A Novel System for Crystal Polymorph Discovery via Selective-Wavelength Infrared Irradiation Using Metamaterials

Abstract

The control of crystal polymorphs is central to the design of pharmaceuticals and functional materials. Conventionally, crystal polymorph production has been controlled primarily by adjusting chemical and thermodynamic parameters. In this study, we developed a device capable of emitting infrared radiation at selected wavelengths using a novel material having a “MIM structure” which is a type of metamaterial. With this device, we propose a new approach to crystal polymorph control through the irradiation of narrow-band infrared radiation that coincides with the infrared absorption band of specific functional groups. In this paper, we present the design and operating principle of a new crystallization system, and as an application example, we report the experimental results of controlling the crystal polymorphs of Ritonavir, an active pharmaceutical ingredient.

1. Introduction

The control of crystal polymorphs is central to the design of pharmaceuticals [1,2,3,4,5,6,7,8,9] and functional materials [10,11,12,13,14]. It is of particular importance in the pharmaceutical field in that, as generally known, differences among crystal polymorphs of the same compound can significantly affect the solubility, stability, absorption, and bioavailability of the compound. Polymorph control is fundamental to the stability of drug efficacy and to formulation development. Crystal polymorphs also represent a key aspect of patent strategies and quality assurance systems, making them critically important for research and development by pharmaceutical companies. Conventionally, crystal polymorph production has been controlled primarily by adjusting chemical and thermodynamic parameters, such as solvent selection, temperature control, cooling rate, and recrystallization method [15,16,17].

Optical control and applied physics have recently seen dramatic progress in research on metamaterials—materials with metal micro-pattern structures on the scale of wavelengths of light or smaller. In this study, we developed a device capable of emitting infrared radiation at selected wavelengths using a novel material having a Metal-Insulator-Metal (MIM) layered structure, which is a type of metamaterial. With this device, we propose a new approach to crystal polymorph control through the irradiation of narrow-band infrared radiation that coincides with the infrared absorption band of specific functional groups.

During crystallization, low- to medium-molecular-weight organic compounds such as active pharmaceutical ingredients form aggregates through various intermolecular interactions or forces, such as the van der Waals force, hydrogen bonds, and π-π interactions, to produce hierarchical crystal structures. Infrared radiation can directly excite the functional groups involved in these intermolecular interactions, thereby inhibiting the formation of specific bonds and the nucleation of specific crystals.

This paper discusses the design and operating principles of a novel crystal precipitation system that incorporates the above-mentioned selective-wavelength infrared irradiation technique. We also present experimental results demonstrating crystal polymorph control for Ritonavir, an active pharmaceutical ingredient, as a potential real-world application.

2. Experimental Apparatus

This crystallization method is based on what is known as the evaporation method, the most distinctive feature of which involves the continuous irradiation with infrared radiation at selected wavelengths during the solvent evaporation process—i.e., during solute recrystallization. The solute crystals are obtained as residues after the solvent is completely removed by drying. Infrared radiation is the electromagnetic radiation emitted by the random thermal vibrations of a molecule. Thus, it typically exhibits very weak wavelength selectivity and directionality in radiation, which is the propagation phenomenon of the emitted electromagnetic waves. In practice, as will be described later, infrared wavelength control requires significant engineering work. While reports do exist on basic research and the processes that account for the spectral affinity between emitters and absorbers [18,19,20,21], very few case studies have been performed to date.

To achieve infrared wavelength selection for the purposes of this study, we developed a selective-wavelength MIM emitter. The main wavelength of the radiation from this MIM emitter can be set arbitrarily within the range of approximately 3.0 μm to 8.0 μm, depending on the structural design. While the half-widths of the main wavelength band vary slightly depending on the wavelength, the bandwidth is typically narrow, at around 0.4 μm, and rarely exceeds 1.0 μm. This enables the emitter to target absorption bands of specific functional groups found in various organic compounds. This feature is expected to have the effect of dissociating the aggregates produced by specific intermolecular interactions through selective excitation of the vibration of the target functional group.

