Liquid crystals in displays

The contents on this page originate to a large extent from the doctoral thesis of Marek Matuszczyk.

The first liquid crystal displays that became successful on the market were the small displays in the digital watches of the eighties. These were simple twisted nematic (TN) devices with characteristics satisfactory for such simple applications but by no means possible to use for larger screens. The vast development of liquid crystal technology that followed has given us both totally new types of liquid crystals (FLC) and much improved techniques based on nematic liquid crystals (STN and TFT). These are the techniques that make today's laptop computers and flat table top displays possible.

The typical liquid crystal cell is a sandwich construction where the liquid crystal substance is encapsulated between two glassplates (substrates). These are coated on the inside with a conductive layer and a polymer layer which are both transparent. The conductive layer (Indium Tin Oxide, ITO) is needed in order to apply an electric field over the cell, and the polymer coating is used to control the orientation of LC molecules next to the substrates. By mechanically rubbing the polymer, fine grooves, along which the LC molecules want to align, are created.

The Twisted Nematic (TN) cell

In 1971 scientists at Hoffmann-Laroche presented the idea behind the twisted nematic display. The rubbed polymer coatings make sure that the LC molecules align parallel to the glassplates in one well defined direction. The cell is assembled such that the rubbing directions on the upper and lower substrates are perpendicular to each other. When the cell is filled with a nematic liquid crystal, these conflicting surface conditions create a continuous twist in the alignment (director) configuration by 90 degrees across the cell (figure 1a).

Figure 1. The construction and operation of the twisted nematic cell.
a) Electric field off. The LC molecules at the upper substrate are forced to orient perpendicularly to those at the lower. The conflicting surface conditions lead to a helical structure with a 90° twist between the substrates. This has a guiding function on the light polarization plane so that the cell appears transparent.
b) Electric field on. The molecules align with the field leading to the disappearance of the helical structure and, with that, the guiding function. Light which passes through the polarizer is now blocked by the analyzer and the cell appears black.

On the outside of the substrates, linear polarizers oriented parallel to the rubbing directions of their respective substrate, are applied. When light enters the cell the first polarizer lets through only the component oscillating parallel to the LC director next to the entrance substrate. During the passage through the cell, the polarization plane is turned along with the director helix, so that when the light wave arrives at the exit polarizer, it passes unobstructed. The cell is thus transparent in this state, called the OFF state.

If we now apply an electric field over the cell, the molecules will want to align parallel to the field (there are nematic substances where the molecules orient perpendicular to the field, but these are of course not used in TN cells). With the molecules oriented in this way the helical structure disappears, and so does the guiding of the light polarization plane. Thus light cannot pass both polarizers and the cell appears black - the ON state. By dividing the electrodes into pixels, different parts of the cell can be independently switched between black and transparent, and an information display can thus be constructed.

The electro-optic response characteristic of the TN cell is asymmetric because only the non-transmissive state can be activated by the electric field. When the electric field goes to zero, the twisted structure, which provides wave-guiding of the incident light, is restored by the elastic torques. The typical switching times of a TN cell are in the millisecond range. Nematic liquid crystals exhibit good mechanical shock stability, because they spontaneously return to the uniform alignment after mechanical distortion. Despite relatively long switching time and limited viewing angle, they are very useful for device applications and have nowadays a broad temperature range, simple construction and low price. The TN technology has, however, very flat threshold characteristics, leading to low contrast and viewing angle in a display scanned line by line. This makes it practically useless for such applications as computer displays.

The Supertwisted Nematic (STN) display

In a normal twisted nematic display the OFF state is characterized by a director twist of 90 degrees. Suppose we would increase this angle to, for instance, 120 degrees. Because the two liquid crystal states n and -n are indistinguishable, this twist would be unstable and tend to jump to the twist -60 degrees which has a lower energy. Hence, 90 degrees is the highest twist we can introduce without any special precautions. In fact, also the 90 degrees case is ambiguous: once we have unwound the twist by a strong field, the liquid crystal does not know whether to go back to a +90 or -90 degree twist, because these are equivalent. This indeed disturbed the operation of the first TN displays (giving irregular domains of twist and anti-twist when relaxing back to the field-off state) until a chiral dopant (if you are not familiar with the concept of chirality, check out the tutorial!) was mixed in, stabilizing the 90 degree twist of only one handedness. Now the liquid crystal has a "memory" telling how to relax back; the "spring" has now a right or left handed torque.

Using chiral dopants we may increase the twist to any value, as long as we increase the amount of dopant. In the end we would arrive at a cholesteric structure. The unwinding of a cholesteric in an electric or magnetic field is a classic problem in liquid crystal physics. There exists a certain threshold below which essentially nothing happens. On approaching the threshold level the helix unwinds almost at once for very small changes in the field. This means that in the supertwisted display the electro-optic characteristic can be made very steep (Figure 2). A steeper characteristic is very desirable for a nematic display. If the threshold characteristics is sharp, the ratio between ON and OFF voltages can be very close to unity and much higher multiplexing rates become possible. The STN display can have about 500 lines, while standard TN displays can hardly reach 100 lines. But a large twist has also disadvantages. One is that part of the waveguiding gets lost and therefore the display becomes chromatic (colored). Another is that the response is slower than for the TN display.

Figure 2. Schematic electro-optic characteristic of the supertwist (STN) cell compared to the one of the twisted nematic (TN) cell.

