Supercritical CO2 versus water as an antisolvent in the crystallization process to enhance dissolution rate of curcumin

Abstract Antisolvent crystallization approach using either water (in conventional crystallization process (WAS)), or supercritical CO2 (in supercritical anti-solvent crystallization (SCAS)), was employed in presence of hydroxypropyl methylcellulose (HPMC) to enhance the dissolution of curcumin. The impact of pressure, temperature and depressurization time on the SCAS process was studied using the Box-Behnken design to achieve the highest saturation solubility. A physical mixture of curcumin-HPMC was prepared for comparison purposes. Saturation solubility, scanning electron microscopy, differential scanning calorimetry, X-ray diffraction analysis and Fourier transform infrared spectroscopy were conducted to characterize the solid-state characteristics of the crystallized samples. Dissolution studies helped in ascertaining the effects of the crystallization techniques on the performance of the formulation. Curcumin crystalized by different antisolvent displayed varied shapes, sizes, saturation solubility’s and dissolution properties. In SCAS process, the maximum saturation solubility (2.83 µg/mL) was obtained when the pressure, temperature and depressurization time were 275 bars, 55 °C, and 22 min respectively. The SCAS samples showed the highest dissolution (70%) in 30 min compared to WAS (27%), physical mixture (18%) and unprocessed curcumin (16%). The improved dissolution rate of SCAS sample originates from the development of sponge-like particles with augmented porosity, decreased crystallinity as well as increased solubility of curcumin.


