by W.E.

Howard
Thin-film-
transistor/
liquid crystal
display
technology-
An introduction

Liquid crystals are simple and very efficient those projected from 35-mm slides. Nevertheless, this
electro-optic transducers,or light valves. Thin- marvelous technology is losing its dominance to relatively
film transistors are simple electronic control new flat-panel technologies, principally because you cannot
devices which can be fabricated on large fit a color CRT in your pocket, nor can you fit one in a
transparent substrates. These two notebook computer.
technologies, when combined, allow the The tremendous progress in integrated electronics has
fabrication of electronic displays which brought us to the point where the electronics of a
challenge the dominance of the cathode ray television receiver is indeed pocketable, and where
tube (CRT). This paper reviews the history of significant computing capability can be packaged in a
this important development, presents the notebook-sized product. This has created a huge and
current status in comparison to the color CRT, rapidly growing demand for light, thin flat-panel displays
and describes the remaining challenges to be which can provide the images needed for these
overcome if the color CRT is truly to be applications.
displaced. Display engineers have, of course, been dreaming of
thin, flat screens since the advent of television. The
dreams have spawned many inventions over the decades,
Introduction but most of these have never attracted any significant
Two classes of electronic display have greatly changed the interest or investment, either because they could not
way we live and work-television receivers and computer match the performance of a CRT or because they could
display monitors. Until recently, both of these have been not compete in cost. In the latter case, it has always been
based upon one technology, the cathode ray tube (CRT), felt that thinness alone would not support much of a cost
which dates from the 19th century. Over the years, CRT premium, and since new technologies typically have high
technology has been refined and extended until, today, initial costs, there was a very major barrier to innovation.
one can buy high-definition television monitors with The market for portables has transformed this situation,
40-in.-diagonal screens which can display images rivaling because the CRT is no longer a competitor.
QCopyrigbt 1992 by International Business Machines Corporation. Copying inprintedform for private use is permitted without payment of royalty provided that (1) each
reproduction is done without alteration and (2) the Journal reference and IBM copyright notice are included on the first page. The title and abstract, but no other portions, of
this paper may be copied or distributed royalty free without further permission by computer-based and other information-sewice systems. Permission to republish any other
portion of this paper must be obtained from the Editor. 3

