Summary

The difference of the intensity operational range results from the various profiles of LC director after polymerizations. In the case of the pi cell stabilized in the RB state, the LC director throughout the device is stabilized in a more relaxed (i.e. lower tilt) state, which thus has higher freedom of switching compared with the pi cell stabilized with conventional polymerization. To further increase the intensity operational range, we have tried to lower the concentration of the reactive mesogen for synchronized polymerization (it was reported in ref. 62 that using lower concentrations results in higher intensity operational range); however, using concentrations lower than 4% produces only partial stabilization of the bend state.

We have demonstrated a synchronized polymer-stabilization technique which can

be used to stabilize the non-permanent states in LC devices. In this thesis, the non-permanent RB state reported to have high brightness [63] is observed and stabilized by our proposed technique. This pi cell stabilized in the RB state has the attributes of high intensity operational range (an enhancement is obtained by a factor of 1.5 compared with the best case of the conventionally polymerized pi cell) and feasibility for full dynamic range optical compensation (no twist state occurs during the device operation). This proposed pi cell device can be used with a switching backlight technique to achieve a complete dark state and lower the power-consumption [70].

Chapter 5

Investigation of symmetric H state in a pi-cell

5.1 Introduction

As mentioned in chapter 4, the normal pi-cell needs to be primed to the bend state and sustained with a critical voltage to prevent the undesirable recovery into the twist or splay state [16, 32, 57, 58, 71-75]. In contrast, the symmetric H (Hs) state (a transient state obtained by a sudden application of voltage to the ground splay device) needs no priming – i.e. the Hs state is continuous with the ground state.

Moreover, the Hs state has been reported to have the merit of very fast switching, on a scale of 1 millisecond (as shown in Fig. 5-1) [14]. Because of these distinguished features, the Hs state has attracted much attention. In this chapter, we will further investigate its possibility for display applications.

Fig. 5-1 Hs state director profile and its response time [14].

The reason that hinders the Hs state from LCD applications is its short life time (of less than 200 ms, in general). In this chapter, our major objective is to come up with a method for extending its lifetime. Before going to design a structure for this purpose, a direct investigation on the LC profile of the Hs is necessary. So, in the beginning of this chapter, we propose a method for verifying the LC profile of the Hs state, which

was understood only by calculation, but we prove it with direct observation.

After confirming the LC profile of the Hs state, we further propose a method to extend its lifetime. The simulation of this proposed method will be reported, and the fabrication process will also be introduced.

5.2 Investigation of the transient symmetric H state 5.2.1 An introduction of the symmetric H state

In previous work, it was suggested that under field application, the Hs state has an internal director structure in which the director in the centre of the device remains parallel to the surfaces. This “de-couples” the two halves of the pi-cell, thus the cell is effectively divided into two half-thickness Fréedericksz devices, from which the fast switching behavior results. The switching rate enhancement, ignoring the flow effect of the pi-cell, can be explained by rotational viscosity, and K11, the splay elastic constant. Thus, the smaller the effective thickness of the switching layer, the faster the device. Some modeling and experimental results have shown the switching rate of Hs state is faster than that of asymmetric H (Ha) state in a pi-cell by a factor of 4, which supports the existence of the central non-switching region, since it de-couples the LC director in the two halves of the cell [14, 76, 77]. This central region and de-coupled switching make the Hs state have a symmetric director profile as shown in Fig. 5-1; however, owing to its short lifetime, no direct evidence has been presented to demonstrate the profile symmetry.

To elucidate the short lifetime issue, a continuous sine waveform is applied to the

pi-cell to observe the phase transition. As shown in Fig. 5-2, when the voltage is applied, the splay (S) state is suddenly switched to the Hs state without nucleation.

After that, the asymmetric H (Ha) state forms in the Hs domain and finally engulfs the whole area of Hs state. Meanwhile, the domain wall between the bend state and the Ha state is still moving toward the Ha state side. When the applied voltage is removed, the Ha state turns back to the S state and the bend (V) state turns back to the twist (T) state.

There are two interesting phase transitions observed. One happens during the recovery form V state to T state; the other, Ha state to S state. As the V state recovers to T state, some textures appear along the rubbing direction, as shown in Fig. 5-2. The texture may result from the collision between the LC directors with different laying down directions. These explanations have demonstrated that the bend state is the most stable state among the states in a pi-cell while applying a voltage larger than the critical voltage. Thus, the bend state tends to engulf the Hs state.