It should be noted that the hydrogen bond is considered to be the strongest among the intermolecular interactions, generally on the order of several tens of kJ/mol. If we were to assume that this energy is 30 kJ/mol, it would be equivalent to approximately 0.31 eV. On the other hand, the photon energy (E) of infrared radiation with a wavelength (λ) of 3 μm can be quantized using the relationship E = hν = hc/λ [J] (where h is Planck’s constant, ν is frequency, and c is the speed of light), which gives the value 0.41 eV.

So theoretically, infrared radiation should have sufficient energy to cause dissociation of an independent hydrogen bond. Figure 1 shows the experimental apparatus equipped with this MIM emitter and the experimental system.

Figure 1. Experimental apparatus (left: schematic illustration; and right: the actual apparatus).

3. MIM Emitter

Figure 2 shows a photo of the actual MIM emitter and a schematic of its structure.

Figure 2. The MIM emitter and its conceptual diagram (when the metal pattern is circular).

While Figure 2 shows a circular metal pattern, other shapes are possible, such as rectangular or cross shapes. For our study, we chose to use a square pattern. The principle of selective infrared irradiation by the MIM radiation plate is as follows: The three-layered structure of this metal-insulator-metal metamaterial is known to exhibit wavelength selectivity due to its magnetic resonance modes [22,23,24,25,26]. When electromagnetic waves arrive at the MIM structure, anti-parallel currents of specific wavelengths are induced in the upper and lower metal layers. These electric fields generate a magnetic field within the insulator; the changes in both fields amplify each other, enabling selective wavelength electromagnetic wave irradiation and absorption. In the method presented in this study, the MIM emitter is designed so that the main irradiation wavelength coincides with the absorption band generated by the molecular vibration of specific functional groups, which are thought to be involved in hydrogen bond formation between the solute molecules. Examples of target absorption bands include the absorption band attributable to the O-H stretching vibration of the carboxy group at around 3.0–3.3 μm and the absorption band attributable to the C=O stretching vibration of the carbonyl group at around 5.8 μm.

To fabricate the MIM structure, a metal layer is sputtered onto a sapphire substrate, followed by the deposition of aluminum oxide. The patterning on the outermost layer is then formed by electron beam lithography. An adhesive layer of Cr or similar material is added between each layer. Figure 3 and Table 1 give the detailed structure of the MIM emitters and the design values of the radiation plates used in the measurements below. The design values given in Table 1 are the specifications for the actual radiation plates used in the experiments described in Section 6.

Figure 3. Example of the cross-sectional structure of a MIM selective radiation plate.

Table 1. Example of the design values of the MIM selective radiation plate.

Emitter Name	w [nm]	h [nm]	d [nm]	f [nm]	Λ [nm]
Emitter 3.3	615	45	46	200	1000
Emitter 5.8	1450	79	124	200	3200
Emitter 6.7	1740	112	121	200	4000

4. Radiation Energy

We measured the normal-incident hemispherical reflectance spectra of the emitter having the fabricated MIM structure using a Fourier transform infrared (FTIR) spectrometer with an integrating sphere. We then calculated the normal emissivity spectra based on assuming the sum of reflectance, absorptance, and transmissivity to be 1, Kirchhoff’s law, and the assumption that transmissivity is 0. Figure 4 shows the normal emissivity of each MIM emitter presented in Table 1. The yellow vertical line in each plot indicates the target wavelengths for each emitter. We measured normal emissivity at five points: center, top, bottom, left, and right. Although the spectral shapes of the five points show a slight deviation at 3.3 μm, they are very similar, and the graphs are displayed almost overlapping. Despite small peaks attributable to higher-order resonance modes on the shorter wavelength side, we confirmed that high emissivity is achieved mainly around the target wavelength. For example, in the case of Emitter 3.3, a peak of normal emissivity occurs at around 3.3 μm with a half width of approximately 0.4 μm. The MIM emitter records normal emissivity exceeding 0.84 at all target wavelengths, achieving high values. Figure 5 compares the spectral radiant energy of the MIM emitter and a conventional ceramic heater at equivalent input energies. We see that the MIM emitter exhibits significantly improved narrow-band radiation spectrum characteristics compared to the conventional ceramic heater. Since the radiation energy is strongly dependent on the temperature of the MIM emitter itself, the emitter is typically heated to approximately 600 °C using a separate heat source in the experiments to achieve the desired effect. The infrared energy on the liquid surface is approximately 0.2 to 0.3 [W/cm²].

