Open Access
Add to BibSonomy
Abstract
We herein propose a multiwalled carbon nanotube (MWCNT) doping method into liquid crystal (LC) alignment polyimides (PIs) with low resistivity for resolving both issues of voltage holding and image sticking in low-frequency-driven fringe-field switching (FFS) LC modes using negative dielectric LCs (n-LCs). By utilizing strong ion trapping ability of MWCNTs, the FFS n-LC cell aligned by low resistivity PIs with 0.05 wt% MWCNT doping exhibited an excellent voltage holding ratio of 99% under an extremely low operation frequency of 0.5 Hz and approximately 23.6 times better surface discharging property than that aligned by high resistivity PIs.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Depending on initial liquid crystal (LC) configurations and field-induced-switching schemes, several types of LC modes have been developed thus far. However, for portable display applications such as mobile phones and tablets requiring high-resolution pixel density (pixel per inch: ppi), several types of fringe-field switching (FFS) LC modes have been widely adopted owing to several positive properties such as wider viewing angle, higher transmittance with possibility of pixel design for higher pixel aperture ratios, and operation by lower power consumption compared with other types of LC modes [1–3]. In current display markets, development of mobile displays with resolutions over quad high definition (QHD) is progressing. Among several key technologies required in developing a higher ppi mobile display panel, a low-frequency-driven scheme for operating an LC display (LCD) panel has attracted significant attention as a promising method for resolving power consumption problems [4–6]. Because the power consumption level required in the driving circuit unit for operating LCD panels is linearly proportional to its image refreshing rate, a low-frequency-driven scheme can be effective as a power-saving approach in a higher ppi panel. These are important issues that must be solved for developing larger mobile display panels at a higher pixel density because of the significantly increased total pixel amount. However, when the FFS LC mode is driven under a lower operating frequency condition than the conventional 60 Hz, image flickering problems can be observed, where two different transmittance levels are exhibited during signal frame change [6–12]. The image flickering observed in the low-frequency-driven FFS LC mode can be attributed to flexoelectric (FE) effects or voltage holding effects [6–12]. These temporal transmittance level fluctuations cannot be observed in the conventional 60 Hz operation considering the dynamic sensitivity of the human visual perception; however, image flickering can be observed clearly at lower operating frequencies [13]. Thus, commercially achievable level of low-frequency-driven still remains between 20 and 30 Hz [14]. A mobile LCD panel, which can be operated at a much lower driving frequency without causing image flickering, must be developed, especially for competing it with organic light-emitting displays in terms of power saving issue.
The image flickering problems can be caused in a different manner depending on the LC types (positive or negative dielectric anisotropy LCs: p-LC and n-LC, respectively) employed in FFS LC modes. Unlike other types of LC modes, the FFS LC modes using p-LCs (FFS p-LC modes) cause highly deformed splay or bend LC distributions near an electrode substrate under the field-on state, which makes the FE effect coupled with the transmittance level fluctuations exhibiting its signal-polarity dependency. Transmittance level variations of average value over each signal frame time and over each entire pixel area, called static flickering, can be minimized by optimizing the signal voltage waveform or electrode structures that generate fringe-field patterns [8–10]. To resolve using LC materials, the FE polarization effects induced by the spatially, highly deformed fringe-field can be reduced by modifying FE coefficients of LCs through doping bent-core LC molecules within nematic p-LCs [6]. Despite these efforts, transient blinking can appear at the moment of signal polarity change. In our previous research, we revealed that this dynamic flickering effect is highly related to the asymmetric dynamic responses between the splay-to-bend and bend-to-splay transitions occurring in positionally different manners at the moment of polarity change of the signal voltage [7]. We demonstrated that the dynamic flickering can be resolved by balancing the FE-coupled LC dynamics through tuning the splay and bend elastic constants (K11 and K33, respectively) of the p-LCs introduced to achieve the symmetric dynamic switching behaviors of the field-induced transitions between the splay- and bend-enhanced positional deformations.
When using n-LCs for the FFS LC modes, the FE-coupled flickering problems can be mitigated better compared with when using FFS p-LC modes because the n-LCs are reoriented horizontally rather than vertically in response to the spatially, highly deformed fringe field distributions, thus generating more suppressed field-induced splay or bend elastic distortions near the LC alignment surface [12,15–17]. Furthermore, owing to the n-LC reorientation that is perpendicular to the applied fields, the FFS LC mode using n-LCs (FFS n-LC mode) can exhibit a higher transmittance and contrast ratio; additionally, a lower mura effect against the point touch contact can be achieved compared with the FFS p-LC modes, which is required for high-resolution mobile displays operated with a touch interface [16–18]. Despite these advantages, it is challenging to employ FFS n-LC modes in low-frequency-driven schemes (under 0.5 Hz driving frequency) for lowering the power consumption level of the panel because of the voltage holding issues of the LC cell. At a lower operation frequency, temporal transmittance variation in the FFS n-LC modes with a lower voltage holding ratio (VHR) and its abrupt instantaneous change in transmittance level at the moment of signal polarity change can occur; this can cause other types of image flickering problems even though FE-coupled flickering behaviors are suppressed [15,19,20]. The main reason for the comparatively lower VHR properties in the FFS n-LC modes is the higher mobile ion density within an LC cell, resulting in a lower resistivity of the LC layer and lower dielectric properties of n-LCs compared with those of p-LCs [20,21]. The mobile ion density in an n-LC cell is highly dependent on the molecular constituents of the employed n-LC; it is also affected by the type of LC alignment polyimide (PI) layer [21–23]. In terms of the VHR issue in the FFS n-LC modes, a high resistivity PI (ρ ∼ 1015 Ω·cm) is more suitable as an LC alignment layer rather than a low resistivity PI (ρ ∼ 1013 Ω·cm) because high resistivity PIs contain fewer ion impurities within the PI layer and exhibit lower ion-desorption properties in response to the LC-switching field [22]. However, in general, high resistivity PIs exhibit a lower discharging coefficient, i.e., the characterization factor denoting the charge dissipation ability accumulated within the PI layer, than low resistivity PIs [23]. When a low-frequency-driven scheme is applied, an FFS LC cell employing a high resistivity PI is not free from the severe image sticking issue. In an FFS LC cell, residual charges accumulating within the PI layer during operation can be caused by undesired positional direct-current (DC) offset voltage deviations and imperfect capacitive dielectrics between the pixel and common electrodes of the FFS LC mode [22,23]. Thus, for FFS p-LC modes that are relatively free from the VHR issue owing to a lower mobile ion density within the p-LC, a low resistivity PI is typically preferable for resolving the image sticking problem in LCD industries [23]. However, the PI type for the FFS n-LC modes should be chosen carefully to avoid VHR and image sticking problems, by considering the bulk ion density of the employed n-LC and the discharging coefficient of the employed PI [21,23]. To render the FFS n-LC panels operable at an extremely low frequency (∼ 0.5 Hz) refreshing rate, a more viable method to both decrease the bulk mobile ion density of the n-LC layer and enhance the surface discharging properties of the PI layer must be developed. The investigations on the effects of the LC and PI types on the VHR and image sticking behaviors are also important issues in the in-plane-switching LC modes that should be further developed [21,22,24–28].
