The impact of flexibility of polyimides backbones on the stability of liquid crystal vertical alignment

Abstract

Two series of polyimides (PIs) which contained identical side chains but different backbones were synthesized from the same functional diamine: RPI-X% and SPI-X% (X = 20, 30, 40, X stands for the molar content of functional diamine). Rigid polyimides (RPIs) are PIs whose backbones are composed of wholly rigid aromatic units and soft polyimides (SPIs) are PIs whose backbones are composed of flexible ether units and aromatic units. The surface morphology, chemical composition and molecular orientation of the polyimide (PI) alignment layer surfaces were investigated by atomic force microscopy, X-ray photo-electron spectroscopy and polarized attenuated total reflection Fourier transformed infrared spectroscopy, respectively. The results showed that both RPI and SPI could induce LC to align vertically without a rubbing process when the molar content of the functional diamine reached 30%. However, the rubbed RPI-30% whose side chains were oriented vertically induced vertical alignment of the liquid crystal (LC), while the rubbed SPI-30% whose side chains were oriented parallel induced parallel alignment of the LC. The cause of this distinction was attributed to the different main chain structures: the wholly rigid aromatic units in RPI's main chains could restrict the movements of molecular chains so the side chains can maintain vertical orientation all the time; flexible ether bands in SPI main chains could endow molecular chains with mobility so the side chains fall over easily after rubbing.

---

1. Introduction

In the field of liquid crystal displays (LCDs), the vertical alignment mode which can achieve a wide viewing angle, short response time, and high contrast ratio at the same time has been widely used.^1–5 This mode needs liquid crystal (LC) molecules to be aligned vertically at the field-off state, which means that the pretilt angle should be above 89° after a rubbing process.⁶ Pretilt angle is the tilt angle between the average orientation of LC molecules and the substrate as shown in Fig. 1.^7,8 It is a crucial alignment characteristic which prevents the creation of reverse tilt disclinations in LCDs.⁹ Many researchers have noticed that polyimides (PIs) with long side chains contribute to achieving a high pretilt angle.^10–13 Thus, PIs which possess bulky side chains composed of rigid units and long alkyl chains have been widely used as the alignment layer to achieve pretilt angles above 89°. Although it has been reported that some main-chain-type PIs containing short side groups can obtain high pretilt angles (about 15°), they can hardly make pretilt angles reach 89°.¹⁴ If the pretilt angle can't up to 89°, it will not be utilized for vertical alignment layer and generate the excellent characteristics of display as mentioned in the opening.

Fig. 1 (a) Systematic diagram of rubbing. v: glass substrate moving speed; n: rpm of the rotating roller; M: contact length; r: rotating roller radius + rubbing cloth thickness. (b) A schematic drawing of the structure of a typical liquid crystal cell without twist in an antiparallel assembly.

To better understand the alignment behavior of LC molecules on rubbed PI films, the chemical structure, the length, the conformation or orientation of side chains and the interactions between side chains and LC molecules have been researched deeply for many years.^15–21 However, there were very few researchers investigated the effect of main chains of PIs on vertical LC alignment. Arafune et al. pointed out that the pretilt angle of LC increased with the inclination angle of the backbone structure.²² Seo et al. concluded that PI with trifluoromethyl moieties in its backbone contribute to achieving a high pretilt angle.²³ Chern et al. claimed that PIs containing only short side groups could achieve high pretilt angel because of the flat-straight conformations of backbones.²⁴ Tanaka et al. discovered an alternation of side-chain mesogen orientation caused by the backbone structure in liquid-crystalline polymer thin films.²⁵ These discussions cause us to infer that the main chain structure of side-chain-type PIs may have unexpected effects on vertical LC alignment too. Therefore, we synthesized a series of PIs possessing identical side chains but different main chains. The side chain structure was designed from the models of LCs which contained rod-like components (biphenyl groups) and long tails (alkoxy groups). Meanwhile, the main chains can be classified into two types according to their flexibility: rigid main chains composed of wholly rigid aromatic units; and soft main chains composed of flexible ether units and aromatic units. The PIs containing rigid main chains are named rigid polyimide (RPI). While, the PIs possessing soft main chains are named soft polyimide (SPI). The preliminary experiments showed that LC cells with RPIs and SPIs exhibited absolutely different alignment. The RPIs could induce LCs to align vertically no matter whether they were rubbed or not. Nevertheless, for the SPIs, both parallel and vertical alignment of LCs appeared, even some of them had the ability to induce vertical alignment without rubbing but lost it after rubbing. In order to deeply explore the mechanism, LC cells were observed by polarizing optical microscope (POM) and PI film surfaces were investigated by atomic force microscopy (AFM), X-ray photo-electron spectroscope (XPS) and polarized attenuated total reflection Fourier transformed infrared spectroscopy (ATR-FTIR).

