The high-performance liquid chromatography (HPLC) method was used to determine the peak area of each standard and to generate the calibration curve. The mean regression equation for the concentration range of 5–100 µg/mL of CBD in methanol was run in triplicate, y = 80,926x – 10,018 (R² = 0.9992, n = 7), where x represents the concentration and y represents the peak area of CBD.

The results from the liquid chromatography–mass spectrometry (LC-MS) method were obtained from eight concentration points ranging from 5 ng/mL to 1000 ng/mL, with six replicates (n = 6) for each point. The calibration curve followed the equation y = 101,766x + 398,014, demonstrating strong linearity with an R² value of 0.9994. The limit of detection (LOD) was determined to be 1 ng/mL, while the limit of quantification (LOQ) was 5 ng/mL, and the retention time was 6.26 ± 0.03 min.

To ensure the reliability of the developed liquid chromatography–mass spectrometry (LC-MS) method, we assessed its intra-day accuracy and precision. The intra-day accuracy was evaluated by analysing multiple replicates of the samples under identical conditions within a single day, demonstrating consistent performance across measurements (Table 1). Additionally, the precision of the method was determined by assessing the repeatability and reproducibility of the obtained results, confirming the method’s robustness (Table 2). These findings validate the suitability of the LC-MS method for the accurate quantification of the target analytes.

Initial measurements of the CBD-PLC formulation showed a particle size of 194.3 nm. Particles within the 50 nm to 200 nm range have been demonstrated to facilitate superior intestinal absorption, offering potential advantages for oral drug delivery systems. Conversely, particles exceeding 500 nm are associated with decreased transport efficiency across intestinal barriers, limiting their therapeutic utility [26].

Significantly, prior studies have achieved up to a 7.2-fold increase in oral bioavailability using particle sizes of around 329.63 ± 35.71 nm [27]. However, these results were dependent on complex procedures involving multiple solvents and extensive processing steps. In contrast, our streamlined methodology utilised ethanol, a safer solvent, achieving comparable particle size and enhanced polydispersity index (PDI) with greater stability and fewer processing requirements. Further research demonstrated that a drug–PL complex with a particle size near 7134.79 ± 0.67 nm improved solubility and pharmacological efficacy [28], underscoring the potential of the CBD-PLC formulation to similarly enhance bioavailability.

Additionally, the zeta potential of the formulation was measured at −32.5 mV, corroborating previous findings that a zeta potential above −30 mV is indicative of robust physical stability [29,30].

Figure 1 presents the logarithmic form of the oil/water partition coefficient (P_o/w) profiles of CBD, a physical mixture (PM), and CBD-PLC. The data reveal that CBD has a notably high log P_o/w value of 7.54, highlighting its pronounced lipophilicity [31]. This is consistent with predictions from SwissADME which have shown a comparable log P_o/w value of 6.52 for CBD [32], further corroborating our experimental results. The high log P_o/w value suggests that CBD may experience limited bioavailability due to its poor aqueous solubility, which could hinder its absorption in biological systems, particularly in environments like the GI tract where water solubility is important [31]. In contrast, the PM of CBD and PL shows a significant reduction in the log P_o/w to 0.3. This decrease suggests that the physical presence of PLs in the mixture enhances the overall hydrophilic–lipophilic balance, likely due to the amphiphilic nature of the PLs, which can act as surfactants, increasing the solubility of lipophilic compounds like CBD by forming micellar structures or by dispersing them more effectively in aqueous environments [33,34].

Most notably, CBD-PLC exhibited an even lower log P_o/w value of 0.13, indicating a substantial reduction in lipophilicity compared to both CBD and the PM. This suggests that the complexation of CBD with PL has significantly altered the solubility profile of the compound. The formation of CBD-PLC likely improves CBD’s solubility in polar environments due to the encapsulating properties of the PL, which can facilitate the formation of a more stable, water-dispersible system [17]. Furthermore, PLCs, particularly those composed of highly saturated PLs, form nano-sized micelles, as demonstrated in the Scanning Electron Microscopy (SEM) images (Figure 2). These micelles significantly increase the equilibrium solubility of CBD-PLC in water [35]. This reduction in lipophilicity not only enhances solubility but could also improve CBD’s bioavailability by enabling better absorption and distribution within the body [17].

