1. Introduction

Colloidal system is an interesting research topic. Recently, the field has been enriched by the introduction of liquid crystal (LC)-colloidal system in which the dispersive medium is LC materials [1]. Different mechanisms, such as dielectrophoretic (DEP) force, electrophoretic (EP) force and the distortion of LC director have been utilized to explain the motion of the nanoparticles in the LC-colloidal systems [2–4]. The DEP motion of the nanoparticles under AC voltage excitation in LC-nanoparticle dispersions has been observed [3,5]. The magnitude and the direction of the DEP force are proportional to the gradient of the square of the electric field and the Clausius-Mossoti factor [6]. Electrophoretically controlled LC-silica dispersion has been used to fabricate the display devices. The hydrophilic silica nanoparticles covered with hydroxyl groups can form, due to hydrogen bonding, agglomerate networks in the LCs. The hydroxyl groups on the silica surface and the polar nature of the LCs yield a homeotropic LC alignment on the silica surface [7]. The scattering type bistable LC cells, the polarity-controlled multistable displays and the LC lens fabricated using the LC-silica dispersions have been demonstrated [8–11]. Application of an AC voltage also reorientates the LC director and then lifts the nanoparticles towards the substrate surface [3]. The lifting force is associated with the elastic interactions between the nanoparticles and the LC director distortion near the substrate surface.

The optically compensated bend (OCB) mode, or the pi cell has been reported to exhibit the fastest switching among the nematic LC modes [12]. To operate the pi cell, the LCs must be transformed from the originally splay state to the bend state. However, owing to the topological difference between the two states, a bias voltage greater than the critical voltage has to be applied to the pi cell to initiate nucleation for the splay-to-bend transformation which take a long time to complete. Many techniques have been proposed to induce a fast and uniform splay-to-bend transformation. These techniques can be classified into two types. The first one is to provide the pi cell an initially non-biased bend state by stabilizing the LCs at the high pretilt angle alignment, which can be achieved by forming polymer structures in the cell or providing nanostructure on the substrate surface [13–15]. The other solution is to create an initially twisted pi state, which is topologically equivalent to the bend state. Doping chiral materials into the LC host is a typical method to create a twisted pi state [16].

When a high voltage is applied to the conventional pi cell, the LCs transform from the splay state to the bend state, since the Gibbs free energy density of the LCs in the splay state is higher than that of the LCs in the bend state [13,17]. Turning off the applied voltage transforms the LCs from the bend state to the temporarily twisted pi state, which is then replaced by the splay state via the propagation of the splay state disclination line. The formation of the polymer structures in the pi cell has been reported to stabilize the LCs, creating the initially twisted pi cell without chiral dopant [18,19]. However, the used monomer materials need special development and the fabrication processes of the twisted pi cells are complicated. In this paper, we demonstrate a simple method to fabricate the initially twisted pi cell without chiral material. The cell was fabricated with the LC-silica dispersions. The obtained results show that the formed initially twisted pi cell is permanent. The possible mechanisms are discussed.

2. Experimental

Figure 1 shows fabrication principle of the silica nanoparticle doped (SND) twisted pi cell. Initially, the silica nanoparticles are uniformly dispersed in the pi cell, as shown in Fig. 1(a). When the AC high voltage is applied to the cell, the LC molecules transform from the splay state to the bend state, creating director distortion next to the substrate surfaces. The director distortion generates a lifting force that moves the doped silica nanoparticles to accumulate on both substrate surfaces [3], as shown in Fig. 1(b). After switching off the supplied voltage, the polarity of the substrate surface traps the accumulated silica nanoparticles on the substrate surface [20], stabilizing the LCs at the twisted pi state, as shown in Fig. 1(c). Notably, the polarity of the substrate surface is mainly attributed to the coated polyimide materials.

 

Fig. 1 Fabrication principle of the silica nanoparticle doped twisted pi cell.

Download Full Size | PDF

To confirm the validity of the proposed model, the LC-silica dispersion comprised the nematic E7 (Merck) and silica nanoparticles Aerosil®R812 (primary particle size is 7 nm, from Degussa-Huls) was filled into an empty cell, which was constructed from two glass substrates coated with indium tin oxide (ITO) and separated by 5.7 μm spacers. The two glass substrates were treated with conventional polyimide layers that were rubbed in parallel directions, providing a low pretilt angle of ~2 degrees for LC molecules. E7 had the ordinary dielectric constant of 5.49 and extraordinary dielectric constant of 19.24, respectively. R812 had a dielectric constant of ~3.8.

