Equilibrium solubility determination, solvent effect and preferential solvation of amoxicillin in aqueous co-solvent mixtures of N,N-dimethylformamide, isopropanol, N-methyl pyrrolidone and ethylene glycol

J. Chem. Thermodynamics 142 (2020) 106010

Contents lists available at ScienceDirect

J. Chem. Thermodynamics journal homepage: www.elsevier.com/locate/jct

Equilibrium solubility determination, solvent effect and preferential solvation of amoxicillin in aqueous co-solvent mixtures of N,N-dimethylformamide, isopropanol, N-methyl pyrrolidone and ethylene glycol Wanxin Li a, Ali Farajtabar b, Rong Xing a, Yiting Zhu a, Hongkun Zhao c,⇑ a

School of Chemistry and Environmental Engineering, Yancheng Teachers University, Yancheng, Jiangsu 224002, People’s Republic of China Department of Chemistry, Jouybar Branch, Islamic Azad University, 4776186131 Jouybar, Iran c College of Chemistry & Chemical Engineering, YangZhou University, YangZhou, Jiangsu 225002, People’s Republic of China b

a r t i c l e

i n f o

Article history: Received 17 October 2019 Received in revised form 12 November 2019 Accepted 13 November 2019 Available online 15 November 2019 Keywords: Amoxicillin Solubility Jouyban-Acree Solvent effect Preferential solvation

a b s t r a c t The mole fraction solubility of amoxicillin in four co-solvent mixtures of N,N-dimethylformamide (DMF) + water (2), isopropanol (1) + water (2), N-methyl pyrrolidone (NMP) (1) + water (2) and ethylene glycol (EG, 1) + water (2) at temperatures from 278.15 K to 328.15 K was determined by means of the shakeflask technique. At the same temperature and composition of DMF, isopropanol, NMP or EG, the solubility magnitude of amoxicillin was highest in the DMF (1) + water (2) mixture, and lowest in the isopropanol (1) + water (2) mixture. Through the Jouyban-Acree model, amoxicillin solubility was well correlated obtaining RAD lower than 4.55% and RMSD lower than 1.96
10-4. Quantitative values for the local mole fraction of DMF (isopropanol, NMP or EG) and water nearby the amoxicillin were computed by means of the Inverse Kirkwood–Buff integrals method. In the DMF (1) + water (2) mixture with compositions 0.20 < x1 < 1.00, NMP (1) + water (2) mixture with compositions 0.165 < x1 < 1.00 and EG/isopropanol (1) + water (2) mixtures with compositions 0.25 < x1 < 1.00, amoxicillin was solvated preferentially by the co-solvent. In addition, solvent effect was modeled by linear solvation energy relationships in terms of KAT solvent polarity descriptors to detect the main intermolecular interactions controlling the solubility variation in solvent mixtures. Results showed that the work for cavity formation in solvent for solute’s accommodation had the most significant effect on solubility variance over the entire composition range in all the mixed solvents. Ó 2019 Elsevier Ltd.

1. Introduction In recent times, research upon solubility of the pharmaceutical intermediates and drugs has become a growing focus in the pharmaceutical fields. The solubility of pharmaceutical intermediates and drugs in co-solvent solutions is an essential physicochemical property, which plays a major role in numerous physical and biological processes [1–4]. In addition, the solubility in aqueous cosolvent mixtures as a function of temperature and composition is important for purification of raw material and understanding the mechanisms concerning physical and chemical stability of the solid dissolutions [3,5,6]. It is usually regarded as a crucial factor in designing a crystallization process, wherein the knowledge of solubility is needed to control the desired polymorphic form, ⇑ Corresponding author. E-mail address: [email protected] (H. Zhao). https://doi.org/10.1016/j.jct.2019.106010 0021-9614/Ó 2019 Elsevier Ltd.

supersaturation, yield and particle size. Co-solvency is an effective and optional solubilization method which is considered to change the solid solubility, as aqueous co-solvent mixtures can be employed as reaction medium of many substances [1,6]. Low solubility in water is likely to result in formulation difficulty or low bioavailability during the clinical development [1,5,7]. Furthermore, drugs solubility in co-solvent mixtures allows us to carry out a thermodynamic analysis to deeply understand the molecular mechanisms relating to the drug dissolution process and evaluate the preferential solvation of a solute by the co-solvent components in solutions [8–11]. Amoxicillin (CAS Reg. No. 26787-78-0, structure shown in Fig. 1) is a commonly used penicillin antibiotic [12–15]. It is used to treat many different types of infection caused by bacteria, such as tonsillitis, bronchitis, pneumonia, gonorrhea, and infections of the ear, nose, throat, skin, or urinary tract. Amoxicillin is also sometimes used together with clarithromycin (Biaxin) to treat

2

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

2.1. Jouyban-Acree model The Jouyban-Acree model, described as Eq. (1), may provide precise mathematical description for the dependence of solute solubility on both solvent composition and temperature for cosolvent solutions [34,35].

lnxw;T ¼ w1 lnx1;T þ w2 lnx2;T þ Fig. 1. The chemical structure of amoxicillin.

stomach ulcers caused by Helicobacter pylori infection. However, the amoxicillin solubility in water is low [16–21]. In the previous works, many methods, such as surfactant addition, cosolvency, complexation, and pH adjustment and so on, are used to solubilize drug candidates with low aqueous solubility [1,5]. The most powerful and effective tool is mixing a miscible and safe co-solvent with water [1–3,5,16–20]. Nevertheless, regardless of the usefulness of amoxicillin, the information on physicochemical properties, e.g. solubility, in mono-solvents and solvent mixtures are very scarce. A thorough literature search indicates that only the amoxicillin solubility in ethanol + water and sodium chloride + water mixtures and supercritical carbon dioxide at several temperatures is available [21–27]. However, the physicochemical property of amoxicillin in solvent mixtures has not yet been investigated systematically. This case stimuluses us to make in-depth research on the solubility and the solute–solvent and solvent–solvent interactions of amoxicillin in aqueous co-solvent mixtures. For the co-solvency method, solvent selection is a vital process. The practicable solvents should be environmentally safe, commercially available, non-corrosive and thermally stable. In the pharmaceutical fields, the commonly employed co-solvents are N,N-dimethylformamide (DMF), ethanol, isopropanol, ethylene glycol (EG), dimethyl sulfoxide (DMSO) and so forth [1,3,5,6]. Isopropanol is a flammable compound with a strong odor. It dissolves a wide range of non-polar compounds. Compared to other alternative solvents, isopropanol is relatively non-toxic. It is often employed solely or in mixtures with other solvents for diverse aims containing in penetration-enhancing pharmaceutical compositions for the topical percutaneous and transepidermal uses [28,29]. DMF is a commonly used co-solvent in investigating the interrelation between solubility of drugs and medium polarity [30]. The solutions of DMF-water present very strong non-ideal, so it may serve in the solute solvation procedure through hydrophobic interactions and preferential solvation [31]. NMP is also a common co-solvent in pharmaceutical industry. It has strong solubilization capacity and is an important solvent in the extraction, crystallization and purification processes of many drugs [32]. EG is a safe and pharmaceutically acceptable solvent for industrial applications [33]. Based on these points-of-view, the main objectives of the present work are to report the solubility of amoxicillin in aqueous co-solvent mixtures of DMF, isopropanol, NMP and EG at temperature range from 278.15 to 328.15 K under atmospheric conditions and estimate some thermodynamic quantities of the mixtures.

