2

JOURNAL OF DISPLAY TECHNOLOGY, VOL. 3, NO. 1, MARCH 2007

Novel Color-Sequential Transective Liquid Crystal Displays

Ju-Hyun Lee, Xinyu Zhu, and Shin-Tson Wu, Fellow, IEEE
AbstractA novel transective liquid crystal display architecture and its system driving schemes are proposed. In the reective mode, the ambient light is used to readout the displayed images. While in the transmissive mode, a color-sequential light emitting diode backlight is used to eliminate the color lters. Under such device conguration, several advantages such as increased brightness and maximized color saturation for both transmissive and reective modes can be achieved. Index TermsColor sequential imaging, liquid crystal display (LCD), reective mode, transective mode, transmissive mode.

I. INTRODUCTION RANSFLECTIVE liquid crystal display (LCD) has been used widely for portable electronic applications because of its large dynamic range for ambient light. A transective LCD can display images in transmissive mode and reective mode, independently or simultaneously. Therefore, such a transective LCD can be used under any ambient lighting circumstances. Because transective LCDs can display images under a bright environment by reective mode using ambient light, it reduces power consumption which is the key issue for mobile and portable applications. Also, transective LCD provides good readability in low light ambient using its own backlight. To realize the dual functions of reective and transmissive displays, each color sub-pixel of a transective LCD is divided into two regions: transmissive (T) and reective (R) regions. Only the reective region has a bumpy metal mirror for reecting the incident light. The reector partially implemented on the reective region in the sub-pixel is called transector [1] or an opening-on-metal transector [2]. The transector provides access to optimize the ratio of image brightness between transmissive and reective modes by simply adjusting the areas of the reective and transmissive regions. Therefore, we can develop an application oriented and optimized transective LCD. However, this advantage could become disadvantage because the light efciency of each transmissive and reective mode is reduced. To realize the bi-functional property of the system several technical challenges have to be overcome without raising manufacturing cost. Among the device performance factors, higher optical efciency and better color reproduction for both T- and

Manuscript received September 16, 2006; revised November 7, 2006. This work was supported by Toppoly Optoelectronics Corporation (Taiwan), R.O.C.. The authors are with the College of Optics and Photonics, University of Central Florida, Orlando, FL 32816 USA (e-mail: [email protected]). Digital Object Identier 10.1109/JDT.2006.890700

R-modes are most critical. Meanwhile, matching the electro-optical property between the T and R modes is another important issue for transective LCDs. The unequal electro-optical response, that is, the discrepancy between the voltage dependent transmittance and reectance, originates from the phase retardation difference between the T and R modes. If the LC layer thicknesses on the T and R modes are equal, then the optical path length of the R-mode would be two times longer than that of the T-mode. As a result, the transmittance and reectance of the LCD will be different. To solve this problem, several techniques have been developed. For example, the dual cell gap structure [3] employs different cell gaps for the T and R modes to compensate the phase difference. This is by far the most popular device structure for commercial transective LCDs. However, fabrication of a dual cell gap structure is complicated. To avoid the fabrication difculty, single cell gap is preferred. One approach is to use different LC modes (dual LC mode) for the T and R modes [4], [5]. However, realizing the dual LC mode in a device is more difcult than fabricating the dual cell gap. Another approach is to use two TFTs (thin lm transistors) for each sub-pixel to independently control the transmissive and reective modes. In this case, we can match the electro-optical responses of the transmissive and reective modes very well by adjusting the driving voltages of each TFT. We call this the dual gamma driving method [6]. However, this method needs two times more data drivers than the conventional system and this method increases the system cost. Unequal color reproduction is another technical issue of the transective LCD which employs a single thickness color lter on the sub-pixel area. In the R region, the incident light passes through the color lter twice, while it passes only once in the transmissive region. Therefore, if we optimize the pigment density or color lter thickness for T-mode, then the color lter on the reective region absorbs more light leading to a darker image than the transmissive one. To solve this problem, different thickness of color lters for reective and transmissive modes has been suggested [7], [8]. This method also increases the fabrication complexity. There exists an interesting imaging technology which improves the optical efciency and color purity remarkably for the direct-view type transmissive LCD. The technology is called color sequential imaging method [9][11]. Color sequential imaging technology is already being widely used in projection displays especially for the single panel DLP (digital light processing) projection TVs [12], [13]. However, color break-up [14], [15] occurs when the response time of the imager is not fast enough. Color break-up is the tri-chromatic separation of the viewers perception when our view-point moving speed is

