*Article* **Chiral** π**-Conjugated Liquid Crystals: Impacts of Ethynyl Linker and Bilateral Symmetry on the Molecular Packing and Functions**

**Atsushi Seki \*, Kazuki Shimizu and Ken'ichi Aoki**

Department of Chemistry, Faculty of Science Division II, Tokyo University of Science, 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan

**\*** Correspondence: a\_seki\_3@rs.tus.ac.jp; Tel.: +81-3-3260-4271

**Abstract:** Recently, various chiral aromatic compounds, including chiral π-conjugated liquid crystals, have been developed for their unique photofunctions. One of the typical photofunctions is the bulk photovoltaic effect of ferroelectric π-conjugated liquid crystals, which integrates a polar environment based on molecular chirality with an extended π-conjugation system. Tuning the spectral properties and molecular packing is essential for improving the optical functions of the chiral π-conjugated liquid crystals. Herein, we examined the effects of an ethynyl linker and bilateral symmetry on the liquid-crystalline (LC) properties and π-conjugated system through detailed characterization via polarizing optical microscopy, differential scanning calorimetry, and X-ray diffraction analysis. The spreading of the π-conjugated system was evaluated using UV–vis absorption and photoluminescence spectroscopy. Bilateral symmetry affects the LC and photoluminescent properties. Hetero-substitution with a sparse ethynyl linker likely allows the formation of an interdigitated smectic LC structure. Because the molecular packing and photophysical properties can affect the photo- and electrical functions, we believe this study can promote the molecular design of novel functional π-conjugated materials, such as chiral ferroelectric π-conjugated liquid crystals, exhibiting the bulk photovoltaic effect.

**Keywords:** molecular chirality; π-conjugated compound; liquid crystal

### **1. Introduction**

Molecular chirality can induce the formation of hierarchical suprastructures, which acts as platforms for biological, pharmacological, chemical, and physical functions [1–7]. A broken-symmetry structure is a self-organized structure that reflects the molecular chirality. Symmetry reduction leads to the stabilization of polar structures. Thus, ferroelectricity can be observed in such chiral suprastructures [8–10]. Another representative chiral supramolecular system is the helical self-assembly. The absolute configuration of the chiral molecules labeled (*R*) or (*S*) reflects their helical structure and axis. Because the helical conformation due to inherent molecular chirality is known to contribute to various functionalities of selfassembled materials, chiral materials, including chiral polymers [6,11], chiral supramolecular polymers [6,12], and chiral liquid crystals (CLCs) [2,13–15], have been extensively developed and studied. In particular, CLCs show sensitive responses to external stimuli, such as temperature and electric fields, because of their dynamic nature [14,15]. In recent years, we have focused on chiral smectic liquid crystals resulting from introducing molecular chirality into smectic liquid crystal systems [16–19]. From both basic scientific and engineering standpoints, the most important and beneficial chiral smectic liquid crystal is a ferroelectric liquid-crystalline material that exhibits a chiral smectic C (SmC\*) phase. In the neutral SmC\* phase, the CLC molecules form a tilted-layer structure with helical twisting of the molecular axis along the normal layer. When an electric field is applied to planar-aligned CLC molecules with a transverse dipole moment in the SmC\* phase, molecular reorientation should occur

**Citation:** Seki, A.; Shimizu, K.; Aoki, K. Chiral π-Conjugated Liquid Crystals: Impacts of Ethynyl Linker and Bilateral Symmetry on the Molecular Packing and Functions. *Crystals* **2022**, *12*, 1278. https:// doi.org/10.3390/cryst12091278

Academic Editor: Charles Rosenblatt

Received: 9 August 2022 Accepted: 6 September 2022 Published: 9 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

to unwind helical structures. Because molecular chirality can stabilize the polar structure owing to the reduction of structural symmetry, ferroelectric properties are often observed in the SmC\* phase. Conventional CLCs have been investigated for their applications such as in optical sensors [20] and high-speed liquid-crystal displays [21]. While conventional CLCs are generally insulators, chiral π-conjugated liquid crystals have the potential to be unique photofunctional materials [16–19,22]. Many π-conjugated liquid crystals have been synthesized, and their electronic charge carrier transport properties have been explored as liquid-crystalline (LC) semiconductors [23–28]. LC materials have some advantages such as improved solubility in common organic solvents, the control of molecular orientation, and the formation of uniform thin films against inorganic semiconductors. Therefore, LC semiconductors have been frequently used as active materials in optoelectronic devices such as bulk heterojunction organic photovoltaic devices [29,30], organic light-emitting diodes [31–33], and organic thin-film transistors [34–36].

This study aimed to develop chiral π-conjugated liquid crystals for novel optoelectronic materials. Recently, exciting applications of chiral π-conjugated liquid crystals have been reported. For instance, Funahashi et al. developed electric-field-responsive CLCs that exhibited an SmC\* phase. As terthiophene-based CLCs show ferroelectricity and photoconductivity in the SmC\* phase, the combined effect of spontaneous polarization and carrier transport results in a bulk photovoltaic effect in the LC phase [37]. The bulk photovoltaic effect based on molecular chirality is a newly discovered type of ferroelectric photovoltaic (FePV) effect, which is classified as one of the bandgap-independent photovoltaic effects [38–41]. As the FePV effect shows unique characteristics, such as ultrafast spontaneous photocurrent [42], low noise current [43], and no dissipation [44], the anomalous photovoltaic effect in ferroelectrics is evidently different from conventional photovoltaic effects based on p-n junctions [40,41]. Therefore, the FePV effect has attracted considerable attention in material chemistry and physics. While the FePV effect in ferroelectric ceramics has been widely investigated for several decades [38–44], reports on the FePV effect in organic materials other than the FePV effect of CLCs [16–19,22] are still limited [45–47]. The FePV effect for organic materials is essential for developing novel high-performance organic photoelectronic devices, including organic photovoltaic cells and organic photodetectors [47]. In fact, the FePV effect with a high open-circuit voltage of over 1 V was recently achieved by using CLCs doped with a fullerene derivative [48]. The exploration of CLCs, which are candidates for the active materials of the FePV effect, has only begun and is still important. In particular, tuning the light absorption property is a significant factor in realizing a large short-circuit current density, resulting in efficient charge carrier generation. The most common approach for tuning spectroscopic properties is expanding the π-conjugated systems, such as by introducing an ethynyl linker.

In this study, we examined the influence of the ethynyl linker between oligothiophene and chiral fluorophenyl units on the LC and photophysical properties. In addition, the impact of bilateral symmetry of the chiral compounds upon those properties were studied. We synthesized three chiral π-conjugated compounds, (*R*)-**1**, (*R*)-**2**, and (*R*)-**3** (Figure 1). Molecular packing in the smectic LC phase and its spectroscopic properties were also investigated.

**Figure 1.** Chemical structures of the chiral π-conjugated compounds (*R*)-**1**, (*R*)-**2**, and (*R*)-**3**.

### **2. Materials and Methods**

### *2.1. General Procedures and Materials*

All reagents were purchased from Sigma-Aldrich Japan (Tokyo, Japan), Tokyo Chemical Industry Co., Ltd.(Tokyo, Japan), Kanto Chemicals (Tokyo, Japan), and FUJIFILM Wako Pure Chemicals (Osaka, Japan) and were used without further purification. All the reactions were performed under an argon atmosphere in a well-dried flask equipped with a magnetic stirring bar. The synthetic scheme for the target compounds is described in the next section (Section 2.2. Synthesis). The details of synthetic conditions are described in the attached Supplementary Files, Section S1. All 1H and 13C NMR spectra were recorded on a Bruker (Osaka, Japan) Biospin AVANCE NEO 400 spectrometer in CDCl3 (400 MHz for 1H NMR spectra, 100 MHz for 13C NMR spectra). All chemical shifts (δ) in the 1H and 13C NMR spectra are quoted in ppm using tetramethylsilane (δ = 0.00) as the internal standard (0.03 vol%). High-resolution mass spectrometry (HRMS) measurements were carried out by electrospray ionization using a SCIEX (Tokyo, Japan) X500R QTOF spectrometer. Elemental analysis was entrusted to A-Rabbit-Science Japan Co., Ltd. (Kanagawa, Japan).

### *2.2. Synthesis*

The chiral π-conjugated compounds (*R*)-**1**, (*R*)-**2**, and (*R*)-**3** were synthesized according to the procedures shown in Scheme 1. All compounds were synthesized via Pd-catalyzed C-C coupling reactions. The chiral starting material (*S*)-2-octanol was purchased from Tokyo Chemical Industry Co., Ltd. (Specification value: chemical purity ≥ 98.0%, optical purity ≥ 98.0%ee). Compounds **4**, (*R*)-**5**, **6**, (*R*)-**7** and **8** were synthesized with reference to literatures [17,19,49–51]. The chiral compound (*R*)-**5** was synthesized via the Suzuki–Miyaura reaction between 2,2'-bithiophene-5-boronic acid pinacol ester and 4-bromo-2-fluoro-1-{(*R*)- 2-octyloxy}benzene. 4-Bromo-2-fluoro-1-{(*R*)-2-octyloxy}benzene was synthesized via the Mitsunobu reaction between 4-bromophenol and (*S*)-2-octanol. It is noted that the optical purity of (*S*)-2-octanol is guaranteed ≥ 98.0%ee by the standard. Because the Mitsunobu reaction generally undergoes the typical SN2 displacement pathway, chiral inversion must be caused [52]. Shi et al. reported that the Mitsunobu reaction using chiral alcohols exhibiting high enantiomeric excess (> 90%ee) with phenol derivatives afford the product with high optical purity (> 90%ee) [53]. Based on these findings, various chiral liquid crystals have been synthesized from (*S*)-2-octanol or (*R*)-2-octanol the several reaction steps including Mitsunobu reaction and C-C cross-coupling reactions [54–57]. 1H-, 13C NMR and HRMS spectra for the target compounds (*R*)-**1**, (*R*)-**2**, and (*R*)-**3** are shown in the ESI, Sections S2 and S3.