As an approach for reducing the mobile ion density within an LC bulk layer, carbon nanomaterial doping (as a mixture) into the LC layer has been proposed for the homogeneously planar configuration of a nematic LC (NLC) cell [29–31]. Carbon nanotubes (CNTs) or fullerenes suspended into NLCs demonstrated superior ion trapping capability with effectively reducing the bulk mobile ion density within the host NLC. However, when the doping density of the carbon nanomaterials is increased, several LC properties affecting the electro–optic properties, such as the LC order parameter and dielectric properties, are varied from those of pristine host NLC [31,32]. In particular, in a higher dispersion condition, the aggregation of nanomaterials should be considered for a reliable operation and long-term stability of the LC cell [31]. For metallic nanomaterials such as multiwalled CNTs (MWCNTs), field-induced LC reorientation behaviors exhibited in them can be distorted even at a lower doping density [32].
In this study, to resolve the VHR and image sticking problems of low-frequency-driving FFS n-LC modes, we propose a modified LC alignment layer prepared by dispersing MWCNTs into low resistivity PI. Owing to the surface charge trapping effect of the MWCNT doped within a low resistivity PI layer, the mobile ion density of the n-LC was significantly reduced by two orders of magnitude from 2.068 nC/mm3 to 0.027 nC/mm3, as a result of the much enhanced VHR property compared with that of the FFS n-LC cell aligned with the pristine material of low resistivity PI. Even at an extremely low operation frequency of 0.5 Hz, the ideal VHR value of 99% could be obtained by only 0.05 wt% of MWCNT doping, whereas the FFS n-LC cell aligned by low resistivity PI without MWCNT doping exhibited 63% VHR. Owing to the low resistivity of the PI employed in our approach and the positive contribution from the metallic MWCNT doping, the presented FFS n-LC cell demonstrated excellent discharging properties unlike the FFS n-LC aligned by high resistivity PI layers.
2. Experimental procedure
Figure 1 shows the schematic of the FFS n-LC cell, where n-LCs are aligned on the MWCNT-doped planar LC alignment PI layer to resolve the VHR and surface charge accumulation problems especially for the low-frequency-driven operation in our approach. As the LC alignment layer, the low resistivity PI layer was used for solving discharging issues in FFS LC modes. Compared with p-LCs, the relatively abundant mobile ions within the n-LCs diffuse to the surface and subsequently trapped by the MWCNTs that are dispersed within the low resistivity PI alignment layer. Owing to their relatively high electron affinity, CNTs possess high ion trapping ability [32]. In particular, the ion trapping efficiency of MWCNTs has been reported to be approximately double that of single-walled CNTs (SWCNTs) from the evaluation of ion trapping coefficient of both types of CNTs [29]. Charges from the surface-adsorbed and trapped ions by the MWCNTs and those from the surface-accumulated leakages passing through the dielectric layer between the pixel and common electrodes can be easily discharged without image sticking problems owing to the low resistivity of PI and the metallic property of the MWCNT employed in the presented scheme. Unlike doping CNTs into the LC layer, the MWCNTs that dispersed well within the PI layer do not cause CNT aggregation and do not disturb the field-switching behaviors of the LC molecules.
Fig. 1. Schematic of low-frequency-driven FFS LC cell, where n-LCs are aligned on MWCNT-doped planar LC alignment PI layer. Mobile ionic charges within n-LCs are trapped by MWCNT dispersed within PI, thus enhancing voltage-holding properties and improving dynamic flickering effects in low-frequency-driven FFS LC cell.
In Fig. 1, the planar LC alignment of SE6514 (ρ ∼ 1013 Ω·cm, Nissan Chemical Corporation) was used as the low resistivity PI. Furthermore, MWCNTs (Chengdu Organic Chemicals Co. Ltd) of lengths 0.5–2 µm were dispersed to the host PI material. To compare the VHR characteristics of the FFS n-LC modes according to the resistivity level of the employed PI layer, an FFS n-LC cell aligned by high resistivity PI was additionally prepared using SE5811 (ρ ∼ 1015 Ω·cm, Nissan Chemical Corporation) without MWCNT doping. Both types of PIs used in our experiments have similar basic chemical structures with the dianhydride and diamine parts. In the low resistivity PI, the diamine part is constituted by more phenyl rings to reduce the PI resistivity level by utilizing the π electrons of the phenyl group. On the other hand, the high resistivity PI is constituted by more alkyl chains in the diamine part. Before mixing the MWCNTs into the PI, the pristine MWCNTs were annealed in a vacuum oven at 200 °C for 4 h to remove any humidity adsorbed on the MWCNT surfaces. After the annealing, the MWCNTs were dispersed in the host PI material by varying the MWCNT doping ratio from 0 to 0.05 wt% to verify the doping-dependent ion trapping capability at the surface. To achieve a homogeneous MWCNT dispersion in the low resistivity host PI material, bath sonication was performed at 46 kHz, 150 W for 30 min. To avoid the sonication-induced heating effect of the suspension, which can cause the pre-curing problem of the host PI material, the temperature of the suspension within the sonication bath was maintained at approximately 0 °C during the sonication. Without any additional surface treatment on the MWCNTs, the MWCNTs were well dispersed in the host PI; this can be attributed to the interaction between the conjugated aromatic parts of the MWCNTs and the PI material [32]. The homogeneous suspension of the MWCNTs and PI mixture was spin-coated on the inner surfaces of the top and bottom substrates of the FFS n-LC cell to form the planar alignment layer. These coated substrates were prebaked at 80 °C for 60 s to evaporate the solvent of the solution and then post-baked at 250 °C for 30 min for PI imidization. After post-baking, both substrates were rubbed using a commercial rubbing cloth before substrate assembly to define the LC alignment direction. As the n-LC, the ML1407 (Merck) NLC was used, which is a fluorinated LC material. The physical properties of the employed n-LC are as follows: dielectric anisotropy of Δɛ = –4.1; birefringence of Δn = 0.1011; elastic constants of K11 = 13.3 pN, K22 = 6.7 pN, and K33 = 15.1 pN.
