1. Introduction

The optically compensated bend (OCB) mode has recently attracted much attention because it exhibits a fast response time and wide viewing angle [1,2]. Operationally, liquid crystals (LCs) must initially transform from the initial splay state to the bend state in the OCB mode. Then, the bright and dark states are a bend state and a homeotropic state, respectively. The transition between these two states occurs rapidly. Additionally, the viewing angle of OCB-LCD is large because of its self phase compensation. However, the transformation from the initial splay state to the bend state is slow. Therefore, a bias voltage must be applied to maintain the cell in the bend state. Many works have found intermediate pre-tilt angles of ~47° from the substrate surface, to fabricate a unbiased OCB cell [8–10].

Given the need to save time and for the mass production of displays, the method of controlling the pre-tilt angle of LCs in an unbiased OCB cell by the evaporation of silicon monoxide (SiOx) is inappropriate [3]. Surface-induced alignment has been suggested to yield intermediate pre-tilt angles. These new approaches include mechanical rubbing of polyimide (PI) [4,5], mixing vertical and horizontal polyimides [6–9], the use of dual alignment layers [10] and the photopolymerization of reactive monomers [11]. The pre-tilt angle can be controlled using the above approaches from about 20° to 80°. However, some approaches are associated with many technical difficulties and disadvantages to be overcome, including alignment stability, high cost of manufacture and so on. The mixture of two PIs is preferred because it easily yields intermediate pre-tilt angles with conventional rubbing without changing the normal fabrication process.

Recently, Yeung et al. described an approach for obtaining intermediate pre-tilt angles based on a mixture of horizontally and vertically align PIs (H+V PI mixture). The mixture was homogeneously mixed, and then underwent phase separation to form nano-sized domains [6,7]. This work also demonstrates a simple approach, based on an H+V PI mixture to obtain an arbitrary pre-tilt angle (~15° to 85°) by controlling the H- to V-PI concentration ratio, the baking temperature and the rubbing strength. Unlike the PI mixture of Yeung et al., the materials adopted herein do not undergo noticeable phase separation as the solvent evaporates, and the alignment layer does not seem to form nano-sized domains, as presented in Fig. 1, which presents the scanning electron microscope (SEM) image of the alignment layer. Experimental results show that an arbitrary pre-tilt angle from ~15° to 85° can be achieved. A variable polarization converter is fabricated by rubbing a mixed PI film with a varying rubbing strength. The polarization of linearly polarized incident light can be converted continuously as the light emerges from the device.

 

Fig. 1. SEM images of (a) rubbed and (b) non-rubbed substrate coated with mixed PI.

Download Full Size | PDF

2. Experiments

The materials adopted herein were horizontal polyimide (H-PI) AL-1426B (from Daily Polymer Corporation) and vertical polyimide (V-PI) AL-00010 (from Japan Synthetic Rubber Company). The concentration ratios of H- to V-PI, the baking temperature, and the strength of rubbing on the substrates were controlled to prepare the substrates of the cell. The homogeneously mixed PI compound was coated on an indium-tin-oxide (ITO) glass substrate by spin coating. After they had been baked and rubbed, two substrates that had been treated identically were assembled to produce an empty anti-parallel LC cell with a ~12 um gap. An empty cell was finally filled with K15 liquid crystal (Merck) to form a LC cell. The pre-tilt angle of an LC cell, which is defined as the angle made between the LC director and the 500nm substrate surface, was measured by fitting the measurements obtained using the crystal rotation method [12] with self-developed Labview software. The measured angle was doubly checked by plotting the measured transmission versus voltage (T-V) curve of the cell. The pre-tilt angle can be obtained by fitting the T-V curve using Dimos software. Finally, a polarization converter was fabricated by applying a varying rubbing strength.

The T-V curve was measured as follows. A light beam from an He-Ne laser (λ=632.8 nm) was passed through the LC sample that was placed between two polarizers, which were oriented at ±45° with respect to the rubbing direction of the sample. The transmission-voltage was then measured.

3. Results and discussion

Three approaches for controlling the pre-tilt angle of an LC cell are investigated herein; they are the control of the concentration ratio of V- to H-PI, the control of the baking temperature, and the control of rubbing strength on the substrates. The rubbing strength is defined by the pile impression, which represents the distance from the flannel to the surface of the substrate measured using the scale provided in the rubbing machine. The pile impression is zero when the flannel just touches the surface of the substrate.