2. Experimental

2.1 Materials

4-Dodecyloxy-diphenyl-4′,4′′-diaminotriphenylamine (N12) was synthesized in our laboratory according to our previous work.²⁶ 2,2′-Bis-(trifluoromethyl)-4,4′-diaminodiphenyl (TFDB), pyromellitic dianhydride (PMDA) and N-methyl-2-pyrrolodone (NMP) were obtained from Aladdin Chemical Reagent Corp. (Shanghai, China). Prior to use, PMDA was purified by recrystallization from acetic anhydride and dried under vacuum. TFDB was purified by recrystallization from toluene. NMP (semiconductor grade, 99.9%) was used as received without further purification. 4,4′-Oxydiphthalic anhydride (ODPA) obtained from Shanghai Research Institute of Synthetic Resins was purified by a recrystallization from acetic anhydride. 4,4′-Oxidianiline (ODA) was provided by Zigong city Zhongtian wins new material technology Co., Ltd. (Zigong, China). Nematic LC E7 (n_o = 1.521, Δn = 0.22, T_n−1 = 60 °C) which is a mixture of four kinds of LC: 4-pentyl-4′-cyanobiphenyl (5CB), 4-heptyl-4′-cyanobiphenyl (7CB), 4-octyl-4′-cyanobiphenyl (8CB) and 4-cyano-4′′-n-pentyl-p-terphenyl (5CT) was purchased from Yantai Xianhua Chem-Tech Co., Ltd. (Yantai, China).

2.2 Polymer synthesis and sample preparations of PI films

The PIs were synthesized from PMDA, ODPA, TFDB, ODA and N12 in different ratios by a routine two-step procedure. Molar ratios of monomers and abbreviations of PIs were listed in Table 1. The polyamic acid (PAA) precursor solution was prepared by solution condensation polymerization at ambient temperature and at a concentration of 15 wt% by NMP. After being diluted, PAA was thermally imidized to obtain PI. The synthesis of SPI-20% was used as an example to illustrate the general synthetic route used to prepare the PIs. After all ODA (0.8 mmol) and N12 (0.2 mmol) were dissolved in NMP (3.27 g), ODPA (1.0 mmol) was added in an ice bath. The solution was stirred for 12 h at room temperature. Then NMP (7.70 g) was added and the solution was stirred for an additional 4 h to yield moderate viscous PAA solution. For the thermal imidization method, the PAA solution (5 wt% in NMP) was spin-coated on glass substrates (for LC cell assembly, XPS) and KBr tablets (for polarized ATR-FTIR), respectively, at a rotation speed of 600 rpm for 15 s and 2500 rpm for 30 s. The substrates were then baked at 80 °C for 5 min and 230 °C for 1 h to produce fully imidized PI films.

