In this paper, the optical performance degradation of a liquid crystal (LC) cell due to the instability of pre-tilt angle and polar anchoring strength of the alignment surface of liquid crystal devices is explored. Under accelerated thermal treatment, changes in both the pre-tilt angle and polar anchoring strength are observed. The impacts of these changes are modeled for both twist nematic (TN) and electrically controlled birefringence (ECB) cells. Through this modeling, we find that a stable surface is very important to the long term performance of liquid crystal devices for the telecommunication applications.
© 2003 Optical Society of America
Liquid crystal (LC) has been widely used in the field of information display because of its low power consumption, light weight, high resolution and high contrast ratio. In recent years, there has been intensive research and development effort to utilize liquid crystal in telecommunication applications such as optical switches and variable optical attenuators (VOA).1,2 The reason to use liquid crystal is simple: liquid crystal display is a mature technology. Many years’ research achievement, both in material and devices, can be easily transferred into the telecommunication area. It can take only weeks from designing to delivering a liquid crystal device for a new application. Liquid crystal device (LCD) has many other advantages: fast speed, small size and fabrication ease. It has become a major player in the optical component field.
The requirements on a liquid crystal device in telecommunication applications are different from those of displays. The most important ones are high optical performance in IR and long-term (>15 years) reliability. For example, in switching and blocking applications, a contrast ratio of higher than 10,000:1 is required while in display 500:1 contrast ratio can be considered very good. It has been demonstrated that liquid crystals are capable of achieving high contrast ratio by good design and optimizing manufacture processes. However, the question remains as to how to keep this performance level over a long period of time. There are numerous parameters that could change over time: thickness, alignment surface, liquid crystal properties and so on. This paper investigates how the pre-tilt angle and polar anchoring strength change under accelerated thermal treatment and how these changes will affect the optical performance of the devices.
2. Pre-tilt angle and anchoring strength experiment
Surface treatment and conditions are critical to the operation of any liquid crystal cell. For both TN and ECB cells, the alignment of the LC molecules at the substrate interface is homogeneous. As indicated in Fig. 1, the rod-like LC molecules in a homogeneous alignment lie approximately parallel to the substrate surface. This alignment is most commonly achieved by rubbing a substrate that has been coated with a polyimide. When the cell is filled with LC, the molecules align along the rubbing direction. As a result of the rubbing, the molecules lie at a slight angle with respect to the substrate surface. This small tilt, θp, is known as the pre-tilt angle. Another surface property critical to the operation of a LC is the anchoring strength. This parameter gives a measure of the binding strength between the alignment layer and the LC molecules. Since our interest is in TN and ECB LCDs where the applied field is basically perpendicular to the substrates, we will focus on polar (out of plane) anchoring strength. In most applications, a hard anchoring approximation is made. This assumes the anchoring strength is strong enough that the molecules at the surface remain fixed even under external force. As will be seen in following sections, this assumption fails when estimating the performance of LC telecommunication devices.
To investigate the changes in pre-tilt angle and polar anchoring strength, samples were made by using fused silica substrates coated with ITO as electrodes. A layer of polyimide was spin coated, cured, and mechanically rubbed to generate the preferred alignment. Two substrates were glued together in such a way that the rubbing directions on two substrates were anti-parallel. The cell thickness was controlled by 50 µm spacers and measured by interference method. A nematic liquid crystal was filled into the cells in a vacuum chamber at room temperature. The clearing temperature of the liquid crystal is higher than 115 C. After filling the cells were plugged and the LC was heated to isotropic phase for 10 minutes to eliminate defects formed during filling.