Fig. 5-2 Phase transition by continuous driving waveform.

5.2.2 Preliminary observation on the symmetric H state

To quantify the lifetime of the Hs state and observe the collapsing mechanism, we make the experiment as follows: Initially, a typical pi-cell (with 2.6 um cell gap, filled with LC material E7 and using parallel rubbed polyimide as the alignment layers) is positioned at 45° between the crossed polarizers and the transmission is measured during signal application. With an impulse voltage signal of 5Vrms, the intensity variation during the state transitions from the ground splay state, to the Hs and into the Ha states is observed by a photo-detector as shown in Fig. 5-3. The Hs state is observed to have a life-time dependence on the applied voltage. As the applied voltage is increased, generally the life-time of Hs state (in the duration of the Hs plateau in Fig. 5-3) will be longer. However, at higher voltages, the bend state rapidly nucleates and engulfs the whole area of the splay states.

Fig. 5-3 The state transition from splay to Hs to Ha state.

5.2.3 Stroboscopic illumination assisted conoscopy

As the aforementioned descriptions, the difficulty in proving this Hs state LC profile by direct observation was due to the short life-time of Hs state (typically around a few tens ms, although it can be present for hundreds ms in certain circumstances). In this section, we report work where we utilize a burst driving method along with stroboscopic light emitting diode (LED) illumination to capture the conoscopic images for symmetric LC director profile of the Hs state.

It is clear that we can verify the LC profile by comparing the measured conoscopic images with the calculated images. Since the conoscopic contour is affected by LC symmetry, as shown in Fig. 5-4, the contours of Hs state and Ha state are different.

The methodology of this verification is shown in Fig. 5-5. However, it was not possible to be carried out because the actual case of measuring conoscopic image takes several minutes, while the lifetime of Hs state is of less than 200 ms. To overcome this issue, we propose to use a stroboscopic illumination along with the conventional conoscope for verifying this transient LC profile.

To observe the conoscopic images of each state during the Hs state collapsing, the device is driven by a burst waveform for Hs state formation and a continuous waveform for Ha state formation. As shown in Fig. 5-6, the burst driving waveform is composed of two parts: an operating time and a delay time. The operating time is determined by the need to switch the device into the Hs state but avoid breaking down into the Ha state(s) (or transitioning into the bend state). The delay time allows recovery of the ground splay state. In this case, the operating time is set as 10 ms, and the delay time is set as 90 ms; meanwhile, the stroboscopic LED illumination is set to delay from the start of the switching signal by 5 ms (to allow the formation of Hs state), then to illuminate for 5 ms and be off for 90 ms (to synchronize with the device switching signal). This synchronized driving scheme ensures that only the Hs state is

captured by the conoscope. To further check the results, the conoscopic images are obtained with two different wavelengths of light source (a blue LED with the wavelength of 436-486 nm and a red LED with the wavelength of 622-654 nm).

Fig. 5-4 LC director profiles of splay, Hs, and Ha states in a pi-cell.

Fig. 5-5 The methodology of measuring the LC director profiles of splay, Hs, and Ha states in a pi-cell.

Fig. 5-6 Burst driving waveform for the device and illuminating LED.

5.2.4 Simulation on transient symmetric H state

Based on the Frank-Oseen continuum theory as discussed in chapter 2 [78, 79], the equilibrium director profiles in the LC cell can be calculated. Using typical LC material parameters (those for E7) and initializing from a ground splay state with a slight asymmetry (to allow eventual formation of the Ha state), we can obtain the director profiles of each state as shown in Fig. 5-7. After calculating the director profiles, an extended Jones matrix technique [80] can be used to determine the conoscopic images of each state for wavelengths corresponding to using blue and red LEDs as the light sources (examples of which are shown in Fig. 5-8).

(a) (b) (c) Fig. 5-7 The director configurations of (a) splay, (b) Ha, and (c) Hs states.