IBM J. RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992 W. E. HOWARD

 Liquid crystal displays
Liquid crystals (see for example [l]) are materials which
exhibit many of the properties of liquids but which differ in
having less symmetry than simple liquids, which are
isotropic in their properties. Among liquid crystals, the
simplest type is called nematic; it is characterized by
having a preferred direction or axis, for asymmetrical
mechanical and optical properties. Such asymmetrical
optical properties are at the root of all display applications.
The first liquid crystal displays were described by
Heilmeier et al. in 1968 [2]. They were based upon a
phenomenon known as dynamic scattering, wherein the
passage of current through a nematic liquid crystal causes
the material to break up into domains having randomly
directed axes. Since the domains are optically
asymmetrical, they scatter light, rendering the material
reflective or cloudy. This was recognized as offering a
potential for electronic displays, even though the voltage
for saturation was fairly large in relation to the threshold
Schematic view of the operation of a twisted nematic (TN)liquid
crystal cell. When no voltage is applied.(a), the polarization of the voltage for change, implying that efficient x-y addressing
transmitted light is rotated, and transmission of the cell is a maxi- would not be practical for large matrices.
mum; with voltage applied (b), the polarization is unchanged, and In 1971, a new type of liquid crystal display was
there is no transmission (see text).
described by Schadt and Helfrich [3]: the twisted nematic
(TN) display. In this display, a nematic liquid crystal is
contained by two closely spaced (5-20 pm apart) glass
The first successful high-information-contentflat panels plates which have been coated with a polymer and rubbed
were plasma panels, which emitted orange light from neon- in such a way that the nematic LC is aligned parallel to the
based gas mixtures. These are still widely used, but full- rubbing direction. If the two plates are rubbed at 90"
color versions have been difficult to develop and tend to be angles to one another, the liquid crystal deforms into a
limited in resolution, so that they are more suitable for twisted structure (see Figure 1). If polarized light is
group viewing than for personal use. They also tend to be incident on such a cell, the plane of polarization follows
power-hungry, with efficiencies typically in the range of the twist of the LC; i.e., it is rotated by 90" in passing
0.2-0.4 lumens per watt. through the cell. If a second exit polarizer (Figure 1) is
Electroluminescent (EL) displays, which are all solid- also rotated 90" with respect to the input polarizer, the
state, use the direct production of light in thin films and light passes through undisturbed. When a strong electric
powders to provide mostly yellow displays of good field is applied across the cell, the liquid crystal molecules
appearance and moderate cost. These compete rotate so as to align themselves with the electric field,
successfully with plasma displays in some applications, but overcoming the influence of the rubbing alignment. This
they also tend to be too power-hungry for battery-powered leads to a disruption of the twist and consequently the
portable products. In addition, a full-color display has been rotation of polarization. The incident light now sees
elusive because of the unavailability of a good blue EL crossed polarizers, and there is no transmission. The
phosphor. transition from full transmission to no transmission takes
The problems of these two technologies have caused place over a voltage range of 1-3 V, so shades of gray
more and more product designers to turn to liquid crystal can be achieved. This device was widely adopted for
(LC) displays, which have consequently experienced use in calculators, watches, and numerical readout
phenomenal growth. The success of liquid crystal displays displays. Twenty years later it is still dominant in these
has fostered continued development, to the point where applications.
full-color video displays have been realized which can rival From the earliest days of liquid crystal displays,
and even surpass the CRT in appearance. It is this fact, however, there has existed a desire to make complex
coupled with relatively good efficiency and the potential matrix displays which could be used, for example, to
for ultimate costs competitive with those of CRTs, which display television images. The simplest structure is an x-y
has led IBM to invest in the technology, in partnership matrix in which one plate carries row electrodes and the
with Toshiba, and which has provided the motivation for other column electrodes. The object is to have a set of
the work described in the papers of this special issue. voltage waveforms for the rows and columns such that any

W. E. HOWARD IBM J. RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992

set of intersections can be activated without turning on
unselected intersections. Since there is a threshold voltage
below which no change in transmission takes place, this 0.4
U-, IJ-, X -
x-y selection should be achievable, provided the matrix is
not too large. It soon became apparent, however, that for
typical TN characteristics, matrices with more than about 0.3
ten rows showed reduced contrast; that is, the unselected
intersections were being partially turned on.
This problem received a general treatment in 1974,in a
classic paper by Alt and Pleshko [4]. Because the TN cell
is known to respond to the square of the applied voltage,
averaged over a time shorter than the transition time, they
showed that the achievable ratio of RMS voltage at a
selected point relative to the RMS voltage at an unselected
point was a simple decreasing function of the number of
rows N being multiplexed. Specifically,
O"[
0.0 I

0
I

1
I

2
L 3 4 5
^^ ."

Voltage (V)

(V:,) - 1 + N"'*
"