2. Theoretical aspects In the present paper, the Jouyban-Acree model [34,35] is employed to mathematically describe the amoxicillin solubility in aqueous co-solvent solutions of DMF, isopropanol, NMP and EG. Moreover, the KAT-LSER model is used to examine the solvent effect on the amoxicillin solubility [36–39].

2 w1 w2 X J ðw1  w2 Þi T=K i ¼ 0 i

ð1Þ

herein xw,T is the mole fraction solubility of amoxicillin in mixtures at temperature T/K; w1 and w2 refer to, respectively, the mass fraction of co-solvents 1 (DMF, isopropanol, NMP or EG) and 2 (water) in the co-solvent solutions free of amoxicillin; x1,T and x2,T stand for the mole fraction solubility of amoxicillin in neat solvents at T; and Ji are the model parameters. 2.2. Solvent effect In developing models based on quantitative structure–property relationships for solvent effect, considerable efforts have been devoted to quantitative description of the solvent polarity [39]. In this regard, Kamelt and Taft et al. introduce their well-known empirical model based on the linear solvation energy relationships concept [36–38]. This model, abbreviated KAT-LSER hereafter, divides the total change in free energy induced by the solvent into some separated intermolecular interaction energy terms. These terms account for both specific (e.g. hydrogen bonding) and nonspecific electrostatic interactions (such as Keesom dipole–dipole, Debye dipole-induced dipole and London instantaneous induced dipole–dipole dispersion) that might occur between solute and solvent molecules. Three empirical solvent parameters named p*, b and a have been introduced to describe the feature of the solvent at the molecular level. According to Kamlet et al., p * represents dipolarity/polarizability as a scale to characterize the solvent’s ability for non-specific interactions; b and a symbolize the capacity of solvent to act as a hydrogen-bond acceptor and hydrogen-bond donor in specific interactions, respectively [40–42]. KAT parameters are derived by the solvatochromic comparison method from a direct measurement of a change in the solute’s electronic transition energy due to corresponding solvent effect. Hence, a linear correlation is expected between interaction energy terms defined by KAT parameters with the change induced by the solvent in Gibbs energy of a property (e.g. solubility). Therefore, the examination of KAT-LSER model gives opportunity to extract detailed information about the nature and significance of different solvation components playing in the solvent effect. In this way, another objective of this paper is to analyze the solvent effect upon the solubility variation of amoxicillin in aqueous co-solvent mixtures of DMF, EG, NMP and isopropanol through the KAT-LSER model in order to explain the relative importance and the nature of intermolecular interactions that bring about the solvent effect. The general form of KAT-LSER model relates the Gibbs energy of solvation of a given solute (presented proportionally by the natural logarithmic form of the mole fraction solubility, lnx) in a linear correlation to different solute–solvent and solvent–solvent interaction energy terms as Eq. (2) [36–38].

lnxw ¼ c0 þ c1 p  þc2 b þ c3 a þ c4

V s d2H 100RT

! ð2Þ

here, the terms c1 p*, c2 b and c3 a relate to the energy for nonspecific and specific solvent–solute interactions; The coefficients ci=1-3 demonstrate sensitivity of the solute solubility to respective V d2

s H in the Eq. (2) refers to the cavity energy term. The last term 100RT term that defines the energy term for solvent–solvent interactions.

3

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

This cavity term evaluates the energy of solute’s accommodation as a product of solute molar volume, Vs, and squared Hildebrand solubility parameter, d2H, of the solvent. The gas constant R and temperature T are considered here in denominator to offer a dimensionless value for the cavity term. Consequently, the coefficient c4 discloses susceptibility of the solute solubility to solvent–solvent interactions. In practice, the mole fraction solubility are first converted to lnxw, and then mathematically correlated to Eq. (2) in terms of multiple linear regression analysis. The objective function used here is

F¼

2 X lnxew;T  lnxcw;T

ð3Þ

In addition, the relative average deviation (RAD) and rootmean-square deviation (RMSD), which are described as, respectively, Eqs. (4) and (5), are employed herein in order to evaluate the Jouyban-Acree model.

1 0 c e  1 X @xw;T  xw;T A RAD ¼ N xew;T

RMSD ¼

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uN uP c u ðxw;T  xew;T Þ2 ti¼1 N

ð4Þ

ð5Þ

where, N refers to the number of data points. xew;T denotes the mole fraction solubility of amoxicillin determined in the present paper; and xcw;T , the mole fraction solubility computed by using the Jouyban-Acree model.

with 



x2 V 2 D Q



x1 V 1 D Q

G1;3 ¼ RT jT  V 3 þ

G2;3 ¼ RT jT  V 3 þ

ð9Þ



ð10Þ

 3    1=3  0:085 V cor ¼ 2522:5 r 3 þ 0:1363 xL1;3 V 1 þ xL2;3 V 2

ð11Þ

where in, jT signifies isothermal compressibility of the co-solvent 



(1) + water (2) mixtures; V 1 and V 2 refer to the partial molar volumes of the components in the aqueous co-solvent mixtures; and 

V 3 , the partial molar volume of amoxicillin in the solutions. The function D expressed as Eqn. (12) is the first derivative of standard molar Gibbs energy of transfer of amoxicillin from neat water (2) to the co-solvent (1) + water (2) mixtures with respect to the cosolvent composition. The function Q expressed as Eqn. (13) is the second derivative of excess molar Gibbs energy of mixing of the two solvents (GExc 1þ2 ) with respect to proportion of water in the cosolvent mixtures. Vcor denotes the correlation volume; and r3, the molecular radius of amoxicillin computed by using Eq. (14), where NAv is the Avogadro’s number.

D¼

@ Dtr Goð3;2!1þ2Þ

!

@x1

ð12Þ T;P

" # @ 2 GExc 1þ2 Q ¼ RT þ x1 x2 @x22

ð13Þ

T;p

2.3. Preferential solvation The Inverse Kirkwood–Buff integrals method {Eq. (6)} describes the local composition of solvent nearby a nonelectrolyte or a weak electrolyte compared with the global solution compositions [8– 11]. It is a valuable tool to study the preferential solvation of nonelectrolyte or weak electrolyte in the aqueous co-solvent mixtures.

Z

Gi;3 ¼ 0

r cor

ðg i;3  1Þ4pr 2 dr

ð6Þ

herein, gi,3 refers to the pair correlation function for solvent i in the co-solvent (1) + water (2) mixtures nearby the amoxicillin (3); r denotes the distance between the molecule centers of co-solvent (1) or water (2) and that of amoxicillin (3); rcor is the correlation distance, for which gi,3 (r > rcor) approximately equals to 1. Thus, for r > rcor, the integral value is essentially zero. The preferential solvation parameter (dx1,3) of amoxicillin (compound 3) by the co-solvent (compound 1) in the solutions of cosolvent (1) + water (2) is expressed as [8–11].

dx1;3 ¼ xL1;3  x1 ¼ dx2;3

ð7Þ

wherein xL1;3 refers to the local mole fraction of co-solvent (1) adjacent the amoxicillin (3); and x1, the bulk mole fraction of cosolvent (1) in the initial solutions. If the dx1,3 value is higher than zero, the solute (amoxicillin) is preferentially solvated by the cosolvent (1); while its value is lower than zero, then the solute is said to be preferentially solvated by water (2). The values of dx1,3 can be obtained from the Inverse Kirkwood–Buff integrals method for the individual solvent composition [8–11].

