Supercooled liquid during the evaporation and cooling crystallization of valsartan: a macroscopic observation

Abstract

Macroscopic observation is an interesting topic in the field of non-classical crystallization. Because the glass transition temperature (T_g) of organic molecules is usually not high and can be further decreased by solvation, it is necessary to distinguish between supercooled liquids and amorphous solids during the crystallization process of organic molecules. The difference in the fluidity between these two phases makes it possible to distinguish them through a macroscopic observation. The transition process from a supercooled liquid precursor to the amorphous phase was observed without any additive during the evaporation crystallization process of valsartan in ethanol, as well as in the evaporation crystallization process of glucose in water, via visual inspection of fluidity and fluorescent probe method, and was also observed during the cooling crystallization process of valsartan in water via dissolution experiment and fluorescent probe method. More examples are needed to further refine the principle by which supercooled liquid precursors undergo a transformation to amorphous solids during the crystallization process of organic molecules.

---

Introduction

The classical crystallization process is a one-step procedure involving ions, atoms and molecules, while non-classical crystallization is a complex multi-step process that involves various intermediate particles.^1–3 According to Ostwald's step rule, the form appearing first is the closest to the free energy of the starting phase and progresses to the lower free energy state in a stepwise manner. It is hence expected that crystallization shall proceed in sequential steps, as guided by the free energy regime. The kinetics of these steps can vary based on the crystallization tendency of the compound. Only when the kinetics are slower than the experimental time frame, these events can be captured. Non-classical crystallization processes have been widely observed in the solution crystallization of biominerals and materials.^4,5 The intermediate particles in the non-classical crystallization process can have various phases including clusters, liquid phase, amorphous phase, and nanocrystals.^6,7 The most effective method for observing the non-classical crystallization process is electron microscopy. However, it was only recently that images with sufficient spatial resolution of soft materials sensitive to the beam were obtained.⁸ Compared with the crystallization process of inorganic substances, the crystallization process of organic molecules still lacks sufficient understanding.⁹ It is certain that in the crystallization process of organic molecules, non-classical crystallization processes are widespread, such as the crystallization process of the biominerals guanine^10,11 and uric acid.¹² The existence of molecular association, solvation, aromatic interactions, and hierarchy of intermolecular interactions during the crystallization of organic molecules makes the study of non-classical crystallization process complex and challenging.^13–15

In the non-classical crystallization process of organic molecules, one of the challenging issues is how to distinguish between supercooled liquids and amorphous solids. For amorphous solids, if the temperature is above the glass transition temperature (T_g), the solid phase will transform into a supercooled liquid phase.^16,17 The T_g of the intermediate particles during the non-classical crystallization process is difficult to measure, and most of the vary-temperature characterization technologies are not suitable for distinguishing between supercooled liquids and amorphous solids, especially cryo-technologies. The T_g of the pure amorphous solids of organic molecules is not high; it is usually lower than 100 °C or even lower than room temperature.¹⁸ If the amorphous solid contains a high content of solvent, its T_g is likely to be lower than room temperature. At this point, only a supercooled liquid exists.^19,20

The significant difference between supercooled liquids and amorphous solids is their fluidity, that is, supercooled liquid has better fluidity than the amorphous phase. Molecular mobility properties can be presented as mechanical deformation, dissolution behavior, and molecular diffusion rate. Mechanical deformation and dissolution behavior can be observed directly, and the molecular diffusion rate can be assessed using the fluorescent probe technique. Fluorescent probe is a classical method used for studying the recrystallization²¹ and phase separation^22,23 of amorphous phases. The fluidity of supercooled liquids and amorphous solids is different, which means that the diffusion rate of fluorescent molecules into them also varies. At low concentrations, the fluorescence intensity is directly proportional to the concentration of fluorescent molecules, while the fluorescent molecule concentration in the solution is directly proportional to its diffusion rate according to Fick's law. Therefore, we propose that under the same experimental conditions, the fluorescence intensity of the supercooled liquid containing diffusing fluorescent molecules is significantly stronger than that of the amorphous phase.