Table 1 Compositions and abbreviation of PIs

Monomer composition	Abbreviation of PIs	Mole ratio of monomers
PMDA : TFDB : N12	RPI-20%	10 : 8 : 2
	RPI-30%	10 : 7 : 3
	RPI-40%	10 : 6 : 4
ODPA : ODA : N12	SPI-20%	10 : 8 : 2
	SPI-30%	10 : 7 : 3
	SPI-40%	10 : 6 : 4

---

2.3 Fabrication of the LC cells

The prepared PI films were subsequently rubbed with a roller covered commercial rubbing cloth. Two pieces of the rubbed glass substrates were assembled together anti-parallel with respect to the rubbing direction by using 42 μm thick PI film spacers, injected with LC (E7), and then sealed with photo-sensitive epoxy glue, giving antiparallel nematic LC cells. For comparison, some LC cells were assembled from non-rubbed glass substrates using the same method. All LC cells were then heat-treated for 20 min at 80 °C, which is higher than the nematic-to-isotropic transition temperature of E7, to remove flow marks.

2.4 Measurements

The structures of PIs were confirmed by Fourier transform infrared spectra (FTIR) which were recorded on a Nicolet 560 (Madison, WI, USA) FTIR spectrophotometer. The pretilt angles of LC cells were measured by crystal rotation method using a pretilt angle tester from Changchun Institute of Optics, Fine Mechanics and Physics (Changchun, China), and at least five different points on cells were selected for measurement. Optical micrographs were obtained from POM (Shanghai Millimeter Precision Instrument Co. Ltd., Shanghai, China) under room temperature with the magnification of 40 times. Surface morphology of the PI films before and after rubbing was investigated by AFM operating in tapping mode using an instrument with a SPI4000 Probe Station controller (SIINT Instruments Co., Japan) at room temperature. Olympus tapping mode cantilevers with spring constants ranging from 51.2 N m⁻¹ to 87.8 N m⁻¹ (as specified by the manufacturer) were used with a scan rate in the range of 0.8–1.2 Hz. XPS measurements were made with a Kratos XSAM 800 spectrometer (Kratos Analytical Ltd., Manchester, U.K.) employing monochromatic Al K-alpha X-ray radiation (1486.6 eV). The take-off angle was fixed at 60° to ensure the detecting depth unchanged. Polarized ATR-FTIR spectra were recorded on a Nicolet 560 Fourier transform spectrometer using PI films coated on KBr tablets, a series of spectra were record by every 10 deg with the polarized angle varied from 0° to 180°.

3. Results and discussion

3.1 Polymer synthesis

RPIs and SPIs were prepared from PMDA, ODPA, TFDB, ODA and N12 via a conventional two-step procedure as illustrated in Fig. 2. Meanwhile, their structures were characterized by FTIR. In the spectra of RPI-40% and SPI-40% (Fig. 3), the peaks at 1778–1780 (C=O, asymmetric stretching) and 1723–1729 cm⁻¹ (C=O, symmetric stretching) are the typical frequencies of imides. And the C–N–C absorption at 1369–1377 cm⁻¹ (stretching vibration) also confirms the formation of imides. The appearance of these characteristic peaks of PI in the FTIR spectra has been expected to prove the synthesis of the SPI and RPI. Normally, the characteristic peaks of –CONH– (N–H stretching at 3200–3363 cm⁻¹ and C=O stretching at 1650–1674 cm⁻¹) will appear if the polymer is PAA. However, these peaks weren't observed which indicated that the PAAs had been fully imidized into PIs within the detection limit of FTIR.

Fig. 2 Synthetic scheme and chemical structures of PIs.

Fig. 3 FTIR spectra of RPI-40% and SPI-40%.

3.2 Alignment properties of LCs

The pretilt angles of LC cells were measured and the POM was used to verify the alignment capability of PI films (Fig. 4). For the LC cell whose pretilt angle was approximately 90°, a uniform darkness would be observed, and the dark crossed brush would be clearly seen and would not move with the LC cell rotating under conoscope, all of these proved that LC molecules were aligned vertically. On the other hand, for the LC cell whose pretilt angle was less than 20°, instead of a uniform darkness, an alternate and periodic appearance of darkness and brightness would be gained when the LC cells were rotated between crossed polarisers, which indicated homogeneous LC alignment.

