Pseudo-rodlike molecules with hockey-stick-shaped mesogen

ABSTRACT Hockey-stick-shaped molecules were newly synthesised to obtain pseudo-rodlike molecules. The designed molecules consist of a polar terminal ring (i.e. 2,3,4-, 2,4,6- or 3,4,5-trifluorophenyl group), a rigid middle block (i.e. four rings with aligned ester linkages) and a flexible terminal chain (i.e. dodecyloxy group). We found that the compounds with 2,3,4- and 3,4,5-trifluorophenyl groups formed a smectic A mesophase with head-to-head bi-layer building blocks, whereas the compound with 2,4,6-trifluorophenyl group formed a nematic mesophase. This might be concerned with the behaviour of pseudo-rodlike molecules GRAPHICAL ABSTRACT


Introduction
Over the past few decades, many studies have examined the structure-property relationship of bent-core molecules, which can be differentiated from rod-like molecules because the nonlinear mesogenic structures with deficient molecular symmetry yield unconventional mesomorphism such as ferroelectricity, antiferroelectricity, flexoelectricity, cybotacticity, biaxiality etc. [1][2][3][4][5][6][7][8] In general, the two arms of the bent-core molecule are composed of a nonlinear central mesogenic unit with two flexible terminal chains, which are of substantially equal structure. Furthermore, if the two arms of the molecule are roughly different, the hockey-stick-shaped molecules are constructed. [9][10][11][12][13] In other words, this type of molecules are considered to have an intermediate structure between rodlike and bent-core molecules. From a structural viewpoint, hockey-stick-shaped molecules can be classified into two categories. In the first type, a rod-like mesogen possesses two flexible terminal chains: one chain is connected to the para position of one side of the terminal ring, and the other is connected to the meta position of the other side. [14][15][16][17][18] In the second type, the bent-core mesogen consists of two arms with different numbers of aromatic rings that have one or two terminal chains. [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32] Recently, Prasad et al. reported synthesis of a hockey-stick-shaped azo compound in a study of the effect of linkage variation on the mesomophic properties. [12] Strictly speaking, the above two types of molecules can be considered as belonging to the family of non-symmetrical bent-core molecules. On the other hand, if a bent-core molecule possesses two arms of greatly different lengths, the 'real' hockey-stick-shaped molecules can be created because apparently the molecular structure looks like a hockey stick shape. This type can be considered as the third type of hockey-stick-shaped molecules. Surprisingly, very few studies have been done so far to develop the third type of compounds. Samulski's group has reported the synthesis of the compounds based on p-quinquephenyl group. [33] Recently, we have predicted that the compound with polar head may form polar smectic phase as well as we have publicised some part of synthesis and mesomorphic property as a patent. [34][35][36] In this study we have focused on the investigation of 'real' intermediate properties between those of rod-like and bent-core compounds, i.e. the pseudo-rodlike compounds.
In this work, we have synthesised and characterised the third type of hockey-stick-shaped compounds, which we call the 'real' hockey-stick-shaped molecules (Scheme 1). The trifluorinated phenyl ring in the short arm may contribute to enhance the net dipole moment as well as the polarisabilty anisotropy by incorporation with the conjugated aromatic parts in the long arm, which have aligned ester-linkages. The structure of the compounds was identified using IR and NMR spectrometry and an elemental analysis; and their mesogenic properties were investigated using polarised optical microscopy (POM), differential scanning calorimetry (DSC), and X-ray diffractometry (XRD) measurements.

Experimental details
All compounds were prepared as described in [36] (synthesis of compounds 3a-3c is in supplemental data). Scheme 1. Synthesis of the hockey-stick-shaped compounds.

General characterisation
IR and NMR spectra were obtained using a Jasco 300E FT/ IR and a Brucker Advance 400 MHz spectrometer, respectively. Elemental analysis was performed using a Thermo Finnigan EA1108. The transition behaviours were characterised by DSC (Netzsch DSC 200 F3 Maia). DSC measurements were performed in an N 2 atmosphere with heating and cooling rates of 10°C/min. Optical texture observation was conducted using a polarising microscope (Axioskop 40 Pol) with a heating stage (Mettler FP82HT).

