Solubility Determination, Hansen Solubility Parameters and Thermodynamic Evaluation of Thymoquinone in (Isopropanol + Water) Compositions

This article studies the solubility, Hansen solubility parameters (HSPs), and thermodynamic behavior of a naturally-derived bioactive thymoquinone (TQ) in different binary combinations of isopropanol (IPA) and water (H2O). The mole fraction solubilities (x3) of TQ in various (IPA + H2O) compositions are measured at 298.2–318.2 K and 0.1 MPa. The HSPs of TQ, neat IPA, neat H2O, and binary (IPA + H2O) compositions free of TQ are also determined. The x3 data of TQ are regressed by van’t Hoff, Apelblat, Yalkowsky–Roseman, Buchowski–Ksiazczak λh, Jouyban–Acree, and Jouyban–Acree–van’t Hoff models. The maximum and minimum x3 values of TQ are recorded in neat IPA (7.63 × 10−2 at 318.2 K) and neat H2O (8.25 × 10−5 at 298.2 K), respectively. The solubility of TQ is recorded as increasing with the rise in temperature and IPA mass fraction in all (IPA + H2O) mixtures, including pure IPA and pure H2O. The HSP of TQ is similar to that of pure IPA, suggesting the great potential of IPA in TQ solubilization. The maximum molecular solute-solvent interactions are found in TQ-IPA compared to TQ-H2O. A thermodynamic study indicates an endothermic and entropy-driven dissolution of TQ in all (IPA + H2O) mixtures, including pure IPA and pure H2O.


Introduction
Thymoquinone (TQ, Figure 1) is a naturally-derived bioactive compound, which is isolated mainly from the essential oils of Nigella sativa [1,2]. In the literature, a variety of therapeutic/pharmacological effects of TQ have been reported [3][4][5][6][7]. However, a high amount of TQ is required to produce its therapeutic efficacy, due to its poor water solubility [7,8].
The solubility data of poorly water-soluble naturally-derived bioactive compounds in cosolvent mixtures have a significant role in various industrial processes, such as purification, recrystallization, drug discovery, and formulation development [9][10][11][12][13]. The potential of isopropanol (IPA) has been proved in enhancing the solubility of various poorly soluble bioactive compounds, such as ferulic acid, hesperidin, iminodibenzyl, and quinocetone in the literature [10,[14][15][16]. Various formulation strategies, including nanomedicines and advanced drug delivery systems, have been adopted for the solubility/dissolution and bioavailability improvement of TQ in the literature [17][18][19][20][21][22]. The solubility of TQ in neat water (H 2 O) and neat isopropanol (IPA) has been documented well in the literature [7,23,24]. However, the solubility data of TQ in various (IPA + H 2 O) mixtures is not studied and reported in the literature. Therefore, the solubility data of TQ in various (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O at 298.2-318.2 K and 0.1 MPa, respectively, are studied and reported in this work. The thermodynamic behavior of TQ in various (IPA + H 2 O) compositions, including neat IPA and neat H 2 O, was also evaluated. The solubility parameters for free TQ, neat IPA, neat H 2 O, and various (IPA + H 2 O) mixtures free of TQ were also predicted to correlate with experimental solubility data. The activity coefficients of TQ were also determined to evaluate the TQ-IPA or TQ-H 2 O interactions at the molecular level. The physicochemical data of TQ in various (IPA + H 2 O) combinations reported in this investigation would be applicable in "purification, recrystallization, drug discovery, preformulation studies, formulation development, and preclinical studies" of TQ at laboratory and industrial scale.

