Solubility and Crystallization of Piroxicam from Different Solvents in Evaporative and Cooling Crystallization

Abstract

In this work, the solubility of a non-steroidal anti-inflammatory drug (NSAID), piroxicam, is investigated. The polymorphic form II, which is the most stable form at room temperature, was investigated in seven different solvents with various polarities. It has been found that the solubility of piroxicam in the solvents is in the following order: chloroform > dichloromethane > acetone > ethyl acetate > acetonitrile > acetic acid > methanol > hexane. Crystallization of piroxicam from different solvents has been performed with evaporative crystallization and cooling crystallization; the effects of solvent evaporation rate and solute concentration have also been studied. Both form I and form II could be produced in cooling and evaporative crystallization, and no simple link can be identified between the operating parameters and the polymorphic outcome. Results obtained in the present work showed the stochastic nature of the nucleation of different polymorphs as well as the complexity of the crystallization of a polymorphic system.

1. Introduction

Crystallization often serves as the final separation and purification step in the production of active pharmaceutical ingredients (APIs) [1]. Generally, cooling, solvent evaporation or the addition of an antisolvent are used to create supersaturation in the solution, which will induce nucleation and hence crystallization. A large percent of APIs can exist in different solid states as polymorphs, hydrates, solvates or cocrystals. Different solid forms of an API have different physical and chemical properties, which will influence the further processing and formulation of the API as well as product effectiveness, such as bioavailability [2,3]. Although it is common that the thermodynamically most stable form is selected for formulating the final dosage, metastable forms may be occasionally selected due to their higher solubility and dissolution rate in water. In the latter case, the polymorphic purity of the crystallized product is very important as a trace amount of the stable form can induce and facilitate the transformation of the metastable form to the stable form [4]. Consequently, it is of paramount importance to control the crystallization process so that the desired solid form is produced without any minor contamination of other forms [5].

The formation of a specific crystalline state of the API is influenced by the operating conditions during crystallization, such as solvent, temperature, supersaturation, crystallization method and so on. However, nucleation of different solid forms from a supersaturated solution represents a process with a very complex nature, and it is difficult to identify the underlying mechanism behind the formation of the different solid forms. A link between the solvent and polymorphic form of isonicotinamide (INA) being nucleated was reported by Kulkarni et al. [6]. In their work, the hydrogen bonding capabilities of given solvents were investigated and observed to have an influence on the polymorphic form yielded from the crystallization process. Nevertheless, other process parameters should be considered as these will also influence the polymorphic form. It was observed in our previous study that the formation of INA polymorphs also depended on the solute concentration as well as the temperature at which nucleation was onset [7,8]. The complex nature of the nucleation of polymorphs has been demonstrated in our previous work by the cooling crystallization of piroxicam from acetone–water solutions [9]. A solid form landscape has been established to show the dependence of the nucleated polymorph on the solute concentration. However, a further study revealed the specific challenge for upscaling the cooling crystallization from 100 mL to 2 liters. Batch cooling crystallization with the same operating conditions yielded different polymorphs of piroxicam in the small and large-scale systems [10].

Piroxicam is a non-steroidal anti-inflammatory drug (NSAID), forming four anhydrous polymorphs and one monohydrate [11,12,13,14]. In the present work, the solubility of piroxicam in different organic solvents has been investigated at temperatures ranging from 10 °C to 40 °C. The nucleation of the different solid forms of piroxicam was examined in evaporative and cooling crystallization. The effects of evaporation rate as well as the solute concentration and temperature on the polymorphic outcome were studied. It was observed that both form I (BIYSEH01, 03, 04, 10, 13, 14, 16) and form II (BIYSEH02, 08) of piroxicam could be yielded; the frequency of occurrence of the polymorphs did not show any clear dependence on the characteristics of the solvents and the other operating parameters. This is in agreement with recently published literature [15] in which the stochastic nucleation of INA polymorphs was demonstrated by both experimental and modeling approaches.

2. Experimental Method

2.1. Materials and Equipment

Piroxicam was purchased from Hyper Chemicals Limited, Zhejiang, China, and identified as the polymorphic form II. Analytical grade acetic acid (AcOH), chloroform (TCM), dichloromethane (DCM) and methanol (MEOH) were purchased from Sigma Aldrich (St. Louis, MO, USA). The other solvents, acetonitrile (CAN), ethyl acetate (EtOAc) and hexane, are HPLC standards purchased from VWR Chemicals (Radnor, PA, USA).

