1. Introduction

The sophisticated control and uniform alignment of liquid crystal (LC) molecules [1,2] have long been subjects of interest and are critical development of advanced multitasking, flexible LC displays (LCDs) [3]. Commercially, the mechanical rubbing of spin-coated polyimide (PI) layers has been widely used to provide topographical microgrooves that allow the uniaxial homogeneous or homeotropic alignment of LCs [4]. However, the rubbing process has significant drawbacks such as the generation of debris, electrostatic discharge and local defects as a result of the friction between the rubbing roller and the substrate [5]. Consequently, a number of noncontact alignment methods have been proposed as potential replacements for mechanical rubbing, including the photo-alignment method [6,7], self-assembled monolayers [8] and ion-beam irradiation [9–11]. however, although these noncontact approaches are free from electrostatic charge and dust generation they suffer from alignment instability, low anchoring energy, and image sticking as a result of capacitance hysteresis [12].

As an alternative, researchers have employed specific surface topographies to uniformly align the LC molecules using various methods including laser writing [13], electron-beam writing [14], microrubbing [15], dip-pen lithography [16], rigiflex lithography [17], and soft lithography [18]. Although surface patterning methods are not entirely free from dust generation and contamination, they have been intensively studied because the alignment of the LC molecules in these microscale or nanoscale patterns allows more freedom to control the alignment properties (i.e., pretilt angle, anchoring energy, multistability) than conventional methods. Recently, our group introduced laser interference lithography (LIL) to LCD applications for the first time [19]. Using this nanoimprint lithography, we quickly and uniformly fabricated periodic nanosized structures, over large areas, and transferred this nanopattern onto a homeotropic PI surface for uniform LC alignment. However, the resultant LC device switching (both the low threshold voltage and rapid response time) on such a series of surface patterned alignment layers, including our nanoimprint lithography, was inferior to that of LCDs containing PI alignment layers [19,20]. Moreover, most of the patterned LC alignment layers exhibited weak optical properties because the surface pattern induces light loss due to the low-order-parameter LC alignment and the surface graing generates scattering [21–24]. Therefore, a new fabrication method is needed for advanced LC alignment layers that can simultaneously solve the aforementioned problems.

In this study, we demonstrate an simple strategy for fabricating novel LC alignment layers that ensures perfect LC alignment, superior switching without any capacitance hysteresis, low transmittance loss, and high thermal stability with sufficient anchoring action. Thin films (approximately 10 nm) of reactive mesogens (RMs) were transferred onto conventional homeotropic PIs from a UV-cured RM stamp via contact printing. Using linearly polarized UV exposure, we made optically anisotropic RM stamps with directionally cross-linked networks of photopolymerized RM monomers. When detached from those stamps, the resulting RM films not only successfully align the LC molecules in the homeotropic state, but also fill the microdefects on the surface of PI surfaces leading to excellent thermal stability and hysteresis free operation. Our LC devices on the defect free RM/PI polymeric stacks exhibited superior electro-optic (EO) properties to those containing rubbed PI layers; a 38.3% decrease in the V_th and a 26.1% reduction in the response time were observed for a vertically aligned (VA) mode LCD. Moreover, owing to suppressed light scattering, a flat surfaces of RM layers induced higher optical transmittance (81.4%) than that induced by nanopatterned PIs (77.5%) [19]. The RMs is a excellent candidate in itself as a LC alignment layer [25], but the RM/PI polymeric stacks show a significantly improved properties [26–28].

2. Experimental

To fabricate the RM stamps, RM (RMS03-015, Merck) monomers were coated onto 1-µm-thick flexible films via spin coating (3000 rpm for 30 sec). These films were cured using a linearly polarized ultraviolet (UV) exposure system (Oriel Co.) to obtain a uniform UV energy density with 80% of its original UV energy. The source was a 1 kW mercury lamp connected to the lamp supply. Using a convex lens, the films were exposed to UV radiation at wavelengths of 310~330 nm and an energy density of 7.9 mW/cm². The incident angle of UV the exposure was 90° with an exposure time of 3 min. Using a spin coater, a 50-nm-thick homeotropic polyimide (PI, JALS-696-R2, JSR Co. Ltd) was coated onto indium tin oxide (ITO)-deposited glass substrates followed by pre-baking on a hot plate at 80°C for 10 min and hard baking in an oven at 230°C for 1 h. After these processes, we pressed the prepared RM stamps onto the PI/ ITO target substrate for 10 min and then peeled them away. The rubbing is performed on the PI coated substrates. The rubbed substrates were attached antiparallel direction which cell gap is 60 um. After fabrication, LC is injected into the cell via capillary force to obtain information on the alignment properties which is the photomicroscope (BXP 51, Olympus) image. The negative LCs (Tc = 75 °C, Δε = −4, Δn = 0.777, MJ98468, Merck Corp.) were injected into each cell via capillary injection at room temperature. To confirm the optical anisotropy and the compositional transformations of the RM-transferred PIs, we measured the optical retardation (REMS-100, Sesim) and employed X-ray photoelectron spectroscopy (XPS, VG Microtech ESCA2000), respectively. To measure the electro-optic (EO) properties, we fabricated vertically aligned (VA) LC cells. The VA cells had cell gaps of 5 μm to allow measurement of the EO characteristics and hysteresis. The EO measurements of the voltage-transmittance (V–T) characteristics and response times were confirmed using an LCD evaluation system (LCMS-200). An LCR meter (4284A, Agilent) verified the dielectric constants and the residual DC voltage properties of the VA cells at room temperature.