1551-319X/$25.00 2007 IEEE

LEE et al.: NOVEL COLOR-SEQUENTIAL TRANSFLECTIVE LCDs

Fig. 1. Schematic of the new transective LCD.

faster than the color sub-frame switching speed for following the moving object in the large-sized motion pictures. Basically color sequential imaging requires at least three times faster sub-frame refreshing rate than conventional standard video frame frequency to draw three primary color sub-frames sequentially. To solve the color break-up problem, one approach is to use a very fast sub-frame refreshing rate. Nowadays, many commercial DLP projection engines use very high sub-frame frequency rate, higher than 500 Hz, which is almost ten times faster than the standard video frame rate. However, some sensitive observers are still able to detect the color break. Since most of LCDs have a relatively slow response time, this approach is not amiable for large screen LCDs for HDTV applications. Another approach to solve the color break-up problem is to x our view-point. We will meet this situation when we see a small size image. Therefore, the color sequential imaging is more suitable for small-sized LCDs. Together with the highly saturated color switchable light source, such as an LED (light emitting diode) backlight, the color sequential transmissive LCD has already been demonstrated to have brighter and better color images than other conventional small-sized transmissive LCDs for mobile devices [16]. In this paper, we suggest a new transective LCD system architecture along with its operating methods using a color sequential imaging technique for small-sized LCDs for mobile/ portable applications. II. DEVICE DESIGN Fig. 1 shows the schematic of our device conguration. Here, the dashed box represents one pixel unit. A unique feature of our new LCD, in comparison with the conventional transective LCDs, is its pixel structure. Each pixel has four instead of three sub-pixels in other LCDs. Three of the sub-pixels with

bumpy metal reectors are for reective modes and three different primary color lters for each sub-pixel, respectively. One sub-pixel is for transmissive mode without a color lter. Therefore, the color sequential imaging method is required to produce color images in the transmissive mode. For the color sequential imaging, our new device needs the color switchable backlight such as an LED backlight which is widely used in small sized LCDs for mobile phones. Each sub-pixel in our new device is driven by a single TFT which is independent from the TFTs of other sub-pixels in a pixel unit. All pixel electronic devices such as TFTs and storage capacitors, including those for transmissive sub-pixels, are located under the metal reectors in the reective region as shown in the gure as dotted boxes. Putting electronics under the metal electrodes is the same way as the conventional transective LCDs to maximize the aperture ratio of the pixel. Reective sub-pixels on a row are addressed by a scan line connected to a gate driver for the reective mode, but transmissive sub-pixels are excited by another scan line connected to another gate driver for the transmissive mode. Just as the gate drivers, the data drivers for the transmissive and reective modes are independent of each other. To synchronize the reective image with the transmissive one, a timing control electronic circuit is required which also controls the timing of LED backlights color switching. III. DISCUSSION Because of the aforementioned structural characteristics, our new transective LCD has following advantages: 1) The optical efciency of both transmissive and reective images is enhanced greatly. One pixel of the conventional transective LCD is divided into six regions because each sub-pixel of the three primary color sub-pixels is divided into transmissive and reective regions. However, in our

JOURNAL OF DISPLAY TECHNOLOGY, VOL. 3, NO. 1, MARCH 2007

TABLE I IMPACTS OF INCREASED APERTURE SIZE OF THE NEW LCD SYSTEM

TABLE II SYSTEM PARAMETER COMBINATION TABLE TO DESIGN THE SYSTEM DRIVING SCHEMES

( : one frame scanning time, Ns: # of scan line in conventional TFT LCDs, Nd: # of data line in conventional TFT LCDs)

new device, one pixel is divided by only four sub-pixels because one transmissive sub-pixel shares the area for the three primary colors. Therefore, the aperture size of both transmissive and reective regions increases, as shown in Table I, which also greatly increases the light efciency. In addition, because the transmissive sub-pixel does not have a color lter which absorbs light and decreases optical efciency of the transmissive mode, the light efciency of the transmissive mode increases more noticeably. 2) We can easily optimize and maximize the color purity of both transmissive and reective images without using the complicated color lter fabrication method, such as making different color lter thicknesses for the transmissive and reective regions. We only need to optimize the color lter thickness for the reective mode. 3) To obtain a single electro-optical response, we only need to control the driving voltages of the transmissive and reective modes separately by data processing. Unlike the commercially employed dual-cell gap method, the fabrication process of this new approach is similar to that of single cell gap device. Thus, the fabrication yield should be higher and the manufacturing cost should be lower.