dx1;3 ¼

x1 x2 ðG1;3  G2;3 Þ x1 G1;3 þ x2 G2;3 þ V cor

ð8Þ

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  21 3 3
10 V3 r3 ¼ 4pNAV

ð14Þ

Due to the dependence of jT upon the co-solvent compositions, so the term containing jT is not known for the systems investigated. Alternatively, the contribution of the term RTjT to Inverse Kirkwood–Buff integrals is very small; consequently, the jT values can be approximately calculated through considering an additive property via the compositions of co-solvent solution and the available values for neat solvents by [8–11]:

jT ¼ x1 joT;1 þ x2 joT;2

ð15Þ

where in xi are the mole fractions of component i in the aqueous cosolvent solutions; joT;i refers to isothermal compressibility of the neat component i. Therefore the values of RTjT are attainable in terms of the available joT;i values for DMF (0.653 GPa1), EG (0.364 GPa1), NMP (0.523 GPa1), isopropanol (1.332 GPa1) and water (0.457 GPa1) at 298.15 K [43], taken as independent of the temperature [8].

 x3;2 Dtr Go3;2!1þ2 ¼ RTln x3;1þ2

ð16Þ

The values of Dtr Go3;2!1þ2 are mathematically correlated by using an empirical Eq. (17). x

t 1

Dtr Go3;2!1þ2 ¼ A0 þ A1 e

1

x

t 1

þ A2 e

2

ð17Þ

here A0, A1, t1, A2 and t2 refer to the equation parameters. The definitive correlation volume requires iteration since it is dependent upon the local mole fractions nearby the solute. The iteration can be made by substituting dx1,3 and Vcor in Eqs. (7) and (8) to Eq. (11) to re-calculate the xL1;3 until a non-variant Vcor value is attained.

4

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

3. Experimental 3.1. Materials Amoxicillin was bought from the Sigma Chemical Co., Ltd, China with a mass fraction of 0.982. It was re-crystallized three times in pure ethanol. The final purity of amoxicillin in mass fraction was 0.996 confirmed through a high-performance liquid chromatography (HPLC, Agilent 1260). The organic solvents of EG, DMF, NMP and isopropanol were purchased from the Sinopharm Chemical Reagent Co., Ltd., China, which compositions in mass fraction were all no smaller than 0.994 identified by using a gas chromatography {Smart (GC-2018)}. The twice-distilled water had a conductivity of 2 mS cm1, which was prepared in our laboratory. The detailed descriptions of these substances were summarized as Table 1. 3.2. Solubility measurement The analytical balance having a type of BSA224S was used to prepare the aqueous mixtures. Each mixture in experiment was around 15 mL, which relative standard uncertainty was assessed to be 0.0002. The mass fraction compositions of co-solvent in the mixtures covered the range from 0 to 1. During the experiment procedure, the atmospheric pressure was approximately 101.2 kPa. The determination of amoxicillin solubility in the co-solvent solutions of (DMF + water), (NMP + water), (isopropanol + water) and (EG + water) was performed by using a saturation shakeflask technique [44,45], which reliability was checked by the solubility of benzoic acid in neat toluene [45]. The amoxicillin composition in equilibrium liquor was determined through the Agilent-1260 HPLC. The experiments of amoxicillin solubility determination was carried out at temperatures ‘‘T = 278.15 K to 328.15 K” every 5 K and ambient pressure ‘‘p = 101.2 kPa”. Excess solid amoxicillin was added to a certain amount of aqueous mixtures of co-solvent in triplicates. Each amoxicillin solution was fully mixed and then placed into a thermostatic shaker purchased from the Tianjin Ounuo Instrument Co. Ltd., China. The prepared solution was shaken by using the shaker at a speed of 100 rpm. In an attempt to get equilibration time, 0.5 mL of liquor was withdrawn with a 2 mL syringe every 1 h and then analysed by using the Agilent-1260 HPLC. Results indicated that 16 h was adequate for all the solutions to arrive at equilibrium. After that all solutions were removed from the shaker and allowed to precipitate all undissolved amoxicillin particles from the liquid for 3 h. The liquid was taken out carefully, diluted (if necessary) and texted by using the HPLC. 3.3. Analysis method The composition of amoxicillin in equilibrium liquid was analyzed by the Agilent-1260 HPLC. The UV detector wavelength

was set to 260 nm [46]. The chromatographic column used in determination was a reverse phase column having a type of LPC18 (250 mm
4.6 mm), which temperature was 303 K. The flow rate of mobile phase, neat methanol, was 1.0 mL min1. Each test was carried out three times, and the average of three tests was regarded as the final solubility of the analysis. The relative standard uncertainty is estimated to be no more than 0.0603 for mole fraction solubility. 3.4. X-ray powder diffraction So as to check no existence of the polymorph transformation or solvate formation of amoxicillin during experiments, the solid equilibrated with liquid was identified by using X-ray powder diffraction (XRD). All determinations were carried out upon a HaoYuan DX-2700B (HaoYuan, China) instrument by Cu Ka radiation (k = 0.154184 nm) at a scan speed of 6 degmin1 under local pressure. The tube current and voltage were, respectively, 30 mA and 40 kV. The data were gathered from 10° to 80° (2-Theta). 4. Results and discussion 4.1. X-ray powder diffraction analysis The obtained XRD patterns of raw amoxicillin together with the solids in equilibration with liquids are graphically shown in Fig. S1 of Supporting Material. It is shown that all XRD patterns of solid of amoxicillin in equilibrium with its mixture present the characteristic peaks similar with the raw amoxicillin. Therefore, the solvate formation or polymorph transformation doesn’t occur during the entire experiments. As is evident that amoxicillin may present the form of amoxicillin trihydrate in aqueous solutions [47,48]. However no amoxicillin trihydrate was found during the entire process of experiments. Perhaps the amoxicillin trihydrate is formed under certain conditions, such as adding an acidic solution into an aqueous solution of crude amoxicillin, and performing dissolved clarification to obtain a dissolved-clarified aqueous solution of the crude amoxicillin; performing cooling on the dissolvedclarified aqueous solution of the crude amoxicillin from 291 to 303 K to 273–278 K within 0.5–1.5 h, and adjusting the pH to 5– 6 in the cooling process [47]. Nevertheless, the equilibrium solid phase of amoxicillin in equilibrium with its aqueous solution is amoxicillin in the present work. 4.2. Solubility data The determined solubility of amoxicillin in mole fraction in mixtures of (DMF + water), (NMP + water), (EG + water) and (isopropanol + water) is listed in Tables 2–5, respectively. As well, the relationships between the temperature and solvent composition and the mole fraction solubility are graphically shown in

Table 1 Comprehensive information on the materials used in the work.

a b

Chemicals

Molar mass/ gmol1

Source

Initial mass fraction purity

Final mass fraction purity

Purification method

Analytical method

amoxicillin isopropanol DMF NMP EG water

365.4 60.10 73.09 99.13 62.07 18.02

Sigma Chemical Co., Ltd Sinopharm Chemical Reagent Co., Ltd.

0.982 0.995 0.995 0.995 0.994

0.996 0.995 0.995 0.995 0.994 Conductivity < 2 mScm1

Recrystallization none none none none Distillation

HPLCa GCb GC GC GC Conductivity meter

Our lab

High-performance liquid phase chromatography. Gas chromatography.