**Scheme 1.** Synthesis routes of compounds (*R*)-**1**, (*R*)-**2**, and (*R*)-**3**.

### 2.2.1. Characterization of (*R*)-**1**

5-Octyl-5"-{3-fluoro-4-[(*R*)-2-octyloxy]phenyl}-2,2':5',2"-terthiophene: (*R*)-**1**

1H NMR (400 MHz, CDCl3): δ [ppm] = 7.30 (dd, 1H, *J* = 12.2, 2.2 Hz), 7.25 (ddd, 1H, *J* = 8.4, 2.4, 1.2 Hz), 7.10 (d, 1H, *J* = 3.6 Hz), 7.08 (d, 1H, *J* = 3.6 Hz), 7.05 (d, 1H, *J* = 3.6 Hz), 6.99 (d, 1H, *J* = 4.0 Hz), 6.98 (d, 1H, *J* = 3.2 Hz), 6.95 (t, 1H, *J* = 8.6 Hz), 6.68 (d, 1H, *J* = 3.6 Hz), 4.37 (sextet, 1H, *J* = 6.0 Hz), 2.79 (t, 2H, *J* = 7.4 Hz), 1.86–1.55 (m, 4H), 1.51–1.20 (m, 18H), 1.33 (d, 3H, *J* = 6.0 Hz), 0.89 (t, 6H, *J* = 7.0 Hz); 13C NMR (100 MHz, CDCl3): δ [ppm] = 153.8 (d, *J* = 244.2 Hz), 145.8 (d, *J* = 10.9 Hz), 145.7, 141.8 (d, *J* = 2.2 Hz), 136.9, 136.3, 135.4, 134.4, 127.8 (d, *J* = 7.2 Hz), 124.9, 124.3, 124.1, 123.5 (d, *J* = 15.9 Hz), 123.4, 121.4 (d, *J* = 3.6 Hz), 117.9 (d, *J* = 2.7 Hz), 113.8, 113.6, 76.6, 36.5, 31.9, 31.8, 31.6, 30.2, 29.3, 29.3, 29.2, 29.1, 25.4, 22.7, 22.6, 19.8, 14.1, 14.1; HRMS (ESI): molecular weight: 582.8954 (C34H43FOS3); m/z calculated for [C34H43FOS3] +: 582.2455 ([M]+); found: 582.2456; elemental analysis (%) calculated for C34H43FOS3: C 70.06, H 7.44, F 3.26, O 2.74, S 16.50; found: C 69.81, H 7.28.

### 2.2.2. Characterization of (*R*)-**2**

5-Octyl-5"-({3-fluoro-4-[(*R*)-2-octyloxy]phenyl}ethynyl)-2,2':5',2"-terthiophene: (*R*)-**2**

1H NMR (400 MHz, CDCl3): δ [ppm] = 7.25–7.18 (m, 2H), 7.14 (d, 2H, *J* = 3.6 Hz), 7.07 (d, 1H, *J* = 4.0 Hz), 7.03 (d, 1H, *J* = 4.0 Hz), 6.99 (d, 1H, *J* = 4.0 Hz), 6.99 (d, 1H, *J* = 3.6 Hz), 6.91 (t, 1H, *J* = 8.8 Hz), 4.39 (sextet, 1H, *J* = 6.0 Hz), 2.79 (t, 2H, *J* = 7.6 Hz), 1.85–1.55 (m, 4H), 1.52–1.22 (m, 18H), 1.33 (d, 3H, J = 6.4 Hz), 0.88 (t, 6H, *J* = 6.8 Hz); 13C NMR (100 MHz, CDCl3): δ [ppm] = 152.9 (d, *J* = 244.8 Hz), 147.0 (d, *J* = 10.9 Hz), 146.0, 138.7, 137.6, 134.7, 134.2, 132.7, 127.9 (d, *J* = 3.6 Hz), 124.9, 124.8, 123.6 (d, *J* = 3.2 Hz), 123.2, 121.7, 119.4, 119.2, 116.7 (d, *J* = 2.4 Hz), 115.3 (d, *J* = 8.3 Hz), 93.2, 82.0, 76.3, 36.4, 31.9, 31.8, 31.6, 30.2, 29.3, 29.2, 29.1, 25.4, 22.7, 22.6, 19.8, 14.1, 14.1; HRMS (ESI): molecular weight: 606.9174 (C36H43FOS3); m/z calculated for [C36H43FOS3] +: 606.2455 ([M]+); found: 606.2453; elemental analysis (%) calculated for C36H43FOS3: C 71.24, H 7.14, F 3.13, O 2.64, S 15.85; found: C 71.25, H 7.16.

### 2.2.3. Characterization of (*R*)-**3**

5,5'-Bis({3-fluoro-4-[(*R*)-2-octyloxy]phenyl}ethynyl)-2,2'-bithiophene: (*R*)-**3**

1H NMR (400 MHz, CDCl3): δ [ppm] = 7.25–7.19 (m, 4H), 7.15 (d, 2H, *J* = 3.6 Hz), 7.07 (d, 2H, *J* = 4.0 Hz), 6.92 (t, 2H, *J* = 8.6 Hz), 4.40 (sextet, 2H, *J* = 6.2 Hz), 1.85–1.72 (m, 2H), 1.67–1.55 (m, 2H), 1.52–1.24 (m, 16H), 1.33 (d, 6H, *J* = 6.4 Hz), 0.88 (t, 6H, *J* = 6.8 Hz); 13C NMR (100 MHz, CDCl3): δ [ppm] = 152.9 (d, *J* = 245.4 Hz), 147.1 (d, *J* = 10.8 Hz), 138.0, 132.7, 128.0 (d, *J* = 3.0 Hz), 124.0, 122.5, 119.3 (d, *J* = 20.2 Hz), 116.7 (d, *J* = 2.4 Hz), 115.2 (d, *J* = 8.5 Hz), 93.5 (d, *J* = 2.8 Hz), 81.8, 76.3, 36.4, 31.8, 29.2, 25.4, 22.6, 19.8, 14.1; HRMS (ESI): molecular weight: 658.9068 (C40H44F2O2S2); m/z calculated for [C40H45F2O2S2] +: 659.2824 ([M+H]+); found: 659.2828; elemental analysis (%) calculated for C40H44F2O2S2: C 72.91, H 6.73, F 5.77, O 4.86, S 9.73; found: C 72.99, H 6.86.

### *2.3. Characterization of LC Properties*

The LC properties of chiral π-conjugated compounds were characterized using differential scanning calorimetry (DSC), polarizing optical microscopy (POM), and X-ray diffraction (XRD). DSC measurements were conducted using a SHIMADZU (Kyoto, Japan) DSC-60 system equipped with a liquid nitrogen auto-cooling system (TAC-60L). Approximately 2–3 mg of each sample was sealed in an aluminum pan. The optical texture was observed using a polarizing optical microscope (Olympus BH2, Olympus Corporation, Tokyo, Japan) equipped with a digital camera (AS ONE HDCE-X1 (AS ONE Corporation, Osaka, Japan) and a temperature control system (METTLER TOLEDO FP90 and FP82HT). Indium tin oxide (ITO) sandwich cells filled with chiral π-conjugated compounds were used for POM observations. Empty ITO sandwich cells (KSSO-02/A311P1NSS05, cell gap: 2 μm) were purchased from EHC Corporation (Tokyo, Japan). The ITO surface without a polyimide was rubbed to assist in the planar orientation of the smectic phases. The scan rate of DSC measurements and POM observations was 10 ◦C min<sup>−</sup>1. XRD measurements were performed using a Rigaku RINT-2500 (Ni-filtered Cu Kα radiation, Rigaku Corporation,

Tokyo, Japan) equipped with a custom-made thermal control system composed of a silicone rubber heater, thermocouple sensor, and PID-type thermal controller (AS ONE TJA-550).

### *2.4. Characterization of Spectroscopic Properties*

UV–vis absorption spectra were recorded using a JASCO (Tokyo, Japan) V-650 spectrometer. UV-vis absorption spectra were measured using a pair of quartz cells (cell gap: 1 cm). The photoluminescence (PL) emission spectra were recorded using a SHIMADZU (Kyoto, Japan) RF-6000 spectrometer. Emission spectra were measured using a pair of quartz cells (cell gap: 1 cm).

### **3. Results and Discussion**

### *3.1. Liquid-Crystalline Properties*

### 3.1.1. Polarizing Optical Microscopy

In the POM observation of chiral phenylterthiophene derivative (*R*)-**1**, a broken fan-like texture with stripes was observed in the area where the sample was sandwiched between two ITO electrodes, at approximately 140 ◦C upon cooling from the isotropic liquid (IL) state (Figure 2a). The broken fan-like domains suggest the formation of an LC tilted-layer structure. Furthermore, the stripe pattern in each fan-shaped domain should be derived from the disclination. Therefore, the characteristic optical textures indicated the appearance of a chiral smectic C (SmC\*) phase with a nonpolar helical structure. When the sample was cooled to approximately 130 ◦C, the polarized optical texture was transformed, and tile-like domains were observed in the homeotropic domains of the SmC\* phase between the two glass substrates (Figure 2b). This change in texture corresponds to a phase transition from SmC\* to ordered smectic phases. Upon further cooling to room temperature, the domain shapes were maintained without drastic textural changes (Figure 2c). In the POM study of compound (*R*)-**2**, an ethynyl linker introduced between the terthiophene and chiral fluorophenyl units, we observed a fan-shaped texture upon cooling from the IL phase (Figure 3a). A typical fan-shaped texture shows the appearance of a smectic LC phase. Because the color and contrast of the optical texture vary as the sample temperature of (*R*)-**2** decreases to approximately 110 ◦C (Figure 3b), the high-temperature smectic LC phase is changed to another smectic LC phase at this temperature. After cooling below 100 ◦C, the stripes appeared in fan-shaped domains (Figure 3c). In the polarized optical texture of (*R*)-**2** at 45 ◦C, the stripes of fan-shaped domains are more conspicuous (Figure 3d). The slight change of optical texture is probably due to rearrangements in the intralayer molecular packing. These results support that (*R*)-**2** exhibited several smectic LC phases. In the POM observation of compound (*R*)-**3** modified with chiral fluorophenyl units on both wings of the 2,2'-bithiophene core, two types of optical textures were observed during cooling from the IL phase (Figure 4a,b). However, these textures differ from the distinctive textures of LC phases. Therefore, we conclude that (*R*)-**3** does not exhibit LC properties.

**Figure 2.** POM images of (*R*)-**1** at (**a**) 140 ◦C, (**b**) 95 ◦C, and (**c**) 40 ◦C. The black arrows indicate the border of ITO electrode and glass surface.

**Figure 3.** POM images of (*R*)-**2** at (**a**) 120 ◦C, (**b**) 105 ◦C, (**c**) 90 ◦C, and (**d**) 45 ◦C.

**Figure 4.** POM images of (*R*)-**3** at (**a**) 70 ◦C, and (**b**) 40 ◦C.

3.1.2. Differential Scanning Calorimetry

The DSC thermogram of (*R*)-**1** exhibits two distinct endothermic peaks due to firstorder phase transitions during the second heating (Figure 5a). Although a crystal-LC phase transition peak is found at 63.3 ◦C on the first heating, no endothermic peak of crystal-LC transition is seen on subsequent heating scans. Similar phase transition behaviors are found in analogous phenylterthiophene derivatives [17,37]. The inconsistency in the number of first-order phase transition peaks between first cooling and second heating scans

suggests the existence of a monotropic metastable mesophase during cooling. When we consider the results of POM and DSC studies of (*R*)-**1**, the exothermic peak at 144.2 ◦C during cooling indicates the IL–SmC\* phase transition. The following peak observed at 131.3 ◦C corresponds to the transition from SmC\* to metastable ordered smectic phases. The metastable smectic phase transforms to a more stable ordered smectic phase at 66.5 ◦C during cooling. On the second heating scan, the endothermic peak of the ordered smectic-SmC\* phase transition is observed at 132.5 ◦C, and the SmC\*-IL phase transition follows at 144.8 ◦C. In the DSC thermogram of (*R*)-**2**, three peaks are observed during the second heating (Figure 5b). These peaks correspond to the two LC–LC phase transitions and an LC–IL phase transition based on the results of the POM study. The middle LC phase changes to the low-ordered smectic phase at 111.5 ◦C after the first LC–LC phase transition occurs at 101.3 ◦C. The smectic LC structure and the molecular order collapses at 126.0 ◦C. The broad exothermic peak, at approximately 67 ◦C, is observed on the first cooling scan, corresponding to the transition from metastable to stable states (Figure 5b). A metastable LC phase is observed during cooling for compound (*R*)-**2**. Therefore, we conclude that compounds (*R*)-**1** and (*R*)-**2** exhibit similar monotropic behavior. The broad tolerance of molecular packing style in (*R*)-**2** should be also reflected to the complicated phase transition behaviors on the first heating process. Compound (*R*)-**3** also shows several phase transitions at 52.3 and 77.2 ◦C, as observed in the second heating scan (Figure 5c). The DSC and POM studies support that compound (*R*)-**3** exhibits crystalline polymorphism.