This study represents a significant advancement in the development of an oral CBD delivery system. The findings provide critical insights that will guide the optimisation of PL selection for further refinement using response surface methodology (RSM) within the Stat-ease 360 software (version: 23.1.3) framework. Notably, during the initial experimental phase, L-α-Phosphatidylcholine derived from soybean was excluded due to its suboptimal characteristics. It exhibited the largest particle size among the five PLs evaluated and showed the lowest solubility (log P_o/w) in Phosphate-Buffered Saline (PBS) at pH 6.8, which is critical for effective CBD absorption in the small intestine [36]. These results underscore the importance of careful PL selection to enhance water solubility and potentially the bioavailability of oral CBD formulations.

The results from Table 3 and Table 4 indicate that although PL-100M exhibited the lowest Log P_o/w, its particle size and PDI were not as favourable as those of PC-98T and L-α-Phosphatidylcholine derived from dried egg yolk. Particle size and PDI analysis demonstrated that the encapsulation of CBD with PC-98T resulted in particles with an average size of approximately half that of pure CBD (Table 4). Based on these findings, we employed the Design of Experiments to determine the optimal ratio (w/w) between the PL (PC-98T) and CBD for the most effective particle size, PDI and Log P_o/w (Table 5).

The optimisation table (Table 6) demonstrates that our results fall within the range proposed by the DoE approach.

Among the PLs tested, PC-98T emerged as the most promising, showcasing superior attributes, as evidenced in Table 3 and Table 4. This led to its selection for subsequent optimisation using the DoE approach to determine the optimal CBD-to-PL ratio. The rigorous evaluation identified a specific ratio that accentuated the desired properties of CBD-PLC, as outlined in Table 6.

However, sourcing PC-98T posed a logistical challenge due to its limited availability in Australia. This necessitated exploring an alternative—L-α-Phosphatidylcholine from dried egg yolk. Preliminary observations supported this switch, showing that increasing the amount of any tested PL consistently amplified solubility (Table 3). This uniform behaviour underscores the foundational trait of PLs in the CBD formulation, transcending specific varieties.

Given this pervasive trend, it is plausible to infer that the optimised ratio obtained for PC-98T might be transferable to L-α-Phosphatidylcholine from dried egg yolk. While molecular interactions between PC-98T and L-α-Phosphatidylcholine from dried egg yolk may exhibit nuanced differences, their shared origin—egg yolk phosphatidylcholine—supports this hypothesis. By adopting the optimised ratio for L-α-Phosphatidylcholine from dried egg yolk, we leverage the intensive optimisation efforts previously undertaken, ensuring seamless research progression despite logistical challenges.

Furthermore, the established ratio between CBD and PL, 0.6783:20, as determined by the DoE, aligns with the maximum daily exposure (MDE) guidelines set forth by the US FDA. Notably, among all PLs analysed, L-α-Phosphatidylcholine from dried egg yolk delivered commendable secondary results in terms of particle size and PDI at pH 6.8 (Table 4), reinforcing its suitability as an alternative to PC-98T.

The PM of CBD and PL reveals that CBD forms crystals characterised by flat surfaces and sharp edges of irregular lengths (Figure 2a–c), while PL appears as rounded and irregularly shaped particles (Figure 2d–f). In contrast, CBD-PLC exhibits doughnut-shaped particles (Figure 2g–i), suggesting the probable dispersion of CBD within the PL matrix. The CBD-PLC particles display a notably smooth and porous topography with rounded edges, presenting a more ordered appearance and regular shapes (Figure 2g–i) compared to the individual CBD (Figure 2a–c) and PL (Figure 2d–f) particles [37,38].