Introduction
The low dissolution rate and poor water solubility of drugs could be improved using several approaches (solid dispersion, crystal engineering, particle size reduction) (Jain et al. 2015). Crystal engineering has attracted great interest to meet this objective over the last few years. Both the top-down and bottom-up approaches have been implemented for crystal manipulation (Hu et al. 2014). Top-down strategies require drug particle breakage physically through ball milling or pressure homogenization while bottom-up approaches utilize drug molecules to fabricate particles via the precipitation process. High energy input requirements, extended processing period, risk of metal contamination and wide particle size distribution are among the disadvantages of top-down strategies (Krause et al. 2000;He et al. 2010). This has aroused much more interest in bottom-up approaches as an alternative to crystal engineering. In the majority of these approaches, the principal objective is to obtain nanosized particles with much more effective surface area and subsequently dissolution rate, meanwhile enhanced wettability and solubility were also pursued. The antisolvent precipitation/crystallization strategy has revealed encouraging outcomes in producing very small size crystals through a controlled crystallization procedure using different stabilizing agents in the non-solvent phase (Thorat and Dalvi 2012;Kim and Yeo 2015). The benefits of the antisolvent precipitation (ASP) approach include the ease of performance, cost-effectiveness and simplicity of scaling up (Ulrich and Jones 2004;Park and Yeo 2012). In this method, stabilizers can slow down or impede particle size growth by overlaying onto the particle surfaces. Nonetheless, instant drying is essential to hinder subsequent particle size enlargement (Thorat and Dalvi 2012;Sinha et al. 2013).
Supercritical fluids (SCF) which could be used as solvent or antisolvent in various processes (including encapsulation, particle nucleation, essential oil extraction, fractionation, separation, micronization and polymerization), are gaining more attention in chemical, environmental, and pharmaceutical applications. Carbon dioxide (CO 2 ) as a safe and cheap gas with low critical pressure (74 bars) and temperature (31.1 C) has been exploited as a suitable SCF for crystal manipulation of drugs (Bouali et al. 2021;Tutek et al. 2021;Esfandiari and Sajadian 2022).
The supercritical anti-solvent (SCAS) approach has been widely exploited to customize the solid-state attributes of active pharmaceutical ingredients (APIs) (Sun 2014;Liu et al. 2017). As a fundamental mechanism of the SCAS method, an API solution is mixed with an SCF, operating as an antisolvent. The SCF drives the supersaturation and further solute precipitation, resulting in particle formation and size enlargement. Several elements, in particular, the nature of organic solvent and antisolvent, temperature, solution velocity, pressure, solute quantity, and nozzle diameter could contribute to particle microstructures during the SCAS method (Reverchon and De Marco 2011;Rossmann et al. 2012). Earlier investigations have indicated that in the SCAS procedure, the generation of foamy or sponge-like particles could increase the solubility and dissolution rate of APIs (Yasuji et al. 2008;Wu and Su 2017). The foamy particles are generated due to the volume augmentation when the supercritical state is depressurized. The quality of foam generated could be modified by handling the pressure and temperature of the procedure in conjunction with the rates of dissolved and depressurized gas (Monnereau et al. 2014;Salerno et al. 2014;Frerich 2015;Chen et al. 2016;Sun et al. 2016).
Curcumin (CUR), a yellow polyphenolic hydrophobic compound, is extracted from the turmeric (Curcuma longa) rhizomes. This substance has shown a wide variety of pharmacological properties such as anti-inflammatory, angiogenesis (at low concentration for wound healing), anti-angiogenesis (at a high concentration as an antitumor), antimicrobial, and antioxidant properties. However, its low and insufficient oral bioavailability originated from its inadequate aqueous solubility (Goel et al. 2008) has resulted in its limited use. In recent years, different strategies have been presented to overcome this challenge. In this regard, particle size reduction (Beloqui et al. 2014;Sadeghi et al. 2016;, solid dispersion preparation (Seo et al. 2012;Li et al. 2015), preparation of lipid-based solid particles (Kakkar et al. 2013), preparation of nanosized micelles (Abouzeid et al. 2014;Song et al. 2014), and cyclodextrin-based nanoparticles approaches (Tønnesen et al. 2002) are amongst methods that have been tried. SCAS technology has been also studied and applied in the processing of curcumin particles for micronization, nanoencapsulation or microencapsulation to improve stability, bioavailability, and water solubility (Kakran et al. 2012;Yadav and Kumar 2014;Zabihi et al. 2014;Jia et al. 2016;Gracia et al. 2018). Different carriers such as polyvinylpyrrolidone (PVP), Eudragit V R L100, and Pluronic V R F127 have been studied for encapsulation of curcumin (Arango-Ruiz et al. 2018). Furthermore, in another study, coprecipitates of curcumin and PVP (guest particles) were simultaneously produced and coated onto microcrystalline cellulose, corn starch and lactose (host particles), by combining the SCAS with two different configurations of fluidized bed (tapered bed and stirred vessel) (Matos et al. 2020). To improve the oral bioavailability of CUR by reducing its particle size to the nano range using the SCAS process, combining Tween 80 as a solubilizer and permeation enhancer with different organic solvents was studied (Anwar et al. 2015). Inhalable dry formulations for pulmonary delivery of curcumin have been produced by SCAS micronization. The antioxidant curcumin was co-processed with hydroxypropyl-beta-cyclodextrin (HP-b-CD) and PVP to form binary and ternary composites with enhanced flow-ability for pulmonary delivery (Kurniawansyah et al. 2015). Cocrystallization with supercritical solvent technique (CSS) using nicotinamide (Ribas et al. 2019a), N-acetylcysteine (Ribas et al. 2019b), and resveratrol (Dal Magro et al. 2021) to improve curcumin dissolution rate and bioavailability were also investigated. In most of the studies listed above supercritical technology has been used for encapsulation of curcumin where a high amount of polymers were employed.
The focus of the current investigation was to manipulate curcumin crystals by antisolvent crystallization in presence of a little amount of a stabilizer where two common antisolvents (water or supercritical CO2 used in conventional or supercritical antisolvent crystallization) were compared in the process of antisolvent crystallization of curcumin under the same conditions. Moreover, it is for the first time the current study introduces HPMC as a stabilizer in supercritical fluid technology to modify curcumin particles. To the best of our knowledge, these points have not been considered in other studies.

Supercritical anti-solvent crystallization of CUR (SCAS)
The instrument design for the antisolvent crystallization using the supercritical procedure employed in this study was similar to previously reported studies ( Figure 1) (Campardelli et al. 2015;Johannsen and Brunner 2018). Initially, the CO 2 (compressed in a cylinder), was introduced into a molecular sieve-filled column and then filtered for its purity enhancement. Then the CO 2 was liquefied using a recirculation cooling bath. The supercritical conversion of the pumped liquefied CO 2 was conducted using a heating step.
A solution containing 1000 mg CUR and 200 mg HPMC in 10 mL acetone was used ) to fill the stainless-steel column, fixed with inlet and outlet filters. During the CO 2 transfer through the column, the pump presented a regular experimental pressure. Preheating coil and oven were used to limit temperature variations. After achieving the appropriate supercritical temperature and pressure, the pump was disconnected and switched off. The dissolution of acetone in the scCO 2 occurred during the soaking time (t s ¼ 1 h). A back-pressure regulator controlled the constant flow rate of supercritical CO 2 at the depressurization time (dynamic time). Ultimately, the CUR particles were developed and deposited on the sides and bottom of the column.
where Y is a response, b o indicates the intercept, b j denotes linear coefficients, b jj represents squared coefficients and b jk stands for interaction coefficients. Levels of certain independent variables are represented by X i , X j 2 , X j , and X k (Kamali et al. 2018). The dependent variable (Y) was calculated using the Design-Expert program (version 11.0.4).