3. Results and discussion

Figure 2 shows the measured transient transmittances of the SND pi cells after AC high voltage pulse (20 V, 1 kHz, 200 ms) excitation. The director of the LCs is parallel with the transmission axes of the two parallel-aligned polarizers. Consequently, the LCs without twisted domains will have a high transmittance. As shown in Fig. 2, at t = 0, the SND pi cells are in the bend state with high transmittances, because the supplied AC high voltage transforms the LCs from the originally splay state to the bend state. As the supplied AC high voltage is suddenly switched off (t > 0), the transmittance of the pristine pi cell (without silica nanoparticles) drops rapidly and then increases slowly following saturation. The rapid drop of transmittance indicates that the LCs have relaxed to the twisted state; the slow increase of transmittance indicates the annihilation of the twisted domains and the growth of the splay domains. The saturated transmittance of the pristine pi cell reveals that the LCs are in a metastable state comprising the splay domains and the twisted domains. The metastable state can be retained ~30 min, and then returns to the originally splay state with an even higher transmittance. For the 0.3 wt% SND pi cell, the saturated transmittance of the cell is less than that of the pristine pi cell, but this saturated transmittance does not return to the originally splay state even after 1 month, indicating that some twisted domains are permanently stabilized. The saturated transmittances of the SND pi cells decrease with increasing silica nanoparticle concentrations, indicating that the stabilized twisted domains increase with increasing silica nanoparticle concentrations. According to our results, the twisted domains can be created when the doped silica nanoparticle is higher than 0.3 wt%. As the doped silica nanoparticle concentration exceeds 0.6 wt%, the saturated transmittance of the SND pi cell does not decrease further. Notably, the transmittance of the pi cell in the twisted state can be similar to that of the pi cell in the splay state or the bend state in the case of the adiabatic following condition. However, in this experiment, the used pi cell has a thin cell gap, which is far below the adiabatic following condition. Consequently, the twisted state of the pi cell can be easily examined by the rapid drop of transmittance.

 

Fig. 2 Measured transient transmittances of the SND pi cells after AC high voltage pulse (20 V, 1 kHz, 200 ms) excitation.

Download Full Size | PDF

Figure 3 show the polarized optical microscopic (POM) images of the SND pi cells after AC high voltage pulse (20 V, 1 kHz, 200 ms) excitation. The applied AC high voltage causes the LCs to have a director distortion near the substrate surface, generating a lifting force that accumulates the nanoparticles on the substrate surface of the cell. In Figs. 3(a)-3(c), the LC director is parallel with one of the transmission axes of the crossed polarizers. The POM images of the cells after high voltage pulse excitation are taken when the LC alignment reaches stable state (~3 hr after the voltage pulse excitation). For the pristine pi cell, the POM images before and after high voltage pulse excitation are in the same dark state, as shown in Fig. 3(a), owing to the lack of the twisted domains. However, as can be seen in Figs. 3(b) and 3(c), for the SND pi cells, the POM images after AC high voltage pulse excitation show leakage of the light, which increase with silica nanoparticle concentrations. The leakage of the light is resulted from the twisted domains caused by the doped silica nanoparticles, which are deposited on both substrate surfaces under AC high voltage application. In Figs. 3(d)-3(f), the LC director has an angle of 45° with respect to the transmission axes of the crossed polarizers. The dark domains appear in the POM images of the SND pi cells, as shown in Figs. 3(e) and 3(f). Rotating the cells revealing that some dark domains are invariant, indicating that the LCs in these domains are homeotropically aligned, which is caused by the silica nanoparticles deposited on the substrate surfaces. The homeotropically-aligned LC domains increase with silica nanoparticle concentrations, due to the increased amounts of the silica nanoparticles deposited on the substrate surface. The applied AC high voltage may create DEP force to moves the nanoparticles. However, in our experiment, the dielectric constant of the silica nanoparticle is less than that of the LCs; the created DEP force will push the silica nanoparticles away from the substrate surface. Hence, the DEP force is failed to trap the silica nanoparticles on the substrate surface. Furthermore, the supplied AC voltage also suppresses the EP force of the silica nanoparticles in the SND pi cell.

 

Fig. 3 POM images of the SND pi cells after the AC high voltage pulse (20 V, 1 kHz, 200 ms) excitation: (a), (d) 0 wt%; (b), (e) 0.3 wt%; (c), (f); 0.9wt%. P and A indicate the transmission axes of the polarizer and analyzer, respectively.