Valsartan (C₁₄H₂₉N₅O₃, Val, Scheme 1) is a small molecule that has been used for the treatment of hypertension. Val belongs to biopharmaceutics classification system (BCS) II drug with high permeability and poor aqueous solubility.^24,25 Understanding the crystallization process and mechanism of Val is helpful for developing new formulations to enhance its solubility in water and oral bioavailability. The glass-forming ability of Val is very high,²⁶ and the most common dosage form of Val is the amorphous phase,^27,28 which is suitable for the study of which phase (amorphous phase or supercooled liquid) appears first during the crystallization process without interference from recrystallization. The T_g of amorphous Val is about 76 °C, and the presence of moisture significantly decreases this value.^29,30

Scheme 1 Molecular structures of valsartan (Val), rhodamine B and curcumin.

In this study, a macroscopic observation confirmed the formation of a supercooled liquid precursor to an amorphous solid during evaporation crystallization of Val in ethanol and cooling crystallization in water, and this was further extended to the evaporation crystallization process of glucose (D-(+)-glucose, C₆H₁₂O₆, Glu) in water. Mechanical deformation, dissolution behavior and fluorescent probe with rhodamine B and curcumin (Scheme 1) were used for the identification of supercooled liquid. The novelty of this work is that we observed the differences between a Val supercooled liquid and an amorphous glass in the macroscopic scale. The crystallization phenomenon of the supercooled liquid precursor to the amorphous solid in the organic molecular system can promote the development of non-classical crystallization principles.

Experimental section

Materials

Valsartan (Val, C₁₄H₂₉N₅O₃, HPLC, 98%), rhodamine B (C₂₈H₃₁ClN₂O₃, AR), curcumin (C₂₁H₂₀O₆, 98%), and glucose (Glu, D-(+)-glucose, C₆H₁₂O₆, AR) were bought from Aladdin. Ethanol was bought from Shanghai Hushi. All the reagents were used without any treatment.

Methods

Valsartan evaporation crystallization. 40% Val ethanol solution was prepared by dissolving 1.00 g of raw Val (Val-raw) in 1.5 g of ethanol. 40% Val ethanol solution (1 g) was evaporated at 60 °C in the air on a macroscopic thermobalance (OHAUS MB120ZH, aluminium disc). The weight change of the sample over time was recorded. The product extracted at 1 hour quenched at room temperature is a typical supercooled liquid (Val-SL), and the product extracted at 8 hours quenched at room temperature is a typical amorphous glass (VaL-G). The ethanol content of Val-SL and Val-G obtained from different batches varies slightly due to the environmental difference.

Glucose evaporation crystallization. 10% Glu aqueous solution was prepared by dissolving 0.50 g of raw Glu (Glu-raw) in 4.5 g of water. 10% Glu aqueous solution (1 g) was evaporated in air at 140 °C on a macroscopic thermobalance (OHAUS MB120ZH, aluminium disc). The product extracted at 0.5 hour quenched at room temperature is a typical supercooled liquid (Glu-SL), and the product extracted at 1 hour quenched at room temperature is a typical amorphous glass (Glu-G).

Mechanical deformation. The product (Val-SL, Val-G, Glu-SL, and Glu-G) extracted from the macroscopic thermogravimetric analysis (Macro-TGA) placed on the aluminium disc was dipped using a tweezer and then photographed.

Fluorescent probes in evaporation crystallization. Raw rhodamine B or raw curcumin (∼1 mg) was physically mixed with 0.10 g of the evaporated product by grinding using an agate mortar for 3 minutes and then left to stand for 1 hour. The samples were characterized via fluorescence spectroscopy (Gangdong Technology, F-320A). The excitation wavelength (λ_ex) was selected as 530 nm with a scanning range of 545–700 nm for rhodamine B systems, and the λ_ex was 430 nm with a scanning range of 440–700 nm for curcumin systems.

Valsartan cooling crystallization. Val-raw (0.10 g) was dissolved in 100 mL of boiling water, and then the solution was kept in an ice-water bath for 1 hour, after which a Val suspension was obtained. The product (Val-wet) obtained by filtering using a common qualitative filter paper (pore size ∼20 μm) was further dried at 50 °C overnight, and the dried product of Val-wet was named as Val-dry.

Fluorescent probes in cooling crystallization. Approximately 0.1 mL of rhodamine B aqueous solution (1 mg mL⁻¹) or curcumin aqueous dispersion (1 mg of raw curcumin dispersed in 1 mL of water) was added to the Val-wet suspension, and then the mixture was stirred for 1 hour. Next, 0.1 mL of the rhodamine B solution or curcumin dispersion mentioned above and 0.10 g of Val-dry or Val-raw were added to 100 mL of water, and then the mixture was stirred for 1 hour. The samples were collected by filtering and then characterized via fluorescence spectroscopy as mentioned above.

