History Crystallized: A First-Person Account of the Development of Matrix-Addressed LCDs for television at RCA in the 1960s
The path to ubiquity for LCD TVs began 45 years ago at the RCA Laboratories in Princeton, New Jersey. RCA veteran Bernard J. Lechner was there, and he takes a look back at the earliest days of LCD research and development.
by Bernard J. Lechner
ON May 28, 1968, RCA held a press conference in New York City to present the results of a major project to make liquid-crystal displays (LCDs), including television displays. The work had been conducted in secret for more than 3 years at the RCA Laboratories, David Sarnoff Research Centerin Princeton, New Jersey. At that press conference, I gave the first public demon-stration of an LCD, albeit only 36 pixels, reproducing a moving gray-scale image at full NTSC television rates. (More about the press conference later.)
The Beginning
The liquid-crystal story at RCA Laboratories began for me in the spring of 1962. My rec-ollection of first seeing an electro-optic effect in a liquid crystal is very clear. My boss, Jan Rajchman, walked into my lab and said, "Si (Simon) Larach has something to show us." We went upstairs to Larach's lab where he had two pieces of glass with a transparent conductive coating, some yellowish powder, a hot plate, a microscope lamp, a couple of clip leads, and a 22.5-V "C" battery. He placed the powder (para-azoxyanisole) between the two pieces of glass, put the resulting sandwich on the hot plate, and shined the microscope lamp on it. Carefully adjusting the hot plate so the powder just melted (para-azoxy-anisole is a liquid crystal over a very narrow temperature range at about 120°C), he then used the clip leads and the battery to demonstrate the electro-optic effect that later became known as dynamic scattering. When the electric field was applied, there was a dramatic change – the liquid turned from clear to milky white.
What he showed us was based on the work of Richard "Dick" Williams, who worked for Larach at the time. Williams had discovered an electro-optic effect in para-azoxyanisole earlier that spring. At a field of about 1000 V/cm, he observed that visible domains, which became known as "Williams Domains," formed in the liquid-crystal layer. Williams described these results in a 1963 article,1 where he also explained that at higher fields (e.g., 2000–3000 V/cm), a stronger effect was observed characterized by vigorous agitation and stirring of the liquid. Based on recent conversations with Williams, I am certain that it was the stronger effect, which occurs under high fields, that Larach demonstrated to Rajchman and me.
Larach did not promote the discovery further, and Williams stopped working on liquid crystals in mid-1962 when he took on a new assignment at RCA in Zurich. He did, however, file a patent application on November 9, 1962. The patent2 did not issue until May 30, 1967, so its contents remained secret until then. It was RCA's first LCD patent and, although there were other disclosures and applications pending, it was RCA's only issued LCD patent at the time of the press conference the following May. Figures 1–3 of the Williams patent are reproduced in Fig. 1. Figure 3 of the patent shows the structure of a basic x-y addressed liquid-crystal matrix display.a Williams described both transmissive and reflective modes of operation in the patent, and they are illustrated in Figs. 2 and 1(b) of the patent, respectively.
A New Start
Jan Rajchman and I were excited by what we saw. The immediate problem, of course, was the high operating temperature required. Rajchman urged that the RCA Laboratories initiate a program to search for or synthesize materials that would operate at room temperature and over a wide temperature range. No further work was done on liquid crystals, however, until the fall of 1964, when George Heilmeier revisited Williams's discovery and began to investigate the physical mechanism and optical properties of the electro-optic effect in far more detail. Heilmeier worked primarily at the higher fields and called the effect "Dynamic Scattering," consistent with the turbulent nature of the activity in the liquid at the higher fields.3
As a result of Heilmeier's work and his enthusiasm about the potential of LCDs, in early 1965 RCA Laboratories' management decided to initiate a formal liquid-crystal project to understand the electro-optic effects, find or create room-temperature materials, and ultimately make a flat-panel TV display. Heilmeier and his colleagues, including Louis Zanoni and Lucian Barton, continued to investigate the electro-optic properties of various liquid-crystal materials and cell configurations. They and others at RCA Laboratories also began to make a variety of displays, mostly based on seven-segment numeric indicators, to demonstrate the potential of LCDs for clocks, wristwatches, and similar applications. Electronic shutters and dimmable mirrors were also explored. RCA made it a secret research project, where the internal progress reports were numbered and limited in distribution.4
The Chemists Come Through
Two chemists, Joseph Castellano and Joel Goldmacher, joined the project in 1965 to seek or make new liquid-crystal materials that would work at room temperature. By 1966, Castellano and Goldmacher had synthesized liquid-crystal materials that would operate at room temperature and over a sufficiently wide temperature range to be useful in practical displays. The formula for success turned out to be to use mixtures of materials that were not, on their own, liquid crystals at room temperature. Figure 2 illustrates the first nematic liquid-crystal room-temperature mixture.5 Now, it was time to get serious about applying liquid crystals to practical displays, including television.
