration curve was as follows: CBDA, R 2 = 0.9916; CBD, R 2 = 0.9993; ∆9-THC, R 2 = 0.9944; and ∆9-THCA, R 2 = 0.9994. Thus, we observed excellent linearity above R 2 ≥ 0.9900. However, in the case of CBDA and ∆9-THC, the R 2 values were 0.9999 and 0.9986, respectively, when the maximum concentration of 640 µg/mL was removed. This observation was associated with the high concentration, which violates the Beer-Lambert law because of the characteristics of the calibration curve. In the case of quantitative analysis, CBDA and ∆9-THC should be analyzed by excluding 640 µg/mL ( Figure 5).

Evaluation of Industrial Decarboxylation Efficiency of Hemp Inflorescence
Heat is a major factor in cannabis decarboxylation [24]. Therefore, we analyzed the decarboxylation graph according to temperature and heating time using the newly established simultaneous analysis conditions. Figure 6 shows the decrease in the acidic form of cannabinoid and the increase in its neutral form. This decarboxylation reaction was accelerated as the temperature and time increased and showed a pattern similar to that reported in previous studies [21,25].

Linearity
To test linearity, standard products were prepared at concentrations of 20, 40, 80, 160, 320, and 640 μg/mL, and calibration curves were measured using newly developed analysis methods. The coefficient of determination (R 2 ) of the evaluated calibration curve was as follows: CBDA, R 2 = 0.9916; CBD, R 2 = 0.9993; Δ9-THC, R 2 = 0.9944; and Δ9-THCA, R 2 = 0.9994. Thus, we observed excellent linearity above R 2 ≥ 0.9900. However, in the case of CBDA and Δ9-THC, the R 2 values were 0.9999 and 0.9986, respectively, when the maximum concentration of 640 μg/mL was removed. This observation was associated with the high concentration, which violates the Beer-Lambert law because of the characteristics of the calibration curve. In the case of quantitative analysis, CBDA and Δ9-THC should be analyzed by excluding 640 μg/mL ( Figure 5). Hemp without additional processing also undergoes decarboxylation in its natural state owing to the effects of environmental factors such as light, and also generates neutral cannabinoids [26]. We selected the DT 50 parameter (when acid cannabinoids are reduced by half; half-life) to accurately evaluate the rate of decarboxylation by removing the cannabinoid factors that are oxidized during heating and naturally generated neutral cannabinoids. DT 50 was not achieved at 90 • C in either CBDA or ∆9-THCA after 60 min. The DT 50 was 44.4 for CBDA and 39.1 for ∆9-THCA at 105 • C, 32.6 for CBDA and 27.3 for ∆9-THCA at 120 • C, and 25.8 for CBDA and 21.1 for ∆9-THCA at 135 • C. As for decarboxylation, ∆9-THCA had a DT 50 value of approximately 5.1 ± 0.3 at the same temperature as CBDA, suggesting that ∆9-THCA decarboxylated faster than CBDA did under the same conditions (Figure 7).
The purpose of this study was to compare the loss of neutral cannabinoids from natural products and extracts in response to heating at a maximum temperature of 135 • C for 30 min using randomly selected CW21-13 and RDA-35 varieties. The DT 50 values of the acidic cannabinoids were not noticeably different for both extracts and natural products. Thus, neither the extract nor the natural product had any effect on the decarboxylation rate.
Heat is a major factor in cannabis decarboxylation [24]. Therefore, we analyzed the decarboxylation graph according to temperature and heating time using the newly established simultaneous analysis conditions. Figure 6 shows the decrease in the acidic form of cannabinoid and the increase in its neutral form. This decarboxylation reaction was accelerated as the temperature and time increased and showed a pattern similar to that reported in previous studies [21,25]. Hemp without additional processing also undergoes decarboxylation in its natural state owing to the effects of environmental factors such as light, and also generates neutral cannabinoids [26]. We selected the DT50 parameter (when acid cannabinoids are reduced by half; half-life) to accurately evaluate the rate of decarboxylation by removing the cannabinoid factors that are oxidized during heating and naturally generated neutral cannabinoids. DT50 was not achieved at 90 °C in either CBDA or Δ9-THCA after 60 min. The  Upon incubation for 30 min at 135 • C, the yield of neutral cannabinoids increased in CW21-13. The CBD content of the raw material and extract was 33,786 and 10,046 µg/mL (about 3.4 times), respectively, and the ∆9-THC content was 956 and 235 µg/mL (about 4.0 times), respectively. In the case of RDA-35, the CBD and ∆9-THC contents of the raw material were 19,756 and 792 µg/mL, respectively, and those of the extract were 6306 µg/mL (about 3.1 times) and 155 µg/mL (about 5.1 times), respectively. Thermal destruction of neutral cannabinoids in the raw material was significantly lower in both varieties ( Figure 8). Even during visual evaluation, the neutral cannabinoid levels increased over time in the extract and decreased in the second half. On the other hand, the natural product had a pattern of increase or did not continue to decrease even at 30 min, the end-point of the experiment. DT50 was 44.4 for CBDA and 39.1 for Δ9-THCA at 105 °C, 32.6 for CBDA and 27.3 THCA at 120 °C, and 25.8 for CBDA and 21.1 for Δ9-THCA at 135 °C. As for decar tion, Δ9-THCA had a DT50 value of approximately 5.1 ± 0.3 at the same tempera CBDA, suggesting that Δ9-THCA decarboxylated faster than CBDA did under th conditions (Figure 7). The purpose of this study was to compare the loss of neutral cannabinoids fro ural products and extracts in response to heating at a maximum temperature of 13 30 min using randomly selected CW21-13 and RDA-35 varieties. The DT50 values of the acidic cannabinoids were not noticeably different for both extracts and natural products. Thus, neither the extract nor the natural product had any effect on the decarboxylation rate. Upon incubation for 30 min at 135 °C, the yield of neutral cannabinoids increased in CW21-13. The CBD content of the raw material and extract was 33,786 and 10,046 μg/mL (about 3.4 times), respectively, and the Δ9-THC content was 956 and 235 μg/mL (about 4.0 times), respectively. In the case of RDA-35, the CBD and Δ9-THC contents of the raw material were 19,756 and 792 μg/mL, respectively, and those of the extract were 6306 μg/mL (about 3.1 times) and 155 μg/mL (about 5.1 times), respectively. Thermal destruction of neutral cannabinoids in the raw material was significantly lower in both varieties ( Figure  8). Even during visual evaluation, the neutral cannabinoid levels increased over time in the extract and decreased in the second half. On the other hand, the natural product had a pattern of increase or did not continue to decrease even at 30 min, the end-point of the experiment.