To characterize the mobile ion density variation within the LC bulk according to the surface condition of the LC alignment layer, electrically controlled birefringence (ECB) LC cells were prepared by changing the MWCNT doping density within the host PI at the same LC condition. For this experiment, vertical electric fields were applied to the ECB cells by patterned indium-tin-oxide (ITO) electrodes (5 mm × 5 mm in pattern size), and the thickness of the LC layer was 5.5 µm. Because the n-LC is reoriented perpendicularly to the applied field, the n-LC molecules cannot be reoriented in response to the applied field in the ECB configuration, and the effect of dielectric variation of the n-LC layer can be disregarded in our mobile ion density characterization with capacitive current measurement.
The VHR properties of the FFS n-LC cells were characterized by changing the MWCNT doping ratio within the low resistivity PI. The VHR-dependent image flickering effects were analyzed by varying the operation frequency of the signal polarity inversion. For the sets of FFS n-LC cells, the discharging coefficients of the PI layers were evaluated as well. For all sets of the experiments, the total size of the FFS n-LC cell was 30 mm × 30 mm, and the thickness of the LC layer was 4 µm. The FFS n-LC cells were arrayed with the pixels and each pixel size was 64 µm × 192 µm. For the FFS LC operation, the fringe fields were formed by the stacked electrode structure of the top interdigitated ITO patterns and the bottom non-patterned ITO at the bottom substrate, as shown in Fig. 1. Between the top and bottom electrodes, a SiNx (ɛ = 6.7, ρ ∼ 1014 Ω·cm) layer was deposited as an insulator and its thickness was 0.5 µm. Each pixel of the FFS n-LC panels does not contain a TFT for operation. The operating field was directly applied to the stacked electrodes using a signal generator to exclude the evaluation errors from the charging and capacitive holding effect by the TFT imperfection in our VHR and discharging characterization according to the surface alignment layer conditions. The operating signal was applied to the patterned ITO layer as the pixel electrode, and the non-patterned bottom ITO layer was used as the common electrode. The patterned pixel electrodes had a width of 3 µm, and their spacing between the ITO patterns was 4 µm. For the top alignment substrate, the same PI layer as that used for the bottom alignment substrate for each set of experiments was spin-coated without ITO deposition. For all sets of the FFS n-LC cells, the rubbing direction of the PI layers was 7° with respect to the interdigitated pattern direction of the pixel electrode.
3. Result and discussion
3.1 Flickering by ionic charge transfer effect in FFS n-LC cell
Considering the mobile ions within an LC layer and their field-induced drift behaviors, the effective field applied to the LC cell decreases gradually at a fixed DC field. This results in a voltage loss (Vloss) compared to the initial voltage level applied to the LC cell; the Vloss(t) level can be described by the amount of mobile ion charge transfer during the signal voltage duration in response to an applied field, as follows [21,33]:
where q is the charge of an electron (1.602 × 10−19 C), na(t) is the surface-adsorbed charge density by the field-drifted mobile ions at the PI layer, CLC is the sheet capacitance of an LC layer, and f is the operation frequency of the signal field applied to the LC cell. Furthermore, CLC(t) is a temporally varying amount according to the LC configurations reoriented in response to the applied field. As shown in Eq. (1), the amount of Vloss(t = 1/2f) increases at LC cell conditions of higher na(t) and lower CLC(t) and the operation condition of lower f because of the increasing na(t) with time while the signal polarity of the applied voltage remains constant. The Vloss-related na(t) amount depends on the inherent mobile ion density, NLC(t = t0), within the LC and the surface density of the impurity charges, nd(t), desorbed from the alignment layers. Compared with the p-LCs, the resistivity levels of the n-LCs are lower because of higher mobile ions within LCs. Due to the LC chemical structure for exhibiting the negative dielectric anisotropy, the dielectric constants of the n-LCs are also lower than those of the p-LCs, which results in the lower LC capacitance levels of the n-LC cells. As shown in Eq. (1), the FFS n-LC modes cannot avoid the VHR issues. Especially in the low-frequency-driven FFS n-LC modes, Vloss(t)-related transmittance variations occur in the form of the flickering images considering the increased Vloss(t) amount and the dynamic sensitivity of the human eyes at a lower operation frequency.Figure 2 shows the voltage holding properties of the FFS n-LC cells operated at 0.5 Hz driving frequency, in which Figs. 2(a) and 2(b) present the temporal transmittance variations for the cells prepared with low (ρ ∼ 1013 Ω·cm) and high (ρ ∼ 1015 Ω·cm) resistivity PIs, respectively. In Fig. 2, for both alignment layers, the pristine PI materials were used as purchased without MWCNT doping. For characterizing the voltage holding property according to the employed PI type, the transmittance variations of the FFS n-LC cells placed between the crossed polarizers were measured using a photodetector (Model 2031, New Port Inc.) and oscilloscope (DSO1052B, Keysight Technologies Inc.), as shown in the schematic diagram of Fig. 3. A square wave signal of V20 was applied to the cells using a signal generator (33500B, Keysight Technologies Inc.). V20 is the voltage level required for a 20% transmittance of the full brightness level that corresponds to the half-gray level of the gamma 2.2 grayscale; it is the signal voltage level typically employed in evaluating image flickering properties considering human gamma sensitivity [7]. To evaluate the field-dependent transmittance variation, the signal voltage applied to the cells was measured using an oscilloscope with synchronized channels. For this characterization, a He–Ne laser (λ = 633 nm) was used as the input source.
Fig. 2. Dynamic transmittance variations in low-frequency-driven (0.5 Hz) FFS n-LC modes exhibiting different voltage holding properties and flickering effects according to resistivity of LC alignment surfaces: (a) low resistivity PI and (b) high resistivity PI layers.
Fig. 3. Schematic diagram of electro–optic measurement system used for characterizing voltage holding properties in FFS n-LC cells according to alignment layer conditions.
As shown in Fig. 2, when n-LC for the FFS LC cells, both samples exhibited instantaneous transmittance level variations at moments of signal polarity changes because of voltage holding. The amount of Vloss, represented by the continuously decreasing transmittance level during each signal duration time for a signal polarity (Fig. 2), was much higher in the FFS n-LC cell aligned with low resistivity PI than that aligned with high resistivity PI. The results of Fig. 2 indicate that the field-driven desorption charges from the PIs can significantly affect the VHR properties as much as the initial mobile ion density of NLC(t = t0) within an LC. It has been reported that high resistivity PIs possess fewer charge impurities than low resistivity PIs and that their charge desorption coefficients are relatively lower [22]. However, high resistivity PIs cannot easily provide discharging path layers for the surface charges accumulated from the imperfect capacitive layer between the pixel and common electrodes of the FFS modes [22]. Image sticking problems can be easily encountered in long-term operations. Meanwhile, using low resistivity PIs for low-frequency-driven FFS n-LC modes results in flickering behaviors because of the poor VHR properties, as shown in Fig. 2(a). A fundamental solution for resolving the imaging sticking and flickering issues should be proposed for LC alignment layers in low-frequency-driven FFS n-LC modes.