(VE,) 1 - N"'*' Transmission vs. voltage for a twisted nematic liquid crystal cell,
for different horizontal viewing directions, measured from normal
Figure 2 shows the transmission vs. voltage (T-V)
characteristics for a typical TN cell, for various viewing
angles. It is clear that, even for normal viewing, a Vo,/Vo,
ratio of at least 1.4:l would be required, corresponding to
a limit of about nine rows in a matrix. Maintaining good The SBE concept was soon exploited in a flood of new
contrast over a wide range of viewing angles would reduce displays with twist angles less than 270" and greater than
further the number of rows allowed. 180". These are generically referred to as supertwisted
Great efforts were made in the '70s and early '80s to nematic, or STN, displays. Providing rather steep T-V
develop liquid crystal formulations for which the transition curves, at least for normal incidence viewing, these
was more abrupt (i.e., for which the T-V curve was displays have allowed multiplex ratios to be pushed to the
steeper), so that larger matrix displays could be produced. vicinity of 400. The contrast of such displays falls off
Indeed, considerable progress was made, so that by 1983 rapidly as the viewing angle is changed from normal
multiplex ratios of 1OO:l were being achieved, with incidence, but continued improvements based upon the use
contrasts of 1O:l or more [5]. of compensating filters and new LC materials have
Just when it appeared that improvements in 90" TN cells maintained this technology as the leading low-cost
had leveled off, the prospects for high-information-content high-information-content display.
matrixes were given a tremendous boost with the
introduction, in 1984,of the supertwisted birefringent Thin-film transistors
effect (SBE) display, in which the twist of the LC was Thin-film transistors (TFTs), or, more precisely, insulated-
increased to 270".While such a twist is not stable in a gate thin-film transistors, date back to the early OS, when
purely nematic LC, since it prefers to twist 90" in the P. K. Weimer [7]reported devices using CdS as the active
opposite direction, with the addition of cholesteric material (the concept is even older, dating back to 1934).
components (that is, LC materials with a built-in twist) a Figure 3 shows a schematic view of one common type of
270" twist can be stabilized. The T-V transition in such a TFT, a so-called inverted staggered TFT. In such a device
cell can be extremely steep; indeed, it can be bistable, Le., there is a gate electrode on the bottom, which is covered
with no stable intermediate states. The optical behavior of with an insulator, followed by the active semiconductor
SBE displays is more complex than that of TN displays material and a top passivation insulator. The passivation
and involves the birefringence of the LC, as suggested by insulator is etched back to allow source and drain contacts
the name. SBE displays also are not easy to fabricate, to be made to the semiconductor, completing the device.
since they require an expensive alignment process and When a voltage is applied to the gate, charge is induced in
very tight gap control. Nevertheless, they offered the the normally resistive active layer, making it conducting.
prospect of matrix displays with more than 100 rows with The source-drain resistance may thus vary by a factor of
good contrast. lo6 or more, providing a good switch. 5