Dissolution profiles. Val-raw, Val-wet, or Val-dry (0.1 g) was added to 100 mL of water, and then the dispersions were left to stand in a water bath at 37 °C. The supernatant (5 mL) was taken out at predetermined time points (0.5, 1, 2, 4, 8, 12, and 24 h) without fresh water added and then filtered with PES membrane (0.22 μm). The filtrate (2 mL) was mixed with 2 mL of ethanol, which was used for the ultraviolet-visible absorption spectrum (UV-vis, Shimadzu UV-2550, Shimadzu, Japan) test to determine the supernatants' drug contents. The dissolution experiments were performed in triplicate.

Conventional characterization. The samples were characterized via powder X-ray diffraction (PXRD, Philips X'Pert Pro, Cu Kα, 40 kV, 30 mA, 5–30°, 0.5° min⁻¹), Fourier transform infrared spectroscopy (FTIR, Shimadzu, IRAffinity-1S, Attenuated Total Reflectance [ATR] module, 400–4000 cm⁻¹, 2 cm⁻¹), confocal Raman spectroscopy (Thermo Fisher Scientific, DXR3xi, 532 nm, 40 mW, 0.002 s, 1000 scans, 50–3400 cm⁻¹), and scanning electron microscopy (SEM, Thermo Fisher Scientific, Apreo 2C, 2 kV). These characterizations were carried out according to the conventional operation methods without any additional treatments.

Results

Valsartan evaporation crystallization

The macroscopic thermogravimetric analysis (Macro-TGA) result (Fig. 1a) of 40% Val ethanol solution indicates that the evaporation rate of ethanol is relatively fast in the initial stage. Within ten minutes, the ethanol content dropped from the initial 60% to 21%. Subsequently, the evaporation rate of ethanol slowed down. At 1 hour, the ethanol content was 13%, and by 8 hours, it dropped to 8%. The product extracted at 1 hour quenched at room temperature from the Macro-TGA analysis process was supercooled liquid, as this product generates ripples under the action of mechanical force (the left optical photograph in Fig. 1a), which is termed as Val-SL in this study. The solubility of Val in ethanol at 298.15 K was 0.097681 in mole fraction,³¹ corresponding to a mass fraction concentration of 50.6%. Val-SL only contains 13% ethanol; thus, it should be a supercooled liquid. The product extracted at 8 hours quenched at room temperature was amorphous glass, termed as Val-G, as the product breaks under the action of mechanical force (the right optical photograph in Fig. 1a). The ethanol contents of the evaporation products of Val solution determine the existence form of the quenching products, while the ethanol contents of the evaporation products of Val solution correspond to specific extraction times. The reason for extracting the samples at 1 hour and 8 hours is that the evaporation products at these two times after quenching show an obvious property of supercooled liquid or amorphous glass. Both Val-SL and Val-G are transparent and cannot be visually distinguished from each other without external disturbances. The powder X-ray diffraction (PXRD) pattern of Val-raw showed a basically amorphous characteristic with some broad diffraction peaks at 2θ of 5.6° and 17.7° (Fig. 1b), consistent with the reported results.^27,28,32 The PXRD patterns of Val-SL and Val-G were almost the same without any diffraction peaks, consistent with their liquid or amorphous features; however, they cannot be distinguished from each other via PXRD. The low-frequency Raman spectra showed that both Val-SL and Val-G do not have lattice vibrations (Fig. S1), which cannot be used to distinguish them. Raman and IR spectra, which are related to molecular vibrations, cannot also distinguish between Val-SL and Val-G but can prove that the ethanol content of Val-SL is higher than that of Val-G (Fig. S2, Table S1).

Fig. 1 Supercooled liquid (Val-SL) and amorphous glass (VaL-G) observed in the evaporation crystallization process of valsartan (Val). (a) Macroscopic thermogravimetric analysis (Macro-TGA) curve of the 40% Val ethanol solution obtained at 60 °C in air; inset: the optical images of Val-SL and Val-G. (b) Powder X-ray diffraction (PXRD) patterns of raw Val (Val-raw), Val-SL and Val-G.