3. Results and discussion

Figure 1 depicts the complete process for the preparing the RM stamps and their use in fabricating the VA-LCDs as a test type. Optically anisotropic liquid crystalline RM stamps were fabricated under a linearly polarized UV exposure system that ensures directional photo-polymerization of spin-coated RM monomers on a flexible substrate. The base homeotropic PI layers were then spin-coated onto indium tin oxide (ITO)-deposited glass as target substrates, which were subsequently soft and hard baked. To transfer the RM thin films from the flexible stamps onto base PI substrates, the RM stamps were placed into contact with the PI-coated ITO substrates for 10 min and then peeled away. Using this contact imprinting method, a 10-nm-thick RM layer was printed onto the target PIs for LC alignment and switching. Finally, negative LCs were injected into sandwiched LC cells using RM transfer to prepare the VA-LCDs. The resultant LCDs exhibited perfect LC alignment and higher EO characteristics than the conventional VA-LCDs with rubbed or nanopatterned PIs.

 

Fig. 1 Schematic representation of the RM imprinting processes. Thin films of RMs were transferred from RM stamps to base PI layers to fabricate high-performance VA-LCDs.

Download Full Size | PDF

To investigate the alignment states of the LCs on PIs and the transferred RM surfaces, we used polarized optical microscopy (POM), as shown in Figs. 2(a) and 2(e). Whereas a schlieren texture was observed in the LCs (random alignment of the LCs, Fig. 2(a)) on the pure PIs without any surface treatment, a local alignment of the LCs was induced on the surface of the RM layers transferred from the nonpolarized UV-cured RM stamps (Figs. 2(b) and 2(c)). To achieve uniform and full alignment of the LCs, we created RM stamps via linearly polarized UV curing and applied them to the fabrication of LC alignment layers to obtain defect-free, large-area alignment of the LC molecules (Figs. 2(d) and 2(e)), which is essential for the various display applications.

 

Fig. 2 (a) Random orientation of LCs on a pure PI surface. (b) Locally aligned LCs with a crossed polarizer and (c) a parallel polarizer in the same region. The areas within the dotted-lines indicate the local alignment state of the LCs. Perfectly aligned LCs with crossed (d) and parallel polarizers (e). (f) Rotation angle dependence of the optical retardation for rubbed, RM transferred, and pure PI surfaces.

Download Full Size | PDF

We then measured the optical retardations of the target substrates to trace the LC alignment behavior on the different PI and RM surface layers as shown in Fig. 2(f). Optical and topographical anisotropies of the alignment layers are key factors in ability of LC molecules to align. In particular, optical retardation measurements using ellipsometry for visible light can evaluate the anisotropy of target substrates, and thus determine the optimization conditions for fabricating UV-cured RM stamps for LC alignment and switching [29,30]. As a result of the optical retardation measurement, almost no retardation was observed in the pure homeotropic PI films, indicating that we cannot align the LCs on a PI surface without an alignment process. In contrast, the value of the retardation increased as we applied RM imprinting to this optically isotropic PI base substrate. We slightly increased the retardation of the LC alignment layers by transferring the RM layers from the RM stamps under nonpolarized UV curing. However, compared with the retardation of conventional rubbed PIs for LC alignment, this value is insufficient for achieving full alignment of the LCs. To increase the anisotropic characteristics of the RMs to the level of the rubbed PIs, we imprinted the RM layers using the stamps fabricated under linearly polarized UV exposure. The resulting RM layers exhibited significantly enhanced optical retardation and anisotropy, which can induce perfect LC alignment as with conventional rubbed PIs. These results were in good agreement with the LC alignment trends presented in the POM images.