**Figure 5.** DSC thermograms of (**a**) (*R*)-**1**, (**b**) (*R*)-**2**, and (**c**) (*R*)-**3** at a scanning rate of 10 ◦C min<sup>−</sup>1.

### 3.1.3. X-ray Diffraction

The variable-temperature XRD measurements were conducted for the chiral π-conjugated compounds (*R*)-**1**, (*R*)-**2**, and (*R*)-**3** to gain insight into their molecular packing and selfassembled structures. The XRD pattern of (*R*)-**1** in the SmC\* phase at 139 ◦C (Figure 6a, upper) exhibits diffraction peaks at 2*θ* = 3.12◦, 6.10◦, 12.11◦, 15.15◦, and 18.21◦, which correspond to diffractions from the (001), (002), (003), (004), (005), and (006) planes, respectively. Because all *d*-spacings estimated from these diffractions can be expressed as integer ratios, the XRD pattern also indicates that a smectic-layer structure at approximately 140 ◦C can be formed. The molecules in the smectic-layer structure should be tilted with respect to the normal of the layer because the layer spacing (29 Å) is shorter than the theoretical extended molecular length (35 Å) of (*R*)-**1** estimated by the molecular mechanics calculation (Energy minimization calculation, MM2 force field, PerkinElmer, Chem3D 18.1). The SmC\* phase is observed between 144 and 131 ◦C upon cooling of (*R*)-**1**. This observation coincides with those of the preceding POM and DSC studies. In the XRD pattern, during cooling, we observed a diffraction peak with a low intensity at 95 ◦C in the wide-angle region (2*θ* = 19.30◦) and several other diffraction peaks that represented from the smectic-layer structures (Figure 6a, middle). The low-intensity peak can be assigned to the (010) plane, reflecting the intralayer order. From the periodic diffraction peaks corresponding to the (001), (002), (003), (004), and (006) planes, the layer spacing is estimated to be 29 Å. Therefore, the tilt angle remains unchanged through the SmC\*–LC phase transition. As no other peaks are observed in the wide-angle region, the intralayer order in the metastable smectic phase should be confined in the short range. Therefore, we consider the metastable phase at 95 ◦C to be an ordered chiral smectic (SmX1\*) phase which is probably either chiral smectic F or chiral smectic I phase [57–60]. The XRD profile of the more stable highly ordered chiral smectic (SmX2\*) phase at 27 ◦C differed from those of the SmC\* and SmX1\* phases (Figure 6a, lower). The increase in the intralayer molecular order is indicated by a broad peak observed at 2*θ* = 18.64◦ for the (100) plane and by an increase in the relative peak intensity of the (010) diffraction. The shorter layer spacing of 29 Å and the calculated molecular length indicate that the tilted-layer structure is maintained even in the ordered smectic phase. This ordered smectic (SmX2\*) phase should be one of the chiral smectic G, chiral smectic J or chiral smectic H phase, as determined by the general phase transition sequence [57–61].

The XRD pattern of the LC phase of (*R*)-**2** at 124 ◦C (Figure 6b, upper) exhibits several peaks at 2*θ* = 2.16◦, 4.36◦, and 6.49◦. These three peaks are attributed to the (001), (002), and (003) planes with diffractions derived from the periodicity of the smectic-layer structure. Although the extended molecular length of (*R*)-**2** is estimated to be 38 Å by MM2 calculations, the experimentally obtained layer spacing is 41 Å. The layer spacing is greater than the theoretical molecular length that an interdigitated layer structure can achieve. The halos observed at approximately 2*θ* = 12◦, 20◦, 24◦, and 26◦ also confirm the interdigitated organization of (*R*)-**2** molecules. Because the POM textures of (*R*)-**2** at a comparable temperature are typical for a low-ordered smectic phase and not for a characteristic texture for a highly ordered smectic phase, the formation of a highly ordered smectic phase is uncertain. In addition, the sample of (*R*)-**2** shows fluidity in the LC phase. These behaviors can be observed in a low-ordered interdigitated smectic phase. While the halo at 2*θ* = 12◦ can be ascribed to the disordered aggregation of bulky chiral alkyl chains based on steric effects, the series of halos between 2*θ* = 18◦ and 2*θ* = 30◦ probably resulted from the disordered aggregation of the linear aliphatic chains and interaction between aromatic units. Thus, the appearance of several halos suggests that each of the rigid aromatic units and mobile chiral alkyl chains is segregated and gathered in different periodicities. The integrated molecules of (*R*)-**2** in the LC phase should be tilted with respect to the layer normal, considering the molecular packing model (Figure 7). Therefore, upon cooling, we assigned the LC phase of (*R*)-**2** between 126 ◦C and 112 ◦C to an interdigitated chiral smectic C (SmC*d*\*) phase [62–67]. When the XRD sample of (*R*)-**2** was cooled to 108 ◦C, the normalized intensities of the (002) and (003) diffraction peaks in the XRD profile (Figure 6b, middle) were higher than those of the same peaks in the XRD pattern of the

SmC*d*\* phase (Figure 6b, upper). The absence of sharp diffraction peaks in the wide-angle region indicates the absence of long-range intralayer order in the middle-temperature LC phase. Because the interlayer spacing undergoes a slight change of 41–42 Å via the SmC*d*\*–LC phase transition, the interdigitated layer structure is maintained. Based on these results, we believe that the middle-temperature LC phase is a chiral smectic (SmX*d*1\*) phase, in which the interdigitated LC structures have short-range intralayer order. In the XRD pattern of (*R*)-**2** cooled to room temperature (34 ◦C), additional weak diffraction peaks were observed at 2*θ* = 10.6◦ and 19.4◦ (Figure 6b, bottom). These peaks originated from the (005) and (010) diffraction planes. Because the sharp (010) diffraction peak indicates growing intralayer-bond order, the room-temperature LC phase is identified as a highly ordered interdigitated chiral smectic (SmX*d*2\*) phase. The small difference of XRD patterns between SmX*d*1\* phase and SmX*d*2\* phase suggests the slight structural change through phase transitions via the metastable state. The POM study of (*R*)-**2** on cooling process (Figure 3b–d) also supports this consideration. The metastable phase of (*R*)-**2** probably appears while the rearrangement of intralayer molecular packing proceeds under the influence of spatial factors by bulky chiral unit and sparse ethynyl moiety.

**Figure 6.** Variable-temperature XRD profiles of (**a**) (*R*)-**1**, (**b**) (*R*)-**2**, and (**c**) (*R*)-**3**.

**Figure 7.** Schematic illustrations of the molecular packing models in the (**a**) SmC\* phase of (*R*)-**1** (monolayer structure), and (**b**) SmC*d*\* phase of (*R*)-**2** (interdigitated layer structure) assumed from the XRD profiles.

Compound (*R*)-**3** exhibited complicated XRD patterns, indicating crystalline molecular packing below 74 ◦C (Figure 6c). Although several phase transition peaks were observed in the DSC thermogram, bis({fluorophenyl}ethynyl) bithiophene (*R*)-**3** showed crystal polymorphism and did not show any thermodynamically stable LC phase.

The phase transition behaviors of (*R*)-**1**, (*R*)-**2**, and (*R*)-**3** are summarized in Table 1. For each of LC compounds (*R*)-**1** and (*R*)-**2**, the initial crystalline precipitates for the characterization of LC properties were obtained by recrystallization. It is noted that both LC compounds (*R*)-**1** and (*R*)-**2** exhibit the crystalline–LC phase transition only in the first heating process. Once the precipitates melted to the IL phase, no crystallization occurred during the cooling process below −<sup>50</sup> ◦C at a scanning rate of 10 ◦C min−1. These results show that a bilateral asymmetric molecular structure is effective for liquid crystallinity. In addition, we consider that the interplay of bulky chiral unit and sparse ethynyl moiety prominently causes a variety of molecular packing as well as the formation of interdigitated structures.


**Table 1.** Phase transition behavior of (*R*)-**1**, (*R*)-**2**, and (*R*)-**3**.

The abbreviations Cr, Cr1, Cr2, IL, M, SmC\*, SmC*d*\*, SmX1\*, SmX2\*, SmX*d*1\*, SmX*d*2\*, and SmX*m*\* denote crystalline, crystalline 1, crystalline 2, isotropic liquid, unidentified ordered, chiral smectic C, chiral interdigitated smectic C, unidentified ordered chiral smectic 1, unidentified ordered chiral smectic 2, unidentified interdigitated ordered smectic 1, interdigitated ordered smectic 2, and metastable interdigitated ordered smectic phases, respectively.