Polymorphism is an essential part of pharmaceutical drug development since it can affect drug performance; for example, a difference of 4-fold in solubility can occur in different forms due to physicochemical properties and dissolution rate variances [39]. These dual or more different crystalline structures can potentially impact drug absorption and bioavailability [39]. Therefore, it is plausible to hypothesise that CBD-PLC particles may enhance solubility and dissolution rates.

The CBD-PLC and PL were found to be insoluble in acetonitrile, whereas pure CBD remained soluble in this solvent. This difference in solubility allowed for the quantification of pure CBD by measuring its concentration in the supernatant after centrifugation. All experiments were performed in triplicate to ensure accuracy and reproducibility. The EE% was calculated to be 78.53 ± 6.40%, demonstrating a highly efficient encapsulation of CBD within CBD-PLC.

The absorption curves of the PM (red) and CBD-PLC (light green) are almost indistinguishable (Figure 3). Conversely, the PL (dark green) and CBD (blue) displayed distinct absorption bands. This variation in the UV spectra of CBD upon interaction with the PL suggests potential chemical reactions between CBD and the PL’s functional groups [40].

The DSC analysis of CBD, the PL, the PM, and the complex (CBD-PLC) revealed distinct thermal behaviours for each sample (Figure 4). The CBD showed a sharp endothermic peak at around 66 °C, which is consistent with its melting point, as reported in the literature [13]. The PM of CBD and the PL exhibited the CBD peak, suggesting that the two components retain their individual characteristics in the mixture. However, the complex showed a significant change in the DSC curve, with the disappearance of the CBD peak. This could be indicative of the incorporation of CBD into the PL structure, possibly due to the formation of the complex. This observation aligns with the findings of Shilo-Benjamini, Y. et al. [41], who reported the successful encapsulation of CBD. This could have significant implications for the development of CBD formulations, which have been shown to enhance the bioavailability and therapeutic efficacy of CBD [13,41].

FTIR offers a profound understanding of molecular interactions between active pharmaceutical ingredients and excipients. The spectra of CBD, the PM (comprising CBD and the PL at the optimised ratio of 0.6783:20), and CBD-PLC were studied to discern potential interactions between CBD and the PL (Figure 5). The FTIR spectrum of CBD (grey) reveals significant molecular vibrations within the range of 3407 to 3547 cm⁻¹ which correspond to the O-H (aromatic) stretching vibrations [42]. While these molecular vibrations are evident within the range of the pure drug (79% T) and the PM (92%T), a shift to a diminished peak at 3379 cm⁻¹ in the complex is observed, aligning with the peak at 3376 cm⁻¹ exclusive to the PL. Zhang et al. (2014) suggested that several signals in the initial segment of the FTIR spectrum arise from O-H and N-H stretching vibrations [43]. Their disappearance in the complex, compared to the PM, potentially indicates the formation of hydrogen bonds between CBD and PC [43]. This suggests that the hydroxyl (O-H) or amine (N-H) groups present in the PM might be engaged in hydrogen bonding within the complex, leading to the disappearance or shift of these peaks [43]. Furthermore, the bands proximate to 3000 cm⁻¹ are ascribed to C-H stretching (phenyl) [42], and the sharp peak around 2920 cm⁻¹ indicates the presence of methyl and methylene groups [42]. In addition, the C-O stretching vibrations are detected at approximately 1215 cm⁻¹ [42].