Figure 4. Normal emission spectra of the MIM emitters.

Figure 5. Comparison of emission spectra.

5. Experimental Method

We performed an IR-induced crystallization experiment for crystal polymorphs using the experimental apparatus described in Section 2 and Section 3. The apparatus shown in Figure 1 is housed in a PET resin case that can isolate it from the external air. Dry air (at a dew point of approx. −5 °C) is fed into the case at approximately 15 L/min. A Petri dish containing the sample is placed on the Peltier temperature controller, directly below the ZnSe plate serving as the radiation transmission window. The sample consists of a solution in which the target substance to be crystallized (solute) is completely dissolved (confirmed visually) in the specified solvent. While the amount of solvent and solute can be set arbitrarily, 20 to 200 mg of a solute is typically dissolved in about 1 mL of solvent (See Supplement S4 for details).

The surface of the MIM emitter is heated using a ceramic heater attached to the back side and maintained at approximately 400 to 600 °C throughout the experiment. However, since the MIM emitter is housed in a vacuum chamber in which convective heat transfer is blocked, the temperatures around the sample can be freely adjusted, regardless of heater temperature. The top part of the Petri dish is in contact with the air inside the case. Heat exchange occurs between the solution and the air through natural convective heat transfer. The solvent evaporated from the liquid surface is transported and diffused inside the container by natural convection and removed from the system through the exhaust port attached to the case. Thus, the solution temperature is determined by a complex physiochemical process that involves a combination of multiple factors, including the selective wavelength radiation from the ZnSe window, convective heat transfer within the case, the latent heat of evaporation of the solvent, the heat of dissolution and solidification of the solute, and heat transfer at the contact between the Petri dish and the Peltier temperature controller, all of which continuously change with time. After the completion of solvent crystallization (absolute dryness), the crystals in the Petri dish are collected and their crystal forms verified by powder XRD and other analytical methods.

6. Results of the Experiment (Separation of Ritonavir Crystal Polymorphs)

As an example of IR-induced crystallization with our experimental apparatus, we present the results of a series of experiments targeting Ritonavir, a protease inhibitor used as an HIV treatment drug. Crystal form I, believed at the time to be the most stable phase, was the form initially marketed as Ritonavir. However, following its release, an even more stable phase (crystal form II) was discovered, and the manufacturer was forced to redesign their entire pharmaceutical production process. This incident has made Ritonavir a famous example of the problems posed by polymorphs (Figure 6).

Figure 6. Molecular structure of Ritonavir (molecular weight: 721).

Although slight differences can be observed between the intensity of the infrared absorption spectra of crystal forms I and II, they also share many characteristics at the peak positions (Figure 7). In the present experiment, we selected the main irradiation wavelength by taking into consideration the absorption bands for the amide group (3.0 μm/3300 cm⁻¹), carbamate group (3.0–3.3 μm/3300–3000 cm⁻¹), hydroxyl group (3.3 μm/3000 cm⁻¹), and thiazole group (3.0 μm/3300 cm⁻¹). These are the representative functional groups that form the intermolecular hydrogen bonds of Ritonavir.

Figure 7. Comparison of the absorption spectra of individual crystal forms.

This experiment involved dissolving 200 mg of Ritonavir powder in a Petri dish containing 1 mL of solvent at 85 °C until visual confirmation of complete dissolution, followed by crystal precipitation performed multiple times while changing the parameters. The parameters were the main irradiation wavelength and the temperature of the Peltier controller. Table 2 shows the other experimental conditions and the crystal forms obtained from the experiment. Note that the temperature of the solution was maintained at 50 °C using the Peltier controller during crystal precipitation under each condition. For details of other conditions, please see the Supplementary Materials.

Table 2. Crystal forms obtained under each experimental condition.

Experiment No.	Main Wavelength	Heater Temperature [°C]	Peltier Temperature * [°C]	Precipitated Crystal Form
1	None	None	53	I (pseudo stable)
2	3.3	600	25	II (most stable)
3	5.8	600	25	II (most stable)
4	6.7	600	53	I (pseudo stable)

* Peltier temperature to bring the solution temperature to 50 °C.