### *3.2. Spectroscopic Properties*

Figure 8a shows the UV–vis absorption and PL spectra in a dilute THF solution of (*R*)-**1**, (*R*)-**2**, and (*R*)-**3**. The absorption spectrum of (*R*)-**1** in THF (10 μM) showed a quasiunimodal absorption band corresponding to the π–π\* transition of the terthiophene unit between 330 and 450 nm. The absorption maximum was 393 nm with a molar absorption coefficient of 4.0 × 104 L mol−<sup>1</sup> cm−1. By comparing the absorption spectra of (*R*)-**<sup>1</sup>** and

(*R*)-**2** in THF dilute solutions, a slight shift in the absorption band of (*R*)-**2** is observed towards the longer wavelength region. This result suggested that introducing an ethynyl linker to the mesogenic core accurately extended the effective π-conjugation length. In addition, the THF solution of compound (*R*)-**2** showed a higher molar absorption coefficient of 4.7 × 104 L mol−<sup>1</sup> cm−<sup>1</sup> at the absorption maximum (*λ*abs = 397 nm) compared to those of the solution of (*R*)-**1**. The absorption spectrum of (*R*)-**3** in THF (10 μM) displays the π–π \* transition band of the bithiophene core with an absorption maximum of 390 nm. For a dilute solution of (*R*)-**3**, the molar absorption coefficient attained 5.5 × 104 L mol−<sup>1</sup> cm−1. The absorption edges in the THF solutions of (*R*)-**1**, (*R*)-**2,** and (*R*)-**3** are 446, 450, and 455 nm, respectively (Figure 8b). Because the order of the absorption maxima and edges reflects the π-conjugation length, (*R*)-**2** should have the longest effective π-conjugation length among the three compounds. Each fluorescence spectrum indicates well-resolved vibrational structures (Figure 8a). The maximum PL intensity of (*R*)-**3** in the THF solution was more than twice those of (*R*)-**1** and (*R*)-**2**. This result indicates that modifying phenylethynyl units on both wings of the 2,2'-bithiophene core enhances fluorescence emission. The luminescence enhancement appears to result from the suppression of thermal relaxation, and an increase in the oscillator strength is observed. In the case of compound (*R*)-**3**, introducing an ethynyl linker may reduce steric interactions and extend the π-conjugation length [68]. Table 2 lists the spectroscopic parameters. The disubstituted compound (*R*)-**3** showed the slightest Stokes shift among the three compounds, and phenylethynyl terthiophene (*R*)-**2** displayed the most significant Stokes shift. According to previous reports, we considered that the difference in Stokes shift originated from conformational changes rather than solvent effects [68]. The difference in Stokes shift implies a variation in the conformational change between the ground and excited states. The modification of the chiral phenylethynyl units showed different effects on excitation and emission in (*R*)-**3** and (*R*)-**2**. Because interorbital electronic interactions are sensitive to modifications, the principal cause of the difference in spectroscopic properties of (*R*)-**2** and (*R*)-**3** seems to be left-right asymmetrical hetero-substitution. The lack of a drastic increase in the PL intensity of (*R*)-**2** also supports this hypothesis.


**Table 2.** Spectroscopic properties of (*R*)-**1**, (*R*)-**2**, and (*R*)-**3** in a dilute THF solution (10 μM).

(a) Excitation wavelength of 393 nm. (b) Excitation wavelength: 397 nm. (c) The excitation wavelength was 390 nm.

**Figure 8.** (**a**) UV–vis absorption and photoluminescent spectra in THF dilute solution (10 μM) of (*R*)-**1**, (*R*)-**2**, and (*R*)-**3**. (**b**) The magnified UV–vis absorption spectra in (**a**). Each triangle marks in the inset of absorption spectra depicting the absorption edge of (*R*)-**1** (blue), (*R*)-**2** (red), and (*R*)-**3** (green).

### **4. Conclusions**

We synthesized three chiral oligo-thiophene derivatives, (*R*)-**1**, (*R*)-**2**, and (*R*)-**3**. While compounds (*R*)-**1** and (*R*)-**2** exhibited chiral smectic LC phases, the disubstituted bithiophene analog (*R*)-**3** showed only crystalline polymorphism. This outcome suggests that the bilateral symmetry hinders liquid-crystallinity. In other words, molecular structuring with left-right asymmetry promotes the formation of LC structures. The ethynyl-inserted monosubstituted compound (*R*)-**2** forms LC interdigitated layer structures due to the steric effect of the spatially sparse ethynyl linker and the bulky chiral moiety. In addition, the interplay of both units could effectively contribute to the formation of various smectic LC structures. The UV–vis absorption and PL spectra in a dilute THF solution indicate that (*R*)-**2** has a more expansive π-conjugation system than (*R*)-**1** because of the introduction of the ethynyl linker. The ethynyl linker also contributed to an increase in the molar absorption coefficient in the visible region. Because the molecular packing and photophysical properties affect the photoconductive properties, introducing an ethynyl linker in the central π-conjugated core causes drastic changes in the performance of organic optoelectronic devices. We believe our results can help in the molecular design of novel functional chiral π-conjugated liquid crystals, including ferroelectric π-conjugated liquid crystals that exhibit the FePV effect.

**Supplementary Materials:** The following supporting information can be downloaded from https: //www.mdpi.com/article/10.3390/cryst12091278/s1, Section S1. Synthetic procedure; Section S2. 1H and 13C NMR spectra (Figure S1. 1H NMR spectrum of (*R*)-**1**; Figure S2. 13C NMR spectrum of (*R*)-**1**; Figure S3. 1H NMR spectrum of (*R*)-**2**; Figure S4. 13C NMR spectrum of (*R*)-**2**; Figure S5. 1H NMR spectrum of (*R*)-**3**; Figure S6. 13C NMR spectrum of (*R*)-**3**); Section S3. High-resolution electrospray ionization (ESI) mass spectra (Figure S7. High-resolution ESI mass spectrum of (*R*)-**1**; Figure S8. High-resolution ESI mass spectrum of (*R*)-**2**; Figure S9. The high-resolution ESI mass spectrum of (*R*)-**3**).

**Author Contributions:** Conceptualization, A.S.; methodology, A.S.; validation, A.S., K.S. and K.A.; formal analysis, A.S. and K.S.; investigation, A.S. and K.S.; resources, A.S. and K.A.; data curation, A.S. and K.S.; writing—original draft preparation, A.S.; writing—review and editing, A.S. and K.S.; visualization, A.S. and K.S.; supervision, A.S. and K.A.; project administration, A.S.; funding acquisition, A.S. and K.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was financially supported by a research fund from the Tokyo University of Science for A.S. and K.A. and a research grant from the Amano Institute of Technology, Japan for A.S.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors thank Khoa V. Le and T. Sasaki at the Tokyo University of Science for their help with POM observations. We also greatly appreciate the help of HRMS measurements by Y. Yoshimura at the Tokyo University of Science.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


## *Article* **Chromonic Ionic Liquid Crystals Forming Nematic and Hexagonal Columnar Phases**

**Takahiro Ichikawa \*, Mei Kuwana and Kaori Suda**

Department of Biotechnology, Tokyo University of Agriculture and Technology, Nakacho, Koganei, Tokyo 184-8588, Japan

**\*** Correspondence: t-ichi@cc.tuat.ac.jp

**Abstract:** We designed an ionic salt by combining a π-conjugated anion and a cholinium cation. It formed homogeneous mixtures with water in various weight ratios. The obtained mixtures showed chromonic liquid-crystalline behavior in a wider concentration range as compared to analogous compounds with inorganic cations. Although only an exhibition of nematic phases was previously reported by Kasianova et al. for analogous compounds with an inorganic cation in 2010, the ionic salt with a cholinium cation showed not only nematic phases but also hexagonal columnar phases. The formation of hexagonal columnar phases is attributed to its ability to form mesophases even in a high concentration range, which enables the cylindrical aggregates of the π-conjugated anions to form dense packing. By examining the states of the water molecules, we revealed that the ability of the cholinium cation to form a hydrated ionic liquid state strongly contributes to the widening of the concentration range forming chromonic liquid-crystalline behavior.

**Keywords:** chromonic liquid crystal; ionic liquid; nematic; hexagonal columnar phase

### **1. Introduction**

Chromonic liquid crystals are a class of lyotropic liquid crystals. A unique point of chromonic liquid crystals is that they have molecular structures composed of a polycyclic aromatic core having several ionic and/or hydrophilic groups at its periphery [1–8]. It has been generally understood that the aromatic core plays a key role for the formation of self-assembled cylindrical aggregates through π-π interactions and/or other interactions. The hydrophilic groups are important for solubility into water. To date, there have been many reports of ionic compounds exhibiting chromonic liquid-crystalline (LC) behavior. Most of them are composed of π-conjugated mesogens with anionic groups and inorganic cations [1–6] while, in some case, chromonic liquid crystals composed of π-conjugated cations and inorganic anions have been also reported [7,8].

On the other hand, in the several decades of studies, there have been a growing interest on the use of organic cations for creating functional ionic compounds, such as ionic liquids [9–11], ionic liquid crystals [12–15], ionic plastic crystals [16–18], and ionic crystals [19]. In the course of studies on ionic liquids, it has been revealed that there is a potential that a slight difference of the organic cation structures results in the large difference of physicochemical properties and functions. For example, imidazolium cations are recognized as one of the most suitable cations for designing ionic liquids dissolving cellulose [20,21]. On the other hands, the use of cholinium cation has attracted an increasing attention for yielding hydrated ionic liquids for dissolving some bio-functional polymers [22]. For example, Fujita and Ohno reported that hydrated ionic liquids have a great potential as liquid media for enzyme storage [23–25]. One of the important characteristics of hydrated ionic liquids is that they maintain liquid states even in quite high concentration conditions. This characteristic leads us to envision that the employment of suitable organic cations would be one of an advanced strategy for controlling the chromonic LC behavior of π-conjugated compounds with anionic groups.

**Citation:** Ichikawa, T.; Kuwana, M.; Suda, K. Chromonic Ionic Liquid Crystals Forming Nematic and Hexagonal Columnar Phases. *Crystals* **2022**, *12*, 1548. https://doi.org/10.3390/ cryst12111548

Academic Editor: Alberta Ferrarini

Received: 12 October 2022 Accepted: 28 October 2022 Published: 29 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

As an anion with π-conjugated structure, we have employed 4,4-(5,5-dioxidodibenzo[b, d]thiene-3,7-diyl)dibenzenesulphonic acid (**pQpdS**) anion. It is an anion whose Cs salt was reported to exhibit chromonic LC behavior at a water content of 85 wt% by Kasianova et al. in 2010 [26]. As an organic cation, a cholinium (Ch) cation has been selected. By combining these cation and anion, we have synthesized an organic salt, **pQpdS-Ch** (Figure 1). Its chromonic LC behavior in water has been examined using polarized optical microscopy (POM) observation, differential scanning calorimetry (DSC), and X-ray diffraction (XRD) measurements.

**Figure 1.** Molecular structure of **pQpdS-Ch**.

### **2. Materials and Methods**

The synthesis scheme of **pQpdS-H** is shown in Scheme 1. To an aqueous solution of choline hydroxide, an equimolar amount of 4,4-(5,5-dioxidodibenzo[b,d]thiene-3,7 diyl)dibenzenesulphonic acid (**pQpdS-H**) was added. The solution was stirred until the white solid of **pQpdS-H** dissolved into the solution. Evaporation of water yielded a **pQpdS-Ch** as a white solid.

**Scheme 1.** Synthesis of **pQpdS-Ch**.

1H NMR (400MHz, D2O): *δ* = 7.78 (s, 2H), 7.64 (d, *J* = 8.4 Hz, 4H), 7.41–7.35 (m, 8H), 3.92–3.89 (m, 4H), 3.37–3.35 (m, 4H), 3.05 (s, 18H).

