Three kinds of LCD modes are currently in the mainstream of the LCD industry.
They are multi-domain vertically aligned (MVA) [9], in-plane switching (IPS) [10], and optically compensated bend (OCB) [11] mode LCDs, whose cell structures and driving concepts are shown in Fig. 1-8. Their distinguished feature of “wide viewing angle” is an important requirement for LCD-TV applications. Therefore, these three wide-viewing LC modes are selected to develop the further fast-switching requirement.
Though IPS and MVA modes have very outstanding wide-viewing characteristic, their response times are not fast enough for high-frame-rate display applications, such as field sequential color LCD or temporally multiplexing 3D display. Therefore, a novel technique, Pi-cell also known as OCB mode, was proposed to fulfill these requirements. The comparisons of different LC modes are listed in Tab. 1-1. Though OCB-LCD is regarded as the fastest switching LCD mode among the LCDs, some issues still need to be resolved. Among them, the most critical issue is the recovery from the bend state back to the splay or twist states.
As shown in Fig. 1-8 (c), the Pi-cell is normally operated in the bend state;
however, because of the topological difference between the ground splay state and the bend state, a nucleation transition has to be completed before operation. The transition
can be initiated by applying a critical voltage (Vcr), and then this voltage is held to sustain the device in the bend state. However, the Pi-cell still has the tendency of relaxing into the splay state. Therefore, we will propose two novel Pi-cell structures to resolve transition issue: one is to form a nanostructure by coating nano-particles before PI treatment to enhance the transition rate from splay to bend state; the other one is to use reactive monomers to modify the pretilt angle of alignment layer whose anchoring energy may be changed for the contrast ratio improvement.
Tab. 1-1 The characteristics comparisons of different LC modes.
(Data source: Chung Hwa Picture Tube, LTD)
◎
◎: Excellent ○: Good △: Acceptable X: Poor
(a) (b)
(c)
Fig. 1-8 The cell structures and driving schemes of (a) IPS, (b) MVA, and (c) Pi-cell.
1.5 Motivation and objective of this dissertation
With the growth of the information industry, the display technique requirements have become increasingly demanding. The desirable features are high contrast ratio, high power efficiency, high resolution and good visual quality. To fulfill these features, we designed and demonstrated the new technologies to achieve high contrast ratio and clear motion image quality. We did this by using conventional Pi-cell, also named OCB mode, in the OCB-LCD TV. To evaluate the motion quality of LCD TVs, we used the Moving Picture Response Time (MPRT). By applying the Dynamic Scanning Backlight with Black Insertion (DSBBI) technology previous proposed by our research team, the MPRT of a 32-inch OCB-LCD TV can achieve 5.9 ms in average gray-to-gray levels, which is comparable to CRT TV (MPRT~4ms) but better than other LCD modes [12]. By changing the structure of the LC cell, the contrast ratio of OCB-LCD TVs can be improved by 60% by adjusting the gamma curves with respect to R, G and B individually. With dynamic controlled backlight luminance, the contrast
V>Vcr
Transition Spot Splay state
Polarizer Retarder Glass/PI
LC
V= 0
Polarizer
Bend state(Driving Mode)
Vcr Von
Fast response
ratio of OCB-LCD TVs can even achieve greater than 1000:1 at 50% dimming setting [3].
To further realize the features of high power efficiency, adopting the FSC-LCD structure is a good candidate. The requisite components of FSC-LCD are the fast switching modes of liquid crystal and the backlight module. Among these requirements, the most urgent one is the fast response LC mode because the current research in the field is still short of realizing the FSC-LCD without the color break-up (CBU) phenomenon. The OCB-LCD modes possess the possibility to achieve the FSC-LCD.
In the demonstration of a 32-inch OCB-LCD TV, using conventional Pi-cell, we found two key issues that need to be resolved. First, applying the critical voltage bias for the LC state transition from splay state to bend state and keeping it in stable bend state are strict. To hold the Pi-cell in stable bend state, the applied voltage should be larger than the critical voltage of blue light. As the result, the transmittance and contrast ratio of the Pi-cell must be degraded. Therefore, it is necessary to develop a novel Pi-cell that is stable in bend state without bias voltage [13]. Second, in order to obtain good dark state and wide viewing angles in all viewing directions, we used commercial OCB-WV films to compensate for the viewing angle dependence of bend alignment LC cell [14]. However, the retardation mismatch occurs when the OCB-LCD shows the low gray-level images or patterns at large viewing angles and results in the red and blue light leakages corresponding to the right and left sides of rubbing direction respectively. Therefore, to design a high contrast ratio and transmittance of Pi-cells without using OCB-WV films is a challenge. In order to realize high image quality displays which fulfill the aforementioned features, it is very important to find the solutions to the two key issues mentioned above.