IBM J . RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992 W. E. HOWARD

 improved threshold characteristic, as in the case of putting
Top dielechic, a-SkH a nonlinear resistor in series with each LC element, or
direct control of the LC voltage, as can be achieved by
using a TFT.
A classic paper on this subject is by Lechner et al. [9],
Glass subshate who in 1971 analyzed the requirements of an active-matrix
Gate dielectric LC display and proposed numerous possible approaches,
including the use of TFTs. It is a tribute to these authors that
almost every one of their schemes has since been
implemented in workinghigh-information-content LC
Schematic cross section of an inverted-staggered a-Si:H thin-film
displays. These include a two-diode approach, currently
transistor (TFT),characterized by a bottom gate electrode and top being pursued by Philips [lo], a ring diode approach,
source-drain contacts. employed by Citizen [ll], and a gas-discharge switch,
used by Tektronix [12] (although not inquite the same way).
Also in 1971, T. P. Brody and others at Westinghouse,
whose TFTs had been in development for some years,
Thin-film transistors were initially viewed as low-cost received a U.S. Air Force contract to build a TFTLC
alternatives to single-crystal transistors, but they were display. By 1973, a 6-in. X 6-in. display had been built
soon found to suffer from significant performance [14], although most of the effort of the group at that time
drawbacks, especially in terms of their stability and their was in fact focused on TFT/electroluminescent displays.
switching speed. This was largely due to the immature T. P. Brody is without doubt the principal pioneer of
technology of deposited insulators and the large density of TFTLC display technology. He had an unshakable vision
traps in the CdS and, later, CdSe films. of the potential of TFT technology to parallel the
Various active materials have been used in addition to development of integrated circuits, with ever-increasing
CdS: CdSe, Te, polysilicon, amorphous silicon (a-Si), complexity and perfection. He and his colleagues initially
amorphous germanium, etc. Among these, the most widely focused on Te as the active material for their TFTs, but
used today are a-Si and polySi, with CdSe a somewhat later they switched to CdSe, which offered good mobility
distant third. and which fitted well their fabrication strategy of
From a practical standpoint, the report of an amorphous patterning devices with evaporation shadow masks.
silicon TFT by LeComber et al. [8] in 1979 must be By 1979, Westinghouse had decided not to continue the
considered a major milestone. As they recognized, the TFT work. Brody, however, had not given up, and in 1981
characteristics of a-Si TFTs are remarkably well matched he founded Panelvision in order to pursue the development
to the requirements of liquid crystal driving, since they of TFT/LC displays. Panelvision ultimately became the
combine low OFF current with good ON/OFF ratios. first company to produce a commercial TFT/LC product,
Moreover, it was recognized immediately that they could in 1984, but by that time there had been a major upsurge in
benefit from the tremendous investment in a-Si solar cell TFTLC activity in Japan. The work in Japan was heavily
technology, which could provide uniform, reproducible focused on a-Si:H TFTs, although the first product, a 1-in.
film quality over large areas using plasma-enhanced pocket TV from Seiko-Epson, employed polysilicon TFTs.
chemical vapor deposition (PECVD). The same PECVD In a TFTLC display, each cell of a matrix has a TFT,
processes could also be used to deposit the gate insulator the gate of which is attached to a horizontal row electrode
material, in the same system, so that contamination of the and the drain of which is attached to a vertical column
critical interface could be avoided. Perhaps even more electrode. The source of the TFT is attached to the liquid
important, a-Si TFTs can be made at low temperatures crystal electrode (see Figure 4). The display is activated
(25O-35O0C), thus allowing the use of inexpensive glass a row at a time, by activating the gate lines. The column
substrates. electrodes carry the data voltages, synchronized to the
gate pulses, so that when a given TFT is turned on, the
TFT/LC displays data line charges up the liquid crystal capacitor, formed
It was recognized early in the history of liquid crystal between the liquid crystal picture element electrode and an
displays, even before the work of Alt and Pleshko, that opposing ground plane electrode, to the appropriate
simple matrix multiplexing of liquid crystals would not voltage on the data line at that time. Then the TFT is
permit high-information-contentdisplays, especially turned off, so that the charge is held on the capacitor until
television displays, to be made with acceptable contrast. the next refresh cycle. To avoid any flicker in the image,
Thus, there arose the idea of using thin-film devices to typical refresh rates are 60-70 Hz. The data voltage is
6 make an active matrix which would provide either an isolated from the other rows of the display by the TFTs,

W. E. HOWARD IBM J. RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992

 Analog DRAM (write. only)

Typical waveforms for a thin-film-transistorfliquidcrystal display (TlTiLCD), illustrating an ON cell (first row) and an OFF cell (second
row). The voltage V, (ON) for the ON cell is also shown.

which are turned off, so that crosstalk is very low, even reported in 1981 an approach to color LC displays using
when the number of rows in the matrix is very large, more micro color filters incorporated inside the display, to avoid
than one thousand. any problems of parallax. He argued that by putting a
The overall effect is almost the same as being able to white light behind the display and having red, green, and
control, individually and independently, the voltage at each blue filters on separate electrodes within each picture
liquid crystal element. Thus, any point on the transmission element (pixel), a color display could be made which
voltage curve is accessible, leading to good ON/OFF would be analogous to a color CRT (see Figure 5).
contrast and good gray-scale control. This approach was adopted by Morozumi et al. [16] of
Seiko-Epson when, in 1983, they captured the imagination
Full-color liquid crystal displays of the display community with a 1-in. TFT/LC color TV.
Not all liquid crystal development was directed toward the For anyone with doubts about what liquid crystal displays
addressing problem. T. Uchida of Tohoku University [15] could do, the Morozumi demonstration was persuasive. It 7

IBM J . RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992 W. E. HOWARD

 a-Si TFT

TFT substrate

Gate line

Polarizer Common substrate

Schematic rendering of a portion of the structure of a complete TFTLCD panel. Light from a fluorescent lamp passes through individual
liquid crystal cells, each of which is controlled by a thin-film transistor and each of which has associated with it a color filter. The overall
result is an array of individually controlled red, green, and blue light sources which can reproducean arbitrary image, in analogy with a color
cathode ray tube.