The pre-tilt angle and polar anchoring strength were measured after the samples were made. Then the samples were divided into three groups. Samples in group A were put in an oven with 95 C while samples in Group B were in an oven with 115 C. Samples in Group C were kept at room temperature all the time. After every seven days’ thermal treatment, the pre-tilt angle and polar anchoring strength were measured again. All the measurements were performed at room temperature. The pre-tilt angles were measured by using the magnetic null method3 and the polar anchoring strengths were measured by using the retardation vs. voltage (RV) technique.4
The pre-tilt angle results are shown in Fig. 2. Two conclusions can be drawn from the results. First, the pre-tilt angle increases in the first week and then stabilizes around 1.2°. Second, thermal treatment temperature has little influence on the pre-tilt angle change. Our experience with main chain polyimide is that pre-tilt angle decreases with time, rather than increase as seen in this experiment. The magnitude of the change is well above the measurement error (±0.1°). Since voltage was applied during the anchoring strength measurement, a separate experiment was conducted to investigate if this voltage played a role in the pre-tilt change. In this experiment, we made two samples and measured the pre-tilt angles. Then we applied a 50 V ac voltage to the samples for 16 hours and measured the pretilt angles again. The difference between the two measurements for each sample was within the measurement error. We concluded that voltage had little to no impact on the increase of the pre-tilt angle. The mechanism responsible for pre-tilt angle increase is still not understood.
The results of the polar anchoring strength measurements are shown in Fig. 3. Generally speaking, the second measurement gives higher values than the first and third measurements. When comparing the final values to the initial ones, five samples show increase and three samples show decrease in polar anchoring strength. Considering the relatively large measurement error (±20%), the polar anchoring strength doesn’t change too much. Another important feature is that different groups follow similar trend which means thermal treatment temperature has little impact on the change of the polar anchoring strength.
When measuring the samples for the second and third time, there was noticeable difference in how well the samples could be measured. In most cases, the data range that could be used to calculate the anchoring strength became narrower which made the measurement less accurate. In a few samples, measurements were possible after the thermal treatment when prior they weren’t. Observed changes in some samples were obvious under examination with a polarizing optical microscope. All these changes can complicate the interpretation of the result as will be discussed later in this paper.
3. Optical performance modeling
3.1 Modeling method
We have learned that there exist some changes in time in the pre-tilt angle and polar anchoring strength of the liquid crystal cells. In this study, we investigate the impact of these changes on the performance of liquid crystal devices. We first numerically calculate the liquid crystal configurations under different voltages and boundary conditions. Then the results are used to calculate the optical performance of the devices. The modeling on liquid crystal configuration can be described as following.
Consider a nematic liquid crystal confined between two alignment surfaces at z=0 and z=t, the free energy per unit area of the liquid crystal can be expressed as
in which fbulk is the bulk free energy density, and fs is the surface energy density. The bulk free energy density can be written as
in which n is the liquid crystal director and can be represented by polar angle θ and azimuthal angle φ, E is the electric field, and K11, K22, K33 are elastic constants for splay, twist and bend, respectively. To minimize the free energy and calculate the liquid crystal equilibrium configuration, we apply Euler equation,
The surface anchoring energy density is
in which θp is the pre-tilt angle and W is the polar anchoring strength. The boundary conditions at the surfaces are
Using a relaxation technique, the liquid crystal configuration can be numerically solved.5 The liquid crystal used in this modeling is 4-n-pentyl-4’-cyanobiphenyl (5CB) from EM industries. 5CB can be a key component in the liquid crystal mixture for telecommunication applications because it has high dielectric anisotropy and large birefringence. The elastic constants used in the modeling for splay, twist and bend are 6.4×10-12 N, 3.0×10-12 N and 10.0×10-12 N, respectively. The dielectric constants ε// and ε⊥ are 19.2 and 6.7, respectively. Considering that the operation wavelength is around 1.55 µm, the birefringence is estimated to be 0.16.
Extended Jones matrix method6 is used to model the optical performance of the liquid crystal devices. In the modeling, LC devices are placed between cross polarizers. The polarizers are assumed to be perfect: loss-less and infinitely large polarization efficiency. Two types of liquid crystal modes have been studied: TN mode and ECB mode.
Assuming the powers at input and output are Pin and Pout respectively, the transmission is expressed in decibel:
Contrast ratio at a certain voltage is defined as the difference between the maximum transmission and the transmission at that voltage.
3.2 TN LCD
In TN LCD, the rubbing directions on the two surfaces are perpendicular to each other. The liquid crystal molecules undergo a 90° rotation when moving from one surface to the other. In the optical modeling, the rubbing direction of the TN LCD at the first surface is arranged to be parallel to the polarization direction of the first polarizer. At off state (0 V), the polarization of the incident light follows the rotation of the LC molecules and the light goes through the second polarizer without loss. To achieve a 90° polarization rotation at 1550 nm wavelength, the thickness of the TN LCD needs to be 8.4 µm.
