Advanced Solid-State NMR Analysis of Two Crystal Forms of Ranitidine Hydrochloride: Detection of ¹H–¹⁴N Intra-/Intermolecular Correlations

Abstract

Understanding the characteristics of crystal polymorphism of active pharmaceutical ingredients and analyzing them with high sensitivity is important for quality of drug products, appropriate characterization strategies, and appropriate screening and selection processes. However, there are few methods to measure intra- and intermolecular correlations in crystals other than X-ray crystallography, for which it is sometimes difficult to obtain suitable single crystals. Recently, solid-state NMR has been recognized as a straightforward method for measuring molecular correlations. In this study, we selected ranitidine hydrochloride, which is known to exist in two forms, 1 and 2, as the model drug and investigated each form using solid-state NMR. In conducting the analysis, rotating the sample tube, which had a 1-mm inner diameter, increased the solid-state NMR resolution at 70 kHz. The ¹H–¹⁴N dipolar-based heteronuclear multiple quantum coherence (D-HMQC) analysis revealed the intermolecular correlation of Form 1 between the N atom of the nitro group and a proton of the furan moiety, which were closer than those of the intramolecular correlation reported using single X-ray crystal analysis. Thus, ¹H–¹⁴N D-HMQC analysis could be useful for characterizing intermolecular interaction in ranitidine hydrochloride crystals. In addition, we reassigned the ¹³C solid-state NMR signals of ranitidine hydrochloride according to the liquid-state and multiple solid-state NMR experiments.

Introduction

Crystal polymorphs of an active pharmaceutical ingredient (API) exhibit specific physicochemical properties, such as melting point, solubility, dissolution rate, and stability.^1,2) In addition, the crystal form of APIs impact the properties and quality of drug products. Therefore, it is important to develop advanced analytical methods for analyzing crystal polymorphs to properly characterize APIs. The techniques commonly used to determine the crystal forms of APIs include melting point; thermal; mid-IR, near-IR, terahertz, and Raman spectroscopy; powder X-ray diffraction (pXRD), optical microscopy, and solid-state NMR (SSNMR) analyses.^1,2)

Characterization techniques are categorized into those involving analyses at the molecular, particulate, and bulk levels.³⁾ Techniques for the molecular structural analysis of crystalline forms of APIs include ¹H- and ¹³C-SSNMR, and single-crystal XRD (scXRD), which can be used to elucidate crystal structures.^4–6) Although these techniques are useful, only scXRD can be used to analyze intra- and intermolecular correlations in crystals. However, preparing suitable single crystals for X-ray crystallography is challenging. Recently, the measurement of molecular correlations using SSNMR has been reported as straightforward method not requiring the preparation of a single crystal,^7,8) but few cases have been reported and it is not yet commonly used for measuring APIs.

Therefore, in this study, we analyzed ranitidine hydrochloride as a model drug known to exist as crystal polymorphs. The two forms (1 and 2) of ranitidine hydrochloride differ in melting point and water solubility, but are known to be therapeutically equivalent.⁹⁾ Crystal forms of ranitidine hydrochloride have been identified using discriminatory methods, such as diffuse reflectance IR Fourier transform spectroscopy (DRIFTS), Raman spectroscopy, optical microscopy, differential scanning calorimetry (DSC), scanning electron microscopy (SEM), scXRD, and pXRD.^2,10–12) Analyses of ranitidine hydrochloride and its crystal forms using SSNMR methods, including ¹³C and ¹⁴N one-dimensional (1D) and ¹H–¹⁵N 2D analyses, have also been reported.^13,14) However, the intermolecular correlations in the crystalline state have not been examined.

¹H–¹⁴N dipolar-based heteronuclear multiple quantum coherence (D-HMQC) can be used to detect intermolecular interactions.^7,15–26) However, there are few pharmaceutical applications for ¹H–¹⁴N D-HMQC, although studies have reported ¹H–¹⁴N D-HMQC measurements of biomolecules containing N atoms such as amino acids and nucleic acid derivatives.^7,26) Thus, we used ¹H–¹⁴N D-HMQC to investigate the intermolecular correlation of ranitidine hydrochloride crystal polymorphs, Form 1 and 2, and compared the results with the structure reported using scXRD measurements.

Because we performed the ¹H–¹⁴N D-HMQC with an increased resolution achieved by rotating the sample tube, which had an inner diameter of 1 mm, at 70 kHz magic angle spinning (MAS). Furthermore, a similarly fast rotation can also increase the resolution of ¹H- and ¹³C-NMR. Therefore, we also reassigned the ¹³C-NMR spectra of ranitidine hydrochloride including 2D measurements.