Fig. 1: U.S. Patent 3,332,485 to Richard Williams, filed Nov. 9, 1962.
Fig. 2: Composition of first room-temperature nematic liquid-crystal mixture. (Source: Joseph Castellano.)
Time For Television
In early 1966, I was asked formally to consult to the project concerning the matrix-addressing issues essential to making a television display. At the time, I was just completing a program to develop a 1200-element ferroelectric-controlled electroluminescent matrix-addressed flat-panel display that produced moving gray-scale television images in real time.6 Juri Tults and I delivered the working display to Wright-Patterson Air Force Base in Ohio in the Spring of 1966. By that Summer, a significant fraction of my group, including Frank Marlowe, Ed Nester, and Juri Tults, and I had begun a full-time effort to solve the matrix-addressing problem for LCDs and to build a liquid-crystal television display. We had determined that individual liquid-crystal cells could be addressed with 60-μsec pulses, thus allowing line-at-a-time TV addressing. The cells did not, however, have the threshold required for matrix addressing and, other than the dielectric relaxation time of the liquid-crystal material, there was no storage. The simple approach, which bore some resemblance to the electroluminescent matrix display and which I called the "classical" method, was to use a series diode to provide a threshold and possibly to add a parallel capacitor to provide storage.
Despite efforts during 1966 to fabricate integrated thin-film diodes on glass and sapphire substrates, it became clear that using just a diode, with or without a capacitor, would not work. To overcome the poor performance of the simple classical method, I decided to add a second diode and a reset generator as shown in Fig. 3, which is reproduced from my 1969 University of Illinois paper.7 This allowed a larger capacitor and lower-amplitude addressing pulses to be used and resulted in a significant improvement in performance. This scheme was called D2C, and Philips used a variant of it in an LCD product during the 1990s.
I soon realized that a better approach would be to replace the diodes and the reset generator with a field-effect transistor. This arrangement, shown in Fig. 4, also reproduced from my 1969 University of Illinois paper,7 was called FETC. Since the field-effect transistor conducts current in both directions, the reset generator and extra bus bar are not needed. Also, reversing the polarity on alternate fields to provide AC drive is simpler. The FETC scheme is basically the same circuit configuration used in all LCD TV and computer displays today. To test and demonstrate these schemes, as well as any others that might be proposed, Tults designed and built a set of electronics, which we called "the exerciser," to drive a 2 x 18 matrix at full NTSC television rates and produce moving gray-scale images. The 36 elements were electrically buried in a full NTSC-resolution panel and addressed using the line-at-a-time addressing technique that had been developed earlier for the ferroelectric-controlled electroluminescent matrix display.