Some drawbacks are remaining. One drawback is the requirement of having 33% more data drivers which raises the cost of the nal product. However, it is smaller than the case of the double TFT approach for conventional transective LCDs which needs two times more data drivers. In addition, if we adopt the system-on-glass technology by using the low temperature poly-silicon process, no more cost-up issue remains. Another disadvantage is that our new system requires complex driving electronics. But it might be acceptable to todays well established ASIC design technologies. For the synchronization of transmissive images with reective images together with LED backlight color switching, several different driving schemes can be considered. In this paper, we use the system parameter combination table as shown in Table II. As a result, we suggest ve different driving schemes which are appropriate to LCDs operation principle, line-at-atime scanning. Fig. 2 shows an example of the operation timing diagram of the system. The top graph is the timing diagram for reective mode and the bottom graph shows the timing diagram for transmissive mode. Along the timing lines, which are shown as slanted lines of reective mode, reective sub-pixels are

LEE et al.: NOVEL COLOR-SEQUENTIAL TRANSFLECTIVE LCDs

Fig. 3. System driving scheme using impulsive type backlight driving method. Fig. 2. System driving scheme using dark sub-frames in transmissive mode.

scanned by scan drivers and one frame of reective image is drawn line by line. During one frame period, the transmissive sub-pixels are scanned six times. The rst sub-frame scanning of the transmissive mode is synchronized with frame scanning of the reective mode. To draw each primary color sub-frame image, the backlight switches its primary color at the time point when the scanning for primary color sub-framesthe rst, third, and fth sub-framesstart. As a result, the backlight is always turned on and only the illumination color is switched. Even numbered sub-frames among six sub-frames of the transmissive mode show black image on the whole LCD area by sending black image data to transmissive sub-pixels during the even numbered sub-frame scanning. This technique is a well known method for color sequential transmissive LCDs [17], [18] and is used to avoid the interference between two sequential color sub-frames. For example, if we do not use the black sub-frame, the latter half period of the last row imaging for a primary color overlaps with the rst half period of the rst row imaging for the next primary color. This is because the operation principle of an LCD is based on hold-type imaging and line-at-a-time scanning. Introducing the black sub-frame solves this problem easily. However, because of the black sub-frame, we lose half of the total light energy. To avoid this energy loss, a color scrolling light illumination technology can be considered which is used in some of the prototyped projection display systems. However, realizing a color scrolling backlight for direct view LCDs is challenging. Another method is to use an impulsive type backlight. Fig. 3 shows the relationship between row scanning of the transmissive mode and the backlight ashing timing. This concept is also used in some recent high-end transmissive LCDs for improving motion picture quality. The backlight is activated between the time point when the last row of the transmissive subpixels are scanned at a sub-frame scanning and the time point when the rst row of the transmissive sub-pixels are scanned at the next sub-frame scanning. This method allows slower scanning speed and LC response time than those of the aforementioned methods. One disadvantage is the requirement of driving the LED with a higher peak power in order to keep the same brightness for the transmissive image. In the previous case, the

Fig. 4. Pixel electronic circuits: (a) conventional circuit for line-at-a-time scanning imaging method, and (b) frame buffer pixel circuit for frame by frame imaging method.

LC response time should be at least six times faster than conventional color sub-pixel based LCDs. However, in the case of the impulsive backlight based system, shorter backlight activation duration permits a slower LC switching speed. In the previous two methods, the employed pixel electronics is a single TFT-based electronics for both T and R-mode subpixels which is common pixel electronics for most of LCDs using line-at-a-time scanning driving scheme. However, there is another pixel electronics which helps to tolerate a slower LC response and higher power efciency of backlighting which is called a frame buffer pixel circuit using more than two TFTs per sub-pixel [19]. Fig. 4 shows the difference between a conventional single TFT based pixel electronics and a frame buffer pixel circuit. In Fig. 4(a), one TFT is connected in parallel with a storage and an electrode which applies the voltage to capacitor the liquid crystal layer (LC). To draw a line of image on a row of pixels in an LCD panel, the data driver sends the pixel data