Because of the different mobilities of Val-SL and Val-G, it is possible to use the fluorescent probe method to distinguish between them based on the different diffusion rates of probes in them. Considering that the fluidity of Val-SL is better than that of Val-G, we speculated that the fluorescence intensity produced by physically mixing Val-SL with fluorescent substances will be higher than that of the corresponding mixture of Val-G. The experiment results are consistent with our speculation. For Val samples containing the same content of rhodamine B as a hydrophilic fluorescent probe,^33,34 Val-raw is slightly purple, Val-G is purple, and Val-SL is fluorescent red (Fig. 2a). The fluorescence spectra indicate that Val-SL shows strong fluorescence, while Val-raw and Val-G do not show fluorescence (Fig. 2b). For Val samples containing the same content of curcumin as the fluorescent probe,³⁵ Val-raw is slightly yellow, Val-G is yellow, and Val-SL is fluorescent orange (Fig. 2c). The fluorescence spectra indicate that Val-SL exhibits strong fluorescence, while Val-raw and Val-G have weak fluorescence (Fig. 2d). The fluorescent probe experiment results confirm that Val-SL is supercooled liquid and that Val-raw and Val-G are amorphous solids. By precisely controlling the solvent content of the evaporation products, Val-SL can be prepared from different Val solutions (Fig. S3).

Fig. 2 Fluorescent probe method for distinguishing between Val-SL and Val-G by physically mixing fluorescent substances with Val samples. Optical images and fluorescence spectra of Val samples with rhodamine B (a and b) and curcumin (c and d) as fluorescent probes.

Valsartan cooling crystallization

Val can be recrystallized by cooling its hot aqueous solution. Val suspension was prepared (Fig. 3) as mentioned in the experimental part. The PXRD pattern of Val-wet shows only two budges. Val-dry shows a similar PXRD pattern. The PXRD patterns of Val-wet and Val-dry are consistent with its liquid or amorphous properties; however, this method cannot distinguish between these two states. The SEM images of Val-wet and Val-dry indicate that it is also difficult to distinguish between them based on morphology (Fig. S4).

Fig. 3 Optical image of the Val suspension, PXRD patterns and optical images of the filtered product of the Val suspension (Val-wet), and the product obtained by drying Val-wet at 50 °C overnight (Val-dry). Val suspension was obtained by dissolving 0.10 g of Val-raw in 100 mL of boiling water and then cooling the solution in an ice-water bath for 1 hour.

The dissolution experiment results (Fig. 4) indicate that Val-raw and Val-dry have similar dissolution profiles and that the dissolution concentration of Val-wet is significantly lower than those of Val-raw and Val-dry. The low solubility of Val-wet indicates that Val-wet is a gel or a supercooled liquid during the dissolution process, which is commonly observed in the dissolution process of amorphous phases or amorphous solid dispersions.^23,36 Considering that Val-wet is the precipitate obtained from the Val suspension and is easy to be redispersed, the possibility of it being a gel can be eliminated. Thus, we proposed that Val-wet is a supercooled liquid and Val-dry is an amorphous glass. Furthermore, the lower dissolution rate of Val-wet is caused by the reduction of surface energy and the decrease in surface area.

Fig. 4 Dissolution profiles of Val-raw, Val-wet and Val-dry at 37 °C (n = 3).

Physical mixing is not suitable for the preparation of samples to distinguish between Val-wet and Val-dry using the fluorescent probe method (Fig. S5) because of the slow mobilities of Val-wet and Val-dry. The slurry of fluorescent probes under stirring was used to accelerate the entry of the fluorescent probes into the Val cooling crystallization samples. After being slurried with rhodamine B probe, Val-raw is purple, Val-dry is purple, and Val-wet is fluorescent pink in appearance (Fig. 5a). Furthermore, Val-wet showed a stronger fluorescence than Val-raw and Val-dry (Fig. 5b). Val-raw and Val-dry exhibited obvious fluorescence because rhodamine B can easily dissolve in water, which is beneficial for the molecular diffusion. For curcumin probe, Val-raw is slight yellow, Val-dry is yellow, and Val-wet is fluorescent yellow (Fig. 5c). Val-wet showed strong fluorescence, while Val-raw and Val-dry showed weak fluorescence (Fig. 5d). The results obtained by using the slurry method to introduce probes (namely rhodamine B and curcumin) confirm that Val-wet is a supercooled liquid, while Val-dry is an amorphous solid.