3.2 Mobile ion density according to MWCNT doping condition in LC alignment layer
In this study, we propose the MWCNT-doped low resistivity PI layer for low-frequency-driven FFS n-LC modes by effectively reducing the mobile ion density within the LC layer by utilizing the excellent ion trapping ability of the MWCNTs dispersed within the PI. To characterize the mobile ion densities dependent on the MWCNT doping condition within the low resistivity PI layer, ECB n-LC cells were prepared with different surface alignment PI conditions by varying the MWCNT doping ratio within the PI, as shown in Fig. 4(a). In evaluating the bulk mobile ion density with the capacitive current variation caused by the ionic charge transfer amounts in response to the signal polarity changes (see the schematic of Fig. 4), applying vertical fields to the ECB n-LC cells can avoid the LC capacitive current orientable from the LC dielectric changes by the field-driven LC reorientations. As the signal voltage, a 0.1 Hz triangle wave with an amplitude of 10 V was applied to the ECB n-LC cells, as shown in Fig. 4(b). For the signal voltage generation and the current measurement in nano-ampere scale, a precision source-measurement unit (A2901, Keysight Technologies Inc.) was used. To evaluate the bulk mobile ion density within the LC layer, an effective LC bulk layer of dimensions 5 mm × 5 mm × 5.5 µm was used, which was defined by the square-patterned ITO layers and the LC layer thickness.
Fig. 4. (a) Schematics of homogeneously planar-aligned n-LC cell used for characterization of time-dependent variation of mobile ionic charge density (NLC(t)) within n-LC bulk caused by polarity-changing field-induced desorption and adsorption of mobile ions at LC alignment surfaces. (b) Triangular AC waveform (0.1 Hz) applied to n-LC cell for measuring surface-dependent mobile ionic charge density within n-LC bulk.
In Fig. 4(a), the field-dependent temporal variations of nd, na, and NLC are expressible as follows [21,33]:
where Rd and ni in Eq. (3) are the desorption coefficient and the initially possessed ion density in the alignment layer, respectively. In Eq. (4), μ, Va, and d represent the ion mobility within the LC layer, the applied voltage, and the cell gap condition of the ECB LC cell, respectively. In response to the signal polarity of an applied field, the mobile ions transfer and adsorb on the PI surfaces, and the amount is dependent on the desorbed charges from the PI, as shown in Eqs. (2) and (3).Figure 5(a) shows the capacitive current measurement result for the ECB n-LC cell aligned by the pristine low resistivity PI alignment layers used without MWCNT doping. The applied signal voltage was swept with a periodically increasing and decreasing field, as shown in Fig. 4(b), and the capacitive current result of Fig. 5(a) includes two transient current peaks owing to the field-driven mobile ions within the LC cell. In our measurement scheme, the capacitance current terms related to the field-induced n-LC reorientation at LC bulk or surface LCs can be excluded [34,35], and two transient current peaks presented in Fig. 5 can be attributed to the results of the mobile ion movements within the n-LC in response to the signal polarity change. In this graph, the filled areas of the measured current peaks represent the total charges by the field-driven ionic charge flows. The mobile ion density can be evaluated by dividing the measured total charges with the effective LC volume. The evaluated value of the mobile ion density for the LC cell prepared without MWCNT doping into the low resistivity PI was 2.068 nC/mm3. Figure 5(b) shows the capacitive current measurement results obtained for several sets of ECB n-LC cells aligned in different PI conditions through varying their MWCNT doping ratios; the results are summarized in Fig. 5(c). Compared with the n-LC cell aligned by the pristine PI, Fig. 5(b) shows that a minute amount of MWCNT doping can effectively reduce the mobile ion density within the LC layer. At 0.001 wt% of MWCNT doping, the LC mobile ion density can be decreased to 1.014 nC/mm3 by 49% compared with the non-doping case. By increasing the MWCNT doping ratio, both the transient current peaks in Fig. 5(b) and the evaluated mobile ion densities of Fig. 5(c) decreased. At 0.05 wt% of MWCNT doping, the measured mobile ion density was 0.027 nC/mm3, i.e., 1.3% compared with the LC cell without MWCNT doping into the low resistivity PI layer. This value, obtained for the n-LC, was much lower (∼ 1/10) than the inherent mobile ion density levels measured for p-LCs [31,36], thus indicating that the small amount of MWCNT doping into low resistivity PI layers can provide a viable solution for resolving VHR issues in low-frequency-driven FFS n-LC modes through depleting the mobile ion charges within the LC layer, owing to the excellent ion trapping ability of the MWCNT. The thickness of the PI layer was about 130 nm. Figure 5 shows that the MWCNTs dispersed within the sufficiently thin PI layer can effectively play a role as the distributed charge trapping sites. Note that the trapped charges by the MWCNTs should not be accumulated at the LC alignment surface and should be dissipated considering the image flickering issues at a low-frequency operation and the image sticking issue. For this purpose, instead of the SWCNTs [29,32], the metallic MWCNTs were doped into the PI layer. The discharging properties of the MWCNT-doped PI will be discussed in subsection 3.4 in detail.
Fig. 5. (a) Capacitive current measurements of homogeneously planar-aligned n-LC cells in response to triangular AC waveform (0.1 Hz), obtained for pristine low resistivity LC alignment PI. (b) Capacitive current variation according to doping conditions of MWCNTs within low resistivity LC alignment PI. (c) Mobile ion densities within n-LC cells according to MWCNT doping ratio within PI, which are extracted values from the measurements of Fig. 5(b).
3.3 Surface-dependent VHR in low-frequency-driven FFS n-LC mode
For the FFS n-LC cells prepared with different alignment conditions by varying the MWCNT doping ratio within the low resistivity PI layer, the time-varying transmittance curves according to the signal polarity change were measured under different signal frequency conditions, as shown in Fig. 6(a). In Fig. 6(a), as representative examples, the dynamic transmittance curves for 0.5 and 0.2 Hz operations are presented. For this evaluation, the measurement setup and the applied signal voltage conditions were the same as those employed for Figs. 2 and 3. Using the dynamic transmittance curves, the VHR values were evaluated by dividing the minimum transmittance level of the FFS n-LC cells with their maximum transmittance level for each signal frame time at different PI and signal frequency conditions, as summarized in Fig. 6(b). The result for the low resistivity PI layer without MWCNT doping is presented in Fig. 2(a).
Fig. 6. (a) Measurements of dynamic transmittance variations in low-frequency-driven (0.5 Hz and 0.2 Hz square waveforms) FFS n-LC modes, which are aligned on low resistivity LC alignment PI layers prepared with different MWCNT doping conditions. (b) Voltage holding ratios of FFS n-LC modes according to MWCNT doping conditions of low resistivity PIs and operating signal frequencies.