Conventional antisolvent crystallization using water as antisolvent (WAS)
According to the procedure, CUR (1000 mg) was dissolved in acetone (10 mL) in presence of 200 mg HPMC. To perform antisolvent crystallization, CUR acetonic solution was added to distilled water (100 mL in a beaker) at a steady rate (10 mL/min) at room temperature while stirring using a magnet stirrer (800 rpm) (Homayouni et al. 2014). Then the mixture was instantly cooled (-20 C) and then freeze-dried for 48 h. The freeze-dried sample was kept in a desiccator until further experiments.

Preparation of CUR-HPMC physical mixture
The CUR and HPMC (at a ratio of 5:1) sieved fractions (less than 250 mm) were mixed for 20 min in mortar and pestle to achieve a homogenous physical mixture (PM).

Scanning electron microscopy
The morphology of CUR as well as the samples obtained via WAS and SCAS methods were analyzed by scanning electron microscopy (SEM) technique (TESCAN, Czech Republic). Before SEM imaging, the samples were made electrically conductive by coating them with gold-palladium alloy (Hu et al. 2011).

DSC analysis
Differential scanning calorimetry (DSC) (Mettler Toledo, GmbH, Greinfensee, Switzerland) was carried out to analyze unprocessed CUR, WAS and the optimized samples obtained via SCAS. To this end, a certain amount of samples (2-3 mg) was heated in aluminium pans (10 C/min) in the range of 0-250 C with N 2 purging at about 80 mL/min. A reference standard was an empty aluminium pan. A STARe software (version 10.00) was applied to analyze the samples.
2.5.3. X-Ray diffraction analysis (XRD) X-ray powder diffraction patterns of all powders including CUR, WAS, the optimized SCAS, HPMC, and PM were obtained on a diffractometer (Bruker, Karlsruhe, Germany) using Cu Ka radiation (range of 5-40 (2h) with 0.05 steps).

Fourier transform infrared spectroscopy (FTIR)
The FTIR spectra of unprocessed CUR, WAS, optimal SCAS, HPMC and PM were recorded by an FT-IR spectrometer (Perkin Elmer spectrum II, PerkinElmer, Waltham, MA) in a range of 450-4000 cm À1 , with a resolution of 1.0 cm À1 , and a number of reference scans of 10. After mixing samples with potassium bromide (KBr) at a weight ratio of 5:1 (KBr: sample), the mixture was compressed by a hydraulic press (7 tons, 2 min) and the FTIR spectra were recorded. The control was an empty KBr disk.

Porosity determination
Equation (2) was utilized to determine the porosity of samples (as the effective percentage) (Young et al. 2005;Goimil et al. 2019): where E represents porosity (%); q b is bulk density, and q t shows true density. Helium pycnometer (Micrometrics1 AccuPyc 1330 pycnometer; Norcross, GA) was utilized to ascertain the true density of samples (three replications), which were cryogenically transformed into a fine powder. On the other hand, bulk density was measured by the Mercury porosimetry technique (Micrometrics1 PoreSizer 9320; Norcross, GA) after three replications. At 72 h prior to determining density parameters, the samples were lyophilized.

Solubility measurement
To determine the saturation solubilities of crystallized CUR samples, unprocessed CUR and PM, each sample was diluted in distilled water (10 mL) in excess quantities, followed by incubation at 25 C for 48 h while shaking at 200 rpm. Then the mixture was membrane filtered.
The CUR absorbance at a wavelength of 426 nm was determined and used to obtain the saturation solubility of each sample. The sensitivity of the method was determined with respect to the limit of detection (LOD) and limit of quantification (LOQ). A series of concentrations of drug solutions (0.01-7 mg/mL) were used and analyzed to construct a calibration curve and determine LOD and LOQ.