Fig. 5 Fluorescent probe method for distinguishing between Val-wet and Val-dry. Optical images and fluorescence spectra of Val samples prepared by slurrying with (a and b) rhodamine B and (c and d) curcumin as the fluorescent probes.

Phenomenon robustness

The formation of a supercooled liquid precursor to an amorphous solid can also be easily observed during the evaporation crystallization process of glucose (Glu) aqueous solution (Fig. 6). Glu was selected in this study because its amorphous phase can be easily prepared and it has good stability. The supercooled liquid of Glu (Glu-SL) can be obtained by quenching the 10% Glu aqueous solution after being heated at 140 °C in the air for 0.5 h at room temperature, with the solution containing 5.2% water and may form ripples under mechanical force (Fig. 6a, left optical photograph). The amorphous solid of Glu (Glu-G) can be obtained by quenching the 10% Glu aqueous solution at room temperature after the heat treatment at 140 °C in the air for 1 h, which contains 0.8% water content and may fracture under mechanical force (Fig. 6a, right optical photograph). Raw Glu (Glu-raw) was in the form of crystal with sharp diffraction peaks, and Glu-G only showed a board budge in the PXRD pattern (Fig. 6b). The PXRD pattern of the freshly prepared Glu-SL did not show sharp diffraction peaks, and that of the stored Glu-SL showed obvious diffraction peaks consistent with that of Glu-raw, indicating that Glu-SL is easy to recrystallize. Glu-G remained stable during storage. The fluorescent probe experiment result (Fig. 6c and d) using samples prepared by physical mixing with rhodamine B indicated that fresh Glu-SL showed much stronger fluorescence than Glu-raw and Glu-G. The fluorescence of the recrystallized Glu-SL was much lower than that of fresh sample, even lower than that of Glu-G. The fluidity and fluorescent probe results confirmed that Glu-SL was supercooled liquid and Glu-G was amorphous glass. Glu can crystallize from its solution upon cooling; however, the cooling crystallization system of Glu is not suitable for this study.

Fig. 6 Supercooled liquid (Glu-SL) and amorphous glass (Glu-G) observed in the evaporation crystallization process of glucose (Glu) aqueous solution. (a) Macro-TGA curve of 10% Glu aqueous solution obtained at 140 °C in the air; inset: the optical images of Glu-SL and Glu-G. (b) PXRD patterns of raw Glu (Glu-raw), fresh Glu-SL, recrystallized Glu-SL (stored for 1 hour) and Glu-G. (c) Fluorescence spectra of Glu-raw, fresh Glu-SL, recrystallized Glu-SL and Glu-G with rhodamine B as the fluorescent probe prepared via physical mixing. (d) Optical images of Glu-raw, Glu-G, fresh Glu-SL, and recrystallized Glu-SL physically mixed with rhodamine B.

Discussion

Fluorescent probe optimization

For the fluorescence probe method, only a limited number of molecule systems are applicable as the supercooled liquids and amorphous glasses of many molecules have a strong tendency to crystallize. Certain experimental conditions of the fluorescent probe method require special attention to avoid obtaining unreliable experimental results.

(1) Avoid the recrystallization of supercooled liquid or amorphous glass. Recrystallization can significantly reduce the fluorescence intensity.

(2) Avoid an extremely high concentration of fluorescent probes. A high concentration of fluorescent probes may lead to aggregation-caused quenching (ACQ). Nevertheless, the ACQ phenomenon has little impact on our experiment. Even if the ACQ phenomenon occurs, the difference in fluorescence can still be observed. Instead, the critical issue is the insufficient concentration of fluorescent molecules.

(3) Avoid the temperature from exceeding T_g. If the temperature is above T_g, amorphous glass will transform into a supercooled liquid.

(4) It is still necessary to compare samples with the reference sample of pure amorphous glass.

How to distinguish samples near T_g is still a problem. Currently, only samples with significant fluorescence differences can be distinguished. Quantitative analysis of the diffusion rate is a feasible means, and the fluorescent probe may be used to measure the diffusion rate.