3.1 Variation of V-PI to H-PI concentrations

The PI-coated substrate was treated at a baking temperature of 200 °C for 1 h and a pile impression of ~100 µm was produced to fabricate an empty LC cell. Various concentration ratios of V- to H-PI were adopted to prepare the PI that was coated on the substrate. Figure 2 plots the measured variation of the pre-tilt angle with the concentration of the V-PI. The result indicates that the pre-tilt angle increases monotonically with the increasing concentration of the V-PI. A wide range of pre-tilt angles (~20° to 60°) can be obtained by varying the V-PI concentration from ~3.57 to 4.55 wt%. Since higher concentrations of H-PI give rise to lower concentrations of side chains that are associated with vertical PI, the pre-tilt angle reasonably declines as the concentration of the horizontal PI increases [8].

 

Fig. 2. Variation of pre-tilt angle as a function of concentration of V-PI in the mixture.

Download Full Size | PDF

3.2 Baking temperature effect

A substrate that was coated with a mixed PI layer with a V-PI concentration of ~4.55 wt% (i.e. H-PI ~95.45 wt%) was baked at various temperatures from 180 to 240 °C for 1 h. It was then rubbed with a pile impression of ~0.1 µm, representing the flannel’s just touching the surface of the substrate, causing the cell to have a tilt-angle in the same direction. Figure 3 plots the measured variation of the LC pre-tilt angle with the baking temperature. It indicates that a wide range of pre-tilt angles (~15° to 75°) can be obtained using this approach, and the pre-tilt angle declines monotonically as the baking temperature increases, for the following reason. Baking temperatures from 180 to 240 °C are higher than that prescribed (180 °C) by the manufacturer for the vertical alignment of AL-00010. Such over-baking in the polyimide has two effects; 1) it causes the further imidization of the backbones of V-PI, promoting the planar alignment; 2) it cleaves away a proportion of the side chains of the V-PI (AL-00010) component, weakening the vertical alignment. Therefore, the pre-tilt angle declines as the baking temperature increases [8].

 

Fig. 3. Variation of pre-tilt angle as a function of baking temperature.

Download Full Size | PDF

3.3 Rubbing strength effect

The substrate coated with a mixed PI layer, with a concentration of V-PI ~4.55 wt%, was baked at 200 °C for 1 h, and was rubbed using various pile impressions. Figure 4 plots the measured variation of the LC pre-tilt angle with the rubbing strength. Figure 4 indicates that the pre-tilt angle declines monotonically as the rubbing strength increases. Pre-tilt angles from about 15o to 90° are obtained by varying the rubbing strength. This result is reasonable, since the pre-tilt angle is well known to drop as the rubbing strength increases. The liquid crystals tend to align in two preferred directions - one is approximately vertical because of the V-PI side chains, and the other is approximately planar because of the rubbed horizontal polyimide. The liquid crystals align in a direction that is determined by the equilibrium between the two orthogonal easy axes. For weak rubbing, the V-PI side chains dominate, and the formed LC cell is homeotripical, θ~90°. As the rubbing strength increases, the effect of the planar easy axis increases, reducing the pre-tilt angle of the cell.

 

Fig. 4. Variation of pre-tilt angle as a function of rubbing strength.

Download Full Size | PDF

3.4 Application: Fabrication of a polarization converter

In this study, an LC cell was fabricated for use as a polarization converter, as presented in Fig. 5. Notably, LCs on the bottom substrate of the cell are aligned at continuously varying tilt angles, while those on the top substrate are vertically aligned. This cell was fabricated as follows. The top substrate was coated with a vertical layer, and the bottom substrate was coated with a mixed PI layer with a V-PI concentration of ~4.55 wt%, and baked at 200 °C for 1 h. It was then rubbed along the x axis with increasing rubbing strength with pile impressions from 0 to 350 µm.

 

Fig. 5. LC director configurations in cell with one substrate rubbed with increasing strength.