We also conducted a compositional investigation of the RM transfer on PI surfaces via XPS analysis [31]. First, we verified the photopolymerization of the RM monomers under polarized UV exposure to fabricate the RM stamps. The XPS spectra for the C 1s and O 1s of the RM stamp before and after UV exposure are provided in Figs. 3(a) and 3(c), respectively. The dotted envelopes from each of the Fig. 3 were fit well with experimental data. The XPS spectra were analyzed with a fitting routine, which can decompose each spectrum into individual sub-spectra for the information of specific bonding. The raw experimental data were used with no preliminary smoothing. The core-level XPS spectra of the C 1s were decomposed into three components as shown in Fig. 3(a). The low-binding-energy component peak centered at 284.6 eV is due to the C-C bond [32], and the peak at 286.2 eV is related to the C-O bond [33]. Another component peak at 288.5 eV corresponds to the C = O double bond [34] in the RM monomers. After polarized UV exposure, the atomic percentage of the C-C single bond was markedly increased to approximately 20.6%, and the C-O single bond increased to approximately 8.7%, whereas that of the C = O double bonds decreased to 13.6%. These trends indicated that the polarized UV exposure selectively broke the C = O double bond to generate radicals that then formed C-C and C-O single bonds, forming linear cross-links between the RMs via photopolymerization. In addition, Fig. 3(c) provides the XPS spectra of O 1s. The component peak centered at 531.6 eV corresponds to the O = C double bond [35], and the peak at 533.0 eV is due to the O-C single bond [36]. Under UV exposure, the atomic percentages of the O = C and O-C were reduced to 54.0,% and 27.0%, respectively, because the weak O 1s bonds were destroyed in the photopolymerization of the RMs [37]. The LC core structures of the RMs may have remained ordered in a homeotropic state due to the linear polarization of the UV radiation, but they ceased to move freely in the LC phase leading to the fabrication of the RM stamp films. (no chemical modification of the LC core occurred during the polymerization.)

 

Fig. 3 XPS spectra for C 1s and O 1s peaks confirm the photo polymerization of the RM stamps and their transfer onto the PI films with contact imprinting. (a) C 1s peaks for the RM stamp before and after UV exposure. (b) C 1s peaks for the base PI surface before and after RM stamp imprinting. (c) O 1s peaks for the RM stamp before and after UV exposure. (d) O 1s peaks for base PI surface before and after RM stamp imprinting. The insets indicate the changes in the RM and LC alignment states before and after UV exposure (a and c), and demonstrate the transfer of RM thin films using contact imprinting.

Download Full Size | PDF

Next, we compared the XPS spectra of the PI substrates before and after imprinting to confirm is that the thin layers of RMs were transferred from the stamps to the base PIs via contact imprinting. Figure 3(b) presents the changes in the core-level XPS spectra of the C 1s peaks on the PI-coated substrate. The component peaks centered at 284.6, 286.2 and 288.5 eV correspond to the decomposed peaks of the RM stamps. The atomic percentage of the C-C single bond increased by approximately 19.7%, and that of the C-O single bond increased significantly by approximately 25.8%. The C = O double bond atomic percentage was also increased by 36.4%, indicating that the increased atomic percentages of the C-C, C-O, and C = O bonds were attributed to the transfer of the RM layers onto the PIs through contact imprinting with the UV-cured RM stamp. Similarly, after imprinting, the atomic percentage of the O = C and O-C bonds on the PIs increased by 20.9% and 34.2%, respectively, due to the transferred RM layers, as shown in Fig. 3(d).

The optical loss is critical factor that should be considered before applying novel surface patterns or imprinting methods to real LCD structures because the surface patterns or imprinted layers can decrease the transparency by generating light scattering through the surface gratings. To compare the optical transmittance of our RM/PI alignment layers with conventional rubbed PI, we employed an ultraviolet-visible spectrophotometer (V-650, JASCO, the range of visible region is 380-780 nm) as shown in Fig. 4(a). The average optical transmittance of the RM/PI and rubbed PI were 81.4% and 81.2%, respectively, indicating that no optical degradation occurred due to the imprinted RM layers on the base PI. Because the imprinted RM thin films have no topographical structures, we can minimize the optical loss within the interface between the alignment layers and the LC media. In fact, our previous work on nanopatterned PIs using LIL recorded an average transmittance of 77.47% due to the light scattering [19].

 

Fig. 4 (a) Optical transmittance spectra of an ITO/Glass, PI/ ITO/Glass, Polarized UV imprinted cell, and rubbed cell. The EO characteristics of the VA-LCDs based on RM imprinting, nanopatterned PIs using LIL and conventional rubbed PIs are shown in (b-c). (b) V-T characteristics for VA-LCDs demonstrating threshold voltages for device operation. (c) Rise and fall response times of each VA-LCD. (d) CV capacitance hysteresis curves for both the VA-LCDs with RM imprinting and the rubbed PIs that indicate the residual DC and possibilities for image sticking. Imprinted RMs cover the surface defects on the base PI, leading to high-performance switching and hysteresis-free operation.