Principle of forming amorphous glass from the supercooled liquid precursor

The presence of intermediate particles is a key characteristic of the non-classical crystallization process. The amorphous phase is one of the common intermediate particles. Since the amorphous phase and the supercooled liquid have very similar properties except for their fluidity, the conventional characterization methods are usually unable to distinguish them in solution systems. The T_g of small molecules is not high and can be reduced to below room temperature by mixing small molecules with a small amount of solvent, and a supercooled liquid can serve as a precursor to the amorphous phase. We proposed that this process is easily understandable (Scheme 2). During the crystallization process of organic small molecule solutions, the solvated molecules in the system are more likely to form a supercooled liquid first, because the process of desolvation is a continuous one. In this study, the cooling rate was controlled as quickly as possible, as a faster cooling rate is conducive to the formation of supercooled liquids and amorphous glasses. It is easier to observe a supercooled liquid in the evaporative crystallization process than in the cooling crystallization process. More characterization methods like fluorescent probes need to be developed, which remains challenging. More examples are needed to refine the principle of forming amorphous glass from the supercooled liquid precursor during the small-molecule crystallization process.

Scheme 2 Transition process from a supercooled liquid precursor to the amorphous phase.

Amorphous glass, supercooled liquid, and gel

Several methods were proposed in this study to distinguish between amorphous glass and supercooled liquid. For transparent bulk samples, an amorphous glass and a supercooled liquid can be distinguished via mechanical force and fluorescence probes. For samples composed of particles, an amorphous glass and a supercooled liquid can be distinguished via dissolution profiles and fluorescence probes. The dissolution profiles are not always working if the amorphous particles form the gel. The fluorescence probes may be a possible means to distinguish between the gel and the supercooled liquid.

Conclusions

The phenomenon that supercooled liquid serves as precursor to the amorphous solid was observed in valsartan and glucose evaporation crystallization process, confirmed by visual fluidity and fluorescent probe method, and in valsartan cooling crystallization process, confirmed by the dissolution profile and fluorescent probe method. The transition process from a supercooled liquid precursor to the amorphous phase may be one of the organic non-classical crystallization principles.

Author contributions

Rongrong Xue: conceptualization, methodology, project administration, writing – original draft, funding acquisition. Tuanjia Li: investigation, validation. Yingying He: investigation. Fenghua Chen: conceptualization, methodology, validation, data curation, supervision, writing – review & editing, funding acquisition. Wangchuan Xiao: resources, supervision, funding acquisition.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI).

Supplementary information: extended characterization details. See DOI: https://doi.org/10.1039/d5ce01019e.

Acknowledgements

This work was supported by the National Natural Science Foundation of China [Grant No. 22005175], the Central Government Guided Local Special Foundation [Grant No. 2022L3045], and the Natural Science Foundation of Fujian Province [Grant No. 2020J01374].