### **3. Results and Discussion**

Mixtures of **pQpdS-Ch** and H2O in 100–X:X weight ratios (X = 90, 80, 70, 60, and 50) were prepared by adding two components into Eppendorf tubes. In order to increase the homogeneity of the mixtures, the tubes were vibrated and centrifugation was performed. We could obtain the homogeneous mixtures when 90 ≥ X ≥ 50 while it was not obtained when X ≤ 40. Small amounts of the homogeneous mixtures were put on a slide glass and covered with a cover glass. Polarizing optical microscope (POM) observation was carried out for them while cooling from isotropic phases observed at around 80 ◦C. The obtained textures are shown in Figure 2. It has been found that the samples with X ≥ 90 shows no birefringence in the temperature range, indicating that mesomorphic behavior is not induced when X ≥ 90. On the other hand, the mixture with X = 80 shows a schlieren texture, which is a characteristic of nematic phases. This behavior is similar to that reported by Kasianova et al. for **pQpdS** with Cs cation [26]. A notable difference has been observed when X ≤ 70. These mixtures show focal conic fan-textures, which are indicative of the formation of columnar LC phases. The thermotropic phase transition behavior of the mixtures is summarized in a bar graph (Figure 3). The formation of the nematic phases results from the cylinder aggregation of the **pQpdS** anions and the subsequent axial alignment of the cylinders. That of the columnar phases can be attributed to the formation of the positional order of the cylinders as the decrease of the inter-cylinder distance.

**Figure 2.** Polarized optical microscopic images of the mixtures of **pQpdS-Ch** and H2O in the 100–X:X weight ratios. (**a**) X = 90 at 25 ◦C, (**b**) X = 80 at 10 ◦C, (**c**) X = 70 at 25 ◦C, (**d**) X = 60 at 25 ◦C, and (**e**) X = 50 at 25 ◦C.

**Figure 3.** Bar graph of the thermotropic phase transition behavior of the mixtures of **pQpdS-Ch** and H2O in the 100–X:X weight ratios on cooling.

In order to confirm the formation of columnar phases, we have performed XRD measurements for the mixtures at 30 ◦C. For avoiding the evaporation of water from the sample, the LC samples were put on an aluminium pan and rapidly covered by a polymer film (DURA SEAL, DIVERSIFIED BIOTECH). A XRD pattern observed for the mixture (X = 60) is shown in Figure 4. An intense peak and two weak peaks were found in the small angle region. The *d*-values estimated from the peak position *θ* values are 32.7, 18.7, and 16.0 Å, respectively. These peaks can be indexed as (100), (110), and (200) reflections of a hexagonal structure, which lead us to identify the columnar phase as a hexagonal columnar (Colh) phase. The intercolumnar distances in the Colh LC phase can be calculated to be 37.8 Å from the *d*-values. The formation of Colh phases for bent shaped chromonic liquid crystals has been also reported by Wang et al. in 2018 [27], which supports our characterization.

**Figure 4.** X-ray diffraction (XRD) pattern of **pQpdS-Ch**/H2O (X = 60) weight ratio at 30 ◦C.

POM observation for macroscopically aligned samples is a useful strategy for deducing the molecular assembled structures in LC states. In order to employ this strategy for the present materials, we sandwiched a small amount of a **pQpdS-Ch**/H2O (X = 60) mixture between a cover glass and a slide glass and then added a mechanical shearing to the cover glass. It is a technique to align the column axis to the shearing direction [28]. As expected, the sheared sample show a homogeneous texture under POM observation, which is indicative of the formation of 1D-aligned columnar phases. The aligned samples were observed under POM with a 530 nm retardation plate inserted in the optical path at 45 degrees. The shearing direction is set parallel and perpendicular to the slow axis direction of the retardation plate. It has been found that, when these two directions are parallel, the texture is observed in a yellow (Figure 5a). It turns into in a blue as the rotation of the sample through 90 degrees (Figure 5b). These results mean that the slow axis of the Colh liquid crystals is perpendicular to the column axis, namely the **pQpdS-Ch**/H2O (X = 60) mixture has a negative birefringence. It is consistent with the results obtained for analogous compound with the Cs cation in a nematic phase that was reported by Kasianova et al. [26].

**Figure 5.** Polarized optical microscopic images of the mixtures of **pQpdS-Ch**/H2O (X = 60) in the Colh phase after shearing. (**a**) The shearing direction is parallel to the slow axis of the retardation plate. (**b**) The shearing direction is perpendicular to the slow axis of the retardation plate.

In order to further confirm the molecular assembled structure of the **pQpdS-Ch**/H2O mixtures in the Colh phase, we have performed polarized IR measurements. A 1D-aligned sample of the **pQpdS-Ch**/H2O (X = 70) mixture was prepared by the same method. IR absorbance was measured with setting the angle of the polarizer (*θ*p) in the range from 0 to 180 degrees. The Colh LC sample was set in such a way that its column axis was parallel to *θ*<sup>p</sup> = 0 degree. While the S=O stretching vibration (*ν*S=O) of the **pQpdS** molecules was observed at 1301 cm–1 independent of the *θ*<sup>p</sup> angles, the peak strength of *ν*S=O clearly depended on the *θ*<sup>p</sup> angles. For example, the absorbance of *ν*S=O was 0.20 when *θ*<sup>p</sup> = 0 degree, which increases as the increase of *θ*<sup>p</sup> (Figure 6a). For further clarify *θ*p-dependence of the *ν*S=O absorbance, we have constructed a polar plot (Figure 6b). It indicates that the sulfonyl groups of the **pQpdS** molecules are oriented perpendicular to the 1D column axis.

**Figure 6.** (**a**) *θ*p-dependence of IR spectra of the **pQpdS-Ch**/H2O (X = 70) mixture in the Colh phase. (**b**) A polar plot of the absorbance of the S=O stretching vibration (*ν*S=O) observed at 1301 cm–1.

The phase transition behavior of these mixtures has been further examined using DSC measurements. The DSC measurements were performed in the temperature range from 0 to 80 ◦C at the heating/cooling rate of 10 ◦C min<sup>−</sup>1. The obtained DSC charts on cooling and heating are shown in Figure 7a,b, respectively. In the cooling process, an exothermic peak is found for each sample when X ≤ 80. These peaks can be attributed to the enthalpy change at the phase transition from an isotropic phase to an LC phase. It can be seen that the peak position shifts to higher temperature region as the decrease of X, which is consistent with the isotropization temperatures observed by POM observation. The thermal stabilization

of the mesophases upon the decrease of X can be explained by the increase of the packing density of the cylinder aggregates. Another notable trend is that the peak area increases as the decrease of X. For example, the peak area of the phase transition from the Colh to Iso phases is 0.83 mJ/mg for the **pQpdS-Ch**/H2O (X = 70) mixture while it increases to 2.16 mJ/mg for **pQpdS-Ch**/H2O (X = 50). These enthalpy changes can be mainly ascribed to the cleavage of the dipole–dipole interaction between the sulfonyl groups. Rough calculation is described in the Supplemental Information.

**Figure 7.** DSC thermograms of the **pQpdS-Ch**/H2O mixtures: (**a**) on cooling and (**b**) on heating.

It is considered that the water molecules in the mixtures exist as bound water and/or free water. With an aim to investigate the states of water in the mixtures, we have performed DSC measurements in lower temperature region. The DSC measurements were carried out from room temperature to –80 ◦C. A peak corresponding to crystallization of free water was found at a temperature lower than 0 ◦C. By estimating the amount of free water in the mixtures from the peak area, the molar ratios of bound water and free water in the mixtures (X = 50–90) were investigated and summarized in Table 1. It has been found that 15–25 water molecules strongly interact with a **pQpdS-Ch** molecule and then exist as bound water. These results are consistent with the number of hydration water molecules reported for cholinium-based hydrated ionic liquids [29]. We assume that the cylindrical aggregates formed by the **pQpdS** anions are surrounded by sheath of hydrated ionic liquids that produce liquidity and prevent crystallization even in the water poor condition (70 ≥ X ≥ 50).


**Table 1.** Weight ratios and molar ratios of **pQpdS-Ch**, free water, and bound water.

Based on the results of POM observation, DSC, and XRD measurements, here we discuss the molecular assembled structure of the **pQpdS-Ch**/H2O mixtures. For assuming the molecular assembled structures, an important characteristic of **pQpdS-Ch** is that it has a strong dipole moment at the sulfonyl group, which can be calculated

to be 5.2 D by DFT calculation (Figure S1) (see supplementary materials). Therefore, it is expected that it forms a dimer in the dissolved state as well as in the assembled states in water. The size of the **pQpdS-Ch** anion is about 20 Å. Based on these results, here we imagine a molecular assembled structure of the **pQpdS-Ch**/H2O (X = 60) mixture in the Colh phase. The number of bound water per the dimer of the **pQpdS-Ch** molecules calculated from the endothermic peaks in the DSC charts is 20×2 = 40. The inter columnar distance is calculated to be 37.8 Å as explained in the above paragraph. Considering these data and the component weight ratio, the molecular assembled structures of the **pQpdS-Ch**/H2O (X = 60) mixture in the Colh phase is drawn as shown in Figure 8.

**Figure 8.** A schematic image of the molecular assembled structure of the **pQpdS-Ch**/H2O (X = 60) mixture in the Colh phase.

In order to further confirm the effects of the cation species, we have also prepared analogous compounds with other inorganic cations, such as Li, Na, and K cations. **pQpdS-Y** (Y = Li, Na, and K) were prepared according to the same procedure used for **pQpdS-Ch**. They were obtained as white or slightly yellowish white compounds (Figure S2). The mixtures of these compounds and water were prepared with varying the component ratios and their phase transition behavior was examined by POM observation.

It has been found that the exhibition of N phases was observed for the **pQpdS-Li/**H2O mixtures when the water content value is 90 ≥ X ≥ 85 and that of Colh phases was observed when X = 80 (Figure 9). The water content dependence of the mesophase pattern is similar to that of the **pQpdS-Ch/**H2O mixtures. These results indicate that the formation of Colh is a phenomena that is observed not solely for **pQpdS-X** with organic cations but also for **pQpdS-X** with inorganic.

**Figure 9.** Polarized optical microscopic images of the **pQpdS-Li**/H2O mixtures in the 100–X:X weight ratios. (**a**) X = 95 at 25 ◦C, (**b**) X = 90 at 20 ◦C, (**c**) X = 85 at 25 ◦C, (**d**) X = 80 at 50 ◦C, and (**e**) X = 70 at 25 ◦C.

On the other hand, we have found that the **pQpdS-Na/**H2O mixtures forms only N phases when 95 ≥ X ≥ 80 (Figure 10) and those with 70 ≥ X form crystalline states. Comparing the water content range forming mesophases for the **pQpdS-X/**H2O mixtures (Figures 3 and 11), it can be seen that the employment of the cholinium cation provides chromonic liquid crystals showing LC behavior in the widest water content range. It is attributed to the higher solubility of **pQpdS-Ch** into water that results from its lower crystallinity than those with inorganic cations. Namely, the employment of the cholinium cation increases the conformational degrees of freedom, which contributes to the inhibition of the crystallization. The melting point of **pQpdS-Ch** is higher than 100 ◦C (Figure S3) that is the important temperature of the definition of ionic liquids. However, considering the recent studies on ionic liquids where hydrated organic salts are called hydrated ionic liquids [23–25,30], we expect that **pQpdS-Ch**/H2O mixture can be regarded as hydrated ionic liquids exhibiting chromonic LC behavior.