The FTIR spectra of CBD-PLC predominantly reflect the peaks of CBD and the PM. For the pure drug and PM, the region ~1580 cm⁻¹ illustrates C=C stretching (phenyl ring) [42], with benzene ring stretching evident at 1646 cm⁻¹ [27]. The absence of these peaks in the formulation and their nonexistence in the PL suggest a modification in the molecular environment surrounding the aromatic rings of CBD upon complexation with PC [42]. Such alterations might be indicative of the formation of CBD-PLC through weak physical interactions and could be attributed to π-π stacking interactions, van der Waals forces, or other non-covalent interactions influencing the vibrational frequencies of these bonds, such as hydrogen bonding or van der Waals interactions between the –OH group of CBD’s phenyl rings and the polar end (P=O group) of the PLs, as proposed in Figure 6.

The peaks observed in Figure 5 at 2921 cm⁻¹, 2852 cm⁻¹, and 1738 cm⁻¹ could be assigned as characteristic C-H stretching bands of long-chain fatty acid and carbonyl (C=O) stretching bands of a PL, respectively [27]. The absorption profiles of these PL peaks remain unchanged in the FTIR spectra of CBD-PLC (Figure 5) and the PM (Figure 5). This observation highlights that the non-polar saturation of the long-chain fatty acids intrinsic to the PL did not participate in the formation of the complex [27]. Moreover, the FTIR analysis confirms that the structure of CBD remains unaltered post-complexation. However, peak shifts potentially due to hydrophobic interactions are evident, culminating in the formation of CBD-PLC [27].

To investigate the structural attributes of the inclusion complexes, the ¹H NMR spectra of CBD, the PL, the PM, and the resultant inclusion complexes were analysed. The ¹H NMR spectrum of CBD (Figure 8) aligns with the established spectral characteristics documented in the literature [47,48,49,50,51]. Pertaining to the complex, notable observations include the spectral behaviour of protons H-3′ and H-5′ and hydroxyl groups OH-2′ and H-2 (Figure 9). The peaks characteristic of the pure drug spectra are conspicuously absent in the spectrum of the complex. This observation lends substantial support to our proposed interaction between the PL and CBD, suggesting that the interaction indeed alters the spectral signature in a manner consistent with our hypothesis.

In the context of the PL, it is evident that chemical shifts play a pivotal role in NMR spectroscopy, serving as indicators of the electronic environment surrounding nuclear spins. Shifts in peak positions, particularly within the complex, are indicative of alterations in this electronic environment, potentially due to molecular interactions [53]. Furthermore, variations in peak intensities are observed, which are often reflective of changes in the relaxation properties of nuclei. These variations can be attributed to molecular interactions [54].

Observational Data:

• PL: Peaks at 4.4025 ppm (intensity: 6.16 × 10⁻¹), 4.1574 ppm (1.80 × 10⁻¹), 3.9972 ppm (4.60 × 10⁻¹), and 3.9123 ppm (4.82 × 10⁻¹).
• PM: Peaks at 4.4587 ppm (1.99 × 10⁻¹), 4.3388 ppm (9.82 × 10⁻²), 4.2625 ppm (4.33 × 10⁻¹), 4.0880 ppm (1.64 × 10⁻¹), and 3.9136 ppm (1.99 × 10⁻¹).
• CBD-PLC: Peaks at 4.4805 ppm (5.02 × 10⁻²), 4.4369 ppm (4.23 × 10⁻²), 4.3497 ppm (1.49 × 10⁻¹), 4.2516 ppm (4.84 × 10⁻¹), and 4.0662 ppm (2.56 × 10⁻¹).
• CBD: Peaks at 4.5940 ppm (1.85) and 3.9095 ppm (4.79 × 10⁻¹).

Figure 10 demonstrates a release rate of 44.7% at 2 h and 67.1% at 3 h, highlighting the efficacy of PLs as a delivery mechanism compared to the 0% release rate for pure CBD and 3.5% and 7.2% for the PM at the same time points. These results align with the previous literature. Zhenhai Zhang et al. reported a release rate of approximately 20% after 2 h [55], while Yingpeng Tong et al. described a similar release profile, characterised by an initial rapid dissolution (~50%) within the first 0–2 h, followed by a more gradual increase between 2 and 12 h, culminating in near-complete dissolution of the PLC by 12 h [40]. Notably, their data indicated a 50% release at 2 h, progressing to around 60% by 6 h, a trend that closely mirrors our observations. In studies utilising comparable methodologies, dissolution rates of 35–50% at the 2 h mark were commonly observed.