Conclusions
We have successfully developed a fast and reliable analytical method using the widely employed HPLC-UV/Vis system. The validation tests demonstrated the high suitability of the method. The decarboxylation process of hemp was analyzed using this chromatographic method, which successfully isolated four major cannabinoids within 10 min. The overall decarboxylation rate may vary depending on factors such as the volume of the reactor due to the nature of the closed oven [27]. However, Δ9-THCA occurred more rapidly at a lower temperature than CBD did in natural products under the same conditions. The thermal destruction of cannabinoids occurred more in extracts than in natural

Conclusions
We have successfully developed a fast and reliable analytical method using the widely employed HPLC-UV/Vis system. The validation tests demonstrated the high suitability of the method. The decarboxylation process of hemp was analyzed using this chromatographic method, which successfully isolated four major cannabinoids within 10 min. The overall decarboxylation rate may vary depending on factors such as the volume of the reactor due to the nature of the closed oven [27]. However, ∆9-THCA occurred more rapidly at a lower temperature than CBD did in natural products under the same conditions. The thermal destruction of cannabinoids occurred more in extracts than in natural products. The results of this study are of great value. This method allows for quick quantification and selection of the main cannabinoids in hemp breeding for industrial applications. In addition, it provides justification about the various processes that occur during the decarboxylation of hemp in its natural state. Funding: This research did not receive any specific grant.
Institutional Review Board Statement: Not applicable.

Informed Consent Statement:
The standard products and samples used in this study were received and tested with the approval of the Seoul Regional Food and Drug Administration (approval number: 1806).
Data Availability Statement: Not applicable.