5

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

Table 2 Mole fraction solubility (xeT;W  104 ) of amoxicillin in mixed solvent of DMF (w) + water (1-w) with various mass fractions within the temperature range from T/K = (278.15 to 328.15) under p = 101.2 kPa.a xeT;W  104 T/K

w 0

0.1000

0.2000

0.3000

0.4000

0.5000

0.6000

0.7000

0.8000

0.9000

1

278.15 283.15 288.15 293.15 298.15 303.15 308.15 313.15 318.15 323.15 328.15

0.6174 0.7245 0.8497 1.021 1.252 1.516 1.805 2.125 2.512 3.021 3.572

1.099 1.414 1.740 2.144 2.576 3.052 3.614 4.227 4.952 5.795 6.875

1.699 2.187 2.651 3.253 3.932 4.848 5.797 6.788 7.853 9.153 10.86

2.568 3.306 4.111 5.184 6.209 7.418 8.784 10.26 12.09 14.23 16.74

3.783 4.768 6.021 7.459 9.244 11.22 13.35 15.81 18.79 21.93 25.57

6.178 7.795 9.781 12.27 14.88 17.88 21.27 25.47 30.26 35.71 41.28

10.34 13.09 16.32 19.92 23.59 28.07 32.97 38.46 44.32 50.98 58.65

17.21 22.19 27.35 32.13 37.52 43.34 49.54 56.99 65.03 74.84 86.54

27.32 34.35 42.75 49.92 57.64 67.62 78.33 89.35 100.8 113.7 127.3

40.33 48.51 60.32 70.44 81.25 95.67 109.6 121.4 137.3 150.9 167.5

55.23 63.93 76.07 89.84 105.6 121.8 138.9 156.7 175.7 198.2 223.2

a Standard uncertainties u are u(T) = 0.02 K, u(p) = 0.45 kPa; Relative standard uncertainty ur is ur (x) = 0.0603. Solvent mixtures were prepared by mixing different masses of the solvents with relative standard uncertainty ur(w) = 0.0002. w represents the mass fraction of DMF in mixed solvents of DMF(w) + water (1-w).

Table 3 Mole fraction solubility (xeT;W  104 ) of amoxicillin in mixed solvent of NMP (w) + water (1-w) with various mass fractions within the temperature range from T/K = (278.15 to 328.15) under p = 101.2 kPa.a xeT;W  104 T/K

278.15 283.15 288.15 293.15 298.15 303.15 308.15 313.15 318.15 323.15 328.15

w 0

0.1000

0.2000

0.3000

0.4000

0.5000

0.6000

0.7000

0.8000

0.9000

1

0.6174 0.7245 0.8497 1.021 1.252 1.516 1.805 2.125 2.512 3.021 3.572

0.9271 1.173 1.522 1.866 2.281 2.715 3.239 3.762 4.377 5.047 5.828

1.423 1.767 2.281 2.841 3.482 4.211 4.954 5.807 6.816 7.916 9.404

2.041 2.592 3.312 4.077 5.008 5.943 7.135 8.488 9.988 11.91 13.34

2.868 3.724 4.758 5.857 7.041 8.357 9.982 11.82 13.73 15.84 18.67

4.324 5.583 6.983 8.415 10.04 11.82 13.91 16.25 18.83 21.78 25.43

6.311 7.701 9.508 11.51 13.82 16.23 19.11 21.29 24.73 28.72 33.61

9.067 10.74 13.31 16.21 19.61 23.54 27.51 31.44 35.64 41.28 49.02

13.02 16.62 20.69 24.62 29.21 34.46 39.70 45.21 52.43 59.25 68.93

19.41 24.46 29.55 35.65 42.81 50.33 58.62 67.62 76.94 88.44 99.64

26.62 33.84 43.23 52.11 61.15 70.61 80.92 92.05 105.1 117.9 131.2

a Standard uncertainties u are u(T) = 0.02 K, u(p) = 0.45 kPa; Relative standard uncertainty ur is ur (x) = 0.0603. Solvent mixtures were prepared by mixing different masses of the solvents with relative standard uncertainty ur(w) = 0.0002. w represents the mass fraction of NMP in mixed solvents of NMP (w) + water (1-w).

Table 4 Mole fraction solubility (xeT;W  104 ) of amoxicillin in mixed solvents of EG (w) + water (1-w) with various mass fractions within the temperature range from T/K = (278.15 to 328.15) under p = 101.2 kPa.a xeT;W  104 T/K

278.15 283.15 288.15 293.15 298.15 303.15 308.15 313.15 318.15 323.15 328.15

w 0

0.1000

0.2000

0.3000

0.4000

0.5000

0.6000

0.7000

0.8000

0.9000

1

0.6174 0.7245 0.8497 1.021 1.252 1.516 1.805 2.125 2.512 3.021 3.572

0.9271 1.232 1.511 1.884 2.271 2.675 3.192 3.771 4.447 5.221 6.197

1.332 1.693 2.163 2.721 3.384 4.075 4.887 5.807 6.807 7.952 9.287

1.795 2.334 2.977 3.701 4.611 5.619 6.775 8.053 9.553 11.37 13.44

2.579 3.208 4.099 5.154 6.278 7.605 9.324 11.29 13.48 16.15 18.95

3.626 4.609 5.765 7.096 8.828 10.93 13.24 15.66 18.57 21.95 25.65

4.992 6.345 7.936 9.361 11.69 14.41 17.72 21.28 25.64 30.65 36.63

7.326 9.115 11.42 13.74 16.73 19.84 24.55 29.91 35.46 43.42 56.08

10.34 12.82 16.38 20.16 24.04 29.75 36.04 42.96 51.15 62.38 78.86

14.18 18.03 22.07 27.76 33.84 41.83 49.61 59.99 73.25 91.55 113.3

20.23 25.62 31.88 40.32 50.14 60.07 72.74 86.81 105.2 128.7 160.2

a Standard uncertainties u are u(T) = 0.02 K, u(p) = 0.45 kPa; Relative standard uncertainty ur is ur (x) = 0.0603. Solvent mixtures were prepared by mixing different masses of the solvents with relative standard uncertainty ur(w) = 0.0002. w represents the mass fraction of EG in mixed solvents of EG (w) + water (1-w).

Fig. 2. It reveals that, for the four co-solvent mixtures studied, the amoxicillin solubility presents a function of solvent composition and temperature. The mole fraction solubility of amoxicillin increases with an increase in temperature and mass fraction of DMF (isopropanol, NMP or EG). The maximum solubility of amoxicillin in mole fraction appears in neat DMF (isopropanol, NMP or EG). It can also be observed from the Tables 2–5 that the amoxicillin solubility in (DMF + water) mixture is larger than those in the other three co-solvent mixtures at the same co-solvent composition and temperature.

The solubility of amoxicillin in neat water determined in this paper as well as that obtainable in the literatures [21–26] is graphically shown in Fig. S2 of Supporting material for comparison. It is found that the mole fraction solubility of amoxicillin in water at 298.15 K determined by us is 1.252
104, which is very close to 1.2006
104 reported by Felix [21] and 1.23
104 reported by Diender [26]. Some large difference is found between the solubility attained by us and the reported ones by Dave [22], Rudolph [23] and Francesco [25], which are 1.724
104, 1.7
104 and 0.7413
104, respectively. However at the other temperatures,

6

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

Table 5 Mole fraction solubility (xeT;W  104 ) of amoxicillin in mixed solvents of isopropanol (w) + water (1-w) with various mass fractions within the temperature range from T/K = (278.15 to 328.15) under p = 101.2 kPa.a T/K