5.2.5 Results of the investigation on symmetric H state

The conoscopic measurement results within a viewing cone of around 30° show directly the symmetry of ground splay and transient Hs states, and also the asymmetry of the Ha state. In addition, the modeling based on the theory outlined above is used to determine the conoscopic images within the viewing cone of around 30°. The experimental and theoretical results are shown in Fig. 5-8, the measured conoscopic images (a)~(f) are in good agreement with the modeling results (g)~(l) (at least for the restricted viewing cone of around 30° – it is difficult to obtain good agreement over very wide viewing cone angles due to the illumination system used). The asymmetry in the Hs state is evident in both the experimental results (Fig. 5-8 images (c) and (d)) and the modeled images (Figs. 5-8 (i) and (j)). The symmetry of the ground splay state is as expected (Figs. 5-8 (a), (b), (g) and (h)). More important for this work is the symmetry evidence in Figs. 5-8 (e) and (f) (experimental) and (k) and (l) (theoretical).

The good agreement between these data is the evidence that the director profile of the Hs state is indeed as expected, with decoupling of the director in the two halves of the cell. Moreover, according to our modeling, if the center point (zero-tilt) of the director structure is off center by 5% of the thickness of the device, then the conoscopic image is off axis by 15°. Thus, we can be confident that the method used here is very sensitive to asymmetry in the structure.

We have confirmed the existence of the symmetric profile of the LC director in the Hs state by stroboscopic conoscopic imaging. Along with the modeling, the transient non-switching director in the center of the device and the consequent de-coupling of the directors in the two halves of the device have been verified to occur. These de-coupled directors divide the device into two half-thickness Fréedericksz layers which result in the fast-switching behavior.

(a) (b) (c) (d) (e) (f)

(g) (h) (i) (j) (k) (l) Fig. 5-8 Measured conoscopic images of device: (a) and (b) in splay state; (c) and

(d) in Ha state; and (e) and (f) in Hs state. Simulated results: (g) and (h) in splay state; (i) and (j) in Ha state; and (k) and (l) in Hs state. The cases of (a), (c), (e), (g), (i) and (k) are illuminated/modeled with a blue LED, while the cases of (b), (d), (f), (h), (j) and (l) are illuminated/modeled with a red LED. (Each set of data has been normalized to optimize the image and exploit the image’s full dynamic range.)

5.3 Design for extending the lifetime of symmetric H state

5.3.1 A proposed structure for fixing the central LC director of Hs state With the stroboscopic illumination assisted conoscopy, we have verified the LC director profile of the Hs state. The central LC director drifting to one of the substrates results in the collapsing of Hs state. Based on this observation, it is possible to extend the lifetime of Hs state if we can fix the central LC director.

According to the report [81], the LC director tends to follow the polymerized structure in the LC cell. Therefore, if we can form the structure to sustain the Hs state, then its lifetime can be extended. Referring to the method that we used in chapter 4, if we use the synchronized polymerization along with photo-mask, then we will be able to form the structure as shown in Fig. 5-9, which thus fixes the central LC director and further extends the lifetime of Hs state.

Fig. 5-9 Concept of Hs lifetime extension by polymer stabilization.

5.3.2 Simulation on the lifetime of proposed device structure

Before going for the execution of polymerization, we have also done some preliminary simulations to find out the best parameters for the shape of polymer wall.

The simulation of polymer wall can be simplified by using a boundary condition. This concept is shown in Fig. 5-10, where the top and bottom substrates are set as strong anchoring with pretilt of +2° and -2°. Besides, the polymer walls in the middle of the cell are also set as strong anchoring with pretilt of 90° (perpendicular to the polymer wall). In this simulation, the pixel size (length A), cell gap (length B), and with or without polymer walls are altered to figure out the relation between the lifetime of Hs state and these factors.

Fig. 5-10 Structure used for simulating the Hs lifetime extension.

The simulation results show the possibility of realizing the Hs state lifetime extension. Initially, the LC director profile is set as splay state as shown in Fig. 5-11 (a). During the following 10 ms, the LC director profile is then gradually shaping into the Hs state as shown in Fig. 5-11 (b). After the Hs state formed, the LC directors then collapse via two paths: one is via twist motion as shown in Fig. 5-12 (a); the other is via central director drifting as shown in Fig. 5-12 (b). To model the actual pi cell as a reference case, a 20 um pixel size and 4 um cell gap without polymer wall is simulated, and the electric field is set as 1.25 V/um. This referenced case collapses into twist motion within 40 ms and its central directors are drifted to the substrate within 150 ms.

