Solubility measurement, model evaluation and thermodynamic analysis of rivaroxaban polymorphs in organic solvents

https://doi.org/10.1016/j.jct.2016.09.036Get rights and content

Highlights

  • The solubility of form II of rivaroxaban in organic solvents was firstly reported.

  • Wilson, NRTL and UNIQUAC were evaluated to model the solubility.

  • The apparent standard enthalpy change of the dissolution of both forms was analyzed.

Abstract

This work measured the solubility of forms I and II of anticoagulant drug rivaroxaban in a series of organic solvents in which the solubility of form II was firstly reported as supplementary to the work of Zhang et al. [22]. A dynamic dissolution method was used to determine the solubility of both forms in acetonitrile, methyl acetate, acetone, 2-butanone, 3-pentanone and 2-hexanone from (273.15 to 318.15) K and in 1,4-dioxane from (288.15 to 333.15) K. The results showed that, the solubility of both forms generally increased with the temperature and decreased with an increase in the carbon number of ketones at the same temperature. The two forms were monotropically related and form II was the metastable form giving a higher solubility than form I. Meanwhile, such semiempirical thermodynamic models based on local composition concept as Wilson, nonrandom two-liquid (NRTL) and universal quasi chemical (UNIQUAC) equations with the correlation between model parameters and temperature were employed to model the studied solid-liquid equilibrium. It was found that UNIQUAC was the most appropriate model among the selected thermodynamic equations for the correlation of the solubility of rivaroxaban polymorphs in organic solvents. Finally, the apparent standard enthalpy change of dissolution was acquired from the experimental solubility data and the positive values of the apparent standard enthalpy change revealed that the dissolution of both forms of rivaroxaban in the selected solvents was an endothermic process.

Introduction

Crystallization refers to the formation of solid particles from a vapor, the solidification from a liquid melt or the formation of dispersed solids from a solution, which is extensively applied in food, pharmaceutical and chemical industries [1], [2]. The solubility of a compound can provide vital thermodynamic information required to design crystallization processes, engineer crystal size distribution and understand phase behaviors in multiple-phase systems [3], [4]. Meanwhile, as to a given compound particularly organic one, it usually has different polymorphs which normally have different solubilities. Therefore, it is necessary to acquire the solubility data of different polymorphs of a compound in various solvents over an abroad temperature range for the purpose of separation and purification through crystallization process [5], [6].

However, the experimental determination of solubility in pure or mixed solvents is usually time-consuming, expensive and sometimes impracticable when a large quantity of the objective compound is unavailable. Reliable solubility prediction using validated thermodynamic models is being received much attention currently [7]. To date, a number of thermodynamic models that can be used to model liquid-liquid phase equilibrium have been improved to model solid-liquid equilibrium, including semiempirical models founded by local composition concept such as Wilson equation [8], [9], UNIQUAC equation [10], [11], NRTL equation [12] and UNIQUAC functional group activity coefficients (UNIFAC) equation [13], though their modelling performance is system-dependent. Han et al. [14] employed van’t Hoff equation, modified Apelblat equation, λh equation, Wilson equation and NRTL equation to correlate the experimental solubility of ethyl 5-amino-4-cyano-3-(2-ethoxy-2-oxoethyl)-2-thiophene carboxylate in methanol, ethanol, 1-butanol, n-propanol, isopropanol, toluene, ethyl acetate, acetonitrile and acetone, and found that all the equations could agree well with the experimental data in which Wilson equation gave better correlation results than other models. Long et al. [15] compared the correlated solubility of benzoic acid in acetone, 2-propanol, acetic acid and cyclohexane by van’t Hoff equation, λh equation, Wilson equation, NRTL equation and UNIQUAC equation with the experimental data, and concluded that the three-parameter NRTL equation could give the best correlation results. Yu et al. [16] used Scatchard-Hildebrand, Wilson, NRTL and UNIQUAC equations to correlate the experimental solubility data of phosphoric acid, P, P′-1,4-phenylene P, P, P′, P′-tetraphenyl ester (PAPTE) in methanol, ethanol, acetone, ethyl acetate, acetonitrile, methyl acetate, toluene, ether, n-propanol and i-propanol, and declared that all the equations could give satisfying correlations in which UNIQUAC worked best.

Rivaroxaban (5-chloro-N-{[(5S)-2-oxo-3-[4-(3-oxomorpholin-4-yl)phenyl]-1,3-oxazolidin-5-yl] methyl}thiophene-2-carboxamide, as shown in Fig. 1) is a low molecular weight oral anticoagulant drug, which can directly inhibit the active form of serine protease Factor Xa (FXa) [17]. Today rivaroxaban has been used for the prevention and treatment of thromboembolic diseases, deep vein thrombosis (DVT) [18], pulmonary embolism (PE) and peripheral arterial occlusive diseases [19]. Rivaroxaban has been reported to have forms I, II, III in which form I (medicinal form) is regarded as the most stable form and form III seems a disappearing form [20], [21]. In our previous work, Zhang et al. [22] have reported the solubility of form I of rivaroxaban in the binary mixtures of ethyl acetate with N,N-dimethylformamide, N,N-dimethylacetamide and tetrahydrofuran at temperatures ranging from (278.15 to 318.15) K. Sun et al. [23] have measured the solubility of the same form in the binary solvent mixtures of 1-methyl-2-pyrrolidinone with water, methanol, ethanol and isopropanol from (273.15 to 323.15) K. But, to the best of our knowledge, there is not any report on the solubility of form II. For expanding the fields of usage and the purification of rivaroxaban polymorphs, this work measured the solubility of forms I and II of rivaroxaban in acetonitrile, methyl acetate, acetone, 2-butanone, 3-pentanone and 2-hexanone from (273.15 to 318.15) K and in 1,4-dioxane from (288.15 to 333.15) K. Meanwhile, the applicability of Wilson, NRTL and UNIQUAC in the modelling of studied solid-liquid equilibrium was evaluated using the experimental solubility data. Finally, the apparent standard enthalpy change of dissolution was calculated by means of the modified van’t Hoff analysis.