Cooling crystallization experiments were conducted using a Mettler Toledo EasyMax 102 Advanced Synthesis Workstation with two 100 mL glass reactors (Mettler Toledo, Columbus, OH, USA), an overhead stirrer and a solid-state thermostat cooling/heating jacket. The setup was controlled using N₂ as the purge gas and using iControl software. Evaporative crystallization experiments were conducted using a Buchi Rotavapor R-210 rotating evaporator (BÜCHI Labortechnik AG, Flawil, Switzerland) and using a temperature-controlled water basin.

Solvates were characterized using a simultaneous thermal analyzer (STA 449 F3 Jupiter®) from NETZSCH (Erich NETZSCH GmbH & Co. Holding KG, Selb, Germany). Samples were heated from room temperature up to 250 °C in a ceramic crucible at a heating rate of 5 °C/min. It has been confirmed in our previous work that the polymorphic forms of piroxicam have very distinguishable Raman spectra [9,10,16]. Raman spectroscopy was used in this work to identify the polymorphic forms obtained from the crystallization processes. A Bruker Senterra Dispersive Raman microscope (Bruker, Billerica, MA, USA) with a 785 nm laser operating at 100 mW with a 5 s integration time and two scans was used to collect the spectra.

2.2. Solubility Measurement

The gravimetric method was used to measure the solubility of piroxicam in seven different solvents: dichloromethane, chloroform, ethyl acetate, acetonitrile, acetic acid, methanol and hexane. The solubility of piroxicam form II was measured at temperatures 10, 20, 30 and 40 °C, except for dichloromethane, where the solubility was measured up to 30 °C as further temperature increase would exceed the boiling point. A suspension with an excess amount of solute (piroxicam) was prepared in 8 mL glass vials with 5 mL solvent. The vials were sealed and maintained under stirring for 24 h at a constant temperature to ensure that solid–liquid equilibrium was reached. The clear solution was sampled with a syringe filter and weighed. After the solvent was evaporated, the mass of the dried solid was measured to obtain the solubility of piroxicam. The experiments were reproduced in triplets for all solvents at the different temperatures and solubility profiles made.

2.3. Cooling Crystallization and Nucleation Kinetics of Piroxicam by Metastable Zone Width Measurement

The outcome of the solubility measurement showed the solubility of piroxicam increased with increasing temperature in five solvents: dichloromethane, ethyl acetate, acetonitrile, chloroform and acetic acid. These were then chosen as the solvents for performing cooling crystallization. The effects of the five different solvents on the nucleation of piroxicam have been investigated by measuring the Metastable Zone Width (MSZW) in the given solvents. Saturated solutions at 10 °C or 30 °C in the five solvents were prepared by dissolving an appropriate amount of piroxicam (form II) in approximately 90 g of solvent in the reactor. The solutions were heated to 40 °C and kept at this temperature for 30 min to ensure the complete dissolution of piroxicam. Then the solution was cooled linearly at 0.5 °C/min. The experiments were performed as duplicates. The MSZW was characterized by the visually observed sudden increase in turbidity of the solution. Nucleation was also verified from the temperature profile of the thermostat as the exothermic nucleation caused a sudden deviation of the reactor temperature from the cooling profile. After complete crystallization, the suspension was filtered with a 250 mL Büchner funnel, the crystals were dried off at room temperature and analyzed with Raman spectroscopy to identify the polymorphic form.

2.4. Fast Evaporative Crystallization

The evaporative crystallization of piroxicam was performed from six solvents: dichloromethane, chloroform, ethyl acetate, acetonitrile, acetic acid and methanol. Saturated solutions at both 20 °C and 30 °C in the six solvents were prepared by dissolving an appropriate amount of piroxicam (form II) in approximately 90 g of solvent. To ensure complete dissolution, the temperature was elevated during the solution preparation. When complete dissolution occurred, the solution was transferred to the rotating evaporator, and the pressure was set to the saturation vapor pressure of the corresponding solvent. The temperature of the solution was controlled by a water basin at 30 °C. The conditions were kept until all of the solvent had evaporated. These experiments were conducted as duplicates. The crystal form was investigated by Raman spectroscopy to determine the polymorphic form.

2.5. Slow Evaporative Crystallization

The effect of solvent evaporation rate was investigated by performing slow evaporative crystallization. Saturated solutions prepared from piroxicam (form II) at room temperature with the abovementioned six solvents were prepared by adding 5 mL of solvent and the appropriate amount of piroxicam to an 8 mL vial. The samples were filtered with 2 μm cellulose membranes to new vials to ensure the removal of any non-dissolved particles. The filtered samples were left uncapped, wrapped in aluminum foil to protect the samples from light. The crystal form was obtained after all solvents had evaporated and was analyzed with Raman spectroscopy.