JOURNAL OF DISPLAY TECHNOLOGY, VOL. 3, NO. 1, MARCH 2007

Fig. 5. Schematic of the system which uses the frame buffer pixel circuit only for the transmissive sub-pixels.

through data lines (D) of the sub-pixels on the same row and the gate driver triggers the TFTs by sending a signal through a gate line (G). Then data are transferred to the storage capacitors through TFTs and the stored data in the storage capacitors provide a voltage to the liquid crystal layer during one frame period. In the gure, the dotted box means a pixel electronic circuit and the right side square block is an equivalent circuit for the left side pixel electronic circuit. Fig. 4(b) shows the basic concept of LCD driving based on frame buffer pixel circuit which has two TFTs in one sub-pixel. Each of these two TFTs has its own function: one function as a memory part to store the image data and the other as an imaging part to control the director orientation of the liquid crystal layer . In the by using the data stored in the memory capacitor gure, the data line (D) of the rst TFT (TFT1) is connected to the data driver. The gate line (G) of the rst TFT is connected to the gate driver. The drain of the rst TFT is connected to the source of the second TFT (TFT2) which is also connected to the rst storage capacitor. When the scan driver scans from the rst row to the last row, image data are transferred to the rst storage capacitor through the rst TFT. The stored data in the rst capacitor are not transferred to the liquid crystal layer immediately because the second TFT is not activated yet during the scanning time. Therefore, the combination of the rst capacitor and the rst TFT functions as a memory buffer. After scanning all rows, that is, after nishing writing one frame image data into the frame buffer, all second TFTs are activated simultaneously by triggering the gate lines VS. Consequently, the stored image are transferred to the second cadata in the rst capacitor to control the liquid crystal layer. Because each pacitor sub-pixel has two gate input lines G and VS, two gate lines in each sub-pixel should be connected to the gate driver. This frame buffer pixel circuit based device architecture is used in some of liquid-crystal-on-silicon devices to realize single panel based optical engine for system cost-down. Because of several

TFTs and storage capacitors in a sub-pixel, the frame buffer architecture reduces the aperture ratio if it is used for the transmissive LCDs. However, if it is used for the reective LCDs or transective LCDs, all pixel electronics can be hidden under the opaque metal reector and no loss of aperture ratio arises. By adopting the frame buffer architecture, we could design three more driving schemes which will benet lowering the power consumption and relaxing the requirement of fast LC switching. Fig. 5 shows the electronics schematic of the system which uses the frame buffer electronic circuit only for the transmissive sub-pixels. Because of the frame buffer concept, one more gate line from transmissive sub-pixels is connected to the gate driver for the transmissive mode and it increases the complexity of gate driver design. However, the effective aperture ratio of the display panel can be kept almost the same as that of the system, which does not use the frame buffer architecture, by adjusting the physical sub-pixel positions to locate all the gate lines and pixel electronics of the transmissive sub-pixel under the opaque metal reectors of the reective sub-pixels. In this system, we can design driving schemes with slower liquid crystal mode in comparison with aforementioned driving schemes based only on the line-at-a-time scanning technique in Figs. 2 and 3. In the case of the driving scheme using six sub-frames with dark sub-frame illustrated in Fig. 2, the response speed of liquid crystal should be at least six times faster than conventional LCDs. By using an impulsive type backlight in Fig. 3, we can relieve this strict requirement of liquid crystal switching speed. However, we still need to drive the liquid crystal faster than three times of the conventional LCDs because we have to secure the time for the duration of backlight illumination. Fig. 6 shows the timing diagram of the device operation based on the frame buffer architecture. During one frame period, transmissive sub-pixels are scanned three times for three primary color sub-frames. The rst scanning for the rst primary color

LEE et al.: NOVEL COLOR-SEQUENTIAL TRANSFLECTIVE LCDs

Fig. 6. System driving scheme using the frame buffer architecture only for the transmissive mode.

Fig. 8. System driving scheme using the frame buffer architecture for both transmissive and reective modes.

