According to the literature, the cyclization reaction of CBD seems to occur following an acid-catalyzed activation of a specific double bond.^9,10 A dihydrobenzopyran ring moiety is formed by internal ether formation of one of the phenolic groups with one of the double bonds. The two double bonds in the CBD structure are responsible for the formation of two different compounds (Scheme 1). If the activation occurs on the Δ⁸ double bond, the products show the THC scaffold (Δ⁹-THC, path b); otherwise, the Δ¹ double bond activation leads to the formation of the iso-THC scaffold (Δ⁸-iso-THC, path a). The latter cyclization is much less frequent. However, acidic conditions are responsible for further isomerization toward the corresponding thermodynamically more stable compounds, Δ⁸-THC and Δ⁴⁽⁸⁾-iso-THC, respectively.¹¹

Although Δ⁹-THC and its derivatives have been widely explored and recognized as the major psychoactive Cannabis constituents, the iso-THC isomers have received little attention. For this reason, we wish to fill the literature gap, especially regarding the provision of full NMR data.

To investigate the susceptibility and selectivity of CBD cyclization, different reactions, including the use of Lewis and protic acids in different solvents and varying the temperature and reaction time, were performed (Scheme 2).

The Lewis acids were evaluated first, starting with the recorded use of BF₃·OEt₂.^12,13 The data suggest that performing the reaction with BF₃·OEt₂ in CH₂Cl₂ at a low temperature affords Δ⁹-THC as the main product, but increasing the temperature and reaction time results in preferential formation of the more stable Δ⁸-THC. The results support this assertion (Table 1, entries 1 and 2). Lowering the temperature also lowers the yields (Table 1, entry 3). Using other solvents gave different degrees of selectivity. In particular, toluene gave results similar to those in CH₂Cl₂, but iso-THCs always accompanied the Δ⁸- and Δ⁹-THCs (Table 1, entries 4 and 5). A reaction conducted in MeCN at −10 °C for 6 h yielded Δ⁸-iso-THC as the main product accompanied by trace amounts of Δ⁴⁽⁸⁾-iso-THC (Table 1, entry 7).

To enhance the yields and the selectivity of the process, a series of tests with different acids were conducted, following the hypothesis that other Lewis acids could actively induce cyclization. Starting from the positive literature results regarding the use of TMSI (trimethylsilyl iodide), which showed a high yield of Δ⁹-THC formation without isomerization,¹⁴ TMSOTf (trimethylsilyl triflate) was tested as an acidic reagent. Contrary to the expectations, it displayed a high affinity for the formation of Δ⁸-THC when CH₂Cl₂ or toluene were used as solvents, even at a low temperature (Table 1, entries 8 and 9).

In(OTf)₃ in CH₂Cl₂ converted CBD into Δ⁹-THC at a low temperature in a better yield than that of BF₃·OEt₂ (Table 1, entry 10). As in previous tests, higher temperatures caused the production of the thermodynamically more stable isomer in higher yields (Table 1, entry 11). Using toluene, the selectivity shifted to the formation of Δ⁸-THC in excellent yields (Table 1, entry 13). The use of ZnBr₂ in CH₂Cl₂ did not promote the cyclization reaction even at room temperature (Table 1, entry 14), while TiCl₄ showed a trend similar to that of BF₃ (Table 1, entry 15). The activity of AlCl₃, AgOTf, and Ti(OiPr)₄ was also investigated, without any noteworthy results. Considering these outcomes, a unique preferential formation path for any of the possible isomers cannot be determined based on the characteristics of the Lewis acid used to induce the cyclization.

Subsequently, protic acid screening was performed (Table 2). The best results for CBD conversion were obtained with HCl, pTSA (p-toluenesulfonic acid), and CSA (camphorsulfonic acid). As reported,⁹ pTSA in CH₂Cl₂ led directly to the formation of Δ⁸-THC as the sole product (Table 2, entry 2). The nature of the solvent clearly affected the reaction outcome. The reaction in n-hexane afforded a mixture of Δ⁹-THC, Δ⁸-THC, and Δ⁸-iso-THC in a ratio of 1:5:1 (Table 2, entry 3), while the reaction in toluene gave a higher selectivity. pTSA gave different isomers in different percentages depending on the solvent and reaction time (Table 2, entries 2–5). The best selectivity of Δ⁹-THC and Δ⁸-THC formation was obtained with toluene and CH₂Cl₂, respectively. On the contrary, the use of 10% mmol catalytic amounts of acid in toluene resulted in almost complete isomerization of the double bond because of the increased reaction time that shifted the outcome to the thermodynamic isomer (Table 2, entry 6). Interestingly, CSA promoted the cyclization of CBD to Δ⁹-THC with complete selectivity and satisfactory yields regardless of the reaction time (Table 2, entry 7). Other protic acids gave worse results for the CBD conversion (Table 2, entries 8–15).

Some Δ⁴⁽⁸⁾-iso-THC formation was detected in three cases (Table 1, entries 5 and 7; Table 2, entry 8), and this compound was isolated and characterized.

The experimental results indicate that toluene is the most suitable solvent for the conversion of CBD into THC isomers. This solvent particularly affects the selectivity of the isomers according to the other experimental conditions (the reaction temperature and nature of the acid cyclization promoter). Increasing the temperature reduces the selectivity of the activation of the double bond and favors the formation of the corresponding most stable isomers. The Lewis acids BF₃·OEt₂, In(OTf)₃, and TMSOTf have a proven effect and affected the major formation of Δ⁸-THC and the product mixtures. As for protic acids, pTSA promotes the reaction to selectively afford Δ⁹- and Δ⁸-THCs, depending on the reaction time. CSA emerges as an interesting cyclization inducer, giving Δ⁹-THC in good yields through readily accessible reaction conditions.