278.15 283.15 288.15 293.15 298.15 303.15 308.15 313.15 318.15 323.15 328.15

w 0

0.1000

0.2000

0.3000

0.4000

0.5000

0.6000

0.7000

0.8000

0.9000

1

0.6174 0.7245 0.8497 1.021 1.252 1.516 1.805 2.125 2.512 3.021 3.572

0.9172 1.197 1.521 1.844 2.223 2.685 3.202 3.839 4.481 5.204 6.053

1.269 1.626 2.084 2.682 3.334 4.061 4.951 5.941 7.003 8.351 9.748

1.724 2.172 2.781 3.515 4.364 5.322 6.608 7.913 9.649 11.35 13.72

2.301 2.948 3.644 4.525 5.724 6.853 8.353 10.19 12.21 14.35 17.03

3.073 3.864 4.776 5.824 7.231 8.661 10.56 12.88 14.87 17.81 21.14

4.024 4.974 6.148 7.363 8.978 10.75 12.88 15.73 18.47 22.12 25.78

5.087 6.176 7.913 9.308 11.56 13.84 16.88 20.21 23.77 28.47 33.19

6.667 8.241 10.23 12.21 15.15 18.47 21.33 25.55 30.62 35.99 41.95

8.429 10.84 13.35 16.28 19.52 23.35 27.46 32.89 38.68 45.51 54.36

10.29 13.18 16.29 19.58 24.48 29.59 36.17 42.96 51.33 61.32 72.25

a Standard uncertainties u are u(T) = 0.02 K, u(p) = 0.45 kPa; Relative standard uncertainty ur is ur (x) = 0.0603. Solvent mixtures were prepared by mixing different masses of the solvents with relative standard uncertainty ur(w) = 0.0002. w represents the mass fraction of isopropanol in mixed solvents of isopropanol (w) + water (1-w).

Fig. 2. Mole fraction solubility (x) of amoxicillin in (a) DMF (w) + water (1-w), (b) NMP (w) + water (1-w), (c) EG (w) + water (1-w) and (d) isopropanol (w) + water (1-w) solutions with various mass fractions at different temperatures: w, mass fraction; j, w = 0; d, w = 0.1000; ▲, w = 0.2000; ◆, w = 0.3000; ., w = 0.4000; w, w = 0.5000; 4, w = 0.6000; s, w = 0.7000; q, w = 0.8000; J, w = 0.9000; h, w = 1. —, calculated curves by the Jouyban  Acree model.

the difference is very small. The deviation is perhaps caused by lots of factors, e.g. equilibration time, determination method, purity, analysis method, sampling and so on. It should be noted that the pH value of amoxicillin is generally in the range from 3.5 to 5.5 [49]. The pH of amoxicillin (>98.3% in mass fraction purity) on market is 5.2–5.3 [50]. More importantly, the solubility data vary little within the pH values from 4 to 6 based on the Refs. [21], [24] and [26] and so forth. So the amoxicillin solubility reported in the

literatures at the value of pH = 5.2 is used for comparison in each case in the present work. 4.3. Solvent effect The first step in KAT-LSER analysis of solvent effect is to obtain dHand KAT parameters a, b and p * for mixtures and Vs for amoxicillin. The KAT parameters a, b and p * can be available for (DMF

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

+ water), (EG + water), (NMP + water) and (isopropanol + water) in literatures [51–54]. The gathered data shows slight discrepancy for pure solvents. Thus, for a homogenous tabulation, the accepted KAT parameters for pure solvents in Ref. [55] are used as reference points, and then the initial data for mixtures are corrected by the method explained previously [56,57]. Table S1 of Supporting material presents these data. For binary mixtures, dHis not available, however it can be readily estimated by /1dH1 + /2dH2 from the volume fraction, /i, and dHi of pure solvents [8]. For the neat solvents, dH is available from Hansen [58]. The dHvalues for binary mixtures are also presented in Table S1. The molar volume of amoxicillin is 236.2 cm3mol1 taken from the Scifinder database [59]. Eq. (2) presents the general form of KAT-LSER model having all molecular descriptors. Therefore, for a comprehensive analysis, we made 15 expressions of KAT-LSER from different combination of solvent descriptors. These equations are fitted to experimental solubility data of amoxicillin in binary solvent mixtures by using the multiple linear regression analysis, MLR. The regressed results are presented in Tables S2–S5 of Supporting Material. We know that any solute–solvent interaction has a positive effect on the solubility, thereby the value of ci corresponding to a, b and p * is expected to be positive in the KAT-LSER models. In contrast, solvent–solvent interactions do play negative role in the solute solubility, because the hole creation within the solvent is an energy-consuming step in the solubilization process, and therefore c4 is expected to be negative in a proper KAT-LSER model. Among models meet these physical criteria, the best is one that shows statistically the highest F-statistic. The best KAT-LSER model for each binary mixture is bolded in Tables S2–S5. Table S2 shows that the solubility of amoxicillin is well correlated with a, b and cavity term in three-parametric, and with b and p * or b and cavity term in two-parametric KAT-LSER model in water + DMF mixtures. However, the standard deviations obtained for their coefficients, particularly for c1 and c3, are very large, thereby the corresponding expressions are improper in description of solvent effect. It means that a and p * appear to have less significant effect on the variance of solubility in the mixtures. There are also excellent correlations between lnx with b and cavity term in two separate single-parametric KAT-LSER model. A comparison of intercept in these two expression helps, along with above-mentioned criteria, to identify the best descriptive model. The intercept coefficient, c0, can be assumed as the predicted ideal solubility, lnxid, of the solute. This quantity can be determined by Eq. (18) [60].

lnxid ¼ 

  DH m T m  T DC p T m  T Tm  ln þ T mT T R R T

ð18Þ

herein, DHm, Tm and DCp denote the enthalpy of fusion, melting point and difference in heat capacity of the solute between its solid and liquid forms, respectively. The second term on the right hand side of Eq. (18) has small impact on solubility, and can safely be neglected. In addition, DHm/Tm is equal to the entropy of fusion, DSm, at the melting point. Thus, Eq. (18) can be written as Eq. (19)

lnxid ¼ 

 DSm T m  T T R

ð19Þ

The melting point of amoxicillin is 467 K [61]. No experimental data on DSm of amoxicillin is available; however, it can be approximated at 56.5 JK1mol1 by Walden’s rule [62]. Introducing these data into Eq. (19) results in lnxid = 3.85 for amoxicillin. The intercept c0 for single-parametric KAT-LSER model involving b is abnormally low considering the already obtained lnxid. In contrast, the single-parametric KAT-LSER model involving cavity term