Previous dissolution outcomes did not align with the anticipated profile, which expected a rapid release within the first two hours, followed by a gradual increase extending to the 12 h mark [49]. Upon recognising that the dissolution medium might be a limiting factor, solubility studies of CBD-PLC were undertaken (Table 7). Various dissolution media, designed to simulate biological environments, were assessed to identify the most suitable medium for in vitro drug dissolution studies. CBD-PLC exhibited poor stability in PBS, whether prepared with Reverse Osmosis (RO) water or MilliQ water, likely due to ionic interactions that could interfere with the complex and produce inaccurate results. The presence of ions in PBS can interfere with lipid dissolution tests due to specific ionic interactions with PL components. For example, the study by Brinkmann-Trettenes and Bauer-Brandl (2014) found that the dissolution rate of griseofulvin in solid PL nanoparticles was significantly different when tested in PBS compared to other media, indicating that the ionic content of PBS can alter the dissolution properties of lipid-based systems [56]. Similarly, Salehi et al. (2020) highlighted that the dissolution of ionizable drugs was slower in phosphate buffer compared to bicarbonate buffer, due to differences in buffer capacity at the interfacial layer [57]. This suggests that the phosphate ions in PBS could similarly affect the stability and dissolution of phospholipid-based formulations [57]. Due to the significantly enhanced recovery of CBD in MilliQ water, as demonstrated in Table 7, this medium was selected for subsequent dissolution studies, aligning with the best practices outlined in USP guidelines [58].

As part of our preliminary studies, we conducted an integrity evaluation of the selected capsule in media with different pH values. The capsules were filled with blue food colouring pellets and the blue colour was used as a capsule release indicator (Figure 11). The capsules were placed in a pH 1.2 solution and MilliQ water (pH 5.1) and incubated at 37 °C under constant agitation for 2 h to evaluate their integrity.

In MilliQ water (pH 5.1), complete pellet release occurred within 15 min, indicating rapid capsule disintegration. In contrast, in the acidic medium (pH 1.2), pellet release was significantly delayed, with only a tiny amount of colour diffusion observed after 1 h. By the 2 h mark, partial dissolution of the release medium had led to some internal degradation of the pellets, as evidenced by the blue coloration; however, the capsules remained structurally intact (Figure 12).

Given that the pH of the MilliQ water used in the experiments fluctuated between 5 and 6, additional release studies at pH 4.5 and 6.8 were deemed unnecessary. Similarly, dissolution studies at pH 1.2 were not conducted, as the capsules can retain their integrity for at least 2 h in this acidic condition.

CBD-PLC’s initial concentration was established. CBD-PLC remained to be stable after 12 months at 4 °C and 25 °C. The values in Table 8 fall well within the stability range set by both the FDA and TGA, indicating good stability at these temperatures.

However, at 40 °C, CBD-PLC retained only 86.75 of its initial concentration, which is below the stability threshold set by the FDA (90–110%) [59] and TGA (92.5–107.5%) [60]. This reduced stability at 40 °C is further corroborated by a noticeable change in colour to a darker hue. This physical change, along with the decrease in concentration, suggests that CBD-PLC demonstrates less stability at higher temperatures, particularly at 40 °C, over a 6-month period. Despite this decrease in concentration, it is noteworthy that there were no significant changes in other characteristics such as PDI and zeta potential. An example particle size and zeta potential distribution of CBD-PLC on day 0 is shown in Supplementary Figures S1 and S2, respectively.