The modeling results on pre-tilt angle change are shown in Fig. 4. The anchoring strength is assumed to be strong in this modeling. For optical switching applications, only two states are needed: high transmission (0 dB) and low transmission (<-40 dB). When a high voltage is applied, low transmission is guaranteed by the combination of cross polarizers and the unique configuration of TN LCD. An increase in pre-tilt angle can only improve the contrast ratio because of less retardance of liquid crystal. For off state, the transmission decreases when the pre-tilt angle is higher but the amplitude of the change is small when the increase of the pretilt angle is less than 10°. We conclude that the performance of TN LCD in switching applications is not sensitive to the pre-tilt angle change.
When used as VOAs, the region from 0 dB to -20dB is the typical operational window. Although the slope seems to be consistent, the actual transmission at a fixed voltage is different for different pre-tilt angles. At 1.4 V, a pre-tilt angle change from 0.5° to 3° corresponds to about 1.1 dB decrease in transmission.
When modeling the effect of polar anchoring strength change, the pre-tilt angle is fixed at 1′. The results are shown in Fig. 5. Similar to the case of pre-tilt angle, TN LCD works well for switching applications even when the polar anchoring strength changes. For VOA applications, however, not only the transmission at fixed voltage but also the slope of the curve change for different anchoring strengths. When the polar anchoring strength changes from 1×10-4 J/m2 to 5×10-5 J/m2, the transmission decreases by 1.4 dB at 1.4 V.
3.3 ECB LCD
Same modeling has been conducted for ECB mode LCD. The cell construction is similar to the ones used for pre-tilt angle measurement in that the rub surfaces are aligned anti-parallel. In the modeling, the LCD is placed in such a way that the rubbing direction is 45° with respect to the polarization directions of the cross polarizers. ECB LCD is basically a tunable wave-plate. At off state, it is designed to be a half-wave plate so it can rotate the polarization of the light 90° and let it go through the second polarizer without loss. At on state, there is always a residual retardance left in the LCD and it is difficult to achieve high contrast ratio.
There are several approaches to achieve high contrast ratio in ECB LCD. The first one is, instead of operating between half wave and close to zero wave, doubling the thickness of the LCD so it will operate between one wave and half wave. The major problem with this approach is that the transmission vs. voltage curve is so sharp between the two states that it will be very difficult to electronically control the device. The second solution is to place the ECB between parallel polarizers and use the off state for high contrast ratio. The third solution is to place a retarder after the LCD with its slow axis perpendicular to the rubbing direction of the LCD. This retarder will compensate the residual retardance of the LCD and a high contrast ratio can be achieved.
The compensator approach is used in this study. The retardation on (δ=Δnd/λ) of the compensator used in the modeling is 0.067. Because of the use of a compensator, the retardation of the ECB LCD at off state is adjusted to 0.567 to achieve high transmission.
The results on the pre-tilt angle study are shown in Fig. 6. When the pre-tilt angle is small, high contrast ratio is achieved when the voltage is around 8 V. At this voltage, the residual retardance of the LC is matched with the compensative retardance. When the pre-tilt angle increases, the retardance of the LCD decreases. The voltage on LC has to be reduced in order to get the right retardance to match the compensative retarder. As a result, the voltage to achieve high contrast ratio decreases with the increase of pre-tilt angle, as shown in Fig. 6. However, there is little difference between the curves with 0.5° and 1° pre-tilt angles. Even when the pre-tilt angle increases from 0.5° to 3°, it is still possible to find a voltage range where better than 40 dB contrast ratio is kept all the time.