Figure 6(a) shows that the amount of Vloss at t = 1/2f, discussed with Eq. (1), increases significantly with decreasing operation frequency because of the increased field-driven mobile ion charges within the LC cells; however, the Vloss characteristics improved considerably with increasing MWCNT doping ratio within the low resistivity PI layer. At the low resistivity PI layer without MWCNT doping, the VHR was 63.1% at 0.5 Hz operation frequency; however, it decreased significantly to 7.8% at 0.2 Hz operation frequency. With only a small amount of MWCNT doping, the VHR characteristics were improved significantly, in which the FFS n-LC cell aligned with the PI doped with MWCNTs of 0.001 wt% mixing ratio exhibited VHR values of 88.9% and 58.3% for the 0.5 and 0.2 Hz operation frequencies, respectively, as shown in Figs. 6(a) and 6(b). When we increased the MWCNT doping ratio in the low resistivity PI up to 0.05 wt%, the excellent VHR of 99% was achieved at a relatively low frequency operation of 0.5 Hz. Despite further reducing the operation frequencies to 0.2 and 0.1 Hz, the voltage holding properties remained and exhibited VHR values of 96.5% and 95.2%, respectively. The operating frequency levels shown in Fig. 6(b), employed for the characterization of surface-dependent VHR properties, were considerably lower than those tested in current industries. When we compared the VHR values obtained for the FFS n-LC cell aligned by the low resistivity PI with MWCNT doping (0.05 wt%) and that aligned by the pristine PI without MWCNT doping, the VHR was improved by approximately 1.6 times and 12.4 times for the operation frequencies of 0.5 and 0.2 Hz, respectively. Despite using the low resistivity PI for the FFS n-LC cells, MWCNT doping within the PI layer can provide excellent VHRs even for extremely low operation frequencies intended for the ultimate power-saving of the FFS LC panels; this can be attributed to the annihilation of mobile ions within the LC layer by the effective ion trapping ability of the MWCNTs doped within the PI layer. Even though the operation frequency was extremely decreased below 0.5 Hz, the transmittance curves of the FFS n-LC cell aligned with the 0.05 wt% of MWCNT doping into the low resistivity PI showed the negligible fluctuation of the transmittance level at the moment of the signal polarity change, which means that the trapped charges by the MWCNTs are not accumulated at the surface and they are dissipated through the metallic MWCNTs.
In Ref. [37], S.-T. Wu et al. discussed the human-perceivable image flickering properties in the FFS LC modes by varying the employed LC types and the operating signal frequencies. To relate the voltage-dependent transmittance level fluctuation with the image flickering level perceivable by human eyes, they defined the flicker parameter as ΔT/Tave = (Tmax-Tmin)/Tave, where Tave is the average transmittance level and ΔT is the transmittance variation by the signal polarity change. In Fig. 7(a), the flicker parameter values evaluated for the FFS n-LC cells aligned by the low resistivity PI layers dispersed with different MWCNT doping density conditions are presented. At operation frequency of 0.5 Hz, the FFS n-LC cells prepared under the MWCNT doping conditions into the PIs over 0.005 wt% exhibited sufficiently low ΔT/Tave levels without causing noticeable image flickering problems considering the flicker sensitivity boundary presented in Ref. [37]. In the case of the MWCNT doping condition of 0.05 wt%, the flicker parameter value of ΔT/Tave = 1.01% was achievable at the 0.5 Hz operation in spite of using the n-LC for the FFS mode. As shown in Fig. 7(b), the value of the flicker parameter increased with further reducing the operation signal frequency.
Fig. 7. Flicker parameter values (ΔT/Tave) of low-frequency-driven FFS n-LC cells, which are aligned by low resistivity LC alignment PI layers prepared with different MWCNT doping conditions. (a) ΔT/Tave value according to MWCNT doping density at the fixed signal frequency (0.5 Hz), (b) ΔT/Tave value according to the signal frequency at the fixed MWCNT doping density condition (0.05 wt%).
Figures 8(a) and 8(b) show the polarizing optical microscopic (POM) images of the low-frequency-driven FFS n-LC cell prepared with the low resistivity PI layer doped with 0.05 wt% MWCNTs. The completely dark and homogeneously bright images obtained at the field-off and field-on (Va = V20 with the 0.5 Hz square waveform) states reveal that the MWCNTs dispersed within the PI layer do not disturb both the initial NLC alignment and the field-induced LC reorientations unlike the CNT mixing approaches into the LC bulk [32]. In Fig. 8(c), the voltage-transmittance (V-T) curves for the FFS n-LC cells according to the MWCNT doping density conditions within the low resistivity PIs. To check the hysteresis properties, the V-T curves were measured by increasing (the solid lines) and then decreasing (the dotted lines) applied bias voltages at a sweep speed of the bias voltage of 150 mV/s. The data presented by the blue color are the V-T curves for the FFS n-LC cell aligned by the low resistivity PI dispersed with 0.05 wt% MWCNT doping. For comparison, the V-T curves for the FFS n-LC cell aligned by the pristine low resistivity PI without MWCNT doping are co-plotted with the red color. In our method, the MWCNTs are dispersed within the PI layers and the PI matrix layers are fully polymerized by the thermal imidization process before the LC cell preparation so that the MWCNTs doped within the low resistivity PIs do not affect the n-LC properties such as the LC ordering, the LC dielectrics, and the LC capacitance unlike the previous approaches of CNT-doping into the LC layer [31,32]. In Fig. 8(c), the FFS n-LC cell (0.05 wt% MWCNT doping into the low resistivity PI) exhibited the nearly identical V-T curves without the hysteresis property like those of the FFS n-LC cell without the MWCNT doping into the PI when the V-T curves were measured by changing the sweep direction of the bias voltages. This hysteresis-free operations of the V-T curves are also achievable in characterizing the V-T curves at different sweep speed condition (600 mV/s). When we measured the V-T curves for the FFS n-LC cell (0.05 wt% MWCNT doping into the low resistivity PI) after one week, there was no reliability issue with exhibiting almost identical V-T curves. When we compare the V-T curves of the FFS n-LC cells according to the MWCNT doping condition within the low resistivity PIs, most of the electro-optic properties such as the threshold voltage (Vth), the maximum transmittance level (Tmax), the voltage condition (Vmax) required for Tmax, and the V20 condition used in our experiment for the flickering evaluation were highly similar each other: Vth = 1.77 V and 1.74 V, Tmax = 14.23% and 14.19%, Vmax = 6.00 V and 5.85 V, and V20 = 2.11 V and 2.09 V for the FFS n-LC cells without and with MWCNT doping within the low resistivity PIs, respectively. The average transmittance levels of the PI layers were 93.74% (without the MWCNT doping) and 93.53% (with the MWCNT doping at 0.05 wt%) over the visible range. At our low doping condition, the light absorption by the MWCNTs within the PI is negligible.