3. Results

3.1. Solubility of Piroxicam

The solubility of piroxicam form II measured in the seven solvents, including dichloromethane, chloroform, ethyl acetate, acetonitrile, acetic acid, methanol and hexane are shown in

Table 1. Solubility of piroxicam form II in different solvents at varied temperatures (in mg/g solvent).

Solvents	10 °C	20 °C	30 °C	40 °C
Chloroform	67.73 ± 9.54	68.79 ± 0.67	78.40 ± 1.22	91.38 ± 2.38
Dichloromethane	61.00 ± 2.71	66.76 ± 0.76	73.80 ± 10.76
Ethyl acetate	12.42 ± 0.24	13.82 ± 0.48	15.55 ± 0.47	17.61 ± 0.67
Acetonitrile	10.73 ± 0.78	10.87 ± 0.34	13.30 ± 0.23	16.61 ± 0.91
Acetic acid	6.82	9.43 ± 0.32	11.97 ± 0.24	15.69 ± 0.04
Methanol	3.19 ± 0.08	4.44 ± 1.53	4.91 ± 0.05	5.08 ± 0.38
n-Hexane	0.68 ± 0.05	0.14 ± 0.25	0.16 ± 0.11	0.29 ± 0.01

and Figure 1a,b. From the figure, it is obvious that the highest solubility of piroxicam was found in dichloromethane and chloroform, while the lowest solubility was found in hexane and methanol. An investigation regarding the solubility of piroxicam in acetone was found from a previous study [9] and shown in Figure 1a. The solubility was measured with the same method but at temperatures 25, 35, and 45 °C. Moreover, it was observed that the solubility of piroxicam in all solvents, except methanol and hexane, increased with increasing temperature. In cooling crystallization, the yield depended on the slope of the solubility curve. It is desirable to perform cooling crystallization using solvents where the solubility increases significantly with increasing temperature. Consequently, the solvents dichloromethane, ethyl acetate, acetonitrile, chloroform and acetic acid were chosen as solvents when conducting the cooling crystallization of piroxicam. Chloroform and acetone showed the largest solubility increase with approximately 20 mg/g solvent from 10–40 °C and 25–45 °C, respectively.

The solubility of a solute in a solvent is determined by how well these materials interact. A way of evaluating these interactions is by the Hansen solubility parameters (HSPs). The HSP propose that the total force of the various interactions between the molecules can be divided into partial solubility parameters, i.e., dispersion forces (δ_d), polar bonding (δ_p) and hydrogen bonding (δ_h) [17]:

$${\mathsf{\delta }}_{\mathrm{T}}^2{=\mathsf{\delta }}_{\mathrm{d}}^2{+\mathsf{\delta }}_{\mathrm{P}}^2{+\mathsf{\delta }}_{\mathrm{h}}^2$$

(1)

A larger similarity between the HSPs of the solvent and solute implies a high degree of similarity of the molecular polarities and hence could imply a higher solubility. The HSPs of piroxicam and the solvents are shown in Figure 2a,b with a plot of δ_h versus δ_T, and δ_d versus δ_p, respectively. It can be seen from Figure 2 that chloroform and dichloromethane are the most similar to piroxicam in terms of the total solubility parameter and the three partial solubility parameters, which are also in agreement with these solvents giving the highest solubility. The solvents including ethyl acetate, acetone, acetonitrile and acetic acid possess a similar total solubility parameter as piroxicam; however, one or two of their partial solubility parameters are significantly different from that for piroxicam. Finally, hexane and methanol have very different HSPs from piroxicam, which was also in agreement with the lowest solubility of piroxicam in these two solvents.

3.2. Piroxicam Polymorphism and Characterization

Piroxicam is a polymorphic compound; it can form at least four anhydrous polymorphs and one monohydrate [11,12,13,14,19]. The focus in this work is on the polymorphic forms I and II. It has been reported that solution crystallization at moderate temperatures yielded either form I or form II, while the preparation of the unstable form III and IV requires more extreme conditions [11,14]. The relative stability of form I and form II has been investigated using suspension conversion and melting temperature [16]. It has been reported that form I and form II are enantiotropically related. Form II is the most thermodynamically stable polymorph at temperatures up to 60 °C, which was confirmed by the suspension conversion method. Form I should be more stable at elevated temperatures as form I has a higher melting temperature than form II [11,16]. The transition temperature of the two forms was found between 60 and 196 °C by [16]. A redetermination of this transition temperature would be preferred to obtain a better insight into the parameters that govern the selectivity between the two polymorphic forms. The difference between the two polymorphs arises from the different intermolecular hydrogen bonding. The orientation of the molecules of the two polymorphic forms are shown in Figure 3. Form I show a head-to-head configuration, while form II show a head-to-tail configuration. Cruz-Cabeza et al. [20] reported that being able to form different intermolecular hydrogen bonding does not lead to any significant higher propensity for the molecule to form polymorphism. However, it could be expected that for the polymorphs with different intermolecular hydrogen bonding, the influence of solvents used in the crystallization process may be more significant on the formation of polymorphic forms because the solvent can affect the self-association of the solute molecules in a supersaturated solution. It has been observed in our previous work [9,16] that the two polymorphs of piroxicam can be identified and characterized using Raman spectroscopy. As shown in Figure 4, several specific Raman shifts for forms I and II are marked with the arrows.