probably triggered more interest and more investment than are very efficient, with efficienciesof 50-70 lumens per
any display prototype in recent memory. In the last eight watt, even for small lamps. On the other hand, backlight
years, more than twenty companies have demonstrated diffusing systems are only 30-50% efficient, so that overall
TFT/LCD prototypes, with the list including all of the efficiencies are in the range of 1-2 lumens per watt for
major electronics companies in Japan, as well as a few color TFT/LC displays. This is still more efficient than
companies (or laboratories) in the U.S. and Europe. most color CRT displays.
The principal drawback to this approach to color The fluorescent lamps which are used in backlighting are
displays is the reduction of efficiency caused by absorption of the three-band type; that is, they emit predominantly in
in the color filters. Even for ideal filters, only one third of red, green, and blue spectral regions. Without this
the lamp spectrum is transmitted by each filter, and since fortunate circumstance, color filter efficiencywould be
ideal polarizers only transmit one half of the light, the considerably lower than 20%.
theoretical efficiency is less than one sixth the efficiency of The penalty in efficiency in achieving color is analogous
the lamp. In practical terms, the filters only transmit about to the situation in color CRTs, where 80-90% of the
25% of white light and polarizers about 40%, and the electron-beam current is intercepted by the shadow mask.
useful area of a cell is 50% or so, leading to an overall In addition to the loss of efficiency, there is a loss of
8 transmission of about 5%. Fortunately, fluorescent lamps resolution in color displays, in comparison to monochrome

W. E. HOWARD IBMDEVELOP.
VOL.
J . RES. 36 NO. 1 JANUARY 1992
Table 1 Comparison of color CRTs andcolorLCDs.

CCRT TFTILCD
Resolution 4/mm 5/mm
Contrast 50: 1 >200: 1
Color gamut Equal
Response time <1 ms <20 ms
Shading Analog Analog
Total content 4 X lo6 pixels >I X lo6 pixels
(low luminance)
Luminance 400 nits (cdlm') >400 nits
(<I x lo6 pixels)
Efficiency 0.5 (high res.) 1-2 lurnenw
-3.0 (low res.) lumenlW
Cost (lO-in.-diag. image) $100-300 $1500
Viewing angle 90" e50" (horiz.)
230" (vert.)

displays, since three cells and three thin-film transistors in coarse ones; TFT/LCD contrast is preserved down to
must be used for each full-color picture element, or pixel. the single-pixel level.
Moreover, the lateral displacement of the three color This set of attributes basically ensures that TFT/LCDs
arrays with respect to one another introduces a visual will replace CRTs in large numbers provided that the cost
error in the image. Finally, the color filter plates are can be reduced to be more competitive. Cost projections
currently quite expensive, so that there is a significant cost for TFT/LC technology predict that the costs will be
premium for color. competitive, in the important 10-14-in. sizes, certainly by
the year 2000 and probably in the late 1990s.
Replacing the color CRT Today, there are products in high-volume manufacture
For any color display technology to replace the color CRT, with 10-in.-diagonal screens and 640 X 480 formats.
it is essential that all of the key characteristics be matched Prototypes have been shown of million-pixel workstation
or exceeded. Resolution and contrast are the most basic. displays in sizes up to 16 in. diagonal.
Resolution must exceed 4/mm in reasonably large displays, The competition of TFTLCD technology with CCRT
with contrast greater than 50:l. TFT/LC displays have technology goes beyond direct-view applications. There is
exceeded these characteristics, especially when one a parallel competition in projection displays, which
considers contrast in high ambient illumination, e.g., in ultimately may have even greater economic impact, as
direct sunlight. In a CRT, the phosphor strongly reflects projection displays with 40-SO-in. screens are expected to
ambient light, whereas in a TFTLCD the color filters be the most popular medium for high-definition television
absorb most of the ambient light. viewing in the late 1990s. Projection cells using TFTLC
The range of colors is another important characteristic, technology, either a-Si:H or polysilicon-based, are already
and in this area the TFTLCD is at least as good as a color at the HDTV level in prototypes, and products are
CRT. Both displays allow continuous shading or gray available which give excellent renderings of standard 525-
scale, with the range being defined by the maximum line TV images. It appears that projection systems using
contrast. Some early TFTLC data displays are deficient in TFT/LC technology will be lighter, cheaper, and more
the number of gray shades allowed only because they accurate than CRT-based systems.
employ digital data drivers with 3 or 4 bits, but this is a
temporary limitation which does not apply, for example, to Remaining challenges
pocket TVs. In addition to cost, a number of challenges must be met if
Since color CRTs are widely used to show television TFT/LC displays are to realize their potential fully. First,
images, it is necessary to have a sufficiently fast response improvements to the array technology are needed if larger
time or update time for an image, usually less than 30 displays with higher content and higher resolution are to
milliseconds. be achieved. Higher-conductivity metals and higher
In several respects the TFTLCD image is superior to mobility in the semiconductor material will enable these
that of a color CRT, as is shown in Table 1. Color CRTs improvements to be made, while at the same time
suffer from defocusing and convergence errors at the edges increasing the aperture ratios, or percentages of active
of the screen, and from nonlinearity in beam positioning. area, for such displays. This is important for decreasing
They also show much lower contrast in fine patterns than power consumption. Since power is so critical for portable 9