7

provides far better result. This leads to the conclusion that solvent–solvent interactions play, effectively, the main role in solubility variation of amoxicillin in water + DMF mixture. Tables S3–S5 show similar results for analysis of solubility of amoxicillin in other binary mixtures. In all binary mixtures, the single-parametric KAT-LSER model involving cavity term is the best expression, and explains more than 99% of solubility variance over the entire composition range in aqueous mixtures of NMP, EG and isopropanol. In compliance with this conclusion, the size of amoxicillin is relatively large, thus it is not surprising that the work for cavity formation within the solvent structure for accommodation of amoxicillin dominates its solubility. 4.4. Solubility modeling Based on the solubility data of amoxicillin in different mixtures, the model parameters in Eq. (1) can be attained via the Mathcad software. The acquired equation parameters as well as the values of RAD and RMSD are tabulated in Table S6 of the Supporting material. The amoxicillin solubility in the four mixtures is backevaluated in terms of the parameters’ values and plotted in Fig. 2. It can be realized from Table S6 that the values of relative average deviations (RAD) and root-mean-square deviation (RMSD) for the four co-solvent mixtures are relative small. The largest values are, respectively, 4.55
10-2 and 1.96
10-4 for RAD and RMSD. In general, the Jouyban  Acree model may be employed to describe the amoxicillin solubility in the binary mixtures of DMF + water, EG + water, isopropanol + water and NMP + water at all initial composition ranges. 4.5. Preferential solvation of amoxicillin The computed values of Dtr Go3;2!1þ2 at 298.15 K by using the Eq. (16) are presented in Table S7 of Supplementary material. In addition, the Dtr Go3;2!1þ2 values are plotted in Fig. S3 of Supporting material. These values are mathematically correlated with the empirical equation {Eq. (17)} for the four co-solvent solutions. The obtained curve-fitting coefficients are tabulated in Table S8 of Supplementary material. So, the values of D are evaluated from the first derivative of Eq. (17) solved in terms of the co-solvent solution composition varying by 0.05 in mole fraction of co-solvent (1) and tabulated in Tables S9–S12 of Supporting material. The molar volume of amoxicillin is regarded as independent upon the studied temperatures and co-solvent compositions. By reason of no partial molar volumes of amoxicillin (3) in the studied mixtures available in publications, in this paper, it is regarded as similar to that for the neat amoxicillin [8,10,11,56,57]. Based on the molar volume of amoxicillin, solute radius value (r3) is calculated by using Eq. (14) as 0.454 nm. Additionally, the values of RTjT and partial molar volumes of two neat solvents in the DMF (1) + water (2), isopropanol (1) + water (2), NMP (1) + water (2) and EG (1) + water (2) mixtures along with the Q values at 298.15 K can be obtainable [10,11,56,57]. As a result, the values of G1,3 and G2,3 in the co-solvent mixtures are obtained and presented in Tables S9–S12 of Supplementary material. It is observed that the G1,3 and G2,3 values are all negative in any cases, which demonstrates that the amoxicillin exhibit affinity for the co-solvents in the four mixtures. The iterated values of Vcor and dx1,3 are presented in the Tables S9–S12 for amoxicillin in DMF (1) + water (2), isopropanol (1) + water (2), NMP (1) + water (2) and EG (1) + water (2) mixtures, respectively. Additionally, the plots of dx1,3 versus co-solvent (DMF, EG, NMP and isopropanol) compositions are graphically shown in Fig. 3. It can be shown from the Figure (3) that the dependence of dx1,3 values on the co-solvent (1) proportion in the

8

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010

isopropanol, amoxicillin is acting as a Lewis acid with DMF, NMP, EG or isopropanol molecules, because the DMF, NMP, EG or isopropanol are more basic than water, as described by the Kamlet– Taft hydrogen bond acceptor parameters, i.e. b = 0.69 for DMF, b = 0.72 for NMP, b = 0.84 for isopropanol, b = 0.52 for EG and 0.47 for water [41,43]. In the water-rich compositions for the DMF (NMP, EG or isopropanol) (1) + water (2) mixtures, where amoxicillin is preferentially solvated by the water, it is conjecturable that amoxicillin act as a Lewis base because of the free electron pairs in oxygens atoms of = O and –OH and nitrogen atoms of –NH2 and > NH (Fig. 1), which interact with acidic hydrogen atoms of water, because the Kamlet–Taft hydrogen bond donor parameters are a = 1.17 for water, 0.00 for DMF and NMP, 0.90 for EG and 0.84 for isopropanol, respectively [42,43], being water more acidic than DMF (NMP, EG or isopropanol). 5. Conclusion Fig. 3. dx1,3 values of amoxicillin (3) in DMF (1) + water (2), NMP (1) + water (2), EG (1) + water (2) and isopropanol (1) + water (2) mixtures at 298.15 K. j, DMF (1) + water (2); d, NMP (1) + water (2); ▲, EG (1) + water (2); ., isopropanol (1) + water (2).

mixtures is non-linear. Addition of the co-solvent makes negative the values of dx1,3 of amoxicillin (3) from the neat water to mixture up to composition x1 = 0.20 mol fraction of DMF for the DMF (1) + water (2) mixture, x1 = 0.165 mol fraction of NMP for the NMP (1) + water (2) mixture, x1 = 0.25 mol fraction of EG and isopropanol for the EG/isopropanol (1) + water (2) mixtures. The smallest negative values are attained with the composition x1 = 0.05 with dx1,3 =  1.123
102 for the DMF (1) + water (2) mixture, dx1,3 =  1.315
102 for the NMP (1) + water (2) mixture, dx1,3 =  0.948
102 for the EG (1) + water (2) mixture and dx1,3 =  1.014
102 for the isopropanol (1) + water (2) mixture. Perhaps the structuring of water molecules around the nonpolar aromatic group of amoxicillin (i.e. hydrophobic hydration of benzene ring) contributes to lowering of the net dx1,3 values to negative in the DMF (NMP, EG or isopropanol) (1) + water (2) mixtures. In the DMF (1) + water (2) mixture with composition 0.20 < x1 < 1.00, NMP (1) + water (2) mixture with composition 0.165 < x1 < 1.00 and EG/isopropanol (1) + water (2) mixtures with composition 0.25 < x1 < 1.00, the local mole fractions of DMF, NMP, EG or isopropanol are greater than that of the solutions and hence the dx1,3 values are positive, which specifies that amoxicillin is preferentially solvated by the co-solvent of DMF (NMP, EG or isopropanol). The co-solvent action to improve the amoxicillin solubility may be caused by breaking of the ordered structure of water nearby the polar moieties of amoxicillin which increases the solvation of amoxicillin exhibiting maximum value in x1 = 0.45 with dx1,3 = 1.156
102 for DMF (1) + water (2), x1 = 0.45 with dx1,3 = 1.198
102 for NMP (1) + water (2), x1 = 0.60 with dx1,3 = 1.009
102 for EG (1) + water (2) and x1 = 0.60 with dx1,3 = 7.102
102 for isopropanol (1) + water (2). It can also be observed from Fig. 3 that in the regions with intermediate composition, the preferential solvation magnitude of amoxicillin by the co-solvent is highest in the isopropanol (1) + water (2) mixture among the four co-solvent mixtures. According to the analysis of structural and functional group, amoxicillin may act as a Lewis acid in mixtures due to the ability of acidic hydrogen atom in its –OH, –NH2, >NH and –COOH groups (Fig. 1) to form hydrogen bonds with proton-acceptor functional group of the DMF, NMP, EG or isopropanol (oxygen or nitrogen atoms in co-solvent). In terms of the preferential solvation results, it is conjecturable that in the region of 0.20 < x1 < 1.00 for DMF, 0.165 < x1 < 1.00 for NMP and 0.25 < x1 < 1.00 for EG and