3.3. Cooling Crystallization and the Effect of Solvent on Nucleation Kinetics of Piroxicam

The effect of a solvent on the nucleation kinetics of piroxicam has been investigated via the measurement of MSZW, the results are shown in Figure 5. The MSZW of piroxicam in acetic acid is relatively narrow compared with the MSZWs in other solvents. The MSZW in solutions with dichloromethane, ethyl acetate and acetonitrile are between 25 °C and 35 °C. Furthermore, piroxicam showed a very wide MSZW in chloroform solution and could not be determined due to the cooling limit of the EasyMax system (approximately −30 °C). The piroxicam–chloroform solution saturated at 30 °C was cooled down to −30 °C and remained as a clear solution. The highly supersaturated solution was left overnight, and the crystallized piroxicam was analyzed with Raman spectroscopy. These observations imply a high energy barrier for the primary nucleation of piroxicam in most of the solvents studied in this work, which could be attributed to the formation of the hydrogen bonds between piroxicam and the solvent. Hydrogen bond formation between the solute and solvent has been observed to have an effect on both the crystallization process and the polymorphic form yielded in other studies [21,22,23]. The relatively wide MSZW of piroxicam in most of the studied solvents could suggest seeding as a feasible strategy for controlling the polymorphism as well as the particle size distribution of piroxicam from a cooling crystallization. The wide MSZW provides a large operating space for selecting optimal seeding parameters, such as seed with the desired polymorph, seed loading and seeding time, to direct the crystallization process towards the desired properties of the crystalline product.

After the measurement of MSZW, the nucleated piroxicam crystals were recovered by filtration and dried at room temperature. Subsequently, the polymorphism of the crystals was analyzed with Raman spectroscopy. It can be seen from Figure 5 that both form I and form II were yielded from the cooling crystallization. It has been discovered in our previous work [9] that the solute concentration has a significant influence on the polymorphism of piroxicam in cooling crystallization from acetone–water mixtures. It was observed in our previous study that low piroxicam concentrations would yield form I, while at higher concentrations (e.g., saturated at 30 °C and 40 °C), form II was obtained [9]. A similar effect of the solute concentration has been observed in the present work with cooling crystallization from ethyl acetate. However, the solute concentration showed no effect on the polymorphism of piroxicam in crystallization from dichloromethane, acetonitrile, chloroform and acetic acid. Form II was solely crystallized out from solutions with dichloromethane and acetonitrile, while form I was solely produced from solutions with acetic acid and chloroform, regardless of the very different solute concentrations in the solutions. Combing the MSZW and the polymorphic forms obtained in the cooling experiments (shown in

Table 2. Piroxicam polymorphic forms obtained from cooling and evaporative crystallizations.

	Linear Cooling Crystallization		Fast Evaporative Crystallization		Slow Evaporative Crystallization
Saturated Solution at Given Temperature (Solubility Rank)	10 °C	30 °C	20 °C	30 °C	20 °C
TCM (1)	-	Form I	-	Form II	Form II
DCM (2)	Form II	Form II	Form II	Form II	Form II
EtOAc (3)	Form I	Form II	Form II	Form II	Form II+I *
ACN (4)	Form II	Form II	Form II	Form II	Form I
AcOH (5)	Form I	Form I	Solvate	Solvate	-
MeOH (6)	-	-	Form II	Form II	Form II

* Form I was observed in one vial and form II in another vial. The dark grey background denotes piroxicam form II.

), it seems that there is no direct link between the nucleation kinetics and the polymorphic outcome. The MSZW in acetic acid is relatively narrow while the MSZW in chloroform is extremely wide (>30 °C, could not be detected), form I was produced from both solvents.