Results and Discussion

¹³C-SSNMR Signal Assignment for Ranitidine Hydrochloride Forms 1 and 2

The nitro-ethylenediamine moiety of ranitidine hydrochloride theoretically allows the formation of three tautomers of enamine, nitronic acid, and imine. The Form 1 polymorph of ranitidine hydrochloride had an enamine structure and Form 2 comprised of two tautomers, presumed to be enamine and nitronic acid according to the single X-ray crystal analysis.^14,27,28) Moreover, the results of the analysis of Forms 1 and 2 were supported by their SSNMR spectra, although that of Form 2 had broad signals all over.¹⁴⁾ Furthermore, the ¹³C assignment may not have been correctly performed previously, therefore, it was repeated in this study.

The chemical structure of ranitidine hydrochloride and its atomic number are shown in Fig. 1. No liquid-state NMR assignment has been reported for ranitidine hydrochloride; therefore, we determined the ¹H and ¹³C liquid-state NMR assignments, detailed in Supplementary Table S1 and Fig. S1. Subsequently, to assign the ¹³C-SSNMR signals of ranitidine hydrochloride, we performed 1D and 2D ¹³C cross polarization MAS (CP/MAS), ¹H–¹⁴N D-HMQC, dipolar dephasing, and ¹H–¹³C double CP/MAS analyses. The ¹³C CP/MAS spectra are shown in Fig. 2, the ¹H–¹⁴N D-HMQC spectra are shown in Figs. 3 and 4, the dipolar dephasing spectra are shown in Supplementary Figs. S2 and S3, and the ¹H–¹³C double CP/MAS spectra are shown in Supplementary Figs. S4 and S5. The data from the liquid-state NMR and multiple SSNMR experiment were used to assign ¹³C-SSNMR signals of both forms of ranitidine hydrochloride and the ¹³C chemical shifts (δ) are presented in Tables 1 and 2. Twelve ¹³C signals were observed in the liquid-state NMR (Supplementary Fig. S1), and 13 and 21 ¹³C signals were observed in SSNMR for Forms 1 and 2, respectively (Fig. 2). In the dipolar dephasing spectra, CH signals at positions 2, 3′, and 4′ and CH₂ at positions 6′, 7′, 8′, and 9′ showed larger decays relative to those of CH₃ at positions 4, 11′, and 12′ and quaternary C atoms at positions 1, 2′, and 5′ in Forms 1 and 2 (Supplementary Figs. S4, S5).

Fig. 1. Chemical Structure of Ranitidine Hydrochloride (C₁₃H₂₂N₄O₃S·HCl)

Fig. 2. ¹³C-SSNMR Spectra of Ranitidine Hydrochloride (A) Form 1 and (B) Form 2 Polymorphs

Fig. 3. ¹H–¹⁴N Dipolar-Based Heteronuclear Multiple Quantum Coherence (D-HMQC) Spectrum of Ranitidine Hydrochloride Form 1 Polymorph

Fig. 4. ¹H–¹⁴N D-HMQC Spectrum of Ranitidine Hydrochloride Form 2 Polymorph

Table 1. ¹³C Solid-State NMR (SSNMR) Signal Reassignment for Ranitidine Hydrochloride Form 1 Polymorph

Chemical shift (δ) (this study)a)	Reassignment (this study)	Chemical shift (δ) (previous report14))b)	Assignment (previous report14))
25.2	CH2 (6ʹ)c)	25.0	CH3 (4)
28.5	CH2 (7ʹ)c)	28.3	CH3 (11ʹ)
31.8	CH3 (4)c)	31.6	CH3 (12ʹ)
36.7	CH3 (11ʹ or 12ʹ)c)	38.5	CH2 (7ʹ)
44.1	CH3 (11ʹ or 12ʹ)c)	43.9	CH2 (8ʹ)
39.8	CH2 (8ʹ)c)	39.6	CH2 (6ʹ)
51.4	CH2 (9ʹ)	51.1	CH2 (9ʹ)
100.3	CH (2)	100.2	CH (2)
111.6	CH (3ʹ)	111.5	CH (3ʹ)
115.0	CH (4ʹ)	114.9	CH (4ʹ)
145.9	C (5ʹ)	145.8	C (5ʹ)
151.0	C (2ʹ)	150.8	C (2ʹ)
156.5	C (1)	156.5	C (1)

a) Recorded in 151 MHz (¹³C) with a spin speed of 70 kHz. b) Recorded in 100 MHz (¹³C) with a spin speed of 20 kHz reported previously.¹⁴⁾ c) The six carbon signals were correspondingly reassigned.