Fig. 7. System driving scheme using the frame buffer architecture only for the transmissive mode with dark sub-frames in the transmissive mode.

sub-frame is synchronized with the scanning of reective mode for a frame. However, during the rst scanning for the transmissive mode, image data are transferred into the frame buffer memory and they do not draw the image for the rst primary color. After nishing the rst scanning, image data stored in the frame buffer memory are transferred to the second capacitor in Fig. 4(b)] through the second TFTs [TFT2 in Fig. 4(b)] [ by triggering the entire second gate lines [G in Fig. 4(b)] of the display panel at a time as shown in Fig. 6 as the leftmost vertical dotted line. As a result, the rst primary color image on the whole area of the panel will be drawn at a time. In the same manner, the second and the third primary color images will be drawn. For this operation concept, the backlight only switches its color the same as the case in Fig. 2. However, different from the case in Fig. 2, there is no light loss here. One concern in this method is a mismatch of imaging timing between transmissive and reective modes because of the different imaging mechanisms: the reective mode draws the images line by line while the transmissive mode draws the images frame by frame. As a result, the last primary color sub-frame image of the transmissive mode is not matched to the reective image. The psychological inconvenience of watching images based on this driving scheme should be studied further.

To solve the mismatch image problem, we suggest a modied driving scheme as shown in Fig. 7. In this case, one frame period is divided by four sub-frames for the transmissive mode, instead of three in the previous method. Among four sub-frames, the last one is a dark sub-frame which can be realized by either turned-off backlight or dark image data signal. Using the dark sub-frame, all primary color sub-frames are well drawn during one frame of a reective image is being shown. However, this method requires faster liquid crystal response than previous technique. Moreover, we lose about 25% light in comparison with the previous method. The last method uses the frame buffer architecture for both reective and transmissive modes. The reective image is also drawn frame by frame as Fig. 8 shows. The time point of starting the rst primary color image of the transmissive mode is synchronized with that of the frame image of the reective mode. By using this method, the reective images and the transmissive images are synchronized perfectly without brightness loss. In addition, the required LC response speed is only three times faster than a conventional LC. IV. CONCLUSION We have developed a new transective LCD system using color sequential imaging for the transmissive mode. Considering several design parameters such as: 1) usage of dark sub-frame; 2) device operation mechanismsline-at-a-time scanning and frame buffer architecture; and 3) the backlight driving method, we have suggested ve different system driving schemes. Because our new system has several advantages in comparison with conventional transective LCDs, we believe our new system will be promising for small-sized mobile display applications and hope that our design will be reduced to practice by LCD manufacturers. REFERENCES
[1] R. D. Ketchpel and S. Barbara, Transector, U.S. Patent 4 040 727, Aug. 9, 1977. [2] M. Shimizu, Y. Itoh, and M. Kubo, Liquid crystal display device, U.S. Patent 6 341 002 B1, Jan. 22, 2002.