The crystallization of polymorphs and solvates from organic compounds is a complex process and represents a particular challenge in the production of APIs. It has been observed that the solute–solvent interaction has an effect on the formation of the self-associations of the solute molecules. Certain solvents can facilitate the formation of head-to-head dimers of the solute molecules; however, the exiting of these dimers does not necessarily promote the nucleation of the polymorphs with similar head-to-head molecular configurations [7]. The stochastic nature of the nucleation of different polymorphic forms in solution crystallization has been demonstrated in the crystallization of isonicotinamide and piroxicam, which has been reported in our previous work [7,9,10] as well as in studies from other groups [6,14].

3.4. Evaporative Crystallization with Fast and Slow Evaporation Rates

It has been observed in our previous work [9] that the crystallization method has a significant effect on the yielded polymorphism of piroxicam when cooling and anti-solvent crystallization was investigated in the solvent system of acetone and water. Piroxicam monohydrate crystallized dominantly from cooling crystallization of piroxicam from acetone–water solutions if the concentration of water was higher than 10 wt.%. However, the anhydrous form I of piroxicam was formed in anti-solvent crystallization from acetone solutions using water as the anti-solvent regardless of the water concentration exceeding 10 wt.%. In this study, the crystallization of piroxicam polymorphs in evaporative crystallization from different organic solvents was investigated. The effects of solvent and evaporating rates have been studied. The overview of polymorphic forms obtained in both cooling and evaporative crystallization of piroxicam from the various solvents is shown in Table 2. The dark grey defines piroxicam form II, and the parentheses represent the given solvent ranking according to the solubility measurement. The temperature listed in the table denotes the temperature at which the saturated solution is prepared.

It is shown in Table 2 (Raman spectra of the samples shown in Table 1 are included in Supplementary Materials Figure S1) that the polymorph formed from the different experiments changes with the crystallization technique (cooling or evaporative) and with the solvent utilized. The cooling crystallization of piroxicam from ethyl acetate and chloroform could produce either form I or form II; evaporative crystallization of piroxicam from all studied solvents (except acetic acid) yielded form II when fast solvent evaporation was applied. Slowing down the evaporation rate led to the formation of form I in solutions with acetonitrile and ethyl acetate. Interestingly, form I was obtained from the cooling crystallization of acetic acid solutions regardless of the different initial solute concentrations, and evaporative crystallization produced a solvate (Raman spectra shown in Figure 4). As shown in Figure 6, the solid form was analyzed by thermogravimetric analysis, confirming the acetic acid solvate with the ratio of 1:1. This observation is in agreement with the literature [11], where a mono-solvate of piroxicam with acetic acid was discovered. The solubility of the solvent is shown in parentheses in Table 2 where one corresponds to the solvent where piroxicam was most soluble, and six to the least soluble. No correlation can be drawn between the solubility of piroxicam in the solvents with the polymorphic form obtained from cooling and evaporative crystallization.

The polymorphic form overview in Table 2 clearly demonstrates the stochastic nature of the nucleation of different polymorphic forms during solution crystallization and the complex interplay of the operation parameters that affect the outcome of a crystallization process. It has been hypothesized that the solvent–solute interaction may have a more significant effect on determining the polymorphism of the crystallized piroxicam, which have different intermolecular hydrogen bonding and configurations of molecules. However, the results obtained in the present work do not verify this hypothesis, as both form I and form II can be crystallized out of several solvents (see Table 2).

4. Conclusions

The solubility of piroxicam form II was measured in seven different solvents of varying Hansen’s solubility parameters at different temperatures (10, 20, 30 and 40 °C). The highest solubility of piroxicam form II was found in dichloromethane and chloroform, while the lowest solubility was found in methanol and hexane. These observations of solubility are also in accordance with the HSPs, where the similarity of parameters between piroxicam and the solvent resulted in higher solubility and vice versa.

The energy barrier for the nucleation of piroxicam depends on the solvent. Solutions with dichloromethane, ethyl acetate and acetonitrile showed a metastable zone width (MSZW) of approximately 30 °C. A narrow MSZW was observed when using the solvent acetic acid, while a very broad MSZW was encountered for chloroform. Polymorphism of piroxicam from cooling and evaporative crystallization could be affected by the operation parameters. Comparing linear cooling crystallization with fast and slow evaporative crystallization indicated some tendencies; form II was favored from both fast and slow evaporative crystallization techniques. Regardless of the crystallization method used, form II was yielded from dichloromethane and methanol. Form II was also yielded by all studied solvents (except acetic acid yielding a solvate) when applying fast evaporative crystallization. This indicates that fast evaporative crystallization is a method that can be used for the production of Form II. However, the obtained results showed the complex and stochastic nature of the nucleation of different polymorphic forms in solution crystallization. No simple link can be suggested to correlate the polymorphism outcome with any operation parameter, such as solvent, solute concentration, or how fast the supersaturation was generated.