In the course of studies on ionic liquids, they have been used in a wide range of fields, including electrochemistry, analysis, catalysis, and solvents. Focusing on hydrated ionic liquids, they have been expected as potential solvent for biomolecules [23,31]. On the other hand, chromonic liquid crystals have been investigated as sensors [32] and optical materials [33]. We believe that the present material design will attract attention in a wide field of research ranging from biotechnology to material chemistry.

**Figure 10.** Polarized optical microscopic images of the **pQpdS-Na/**H2O mixtures in the 100–X:X weight ratios. (**a**) X = 95 at 25 ◦C, (**b**) X = 90 at 25 ◦C, (**c**) X = 85 at 25 ◦C, and (**d**) X = 80 at 25 ◦C.

**Figure 11.** Bar graphs of the thermotropic phase transition behavior of; (**a**) the **pQpdS-Li/**H2O mixtures on cooling; (**b**) the **pQpdS-Na**/H2O mixtures on cooling.

### **4. Conclusions**

We have succeeded in preparing a new class of an organic salt **pQpdS-Ch** showing chromonic liquid-crystalline (LC) behavior. This compound is composed of a rod-shaped aromatic anion with a strong dipole moment and cholinium cations. Both of two components owes their specific tasks. The former plays a key role for the formation of cylindrical aggregates via dipole–dipole interactions. The latter plays an important role for the formation of hydrated states. Moreover, since the hydrated cholinium cation has a larger positional and conformational degrees of freedom than inorganic cations, it results in the inhibition of the crystallization of the cylindrical aggregates. This effect enables to form chromonic LC mesophases even in a wider concentration range than a series of analogous compounds with inorganic cations, which leads to the exhibition of chromonic hexagonal columnar phases.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cryst12111548/s1, Figure S1: A schematic image of the molecular assembled structure. Figure S2: Pictures of the synthesized compounds; Figure S3: TG/DTA result.

**Author Contributions:** T.I., M.K. and K.S. conceived and designed the experiments; T.I., M.K. and K.S. performed the experiments and analyzed the data; T.I., M.K. and K.S. wrote the paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by JSPS KAKENHI numbers JP21H02010, and JP22H04526 from the Japan Society for the Promotion of Science.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** This work was supported by JSPS KAKENHI numbers JP21H02010, and JP22H04526 from the Japan Society for the Promotion of Science. This work was partly supported by the financial support from FUJIFILM.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Yuki Arakawa \*, Yuto Arai, Kyohei Horita, Kenta Komatsu and Hideto Tsuji**

Department of Applied Chemistry and Life Science, Graduate School of Engineering, Toyohashi University of Technology, 1-1 Hibarigaoka, Tempaku-cho, Toyohashi 441-8580, Japan

**\*** Correspondence: arakawa@tut.jp

**Abstract:** The twist–bend nematic (NTB) phase is a liquid crystal (LC) phase with a heliconical structure that typically forms below the temperature of the conventional nematic (N) phase. By contrast, the direct transition between the NTB and isotropic (Iso) phases without the intermediation of the N phase rarely occurs. Herein, we demonstrate the effects of linkage type (i.e., methylene, ether, and thioether) on the typical Iso–N–NTB and rare direct Iso–NTB phase-transition behaviors of cyanobiphenyl (CB) dimers CB3CB, CB2OCB, and CB2SCB bearing three-atom-based propane, ethoxy, and ethylthio spacers, respectively. In our previous study, CB2SCB exhibited the monotropic direct Iso–NTB phase transition. In this study, we report that CB3CB also shows the direct Iso–NTB phase transition, whereas CB2OCB exhibits the typical Iso–N–NTB phase sequence with decreasing temperature. The Iso–LC (Iso–NTB or Iso–N) phase-transition temperatures upon cooling show the order CB2OCB (108 ◦C) > CB3CB (49 ◦C) > CB2SCB (43 ◦C). The thioether-linked CB2SCB is vitrifiable, whereas CB3CB and CB2OCB exhibit strong crystallization tendencies. The phasetransition behaviors are also discussed in terms of the three bent homologous series with different oligomethylene spacers *n*: CB*n*CB, CB*n*OCB, and CB*n*SCB.

**Keywords:** twist–bend nematic phase; direct twist–bend nematic phase transition; liquid crystal dimer; cyanobiphenyl dimer; short spacer

### **1. Introduction**

Liquid crystal (LC) phases are mesophases between anisotropic crystals and isotropic liquids. A nematic (N) phase generally does not have an apparent layered structure and is recognized as the most fluid LC phase. After the heliconical twist–bend nematic (NTB) phase was predicted [1–3], it was experimentally assigned to unknown N (NX) phases below the conventional N phase of bent molecules (dimers) in the last decade [4,5]. The NTB phase possesses a heliconical director precession with a pitch of approximately 10 nm [6–8]. The heliconical structures of the NTB phase result in optical textures and physical properties like those of layered smectic (Sm) phases rather than the conventional N phase [9–13]. Therefore, the NTB phase is often considered a pseudo-layered phase. However, the phase identification of the NTB phase for the NX phase is still under discussion and further study is required to elucidate the detailed structure owing to its elusive nature. Alternatively, a polar twisted nematic (NPT) phase model for the NX phase instead of the NTB phase has been proposed [14,15]. In this paper, the widely recognized term, NTB phase, is used.

The NTB phase can be formed only by bent molecules, such as bent dimers [4,5,16–39], linear oligomers (e.g., trimers, tetramers, and hexamers) [40–49], duplexed hexamers [50], polymers [51], hydrogen-bonded dimers and trimers [52,53] with odd-number atom spacers, and bent-core molecules [54–56]. Mandle reviewed the structure–property relationship of the NTB phase and summarized the recent progress in this topic [57]. Theoretical simulation studies have examined the relationship between the curvature of various bent dimers and the incidence of the NTB phase [58–61]. In nearly all cases of the reported bent

**Citation:** Arakawa, Y.; Arai, Y.; Horita, K.; Komatsu, K.; Tsuji, H. Twist–Bend Nematic Phase Behavior of Cyanobiphenyl-Based Dimers with Propane, Ethoxy, and Ethylthio Spacers. *Crystals* **2022**, *12*, 1734. https://doi.org/10.3390/cryst12121734

Academic Editors: Shigeyuki Yamada, Kyosuke Isoda, Takahiro Ichikawa, Kosuke Kaneko, Mizuho Kondo, Tsuneaki Sakurai, Atsushi Seki, Mitsuo Hara and Go Watanabe

Received: 26 October 2022 Accepted: 20 November 2022 Published: 1 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

molecules, the NTB phase continuously formed at a temperature below the temperature of the conventional N phase, resulting in a typical isotropic (Iso)–N–NTB phase sequence with decreasing temperature. For example, the homologous series of bent symmetric methyleneand asymmetric methylene-/ether- and methylene-/thioether-linked cyanobiphenyl (CB) dimers with *n* number of carbon atoms in the oligomethylene spacers, CB*n*CB (*n* = 5, 7, 9, 11, and 13) [17,25], CB*n*OCB (*n* = 4, 6, 8, and 10) [24,25], and CB*n*SCB (*n* = 4, 6, 8, and 10) [62], respectively, are known to exhibit the typical Iso–N–NTB phase sequence (Figure 1a).

**Figure 1.** Bent dimer systems exhibiting the typical Iso–N–NTB and rare direct Iso–NTB phase transitions investigated in previous studies and the present study. (**a**) Previous work on bent-shaped CB dimer homologs with longer spacers, including methylene-linked CB*n*CB (odd *n* = 5, 7, 9, 11, and 13) [17,25], methylene-/ether-linked CB*n*OCB (even *n* = 4, 6, 8, and 10) [24,25], and methylene- /thioether-linked CB*n*SCB (*n* = 4, 6, 8, and 10) [62] that exhibit the typical Iso–N–NTB phase sequence. (**b**) Previous work on dimers showing the rare Iso–NTB phase sequence from binary mixtures reported by Archbold et al. [63] (**right**) and single-component dimers reported by Dawood et al. [64,65] (iminelinked dimers, **top left**), our group [62] (ethylthio-linked CB2SCB, **center left**), and Wang et al. [66] (a phosphine-bridged dimer, **bottom left**). (**c**) Single-component dimers CB3CB and CB2OCB in the present study.

However, in a few cases, an NTB phase can be directly formed from the Iso phase without the intermediate N phase. Archbold et al. reported that binary mixtures of a dimer with the Iso–N–NTB phase sequence and a chiral dopant (6–10 wt%) exhibit the direct Iso– NTB phase transition (Figure 1b, right) [63]. Dawood et al. reported that two imine-linked dimers with a central propane spacer and terminal methoxy or ethoxy groups (*m =* 1 and 2, respectively) (Figure 1b, top left) show the direct Iso–NTB phase transition [64,65]. We reported that bent CB*n*SCB (*n* = 4, 6, 8, and 10) demonstrates the Iso–N–NTB phase sequence, as described earlier, whereas only the shortest ethylthio-linked CB2SCB exhibits the direct Iso–NTB phase transition upon cooling (Figure 1b, center left) [62]. Shortening the flexible

spacer *n* of the CB*n*SCB dimers lowers the Iso–N phase-transition temperature (*T*IN) upon cooling, thereby narrowing the intermediate N-phase temperature range (Δ*T*N), which leads to the direct Iso–NTB phase transition of CB2SCB. Moreover, Wang et al. recently disclosed that a phosphorus-bridged LC dimer exhibits the direct Iso–NTB phase transition (Figure 1b, bottom left) [66].

Considering that, in our previous study, CB2SCB demonstrated the rare direct Iso– NTB phase transition [62], propane-linked CB3CB and ethoxy-linked CB2OCB bearing the same three-atom-based spacers (as shown in Figure 1c) are worthy of further investigation whether they exhibit the rare Iso–NTB phase or not. The influence of the different linkage types on the phase transition of such short-spacer dimers has yet to be reported. In this study, we investigated the phase-transition behaviors of three CB-based dimers with different three-atom-based spacers, namely CB3CB, CB2OCB, and CB2SCB. The phase-transition behavior of the previously reported CB3CB that did not show the LC phase [67] was reinvestigated, while that of CB2OCB was explored for the first time. The phase-transition data of CB2SCB were obtained from our previous study [62]. We then compared the phasetransition behaviors of three series of CB*n*CB, CB*n*OCB, and CB*n*SCB homologs. Finally, the effects of linkage-type on the NTB phase-transition behavior of dimers with short spacers, particularly on the occurrence of the direct Iso–NTB phase transition, were explored.

