Thin Film Transistor

TFT

The TFTs in active-matrix LCD act as simple ON/OFF switches, at different speeds which depend on the refresh rate of the LCD, for example 60Hz. Figure 1 shows a simple structure of TFT, it consists of three terminals: the gate, the source and the drain. 

Figure 1. A simple Thin-Film-Transistor (TFT) structure

Why TFT

 

In TFT technology, a separate miniscule transistor works for each pixel on the display. As the transistors are so tiny, the charge required to operate it is very small too. This way, the display gets refreshed several times per second, ensuring great visual clarity.

In Passive Matrix LCD monitors that were in use before TFT, fast moving images could not be represented with adequate clarity. For instance, a body in motion from point A to point B would disappear between the two rest points. TFT could meet this challenge as each pixel is backed up with a transistor, and thus track the body throughout the screen. Thus TFT monitors are ideal for games, video displays and everything involving multimedia.

Working 

In a simple TFT, for example N-channel TFT, a positive voltage is applied on the gate in order to switch it ON; the insulation layer can be considered as the dielectric layer in a capacitor, hence negative charges are induced on the semiconductor channel, which is the region between source and drain; these negative charges create a electrons flow from source to drain to make the channel conductive. When a negative voltage is applied on the gate, electrons are depleted in the channel, hence almost no current is present. The ON current depends on different parameters, for example channel width, channel length, gate voltage and the threshold voltage of the TFT.

When the TFT is switched ON, a data voltage is applied on the source, the drain with the LC load capacitance will charge up to the voltage with same amplitude, i.e. transferring the data voltage from the data line to the pixel electrode. When switched OFF, no current in the channel, and data voltage cannot be transferred.

The first TFT for LCD was made of Cd-Se semiconductor thin films, however this is not compatible with normal process. In spite AMLCD with Cd-Se has a better performance, its commercialization is still not success.

While a P-channel TFT can be switched ON by applying a negative voltage on the gate, and can be switched OFF by a positive voltage on the gate.

 TFTs can be formed by three different silicons, they are: crystalline silicon, poly-silicon and amorphous silicon; and in practical manufacturing, poly-silicon can also be processed under low and high temperature, i.e. Low-Temperature Poly-Silicon (LTPS) which can be built on common low-cost glasses, and High Temperature Poly-Silicon (HTPS) which needs quartz plate.  

Since the crystalline Si owns higher mobility, it could integrate more peripheral electronics, hence higher pixel-density-required devices, like projection light valves, usually use crystalline Si.  

Amorphous silicon is widely used in LCD monitor and TV because of its easy manufacturing on large glass substrates, but it has a lower mobility; however during manufacturing, the a-Si is formed by using SiH4, the hydrogen enters into the silicon film, and can improve the loosen Si-lattice in a-Si, thus enhance its performance. The a-Si can therefore be also referred as a-Si:H. The normal electron mobility of a-Si:H is ~0.3-1 cm2/Vsec, compared with c-Si’s >500 cm2/Vsec, it is quite small. But for AMLCD’s TFT’s switch, it is enough. On the other hand, its hole mobility is very low, therefore only N-channel TFT can be practically used. Another drawback of a-Si is its high photoconductivity, which cause the undesirable photo-leakage current in the OFF state. To avoid it, a cover layer is used to shield it from ambient and backlight.

Poly-Si can be used to make both P-channel and N-channel TFTs. Because of its relatively high mobility, both row and column drivers can be integrated on the glass, even D/A converters, DC/DC converters and (micro)processors can be integrated too, which significantly cut the cost from external driver and other devices’ chips. However the off current of Poly-Si is much higher than a-Si, i.e. the OFF state is not stable because of the charge on the pixel capacitor cannot be maintained. In order to decrease the OFF current, a dual gate structure and a lightly doped drain (LDD) were proposed. Both methods can effectively lower the OFF current.

Figure 2 shows a typical pixel structure of AMLCD. It is worth to note that a storage capacitor is connected in parallel to pixel capacitor in order to retain the charge at OFF state, and decrease the voltage dependence and leakage current in the LC capacitance, hence the control of RMS voltage on the pixel is easier.

Figure 2. A TFT AMLCD's pixel layout (color-filter substrate is not shown).

Posted in: Technology