Table 2. ¹³C-SSNMR Signal Reassignment for Ranitidine Hydrochloride Form 2 Polymorph

Chemical shift (δ) (this study)a)	Reassignment (this study)	Chemical shift (δ) (previous report14))b)	Assignment (previous report14))
23.6, 24.6, 25.9, 31.8	CH2 (6ʹ and 7ʹ)c)	24.7, 25.7	CH3 (4)
		29.6	CH3 (11ʹ)
29.4, 30.5	CH3 (4)c)	30.7	CH3 (12ʹ)
40.2	CH3 (11ʹ or 12ʹ)c)	40.4	CH2 (7ʹ)
42.1, 42.4	CH3 (11ʹ or 12ʹ)c)	49.8	CH2 (8ʹ)
36.9, 38.4	CH2 (8ʹ)c)	42.4	CH2 (6ʹ)
50.0, 50.6	CH2 (9ʹ)	50.2	CH2 (9ʹ)
98.6, 100.6	CH (2)	98.9 100.9	CH (2)
109.1, 109.4	CH (3ʹ)	109.5	CH (3ʹ)
114.4	CH (4ʹ)	114.6	CH (4ʹ)
144.1	C (5ʹ)	144.3	C (5ʹ)
153.0	C (2ʹ)	153.2	C (2ʹ)
159.4	C (1)	159.3	C (1)

a) Recorded in 151 MHz (¹³C) with a spin speed of 70 kHz. b) Recorded in 100 MHz (¹³C) with a spin speed of 20 kHz reported previously.¹⁴⁾ c) The six carbon signals were correspondingly reassigned.

As shown in Table 1 and Fig. 2(A), chemical shifts of 13 ¹³C signals in Form 1 matched closely to those reported previously.¹⁴⁾ This result was consistent with those reported previously for the enamine structure in the Form 1 polymorph.^14,27,28) Next, the SSNMR data of Form 1 suggested that the CH₃ signals (δ_C: 36.7 and 44.1) were assigned as C11/C12 as per the observation of the C11/C12-H11/H12 correlation in ¹H–¹³C double CP/MAS spectrum (Supplementary Fig. S4) and N10-H11/H12 correlation in the ¹H–¹⁴N D-HMQC spectrum (Fig. 3); the other CH₃ signal (δ_C: 31.8) was assigned as C4. From the ¹H chemical shift in liquid-state NMR (Supplementary Table S1) and observation of the C6′-H6′ and C7′-H7′ correlations in the ¹H–¹³C double CP/MAS spectrum (Supplementary Fig. S4), the CH₂ signals (δ_C: 25.2 and 28.5) were assigned as C6′ and C7′, respectively, while the other CH₂ signal (δ_C: 39.8) was assigned as C8′. CH signals at positions 2, 3′, 4′ and quaternary C atoms at positions 1, 2′, and 5′ were assigned from the ¹³C chemical shift in the liquid-state NMR (Supplementary Table S1). Finally, these SSNMR data revealed that the six ¹³C assignments in Form 1 reported previously were incorrect; e.g., the assigned C4-methyl (δ_C: 25.0) signal was corrected to the C6ʹ-methylene, (δ_C: 25.2) signal in the CP/MAS spectrum¹⁴⁾ (Table 1). The reassignments of ranitidine hydrochloride Form 1 are shown in Table 1.

In a similar manner of Form 1 assignment, the 21 ¹³C-SSNMR signals were assigned in Form 2. As shown in Table 2 and Fig. 2(B), the ¹³C signals at positions 6ʹ, 7ʹ, 4, 11ʹ/12ʹ, 8ʹ, 9ʹ, 2, and 3ʹ were separated into two groups of signals as indicated by this measurement. However, the twelve ¹³C signals of the Form 2 CP/MAS spectrum reported previously were broad singlet signals.^14,28) Compared to previous findings, our results clearly supported the notion that Form 2 was composed of enamine and nitronic acid.^14,27,28) The reassignments of Form 2 revealed that, as in Form 1, the six ¹³C assignments reported previously were incorrect¹⁴⁾ (Table 2).