References

1. J. J. De Yoreo, P. U. P. A. Gilbert, N. A. J. M. Sommerdijk, R. L. Penn, S. Whitelam, D. Joester, H. Zhang, J. D. Rimer, A. Navrotsky, J. F. Banfield, A. F. Wallace, F. M. Michel, F. C. Meldrum, H. Cölfen and P. M. Dove, Science, 2015, 349, aaa6760.
2. M. Jehannin, A. Rao and H. Cölfen, J. Am. Chem. Soc., 2019, 141, 10120–10136.
3. Z. Ma, B. Li and R. Tang, Chin. J. Chem., 2021, 39, 2071–2082.
4. B. Jin, Z. Liu and R. Tang, CrystEngComm, 2020, 22, 4057–4073.
5. L.-A. DiCecco, T. Tang, E. D. Sone and K. Grandfield, Small, 2025, 21, 2407539.
6. D. Gebauer, J. D. Gale and H. Cölfen, Small, 2022, 18, 2107735.
7. C. Jia, A. Xiao, J. Zhao, P. Wang, X. Fang, H. Zhang and B. Guan, Cryst. Growth Des., 2024, 24, 601–612.
8. J. Du, Y. Bae and J. Yoreo, Nat. Rev. Mater., 2024, 9, 229–248.
9. J.-I. Yoo, J. Jang, J. Shim, V. Panov, P. Sundararajan, S. Kim and J.-K. Song, Colloids Surf., A, 2024, 697, 134390.
10. A. Wagner, V. Ezersky, R. Maria, A. Upcher, T. Lemcoff, E. D. Aflalo, Y. Lubin and B. A. Palmer, Adv. Mater., 2022, 34, 2202242.
11. D. Guo, Y. Liu, X. Hou, X. Wang, C. Fan, L. Bao, X. He, H. Zhang and Y. Ma, CrystEngComm, 2023, 25, 4521–4530.
12. C. Rodriguez-Navarro, K. Elert, A. Ibañez-Velasco, L. Monasterio-Guillot, M. Andres, F. Sivera, E. Pascual and E. Ruiz-Agudo, Commun. Biol., 2024, 7, 828.
13. R. J. Davey, S. L. M. Schroeder and J. H. ter Horst, Angew. Chem., Int. Ed., 2013, 52, 2166–2179.
14. X. Li, J. Wang, T. Wang, N. Wang, S. Zong, X. Huang and H. Hao, Sci. China: Chem., 2021, 64, 1460–1481.
15. J. Zhou, Y. Zhou and W. Tang, Crystals, 2022, 12, 980.
16. E. D. Zanotto and J. C. Mauro, J. Non-Cryst. Solids, 2017, 471, 490–495.
17. R. S. Smith, W. A. Thornley, G. A. Kimmel and B. D. Kay, J. Phys. Chem. Lett., 2025, 16, 3198–3205.
18. T. Li, W. Xiao, S. Ren, R. Xue and F. Chen, Crystals, 2024, 14, 1037.
19. J. Han, L. Li, Z. Pang, M. Su, X. He, S. Qian, J. Zhang, Y. Gao and Y. Wei, Eur. J. Pharm. Sci., 2022, 170, 106083.
20. M. S. Bordawekar, M. Pudipeddi, C. E. Ruegger and S. S. Dhareshwar, AAPS PharmSciTech, 2023, 24, 149.
21. C. Li, G. Hong, H. Yu and L. Qi, Chem. Mater., 2010, 22, 3206–3211.
22. A. Mansuri, P. Münzner, A. Heermant, S. Hänsch, T. Feuerbach, B. Fischer, J. Winck, A. W. P. Vermeer, W. Hoheisel, R. Böhmer, C. Gainaru and M. Thommes, Mol. Pharmaceutics, 2023, 20, 2080–2093.
23. A. Deac, Q. Qi, A. S. Indulkar, Y. Gao, G. G. Z. Zhang and L. S. Taylor, Mol. Pharmaceutics, 2023, 20, 2217–2234.
24. W. Xu, Y. Sun, L. Du, Y. S. Chistyachenko, A. V. Dushkin and W. Su, J. Drug Delivery Sci. Technol., 2018, 44, 399–405.
25. Y. Liu, J. Wang, X. Wang, P. Liu and F. Pang, J. Chem. Eng. Data, 2009, 54, 986–988.
26. J. J. Moura Ramos and H. P. Diogo, RSC Adv., 2017, 7, 10831–10840.
27. Y. Guinet, L. Paccou, F. Danède, P. Derollez and A. Hédoux, Int. J. Pharm., 2017, 526, 209–216.
28. Y. Wang, H. Zhang, X. Qiu, Z. Gao, S. Du and F. Xue, Sep. Purif. Technol., 2025, 365, 132635.
29. M. Skotnicki, A. Gaweł, P. Cebe and M. Pyda, Drug Dev. Ind. Pharm., 2012, 39, 1508–1514.
30. M. Skotnicki, D. C. Apperley, J. A. Aguilar, B. Milanowski, M. Pyda and P. Hodgkinson, Mol. Pharmaceutics, 2015, 13, 211–222.
31. H. Zhang, Y. Zhang, F. Xue, H. Chen, Z. Gao, S. Du and Y. Wang, J. Mol. Liq., 2024, 402, 124774.
32. J.-R. Wang, X. Wang, L. Lu and X. Mei, Cryst. Growth Des., 2013, 13, 3261–3269.
33. X. Leng, W. Xu, C. Qiao, X. Jia, Y. Long and B. Yang, RSC Adv., 2019, 9, 6027–6034.
34. C.-M. Wu, Y.-H. Chen, K. Dayananda, T.-W. Shiue, C.-H. Hung, W.-F. Liaw, P.-Y. Chen and Y.-M. Wang, Anal. Chim. Acta, 2011, 708, 141–148.
35. K. I. Priyadarsini, J. Photochem. Photobiol., C, 2009, 10, 81–95.
36. X. Ma, H. Su, Y. Liu, F. Chen and R. Xue, RSC Adv., 2025, 15, 9348–9358.

---