Download Full Size | PDF

Superior EO characteristics, including a decreased V_th and rapid response time, were achieved in the VA-LCDs based on imprinted RMs, as shown in Fig. 4(b). The VA-LCDs containing rubbed and nanopatterned PIs using LIL were used for comparison. The voltage-transmittance (V-T) curves confirmed that the homeotropically aligned LC could be switched from OFF to ON under an external voltage above a certain threshold (V>V_th). The V_th value for the RM based VA-LCDs (at 10% transmittance) for device switching was reduced by 8.4% compared with the conventional VA-LCDs with rubbed PIs, and was 28.9% lower than that of the VA-LCDs containing nanopatterned PIs using LIL. The time-transmittance plots for the imprinted RMs, rubbed PIs, and nanopatternd VA-LCDs yielded the rise and fall response times provided in Fig. 4(c). Of these, the VA-LCDs on the imprinted RMs recorded the most rapid response time of 26.5 ms (rise and fall times of 11.9 and 14.6 ms, respectively). The VA-LCDs on rubbed PIs exhibited a response time of 36.1 ms (rise and fall times of 15.3 and 20.8 ms, respectively), and the VA-LCDs on nanopatterned PIs had a response time of 27.5 ms (rise and fall times of 10.2 and 17.3 ms). We believe that the superior EO characteristics (both the low V_th and rapid response time) of the RM-imprinted LCDs could be attributed to the defect-filling effect of the RMs on the PI surfaces [27]. The imprinted RMs can fill the microdefects on the PI surfaces (generated from the various contact processes and after annealing) that can disturb the molecular reorientation and increase the molecular collision of the LCs within the interface between the LCs and the alignment layers.

Image sticking, in which the image established during the device-on state survives and gradually fades with time even when the device is off, is among the most important issues for developing a high-quality display with long LCD lifetimes. Image sticking can be characterized by the capacitance-voltage (C-V) hysteresis of the LC cells, which arises from the residual charges accumulated in the localized defect regions as the voltage is switched on [38]. The time-dependent dissipation of the trapped charges causes image sticking. Unlike the C-V curves that exhibit characteristic hysteresis from LC cells prepared on conventional rubbed PIs, a nearly hysteresis-free C-V curve for an LC cell was obtained on the RM-imprinted PIs, as shown in Fig. 4(d). This result also indirectly demonstrates that the imprinted RM layers covered the defects on the PI surface, which can trap charges and increase C-V hysteresis. Consequently, the LCDs containing RM-covered PI were nearly free of residual charges within the LC/alignment layer interfaces.

As indicated in Fig. 5, the imprinted RMs that filled the defects on the PI surface also enhanced the thermal stability of the conventional, rubbed, PI-based LC cells. The increased thermal budgets in advanced LCDs, such as microdisplays and high-brightness displays, deteriorate the organic materials, thus decreasing the lifetime of the displays. Thus, novel LC alignment layers with high thermal stability are essential for highly durable displays [39–41]. The LCDs on our RM-imprinted PIs provide improved high-temperature operation over that of pure PIs due to their increased thermal stability, which arises from the aforementioned surface defect filling. The LC cells with rubbed PIs were stable up to 150 °C but became unstable and developed locally aligned states at 210°C, whereas the RM-imprinted cells exhibited improved thermal stability up to 210 °C. Both LC cells were annealed from 90 to 210 °C for 10 min.

 

Fig. 5 Thermal stability tests on the LC cells for conventional rubbed PIs and imprinted RM layers.

Download Full Size | PDF

4. Conclusion

In conclusion, we demonstrated high performance VA-LCDs with RM-transferred PI alignment layers that were contructed using contact imprinting. The fabrication of an RM stamp with polarized UV curing and its successful transfer onto a base PI were important because they resulted in both perfect alignment of the LCs and superior EO characteristics. Moreover, the VA-LC cells with our RM-imprinted alignment layers exhibited reduced optical loss and high thermal stability up to 210°C, even without capacitance hysteresis. We also elucidated the optical and compositional changes in the RM transfer process to target the PI surfaces by employing optical retardation measurement and XPS analysis, which could lead to increased precision in fabrication and the application of RM stamps for advanced flexible LCDs. We believe that by changing the combinations of RMs, base PIs and LCs, high-performance and various-mode LCDs such as twisted nematic (TN), in-plane switching (IPS), and optically compensated bend (OCB) mode LCDs can be practically fabricated.