Fig. 8. Voltage-dependent polarizing optical microscopic images of low-frequency-driven (0.5 Hz square waveform) FFS n-LC cells that are aligned on low resistivity PI prepared with 0.05 wt% MWCNT doping conditions at (a) Va = 0 V and (b) Va = V20 (voltage level required for 20% of the maximum transmittance). (c) Voltage-transmittance (V-T) curves for FFS n-LC cells according to MWCNT doping conditions within PIs, where the V-T curves are measured by increasing (the solid lines) and then decreasing (the dotted lines) applied voltages to characterize hysteresis properties (sweep speed of bias voltage: (+) or (−) 150 mV/s).
When we compare the V-T curves of two types of the FFS n-LC cells, the V-T curve slopes of the FFS n-LC cell aligned by the low resistivity PIs doped with the MWCNTs were slightly higher than those of the FFS n-LC cell aligned by the pristine PIs without the MWCNT doping in the middle ranges of the grayscales. The authors assume that this might be originated from the weakening of the n-LC anchoring after the MWCNT doping into the low resistivity PIs. When we checked the field-off and field-on response times (τoff and τon) for both samples, the response times were τoff = 14.68 ms and 31.49 ms, and τon = 18.10 ms and 16.52 ms for the FFS n-LC cells aligned by the low resistivity PIs without and with the MWCNT doping, respectively. Compared with the response times of the FFS n-LC cells aligned by the pristine PIs, the field-on response time became slightly faster but the field-off response time became significantly slower after the MWCNT doping into the low resistivity PIs. Compared with the field-on response time, the filed-off response time is more sensitive to the LC anchoring condition of the alignment layer [38,39]. Although most of the electro-optic properties of the FFS n-LC cell were much improved after doping the MWCNTs into the PI layer, the weakened LC anchoring needs to be enhanced by optimization of the MWCNT/PI mixtures as the further works.
For a higher MWCNT doping of over 0.1 wt%, the aggregation behaviors of the MWCNTs were observed from the POM images of the FFS n-LC cell at the field-on state. However, a small amount of MWCNT doping below 0.05 wt% is sufficient to reduce the operating frequency to a relatively low level without causing voltage holding issues.
3.4 Discharging properties dependent on MWCNT doping ratio within PI
Finally, we evaluated the discharging properties of the low resistivity PIs according to the MWCNT doping ratio for the FFS n-LC cells. For comparison, the discharging property of the high resistivity PI was measured as well. For this characterization, each FFS n-LC cell was biased by a DC voltage (at the positive or negative polarity of V20 level) for a sufficiently long time until the transmittance level variation becomes saturated by the surface-accumulated charges. Subsequently, by applying an opposite polarity of the DC voltage, the residual voltage curves, which are time varying with surface discharging, were obtained, as shown in Figs. 9(a) and 9(b). For quantitative analysis and comparison on the discharging properties according to PI types, the discharging coefficients (s−1) were obtained as the reversal of the decay time constant in discharging by using Eq. (5) and the residual voltage curves [23].
where Vr is the time-varying residual voltage, Vsat is the saturation voltage before the polarity change, Vrsat is reverse saturation voltage after the polarity change, and α is the discharging coefficient. The fitted values of the discharging coefficients of PIs were summarized in Fig. 9(c) according to the PI types and the MWCNT doping ratios.Fig. 9. Residual voltage curves for analyzing the discharging coefficients of (a) pristine high resistivity (ρ ∼ 1015 Ω·cm) PI and (b) MWCNT-doped low resistivity PIs (ρ ∼ 1013 Ω·cm for the PI without MWCNT doping). (c) Discharging coefficients depending on PI type and the MWCNT doping ratio in low resistivity PI.
Comparing the discharging properties of the pristine high and low resistivity PIs without MWCNT doping, the discharging coefficient value (α ∼ 0.0011 s−1) of the high resistivity PI was much lower than that (α ∼ 0.0084 s−1) of the low resistivity PI. Thus, as explained in Fig. 2(b), the FFS n-LC cell aligned with the high resistivity PI cannot avoid image sticking issues despite its relatively good voltage holding properties. After doping the MWCNTs into the low resistivity PIs, the discharging coefficient values were enhanced by approximately three times compared with that of the pristine low resistivity PI. According to the MWCNT doping condition, the resistivity of the PIs was evaluated by using the precision source-measurement unit (A2901, Keysight Technologies Inc.) after depositing Al electrode patterns on the PIs. Compared with the resistivity value (ρ = 3.47×1013 Ω·cm) of the pristine low resistivity PI, the resistivity of the PI doped with 0.05 wt% MWCNT was reduced to 6.52×1012 Ω·cm. The whole discharging mechanisms in the MWCNT-doped PIs would be complicated in our low MWCNT-doping condition, which would involve the charge drift effects within the PI matrix and the dispersed MWCNTs, and the field-induced charge transfer behaviors between the MWCNTs and the PI until eventual charge dissipation through the patterned top electrodes of the FFS structure. However, the residual voltage curves and the discharging coefficients shown in Fig. 9 obviously indicate that the metallic MWCNTs dispersed within the PI can effectively work as the distributed discharging paths even under a low MWCNT doping condition. In our FFS n-LC cell conditions, the discharging coefficients according to the MWCNT doping ratio ranging between 0.026 s−1 and 0.031 s−1, on average, did not indicate a meaningful difference. More specifically, at 0.05 wt% MWCNT doping that exhibited the most excellent VHR in our experiments, the average value of the discharging coefficient was approximately 0.026 s−1, i.e., 23.6 times higher than that for the high resistivity PI. The results indicate that the MWCNT-doped low resistivity PI layer can provide an effective solution for resolving VHR and image sticking issues in low-frequency-driven FFS n-LC modes.