### **2. Materials and Methods**

### *2.1. General*

The synthetic routes of CB3CB and CB2OCB are shown in Scheme 1. The molecular structures were analyzed using 1H and 13C nuclear magnetic resonance (NMR) spectroscopy on a JNM ECX 500 spectrometer (JEOL Ltd., Tokyo, Japan). Phase identification was conducted via polarized optical microscopy (POM) using a BX50 microscope (Olympus Corp., Tokyo, Japan) on an LK-600 PM hot stage (Linkam, Surrey, UK). The phase-transition temperatures and associated enthalpy changes were determined using differential scanning calorimetry (DSC) on a DSC-60 Plus (Shimadzu Corp., Kyoto, Japan). Calibration was performed using indium, and the measurements were performed over heating/cooling/heating cycles at a rate of 10 ◦C min−<sup>1</sup> under an N2 gas flow (50 mL min−1).

**Scheme 1.** Synthesis of (**a**) CB3CB and (**b**) CB2OCB.

### *2.2. Synthesis*

2.2.1. 1,3-Bis(4-cyanobiphenyl-4 -yl)propane (CB3CB)

(*E*)-1,3-Bis(4-bromophenyl)prop-2-en-1-one

This compound was synthesized referring to a previously reported method [68]. 4 -Bromoacetophenone (2.19 g, 11.0 mmol), *p*-bromobenzaldehyde (2.04 g, 11.0 mmol), sodium ethoxide (NaOEt) (1.12 g, 16.5 mmol), ethanol (EtOH) (30 mL), and distilled water (10 mL) were added to a round-bottom flask. The resultant mixture was stirred at ambient temperature for 1 h. The reaction mixture was filtered, and the residue was rinsed with copious amounts of methanol to afford the target compound as a pale-yellow solid (86%). 1H NMR (500 MHz, CDCl3) δ 7.89 (d, *J* = 9.0 Hz, Ar–*H*, 2H), 7.75 (d, *J* = 16.0 Hz, CO–C*H*, 1H), 7.65 (d, *J* = 9.0 Hz, Ar–*H*, 2H), 7.56 (d, *J* = 9.0 Hz, Ar–*H*, 2H), 7.51 (d, *J* = 9.0 Hz, Ar–*H*, 2H), 7.47 (d, *J* = 16.0 Hz, CO–CH=C*H*, 1H) ppm. 13C NMR (126 MHz, CDCl3) δ 189.1, 143.9, 136.7, 133.7, 132.3, 132.0, 130.0, 129.8, 128.1, 125.1, 121.9 ppm.

### 1,3-Bis(4-bromophenyl)propane

This compound was also synthesized referring to the literature [68]. (*E*)-1,3-Bis(4 bromophenyl)prop-2-en-1-one (1.10 g, 2.99 mmol) was added to a two-necked roundbottom flask, which was then purged with argon gas. Trifluoroacetic acid (TFA) (11 mL) was then added into the flask, followed by adding triethylsilane (TES) (4.8 mL, 30 mmol) dropwise. The resultant mixture was stirred at ambient temperature for 1 h. TES (0.95 mL, 5.94 mmol) was added to the reaction mixture, which was further stirred for 18 h. The reaction mixture was poured into distilled water, extracted with dichloromethane, and washed with brine. The obtained solution was then dried over magnesium sulfate (MgSO4), and the solvent was removed under reduced pressure. The residue was purified by column chromatography on silica gel using hexane as an eluent to afford 1,3-bis(4-bromophenyl)propane as a colorless solid (78%). 1H NMR (500 MHz, CDCl3) δ 7.39 (d, *J* = 8.5 Hz, Ar–*H*, 4H), 7.04 (d, *J* = 8.5 Hz, Ar–*H*, 4H), 2.58 (t, *J* = 7.5 Hz, Ar–C*H*2, 4H), 1.90 (tt, *J* = 7.5 and 7.5 Hz, Ar–CH2–C*H*2, 2H) ppm. 13C NMR (126 MHz, CDCl3) δ 140.9, 131.4, 130.2, 119.5, 34.6, 32.6 ppm.

### CB3CB

1,3-Bis(4-bromophenyl)propane (200 mg, 0.565 mmol), 4-(4,4,5,5-tetramethyl-1,3,2 dioxaborolan-2-yl)benzonitrile (267 mg, 1.17 mmol), cesium carbonate (Cs2CO3) (748 mg, 2.29 mmol), and tetrakis(triphenylphosphine)palladium(0) [Pd(PPh3)4] (41.5 mg, 35.9 μmol) were added to a two-necked round-bottom flask, which was then purged with argon gas. Tetrahydrofuran (THF) (5 mL) was degassed by bubbling with argon gas and added to the flask. The resultant mixture was stirred at reflux temperature for 3 h. Subsequently, Pd(PPh3)4 (39.5 mg, 34.2 μmol) was added to the reaction mixture, which was further stirred for 4 h. The reaction mixture was extracted with dichloromethane and washed with brine. The solution was then dried over MgSO4, and the volatile solvent was removed under reduced pressure. The residue was purified by column chromatography on silica gel using dichloromethane/hexane (1:1, *v*/*v*) and recrystallized in a dichloromethane/hexane mixture to afford CB3CB as a colorless solid (51%). 1H NMR (500 MHz, CDCl3) δ 7.71 (d, *J* = 8.5 Hz, Ar–*H*, 4H), 7.67 (d, *J* = 8.5 Hz, Ar–*H*, 4H), 7.53 (d, *J* = 8.5 Hz, Ar–*H*, 4H), 7.31 (d, *J* = 8.5 Hz, Ar–*H*, 4H), 2.74 (t, *J* = 7.5 Hz, Ar–C*H*2, 4H), 2.04 (tt, *J* = 7.5 and 7.5 Hz, Ar–CH2–C*H*2, 2H) ppm. 13C NMR (126 MHz, CDCl3) δ 145.5, 142.9, 136.7, 132.6, 129.2, 127.5, 127.2, 119.0, 110.6, 35.0, 32.7 ppm.

2.2.2. (4-Cyanobiphenyl-4 -yloxy)-2-(4-cyanobiphenyl-4 -yl)ethane (CB2OCB) 1-Bromo-4-[2-(4-bromophenoxy)ethyl]benzene

4-Bromophenethyl bromide (690 mg, 2.61 mmol), 4-bromophenol (302 mg, 1.75 mmol), and potassium carbonate (K2CO3) (617 mg, 4.46 mmol) were added to a two-necked roundbottom flask, which was then purged with argon gas. *N*,*N*-Dimethylformamide (DMF) (5 mL) was degassed by bubbling argon gas and then added to the flask. The resultant

mixture was stirred at 90 ◦C for 18 h. More 4-bromophenethyl bromide (690 mg, 2.61 mmol) was added to the reaction mixture, which was further stirred for 5 h. The reaction mixture was extracted with ethyl acetate and washed with brine. The obtained solution was dried over MgSO4, and then the solvent was removed under reduced pressure. The residue was purified by column chromatography on silica gel using hexane/ethyl acetate (20:1, *v*/*v*) to afford 1-Bromo-4-[2-(4-bromophenoxy)ethyl]benzene (28%). 1H NMR (500 MHz, CDCl3) δ 7.43 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 7.35 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 7.14 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 6.75 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 4.11 (t, *J* = 6.5 Hz, Ar–O–C*H*2, 2H), 3.03 (t, *J* = 6.8 Hz, Ar–O–CH2–C*H*2, 2H) ppm. 13C NMR (126 MHz, CDCl3) δ 157.8, 137.1, 132.3, 131.6, 130.7, 120.4, 116.3, 113.0, 68.5, 35.1 ppm.

### CB2OCB

1-Bromo-4-[2-(4-bromophenoxy)ethyl]benzene (96.7 mg, 0.272 mmol), 4-(4,4,5,5-tetra methyl-1,3,2-dioxaborolan-2-yl)benzonitrile (140 mg, 0.611 mmol), Cs2CO3 (182 mg, 0.559 mmol), and Pd(PPh3)4 (64.1 mg, 55.5 μmol) were added to a two-necked roundbottom flask, which was then purged with argon gas. THF (3 mL) was degassed by bubbling argon gas and then added to the flask. The resultant mixture was stirred at reflux temperature for 16 h. An arbitrary amount of Pd(PPh3)4 was added to the reaction mixture. The mixture was then stirred for 16 h, extracted with dichloromethane, and washed with brine. The solution was dried over MgSO4, and the volatile solvent was removed under reduced pressure. The residue was purified by column chromatography on silica gel using dichloromethane/hexane (5:1, *v*/*v*) and recrystallized in a dichloromethane/hexane mixture to afford CB2OCB (20%). 1H NMR (500 MHz, CDCl3) δ 7.72 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 7.69 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 7.67(d, *J* = 8.0 Hz, Ar–*H*, 2H), 7.63 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 7.55 (d, *J* = 8.5 Hz, Ar–*H*, 2H), 7.52 (d, *J* = 9.0 Hz, Ar–*H*, 2H), 7.42 (d, *J* = 8.0 Hz, Ar–*H*, 2H), 7.00 (d, *J* = 9.0 Hz, Ar–*H*, 2H), 4.27 (t, *J* = 6.5 Hz, Ar–O–C*H*2, 2H), 3.18 (t, *J* = 6.5 Hz, Ar–O–CH2–C*H*2, 2H) ppm. 13C NMR (126 MHz, CDCl3) δ 159.3, 145.3, 145.1, 138.9, 137.5, 132.59, 132.56, 131.7, 129.8, 128.4, 127.5, 127.3, 127.1, 119.1, 118.9, 115.1, 110.8, 110.1, 68.5, 35.3 ppm.

### **3. Results**

### *3.1. Phase-Transition Behaviors of CB3CB, CB2OCB, and CB2SCB*

As shown in the DSC curves (Figure 2a), the methylene-linked CB3CB sample exhibited a melting temperature (*T*m) of ~149 ◦C, where it transitioned to the Iso phase without forming the LC phase upon heating. Upon cooling, the CB3CB sample crystallized at ~80 ◦C from the Iso phase. These phase-transition temperatures were higher than those (142.1 and 69.1 ◦C, respectively) reported in the literature [67], where CB3CB did not exhibit the LC phase. However, POM observations in this study revealed that in the supercooled Iso phase of CB3CB (Figure 3a), which does not undergo crystallization at ~80 ◦C, birefringent textures appear (Figure 3b), which then grow mixed textures including fan-, focal-conic-, and rope-like domains, as shown in Figure 3c,d. Besides, we did not observe typical N-phase textures such as marble and schlieren textures during this phase transition, as seen in Figure 3. This texture behavior is similar to the direct Iso–NTB phase transition of CB2SCB [62]. Therefore, it was revealed that CB3CB also shows the monotropic direct Iso–NTB phase transition at ~49 ◦C. During the heating after the cooling, the observed NTB phase of CB3CB transitioned to the Iso phase at ~55 ◦C. The CB3CB sample displayed a strong crystallization tendency and did not vitrify upon cooling, even at a higher rate of 30 ◦C min−1. Additionally, the ether-linked CB2OCB did not show LC phases over the first and second heating cycles, where it exhibited different *T*<sup>m</sup> values of 164.4 and 139.7 ◦C, respectively; Figure 2b represents the latter. This result indicates the existence of crystal polymorphs that depend on the crystallization conditions. Upon cooling, most of the Iso-phase domains and droplets of CB2OCB crystallize at ~104 ◦C, as shown by the exothermic peak in Figure 2b. However, the POM images reveal the formation of the N and NTB phases at ~108 and 78 ◦C, respectively, in the supercooled Iso phase, as confirmed

by the marble/schlieren textures (Figure 4a) and blocky texture (Figure 4b), respectively. Because of the strong crystallization tendencies of CB3CB and CB2OCB, as shown by the DSC curves (Figure 2a,b, respectively), their LC phases were not investigated by X-ray diffractometry. The strong crystallization tendency of these molecules differs from the vitrifiable CB2SCB, with a glass transition temperature of ~20 ◦C, as shown in Figure 2c [62]. The *T*<sup>m</sup> values upon the second heating, the associated enthalpy changes (Δ*H*), and the Iso–NTB, Iso–N, and N–NTB phase-transition temperatures upon the cooling (*T*INTB, *T*IN, and *T*NNTB, respectively) of CB3CB, CB2OCB, and CB2SCB are summarized in Table 1. For simplicity, the crystallization and glass transition temperatures upon cooling are not listed in Table 1.

