1. Introduction

The increase in the demand for green displays imposes constraints on the display power consumption. Thus, novel liquid crystal displays (LCDs) such as reflective and bistable LCDs have attracted the attention of researchers from a wide range of applied fields. Reflective LCDs could minimize the power consumption under ambient light because it does not require a backlight unit [1–5]. Bistable LCDs could also reduce power consumption owing to their intrinsic memory characteristics under the absence of external ðeld [6–9]. Consequently, reflective bistable LCDs, i.e., displays with both reflective and bistable characteristics, could drastically curtail power consumption [10,11]. Recently, a reflective dual-mode liquid crystal display (RD-LCD) was proposed, which can operate memory or dynamic mode by selective switching in a reflective display panel [12]. The memory mode is used for still images or text messages with low power consumption. On the other hand, the dynamic mode is relevant in the case of monitor or television applications for high-quality moving images. The proposed RD-LCD utilized three electric fields. In the memory mode, a vertical electric field or a bottom fringe electric field were used for choosing one of the two stable states. In the dynamic mode, adequate upper fringe electric field was required to realize gray scales. However, complex process on upper and bottom substrates and slow response time in the dynamic mode rendered it difficult for real fabrication. Furthermore, liquid crystal (LC) configuration, splay state as a common dark state, in a cell thickness over pitch (d/p) of 0.25 for long memory retention time could cause the quality of contrast ratio (CR) to deteriorate [13].

Herein, we propose a RD-LCD which consists of $\pi$ twist state as a common dark state in both memory and dynamic modes, which can overcome the aforementioned disadvantages. An optical configuration for the $\pi$ twist state makes low light leakages at dark state, utilizing ideal bistable characteristics, d/p = 0.25. In addition, it is designed to utilize two electric fields, a vertical electric field and a fringe electric field. In the memory mode, a vertical or a fringe electric field is used for choosing one of the two stable states. In the dynamic mode, an adequate vertical electric field is utilized to realize gray scales.

2. Principle of operation and optical configurations

To utilize two electric fields, i.e., the vertical and fringe electric field, a three-terminal electrode structure is necessary, as shown in Fig. 1 . The upper substrate consists of a top electrode for vertical or horizontal switching and patterned electrodes for horizontal switching. An insulator is laid between the top and the patterned electrodes to prevent electrical shorts. The lower substrate consists of a bottom electrode for vertical switching. Initial LC configuration in the three-terminal electrode structure is also illustrated in Fig. 1. When appropriate chiral additive with d/p = 0.25 is injected into a splay cell, spontaneous degeneracy is inevitably formed; it forms $\pi$ twist domains and splay domains of equal probabilities [14,15].

 

Fig. 1 A three-terminal electrode structure and initial LC configuration in d/p = 0.25.

Download Full Size | PDF

Figure 2 shows the transition process of the chiral-splay cell in d/p = 0.25. When a vertical electric field is applied between the top and bottom electrode, the coexisted states (Fig. 2(a)) are changed to a bend state (Fig. 2(b)). After the vertical electric field is removed, the bend state is relaxed to a $\pi$ twist state (Fig. 2(c)). It results from topological equivalence [14,15]. Throughout the whole region in the LC cell, the π twist state is uniformly formed. To convert the $\pi$ twist (Fig. 2(c)) into a splay state, a fringe electric field is needed. When the fringe electric field is applied between the top and patterned electrodes, the $\pi$ twist is converted into splay state coexisted with the $\pi$ twist domains (Fig. 2(d)), unlike uniform splay state (Fig. 2(e)). The fringe electric field only influences local region, i.e. near the edge of patterned electrode, because fringe field intensity is maximal near the edge of the patterned electrode but rapidly drops to zero above the center between the electrodes [16]. In addition, the ideal 0.25 d/p characteristics prevent the propagation of disclination lines [17,18]. Hence, the LCs near the edge of the pixel electrodes are only converted into splay state; however, the LCs above the center of the two electrodes remain in the $\pi$ twist state. This non-uniform splay state (Fig. 2(d)) can cause light to leak if it is used as a dark state. To overcome this problem, we considered an optical configuration which use the $\pi$ twist state (Fig. 2(c)) as a dark state.