In this dissertation, I will describe the tasks completed on the aspects of the Pi-cell and the future works for performance improvements.
1.6 Organization of this dissertation
This dissertation is organized as follows: the principle of deformation mechanism of LC cells, the mechanical properties and transition scheme of a Pi-cell are presented in Chapter 2. The continuum theory, anchoring effect, homogeneous and heterogeneous nucleation theory are described. Additionally, this chapter also presents the operation principle and special characteristics of a Pi-cell. In Chapter 3, the standard TFT-LCD fabrication processes are introduced. And the instruments used in experiment for optical property measurements are described. The basic measuring concept and set-up of the measurement systems are also described in details. In Chapter 4, the novel nanostructure enhanced Pi-cell (NE-Pi-cell) is proposed. The fabrication methods of nanostructures are investigated and optimized; the limitation of nano-particle density is also analyzed. In addition, the dynamic observations and transition rate of a conventional Pi-cell and proposed NE-Pi-cell are presented and evaluated. In Chapter 5, the transition-free RMM-Pi-cell is proposed. The novel structure and the fabrications of a RMM-Pi-cell are presented. The performance comparisons between a conventional Pi-cell and a RMM-Pi-cell are investigated. The potential issue of the RMM-Pi-cell is also discussed here. Finally, the summary of the dissertation and future works are given in Chapter 6.
Chapter 2
Theory and Principle
The deformation mechanism of a LC cell and the principle of state transitions in a Pi-cell are introduced. The continuum theory is used to calculate the Gibbs’s energy of a certain state, and Rapini-Papoular approach is used to calculate the alignment effect on the orientation of LC director. In the principle of state transitions in a Pi-cell, the well-known nucleation theory method is illustrated. In order to study the rate of state transitions, we propose hetero-nucleation theory to explain why the heterogeneous surface energy status can shorten the transition time in a Pi-cell.
2.1 Deformation Mechanism of LC cells
In LCDs, an electric field is often applied to cause reorientation of the LC molecules. The switching performance of an LC device relies on its cell structure design. The Ossen-Frank continuum theory [15-18] is well-used by the LCD researchers. A basic continuum theory considers that the elastic constants of LC molecules determine the restoring torques arise when the system is perturbed from its equilibrium configuration. It is the balance between the electric torque and the elastic restoring torque that determines the LC’s static deformation pattern. Taking more factors into consideration, such as the electric field, the magnetic field, and the anchoring effect makes the equation more appropriate to the actual LCD cases. The detailed description of continuum theory is explained as follows.
2.1.1 Continuum Theory
The Ossen-Frank continuum theory is well-known for the calculation of the equilibrium configuration of LC directors in a cell. LCs exhibit curvature elasticity.
When an electric filed is applied to an LCD device, the LC molecules will be reoriented. The electric torque needs to be balanced by elastic restoring torque determines the LC’s deformation pattern. Any static deformation of LCs can be divided into a combination of three basic deformations: splay, twist, and bend, as illustrated in Fig. 2-1.
Fig. 2-1 Schematic drawing of (a)splay, (b)twist, (c)bend in LC cell [19].
We consider a nematic LC cell in the xy plane. The z axis is chosen to be perpendicular to the cell, so that the cell is located between z = 0 and z = d. Initially, the LC molecules are aligned parallel to the xy plane. An electric field is applied in the LC cell. Following the notation of Ossen-Frank continuum theory, the energy densities of a deformed LC and the electromagnetic field can be written as Eq. (2-1) [20].
[ ]
( )2 ) 1 (
) (
) 2 (
1 2
33 2 22
2
11 ∇⋅n + n⋅∇×n + n×∇×n − E⋅D
= k k k
F (2-1)
≠0
⋅
∇ n n⋅∇×n≠0 n×∇×n≠0
where k11, k22, and k33 are the splay, twist and bend elastic constants, respectively and n is the unit vector representing the director distribution in the cell. The elastic constants are strongly temperature dependent. E is the applied electric field and D is the displacement field vector.