Moreover, we model several cases of the LC cell, and the results are listed in table 5-1. The factors of pixel size, cell gap, with or without polymer wall, and the applied electric field are varied. These cases have indicated some interesting results: Cases A and B show that the pixel larger than 10 um has no difference. Cases D and E show that the polymer wall effect is very significant on extending the lifetime of Hs state.

Cases E, F, and G along with cases J and K show the cell gap effect on the lifetime.

Cases H and I show the effect of applied filed. Cases C and E show the effect of pixel size.

In the case of 9 um pixel size, 4 um cell gap, 1.25 V/um electric field, and with strong anchoring polymer wall, the director profile is sustained in the state of Hs even after 2000 ms as shown in Fig. 5-13. Thus we can say that the lifetime is extended to more than 2000 ms (larger than the referenced case by a factor of more than 10), which means our actual cell with lifetime of 100 ms will be possible to extend its lifetime to 1sec (it is long enough for display applications).

(a) (b)

Fig. 5-11 Initial stage of simulation setup, where (a) represents the splay state at t=0 and (b) represents the Hs state at t=10 ms.

(a) (b)

Fig. 5-12 Two cases of Hs state collapsing via (a) twist and (b) Ha state formations.

Fig. 5-13 LC director profile after 2000 ms in the best case of Hs state lifetime.

Table 5-1 Simulation results of Hs state lifetime extension with polymer wall.

Pixel size (um)

Cell gap

(um) polymer wall Vertical field (V/um)

5.3.3 Polymer-stabilization without incorporating a photomask

The concept of the synchronized system is that the mesogen in a pi-cell is only exposed under UV light at the Hs state. Thus, the polymer network will be formed to maintain the transient configuration of Hs state; this means that the Hs state is less likely to collapse into the Ha state, as a result, the lifetime of Hs state will be extended.

This idea can be carried out by inserting a chopper between the UV light source and the pi-cell. When the UV light pass through the chopper, a photo-detector is used to detect the UV light and trigger the pi-cell to be turned on to the Hs state, by which the optical signal coming from UV light and the voltage signal applied to pi-cell will be synchronized, as shown in Fig. 5-14, which has been reported to be able to stabilize the non-permanent states [82].

The oscilloscopic image is captured to explain the idea of synchronized polymerization. As shown in Fig. 5-15, curve 2 is the UV light signal detected by a photo-detector used to trigger the pi-cell. Curve 1 is the light signal from the pi-cell,

and the driving waveform on this pi-cell is set to be 4 Vrms sine-wave operating for a period of 50 ms and then rest for a period of 200 ms to ensure that only Hs state is formed during the operation time. Then, another photo-detector is used to monitor the UV light signal (signal 3) to ensure only Hs state is formed under the UV light exposure. From this data, we can be confident of the accuracy of synchronized polymerization where the pi-cell is only exposed under UV light at the Hs state.

Fig. 5-14 Concept of synchronized polymerization.

Fig. 5-15 Oscilloscopic image of synchronized signals, where curve 1 represents the optical signal of LC device, curve 2, triggering signal, and curve 3, monitored

UV light signal.

We have tried several testing runs of different conditions for extending the lifetime of Hs state in a pi-cell. The results are summarized in table 5-2. We find that while using UV absorbing LC material such as E7, owing to the non-symmetric exposure of two sides of the pi-cell device, it is likely to form an only one Ha state pi-cell after polymerization, whereas in the case of non-UV-absorbing LC material such as ML1001, the device is very likely to form 2 Ha states with spatially homogeneous distribution.

Varying concentration does not bring any systematical results. In the case of 0.5 wt%, the device before being exposed under UV light get slight lifetime extension of Hs state which may result from the change of elastic properties. As for the case of 2.8 wt%, the state is stabilized in the Ha state within 2 minutes. While in the case of 1.5 wt%, we record the variation during the polymerization and found that the intensity is suppressed gradually as shown in Fig. 5-16. This verifies that the LC directors near the substrates are stabilized; however, the lifetime of Hs is still shortened and the

Varying concentration does not bring any systematical results. In the case of 0.5 wt%, the device before being exposed under UV light get slight lifetime extension of Hs state which may result from the change of elastic properties. As for the case of 2.8 wt%, the state is stabilized in the Ha state within 2 minutes. While in the case of 1.5 wt%, we record the variation during the polymerization and found that the intensity is suppressed gradually as shown in Fig. 5-16. This verifies that the LC directors near the substrates are stabilized; however, the lifetime of Hs is still shortened and the