With these encouraging screening results, the influence of CSA was more thoroughly investigated; therefore, the solvent, temperature, and time were considered variables. Using toluene as the solvent at room temperature gave a 61% yield of Δ⁹-THC accompanied by unreacted CBD (Table 2, entry 7). A longer reaction time led to isomerization of Δ⁹-THC and to decomposition of compounds (Table 3, entry 3). Increasing the temperature led to the formation of a mixture that was enriched with the Δ⁸ isomer over time (Table 3, entries 4–8). Increasing the temperature further drastically reduced the CBD conversion time and isomerization, obtaining Δ⁹-THC as a kinetic reaction product (Table 3, entries 9 and 10). In CH₂Cl₂, the reaction was faster and less selective toward Δ⁹-THC formation (Table 3, entries 11–13). n-Hexane and MTBE (t-butyl methyl ether) induced a high degree of CBD conversion without a marked preferential selectivity for the formation of a THC isomer even at a short reaction time (Table 3, entries 14 and 15). In all experiments, iso-THC isomers were not detected except when the reaction was performed in MTBE (Table 3, entry 15).

Toluene showed the best selectivity; however, the long reaction time seemed to be a drawback. CH₂Cl₂ appeared to be promising in this respect, but continuous monitoring of the reaction was required to avoid the prevalence of isomerization.

Δ⁹-THC, Δ⁸-THC, Δ⁸-iso-THC, and Δ⁴⁽⁸⁾-iso-THC were fully characterized using NMR data, and the complete ¹H and ¹³C NMR assignments (Table 4–6) have been determined on the basis of 1D and 2D NMR spectra (¹H and ¹³C NMR, correlation spectroscopy (COSY), heteronuclear single quantum coherence (HSQC), and heteronuclear multiple bond correlation (HMBC)). The data were compared with those available in the literature.^15,16

Diagnostic and distinguishable NMR peaks permit identification of compounds derived from intramolecular cyclization within the crude reaction mixture and also allow the composition percentage to be determined from integration ratios (Figure 2). In CDCl₃, Δ⁹-THC is characterized by the presence of signals at 6.34 ppm (H-10), 3.23 ppm (H-10a), 2.22–2.16 ppm (H-8), and 4.88 ppm (OH). The ¹H NMR data of Δ⁸-THC show two signals for H-10 (3.21 and 2.19–2.15 ppm), while the signal due to H-10a is present at 2.71 ppm. The olefinic (H-8) and the hydroxy proton appear at 5.45 and 4.63 ppm, respectively. For Δ⁸-iso-THC, the corresponding characteristic signals are the doublet at 4.98 ppm (H-9) and the two signals at 3.49 and 2.37 ppm that match the protons H-3 and H-4, respectively. The spectrum of Δ⁴⁽⁸⁾-iso-THC shows a characteristic signal at 4.29 ppm (H-3).

On the basis of the literature data,¹⁷ an HPLC method was used to follow the cyclization of CBD. The reactions monitored via HPLC provided a composition of the reaction mixture comparable with that of the ¹H NMR data analysis.

The analysis was performed on an ASCENTIS RP-C₁₈ column (5 μm × 4.6 × 150 mm). The pressure was set at 101 bar, and the temperature was maintained at 40 °C with a constant flow rate of 0.95 mL/min. UV spectra were recorded at 228.8 nm using a gradient elution method. The mobile phase consisted of a mixture of A (0.1% v/v HCOOH in H₂O) and B (0.1% v/v HCOOH in MeCN). The gradient elution program was adapted to a 30 min duration to obtain RRT 1.00 for CBD and RRT 1.28 for Δ⁹-THC. After 30 min, the column was purged with 100% B in 7 min; subsequently, the system was washed under these conditions for 3 min and restored to the initial conditions. The retention times were CBD, 23.63 min; Δ⁴⁽⁸⁾-iso-THC, 29.62 min; Δ⁹-THC, 29.92 min; Δ⁸-THC, 30.77 min; and Δ⁸-iso-THC, 30.77 min (Figure 3).

The method allowed an excellent separation of CBD from the THC isomers; in particular, it was possible to recognize Δ⁹-THC from Δ⁸-THC and Δ⁸-iso-THC; however, the peaks were quite close. The drawback was that it remained difficult to obtain a better resolution between the peaks of Δ⁹-THC/Δ⁴⁽⁸⁾-iso-THC, and Δ⁸-THC/Δ⁸-iso-THC, which have similar retention times. For this reason, the HPLC results were always compared with those obtained from the ¹H NMR data.

In conclusion, all THC isomers were fully characterized via ¹H and ¹³C NMR spectroscopy. An analytical method was optimized to monitor the course of the reactions. In particular, it was found that CSA in toluene at room temperature (RT) for 96 h and pTSA in toluene for 48 h at RT were the best conditions for the selective formation of Δ⁹-THC. TMSOTf in CH₂Cl₂ at −10 °C for 6 h, In(OTf)₃ in toluene at 0 °C for 24 h, pTSA in CH₂Cl₂ at RT for 36 h, and CSA in toluene at 40 °C for 96 h selectively afforded Δ⁸-THC in high yields. The use of BF₃·OEt₂ in toluene led to the formation of the iso-THC isomer, depending on the reaction temperature. At −10 °C, a separable mixture of Δ⁹-THC and Δ⁸-iso-THC was obtained, whereas a temperature increase to 0 °C shifted the result toward the corresponding most stable isomers, Δ⁸-THC and Δ⁴⁽⁸⁾-iso-THC. CBD is a challenging substrate that permits the chemical reactivity of natural alkenes and phenols to be addressed and exploited.