These results suggest that combining high-resolution SSNMR analysis with sample-spinning at 70 kHz inside a 600 MHz superconducting magnet and dipolar dephasing measurements clarified the difference between the spectra of ranitidine hydrochloride Forms 1 and 2. Furthermore, this strategy improved the accuracy of ¹³C-SSNMR assignments of both forms.

Intermolecular Correlations of Ranitidine Hydrochloride Form 1

Four N atoms were detected using the ¹H–¹⁴N D-HMQC 2D spectrum of the ranitidine hydrochloride Form 1 polymorph (Fig. 3). A strong correlation was observed with a proton derived from hydrogen chloride (HCl) and the N atom at position 10ʹ, which was protonated. Furthermore, the N atom at position 10ʹ interacted with protons of the alkyl group directly attached to the N atom such as the CH₂ at position 9ʹ and CH₃ at positions 11ʹ and 12ʹ (N10ʹ-H9ʹ/H11ʹ/H12ʹ). A few correlations were also observed between two N atoms and two H atoms each at positions 3 and 6 (N3-H3, N3-H6, N6-H3, and N6-H6). Several correlations between the CH₃ at position 4 and NH at position 3 or 6 were observed (N3/N6-H4). The N atom of the nitrogen dioxide (NO₂) group was correlated with the proton of the CH group at position 2 (N5-H2). Furthermore, a correlation was observed between the N atom of the NO₂ group and CH at position 3ʹ of the furan moiety (N5-H3ʹ), and the groups were considered too far apart to interact intramolecularly. Therefore, we compared the correlation with that reported for the single crystal X-ray structure. The X-ray structure of Form 1 was reported by Hempel et al.,^28,29) which was determined to be the enamine form with some disorder. Figure 5 shows the X-ray structure of Form 1 analyzed by Hempel et al.²⁹⁾ The distances between the N5 atom of the NO₂ group and the C3ʹ atom were calculated to be 6.10 and 3.71 Å (estimated distance between N5 and H3ʹ: 5.72 and 2.98 Å) for intramolecular and intermolecular associations, respectively. Thus, the ¹H–¹⁴N D-HMQC analysis showed intermolecular correlation because only the intermolecular distance (2.98 Å) was in the detectable range.³⁰⁾

Fig. 5. Ranitidine Hydrochloride Packing in the Reported Single Crystal X-Ray Structure²⁹⁾

Distance between N5 of nitro group and C3ʹ (H (white), C (gray), N (cyan), O (red), S (yellow), and Cl (green)). Solid orange and dashed black lines show inter- and intramolecular distances, respectively. Green circle shows moiety derived from hydrogen chloride (HCl).

Next, we recorded the ¹H–¹⁴N D-HMQC 2D spectrum of Form 2 as shown in Fig. 4. The N atom at position 10ʹ was protonated by HCl (N10ʹ-HCl) and showed a correlation with the protons of the CH₂ at position 9ʹ and the CH₃ at position 11ʹ,12ʹ (N10ʹ-H11ʹ/H12ʹ) in the Form 2 polymorph, similar to that of Form 1. The observed correlations between the N atom of NO₂ and a proton of CH at position 2 (N5-H2), and between the protons of CH₃ at position 4 and the N atom at positions 3 and 6 (H4-N3/N6) were also similar to those observed with the Form 1 polymorph. However, the correlations between N atoms at positions 3 and 6 and the protons at 9.6 and 0.5 ppm of the X and Y axes, respectively, detected in the Form 1 polymorph (N3-H3, N3-H6, N6-H3 and N6-H6) were not observed in the Form 2 (Figs. 3, 4). This observation suggests that the H bonding styles of the N atoms at positions 3 and 6 might differ between Forms 1 and 2. Furthermore, the results of the analysis of the Form 2 polymorph using scXRD showed that its packing differed from that of Form 1^14,27). The Form 2 polymorph consists of several possible structures. The distance between the N5 atom and the H3ʹ atom in the Form 2 polymorph may be wide, unlike that of the Form 1 polymorph.

An intermolecular correlation was detected in the Form 1 but not the Form 2 polymorph of ranitidine hydrochloride, possibly because of the difference in crystal packing. Therefore, combining 2D SSNMR, which analyzes the structure of compounds in powder form, with crystal structure prediction⁸⁾ could be expected to become an alternative method to scXRD for molecular detection of crystal polymorphs.