Consider twist and tilt mode. The director n is uniformity twisted as a function of z, so the initial director distribution can be expressed,
) A uniform electric field is applied along the z axis. As a result, the directors are tilted toward the direction of the electric field. This leads to redistribution of the director n as a function of z. Therefore, the director can be written,
) where (θ, φ) are the positioned angles of the LC director, denote the polar coordinates rotated about y axis by θ and about z axis by φ. Substituting Eq. (2-3) to Eq. (2-1), the elastic energy density is obtained as:
[ ] [
2 33 2]
2 2The electromagnetic energy density term is
D E⋅
= 2 1
UEM (2-5) Actually, we are care about the change of the electromagnetic energy density due to the change of the dielectric constant ε as the LC director distribution changes.
Considering the first case: a constant voltage, the electric field E is a constant (E=V/d), regardless the magnitude of the dielectric constant. The net change of the
electromagnetic energy density due to a change of the dielectric constant can be
where A is the area of electrodes, d is the separation between the electrodes, and ΔQ is additional charges supplied by power supply. It can be denoted by
d
Substituting Eq. (2-7) into Eq. (2-6), we obtain
2
where ε0is the initial dielectric constant. A larger final dielectric ε leads to a lower energy density.
Considering another general case, the surface charge density is constant.
Therefore, the displacement field D is a constant. The change of electromagnetic energy density due to a change of dielectric constant can be written
0
We postulate the LC cell is uniform in xy plane, and the electric field E is applied along z axis. An application of Maxwell’s equation and the boundary condition on the surface lead to:
=0
= y
x E
E (2-10)
Eq. (2-9) can be written
(
⊥)
⊥ displacement field vector. The second term is a constant independent of the director orientation θ(z). And the component DZ can be writtenE
DZ =(ε⊥cos2θ +ε||sin2θ) (2-12)
The total free energy in the cell is given by
( )
The preferred LC director distribution functions θ(z) and φ(z) can be acquired by minimizing the total free energy using variation method. i.e. Let δU=0 and give suitable boundary conditions, then the Umin. can be obtained at the same time.
2.1.2 Anchoring Effect
The LC alignment affects the equilibrium state, switching property, and director configuration of an LC cell. This alignment is generally achieved by the anchoring effect with a pretreated layer. Most preferably, polyimide is used owing to its low cost, high stability, and easy process. Two major methods are used to pre-treat the alignment material: mechanical rubbing and photo-alignment. The mechanical rubbing method is executed by a roller covered with woolen texture. By brushing in the same direction, the alignment layer is strained, and this strain aligns the LC
director in a certain direction. The other method of pre-treating the alignment material is to use polarized UV light to orient the molecules of alignment layer in a certain direction. This method can be used along with a photo-mask to make multi-directional alignments in one substrate, but the uniformity is still an issue.
This stabilization mechanism can be expressed by Rapini-Papoular approach [21-23] as follows: resulting from the anchoring effect, Fθ denotes the free energy component in terms of the polar angle, Fφ denotes the free energy component in terms of the azimuthal angle.
Wθ and Wφ are the constants denote the interactions between the substrates and the LC directors. θ0 and φ0are the equilibrium angles with respect to the polar and azimuthal dimensions.
Considering the anchoring effect, the continuum equation mentioned in 2.1.1 needs to be modified as follows. The orientation distribution of stabilized LC directors of a LC cell can be illustrated by the equation.
2.1.3 The Mechanical Properties of Pi-cells
In a Pi-cell, the LC directors have a bend distribution with a total bend of 180°
(π) from surface to surface when the Pi-cell is in operation. The alignment arrangements of upper and lower substrates in a Pi-cell are parallel. Generally, the initial state of a Pi-cell prefers to be aligned in splay state. By calculating the Gibbs’s free energy in Eq. (2-17), the stable state corresponding to the applied voltage can be obtained. As shown in Fig. 2-2, the cell is more stable in splay state than in bend state without applying a voltage. While applying a voltage larger than critical voltage, the bend state becomes more stable than the splay state [24-26].
Source: T. Uchida et al., SID’00 Digest 31, 26 (2000).
Fig. 2-2 Gibbs’s free energy of bend and splay states as a function of applied voltage [24].
Take the pretilt angle into account; the Gibbs’s free energy diagram can be used to determine the pretilt angle of a Pi-cell which is in the bend state initially. As shown in Fig. 2-3 (a), if the pretilt angle is much lower than critical pretilt angle, the Gibbs’s free energy of the bend state is too large to be stabilized. In other words, if the pretilt angle is higher than critical pretilt angle, the bend state can be stabilized without applying a voltage. In heterogeneous case, the bend state can be initiated with nucleation by applying a voltage, as shown in Fig. 2-3 (b) [11][24][27-30].