TQ Solubility Determination
The experimental solubilities of TQ in binary (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O at 298.2-318.2 and 0.1 MPa, respectively, are included in Table 1. The solubility of TQ in pure IPA and pure H 2 O are documented well in the literature [7,23,24]. However, there are no reports on solubility values of TQ in binary (IPA + H 2 O) combinations. Some other cosolvents, such as dimethyl sulfoxide, carbitol, and polyethyleen glycol-400, etc., have been studied to solubilize TQ in our previous publication [24]. In this study, only IPA was studied in the solubilization of TQ in H 2 O. The equilibrium solubility of TQ in neat H 2 O at 298.2 K was reported as 740.63 µg mL −1 (equivalent to 7.75 × 10 −5 in mole fraction) by Salmani et al. [23]. The mole fraction solubility of TQ in neat H 2 O at 298.2 K was determined as 8.27 × 10 −5 by Ahmad et al. [24]. The mole fraction solubility of TQ in neat H 2 O at 298.2 K was obtained as 8.25 × 10 −3 in this research ( Table 1). The equilibrium solubility of TQ in neat IPA at 298.2 K was reported as 165.54 mg mL −1 (equivalent to 4.55 × 10 −2 in mole fraction) by Kalam et al. [7]. The mole fraction solubility of TQ in neat IPA at 298.2 K was reported as 4.51 × 10 −2 by Ahmad et al. [24]. The mole fraction solubility of TQ in neat IPA at 298.2 K was recorded as 4.47 × 10 −2 in this work. The mole fraction solubilities of TQ in pure IPA and pure H 2 O at 298.2-318.2 K have also been reported [24]. The graphical analysis between experimental and literature solubility data of TQ in pure H 2 O and pure IPA at 298.2-318.2 are presented in Figure 2A,B, respectively.   The results depicted in Figure 2A,B showed an excellent correlation of experimental solubility data of TQ with its literature solubilities in neat H 2 O and neat IPA at 298.2-318.2 K [24]. The root mean square deviation (RMSD) between experimental and literature solubility values of TQ in H 2 O and IPA was found to be 1.82% and 0.86%, respectively. These observations indicated that the experimental solubility values of TQ obtained in this research were in good agreement with its literature solubility values [7,23,24].
In general, TQ mole fraction solubilities (x 3 ) were found to be increasing significantly (p < 0.05) with the rise in temperature and IPA mass fraction in (IPA + H 2 O) mixtures, and hence, the lowest mole fraction solubility of TQ was recorded in neat H 2 O (x 3 = 8.25 × 10 −5 ) at 298.2 K and the highest one was recorded in neat IPA (x 3 = 7.63 × 10 −2 ) at 318.2 K. The effect of IPA mass fraction on TQ logarithmic solubilities at 298.2-318.2 K is summarized in Figure 3. According to results depicted in Figure 3, TQ logarithmic solubility was found to be increased significantly with an increase in IPA mass fraction at 298.2-318.2 K (p < 0.05). The maximum solubilization/solvation of TQ in IPA could be possible due to lower polarity of IPA, maximum solute-solvent interactions between TQ and IPA, and lower activity coefficient of TQ in IPA compared with water [10]. Accordingly, IPA could be utilized as an efficient cosolvent in TQ solubilization.

Computation of Hansen Solubility Parameters (HSPs)
The HSPs are a very important tool for validating experimental solubility values to find out the best solvent for the solubilization of materials. Therefore, various HSPs for TQ, neat IPA, neat H 2 O, and binary (IPA + H 2 O) compositions free of TQ were determined in this work. The HSPs for binary (IPA + H 2 O) compositions free of TQ at 298.2 K are included in Table 2. The total HSP (δ) value of TQ was predicted as 21.00 MPa 1/2 , indicating that TQ is the low-polar bioactive compound. The HSP values for neat IPA (δ 1 ) and neat H 2 O (δ 2 ) were predicted as 22. 30

Determination of Ideal Solubility and Activity Coefficients
The results of ideal solubility (x idl ) values of TQ are included in Table 1. The ideal solubilities of TQ were estimated as 6.16 × 10 −1 -9.92 × 10 −1 at 298.2-318.2 K. The ideal solubilities of TQ were significantly higher compared to its experimental solubilities in neat IPA, neat H 2 O, and various (IPA + H 2 O) mixtures (p < 0.05). The ideal solubilities of TQ were increased significantly with an increase in temperature, which indicated that the dissolution process of TQ was endothermic. The higher ideal solubilities of TQ compared to its experimental solubilities were possible, due to its low melting temperature (352.8 K) [24].  Table 3. The γ i value of TQ was obtained larger in neat H 2 O at 298.2-323.2 K. However, the γ i of TQ was found lowest in neat IPA at 298.2-318.2 K. The larger γ i values in water were possible, due to the lowest solubility values of TQ in water. However, the smaller γ i values of TQ were due to the highest solubility values of TQ in IPA [10]. It has been reported very well in the literature that low values of γ i are associated with maximum solute-solvent interactions at the molecular level [10,24]. The γ i values of TQ in IPA were lowest compared to the H 2 O; therefore, the maximum solute-solvent interactions at the molecular level were considered in TQ-IPA compared to the TQ-H 2 O.