IBM J. RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992 W. E. HOWARD

 applications, significant improvements will also be needed 7. P. K. Weimer, “The TFT, A New Thin-Film Transistor,”
in backlighting techniques. Flat fluorescent lamps, which Proc. IRE 50, 1462 (1962).
8. P. G. LeComber, W. E. Spear, and A. Ghaith,
are under development, are an example of an innovation “Amorphous Silicon Field-Effect Device and Possible
which could reduce significantly not only the power Application,” Electron. Lett. 15, 179 (1979).
consumption but also the thickness of TFT/LC displays. 9. B. J. Lechner, F. J. Marlowe, E. 0. Nester, and J. Tults,
“Liquid Crystal Matrix Display,” Proc. IEEE 59, 1566
Large areas are another challenge. While glass is readily (1971).
available in large sheets, and thin-film deposition 10. K. E. Kuijk, “DR, a Versatile Diode Matrix Liquid
techniques are well developed for large areas, we need to Crystal Approach,” Eurodisplay ’90, Proceedings of the
10th International Display Research Conference, VDE
develop patterning techniques whichwill be economical for Verlag, Berlin, 1990, p. 174.
large areas. Existing step-and-repeat exposure systems are 11. S . Togashi, K. Sekiguchi, H. Tanabe, E. Yamamoto,
relatively easily scaled to larger sizes, but theywill result K. Sorimachi, E. Tajima, H. Watanabe, and H. Shimizu,
“A 210 x 228 Matrix LCD Controlled by Double Stage
in proportionately higher costs. Also needed are better Diode Rings,” Eurodisplay ’84, Proceedings of the 4th
techniques for making large-area liquid crystal cells while International Display Research Conference, S.E.E. Paris,
maintaining the tight tolerances required on glass spacing. 1984, p. 141.
12. T. S . Buzak, “A New Active-Matrix Technique Using
The viewing angle of today’s twisted nematic displays is Plasma Addressing,” Digest of Technical Papers, 1990
not adequate for all applications. Although the viewing has SID International Symposium, Society for Information
been made quite uniform in the horizontal direction, the Display, 1990, p. 420.
13. T. P. Brody, “The Thin Film Transistor-A Late
satisfactory viewing range in the vertical direction is Flowering Bloom,” IEEE Trans. Electron Devices ED-31,
limited. We need to develop new cell designs and perhaps 1614 (1984).
even new liquid crystal modes to overcome thislimitation, 14. T. P. Brody, F. C. Luo, D. H. Davies, and E. W.
Greeneich, “Operational Characteristics of a 6 in. by 6 in.,
which is especially noticeable for video-image applications TFT Matrix Array, Liquid Crystal Display,” Digest of
and for large screens. Further in the future, perhaps solid Technical Papers, 1974 SZD International Symposium,
materials can be developedwhich could replace the liquid Society for Information Display, 1974, p. 166.
15. T. Uchida, “A Liquid Crystal Multicolor Display Using
crystal as the electro-optic transducer. Then wecould have Color Filters,” Eurodisplay ’81, Proceedings of the 1st
an all-thin-film structure on a single sheet of glass. If such International Display Research Conference, Society for
a device could also modulate unpolarized light, efficiency Information Display, 1981, p. 39.
16. S. Morozumi, K. Oguchi, S . Yazawa, T. Kodaira, H.