CBD crystals in powder form were kindly provided by Green Dispensary Compounding (Adelaide, Australia). HPLC-grade methanol and acetonitrile were obtained from EMD Millipore^® (Billerica, MA, USA). A Sartorius ultra-pure water system was utilised in all studies (Goettingen, Lower Saxony, Germany). Ethanol was purchased from Thermo Fisher Scientific (Melbourne, Australia). Lipoid E80 (80–85% egg phosphatidylcholine (PC)), 7–9.5% phosphatidylethanolamine (PE), 3% lysoPC, 0.5% lysoPE, 2–3% sphingomyelin, 2% water, 0.2% ethanol, and iodine (value 65–69) were obtained from Lipoid GmbH (Ludwigshafen am Rhein, Rheinland-Pfalz, Germany). PC-98T and PL-100M (68% phosphatidylcholine and 20% phosphatidylethanolamine) were obtained from YST PHARMA CO. Ltd. (Goka-machi, Ibaraki, Japan). L-α-Phosphatidylcholine from dried egg yolk, type X-E, ≥40% (enzymatic), L-α-Phosphatidylcholine from soybean, formic acid, Dulbecco’s Modified Eagle Medium (DMEM), Hanks Balanced Salt Solution (HBSS), Dulbecco’s Phosphate-Buffered Saline (DPBS), Fetal Bovine Serum (FBS), l-glutamine, penicillin/streptomycin (10,000 U/mL), trypan blue solution, acidic isopropanol, D-chloroform, and Corning^® Transwell^® polyester membrane cell culture inserts (12 mm Transwell with 0.4 μm pore polyester membrane, TC-treated, sterile) were obtained from Merck Pty Ltd. (Sydney, Australia). N-Octanol was purchased from ChemSupply (Adelaide, Australia). CBD-D3 was purchased from Novachem Pty Ltd. (Melbourne, Australia). Capsules CONI-SNAP size #00 (White Opaque) with acid resistance were obtained from Medisca (Sydney, Australia). We used the Stat-ease 360 software (version: 23.1.3) for the Design of Experiments. The Caco-2 cell line was kindly provided by Anthony Wignall, University of South Australia.

The HPLC method was conducted under the conditions previously studied and optimised by Abdella et al. [67]. Sample analysis was carried out by using an isocratic method using a Luna 5 µm C8(2) 100 Å (250 × 4.6 mm) analytical column at 30 °C, which was connected to an HPLC system (Shimadzu Corporation, Kyoto, Japan) consisting of a photodiode array detector (LC-20ADXR), a degasser (DGU-20A3), a system controller (CBM-20A), an autosampler (SIL-20AHT), a pump (LC20AD), and an LC solution Chromopac data processor. The mobile phase was a mixture of acetonitrile and MilliQ water (8:2, v/v) eluted at a flow rate of 1.0 mL/min. The sample injection volume was 20 µL, and a wavelength of 220 nm was selected for UV detection of CBD’s peak detected at 7.3 min.

The quantities of CBD that permeated through the Caco-2 cells were quantified using LC-MS/MS analysis. The Sciex 6500+ Qtrap LC-MS/MS (Shimadzu, Kyoto, Japan) was used in positive mode with electrospray ionization to analyse the samples. A 5 μL injection volume of each sample was loaded onto a Phenomenex Kinetex C18 analytical column, 100 Å (100 × 2.4 m, 1.7 μm), with a flow rate of 0.4 mL/min. The column temperature was 40 °C. The mobile phases used for LC separation were MilliQ water containing 0.1% formic acid (A) and methanol containing 0.1% formic acid (B). The gradient elution is summarised as follows: 0.1–6.0 min, 2% B; 6.0–7.0 min, 100% B; 8.0–9.0 min, 2% B. The total analysis required 9 min with a dwell time of 100 ms and the source temperature was 400 °C. The MRM transitions for CBD were as follows:

CBD Quantifying 1: 315.1–193.1;

CBD Quantifying 2: 315.1–259.1;

CBD-D3 (Internal Standard) Quantifying 1: 318.1–196.1;

CBD-D3 (Internal Standard) Quantifying 2: 318.1–262.0.