Apparent Thermodynamic Analysis of TQ Solution
The results of various apparent thermodynamic parameters for TQ dissolution in different (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O, are included in Table 4.

System
Jouyban-Acree Jouyban-Acree-van't Hoff Results of the JAVHM model for TQ in binary (IPA + H 2 O) compositions are also summarized in Table 9. An average RMSD for the JAVHM model was estimated as 0.64%. According to these observations and results, all investigated theoretical models showed low values of RMSD, which indicated good solubility correlation for TQ with all six theoretical models evaluated.

Preparation of Solvent Mixtures
All (IPA + H 2 O) solvent mixtures were prepared on mass, using a "Digital Analytical Balance (Mettler Toledo, Greifensee, Switzerland)" with a sensitivity ± 0.10 mg. The mass fraction of IPA for different (IPA + H 2 O) solvent mixtures was varied by 0.10 from 0.10-0.90. Each solvent mixture was obtained in triplicates. The samples were continuously shaken to reach equilibrium. After equilibrium, the samples were taken out from the shaker and centrifuged. The supernatants were carefully withdrawn and diluted with mobile phase and utilized to determine TQ content using the HPLC method at 254 nm, as described in the literature [38]. "Waters HPLC system (Waters Corporation, Milford, MA, USA)" was utilized for TQ analysis. The binary mixture of potassium dihydrogen phosphate buffer (10 mM) and methanol (90:10, v/v) was used as the mobile phase. The flow rate of the mobile phase was kept constant at 1.0 mL/min with an isocratic mode. The injection volume for all samples was 20 µL. The column used for TQ analysis was Waters Symmetry-R RP C-18 (150 × 4.6 mm, 5 µm particle size). The TQ content in the above samples was calculated from the previously constructed calibration curve of TQ. The TQ calibration curve was constructed on a mass/mass basis, which was observed as linear in the range of 1-200 µg/g. The x 3 value of TQ was then calculated using Equations (1) and (2) [10,11]:

Estimation of HSPs
The HSPs of solute, solvent, and cosolvent mixtures are an important tool to predict the best solvent for the solubilization of the solute [39]. Hence, the HSPs for TQ, neat IPA, neat H 2 O, and different (IPA + H 2 O) compositions free of TQ were predicted to compare the experimental solubility values of TQ. The HSP (δ) for TQ, neat IPA, and neat H 2 O was obtained by applying Equation (3) [25,39]: where, "δ = total HSP; δ d = dispersion HSP; δ p = polar HSP and δ h = hydrogen-bonded HSP". The HSPs for TQ, pure IPA, and pure H 2 O were computed by "HSPiP software (version 4.1.07, Louisville, KY, USA)" [39]. However, the HSPs of different (IPA + H 2 O) compositions free of TQ (δ mix ) were obtained using Equation (4) [26,40]: where, α = IPA volume fraction in (IPA + H 2 O) combinations; δ 1 = HSP of pure IPA, and δ 2 = HSP of pure H 2 O.

Determination of Ideal Solubility and Activity Coefficients
The x idl of TQ at 298.2-318.2 K was calculated by applying Equation (5) [31]: In which, T = absolute temperature; T fus = TQ fusion temperature; R = universal gas constant; ∆H fus = TQ fusion enthalpy, and ∆C p = difference in the molar heat capacity of TQ solid state with that of TQ liquid state [9,41]. The quantitative data of T fus = 318.52 K, ∆H fus = 19.44 kJ/mol, and ∆C p = 61.03 J/mol/K were taken from the literature [24]. The x idl value for TQ was calculated by applying Equation (5).
The γ i for TQ in binary (IPA + H 2 O) compositions, including pure IPA and pure H 2 O, were obtained by applying the non-random two-liquid (NRTL) model [42]. Based on the principle of solid-liquid equilibrium, the γ i can be calculated using the NRTL model with the help of Equation (6) [42,43]: The value of ∆C p between the melting and solid states of TQ is very small, and hence, the Equation (6) can be simplified as Equation (7) [43]: The interactions between TQ and the solvents or solvent mixtures at the molecular level were evaluated based on the recorded TQ γ i data.