The results on polar anchoring strength study for ECB LCD are shown in Fig. 7. Unlike TN LCD, the high contrast ratio state of ECB LCD is very sensitive to the change of polar anchoring strength. For example, when the anchoring strength changes from 1×10-4 J/m2 to 5×10-5 J/m2, the voltage to achieve high contrast ratio reduces from 5.4 V to 4 V. A voltage of 5.4 V, used to be able to give <-50 dB transmission, now can only offer -27 dB. However, when the anchoring strength is stronger, the same percentage of change will give much less change in voltage. For example, when the anchoring strength changes from 1×10-3 J/m2 to 5×10-4 J/m2, any voltage between 6.9 V and 8.2 V can offer better than 40 dB contrast ratio
The magnetic null method on pre-tilt angle measurement is very sensitive and the measurement error is small. But the experiment of polar anchoring strength measurement is delicate and the calculation is subjective in determining the upper limit of the voltage. As a result, the error in the final result is relatively large. Additional error was introduced in this experiment when the sample was moved for each measurement. Although care was used to find the same spot, there was always a possibility that we were measuring a different region. When in-plane inhomogeneities4 were presented in the cells, such as a non-uniform alignment surface or a discontinuity of electrodes, the result would not be consistent if different regions were targeted. For example, if part of measured region had no electrodes or a defect showed up, there would be big impact on the accuracy of the measurement.
In both pre-tilt angle and polar anchoring strength measurements, there is no significant difference among three groups of samples that go through different thermal treatments. This is probably because all three treatment temperatures are well below the polyimide glass transition temperature (>30° C).
The stability of pre-tilt angle and polar anchoring strength is very important to the liquid crystal devices for the telecommunication applications because of the much more stringent requirements. An optical switch using ECB LCD with compensation is a good example. A small change in polar anchoring strength will cause significant degradation of contrast ratio and make the optical switch out of specification, as shown in Fig. 7. However, when the same device is used as a display, even if the anchoring strength changes from 1×10-2 J/m2 to 5×10-5 J/m2, the contrast ratio is always higher than 500:1 with 5.3 V applied voltage.
For switching application, TN LCD is a better mode because of its large tolerance on surface property change including pre-tilt angle and polar anchoring strength. A high contrast ratio can be easily achieved and kept for long time. For ECB LCD with compensation, the contrast ratio is very sensitive to the change of surface properties, especially polar anchoring strength. The surface has to be stable in order to make the switch reliable.
For VOA applications, surface properties are important to both TN and ECB LCDs. A change in pre-tilt angle or polar anchoring strength will cause the shift of attenuation level. However, there are differences between the two LCD modes. In the low attenuation region (0 dB,-20 dB), the transmission curve drops sharply for TN LCD but slowly for ECB LCD. As a result, the same change in surface properties will cause less degradation in ECB LCD.
The pre-tilt angle increases initially but stabilizes at about 1.2° after two weeks. This change is not large enough to cause any serious performance degradation. In fact, this problem can be solved by shelving the liquid crystal devices for a couple of weeks.
The polar anchoring strength also changes, although the amplitude of the change is not much. It is difficult to determine exactly how the anchoring strength behaves because of the limitation in the measurement technique. The stability of anchoring strength is very important to the liquid crystal devices for the telecommunication applications. A small change can cause serious degradation in the performance of the devices. When ECB is used in high contrast telecommunication switching applications, initial polar anchoring strength should be greater than ~ 5×10-4 J/m2 to ensure stable operation.
The authors would like to thank Liou Qiu and Dr. Ivan Smalyukh for measuring the samples, and Professor Oleg Lavrentovich for very helpful discussions.
References and links
1. J. Patel et al., “Liquid crystal and grating-based multiple-wavelength cross-connect switch,” IEEE Photon. Technol. Lett. 7, 514–516 (1995). [CrossRef]
2. Al R. Ranalli, Bradley A. Scott, and John P. Kondis, “Liquid crystal-based wavelength selectable cross-connect,” Proc. Of ECOC 1, 68–69 (1999).
3. T. J. Scheffer and J. Nehring, “Accureate determination of liquid crystal tilt bias angles,” J. Appl. Phys. 48, 1783 (1977). [CrossRef]
4. Yu. A. Nastishin, R. D. Polak, S. V. Shiyanovskii, V. H. Bodnar, and O. D. Lavrentovich, “Nemativc polar anchoring strength measured by electric field techniques,” J. Appl. Phys. 86, 4199 (1999). [CrossRef]
5. Rui-Qing Ma, “Polymer networks formed in liquid crystal and their applications,” Kent State University Dissertation, (2000).
6. P. Yeh, “Exrtended Jones Matric Method,” J. Opt. Soc. Am. 72, 507 (1982). [CrossRef]