X-ray measurements
X-ray studies were performed using the reflection mode of a Rigaku 12 kW rotating-anode X-ray (Cu Kα radiation) generator coupled with a diffractometer. The diffraction peak positions and widths were calibrated with silicon crystals in the high-2θ-angle region (>15°) and with silver behenate in the low-2θ-angle region. To monitor the structural evolution under temperature changes, a hot stage calibrated to be within a ±1°C error was coupled to the diffractometer. Samples were scanned across a 2θangle range of 1.5°-35°at a scanning rate of 2°/min. To orient the chains, each sample was mechanically spun from a homemade mini-extruder with a mild external force at an appropriate temperature below the isotropic temperature and was then quenched to room temperature. The XRD patterns for oriented samples were obtained using a Rigaku X-ray imaging system with an 18 kW rotating anode X-ray generator. Silicon crystal powder, which was used as an internal reference, shows a diffraction ring at a 2θ value of 28.466°. At least 30 min of exposure time was required to obtain high-quality patterns. All background scatterings were subtracted from the sample scans.

Results and discussion
In Table 1, compounds 3a and 3b show the melting (T m ) and isotropisation temperatures (T i ) on both heating and cooling, whereas compound 3c exhibits T i only on cooling (DSC themograms of compounds 3a-3c are in supplemental data). The T m s and enthalpy changes for melting (ΔH m ) are ca. 165°C and ca. 55 kJ/mol, respectively. Unlike the T m and ΔH m values, the T i s and enthalpy changes for isotropisation (ΔH i ) are in the range of 156-171°C and 0.6-5.9 kJ/mol, respectively, depending on the structure. In particular, the ΔH i values of 3a and 3b are about 9 times higher than that of 3. This implies that compounds 3a and 3b formed an enantiotropic smectic mesophase, whereas compound 3c formed a monotropic nematic mesophase. Figure 1 displays the optical textures of mesophases. On cooling, compound 3a showed droplets (see Figure 2a in supplemental data) that grew into focal conic texture for smectic phase (Figure 1(a)). On heating, this compound exhibited the texture for misaligned smectic A phase (see Figures 2b and 2c in supplemental data). On cooling, compound 3b showed batonnets at 167°C (see Figure 2d in supplemental data), which grew into bigger droplets at 163°C (see Figure 2e in supplemental data) and a focal conic-like texture with a line pattern of equidistant strips at 150°C (Figure 1(b)). On heating, this compound exhibited the texture for misaligned smectic A phase at 166°C (see Figure 2f in supplemental data) and concentric circles at 168°C (Figure 1(c)). Compound 3c showed the unknown texture for nematic phase on cooling (Figure 1(d)), but exhibited no optical textures on heating. Figure 2 shows X-ray diffraction (XRD) patterns obtained during heating and cooling. At 30°C, compound 3a showed several distinct diffraction peaks indicating crystalline state. During heating below 168°C (the T m defined by DSC) a distinctive peak appeared continuously in the small-angle region, while all peaks in the wide-angle region were gradually merged into a peak with 2θ = 19.7°(d = 0.45 nm). At 165°C, sharp low-angle and broad high-angle reflections appeared at 2θ = 1.96°(d = 4.50 nm) and around 2θ = 19.7°, respectively. This is indicative of smectic mesophase. At over 169°C (the T i defined by DSC) two amorphous halos around 2θ = 2.19°(d = 4.03 nm) and 2θ = 19.1°( d = 0.46 nm) appeared, indicating isotropic liquid. The former corresponds to the average periodicity of electron density fluctuations between the nanophaseseparated rigid aromatic cores and alkyl tails, and the latter to the average lateral distance between the amorphous chains. Subsequently, when the isotropic liquid was cooled, the patterns for smectic phase and crystalline state appeared reversibly. Compound 3b showed the similar diffraction patterns with 3a (see Figure 3 in supplemental data) whereas compound 3c showed the different diffraction patterns from 3a. When Table 1. Transition temperatures (°C) and enthalpy changes (kJ/ mol) a. compound 3c was heated above T m , broad scattering patterns were only observed in the small-as well as wide-angle region, indicating isotropic liquid. Subsequently, when the isotropic liquid was cooled below T i , the pattern for nematic phase appeared monotropically.