Fig. 4 The POM graphics of LC cells fabricated by rubbed PI films: (A) RPI-20%, (B) RPI-30%, (C) RPI-40%; (D) SPI-20%, (E) SPI-30%, (F) SPI-40%. The POM graphics of the same LC cells after rotating the stage by 45°: (a) RPI-20%, (b) RPI-30%, (c) RPI-40%; (d) SPI-20%, (e) SPI-30%, (f) SPI-40%. The inset images are the conoscopic POM images of the corresponding LC cells.

From Table 2, it can be seen that the pretilt angles are approximately 90° whether the RPI-X% (X = 20, 30, 40) films were rubbed or not. But for SPI-X% (X = 20, 30, 40), initially, the PI film without rubbing even couldn't induce LCs alignment when the content of functional diamine was 20%. After the content of N12 increased to 30%, vertical alignment of LC could be observed. However, pretilt angles of SPI-20% and SPI-30% couldn't achieve 90° after rubbing. Only when the content of N12 increased to 40% could they have the same excellent vertical alignment capability as RPI-X%. To the SPI-X%, vertical alignment capability was improved with the content of N12 increasing, which revealed that the increase of side chains contributed to improving the alignment ability of LC molecules and rubbing resistance of vertical alignment films. The PIs (SPI-20% and RPI-20% or SPI-30% and RPI-30%) which had the same content of side chain but different main chains showed totally different LC alignments. These phenomena suggested that the backbone structure of PI played a vital role in the alignment ability of LC molecules. All the above results confirmed that not only the content of side chain but also the backbone structure of PI had effects on the LC vertical alignment.

Table 2 LC pretilt angles on various PIs

Samples	Pretilt angle θp (°)
	Unrubbed	Rubbed
RPI-20%	89.9°	89.6°
RPI-30%	89.9°	89.9°
RPI-40%	90°	89.6°
SPI-20%	No alignment	7.1°
SPI-30%	90°	12.4°
SPI-40%	89.9°	89.1°

---

3.3 Surface morphology of alignment layers

The surface morphology of PI film was examined by AFM in details before and after they had been rubbed. Fig. 5 shows typical topographical images of non-rubbed and rubbed PI films. During the rubbing process, microgrooves appeared because of the deformation responses to the shear forces produced by contact with fibers. Compared to the rubbed films which exhibited groove-like structures, the non-rubbed films showed much smoother surfaces with relatively smaller roughness. The root-mean-square (RMS) surface roughness of non-rubbed SPI-30% and RPI-30% films were 0.405 and 0.308 nm, respectively. While, the RMS surface roughness of rubbed SPI-30% and RPI-30% films were 0.860 and 0.844 nm, respectively. Besides, the size of microgrooves was over 40 nm and larger than that of LC molecules, which implied that microgrooves couldn't interact with LC molecules sufficiently. Thus, it was believed that the microgroove was not the main reason of LC molecules alignment. Actually, there were not big differences between the surface morphology of the rubbed SPI-30% and RPI-30% films and their RMS were similar. However, the two rubbed films induced definitely different LC alignment. So we speculate that the interactions between chemical groups of PI surface and LC molecules like π–π and van der Waals interactions have great influence on LC alignment.²⁷ In order to confirm our speculation, the surface of PI films was further determined by XPS and polarized ATR-FTIR next.

Fig. 5 The AFM images and surface profiles of PI films: (A) non-rubbed SPI-30% film; (B) rubbed SPI-30% film; (C) non-rubbed RPI-30% film; (D) rubbed RPI-30% film; (a), (b), (c) and (d) surface profiles taken along the red line in the AFM images (A), (B), (C) and (D), respectively. The arrows in the AFM images denote the rubbing direction.