4. Conclusion
We herein proposed doping MWCNTs into a low resistivity (ρ ∼ 1013 Ω·cm) PI layer to effectively reduce the bulk mobile ion density within n-LCs for low-frequency-driven FFS n-LC modes. Even though the FFS n-LC modes indicated less FE effects and higher brightness levels than the FFS p-LC modes, the FFS n-LC cell aligned by the low resistivity and high resistivity PI layer separately exhibited low voltage holding and poor surface discharging issues, respectively, thus hindering its application in the low frequency driving scheme for operation power saving. The capacitive current measurement results indicated that the MWCNT-doped low resistivity PI layer could effectively reduce the bulk mobile ion density within the n-LC owing to the excellent ion trapping effects of the MWCNTs, thus resulting in the highly improved VHR of the low-frequency-driven FFS n-LC aligned by the MWCNT-doped low resistivity PI layer. Interestingly, in contrast to the high resistivity PIs, the MWCNT-doped low resistivity PI layers provided high discharging properties with resolving the image sticking issues, which can be attributed to the metallic properties of the MWCNTs dispersed within the PI. More specifically, the n-LC mobile ion density (2.068 nC/mm3) measured for the pristine low resistivity PI condition without MWCNTs could be reduced to two orders of magnitude, i.e., 0.027 nC/mm3 for the n-LC cell aligned by the MWCNT-doped low resistivity PI at the doping ratio of 0.05 wt%. This mobile ion density level obtained for the n-LC cell was even lower by 1/10 compared with that for p-LCs. Owing to the effectively reduced bulk mobile ions within the n-LC, the FFS n-LC cell aligned by the MWCNT-doped low resistivity PI (0.05 wt% doping ratio) could be operated at extremely low operation frequency levels, thus exhibiting the excellent VHRs of 99%, 96.5%, and 95.2% for 0.5, 0.2, and 0.1 Hz operations, respectively. Meanwhile, the VHR of the FFS n-LC cell aligned by the pristine low resistivity PI deteriorated significantly to 63.1% and 7.8% at the operation frequencies of 0.5 and 0.2 Hz, respectively. In addition, the 0.05 wt% MWCNT-doped low resistivity PI layer demonstrated approximately 3.1 times and 23.6 times better discharging properties than the pristine low and high resistivity PIs, respectively, with resolving charge accumulation issues at the PI surfaces.
Funding
LG Display; National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (2019R1A2C1005531); BK21 Plus project funded by the Ministry of Education, Korea (21A20131600011).
References
1. S. H. Lee, S. L. Lee, and H. Y. Kim, “Electro-optic characteristics and switching principle of a nematic liquid crystal cell controlled by fringe-field switching,” Appl. Phys. Lett. 73(20), 2881–2883 (1998). [CrossRef]
2. D. H. Kim, Y. J. Lim, D. E. Kim, H. Ren, S. H. Ahn, and S. H. Lee, “Past, present, and future of fringe-field switching-liquid crystal display,” J. Inf. Disp. 15(2), 99–106 (2014). [CrossRef]
3. H. Kim and J.-H. Lee, “Fast falling time of fringe-field-switching negative dielectric anisotropy liquid crystal achieved by inserting vertical walls,” Appl. Opt. 54(5), 1046–1050 (2015). [CrossRef]
4. M. Kim, H. S. Jin, S. J. Lee, Y.-H. Shin, H. G. Ham, D.-K. Yang, P. J. Bos, J. H. Lee, and S. H. Lee, “Liquid crystals for superior electro-optic performance display device with power-saving mode,” Adv. Opt. Mater. 6(11), 1800022 (2018). [CrossRef]
5. Y. Jiang, G. Qin, X. Xu, L. Zhou, S. Lee, and D.-K. Yang, “Image flickering-free polymer stabilized fringe field switching liquid crystal display,” Opt. Express 26(25), 32640–32651 (2018). [CrossRef]
6. D.-J. Lee, J.-C. Choi, M.-K. Park, J.-S. Park, A. Ranjkesh, H. C. Choi, Y. M. Ha, S.-W. Choi, E.-J. Choi, and H.-R. Kim, “Optical measurement of flexoelectric polarisation change in liquid crystals doped with bent-core molecules using hybrid-aligned structure,” Liq. Cryst. 44(8), 1321–1331 (2017). [CrossRef]
7. D.-J. Lee, G.-Y. Shim, J.-C. Choi, J.-S. Park, J.-H. Lee, J.-H. Baek, H. C. Choi, Y. M. Ha, A. Ranjkesh, and H.-R. Kim, “Transient flickering behavior in fringe-field switching liquid crystal mode analyzed by positional asymmetric flexoelectric dynamics,” Opt. Express 23(26), 34055–34070 (2015). [CrossRef]
8. S.-W. Oh, J.-H. Park, J.-H. Lee, and T.-H. Yoon, “Elimination of image flicker in a fringe-field switching liquid crystal display by applying a bipolar voltage wave,” Opt. Express 23(18), 24013–24018 (2015). [CrossRef]
9. S.-W. Oh, J.-H. Park, J.-M. Baek, T.-H. Choi, and T.-H. Yoon, “Effect of electrode spacing on image flicker in fringe-field-switching liquid crystal display,” Liq. Cryst. 43(7), 972–979 (2016). [CrossRef]
10. M. S. Kim, P. J. Bos, D.-W. Kim, C.-M. Keum, D.-K. Yang, H. G. Ham, K.-U. Jeong, J. H. Lee, and S. H. Lee, “Field-symmetrization to solve luminance deviation between frames in a low-frequency-driven fringe-field switching liquid crystal cell,” Opt. Express 24(26), 29568–29576 (2016). [CrossRef]
11. S.-W. Oh, J.-M. Baek, J.-W. Kim, and T.-H. Yoon, “Dependence of image flicker on dielectric anisotropy of liquid crystal in a fringe field switching liquid crystal cell,” Jpn. J. Appl. Phys. 55(9), 094101 (2016). [CrossRef]
12. H. Lee, H. Kim, J. Kim, and J.-H. Lee, “Dependence of image flickering of negative dielectric anisotropy liquid crystal on the flexoelectric coefficient ratio and the interdigitated electrode structure,” J. Phys. D: Appl. Phys. 49(7), 075501 (2016). [CrossRef]