The equilibrium solubility of amoxicillin in co-solvent mixtures of DMF (1) + water (2), NMP (1) + water (2), isopropanol (1) + water (2) + and EG (1) + water (2) was experimentally determined by the saturation shake-flask technique within the temperatures from 278.15 K to 328.15 K under ambient pressure (101.2 kPa). At the same mass fraction of DMF (NMP, isopropanol or EG) and temperature, the mole fraction solubility of amoxicillin was larger in (DMF + water) mixture than that in the other three mixtures. The values of preferential solvation parameters (dx1,3) were positive in the DMF mixture with composition 0.20 < x1 < 1, NMP mixture with composition 0.165 < x1 < 1 and EG/isopropanol mixtures with composition 0.25 < x1 < 1, indicating that amoxicillin was preferentially solvated by the co-solvent. The higher solvation by co-solvent might be elucidated in terms of the higher basic behavior of the co-solvent which interacted with the Lewis acidic groups of amoxicillin. In addition, the drug solubility was mathematically correlated through the Jouyban-Acree model obtaining average relative deviations lower than 4.55%. KAT-LSER analysis of solvent effect on solubility variation revealed that cavity term played the main role and thereby solvent–solvent interactions had dominant contribution to solvent effect in all binary mixtures. Acknowledgments The authors express sincere thanks to National Natural Science Foundation of China (Project number: 41877118), Natural Science Foundation of the Jiangsu Higher Education Institutions of China (Project number: 17KJB610013), Natural Science Foundation of Jiangsu Province of China (Project number: BK20181479) and Jiangsu Province Education Department Major Project (19KJA 140003) for their support. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jct.2019.106010. References [1] S.H. Yalkowsky, Solubility and Solubilization in Aqueous Media, American Chemical Society and Oxford University Press, New York, 1999, pp. 180–235. [2] R. Sanghvi, R. Narazaki, S.G. Machatha, S.H. Yalkowsky, Solubility improvement of drugs using N-methyl pyrolidone, Aaps Pharmscitech 9 (2008) 366–376. [3] A. Jouyban, Handbook of Solubility Data for Pharmaceuticals, CRC Press, BocaRaton, FL, 2010. [4] F. Martínez, A. Jouyban, W.E. Acree Jr, Pharmaceuticals solubility is still nowadays widely studied everywhere, Pharmaceut. Sci. 23 (2017) 1–2.

W. Li et al. / J. Chem. Thermodynamics 142 (2020) 106010 [5] J.T. Rubino, Cosolvents and Cosolvency, in: J. Swarbrick, J.C. Boylan (Eds.), Encyclopedia of Pharmaceutical Technology, 3, Marcel Dekker, New York, NY, 1988. [6] P. Kolárˇ, J.W. Shen, A. Tsuboi, T. Ishikawa, Solvent selection for pharmaceuticals, Fluid Phase Equilibr. 194–197 (2002) 771–782. [7] M.E. Aulton, Pharmaceutics. The Science of Dosage Forms Design, 2ed, Churchill Livingstone, London, 2002. [8] Y. Marcus, Solvent Mixtures: Properties and Selective Solvation, Marcel Dekker Inc, New York, NY, 2002. [9] Y. Marcus, Preferential solvation in mixed solvents, in: P.E. Smith, E. Matteoli, J.P. O’Connell (Eds.), Fluctuation Theory of Solutions: Applications in Chemistry, Chemical Engineering, and Biophysics, CRC, Press Taylor & Francis Group, BocaRaton, FL, 2013. [10] F. Martínez, A. Jouyban, W.E. Acree Jr, Preferential solvation of etoricoxib in some aqueous binary cosolvent mixtures at 298.15 K, Phys. Chem. Liq. 55 (2016) 291–303. [11] A. Jouyban, W.E. Acree Jr, F. Martínez, Modelling the solubility and preferential solvation of gallic acid in cosolvent + water mixtures, J. Mol. Liq. 224 (2016) 502–506. [12] V.D. Litskas, X.N. Karamanlis, S.P. Prousali, D.S. Koveos, Effects of the antibiotic amoxicillin on key species of the terrestrial environment, B. Environ. Contam. Tox. 100 (2018) 509–515. [13] H.J. Byung, W.Y. Byung, P. Jungsin, H.K. Jung, Y.L. Jun, S.S. Jae, S.P. Min, Pharmacokinetic drug interaction and safety after coadministration of clarithromycin, amoxicillin, and ilaprazole: a randomised, open-label, oneway crossover, two parallel sequences study, Eur. J. Clin. Pharmacol. 74 (2018) 1149–1157. [14] S. Sahara, M. Sugimoto, H. Ichikawa, T. Kagami, Y. Sakao, N. Ohashi, Y. Horio, K. Sugimoto, A. Kato, T. Furuta, H. Yasuda, Efficacy of reduced dosage of amoxicillin in an eradication therapy for helicobacter pylori infection in patients on hemodialysis: a randomized controlled trial, Digestion 97 (2018) 163–169. [15] N. Liu, H.J. Han, H.J. Yin, L.J. Han, G.Q. Huang, Variations in the fate and risk analysis of amoxicillin and its degradation products during pig manure aerobic composting, J. Hazard. Mater. 346 (2018) 234–241. [16] X.B. Meng, X.P. Zhang, Y.X. Wang, X.B. Cai, Y.T. Guo, X. Zhang, X.R. Xu, X.L. Xia, T. Liu, A.L. Liu, X.D. Li, Amoxicillin soluble powder for improving water solubility, CN Patent 107,789,321, Mar 13, 2018. [17] H.L. Xu, Preparation of high water solubility amoxicillin, CN Patent 109,867,687, June 11, 2019. [18] X.B. Zhang, Amoxicillin soluble powder with improved solubility and stability in water, and preparation method thereof, CN Patent 105,796,500, Jul 27, 2016. [19] M.Z. Hanna, N. Shan, M.L. Cheney, D.R. Weyna, R. Houck, Improvement of solubility and permeability of drugs through generating crystalline forms, WO Patent 2,011,097,269, Aug 11, 2011. [20] S. Sakuma, K. Kikukawa, R. Miyasaka, Method for improving the aqueous solubility of poorly-soluble substances, WO Patent 2,011,039,952, Apr 07, 2011. [21] I.M.B. Felix, L.C. Moreira, O. Chiavone-Filho, S. Mattedi, Solubility measurements of amoxicillin in mixtures of water and ethanol from 283.15 to 298.15 K, Fluid Phase Equilibr. 422 (2016) 78–86. [22] R.A. Dave, M.E. Morris, Novel high/low solubility classification methods for new molecular entities, Int. J. Pharm. 511 (2016) 111–126. [23] E.S.J. Rudolph, M. Zomerdijk, M. Ottens, L.A.M. van der Wielen, Solubilities and partition coefficients of semi-synthetic antibiotics in water + 1-butanol systems, Ind. Eng. Chem. Res. 40 (2001) 398–406. [24] I.M. Bezerra, O. Chiavone-Filho, S. Mattedi, Solid-liquid equilibrium data of amoxicillin and hydroxyphenylglycine in aqueous media, Braz. J. Chem. Eng. 30 (2013) 45–54. [25] C. Francesco, C. Daniela, D.S. Concetta, M. Demetrio, S. Silvio, V. Giuseppina, Modeling solubility, acid–base properties and activity coefficients of amoxicillin, ampicillin and (+)6-aminopenicillanic acid, in NaCl(aq) at different ionic strengths and temperatures, Eur. J. Pharm. Sci. 47 (2012) 661–677. [26] M.B. Diender, A.J.J. Straathof, L.A.M. van der Wielen, C. Ras, J.J. Heijnen, Feasibility of the thermodynamically controlled synthesis of amoxicillin, J. Mol. Catal. 5B (1998) 249–253. [27] M. Ahmadi Sabegh, H. Rajaei, A. Zeinolabedini Hezave, F. Esmaeilzadeh, Amoxicillin solubility and supercritical carbon dioxide, J. Chem. Eng. Data 57 (2012) 2750–2755. [28] G.D. Maia, M. Giulietti, Solubility of acetylsalicylic acid in ethanol, acetone, propylene glycol, and 2-propanol, J. Chem. Eng. Data 53 (2008) 256–258. [29] M. Mohammadzade, M. Barzegar-Jalali, A. Jouyban, Solubility of naproxen in 2propanol + water mixtures at various temperatures, J. Mol. Liq. 206 (2015) 110–113. [30] P.B. Rathi, V.K. Mourya, Solubility prediction of satranidazole in aqueous N, Ndimethylformamide mixtures using extended Hildebrand solubility approach, Indian J. Pharm. Sci. 74 (2012) 254–258. [31] R.J. Sindreu, M.L. Moya, B.F. Sanchez, A.G. Gonzalez, Solvent effects on the dissociation of aliphatic carboxylic acids in water-N, N-dimethylformamide mixtures: Correlation between acidity constants and solvatochromic parameters, J. Sol. Chem. 23 (1994) 1101–1109. [32] R.G. Strickley, Solubilizing excipients in oral and injectable formulations, Pharm. Res. 21 (2004) 201–230.