JOURNAL OF DISPLAY TECHNOLOGY, VOL. 3, NO. 1, MARCH 2007

[3] X. Zhu, Z. Ge, T. X. Wu, and S. T. Wu, Transective liquid crystal displays, J. Display Technol., vol. 1, pp. 1529, Sep. 2005. [4] S. H. Lee, K. H. Park, J. S. Gwag, T. H. Yoon, and J. C. Kim, A multimode-type transective liquid crystal display using the hybridaligned nematic and parallel-rubbed vertically aligned modes, Jpn. J. Appl. Phys. Part 1, vol. 42, pp. 51275132, Aug. 2003. [5] C. J. Yu, D. W. Kim, and S. D. Lee, Multimode transective liquid crystal display with a single cell gap using a self-masking process of photoalignment, Appl. Phys. Lett., vol. 85, pp. 51465148, Nov. 2004. [6] K. H. Liu, C. Y. Cheng, Y. R. Shen, C. M. Lai, C. R. Sheu, Y. Y. Fan, C. C. Chen, and I. J. Lin, A novel double gamma driving transective TFT LCD, in Proc. Int. Display Manuf. Conf., Feb. 2003, pp. 215218. [7] K. J. Kim, J. S. Lim, T. Y. Jung, C. Nam, and B. C. Ahn, A new transective TFT-LCD with dual color lter, in Proc. Intl Display Workshops, Dec. 2002, pp. 433436. [8] K. Fujimori, Y. Narutaki, and N. Kimura, High-transmissive advanced TFT LCD technology, Sharp Tech. J., no. 4, Apr. 2003, Special Papers 1. [9] M. G. Clark and I. A. Shanks, A eld-sequential color CRT using a liquid crystal color switch, in SID Symp. Tech. Dig., May 1982, pp. 172173. [10] R. Vatne, P. A. Johnson, Jr., and P. J. Bos, A LC/CRT eld-sequential color display, in SID Symp. Tech. Dig., May 1983, pp. 2829. [11] H. Hasebe and S. Kobayashi, A full-color eld-sequential LCD using modulated backlight, in SID Symp. Tech. Dig., May 1985, pp. 8183. [12] J. Florence and L. A. Yoder, Display system architectures for Digital Micromirror Device (DMD) based projectors, in Proc. SPIE, Mar. 1996, vol. 2650, pp. 193208. [13] D. Armitage, I. Underwood, and S. T. Wu, Introduction to Microdisplays. West Sussex, U.K.: Wiley, 2006. [14] L. Arend, J. Lubin, J. Gille, and J. Larimer, Color breakup in sequentially scanned LCDs, in SID Symp. Tech. Dig., June 1994, pp. 201204. [15] M. Mori, T. Hatada, K. Ishikawa, T. Saishouji, O. Wada, J. Nakamura, and N. Terashima, Mechanism of color breakup on eld-sequential color projectors, in SID Symp. Tech. Dig., May 1999, pp. 350353. [16] See, Samsung SDIs UFS LCD Principle [Online]. Available: http:// www.samsungsdi.com/contents/en/product/lcd/type02.html [17] J. H. Woo, J. J. Yoo, G. H. Kim, S. C. Lee, and C. R. Seo, Optimization of LED backlight ashing time for eld sequential color LCDs, in Proc. IDW, Dec. 2003, pp. 16811682. [18] J. Hirakata, A. Shingai, Y. Tanaka, and K. Ono, Super-TFT-LCD for moving picture images with the blink backlight system, in SID Symp. Tech. Dig., June 2001, pp. 990993. [19] S. Lee, J. C. Morizio, and K. M. Johnson, Novel frame buffer pixel circuits for liquid-crystal-on-silicon microdisplays, IEEE J. Solid-State Circuit, vol. 39, pp. 132139, Jan. 2004.

Ju-Hyun Lee received the B.S., M.S., and Ph.D. degrees in physics from Sogang University, Seoul, Korea, in 1990, 1992, and 1997, respectively. His Ph.D. dissertation was on antiferroelectric liquid crystals and its device applications. After his Ph.D., he worked in industry to commercialize a ferroelectric liquid crystal display and liquid crystal telecommunication devices. In 2002, he joined Samsung Advanced Institute of Technology and researched on display system architectures and optical system design for display applications. In 2005, he transferred to the College of Optics and Photonics/CREOL, University of Central Florida, as a research scientist. His current research activities include the novel LCD architecture and the LC optoelectronic/photonic device design.

Xinyu Zhu received the B.S. degree from Jilin University, China, in 1996, and the Ph.D degree from Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, China, in 2001. His research work in Ph.D involved mainly the reective liquid crystal display with single polarizer. Following the Ph.D. degree, he joined the School of Optics, CREOL, University of Central Florida, Orlando, as a research scientist in 2001. His current research interests include reective and transective liquid crystal displays, liquid crystal on silicon (LCoS) projection display, wide viewing angle liquid crystal displays and adaptive optics application with nematic liquid crystals.

Shin-Tson Wu (M98SM99F04) received the B.S. degree in physics from National Taiwan University, and the Ph.D. degree from the University of Southern California, Los Angeles. He is a PREP professor at College of Optics and Photonics, University of Central Florida (UCF). His studies at UCF concentrate in foveated imaging, bio-photonics, optical communications, liquid crystal displays, and liquid crystal materials. Prior to joining UCF in 2001, he worked at Hughes Research Laboratories, Malibu, CA, for 18 years. He has co-authored 4 books: Fundamentals of Liquid Crystal Devices (Wiley, 2006, with D. K. Yang); Introduction to Microdisplays (Wiley, 2006, with D. Armitage and I. Underwood) Reective Liquid Crystal Displays (Wiley, 2001, with D. K. Yang) and Optics and Nonlinear Optics of Liquid Crystals (World Scientic, 1993, with L. C. Khoo), 5 book chapters, and over 350 journal papers. He has more than 55 issued and pending patents. Several of his patents have been implemented in display and photonic devices. Dr. Wu is a Fellow of the Society of Information Display (SID) and Optical Society of America (OSA).