CBD-PLC was prepared via the solvent evaporation method. Briefly, a predetermined amount of CBD and the PL was accurately weighed and transferred into a 50 mL Falcon tube. Ethanol (30 mL) was then added, and the mixture was vortexed until a homogeneous solution was obtained. The resulting solution was transferred to a 100 mL round-bottomed flask, ensuring complete transfer by rinsing the Falcon tube twice with additional ethanol (5 mL). The flask was placed under reflux and stirred at 40 °C for 30 min.

Following the reaction, the organic solvent was removed using rotary evaporation (Rotavapor R-210, BUCHI Labortechnik AG, Flawil, Switzerland) under reduced pressure, yielding a solid product. The obtained material was further dried under vacuum at room temperature overnight to remove residual solvent. Finally, the dried product was stored at 4 °C in a sealed container until further analysis.

A preliminary investigation was conducted to compare the effects of four different PLs—Lipoid E 80, PC-98T, PL-100M, and L-alpha-phosphatidylcholine from dried egg—when combined with CBD at four different ratios (1:1, 1:5, 1:10, and 1:20). These combinations were then characterised based on their solubility (as outlined in Section 2.7), particle size, and PDI in PBS at pH 6.8.

Approximately 10 mg of CBD-PLC was dissolved in 2 mL of RO water. The solution was completely dissolved by vortexing and then filtered through a 0.2 μm filter. Subsequently, a 1:10 dilution was made with RO water and allowed to stabilise for a few min away from light before the particle size, PDI, and zeta potential were measured.

The solubility profile of CBD-PLC was investigated by adding 100 ± 10 mg of the complex into sealed Falcon tubes containing 5 mL of PBS (pH 6.8), MilliQ water, or n-octanol [68]. The tubes were subjected to gentle agitation on an orbital shaker at room temperature (23 ± 0.5 °C) for a 24 h duration away from light, followed by centrifugation at 3000 rpm for 10 min. The supernatant was subsequently filtered (0.45 μm) and diluted with methanol (1:1) before analysis via HPLC. The P_o/w and its logarithmic form (log P_o/w) are important indicators of the solubility and bioavailability properties of the compounds tested. In this study, three samples (CBD, PM, and CBD-PLC) were evaluated for their partitioning behaviour between octanol and water. The formulas used are as follows:

In this study, we employed the RSM to optimise the ratio of CBD to the selected PL. Our objective was to achieve optimal performance metrics, including enhanced water solubility, reduced particle size, PDI, and optimal zeta potential. We used the Stat-ease 360 software (version: 23.1.3) to implement an L-optimal design with a quadratic model and a randomised subtype, in line with the study’s objectives and the formulation optimisation process. L-optimal designs are commonly used in the RSM to achieve optimal responses with fewer experiments. The quadratic model, a standard in the RSM, elucidates the relationship between the response and independent variables. Randomising the experimental runs helps minimise potential errors and biases. With a total of six experimental runs and no blocks, this design is well suited for a preliminary study exploring the effects of PLs on CBD. We anticipated that this RSM approach would yield reliable results, guiding the development of an effective CBD oral formulation [69,70].

A Zeiss Merlin FE-SEM was used for the microscopic characterisation of CBD-PLC. The complex and PM were separately suspended in RO water, and a drop was placed on a stub and allowed to dry at room temperature, protected from light. Microscopic views of the samples were observed at magnifications ranging from 100 to 1000×. The morphology of the CBD-PLC samples and the PM containing CBD and the PL in the same ratio optimised by DoE (0.6378:20) was captured using the secondary electron detector with a low voltage of 1–2 kV. This microscope features a unique charge compensation system that allows for high-resolution imaging of non-conductive samples by sweeping away accumulated electrons on the sample surface with a fine jet of nitrogen.