3.4 Chemical composition of PI surfaces

XPS spectra were adopted to analyze the chemical composition of PI surfaces before and after rubbing. As shown in Fig. 6(A) and (a), the peak at 284.4 eV was assigned to the C–C/C=C. The peaks at 285.8 eV, 288.8 and 292.5 eV were due to C–O/C–N, C=O and C–F, respectively.^27,28 For RPI-30%, C–C/C=C changed from 41.92% to 41.60% (−0.32%), C–O/C–N changed from 41.27% to 41.70% (+0.43%), C=O changed from 11.76% to 11.63% (+0.13%) and C–F changed from 5.05% to 5.07% (−0.02%). These intervals were very little within the error range, which indicated that the chemical composition of alignment layer's outer layer almost had no variations (Table 3). The corresponding curves of SPI-30% were shown in Fig. 6(B) and (b). Compared with the changes of RPI-30%, all peaks of SPI-30% had relative obvious: C–C/C=C decreased from 70.57% to 67.63% (−2.94%), C–O/C–N increased from 23.19% to 24.93% (+1.74%) and C=O increased from 6.24% to 7.44% (+1.2%). From Fig. 2, it could be clearly seen that all C–C came from the alkoxy of side chains and both C=O and C–N only existed in the main chain. Besides, most C–O belonged to main chain because the molar content of side chains was only 30%. Thus, we concluded that the content decrease of C–C was induced by the movement of side chain and the content increase of C–O, C–N and C=O was attributed to exposure of main chains. The result signified that the content of main chains in the alignment layer's outer layer increased after rubbing.

Fig. 6 The curve-fitted XPS C1s spectra of non-rubbed RPI-30% (A), rubbed RPI-30% (a), non-rubbed SPI-30% (B) and rubbed SPI-30% (b).

Table 3 Results of the peak separation of XPS C1s spectra

PI	Atomic percentage (%)
	C–C/C=C	C–O/C–N	C=O	C–F
a There were not C–F bonds in the SPI.
Non-rubbed RPI-30%	41.92	41.27	11.76	5.05
Rubbed RPI-30%	41.60	41.70	11.63	5.07
Non-rubbed SPI-30%	70.57	23.19	6.24	—^a
Rubbed SPI-30%	67.63	24.93	7.44	—^a

---

As reported in relevant literature, side chains tend to aggregate in the outer layer before rubbing which results in the “enrichment” of side chain on the surface.²⁹ What's more, it benefits the liquid crystal to align vertically on the PI surface with a large pretilt angle. However, side chains may insert into polymer bulk when undergo the rubbing process because of the mobility of molecular chain. Then more main chains will be exposed on the surface thus result in the increase of main chain content in the alignment layer's outer layer. In addition, it has been widely reported that there are many intermolecular interactions existing between LC molecules and side chains of PIs because of the structural similarity of chemical groups on the surface and LC molecule. In our case, they are the dipole–dipole/π–π interaction between the aromatic units of LC molecules and the biphenyl groups of side chains, and the van der Waals forces between the alkoxy end-group of PIs and the aliphatic tail in LC molecules. As the distribution of side chains in PI surfaces varying, we speculate that if side chains adopt vertical orientation, LC molecules will be aligned vertically because of the intermolecular interactions. Likewise, if side chains adopt parallel orientation, LC molecules will be aligned parallel. Notwithstanding, these speculations have to be further confirmed by polarized ATR-FTIR spectroscopy.

3.5 Molecular reorientation

The orientations of molecular chains in the surface of rubbed PI films were further investigated by polarized ATR-FTIR spectroscopy. Fig. 7 presents two ATR-FTIR spectra of rubbed SPI-30% and RPI-30% film which were measured with light polarized at 0° (parallel to the plane of substrate) and 90° (normal to the plane of substrate). For RPI-30%, the peaks at 1726 and 1369 cm⁻¹ were assigned to C=O and C–N–C stretching vibration of imide ring in main chains. And the peak at 1492 cm⁻¹ was associated with C–C stretching vibration of benzene rings skeleton of 1,2,4,5-tetrasubstituted C–C and 1,2,4-trisubstituted C–C in main chains.^30,31 The peaks at 1505 and 1260 cm⁻¹ were assigned to the 1,4-disubstituted C–C stretching vibration of the biphenyl group and –O– stretching vibration of alkoxy group in side chains.^21,32 Similarly to RPI-30%, peaks of SPI-30% at 1722, 1378 and 1500 cm⁻¹ were attributed to C=O and C–N–C stretching vibration of imide rings, and C–C stretching vibration of benzene rings skeleton, respectively. In addition, the characteristic peak of –O– split into three peaks at 1275, 1260 and 1244 cm⁻¹. They were related to –O– of ODPA, N12 and ODA which were structural units of SPI-30%, respectively. It can be clearly seen that the spectra with light polarized at 0° and 90° did not coincide for both RPI-30% and SPI-30%. The results revealed that the chemical groups in the surfaces of rubbed PI films are anisotropic.