By the end of the first half of 1967, both the FETC and D2C schemes had been tested and demonstrated with external discrete components using the exerciser. Figure 5 is a photograph of a 2 x 18 display using the FETC scheme being driven by the exerciser.7The pixels are 0.06 in. square. The exerciser's electronics provided 60-μsec wide column pulses at a rate of 15.75 kHz. All 18 column drivers were activated simultaneously (line-at-a-time addressing). Each successive group of 525 column pulses contained 523 pulses of full amplitude (corresponding to a fully on pixel) followed by two pulses whose amplitude represented the brightness of a particular one of the 36 pixels in the display. The row drivers for the 2 rows each provided full-amplitude 60-μsec pulses during the 524th and 525th column pulses, respectively. The row pulse rate was thus 1/525th of 15.75 kHz, which is the 30-Hz NTSC frame rate. The video or "camera" signal that was supplied to the column drivers for the 524th and 525th column pulses was obtained from a bank of 36 potentiometers that could be set to provide whatever pattern was desired, e.g., the checkerboard and gray-scale bar shown in Fig. 5. The potentiometers were connected to the column drivers through a set of shift registers that permitted the pattern to be moved from left to right at a continuously variable rate from 0 pixels/sec (i.e., a stationary image) to 10 pixels/sec. Thus, the 2 x 18 display was capable of showing moving images and was, in fact, electrically embedded in a full 525-line NTSC display. The liquid-crystal material used had a turn-on time of 3–4 msec and a turn-off time of about 50 msec at room temperature. Thus, there was noticeable smear at motion rates above about 6 or 7 pixels/sec. However, a fairly simple scheme for achieving fast turn off is described in Refs. 7–9 and was actually tested with the 36-element display shown in Fig. 5.
Of course, I was aware of Paul Weimer's work at RCA Laboratories on evaporated thin-film field-effect transistorsb as well as the efforts on monolithic-silicon integrated circuits and silicon-on-sapphire technology, and I recognized that they all were candidates for the construction of fully integrated LCDs. Frank Marlowe worked with Weimer's group using their thin-film fabrication facility to apply the techniques of evaporated thin-film semiconductors, both diodes and field-effect transistors, to LCDs, and Ed Nester worked with the RCA Laboratories Integrated Circuit Center to do the same using a monolithic-silicon substrate as well as silicon-on-sapphire technology.
In mid-1967, we decided to design and build two 1200-element (30 x 40) integrated TV displays, one based on the D2C addressing scheme and one based on the FETC scheme. The liquid-crystal cells were to be 0.015 in. square, resulting in a 3/4-in.-diagonal display. Both displays were to be reflective, and the D2C display was to be made with evaporated thin-film diodes on a glass substrate. The FETC display was to be made using a single-crystal silicon wafer. Figure 6 shows a photograph of a test wafer for the 1200-element FETC display that was, unfortunately, not operative.
The Press Conference
On May 28, 1968, RCA's vice president for RCA Laboratories, James Hillier, presided over the famous LCD press conference during which I described and demonstrated the 2 x 18 matrix display.c Hillier opened the conference with some words about the importance of displays in the "information handling" business. He hailed the low-power reflective LCDs to be described as a breakthrough in solving the man–machine interface problem. Heilmeier then used a series of slides to give a tutorial presentation defining liquid crystals and how they can be used to make displays. He described the dynamic-scattering effect in some detail and summarized the display-related parameters that RCA Laboratories' researchers had thus far achieved, including a contrast ratio of 15:1. Next, Hillier demonstrated several displays, including a simple electronic shutter, a static display (merely a fixed pattern etched onto the electrode), a seven-segment indicator, a side-lit display that did not rely entirely on ambient light, and "an all-electronic clock that has no moving parts whatsoever." He mentioned possible applications such as traffic signs, scoreboards, and automotive and aircraft instrumentation, and said that an all-electronic wristwatch might be possible in a few years.
Hillier then introduced me to describe our work to make a TV display using liquid crystals. I used a series of slides to describe how a matrix-addressed LCD would be constructed, the need for circuitry associated with each display ele-ment in the matrix, and the need for a line-at-a-time addressing scheme. Using the exerciser electronics, Juri Tults and I then demonstrated a 2 x 18 model such as the one shown in Fig. 5. I did not mention the 1200-element displays, and, in fact, Hillier and I both made it clear that a practical liquid-crystal TV display would be very difficult to achieve because of the need to integrate the required addressing circuitry.
The End
The schedule for the 1200-element integrated displays called for them to be operational by the end of the first quarter of 1968. This proved to be overly ambitious, not only for the technical difficulty, but also because of lack of support from the RCA Laboratories Integrated Circuit Center and the thin-film fabrication facility. Neither of these facilities was under the management control of the liquid-crystal project and each had ambitious objectives of its own that did not include LCDs. Delays and technical problems arose each quarter until finally, at the end of 1968, RCA Laboratories' management decided to stop the effort on the 1200-element displays.