Approximately 10 mg of CBD-PLC was mixed with 1 mL of acetonitrile in an Eppendorf tube. The mixture was then vortexed for 1 h at low speed using a multi-tube vortex mixer (Melbourne, Australia) [19,70,71]. The samples were then centrifuged at 16,100 rcf for 20 min, resulting in a clear supernatant. This supernatant was filtered through a 0.45 μm syringe filter (PVDF). The results were calculated using the equations described below:

A total of 5 mg of CBD and an equivalent amount of CBD-PLC, the PL, and the PM were dissolved in methanol and scanned by using a UV spectrometer over the wavenumber range of 200–800 nm.

DSC was performed using a DSC250 (TA Instruments, New Castle, DE, USA) to evaluate the onset temperature, melting point, width of melting events (WME), enthalpy, and crystallinity index (CI) of the pure drug, PM, PL, and CBD-PLC. Approximately 3 mg of each sample was sealed in an aluminium pan and analysed from 25 to 200 °C at a ramp rate of 10 °C/min. Nitrogen was used as a purge gas at a flow rate of 50 mL/min. The WME and CI were calculated using the following equations:

The spectra of pure CBD, the PM, the PL, and CBD-PLC were obtained using an Infrared Spectrophotometer Tensor 27 with an attenuated total reflectance module Specac Golden Gate (ATR-FTIR, Bruker, Ettlingen, Germany). A total of 64 scans were performed in a wavenumber range from 4000 cm⁻¹ to 400 cm⁻¹ at room temperature. The equipment’s diamond window was used for background correction [72].

All ¹H NMR measurements were obtained using a Bruker Avance III HD spectrometer (Bruker Corporation, Faellanden, Switzerland) at 300 MHz and were analysed using a Bruker Topspin 3.2 program. Chemical shifts are reported in parts per million (ppm) and are referenced to ¹H signals of residual nondeuterated solvent. For sample preparation, briefly, 5 mg of each sample (pure CBD, PM, PL, and CBD-PLC) was dissolved in 0.5 mL deuterated chloroform, and a total of 16 scans were obtained for each sample [73].

The dissolution assays were conducted using the paddle methodology (Method II) as specified in the United States Pharmacopoeia (USP 35). The procedure involved placing 90 mL of MilliQ water in a small vessel, which was then submerged in a water bath maintained at a constant temperature of 37 ± 0.5 °C. The paddle speed was set at 100 rpm [74].

Approximately 900 mg of CBD-PLC (the equivalent of 30 mg of CBD), alongside equivalent quantities of pure CBD and a PM of CBD and PL, was encapsulated in acid-resistant size #00 capsules. These capsules were employed as a dosage form, allowing for accurate quantification of the weighed amount of CBD and subsequent calculation of the expected percentage recovery. The capsules were then carefully positioned on the surface of the dissolution medium within each vessel [70].

Samples were collected at predetermined time intervals (0.3, 0.6, 1.0, 1.3, 1.6, 2.0, 3.0, and 6.0 h). To ensure a consistent volume throughout the experiment, fresh dissolution media were added after each sampling. The collected samples were subsequently filtered through a 0.2 μm filter and diluted in a 1:1 ratio with methanol, and 20 μL of the diluted sample was injected into the HPLC system for further analysis.

For the stability study, samples were meticulously incubated within a specialised stability chamber, adhering to the stringent International Council for Harmonisation (ICH) guidelines [75]. HPLC was employed to quantify the drug content. The CBD-PLC samples were placed into small glass vials and subjected to three distinct conditions over a 12-month period:

• Intermediate conditions at 25 °C with 60% relative humidity, devoid of light exposure;
• Accelerated conditions at 40 °C with 75% relative humidity;
• Long-term storage conditions at 4 °C, ensuring optimal preservation.