Fig. 7 ATR-FTIR dichroic spectra of rubbed SPI-30% (a) and rubbed RPI-30% (b). The solid and dashed lines represent the infrared spectra with light polarized at 0° (parallel to the plane of substrate) and 90° (normal to the plane of substrate), respectively.

The variations of the selected absorption peaks of RPI-30% with the angle of polarization of incident polarized light are plotted as polar diagrams (Fig. 8). The polar diagrams of absorption peaks at 1369 and 1492 cm⁻¹ exhibited maximum intensity along the direction 0° ↔ 180° and minimum intensity along the direction 90° ↔ 270°. Because the imide ring, 1,2,4,5-tetrasubstitutedbenzene and 1,2,4-trisubstitutedbenzene lie in the direction of main chain axis, we can consider that the orientation of main chain was parallel to substrate. On the other hand, the polar diagrams of absorption peaks at 1726, 1505 and 1260 cm⁻¹ exhibited maximum intensity along the direction 90° ↔ 270° and minimum intensity along the direction 0° ↔ 180°, which meant that the C=O, biphenyl group and –O– tended to adopt vertical orientation.²¹ As mentioned in our previous article, the C=O deviated from backbone axis although it was a part of main chain and the biphenyl group and –O– were key components of side chains.³² As a result, it was logical that the C=O was preferentially reoriented normal to the plane of substrate as well as side chains.

Fig. 8 Polar diagrams of some specific absorption peaks of rubbed RPI-30% as a function of the angle of polarization of incident polarized light, measured with linearly polarized IR spectroscopy. In main chain: (a) 1369 cm⁻¹ (C–N–C of imide rings), (b) 1492 cm⁻¹ (C–C of benzene) and (c) 1726 cm⁻¹ (C=O of imide rings); in side chain: (d) 1505 cm⁻¹ (C–C of the biphenyl) and (e) 1260 cm⁻¹ (–O– of N12).

Fig. 9 shows polar diagrams of absorption peaks of SPI-30% with respect to the angle of polarization of incident polarized light. The polar diagrams at 1378 and 1500 cm⁻¹ exhibited maximum intensity along the direction 0° ↔ 180° and minimum intensity along the direction 90° ↔ 270°; and the polar diagrams of absorption peaks at 1722 cm⁻¹ exhibited maximum intensity along the direction 90° ↔ 270° and minimum intensity along the direction 0° ↔ 180°. These results were consistent with the results of RPI-30%, which suggested that the main chain preferred to orient parallel to substrate. On the other hand, the polar diagrams at 1275, 1260 and 1244 cm⁻¹ (–O– of ODPA, N12 and ODA, respectively) exhibited maximum intensity along the direction 0° ↔ 180° and minimum intensity along the direction 90° ↔ 270°, which meant that all –O– bonds tended to adopt parallel orientation no matter which chain they belonged to. The results revealed that both the main chain and the side chain were preferentially reoriented parallel to the plane of substrate.

Fig. 9 Polar diagrams of some specific absorption peaks of rubbed SPI-30% as a function of the angle of polarization of incident polarized light, measured with linearly polarized IR spectroscopy. In main chain: (a) 1500 cm⁻¹ (C–C of benzene), (b) 1378 cm⁻¹ (C–N–C of imide rings) and (c) 1722 cm⁻¹ (C=O of imide rings); in side chain: (d) 1275 cm⁻¹ (–O– of ODPA), (e) 1260 cm⁻¹ (–O– of N12) and (f) 1244 cm⁻¹ (–O– of ODA).