Download Full Size | PDF

Figures 4(a) and 4(b) show the optical images (without polarizers) of the silica nanoparticles deposited on the substrate surfaces after AC high voltage pulse (40 V AC 1 kHz) excitation. The 0.9 wt% SND cell is used in this experiment. The optical microscope is focused at the interface between LCs and the substrate. The green images show the nanoparticles accumulated near the substrate surface, but not the nanoparticles exactly on the substrate surface, owing to the limitation of the depth of field of the optical microscope. As shown in Figs. 4(a) and 4(b), the optical images of the top and the bottom substrate surfaces after AC high voltage pulse excitation are similar, owing to the same amounts of the silica nanoparticles deposited on the top and the bottom substrate surfaces by the created lifting force. However, as shown in Figs. 4(c) and 4(d), after DC high voltage pulse excitation, the amounts of the silica nanoparticles deposited on the top substrate are less than those of the silica nanoparticles deposited on the bottom substrate. This is because that under DC voltage excitation, an additional EP force will move the negatively charged silica nanoparticles toward the bottom substrate with positive voltage polarity. Furthermore, as shown in the figures, the deposition of the silica nanoparticles is not uniform. Therefore, there should be some sub-domains in which the LC directors are not in a well controlled condition, and then degrading the optical performance of the SND twisted pi cell.

 

Fig. 4 Optical images of the silica nanoparticles deposited on the substrate surfaces after high voltage pulse excitation. (a) top substrate, 40 V AC 1 kHz; (b) bottom substrate, 40 V AC 1 kHz; (c) top substrate, 40 V DC (negative polarity); (d) bottom substrate, 40 V DC (positive polarity).

Download Full Size | PDF

Figure 5(a) shows the normalized voltage dependent transmittance (VT) curves of the SND twisted pi cells, which are formed by AC high voltage pulse excitation. As revealed, the doped silica nanoparticles cause the SND twisted pi cell to have the lower operation voltage than the pristine pi cell, because of the homeotropically-aligned LC domains in the SND twisted pi cell. The homeotropically-aligned LC domains also cause the SND twisted pi cell to have the better dark state (ex: at 10 V) than the pristine pi cell. However, the scattering of the incident light by the aggregated silica nanoparticles also slightly decreases the maximum transmittance of the SND twisted pi cell. Notably, as shown in Fig. 5(a), there is a significant difference between the initial transmittances of the 0 wt% and the 0.3 wt% cells. However, as shown in Fig. 2, the saturated transmittances of the 0 wt% and the 0.3 wt% cells are almost the same. This is because that the pristine pi cell (0 wt%) measured in Fig. 2 is in the metastable state comprising the twisted domains and the splay domains; hence, its transmittances can be similar to that of the 0.3 wt% SND twisted pi cell. However, in Fig. 5(a), the pristine pi cell is in the stably splay state; hence, its initial transmittance is significantly different from the 0.3 wt% SND twisted pi cell. The rise and fall times of the SND twisted pi cells fabricated by AC high voltage pulse excitation are shown in Fig. 5(b). The cells are operated from 2.5 V and 10 V. As shown in the figure, the rise time and the fall time of the SND twisted pi cells are less than those of the pristine pi cell, because the doped silica nanoparticles decrease the relaxation time constant of the LC-silica dispersions [21]. As shown in Fig. 2, the 0.6 wt% and the 0.9 wt% cells have the similar saturated transmittances, due to the similar amounts of the nanoparticles deposited on the substrates. Hence, the concentration of the residual nanoparticles in the 0.9 wt% cell must be higher than that of the residual nanoparticles in the 0.6 wt% cell. As a consequent, the 0.9 wt% cell has the shorter relaxation time constant and thus the shorter response time than the 0.6 wt% cell. Notably, as shown in Figs. 4(c) and 4(d), the SND twisted pi cell can also be formed by applying DC high voltage pulse. The VT curve of the SND twisted pi cell formed by DC high voltage pulse excitation is similar to that of the SND twisted pi cell formed by AC high voltage pulse excitation. However, the rise time and the fall time of the SND twisted pi cell formed by DC high voltage pulse excitation are longer than those of the SND twisted pi cell formed by AC high voltage pulse excitation.

 

Fig. 5 (a) Normalized voltage dependent transmittance curves of the pristine pi cell and the SND twisted pi cell; (b) measured rise and fall times of the SND twisted pi cells.

Download Full Size | PDF

4. Conclusion

In conclusion, we have demonstrated a simple method to fabricate the initially twisted pi cell using the LC-silica dispersions. The uniformly-dispersed silica nanoparticles are trapped on both substrate surfaces of the cell, due to the lifting force generated by the director distortion of the LCs next to the substrate surfaces and the polarity of the substrate surface. The formed twisted pi state is permanent stabilized, and the operation voltage and the response time of the SND twisted pi cell are less than those of the conventional pi cell. The SND twisted pi cell can also be formed by using DC high voltage pulse excitation. However, the effects of DC high voltage pulse in forming the SND twisted pi cell are worse than those of AC high voltage pulse. According to our results, the SND twisted pi cell can be formed when the silica nanoparticle exceeds 0.3 wt%. Further studies on the physical properties, such as the helical twisting properties, the pretilt angle and the anchoring properties of the SND pi cells are now underway.