 

Fig. 2 The transition process of the chiral-splay cell in the vicinity of d/p = 0.25: (a) initial coexisted states, (b) bend state by vertical electric field, (c) $\pi$ twist state after relaxation, (d) splay state coexisted with the $\pi$ twist domains and (e) uniform splay state by fringe electric field.

Download Full Size | PDF

By means of Jones matrix method, we tried to convert $\pi$ twist state to optically compensated dark state. The following equation is the Jones matrix used to obtain optically compensated $\pi$ twist state [19].

By calculating above equation, we found a design parameter which is comprised a polarizer with 0° transmission axis (TA), 250 nm-thick positive a-plate with a 48° slow axis (film 1), 60 nm-thick positive a-plate with a 0° slow axis (film 2), 484.1 nm-thick retardation twist LC cell with a 108° rubbing direction, and a reflector. It is shown in Fig. 3(a) .

 

Fig. 3 Operation of the proposed device: (a) optically compensated $\pi$ twist state, (b) dynamic mode by vertical switching, (c) memory mode by horizontal switching.

Download Full Size | PDF

The RD-LCD can operate dynamic or memory mode by selective switching. The optically compensated $\pi$ twist state can be used as a common dark state in both the memory and dynamic modes. The $\pi$ twist state is changed to a bend state when a vertical electric field is applied between the top and bottom electrode. (see Figs. 3(a) and 3(b)) After the vertical electric field is removed, the bend state always returned to the initial $\pi$ twist state (Fig. 3(a)). This vertical switching is utilized in the dynamic mode. On the other hand, when a fringe electric field is applied between the top electrode and the patterned electrodes and then removed, the $\pi$ twist state is changed to the splay state coexisted with the $\pi$ twist domains (see Figs. 3(a) and 3(c)). Each of the π twist and splay state has theoretically infinite memory when the d/p is 0.25. This horizontal switching between the $\pi$ twist and the splay state is used in the memory mode. To return to the initial $\pi$ twist state, a vertical electric field is necessary once more.

Figure 4 shows the polarization paths of three lights, 450 nm (blue light), 550 nm (green light), and 650 nm (red light) on the Poincare sphere. The Poincare sphere representation is introduced to explain how the $\pi$ twist state becomes a dark state. The initial position, S1, represents a 0° linear polarization after passing the polarizer with 0° TA, as shown in Fig. 4(a). The three lights located at S1 are shifted by film 1, and then arrived at different positions around –S1. According to wavelength, each of the three lights has different phase retardations, thereby having different positions. Each of three lights is dispersed again by film 2. All of these are shown in Fig. 4(b). Figure 4(c) shows that each position of the three lights is relocated due to angle difference between 0° slow axis of film 2 and 108° rubbing direction of $\pi$ twist LC cell, which is represented as orange color arrow. Besides, Fig. 4(c) shows that three lights from the relocated positions are shifted by the $\pi$ twist LC cell with 484.1-nm retardation, which is represented as two yellow color arrows. As the orange color arrow and yellow color arrows are marked, the dispersed three lights congregates at -S3 on the Poincare sphere. Figure 4(d) shows that the congregated three lights are dispersed again by the $\pi$ twist LC cell due to the reflection, which is represented as two yellow color arrows. After passing the $\pi$ twist LC cell again, dispersed positions of three lights are relocated due to angle difference between 108° rubbing direction of $\pi$ twist LC cell and 0° slow axis of film 2, which is represented as orange color arrow. Figure 4(e) shows that three lights from the relocated positions are shifted by film 2, and then shifted again by film 1. Finally, three lights arrive at -S1 position which is in the opposite direction with S1, as shown in Fig. 4(f). As a result, the three lights are converted into 90° polarization state and they can’t transmit the initial polarizer with the 0° TA. Therefore, the $\pi$ twist state successfully became an optically compensated dark state in the reflective type configuration.

 

Fig. 4 The polarization paths of three lights, 450 nm (blue light), 550 nm (green light), and 650 nm (red light) on the Poincare sphere: (a) from initial position of lights after passing the polarizer with 0° TA, (b) after passing the film 1 and film 2, (c) After passing $\pi$ twist LC cell, (d) After passing again $\pi$ twist LC cell due to a reflection, (e) after passing the film 2 and film 1, and (f) final positions.