Simply put, although the average orientations of backbones of rubbed RPI-30% and rubbed SPI-30% are both vertical, the average orientations of their side chains are vertical and parallel, respectively. It can be concluded that side chains of SPI-30% fell over after rubbing and adopted parallel orientation. In our last article, it was reported that the side chain would fall over after rubbing if they were linked to the backbone with flexible ether bond spacers. However, in this study, side chains were directly linked to main chains without any spacer. So the reason why side chains fall over should be attributed to the flexibility of main chains. There were many flexible ether bonds in the soft main chains which can move easily during the rubbing process and the movement of them caused side chains to fall over. Combined with the results above, we can conclude that LC molecules tend to be aligned in the direction parallel to the side chains with the aid of intermolecular interactions thereby generating alignment of LCs. And rigid main chain is more conducive to achieving stable vertical alignment of LCs than soft main chain.

Taken together, a generation mechanism of the different LC alignments for the PIs with different main chains can be proposed, which is shown in Scheme 1. For RPI, the rigid main chains restrict the mobility of side chains; for SPI, the soft main chains have many flexible ether bands which can move easily during the rubbing process. In order to make it easier for readers to understand, a simplified analogy can be made. The biphenyl unit of side chain is just like the rigid nail which is fixed in the wooden board (rigid main chain) or sponge (soft main chain). If an external force was applied to the rigid nail, the wooden board would be hard to deform but the sponge would be easy to deform and the rigid nail was titled along the external force.

Scheme 1 Schematic mechanism of the macroscopic molecular orientation of PIs surfaces before and after rubbing processes.

4. Conclusions

Two series of PIs processing the same side chain but different main chains (rigid main chain and soft main chain) have been synthesized. RPI-X (X = 20, 30, 40%) and SPI-40% had high pretilt angles about 89.9° whether they were rubbed or not. Before rubbing, SPI-20% couldn't induce homogeneous alignment of LC and SPI-30% induced vertical alignment of LC. While, both of them had low pretilt angles after rubbing, this meant that they induced parallel alignments of LC. From the perspective of SPI, the content of side chains has great influence on the aligning ability of LC (aligned or not aligned) before rubbing and aligning behavior of LC (parallel or vertically) after rubbing. Because the wholly rigid aromatic units in RPI's main chains could restrict the movement of molecular chains, the side chains can maintain vertically orientation all the time. On the other hand, because flexible ether bonds in SPI's main chains could endow molecular chains with mobility, the side chains easily fall over after rubbing. The results confirmed that PI with rigid main chain is more beneficial to making LC molecules aligned vertically and stably.

Acknowledgements

This work was supported by National Natural Science Foundation of China (Grant no. 51173115), the Ministry of Education (the Foundation for Ph.D. training, Grant no. 20110181110030) of China. We acknowledge Prof. Mingbo Yang for characterization.