During 1969, RCA abandoned entirely the objective of making a liquid-crystal TV display, although other applications, e.g., watches, calculators, printers, automobile mirrors, etc., were pursued until 1972. By 1969, RCA's color-TV-receiver business was mature and the smallest consumer product of significance was a 13-in. color set. Because we could not promise to compete with such a product in any foreseeable time frame, management had no interest in investing further. I tried to keep our effort alive by writing a proposal for a government contract to continue work on the 1200-element models. I still have the typewritten original copy, complete with the standard RCA Government Proposal proprietary notice at the bottom of each page, but I do not recall that I ever got management permission to shop it with various government agencies. We did publish a paper in the Proceedings of the IEEE that described the matrix-addressing work in some detail.9
Post Script
Although RCA gave up on liquid crystals for television in 1969 and for all other applications except its fledgling wristwatch businessdin 1972, the message presented at the 1968 press conference was heard around the world, especially in Japan. By the time RCA had stopped, major efforts were under way elsewhere and, as they say, the rest is history. It is interesting that the first important commercial application of liquid-crystal matrix displays was to laptop computers, fulfilling Jim Hillier's prophecy from the 1968 press conference that the technology would solve the man–machine interface problem.
Acknowledgments
I want to thank Joseph Castellano, John van Raalte, and Richard Williams for reviewing the manuscript and providing helpful comments. I especially want to thank Alexander Magoun, Executive Director of the David Sarnoff Library, not only for his review and helpful suggestions, but also for making available the progress reports, laboratory notebooks, and other archival material from the 1960s that enabled me to refresh and confirm my memories of our work during that historic era.
References
1R. Williams, "Domains in Liquid Crystals," J. Chem. Phys. 39, No. 2, 384-388 (15 July 1963). See, also, his first notebook entry on liquid crystals, April 10, 1962, David Sarnoff Research Center Notebook 15811, David Sarnoff Library. 2U.S. Patent 3,332,485 to Richard Williams, filed Nov. 9, 1962 and issued May 30, 1967. 3G. H. Heilmeier, L. A. Zanoni, and L. A. Barton, "Dynamic Scattering: A New Electrooptic Effect in Certain Classes of Nematic Liquid Crystals," Proc. IEEE 56, No. 7, 1162-1171 (July 1968). 4Liquid Crystal/Liquid Crystal Display Progress Reports, RCA Laboratories, 1965-1972, David Sarnoff Library, Princeton, New Jersey. 5J. A. Castellano, Liquid Gold: The Story of Liquid Crystal Displays and the Creation of an Industry (World Scientific Publishing, Singapore, 2005). 6B. J. Lechner, A. G. Samusenko, G. W. Taylor, and J. Tults, "Ferroelectric Controlled Electroluminescent Displays," Proceedings of the National Aerospace Electronics Conference (May 1966). 7B. J. Lechner, "Liquid Crystal Displays," presented at a conference on Pertinent Concepts in Computer Graphics held at the University of Illinois from March 31 to April 2, 1969 and published in Pertinent Concepts in Computer Graphics, edited by M. Faiman and J. Nievergelt (University of Illinois Press, 1969). 8B .J. Lechner, F. J. Marlowe, E. O. Nester, and J. Tults, "Liquid Crystal Matrix Displays," 1969 IEEE International Solid State Circuits Conference Digest of Technical Papers, 52-53 (Feb. 1969). 9B. J. Lechner, F. J. Marlowe, E. O. Nester, and J. Tults,"Liquid Crystal Matrix Displays," Proc. IEEE 59, No. 11, 1566-1579 (Nov. 1971). •
Fig. 5: A 36-element 2 x 18 LCD operated at full NTSC TV rates. [Source: "Liquid Crystal Displays," by Bernard J. Lechner, University of Illinois, March 31, 1969, published in Pertinent Concepts in Computer Graphics (University of Illinois Press, 1969).]
Fig. 6: "Monolithic-Silicon Test Chip for 1200-Element Liquid Crystal Matrix Display." [Source: B. J. Lechner, et al. (unpublished paper, 1969).]