9

[33] F. Shakeel, M.F. Alajmi, N. Haq, N.A. Siddiqui, P. Alam, A.J. Al-Rehaily, Solubility and thermodynamic function of a bioactive compound bergenin in various pharmaceutically acceptable neat solvents at different temperatures, J. Chem. Thermodyn. 101 (2016) 19–24. [34] G.B. Yao, Q.C. Yao, Z.X. Xia, Z.H. Li, Solubility determination and correlation for o-phenylenediamine in (methanol, ethanol, acetonitrile and water) and their binary solvents from T = (283.15–318.15) K, J. Chem. Thermodyn. 105 (2017) 179–186. [35] J. Wang, A.L. Xu, R.J. Xu, Determination and correlation of terephthaldialdehyde solubility in (ethanol, isopropanol, ethyl acetate, isopentanol) + N, N-dimethylformamide mixed solvents at temperatures from 273.15 K to 318.15 K, J. Chem. Thermodyn. 105 (2017) 327–336. [36] R.W. Taft, J.L.M. Abboud, M.J. Kamlet, M.H. Abraham, Linear solvation energy relations, J. Sol. Chem. 14 (1985) 153–186. [37] M.J. Kamlet, R.M. Doherty, J.L. Abboud, M.H. Abraham, R.W. Taft, Solubility: a new look, Chemtech 16 (1986) 566–576. [38] M.J. Kamlet, R.M. Doherty, J.L.M. Abboud, M.H. Abraham, R.W. Taft, Linear solvation energy relationships: 36. Molecular properties governing solubilities of organic nonelectrolytes in water, J. Pharm. Sci. 75 (1986) 338–349. [39] C. Reichardt, Empirical parameters of solvent polarity as linear free-energy relationships, Angew Chem. Int. Ed. Engl. 18 (1979) 98–110. [40] M.J. Kamlet, J.L. Abboud, R.W. Taft, The solvatochromic comparison method. 6. The.pi.* scale of solvent polarities, J. Am. Chem. Soc. 99 (1977) 6027–6038. [41] M.J. Kamlet, R.W. Taft, The solvatochromic comparison method. I. The.beta.scale of solvent hydrogen-bond acceptor (HBA) basicities, J. Am. Chem. Soc. 98 (1976) 377–383. [42] R.W. Taft, M.J. Kamlet, The solvatochromic comparison method. 2. The.alpha.scale of solvent hydrogen-bond donor (HBD) acidities, J. Am. Chem. Soc. 98 (1976) 2886–2894. [43] Y. Marcus, The Properties of Solvents, John Wiley & Sons, Chichester, 1998. [44] A. Jouyban, M.A.A. Fakhree, Experimental, Computational Methods Pertaining to Drug Solubility, in: W.E. Acree (Ed.), Toxicity and Drug Testing, InTech. Rijeka, Croatia, 2012, pp. 187–218. [45] G.Q. Chen, J.H. Liang, J.C. Han, H.K. Zhao, Solubility modeling, solute-solvent interactions and thermodynamic dissolution properties of pnitrophenylacetonitrile in sixteen mono-solvents at temperatures ranging from 278.15 K to 333.15 K, J. Chem. Eng. Data 64 (2019) 315–323. [46] Y. Liu, L. Ding, M.X. Xie, S.Y. Liu, Q.F. Yang, J.H. Shan, Determination of amoxicillin residue in animal tissues by high performance liquid chromatography, Chin. J. Chromatogr. 21 (2003) 541–544 (Chinese). [47] H.L. Xu, Highly-water-soluble amoxicillin and preparation method thereof, CN Patent 109,867,687, June 11, 2019. [48] L. Shahhet, A.D. Alraghban, M.F.D. Chehna, Improvement of the physicochemical properties of amoxicillin trihydrate powder by recrystallization at different pH values, Int. J. Pharm. Pharm. Sci. 3 (2011) 92–100. [49] L.H. Lu, Improvement of process for preparation of amoxicillin granules, Shanghai Med. Pharma. J. 26 (2005) 470–472 (Chinese). [50] P. Cai, L.X. Zeng, Detection of the qualities of amoxicillin granules on market, J. Hainan Med. Coll. 13 (2007) 316–317 (Chinese). [51] Y. Migron, Y. Marcus, Polarity and hydrogen-bonding ability of some binary aqueous-organic mixtures, J. Chem. Soc. Faraday Trans. 87 (1991) 1339–1343. [52] R.J. Sindreu, M.L. Moya, F.S. Burgos, A.G. Gonzalez, Kamlet-Taft solvatochromic parameters of aqueous binary mixtures of tea-butyl alcohol and ethylene glycol, J. Sol. Chem. 25 (1996) 289–293. [53] B. Garcia, S. Aparicio, R. Alcalde, R. Ruiz, M.J. Davila, J.M. Leal, Characterization of Lactam-Containing Binary Solvents by Solvatochromic Indicators, J. Phys. Chem. 108B (2004) 3024–3029. [54] M. Roses, U. Buhvestov, C. Ràfols, F. Rived, E. Bosch, Solute–solvent and solvent–solvent interactions in binary solvent mixtures. Part 6. A quantitative measurement of the enhancement of the water structure in 2-methylpropan2-ol-water and propan-2-ol-water mixtures by solvatochromic indicators, J. Chem. Soc. Perkin Trans. 2 (1997) 1341–1348. [55] Y. Marcus, The properties of organic liquids that are relevant to their use as solvating solvents, Chem. Soc. Rev. 22 (1993) 409–416. [56] X.B. Li, Y.P. Liu, A. Farajtabar, G.X. Ding, G.Q. Chen, H.K. Zhao, Equilibrium solubility, solvent effect and preferential solvation of bioactive iminodibenzyl in aqueous cosolvent mixtures at various temperatures: Experimental data and thermodynamic analysis, J. Chem. Thermodyn. 132 (2019) 206–213. [57] X.B. Li, C. Cheng, Y. Cong, C.B. Du, H.K. Zhao, Preferential solvation of pioglitazone hydrochloride in some binary co-solvent mixtures according to the inverse Kirkwood-Buff integrals method, J. Chem. Thermodyn. 110 (2017) 218–226. [58] C.M. Hansen, Hansen Solubility Parameters: A User’s Handbook, CRC Press, 2007. [59] SciFinder: https://scifinder.cas.org/scifinder/view/scifinder/scifinderExplore. jsf, Oct 15, 2019. [60] K. Shinoda, Principles of Solution and Solubility, M, Dekker, New York, 1978. [61] I. Nugrahani, S. Asyarie, S.N. Soewandhi, S. Ibrahim, The antibiotic potency of amoxicillin-clavulanate co-crystal, Int. J. Pharmacol. 3 (2007) 475–481. [62] S.J. Franklin, U.S. Younis, P.B. Myrdal, Estimating the aqueous solubility of pharmaceutical hydrates, J. Pharm. Sci. 105 (2016) 1914–1919.

JCT 2019-821