Download Full Size | PDF

Figure 6 presents images of the LC cell under an optical polarized microscope (OPM) with a white light source. In the figure, β is the angle made between the rubbing direction and the polarization of the incident beam. The cell in the image in Fig. 6(a) obtained under a parallel-polarizer OPM is fully bright. This result is reasonable, since the beam that emerges from the LC cell has the same polarization as the incident beam, and is transmitted through the analyzer. As expected, the cell is completely dark observed under the crossed-polarizer OPM, as presented in Fig. 6(b). Figures 6(c) and 6(d) plot the effects of phase retardation with β=45° between the parallel-polarizer and the crossed-polarizer conditions, respectively. The color image is produced when white light passes the cell with various phase retardations. Notably, the right regions in Figs. 6(c) and 6(d) are bright and dark, respectively, because LCs are approximately vertically aligned in these regions.

 

Fig. 6. Fabricated variable polarization converter observed under an optical polarized microscope with (a) P‖A, β=0°; (b) P⊥A, β=0°; (c) P‖A, β=45°; (d) P⊥A, β=45°. β is the angle between the rubbing direction and the polarization of the incident beam.

Download Full Size | PDF

To verify the results presented in Fig. 6, the cell with the structure in Fig. 5 was simulated using Jones Matrix method. It is assumed that the light is normally incident onto the uniaxial sample, and the birefringence, |n _e-n _o|≪n _e,n _o in this formulism (n_e and n_o are the extra-ordinary and ordinary refractive indices of a uniaxial medium, respectively) [13]. The pre-tilt angle on the bottom substrate is assumed to be 15° on one side, increasing continuously to 90° on the other side. The conditions presented in Figs. 6(c) and 6(d) were used to simulate the incidence of light of three wavelengths - 450, 550 and 650 nm onto the cell. The reason to perform simulations with three different wavelengths of incident light instead of using white light is for simplifying the calculation. Since the phase retardation, Γ=2πdΔn/λ (d is the thickness of the LC layer, Δ n is the LC birefringence, and λ is the wavelength of the probe beam), which determines the polarization of the transmitted light is a function of wavelength, the use of a single wavelength of light simplifies the calculation. Also, the mixing of these three wavelengths of light gives the white light.

Figure 7 presents the comparisons between experimental (Fig. 6(c)) and simulated results under P‖A and β=45° conditions. As seen, region 1 shows reddish orange color. It results from the mixture of rich red with weak green color. Similar argument is applied to regions 2–5. The experimental results are consistent with the simulated ones.

 

Fig. 7. Simulation of LC polarization converter cell with structure presented in Fig. 5 at P‖A, β=45°, with probe beams at wavelengths of 450, 550 and 650 nm. The pre-tilt angle on the bottom substrate is assumed to be 15° on one side, increasing continuously to 90° on the other side.

Download Full Size | PDF

The light transmittance of the LC cell that was placed between two crossed polarizers was measured using an He-Ne laser (λ=632.8 nm) to further confirm that the cell indeed functions as a variable polarization converter. The polarizer axis was at an angle of 45° with the rubbing direction, and a linearly polarized laser beam was incident normally on the cell from the side of the vertical alignment with its polarization making an angle 45° to the x axis. Again, the polarizations of the transmitted lights are determined by the phase retardation.

The dots in Fig. 8 plot the measured position-dependent transmittance of the LC cell. The position is defined from the left to the right side of the sample, as presented in Fig. 6(d). Because continuously varying tilt angles induced the varying LC birefringenceΔ n, the transmittance through the analyzer changes with the phase retardation along the x axis. An ideal transmittance curve for tilt angles from ~15 to 90° was simulated using the Jones Matrix formulism. Figure 8 gives the results, and clearly indicates that the experimental results are consistent with the simulated results. The error is attributed to the finite spot size (~1 mm) of the probe laser beam.

 

Fig. 8. The transmittance of LC polarization converter, placed between two crossed polarizers with polarization of probe beam at an angle of 45° from rubbing direction, as a function of position of laser beam. The dots and the dotted line represent the experimental and simulated results, respectively.

Download Full Size | PDF

4. Conclusions

In conclusion, this work demonstrates three approaches to controlling the LC pre-tilt angles (~15° to 85°) in a cell by varying the concentration ratio of H- to V-PI, the baking temperature and the rubbing strength. The process is simple, and is compatible with the currently existing manufacturing process. Additionally, a variable-polarization converter, based on the LC cell that is presented in Fig. 5 was fabricated with one substrate rubbed with increasing strength. The polarization of a polarized beam incident on the cell can be converted continuously upon emergence from the device.