Dissolution studies
In an automated Pharma Test dissolution instrument (Germany), dissolution experiments were conducted using a USP apparatus II performed at 50 rpm at 37 ± 0.5 C. Nearly sink condition was preserved by adding SDS (0.25% (w/v)) in distilled water (1000 mL) (Fouad et al. 2011;Lee et al. 2013). After the addition of samples (equivalent to 10 mg of CUR), in each vessel (in triplicate), a peristaltic pump (Alitea, Sweden) and a sintered filter were used to remove samples from the media. To calculate the CUR concentration, the CUR standard curve derived from its absorbance at 426 nm was exploited. A primary measure revealed no interference between CUR absorption at 426 nm with HPMC and SDS (SLS). To compare the dissolution profiles, the dissolution efficiency (DE) was determined (Costa et al. 2003

Statistical analysis
All statistical analyses of the results were performed using a oneway analysis of variance (ANOVA) followed by the Tukey-Kramer post-test. GraphPad Prism version x8 (GraphPad Software, Inc.) was used for the statistical analyses.

Results and discussion
Employing the antisolvent crystallization process (ASP) for particle engineering is a straightforward and efficacious technique to modify the low water solubility of drug particles. In the current research, acetone acted as a solvent for CUR, while water and sc-CO 2 served as antisolvent. Acetone was exploited as the solvent for CUR due to the good solubility of CUR in it and its rapid rate of evaporation. Distilled water and sc-CO 2 were exploited as antisolvents as stated by their safety, ease of access and affordable price. HPMC was also used as a stabilizer to augment the solubility and dissolution rate of CUR according to the study reported by . Stabilizers have been exploited to impede crystal growth during particle precipitation via steric hindrance derived from absorption at the drug-solvent interface (Maximiano et al. 2011;Verma et al. 2011).

Scas process optimization
For multiple runs of tests, the saturation solubility of different samples corresponds to designated values of variables in the SCAS process are represented in Figure 2 and Table 1. In accordance with the desirability value of 1.0, the greatest saturation solubility of 2.83 mg/mL for curcumin was obtained at 55 C, 275 bar, and 22 min for T, P, and td, respectively. To determine saturation solubility as an act of different independent variables, the second-degree polynomial equation can be expressed as: Y ¼ þ 9:82727 -0:505125T -0:006820 P þ 0:259667 t d þ 0:006271 T 2 þ 0:000034 P 2 -0:004659 t d 2 -0:000085 T Â P -0:001155 T Â t þ 0:000056 P Â t where P represents pressure, T represents temperature, td represents depressurization time, and Y represents saturation solubility. ANOVA (residuals analysis) was used to evaluate the RSM discovered by the experimental design. The linear regression coefficient (R 2 ) of 0.9559 indicated that the model performed as expected. The significance of each item in the model may be highlighted by the fact that its ANOVA test p-value was 0.05 with a 95% confidence level. The maximum F-value and the minimum p-value indicate which parameter is more effective. The results for F-value and p-value in the above equation revealed that the linear parameter of time and its square are more effective (Golmakani et al. 2014).  The influence of temperature and pressure on saturation solubility is shown in Figure 2A. The influence of SCF pressure on the solubility of curcumin was explored at pressures ranging from 75 to 275 bars. The increase in the pressure up to 170 bars had no discernible effect on solubility. As the pressure increased from 170 to 275 bars, the increase in the solubility became more noticeable until the maximum solubility was reached. The parameters of organic solvent diffusivity and antisolvent density were significantly affected by increasing the pressure. Pressure elevation, which increases the density of CO 2 may result in increased CO 2 solubility in an organic solvent. As a result, reaching supersaturated conditions takes less time, and this could lead to powder precipitation (Amani et al. 2021). The saturation solubility showed two unique tendencies after the temperature alteration. The solubility decreased with a gentle slope between 35 and 45 C and then increased as the temperature rose to 55 C.
According to the literature, increasing the temperature causes changes in the solubility and diffusivity of organic solvents in sc-CO 2 , as well as the polymer thermodynamics including swelling and plasticization (Ameri et al. 2020). Besides, at raised temperatures, the reduced viscosity might be used to explain particle diameter diminution. Remarkably, decreased CO 2 viscosity with increased temperature, can enhance diffusivity and chain mobility of HPMC and decrease entanglements within polymer chains, therefore inducing increased mass transfer, nucleation rate and porosity (Liu et al. 2020). Indeed, following temperature elevation from 35 to 45 C, diminution of sc-CO 2 density up to 45 C overlaps with the increased CO 2 diffusion rate up to 45 C at which minimum solubility occurs, and afterwards, enhancement of diffusion rate up to 55 C succeeds the density diminution.
Moreover, a reduction in the viscosity with an increase in temperature can promote the polymer matrix's chain mobility, reduce polymer chain entanglements, and make it easier for dispersion or diffusion of API into the inner region of the polymer during the supercritical procedure. This could lead to increased CO 2 mass transfer and significant CO 2 adsorption inside polymer chains and consequently prompt the generation of foamy or porous constructions after depressurization (Liu et al. 2020). In addition, the temperature elevation, and its effects on heat transfer were accountable for the rapid diffusion of the solvent into the SCF, and a favourable influence on the nucleation rate. Finally, the influence of temperature on drug and polymer particle size and architecture is mutual. As a result, it is vital to ponder all of the variables that come into play when polymer and drug coprecipitation occurs during the supercritical procedure (Kanaujia et al. 2015).
The results showed that the elevation of the depressurization time (td) from 10 to 22 min improved the saturation solubility, and then further increase in t d up to 30 min, ultimately decreasing the solubility (Table 1). The highest saturation solubility of 2.83 mg/mL was observed for curcumin at a depressurization time of 22 min as indicated in Figure 2B. It has been demonstrated that the volumetric expansion had a major impact on the degree of supersaturation. It favours supersaturation levels when the rate of added antisolvent is high (low depressurization time (t d )). As a result, more nuclei are produced and particles with a bigger surface area are generated (M€ uller et al. 2000).