Source: T. Uchida et al., SID’00 Digest 31, 26 (2000).
Fig. 2-3 Transition in Pi-cell, (a) is pretilt angle effect and (b) is the nucleation in a transition process [24][27].
Furthermore, the Pi-cell has the rapid response time because the cell is operated between the bend state at the critical voltage (Vcr) and the near homeotropic state at high voltage. The symmetric director configurations in the lower and upper parts of the LC layer, regarded as two halves of a Fréedericksz cell, imply the equivalent cell gap is one half of the real LC layer. Because the response time of a LC cell is proportional to the square of cell gap, the response time of a Pi-cell has improved by a factor of four.. The schematic diagram is shown in Fig. 2-4. Besides, the flow effect of a Pi-cell leads to the response time improvement is more than four times. The LC flow in Pi-cell is toward the same direction; therefore, the backflow effect can be ignored. The comparison of the flow effect in a Pi-cell and the anti-parallel backflow, which slows down the response of LCs, in a TN-cell is shown in Fig.2-5.
In addition to fast response, the optically self-compensated property is also a key design factor in a Pi-cell. Due to the symmetric director configurations of a Pi-cell, the optical retardations are almost the same, no matter people see from left or right sides, as shown in Fig. 2-6. Therefore, Pi-cell is a kind of wide viewing angle technologies.
Critical pretilt angle
Fig. 2-4 Schematic diagram of two equivalent halves of a Fréedericksz cell.
Fig. 2-5 Flow effect in Pi-cell and TN-cell.
Fig. 2-6 The illustration of optically self-compensated property in a Pi-cell.
Pi-cell TN-cell
Long axis Short axis
2.2 State Transition Scheme of a Pi-cell
Pi-cell [31] possesses fast response time, due to lacking of backflow during molecular relaxation [32-36] in its bend state, for liquid crystal display (LCD) application. Generally, the Pi-cell is operated in the bend state; a nucleation transition has to be completed to operate the Pi-cell in the bend state. The Pi-cell, however, has five distinct states in its transitions, illustrated in Fig.2-7. There are (1) the splay ground state, (2) the symmetric splay state (Hs), (3) the asymmetric splay state (Ha), (4) the bend state, and (5) the 180° twist state [37-39]. The lowest free energy state is splay state without an applied voltage. When applied a voltage higher than the first critical voltage VC1, Hs state would be formed under the same surface condition on both sides of a Pi-cell. The Hs state, then, quickly turned into asymmetric Ha state. As the applied voltage was further increased beyond the second critical voltage VC2, also called the critical voltage, denoted Vcr, of a Pi-cell, the bend state became the stable state. The voltage had to be kept larger than the critical voltage to sustain the cell in the bend state; otherwise, the LC molecules would be relaxed into twist or splay states in a Pi-cell. In TFT-LCD applications, due to the limitation of the maximum driving voltage ( ~ 6V) and the requirement of sustained voltage to prevent recovery of the splay state, the Pi-cell is generally driven in the range between 2 ~ 6V [40]. In the transition scheme, there are two nucleation processes included; one is Hs (or Ha)-to-bend state, the other one is bend state relaxed into twist state.
Fig.2-7 The commonly known states form in Pi-cell devices [41].
2.2.1 Nucleation Theory
Classical nucleation is the extremely localized budding of a distinct thermodynamic phase. Some examples of phases that may form via nucleation in liquids are gaseous bubbles, crystals, or glassy regions. Nucleation generally occurs with much more difficulty in the interior of a uniform substance, by a process called homogeneous nucleation. The creation of a nucleus implies the formation of an interface at the boundaries of a new phase. As mentioned above, the splay-to-bend transition in a Pi-cell has to be completed via so-called nucleation process. In the process, an interface forms to be the boundary between splay and bend states. When the extra voltage, higher than the critical voltage, is applied to the Pi-cell, the extra driving power will force for nucleation and bring about a change in free energy per unit volume, Gv, between splay and bend states. This change in free energy is balanced by the energy gain of creating a new volume, and the energy cost due to
Splayed
creation of a new interface. When the overall change in free energy, ΔG is negative, nucleation is favored.
Some energy is consumed to form an interface, based on the surface energy of each phase. If a hypothetical nucleus is too small (known as an unstable nucleus or
"embryo"), the energy that would be released by forming its volume is not enough to create its surface, and nucleation does not proceed. The critical nucleus size can be
"embryo"), the energy that would be released by forming its volume is not enough to create its surface, and nucleation does not proceed. The critical nucleus size can be