Download Full Size | PDF

To confirm the feasibility of the optically compensated $\pi$ twist state, we calculated it with the commercial LCD simulator package (Techwiz 1D). Figure 5 shows reflectance curves as a function of wavelength in the $\pi$ twist state before and after optical compensation. A good black level, i.e., less than 0.03% reflectance can be obtained in the $\pi$ twist state.

 

Fig. 5 Reflectance curves as a function of wavelength in the $\pi$ twist state before and after optical compensation.

Download Full Size | PDF

3. Experimental results

We fabricated a test cell based on the calculation. The used LC mixture for the experiment was ML-0648 (Chisso, $\Delta \epsilon$:10.3, $\Delta n$:0.103) and the homogeneous alignment layer was PIA-5310 (Chisso). The thickness (d) and d/p of the fabricated cell was 4.7 μm and 0.25, respectively. The width and spacing of the patterned electrodes were 4 and 6 μm, respectively. The compensation films employed here were 250 nm-thick positive a-plate with a 48° slow axis and 60 nm-thick positive a-plate with a 0° slow axis. Figure 6 shows the fabricated coexisted cell in the vicinity of d/p = 0.25. Before optical compensation, we obtained the splay state as a dark state and the $\pi$ twist state as a bright state, as shown in Fig. 6(a). This image was obtained in two crossed polarizer condition. After optical compensation, the dark and bright states were reversed, as shown in Fig. 6(b). This shows that the $\pi$ twist state became a dark state and the splay state became a bright state. This image was obtained in single polarizer condition due to reflective type configuration.

 

Fig. 6 The fabricated coexisted cell in d/p = 0.25: (a) before optical compensation, (b) after optical compensation.

Download Full Size | PDF

Figures 7(a) and 7(b) show reflectance curves as a function of the wavelength in the memory mode and dynamic mode, respectively. In the calculation, the optically compensated $\pi$ twist state, common dark state, showed perfect low reflectance over the entire visible wavelength range. Besides, bright state of memory mode, the splay state, and bright state of dynamic mode, the bend state, showed good dispersion characteristics. Thus, the calculated CR for the memory and dynamic mode was 1044:1 and 821:1, respectively. However, experimentally the optically compensated $\pi$ twist state exhibited light leakages. In addition, the reflectance curves of the splay and the bend state were somewhat different from those of calculated ones. These phenomena are regarded as the mismatching of the optic axis in the compensation films during the fabrication process. Nevertheless, the actually measured CR for the memory and dynamic modes was, respectively, 47:1 and 43:1. These CRs are higher than those previously reported for a RD-LCD which was 30:1 [12].

 

Fig. 7 Reflectance curves as a function of the wavelength in (a) the memory mode and (b) dynamic mode.

Download Full Size | PDF

Figures 8(a) and 8(b) show, respectively, the voltage-reflectance (V-R) curves and response time curves measured in dynamic mode. Maximum reflectance was obtained when a 20 voltage was applied. The response time curves were evaluated with the applied voltage changes between 0 and 20 V at 1 kHz. The rising (falling) response time is the time for the change from 10% (90%) to 90% (10%) of maximum transmittance. The dynamic response time was 13.2 ms, the rising time 13 ms and the falling time 0.2 ms. This fast response suggests suitability for moving picture display.

 

Fig. 8 (a) The voltage-reflectance (V-R) curves and (b) response time curves measured in dynamic mode.

Download Full Size | PDF

Figure 9 shows the electro-optical characteristics in the memory mode. The $\pi$ twist to splay transition was observed by fringe voltage of 35 V during 455.6 ms and the splay to $\pi$ twist transition was observed by vertical voltage of 10 V during 455.6 ms. The whole reset time and writing time in memory mode was 1172.1 ms and 477.3 ms, respectively. The evaluated retention memory time for memory mode was over six months. Thus far, studies have confirmed that the retention memory can be maintained at a constant level for at least 6 months.

 

Fig. 9 The electro-optical characteristics in the memory mode.

Download Full Size | PDF

4. Conclusion

In summary, a reflective dual-mode liquid crystal display (RD-LCD) with low power consumption and high contrast ratio under ambient light was proposed. The display has a $\pi$ twist state as a common dark in both dynamic and memory modes because the $\pi$ twist state shows uniform LC texture when the cell d/p ratio is 0.25. We believe that the proposed RD-LCD can meet the demands for green displays with low power consumption as well as for mobile display devices that need high quality moving pictures.