Morphology observations
SEM analysis revealed different morphology for crystallized samples and unprocessed CUR (Figure 3). Unprocessed CUR revealed plate-like crystals ( Figure 3A) while crystals of WAS sample were rod-shaped with approximate dimensions of 12-15 mm length and 1-1.5 mm width ( Figure 3B). However, samples derived from the optimal SCAS process displayed agglomerated particles consisting of very tiny particles (with approximate dimensions of 1-1.2 mm) ( Figure 3C). The SEM record of CUR particles derived from the optimal SCAS method represented a sponge-like arrangement of tiny particles (foamy structure) similar to the results described by Liu et al. for nimesulide particles obtained by the SCAS process (Liu et al. 2020). Given the availability of a large surface area for dissolution, this framework could be beneficial in improving the dissolution rate. In another research, indomethacin-PVP co-precipitate with a reinforced dissolution rate was prepared based on a solvent-free SCF strategy using sc-CO 2 . The SEM record indicated a porous foamed framework for obtained particles (Gong et al. 2005). The immediate CO 2 release during the process accounted for the formation of foam-like particles with porous structures. The optimum SCAS process generated particles with a smaller pore size (as depicted in SEMs), bigger surface area, and greater porosity (Table 2) in comparison with WAS sample (p < 0.05).  Indeed, fast nucleation and CO 2 quenching resulted in the development of agglomerated particles with foamy structures and numerous small pores within (Chen et al. 2016).

DSC
The DSC results of unprocessed CUR and crystalline products derived from diverse strategies are presented in Figure 4. The endothermic peak of unprocessed CUR was close to 169.97 C, which corresponds to its melting point consistent with previous studies (Kakran et al. 2012). The HPMC also exhibited a broad endothermic peak at 62 C which was related to the adsorbed moisture. The endothermic peak observed for PM sample was close to the melting point of CUR (169.40 C). This peak shifted a little bit to the left side possibly due to partial dissolution of CUR in HPMC while it was heated beyond its T g in the heating process. These findings demonstrate the probability of stabilizer/ CUR miscibility. The same results have been reported in a previous study for different API and polymers . In the WAS sample, the peak of CUR melting was observed in the region of 173 C. However, this peak appeared slightly wider than the unprocessed sample of CUR, indicating a decrease in the crystallinity of CUR. An endothermic broad peak at 72 C was probably due to adsorbed moisture in the sample. In the SCAS sample, a peak of CUR melting appeared in the region of 173 C, but the peak was much wider than that observed for WAS sample with low slope at the start, which could indicate more reduction in CUR crystallinity. Another peak was also observed in the region of 102 C, which is probably related due to the loss of free water entrapped in pores of this sample (not adsorbed moisture).

XRD pattern
The XRD patterns for different samples are presented in Figure 5. The characteristic diffraction pattern of unprocessed CUR with sharp peaks at 2Ɵ values (9.15 , 10.3 , 14.2 , 16.9 , 20.1 and 21.1 ) is related to its crystalline nature (Sanphui et al. 2011). CUR diffraction peaks clearly appeared in the XRD of PM. A comparison of the XRD pattern of WAS with PM revealed no considerable differences between them. It appears that WAS strategy could not markedly affect the CUR solid state throughout the crystallization step. However, observation of the same characteristic peaks with considerably reduced intensities in the XRD pattern of the optimal SCAS sample indicated the huge reduction of crystallinity compared to the unprocessed CUR and WAS samples. XRD data were in good agreement with DSC results. Earlier investigations have demonstrated that the stabilizing function of polymer throughout the antisolvent crystallization process could affect drug crystallinity. This phenomenon was derived from the probable absorption of the stabilizer (HPMC) into the curcumin crystals during the antisolvent crystallization process.