Conclusion

In this study, the high-resolution SSNMR analysis using very fast MAS at 70 kHz within the 1 mm MAS probe provided accurate ¹³C assignments of ranitidine hydrochloride Form 1 and 2 polymorphs. Moreover, ¹H–¹⁴N D-HMQC 2D SSNMR analysis provided additional information on the intermolecular correlation in the Form 1 polymorph based on its crystal forms. Further research should focus on analyzing long-range correlation of polymorphs through combining advanced SSNMR and in silico methods, such as crystal structure prediction, by calculating intermolecular distances of 2D SSNMR cross peaks. These methods are expected to be a viable alternative to X-ray crystallography that address the difficulty of forming crystals for single crystal X-rays analysis.

Experimental

Liquid-State NMR Spectroscopy

Ranitidine hydrochloride was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Dimethyl sulfoxide (DMSO)-d₆ (99.9% atom %D; Sigma-Aldrich, part of Merck, Darmstadt, Germany) was the deuterated solvent used. All liquid-state NMR analyses (¹H, ¹³C, ¹H–¹H correlated spectroscopy [COSY], ¹H–¹³C HMQC, ¹H–¹³C heteronuclear multiple bond correlation [HMBC]) were performed using a JNM-ECZ600R/S1 spectrometer equipped with a cryogenic probe (JEOL RESONANCE Inc., Tokyo, Japan) at 22 and 60 °C operating at 600 and 151 MHz for ¹H and ¹³C, respectively. Delta NMR software ver. 6.1 (JEOL RESONANCE Inc.) was used for NMR data analysis.

Preparation of Ranitidine Polymorphs

Ranitidine hydrochloride crystal polymorphs (Forms 1 and 2) were prepared using ethanol as a crystallization solvent as described previously.³¹⁾ The obtained crystal was analyzed using pXRD to confirm the polymorphic form.^{11,12,14,32,33)} The HPLC purity of each crystal form exceeded 99.6% at a UV wavelength of 235 nm.

SSNMR Spectroscopy

The crystals used in this study were prepared using ¹H single pulse and ¹³C CP/MAS 1D SSNMR analyses. Furthermore, ¹H–¹³C double CP/MAS and ¹H–¹⁴N D-HMQC 2D SSNMR analyses were performed using a JNM-ECZL600G spectrometer with a 1-mm HX MAS probe (JEOL RESONANCE Inc.) operating at 600, 151, and 43 MHz for ¹H, ¹³C, and ¹⁴N nuclei, respectively. Cylindrical 1 mm o.d. zirconia rotors were used and spun at 70 kHz. The ¹³C-NMR chemical shift scale was referenced with the ¹³C methyl signal of L-alanine at 19.6 ppm, which was used as the external standard. In the ¹H–¹⁴N D-HMQC analysis, an SR4 pulse sequence was used to recouple the dipolar interaction between ¹H and ¹⁴N, and tau-exc and tau-rec of 0.257 ms were used. Furthermore, 1D SSNMR analysis of the ¹³C CP/MAS and dipolar-dephasing method was performed using a JNM-ECZ400R/S1 spectrometer (¹³C Larmor frequency, 101 MHz; JEOL RESONANCE Inc.) with a 4-mm HX MAS probe (MAS frequency, 15 kHz). Delta NMR software ver. 6.1 (JEOL RESONANCE Inc.) was used for the NMR data analysis.