The Active-Matrix (TFT) Liquid Crystal Display

An active-matrix liquid crystal display is controlled by an array of thin-film transistors (TFTs), each activating a single pixel using a capacitor. TFT display technology uses deposited layers of different materials to constitute the semiconductor, the insulators and the electrodes. In TFTs, as in conventional transistors, two terminals conduct current and the third one (gate) switches the transistor on and off. The electrodes are arranged as gate lines (rows) and addressing lines (columns). One row at a time is activated. In the open state the capacitors within pixels on the chosen (i.e. addressed) row are charged to the desired amount given by the column voltage, resulting in the desired transmission change through the liquid crystal element. When the capacitors have been charged, the transistors are switched off and the pixel capacitor provides a holding voltage over the LC element. The next row can now be addressed, while the remaining rows are effectively isolated from disturbing voltage pulses (cross-talk). The entire display is rewritten after all the rows have been addressed. Such refreshing prevents fading out of the picture which would otherwise happen as charge gradually leaks from the pixel capacitors, reducing their voltages and thus changing the light transmission states of the pixels.

Most of the parts of both transistors and capacitors are formed on the surface of a single, semiconducting crystal, whose electrical properties are modified in particular regions by the addition of charge-donating atoms, called dopants. The TFT can be fabricated on virtually any surface, including glass. One definite disadvantage of TFT displays is that the transistor blocks part of the light-path, thereby limiting the resolution to considerably less than achievable with passive displays (for instance the FLC technique described below).

The inconvenience of the unsharp threshold for the twisted nematic becomes an advantage in connection with the TFT: the TN pixel provides a good grey scale as a function of the charge level of its capacitor and the cross-talk is eliminated by the TFT. Thus, there is no point in combining TFT with STN. The TFT-FLC, however, is a most powerful, though expensive combination. The TFT technology is very complex and the cost increases rapidly with increasing area of the matrix.

The Surface Stabilized Ferroelectric Liquid Crystal (SSFLC) display technique

Ferroelectric liquid crystals are polar materials. The local dipole density that couples to an external electric field must lie perpendicular to the director, otherwise n and -n would not be identical states. Due to the helicoidal structure of a bulk sample, the local polarisation is thus cancelled on the macroscopic scale. However, every uniform surface condition is in conflict with the helical bulk condition and this can be used to unwind the helix. When the cell thickness is chosen below a certain value the torques enforcing the boundary condition will penetrate the whole liquid crystal. Only two selected states will then be possible for the LC director. The spontaneous polarization will no longer be a local property but a macroscopic polarization will appear. When the smectic C* liquid crystal has thus been forced out of its natural crystallographic state by surface torques the Surface Stabilized Liquid Crystal (SSFLC) structure is obtained. In order to preserve the microscopic uniform orientation of the LC director, and hence the macroscopic polarization, the distance between the confining plates has to be less, or at least of the same order, as the natural helical pitch of the smectic C* material (in the order of 1 to 2 micrometers). This is then a necessary condition for bistability. As it turns out, this thickness also allows to meet the optimum optical condition (the lambda/2 condition) without designing molecules with particularly low birefringence.

Figure 3 a) The helical bulk state in the smectic C*. The director n is tilted by an angle theta with respect to the layer normal k; b) The helicoidal structure is incompatible with a homogeneous boundary condition with the director along the surface and can, therefore, be elastically unwound if a cell is sufficiently thin. The molecular dipole moment may thus point DOWN or UP resulting in the macroscopic polarization.

The SSFLC can be considered as a birefringent slab because of the uniform molecular director orientation and the associated optical anisotropy. The optic axis of the system can be rotated between two stable states (with the polarization pointing UP or DOWN) by an external electric field (Figure 4). The switching process is dominated by a strong ferroelectric torque gamma on the polarization (P) by the applied electric field (E),

Both transmitting and non-transmitting state can be activated by electric pulses. The chosen state remains when the external field is removed. In the non-transmitting state the optic axis is along the direction of one of the crossed polarizers. In the transmitting state the optic axis has been switched to the other side of the smectic C* cone. If the angular difference between these two axis directions is 45 degrees and if the optical thickness of the smectic matches the lambda/2-condition, then the polarization plane of the incoming light is rotated 90 degrees to coincide with the pass direction of the analyzer.

Figure 4. Simplified picture of an SSFLC display. The smectic layers have been omitted for clarity. The symmetry axis of the smectic cone corresponds to the layer normal and the cone surface corresponds to the possible directions of the optic axis.

The linear coupling between gamma and E makes the switching active in both directions on field reversal. Therefore, the switching to a new state is always much faster than the intrinsic elasticity-controlled relaxation of a nematic, and so, SSFLC devices can operate orders of magnitude faster than non-ferroelectric liquid crystal devices.

Each FLC element can be considered a single modulator affecting the properties of incident light. However, the actual layer structure is considerably more complicated than indicated in Figure 4. In reality the layers show a kink or fold somewhere in the middle giving a "chevron" structure.

Ferroelectric liquid crystals can be used in a broad range of applications, beginning from single-pixel light choppers and approaching large area HDTV screens, due to the high modulation frequency combined with possibility for long time memorization of a chosen binary state. The pictures below (click on the thumbnails to see a larger picture) show the first Canon FLC display prototype from 1988 and the first commercial FLC screen from 1995.

Check out two of the coolest FLC companies (links to company presentations)!


Previous topic: Nature's ways of applying liquid crystals

Back to contents

Next topic: Thermal sensors