would be improved at the sametime. Oshima, and T. Mano, “B/W and Color LC Video Display
Finally, we must return to cost. The valueof the raw Addressed by Polysilicon TFTs,” Digest of Technical
materials in a TFT/LCD is rather low, so from the driver Papers, 1983SZD International Symposium, Society for
Information Display, 1983, p. 156.
chips to the color filters to the transistor array, even the
liquid crystal, most of the cost is in processing. As we Received November 18, 1991; accepted for publication
continue to simplify and automate processes and reap the February 4, 1992
economies of scale, we have good prospects for reducing
costs. However, experience suggests that such progress
Webster E. Howard IBM Research Division, Thomas J.
does not come easily, from a few brilliant insights, but Watson Research Center, P.O. Box 704, Yorktown
rather requires the efforts of many engineers over many Heights, New York 10598 (HOWARD at Y m ,
years. [email protected]). Dr. Howard received the B.S.
degree in physics from Carnegie Mellon University in 1955 and
the A.M. and Ph.D. degrees in physics from Harvard
References University in 1956 and 1962, respectively. At Harvard he was
1. P. G. de Gennes, The Physics of Liquid Crystals, a National Science Foundation Predoctoral Fellow. He joined
Clarendon Press, Oxford, UK, 1974. IBM in 1961, and since that time has worked at the Thomas J.
2. G. H. Heilmeier, L. A. Zanoni, and L. A. Barton, Watson Research Center as a Research Staff Member. His
“Dynamic Scattering: A New Electrooptic Effect in activities have been primarily in the areas of semiconductor
Certain Classes of Nematic Liquid Crystals,” Proc. ZEEE physics and display technology, including work on electron
56, 1162 (1968). quantization in surfaces, semiconductor superlattices,
3. M. Schadt and W. Helfrich, “Voltage-Dependent Optical heterojunctions, and the band structure of IV-VI compounds
Activity of a Twisted Nematic Liquid Crystal,” Appl. among the topics in physics, and including plasma technology,
Phys. Lett. 18, 127 (1971). electrochromics, thin-film electroluminescence, and
4. P. M. Alt and P. Pleshko, “Scanning Limitations in Liquid thin-film-transistorfliquidcrystal technology among his areas of
Crystal Displays,” ZEEE Trans. Electron Devices ED-21, interest in displays. Currently, he is manager of the Flat Panel
146 (1974). Display Technologies group. Dr. Howard is a Fellow of the
5. Y. Ishii, S . Kozaki, F. Funada, M. Matsumura, and T. American Physical Society, a Fellow of the Society for
Wada, “Optimum Polarizer Condition in TN-LCD for Information Display, a Senior Member of the IEEE, and a
High Level Multiplexing,” Japan Display ’83, Proceedings member of Sigma Xi. He is currently Secretary of the Society
of the 3rd InternationalDisplay Research Conference, for Information Display. In 1981, he was a co-recipient of the
Society for Information Display, 1983, p. 470. Wetherill Medal of the Franklin Institute, and in 1989 he
6. T. J. Scheffer and J. Nehring, “A New, Highly received an IBM Corporate Technical Award for his work on
Multiplexible Liquid Crystal-Display,” Appi. &ys. Lett. 2D electron gases. He holds seven U.S. patents in
10 45, 1021 (1984). semiconductor and display technology.

W. E. HOWARD IBM J. RES. DEVELOP. VOL. 36 NO. 1 JANUARY 1992