Apparent Thermodynamic Analysis of TQ Solution
Apparent thermodynamic analysis of TQ in binary (IPA + H 2 O) compositions, including pure IPA and pure H 2 O, was performed to evaluate TQ dissolution character by applying van't Hoff and Gibbs Equations. The van't Hoff Equation was applied to compute the thermodynamic parameters of TQ in various (IPA + H 2 O) compositions at mean harmonic temperature (T hm ) = 308 K at 298.2-318.2 K. The ∆ sol H 0 value can be obtained using the Equation (8) [41,44]: By plotting ln x 3 versus 1 T − 1 T hm , the ∆ sol H 0 and ∆ sol G 0 data for TQ were determined using Equations (9) and (10), respectively [45]: The ∆ sol S 0 data for TQ in various (IPA + H 2 O) combinations, including neat IPA and neat H 2 O, were computed by applying Equation (11) [41,44,45]:

Solubility Correlation With Theoretical/Computational Models
where, a and b = model coefficients, and these values were predicted using the least square method. The regression between experimental and VHM solubility of TQ was performed using RMSD and R 2 values. The RMSD of TQ was obtained using its standard equation, reported previously [9]. The MAM model is also a semi-predictive model, which correlates the solubility of solutes at different temperatures at the given set of cosolvent mixtures. Therefore, the MAM solubility (x 3 Apl ) of TQ in binary (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O, was predicted by applying Equation (13) [27,28]: where, A, B, and C = the model coefficients, and these values were predicted using nonlinear multivariate regression analysis of experimental solubilities TQ included in where, x 1 = mole fraction solubility of TQ in IPA; x 2 = mole fraction solubility of TQ in H 2 O; m 1 = IPA mass fraction, and m 2 = H 2 O mass fraction. The BKλhM model is based on two parameters λ and h. It is also a semi-predictive model, which is widely used for the solubility correlation at different temperatures at the given set of cosolvent compositions. The BKλhM solubility (x 3 λh ) of TQ was obtained using Equation (15) [30,31]: where, λ and h = the adjustable model parameters of Equation (15). The regression between experimental and BKλhM solubility of TQ was performed using RMSD. The JAM predicts the solubility of solutes in cosolvent mixtures at the given set of temperatures. The JAM solubility (x 3 JA ) of TQ in binary (IPA + H 2 O) composition was predicted using Equation (16) [32][33][34][35][36]46]: where, J i = model coefficient and this value was determined using no-intercept regression analysis, described in the literature [35,47]. The regression between experimental and JAM solubility of TQ was performed using RMSD.
The JAVHM model is based on the combination of VHM and JAM models. It also predicts the solubility of solutes in cosolvent mixtures at the given set of temperatures. The JAVHM solubility (x 3 JAVHM ) of TQ in binary (IPA + H 2 O) compositions were predicted using Equation (17) [48]: where, A 1 , B 1 , A 2 , B 2 , and J i = the model coefficients. The regression between experimental and JAVHM solubility of TQ was performed using RMSD.

Statistical Evaluation
Statistical evaluation was performed by applying a Kruskal-Wallis test, followed by Denn's test using "GraphpadInstat software (Version 9.1.1, San Diego, CA, USA)". The p < 0.05 or equivalent to 0.05 was taken as a significant value.

Conclusions
In this research, the solubility of a naturally-derived bioactive compound TQ in binary (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O, was evaluated at 298.2-318.2 K and 0.1 MPa. The TQ solubility was found to be increasing significantly with the rise in absolute temperature and IPA mass fraction in all (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O. The experimental solubility values of TQ were regressed well with six different theoretical models, including VHM, MAM, YRM, BKλhM, JAM, and JAVHM models. The results of activity coefficient calculations indicated maximum solutesolvent interactions in the TQ-IPA combination compared to the TQ-H 2 O combination. A thermodynamic study suggested an endothermic and entropy-driven dissolution of TQ in all (IPA + H 2 O) mixtures, including pure IPA and pure H 2 O. The obtained solubility data can be utilized for the drug discovery process, formulation development, and preclinical assessment of TQ.