FTIR data
FTIR examination was performed to inspect the feasibility of CURstabilizer hydrogen bonding following antisolvent crystallization. Figure 6 has illustrated the FTIR spectrums of unprocessed CUR and crystalized samples derived in the presence of HPMC. Consistent with the data of previous work, unprocessed CUR demonstrated characteristic peaks at 1601.2 cm À1 , 1427 cm À1 , 1275.9 cm À1 , and 3505 cm À1 derived from aromatic C ¼ C, phenolic C-O, phenolic C-O, and OH stretches respectively. Furthermore, as stated in other studies, CUR revealed a characteristic carbonyl group-originated peak at 1628 cm À1 (Wegiel et al.

Solubility studies
This experiment was carried out to evaluate the impact of antisolvent type and the presence of HPMC on CUR saturation solubility. Data presented in Figure 7A indicated that the extent of saturation solubility for unprocessed CUR was very low (0.23 ± 0.12 mg/ mL) (Sadeghi et al. 2016). It was noted that mixing the HPMC (hydrophilic polymer) with CUR in PM improved CUR solubility significantly (p < 0.05). The solubility of CUR determined in WAS sample (1.30 ± 0.15 mg/mL) was slightly higher than PM (1.02 ± 0.22 mg/mL) (p < 0.05). Whilst the optimal SCAS sample revealed an approximately 11.8-fold increase in solubility (2.83 ± 0.25 mg/mL) compared to unprocessed CUR and also much  more than the PM and WAS samples ( Figure 7A). This suggests the potential of sc-CO 2 as an antisolvent compared to water in the crystallization of CUR.

Dissolution studies
CUR dissolution improvement is one of the primary goals of this research. In this regard, dissolution tests were carried out to explore the impact of water and sc-CO 2 as antisolvent on the CUR dissolution rate. As shown in Figure 7B, the unprocessed CUR demonstrates a considerably poor dissolution rate so less than 20% of CUR dissolved after 40 min. Although it was expected that the wetting behavior of hydrophobic pharmaceuticals such as CUR and consequently its dissolution rate could be strengthened in the presence of HPMC as hydrophilic material, this was not the case for the PM as the CUR dissolution rate was not affected significantly by the presence of HPMC in the sample. These findings showed that the sole presence of HPMC could not efficiently improve the dissolution rate of CUR. Samples obtained through WAS process presented a 2-fold increase in DE in comparison to unprocessed CUR. As shown in Table 2, the DE for WAS sample was higher and its MDT was lower than either PM or unprocessed CUR. A comparison of the dissolution profiles for the WAS and SCAS samples demonstrated that the SCAS strategy effectively improved the CUR dissolution rate ( Figure 6B). In this regard, within 30 min, the SCAS sample released > 70% of CUR while this value was lower than 27% for WAS, 18% for PM, and 16% for unprocessed CUR. In addition, DE for the SCAS sample was significantly higher than other samples (WAS, PM, and unprocessed CUR) (p-value < 0.05). This marked increase in the dissolution rate of CUR could originate from greater solubility and porosity in the SCAS sample compared to WAS, PM, and unprocessed CUR. Additionally, in accordance with DSC and XRD data, a considerable reduction of CUR crystallinity in the SCAS sample could also contribute to the dissolution enhancement of curcumin in SCAS. Altogether, these findings validate the elevated dissolution rate and solubility of crystallized CUR in the SCAS process.

Conclusion and future perspectives
This research was conducted to compare the antisolvent efficacy of sc-CO 2 and water during the CUR antisolvent crystallization in presence of HPMC as a stabilizer. The outcomes confirmed the potential of sc-CO 2 compared to water as antisolvent in the antisolvent crystallization procedure. Applying sc-CO 2 in the SCAS process has a great influence on the shape, crystallinity, solubility the dissolution rate of CUR. The dissolution rate of SCAS crystallized samples could be controlled by changing temperature, pressure, and depressurization time during the process. The SCAS procedure and applied sc-CO 2 as an antisolvent is a promising strategy to boost the dissolution rate of curcumin.

Disclosure statement
No potential conflict of interest was reported by the author(s).