The XRD measurement results that served as the basis for determining the precipitated crystal forms in Table 2 are shown in Supplement S2. Form I: 6.8°, Form II: 9.5°, 10°, and 11° were focused on, and the presence or absence of peaks near those positions was determined.

Under normal crystallization conditions, the pseudo-stable crystal form I is precipitated. In other words, crystal form I is the more favored form compared to crystal form II in terms of precipitation rate. As in other studies, crystal form I is the phase obtained under conditions without infrared irradiation. On the other hand, applying infrared irradiation at wavelengths of 3.3 μm and 5.8 μm produces the stable crystal form II. Further, applying infrared radiation at wavelengths of 6.7 μm results once again in the precipitation of pseudo-stable crystal form I. This suggests that the crystal form obtained can be controlled with the wavelength of the applied infrared radiation. The mechanism underlying this phenomenon is thought to be as follows.

From Figure 7, we see that, while the spectral forms of crystal forms I and II are essentially consistent with each other, the strongest hydrogen bonds between the functional group corresponding to the main absorption band for crystal form II (the most stable form) are those formed between the amide group and the carbamate group. Although the hydrogen bonds between the amide groups are also present in crystal form I (pseudo-stable form) and the absorption band for that bond resembles that of the hydrogen bond between the amide group and the carbamate group, this bond is weaker [27,28]. Based on the observation that the infrared absorption peaks of the associated functional groups can generally be found in the ranges of 3.0–3.4 μm or 5.7–6.0 μm and that the infrared absorption peaks in the range between 3100 [cm⁻¹] (3.2 μm) and 3300 [cm⁻¹] (3.0 μm) are wider and more pronounced for crystal form II than for crystal form I, we can conclude that a much stronger bond has been formed in crystal form II. This conclusion is also supported by an analysis of the crystal structures of crystal forms I and II. This suggests that the nucleus for crystal form I is more likely to undergo dissociation due to vibrational excitation of the functional group induced by infrared irradiation at specific wavelengths.

Thus, under conditions 2 and 3, in which infrared irradiation is performed continuously in the infrared absorption wavelength regions corresponding to the individual functional group, we believe that the nucleation of crystal form I, which is favored in terms of precipitation rate (but is not favored thermodynamically), is inhibited during the initial stages of crystallization, while that of crystal form II, which is more thermodynamically stable, is promoted. On the other hand, in the wavelength region of 6.7 μm under condition 4, no absorption by the hydrogen bonds of the functional groups is observed. We believe this suggests that the formation of crystal form I, which is favored in terms of precipitation rate, proceeds.

7. Conclusions

In this study, we proposed a novel crystallization method that incorporates infrared irradiation using a metamaterial-type emitter (MIM emitter) capable of emitting infrared radiation at selected wavelengths. We demonstrate its potential application to the screening of crystal polymorphs of active pharmaceutical ingredients. Our crystallization method, based on the evaporation method, which involves infrared irradiation at a specific wavelength during the recrystallization process, can contribute to the ultimate control of crystal polymorphs by directly influencing crystal nucleation due to intermolecular interactions.

In particular, since MIM emitters are capable of narrow-band and wavelength-selective infrared irradiation, wavelengths that coincide with the vibrational absorption bands of functional groups can be selected to achieve selective excitation at the molecular level. In an experiment with the protease inhibitor Ritonavir, we were able to induce the selective formation of a pseudo-stable phase (crystal form I) or the most stable phase (crystal form II) by selecting the irradiation wavelength, suggesting that selective infrared wavelength absorption by intermolecular hydrogen bond networks may play an essential role in the control of crystal polymorphs.

This study introduces infrared irradiation as a new control factor in addition to solvent, temperature, precipitation rate, and other such conventional factors for crystallization. We anticipate our method to become a promising technique for strategically controlling crystal polymorphs in pharmaceutical manufacturing. In future studies, we plan to further analyze the mechanism by which infrared irradiation influences the crystal nucleation process using spectroscopic methods and simulations, with the goal of developing a new and universally applicable crystallization control method.