References

1. K. H. Kim and J. K. Song, NPG Asia Mater., 2009, 1, 29.
2. Y. J. Lee, Y. K. Kim, S. I. Jo, J. S. Gwag, C. J. Yu and J. H. Kim, Opt. Express, 2009, 17, 10298.
3. Q. Hong, T. X. Wu, X. Zhu, R. Lu and S. T. Wu, Appl. Phys. Lett., 2005, 86, 121107.
4. J. J. Lyu, J. Sohn, H. Y. Kim and S. H. Lee, J. Disp. Technol., 2007, 3, 404.
5. R. Lu, S. T. Wu and S. H. Lee, Appl. Phys. Lett., 2008, 92, 051114.
6. C. Cai, A. Lien, P. Andry, P. Chaudhari, R. John, E. Galligan, J. Lacey, H. Ifill, W. Graham and R. Allen, Jpn. J. Appl. Phys., 2001, 40, 6913.
7. K. Sakamoto, R. Arafune, N. ITo, S. Ushioda, Y. Suzuki and S. Morokawa, Jpn. J. Appl. Phys., 1994, 33, 1323.
8. J. Stöhr and M. G. Samant, J. Electron Spectrosc. Relat. Phenom., 1999, 98–99, 189.
9. J. J. Ge, C. Y. Li, G. Xue, I. K. Mann, D. Zhang, S. Y. Wang, F. W. Harris, S. D. Cheng, S. C. Hong, X. Zhuang and Y. R. Shen, J. Am. Chem. Soc., 2001, 123, 5768.
10. S. W. Lee, S. I. Kim, Y. H. Park, M. Reea, Y. N. Rim, H. J. Yoon, H. C. Kim and Y. B. Kim, Mol. Cryst. Liq. Cryst., 2000, 349, 279.
11. B. Chae, S. Kim, S. Lee, S. Kim, W. Choi, B. Lee, M. Ree, K. Lee and J. Jung, Macromolecules, 2002, 35, 10119.
12. Y. J. Lee, Y. W. Kim, J. D. Ha, J. M. Oh and M. H. Yi, Polym. Adv. Technol., 2007, 18, 226.
13. S. I. Kim, M. Ree, T. J. Shin and J. C. Jung, J. Polym. Sci., Part A: Polym. Chem., 1999, 37, 2909.
14. D. S. Seo, Liq. Cryst., 2000, 27, 883.
15. J. Stöhr, M. G. Samant, A. Cossy-Favre, J. Díaz, Y. Momoi, S. Odahara and T. Nagata, Macromolecules, 1998, 31, 1942.
16. S. G. Hahm, T. J. Lee, T. Chang, J. C. Jung, W. C. Zin and M. Ree, Macromolecules, 2006, 39, 5385.
17. B. Lee, G. Shin, J. Chi, W. Zin, J. Jung, S. Hahm, M. Ree and T. Chang, Polymer, 2006, 47, 6606.
18. Y. Lee, J. Choi, I. Song, J. Oh and M. Yi, Polymer, 2006, 47, 1555.
19. B. Chae, S. Lee, B. Lee, W. Choi, S. Kim, Y. Jung, J. Jung, K. Lee and M. Ree, Langmuir, 2003, 19, 9459.
20. X. Wang, P. Zhang, Y. Chen, L. Luo, Y. Pang and X. Liu, Macromolecules, 2011, 44, 9731.
21. J. Wang, L. Wang, Y. Zeng, Y. Fang, Q. Zhang and Y. Wang, Liq. Cryst., 2010, 37, 271.
22. R. Arafune, K. Sakamoto and S. Ushioda, Appl. Phys. Lett., 1997, 71, 2755.
23. D. S. Seo, J. Appl. Phys., 1999, 86, 3594.
24. Y. T. Chern and M. H. Ju, Macromolecules, 2009, 42, 169.
25. D. Tanaka, Y. Nagashima, M. Hara, S. Nagano and T. Seki, Langmuir, 2015, 31, 11379.
26. M. Liu, X. Zheng, S. Gong, L. Liu, Z. Sun, L. Shao and Y. Wang, RSC Adv., 2015, 5, 25348.
27. H. Kang, J. S. Park, E. Sohn, D. Kang, C. Rosenblatt and J. Lee, Polymer, 2009, 50, 5220.
28. T. Solomun, A. Schimanski, H. Sturm and E. Illenberger, Macromolecules, 2005, 38, 4231.
29. X. Wang, H. Wang, L. Luo, J. Huang, J. Gao and X. Liu, RSC Adv., 2012, 2, 9463.
30. H. Ishida, S. T. Wellinghoff, E. Baer and J. L. Koenig, Macromolecules, 1980, 13, 826.
31. R. Nuzzo, L. Dubois and D. Allara, J. Am. Chem. Soc., 1990, 112, 558.
32. X. Che, S. Gong, H. Zhang, B. Liu and Y. Wang, Phys. Chem. Chem. Phys., 2016, 18, 3884.

---