Section snippets

Materials

Rivaroxaban (mass fraction purity greater than 0.995) was provided by Miss Boyle Chemical Co., Ltd. (Shanghai, China). All solvents used in this work were purchased from Sinopharm Reagent Co., Ltd. (Shanghai, China) with analytical reagent grade and used as received without further purification. More detailed information about the materials used in this work was listed in Table 1.

Preparation and characterization of polymorphs

Form I was obtained as the raw material, whereas form II was prepared in-house as described in reference [20]. Both

Models for solid-liquid equilibrium

When a solid phase is in equilibrium with a liquid phase in a nonideal system, the relationship between the fugacities of solute in the two phases is given by [26]:lnfiLfiS=ΔHfusRTtpTtpT-1-ΔCpRTtpT-1+ΔCpRlnTtpTwhere the superscript i refers to the solute, fiL is the fugacity of the solute in the liquid phase, fiS is the fugacity of the solute in the solid phase, ΔHfus is the enthalpy of fusion of the solute, and R is the ideal gas constant. T and Ttp are the experiment temperature and the

Measured solubility and modelling results

Taking into account that the binary interaction parameters aij (i  j) are the function of temperature, the mij and nij of forms I and II of rivaroxaban in Eq. (19) are listed in Table 5, Table 6, respectively.

The experimental and calculated solubility data of the two forms of rivaroxaban in the studied solvents are listed in Table 7, Table 8, respectively. In general, the solubility of both forms increases with the increasing temperature. In the same solvent and at the same temperature the

Conclusions

The solubility data of monotropically-related two forms of rivaroxaban in seven solvents were determined by a dynamic dissolution method at atmospheric pressure. The solubility of both forms of rivaroxaban increased with the temperature and form II had a higher solubility than form I demonstrating that form II was the metastable form. With the correlation between model parameters and temperature, Wilson, NRTL and UNIQUAC equations could well model the studied solid-liquid equilibrium in an

Acknowledgments

This research work was financially supported by the National Natural Science Foundation of China (Nos. 21176102, 21176215 & 21476136), the Natural Science Foundation of Jiangsu Province (No. BK20131100), the Connotation Construction Project of SUES (No. Nhky-2015-05), Science and Technology Commission of Shanghai Municipality (No. 15430501200) and the Sino – German Center for Research Promotion (No. GZ935).

References (46)

  • L. Zhu et al.

    J. Chem. Thermodyn.

    (2015)
  • K.D. Bhesaniya et al.

    J. Chem. Thermodyn.

    (2014)
  • C. Christov

    J. Chem. Thermodyn.

    (2011)
  • C. Du et al.

    J. Chem. Thermodyn.

    (2016)
  • J. Ouyang et al.

    J. Chem. Thermodyn.

    (2015)
  • J. Fang et al.

    J. Chem. Thermodyn.

    (2015)
  • R. Kato et al.

    J. Chem. Thermodyn.

    (2005)
  • S. Han et al.

    J. Chem. Thermodyn.

    (2016)
  • B. Long et al.

    Fluid Phase Equilib.

    (2010)
  • G.M. Yu et al.

    Fluid Phase Equilib.

    (2015)
  • M. Zhang et al.

    J. Chem. Thermodyn.

    (2016)
  • F. Zou et al.

    J. Chem. Thermodyn.

    (2014)
  • A. Grunenberg et al.

    Int. J. Pharm.

    (1996)
  • M. Singh

    J. Chem. Thermodyn.

    (2007)
  • H.H. Tung et al.

    J. Pharm. Sci.

    (2008)
  • R.P. Danner et al.

    Fluid Phase Equilib.

    (1986)
  • A. Beerbower et al.

    J. Pharm. Sci.

    (1984)
  • M.D. Cheng et al.

    J. Chem. Thermodyn.

    (2010)
  • K. Wang et al.

    J. Chem. Thermodyn.

    (2012)
  • J. Lu et al.

    Curr. Med. Chem.

    (2009)
  • J. Chen et al.

    Cryst. Growth Des.

    (2011)
  • A. Nangia

    Acc. Chem. Res.

    (2008)
  • M.A. O’Mahony et al.

    Org. Process Res. Dev.

    (2013)
  • Cited by (27)

    View all citing articles on Scopus
    View full text