13. R. R. Hainich and O. Bimber, Displays: Fundamentals and Applications (CRC, 2016).
14. M. S. Kim, P. J. Bos, D.-W. Kim, D.-K. Yang, J. H. Lee, and S. H. Lee, “Flexoelectric effect in an in-plane switching (IPS) liquid crystal cell for low-power consumption display devices,” Sci. Rep. 6(1), 35254 (2016). [CrossRef]
15. R. Hatsumi, S. Fukai, Y. Kubota, A. Yamashita, M. Jikumaru, H. Baba, K. Moriya, D. Kubota, K. Kusunoki, Y. Hirakata, J. Koyama, S. Yamazaki, Y. Chubachi, and C. Fujiwara, “FFS-mode OS-LCD for reducing eye strain,” J. Soc. Inf. Disp. 21(10), 442–450 (2013). [CrossRef]
16. Y. Chen, Z. Luo, F. Peng, and S.-T. Wu, “Fringe-field switching with a negative dielectric anisotropy liquid crystal,” J. Disp. Technol. 9(2), 74–77 (2013). [CrossRef]
17. Y. Chen, F. Peng, T. Yamaguchi, X. Song, and S.-T. Wu, “High performance negative dielectric anisotropy liquid crystals for display applications,” Crystals 3(3), 483–503 (2013). [CrossRef]
18. H. J. Yun, M. H. Jo, I. W. Jang, S. H. Lee, S. H. Ahn, and H. J. Hur, “Achieving high light efficiency and fast response time in fringe field switching mode using a liquid crystal with negative dielectric anisotropy,” Liq. Cryst. 39(9), 1141–1148 (2012). [CrossRef]
19. I. Miyake, H. Asagi, M. Shimizu, K. Matsumoto, Y. Nishihara, Y. Hashimoto, T. Matsushita, S. Shimada, and Y. Ito, “Photo-alignment and n-FFS LCD technologies with IGZO-TFT applied to generation eight factory,” Dig. Tech. Pap. - Soc. Inf. Disp. Int. Symp. 47(1), 592–595 (2016). [CrossRef]
20. S.-S. Yoon, S.-H. Ahn, W.-Y. Choi, J.-H. Lee, H.-S. Kim, M.-C. Jun, and I.-B. Kang, “The study on DC resistibility of positive and negative dielectric anisotropy liquid crystal in AH-IPS mode,” Dig. Tech. Pap. - Soc. Inf. Disp. Int. Symp. 47(1), 1697–1699 (2016). [CrossRef]
21. D. Xu, F. Peng, H. Chen, J. Yuan, S.-T. Wu, M.-C. Li, S.-L. Lee, and W.-C. Tsai, “Image sticking in liquid crystal displays with lateral electric fields,” J. Appl. Phys. 116(19), 193102 (2014). [CrossRef]
22. T. R. Lee, J. H. Kim, S. H. Lee, M. C. Jun, and H. K. Baik, “Investigation on newly designed low resistivity polyimide-type alignment layer for reducing DC image sticking of in-plane switching liquid crystal display,” Liq. Cryst. 44(4), 738–747 (2017). [CrossRef]
23. D. H. Kim, J. H. Kim, Y. R. Kwon, S. H. Ahn, A. K. Srivastava, and S. H. Lee, “Investigation on ion movement in the fringe-field switching mode depending on resistivity of alignment layer and dielectric anisotropic sign of liquid crystal,” Liq. Cryst. 42(4), 486–491 (2015). [CrossRef]
24. W. Tie, I. H. Jeong, I. W. Jang, J. S. Han, Y. Liu, X.-D. Li, M.-H. Lee, K.-U. Jeong, and S. H. Lee, “Reducing driving voltage and securing electro-optic reliability of in-plane switching liquid crystal display by applying polysulfone photoalignment layer with photo-reactive mesogens,” Liq. Cryst. 41(8), 1057–1064 (2014). [CrossRef]
25. H.-Y. Kim, Y.-J. Jeon, J.-Y. Hwang, J.-H. Kim, and D.-S. Seo, “Residual DC characteristics in the IPS-LCD by capacitance–voltage hysteresis method on a polymer layer,” Curr. Appl. Phys. 2(3), 237–240 (2002). [CrossRef]
26. M.-C. Tseng, O. Yaroshchuk, T. Bidna, A. K. Srivastava, V. Chigrinov, and H.-S. Kwok, “Strengthening of liquid crystal photoalignment on azo dye films: passivation by reactive mesogens,” RSC Adv. 6(53), 48181–48188 (2016). [CrossRef]
27. J. M. Lee, J. H. Kim, H. J. Kim, Y. J. Lim, S. Kundu, M. S. Kim, M.-H. Lee, J. H. Lee, G.-D. Lee, and S. H. Lee, “Achieving a robust homogenously aligned liquid crystal layer with reactive mesogen for in-plane switching liquid crystal displays,” Liq. Cryst. 44(7), 1194–1200 (2017). [CrossRef]
28. Y.-J. Jeon, J.-Y. Hwang, and D.-S. Seo, “Voltage holding ratio and residual DC property of the IPS-LCD on rubbed polymer layers by voltage-transmittance hysteresis method,” Mol. Cryst. Liq. Cryst. 410(1), 369–380 (2004). [CrossRef]
29. C.-W. Lee and W.-P. Shih, “Quantification of ion trapping effect of carbon nanomaterials in liquid crystals,” Mater. Lett. 64(3), 466–468 (2010). [CrossRef]
30. F.-C. Lin, P.-C. Wu, B.-R. Jian, and W. Lee, “Dopant effect and cell-configuration-dependent dielectric properties of nematic liquid crystals,” Adv. Condens. Matter Phys. 2013, 1–5 (2013). [CrossRef]
31. B.-R. Jian, C.-Y. Tang, and W. Lee, “Temperature-dependent electrical properties of dilute suspensions of carbon nanotubes in nematic liquid crystals,” Carbon 49(3), 910–914 (2011). [CrossRef]
32. I.-S. Baik, S. Y. Jeon, S. H. Lee, K. A. Park, S. H. Jeong, K. H. An, and Y. H. Lee, “Electrical-field effect on carbon nanotubes in a twisted nematic liquid crystal cell,” Appl. Phys. Lett. 87(26), 263110 (2005). [CrossRef]
33. M. Mizusaki, T. Miyashita, T. Uchida, Y. Yamada, Y. Ishii, and S. Mizushima, “Generation mechanism of residual direct current voltage in a liquid crystal display and its evaluation parameters related to liquid crystal and alignment layer materials,” J. Appl. Phys. 102(1), 014904 (2007). [CrossRef]
34. K. Miyasato, S. Abe, H. Takezoe, A. Fukuda, and E. Kuze, “Direct method with triangular waves for measuring spontaneous polarization in ferroelectric liquid crystals,” Jpn. J. Appl. Phys. 22(Part 2), L661–L663 (1983). [CrossRef]
35. H. Takezoe, “Polar liquid crystals – ferro, antiferro, banana, and columnar –,” Mol. Cryst. Liq. Cryst. 646(1), 46–65 (2017). [CrossRef]
36. Y.-C. Hsiao, S.-M. Huang, E.-R. Yeh, and W. Lee, “Temperature-dependent electrical and dielectric properties of nematic liquid crystals doped with ferroelectric particles,” Displays 44, 61–65 (2016). [CrossRef]
37. H. Chen, F. Peng, M. Hu, and S.-T. Wu, “Flexoelectric effect and human eye perception on the image flickering of a liquid crystal display,” Liq. Cryst. 42(12), 1730–1737 (2015). [CrossRef]
38. X. Nie, R. Lu, H. Xianyu, T. X. Wu, and S.-T. Wu, “Anchoring energy and cell gap effects on liquid crystal response time,” J. Appl. Phys. 101(10), 103110 (2007). [CrossRef]
39. S.-W. Oh, D.-J. Lee, M.-K. Park, K. H. Park, J.-H. Lee, B. K. Kim, and H.-R. Kim, “Enhancement of viewing angle properties of a single-domain fringe-field switching mode using zero pretilt alignment,” J. Phys. D: Appl. Phys. 48(40), 405502 (2015). [CrossRef]