Code DSC run
In Table 2 the layer spacing (d) values of smectic phase are in a range of 4.20-5.34 nm. For comparison, the d values for crystalline, isotropic and nematic phases in the low-angle region are also included in this table. We have estimated and could match the length (l) and bent angle (ϕ) of molecules with the all-trans conformation by using the COMPASS force field of the Cerius 2 4.6 software at 0 K (see Figure 3). To account for stable conformation, we postulated that phenyl rings are coplanar when carbonyl groups are attached to them, and alkyloxy groups in the all-trans conformation are equiplanar to the benzene rings. [37,38] Probably the most striking result is that the fine conformation could be controlled by the adjustment of the bending angle of the real hockey-stickshaped molecules through varying the substitution position of three fluoro atoms in a terminal ring. Moreover, we postulate that this change in the bending angle can be the main factor explaining why 3c formed nematic mesophase, while 3a and 3b formed smectic mesophase. Figure 4 presents 2D XRD patterns of the oriented samples. In Figures 4(a,e), 3(a,b) showed strong diffraction spots in the low-angle region on the equator. This indicates that the layer normal of crystalline phase is perpendicular to the shear direction, which is along the meridian. At room temperature, 3a and 3b showed the highly 3D ordered hierarchical layered crystalline structures instead of a simple layered crystalline structure. In contrast, at room temperature, 3c exhibited the single molecular building block with no layered structure in spite of a complicated diffraction pattern including an intense one at wide angle region around 2θ = 19.07° (Figure 4(h)). At 150°C, the high-angle scattering for 3a becomes much more diffusive while the low-angle diffraction appears almost unchanged (Figure 4(c)). This implies that the 3D ordering on the sub-nanometre scale is lost while the basic structure remains. In contrast, at 157°C the scattering pattern for 3b was scarcely changed compared with that of room temperature (Figure 4(f)). This indicates that almost no change occurred in the crystalline structure during the transition. In Figures 4(d,g) and 3(a,b) showed that the small angle reflections split into two isolated spots centred on the equator and perpendicular to two diffused wide-angle reflections centred on the meridian. This is indicative of the oriented smectic A mesophase.   The data were obtained from the XRD profiles shown in Figure 2  It is worth mentioning that in ambient temperature, the 2D XRD patterns of crystalline state for 3a show the d = 7.68, 3.84 and 2.56 nm reflections, which could be indexed by (001), (002) and (003)type planes. Considering that the molecular length of 3a was about 4.2 nm, we claim that the head-to-head type structure (about 8.4 nm) consisting of two molecules can be formed in the layer. Practically, when the bi-layer becomes tilting or inter-digitation, the (001) diffraction can be observed at about 8.0 nm of d. If you assume that smectic phase forms the head-to-tail single-layer structure, the d value at 5.34 nm would be indexed as the (001)-type plane. However, this assumption can be excluded because the free volume creations and conformation changes for this phase transition should require an enormous amount of energy. Note that the head-to-head bi-layer structure in crystalline state cannot turn into the head-to-tail single layer structure in a smectic phase. Therefore, the diffraction patterns for compound 3a on the meridian at 2θ = 1.65°should be assigned as (002), and the odd-numbered diffractions were extinct a b c   owing to the identical electron density distribution within the layers. The d value in the crystalline state is smaller than that in the smectic phase because during crystallisation alkyl chains and phenyl rings are more closely packed and the contraction between the layers may occur. Finally, we have concluded that 3a and 3b compounds form the smectic A phase with the head-to-head bi-layer building block (see Figure 5). In contrast, the nematic phase of 3c showed no building blocks. To account for a possible superstructure, we shall now discuss more carefully whether the polar head in our molecules will lean towards the same direction or towards the opposite direction in the head-to-head bi-layer building blocks. We assume that for compound 3b the 3,4,5-trifluorophenyl head groups with longitudinal symmetry are placed in the same direction, simultaneously keeping the bi-layer smectic A structure ( Figure 5(b)). Unlike compound 3b, the frustrated phases are not observed in the case of compound 3a, because the dipole moment induced by 2,3,4-trifluorophenyl head groups with transverse symmetry may be cancelled out to make the system stable ( Figure 5(a)).

Conclusions
Three pseudo-rodlike molecules with a hockey-stickshaped mesogen were studied. The long arm was composed of four rings with aligned ester linkages and connected to dodecyloxy end group. The short arm was composed of one ring with 2,3,4-, 3,4,5-or 2,4,6-trifluoro substituents. Our results suggest that the bending angle of stable conformation can be adjusted by varying the substitution position of the trifluoro substituents, which can significantly affect the property of mesophase. Compounds with 2,3,4and 3,4,5-trifluorophenyl groups formed the enantiotropic smectic A phase with head-to-head bi-layer building blocks while compound with 2,4,6-trifluorophenyl group formed the monotropic nematic phase with no building blocks. Meanwhile, there was a report that the compound 3c formed a smectic phase. [34][35][36] However, in this study we found that this compound forms a nematic phase instead of a smectic phase. In addition, it was reported that the compounds 3a and 3b may form a polar smectic mesophase showing a single reverse current with a surface alignment. [34] However, it has not been confirmed as yet because the reproducibility for the reverse current were very sensitive to the alignment condition. This result might be attributed to the fact that the polar heads of such pseudo-rodlike molecules can be oriented either in the promotive or destructive direction of the net dipole moment in the smectic A, head-to-head bi-layer build block. Nevertheless, it is still an unsettled question, and further study is necessary to ascertain this point.

Disclosure statement
No potential conflict of interest was reported by the authors.