**Figure 2.** DSC curves of (**a**) CB3CB, (**b**) CB2OCB, and (**c**) CB2SCB upon the second heating (red lines) and cooling (blue lines) cycles at a rate of 10 ◦C min<sup>−</sup>1. Cr and G denote the crystal phase and glassy state, respectively. Panel (**c**) is reproduced from Ref. [62].

**Figure 3.** POM images during the Iso–NTB phase transition of CB3CB: (**a**) the Iso phase (52 ◦C), (**b**) the NTB texture appearance in the Iso phase (50 ◦C), (**c**,**d**) growth in the NTB texture (48.5 and 47.5 ◦C, respectively).

**Figure 4.** POM images of (**a**) the N phase (104 ◦C) and (**b**) the NTB phase (69 ◦C) of CB2OCB.

**Table 1.** *T*<sup>m</sup> and the associated Δ*H* upon second heating and *T*INTB, *T*IN, and *T*NNTB upon cooling for CB3CB, CB2OCB, and CB2SCB.


<sup>a</sup> Determined by POM. <sup>b</sup> Obtained from Ref. [62]. <sup>c</sup> Obtained upon first heating.

Thus, CB3CB and CB2SCB exhibited the direct Iso–NTB phase transition, whereas CB2OCB showed the typical Iso–N–NTB phase transition. Moreover, the CB3CB and CB2OCB samples displayed a stronger crystallization tendency compared with the thioetherlinked CB2SCB, which had a vitrifiable NTB phase [62]. The Iso–LC phase-transition temperatures (i.e., *T*INTB or *T*IN) showed the order CB2OCB (108 ◦C) >> CB3CB (49 ◦C) > CB2SCB (43 ◦C), which translated to the order ether >> methylene > thioether in terms of linkage type. Naturally, the NTB phase-transition temperatures (i.e., *T*INTB or *T*NNTB) also showed the same order: ether >> methylene > thioether. This particularly high phase-transition temperature (especially for the LC–Iso or Iso–LC phase transition) with the ether linkage is typical for usual calamitic LCs [69–71], including bent LC dimers [28,30,31,33,37,39]. Additionally, CB2SCB is vitrifiable, whereas CB3CB and CB2OCB strongly crystallize. These trends in the LC phase-transition temperatures and crystallization or vitrification abilities of the three dimers could be attributed to their different linkages, that is, ether (C–O–C), methylene (C–CH2–C), and thioether (C–S–C). The C–O–C bond angle (118◦) is larger than those of the other linkage types, which renders CB2OCB more anisotropic [72]. In addition, The higher rotational barrier [73,74] and stronger electron-donating property of the Ph–O bond may contribute to the molecular rigidity and intermolecular interactions of CB2OCB, respectively. These could result in higher *T*IN and *T*NNTB, a larger Δ*T*N, and, possibly, at least in part, a stronger crystallization tendency. The higher *T*INTB of CB3CB compared with that of CB2SCB is attributed to its higher molecular anisotropy owing to the larger C–CH2–C bond angle (~110◦) compared with the C–S–C angle (~100◦), as well as higher rigidity due to the higher rotational barrier of the C–CH2 bond compared with that of the C–S bond [75]. The higher *T*<sup>m</sup> and *T*IN and stronger crystallization tendency of CB3CB compared with those of CB2SCB could also be attributed to the symmetric molecular structure of the former. A more bent, flexible C–S–C linkage and molecular asymmetry endow CB2SCB with lower phase-transition temperatures and a vitrification ability compared with CB3CB and CB2OCB [33,39,62].

### *3.2. Phase-Transition Behaviors of CBnCB, CBnOCB, and CBnSCB*

As described in the Introduction, each bent CB-based dimer homolog series with the longer spacers CB*n*CB (*n* = 5, 7, 9, 11, and 13) [17,25], CB*n*OCB (*n* = 4, 6, 8, and 10) [24,25], and CB*n*SCB (*n* = 4, 6, 8, and 10) [62] exhibit the typical Iso–N–NTB phase sequence. The

*T*IN, *T*NNTB, and Δ*T*<sup>N</sup> of these homologous series are plotted in Figure 5a–c, respectively, as a function of the total number of atoms in the spacer chain lengths, i.e., *n* for CB*n*CB and *n* + 1 for CB*n*OCB and CB*n*SCB, including the O and S atoms. The *T*INTB values of CB3CB and CB2SCB are included in the plots shown in Figure 5a,b.

**Figure 5.** (**a**) *T*IN, (**b**) *T*NNTB, (**c**) Δ*T*N, and (**d**) schematic models of dimers with shorter and longer spacers showing anisotropy and flexibility. In (**a**,**b**), the *T*INTB values are plotted for CB3CB and CB2SCB because they exhibit the Iso–NTB phase transition. Panel (**d**) was reproduced from Ref. [39].

Overall, the *T*IN and *T*NNTB values were approximately in the order of CB*n*OCB > CB*n*CB > CB*n*SCB, which could be similarly ascribed to the characteristics of the linkers described in Section 3.1 for the shortest homologs. Nevertheless, the *T*NNTB values of the ether-linked CB*n*OCB were relatively close to or partly lower than those of the methylenelinked CB*n*CB. This observation may partly be attributed to the characteristics of the more anisotropic structure of CB*n*OCB because the NTB phase formation for a more anisotropic molecular structure likely requires greater supercooling of the N phase compared to a more bent one [33]. Consequently, the Δ*T*<sup>N</sup> values are in the order of CB*n*OCB > CB*n*SCB > CB*n*CB for all *n*, as shown in Figure 5c. The Δ*T*<sup>N</sup> values of the ether-linked CB*n*OCB homologs are significantly larger than those of the others for all *n* owing to their high *T*IN values.

Next, we investigated the *n* dependence of the phase-transition temperatures of the three homologs. The *T*IN (or *T*INTB) and *T*NNTB values for all the CB*n*CB, CB*n*OCB, and CB*n*SCB homologs increase with increasing *n* and then level off or gradually decline with a further increase in *n*, as shown in Figure 5a,b. Consequently, with increasing *n*, Δ*T*<sup>N</sup> increases for all the homologs, reaches a maximum, and then gradually declines for CB*n*CB and CB*n*OCB, as shown in Figure 5c. These trends of *T*IN (or *T*INTB), *T*NNTB, and Δ*T*<sup>N</sup> for all the dimer homologs could be attributed to the average molecular shape and flexibility with increasing/decreasing *n*, as shown in Figure 5d [39]. The shortest dimers could be more bent (strong biaxiality); hence, their *T*IN (or *T*INTB) and *T*NNTB values were lower than those of the longer dimers. With increasing *n*, the average molecular shape of the bent dimer homologs becomes more linear (or anisotropic), thereby increasing *T*IN and *T*NNTB to some extent. However, further lengthening of the central spacer could enhance the molecular flexibility and dilute the polarizable mesogenic arms that increase the phasetransition temperatures; hence, these phase-transition temperatures nearly remain constant or gradually decline. Thus, the molecular biaxiality (molecular curvature) of the dimer homologs increases with decreasing *n*, which principally decreases *T*IN, and consequently, Δ*T*N. This results in Δ*T*<sup>N</sup> = 0, i.e., the direct Iso–NTB phase transition for the shortest CB3CB and CB2SCB [62].

### **4. Conclusions**

In this study, we evaluated the phase-transition behaviors of three CB-based dimers with propane, ethoxy, and ethylthio spacers. Analogous to the previously reported CB2SCB, the short CB3CB exhibited the rare direct Iso–NTB phase transition, whereas CB2OCB showed the typical Iso–N–NTB phase transition. CB3CB and CB2OCB have strong crystallization tendencies, whereas the thioether-linked CB2SCB exhibited a vitrifiable NTB phase. The NTB phase-transition temperature (*T*INTB or *T*NNTB) decreased in the order CB2OCB (76 ◦C) > CB3CB (49 ◦C) > CB2SCB (43 ◦C). The phase-transition behaviors of all the CB*n*CB, CB*n*OCB, and CB*n*SCB homologs, including those with longer chains, were comprehensively examined. The more anisotropic ether-linked CB*n*OCB series showed significantly higher *T*IN and wider Δ*T*<sup>N</sup> for all *n*. Regarding shorter spacers, the phase-transition temperatures decreased, especially *T*IN. Hence, the Δ*T*<sup>N</sup> for all three homologous series decreased, resulting in the direct Iso–NTB phase transition for the short-spacer-bearing CB3CB and CB2SCB. This phenomenon could partly be ascribed to their bent molecular geometry or enhanced molecular biaxiality owing to their short lengths. Our findings provide new insights into the effects of linkage types on the molecular design of LC dimers that exhibit the direct Iso–NTB phase transition.

**Author Contributions:** Conceptualization, Y.A. (Yuki Arakawa); methodology, Y.A. (Yuki Arakawa); validation, Y.A. (Yuki Arakawa); formal analysis, Y.A. (Yuki Arakawa), Y.A. (Yuto Arai), K.H., and K.K.; investigation, Y.A. (Yuki Arakawa), Y.A. (Yuto Arai), K.H., and K.K.; resources, Y.A. (Yuki Arakawa) and H.T.; data curation, Y.A. (Yuki Arakawa); writing—original draft preparation, Y.A. (Yuki Arakawa); writing—review and editing, Y.A. (Yuki Arakawa), Y.A. (Yuto Arai), K.H., and H.T.; supervision, Y.A. (Yuki Arakawa) and H.T.; project administration, Y.A. (Yuki Arakawa); funding acquisition, Y.A. (Yuki Arakawa). All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Japan Society for the Promotion of Science (KAKENHI grant numbers 17K14493 and 20K15351) and Toyohashi University of Technology.

**Data Availability Statement:** Data are presented in the article.

**Conflicts of Interest:** The authors declare no financial conflict of interest.

### **References**