Acknowledgments

The authors would like to express their deepest appreciation to Dr. Koji Yazawa and Mr. Yutaro Ogaeri (JEOL RESONANCE Inc.) for their assistance in conducting the solid-state NMR analysis. This study was partially supported by the Japan Agency for Medical Research and Development (AMED) under Grant Number JP22ak0101190.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1) Chieng N., Rades T., Aaltonen J., J. Pharm. Biomed. Anal., 55, 618–644 (2011).
• 2) Vippagunta S. R., Brittain H. G., Grant D. J., Adv. Drug Deliv. Rev., 48, 3–26 (2001).
• 3) Brittain H. G., Bogdanowich S. J., Bugay D. E., DeVincentis J., Lewen G., Newman A. W., Pharm. Res., 8, 963–973 (1991).
• 4) Harris R. K., Analyst, 131, 351–373 (2006).
• 5) Harris R. K., J. Pharm. Pharmacol., 59, 225–239 (2007).
• 6) Geppi M., Mollica G., Borsacchi S., Veracini C. A., Appl. Spectrosc. Rev., 43, 202–302 (2008).
• 7) Rossi F., Cerreia Vioglio P., Bordignon S., Giorgio V., Nervi C., Priola E., Gobetto R., Yazawa K., Chierotti M. R., Cryst. Growth Des., 18, 2225–2233 (2018).
• 8) Bravetti F., Bordignon S., Alig E., Eisenbeil D., Fink L., Nervi C., Gobetto R., Schmidt M. U., Chierotti M. R., Chem. Eur. J., 28, e202103589 (2022).
• 9) Madan T., Kakkar A., Drug Dev. Ind. Pharm., 20, 1571–1588 (1994).
• 10) Stephenson G. A., Forbes R. A., Reutzel-Edens S. M., Adv. Drug Deliv. Rev., 48, 67–90 (2001).
• 11) Yamamoto E., Takeda Y., Ando D., Koide T., Amano Y., Miyazaki S., Miyazaki T., Izutsu K.-i., Kanazawa H., Goda Y., Int. J. Pharm., 605, 120834 (2021).
• 12) Chieng N., Zujovic Z., Bowmaker G., Rades T., Saville D., Int. J. Pharm., 327, 36–44 (2006).
• 13) Veinberg S. L., Johnston K. E., Jaroszewicz M. J., Kispal B. M., Mireault C. R., Kobayashi T., Pruski M., Schurko R. W., Phys. Chem. Chem. Phys., 18, 17713–17730 (2016).
• 14) Mirmehrabi M., Rohani S., Murthy K., Radatus B., J. Cryst. Growth, 260, 517–526 (2004).
• 15) Nishiyama Y., Endo Y., Nemoto T., Utsumi H., Yamauchi K., Hioka K., Asakura T., J. Magn. Reson., 208, 44–48 (2011).
• 16) Vitzthum V., Caporini M. A., Ulzega S., Bodenhausen G., J. Magn. Reson., 212, 234–239 (2011).
• 17) Vitzthum V., Caporini M. A., Ulzega S., Trébosc J., Lafon O., Amoureux J. P., Bodenhausen G., J. Magn. Reson., 223, 228–236 (2012).
• 18) Pell A. J., Sanders K. J., Wegner S., Pintacuda G., Grey C. P., J. Chem. Phys., 146, 194202 (2017).
• 19) Lu X., Trébosc J., Lafon O., Carnevale D., Ulzega S., Bodenhausen G., Amoureux J. P., J. Magn. Reson., 236, 105–116 (2013).
• 20) Li Y., Trébosc J., Hu B., Shen M., Amoureux J. P., Lafon O., J. Magn. Reson., 294, 101–114 (2018).
• 21) Aleksis R., Pell A. J., J. Chem. Phys., 153, 244202 (2020).
• 22) Rankin A. G. M., Trébosc J., Paluch P., Lafon O., Amoureux J. P., J. Magn. Reson., 303, 28–41 (2019).
• 23) Brinkmann A., Kentgens A. P., J. Am. Chem. Soc., 128, 14758–14759 (2006).
• 24) Lu X., Lafon O., Trébosc J., Tricot G., Delevoye L., Méar F., Montagne L., Amoureux J. P., J. Chem. Phys., 137, 144201 (2012).
• 25) Shen M., Trébosc J., Lafon O., Pourpoint F., Hu B., Chen Q., Amoureux J. P., J. Magn. Reson., 245, 38–49 (2014).
• 26) Duong N. T., Nishiyama Y., Phys. Chem. Chem. Phys., 24, 10717–10726 (2022).
• 27) Ishida T., In Y., Inoue M., Acta Crystallogr. C, 46, 1893–1896 (1990).
• 28) de Armas H. N., Peeters O. M., Blaton N., Van Gyseghem E., Martens J., Van Haele G., Van Den Mooter G., J. Pharm. Sci., 98, 146–158 (2009).
• 29) Hempel A., Camerman N., Mastropaolo D., Camerman A., Acta Crystallogr. C, 56, 1048–1049 (2000).
• 30) Tatton A. S., Bradley J. P., Iuga D., Brown S. P., Z. Phys. Chem., 226, 1187–1204 (2012).
• 31) Guerrieri P., Salameh A. K., Taylor L. S., Pharm. Res., 24, 147–156 (2007).
• 32) Mirmehrabi M., Rohani S., Murthy K. K., Radatus B., J. Pharm. Sci., 93, 1692–1700 (2004).
• 33) Chieng N., Rades T., Saville D., Eur. J. Pharm. Biopharm., 68, 771–780 (2008).