Professional Documents
Culture Documents
Liquid crystal was discovered by the Austrian botanist Fredreich Rheinizer in 1888. "Liquid crystal" is
neither solid nor liquid (an example is soapy water).
In the mid-1960s, scientists showed that liquid crystals when stimulated by an external electrical
charge could change the properties of light passing through the crystals.
The early prototypes (late 1960s) were too unstable for mass production. But all of that changed
when a British researcher proposed a stable, liquid crystal material (biphenyl).
Today's color LCD TVs and LCD Monitors have a sandwich-like structure (see figure below).
TFT LCD (Thin Film Transistor Liquid Crystal Display) has a sandwich-like structure with liquid
crystal filled between two glass plates.
TFT Glass has as many TFTs as the number of pixels displayed, while a Color Filter Glass has color
filter which generates color. Liquid crystals move according to the difference in voltage between the
Color Filter Glass and the TFT Glass. The amount of light supplied by Back Light is determined by
the amount of movement of the liquid crystals in such a way as to generate color.
The most common liquid-crystal displays (LCDs) in use today rely on picture elements, or pixels,
formed by liquid-crystal (LC) cells that change the polarization direction of light passing through them
in response to an electrical voltage.
As the polarization direction changes, more or less of the light is able to pass through a polarizing
layer on the face of the display. Change the voltage, and the amount of light is changed.
There are two ways to produce a liquid-crystal image with such cells: the segment driving method
and the matrix driving method.
The segment driving method displays characters and pictures with cells defined by patterned
electrodes.
The matrix driving method displays characters and pictures in sets of dots.
The segment drive method is used for simple displays, such as those in calculators, while the dot-
matrix drive method is used for high-resolution displays, such as those in portable computers and
TFT monitors.
Two types of drive method are used for matrix displays. In the static, or direct, drive method, each
pixel is individually wired to a driver. This is a simple driving method, but, as the number of pixels is
increased, the wiring becomes very complex. An alternative method is the multiplex drive method, in
which the pixels are arranged and wired in a matrix format.
To drive the pixels of a dot-matrix LCD, a voltage can be applied at the intersections of specific
vertical signal electrodes and specific horizontal scanning electrodes. This method involves driving
several pixels at the same time by time-division in a pulse drive. Therefore, it is also called a
multiplex, or dynamic, drive method.
In passive-matrix LCDs (PMLCDs) there are no switching devices, and each pixel is addressed for
more than one frame time. The effective voltage applied to the LC must average the signal voltage
pulses over several frame times, which results in a slow response time of greater than 150 msec and
a reduction of the maximum contrast ratio. The addressing of a PMLCD also produces a kind of
crosstalk that produces blurred images because non-selected pixels are driven through a secondary
signal-voltage path. In active-matrix LCDs (AMLCDs), on the other hand, a switching device and a
storage capacitor are integrated at the each cross point of the electrodes.
The active addressing removes the multiplexing limitations by incorporating an active switching
element. In contrast to passive-matrix LCDs, AMLCDs have no inherent limitation in the number of
scan lines, and they present fewer cross-talk issues. There are many kinds of AMLCD. For their
integrated switching devices most use transistors made of deposited thin films, which are therefore
called thin-film transistors (TFTs).
An alternative TFT technology, polycrystalline silicon - or polysilicon or p-Si-is costly to produce and
especially difficult to fabricate when manufacturing large-area displays.
Nearly all TFT LCDs are made from a-Si because of the technology's economy and maturity, but the
electron mobility of a p-Si TFT is one or two orders of magnitude greater than that of an a-Si TFT.
This makes the p-Si TFT a good candidate for an TFT array containing integrated drivers, which is
likely to be an attractive choice for small, high definition displays such as view finders and projection
displays.
A TFT LCD module consists of a TFT panel, driving-circuit unit, backlight system, and assembly unit.
1. LCD Panel
- TFT-Array Substrate
- Color Filter Substrate
2. Driving Circuit Unit
- LCD Driver IC (LDI) Chips
- Multi-layer PCBs
- Driving Circuits
3. Backlight & Chassis Unit
- Backlight Unit
- Chassis Assembly
It is commonly used to display characters and graphic images when connected a host system.
The TFT LCD panel consists of a TFT-array substrate and a color-filter substrate.
The TFT-array substrate contains the TFTs, storage capacitors, pixel electrodes, and interconnect
wiring. The color filter contains the black matrix and resin film containing three primary-color - red,
green, and blue - dyes or pigments. The two glass substrates are assembled with a sealant, the gap
between them is maintained by spacers, and LC material is injected into the gap between the
substrates. Two sheets of polarizer film are attached to the outer faces of the sandwich formed by
the glass substrates. A set of bonding pads are fabricated on each end of the gate and data-signal
bus-lines to attach LCD Driver IC (LDI) chips
Driving an a-Si TFT LCD requires a driving circuit unit consisting of a set of LCD driving IC (LDI)
To reduce the footprint of the LCD module, the drive circuit unit can be placed on the backside of the
LCD module by using bent Tape Carrier Packages (TCPs) and a tapered light-guide panel (LGP).
A TFT LCD panel contains a specific number of unit pixels often called subpixels.
Each unit pixel has a TFT, a pixel electrode (IT0), and a storage capacitor (Cs).
For example, an SVGA color TFT LCD panel has total of 800x3x600, or 1,440,000, unit pixels.
Each unit pixel is connected to one of the gate bus-lines and one of the data bus-lines in a 3mxn
matrix format. The matrix is 2400x600 for SVGA.
Because each unit pixel is connected through the matrix, each is individually addressable from the
bonding pads at the ends of the rows and columns.
The performance of the TFT LCD is related to the design parameters of the unit pixel, i.e., the
channel width W and the channel length L of the TFT, the overlap between TFT electrodes, the sizes
of the storage capacitor and pixel electrode, and the space between these elements.
The design parameters associated with the black matrix, the bus-lines, and the routing of the bus
lines also set very important performance limits on the LCD.
In a TFT LCD's unit pixel, the liquid crystal layer on the ITO pixel electrode forms a capacitor whose
counter electrode is the common electrode on the color-filter substrate.
A storage capacitor (Cs) and liquid-crystal capacitor (CLC) are connected as a load on the TFT.
Applying a positive pulse of about 20V peak-to-peak to a gate electrode through a gate bus-line turns
the TFT on. Clc and Cs are charged and the voltage level on the pixel electrode rises to the signal
voltage level (+8 V) applied to the data bus-line.
The voltage on the pixel electrode is subjected to a level shift of DV resulting from a parasitic
capacitance between the gate and drain electrodes when the gate voltage turns from the ON to OFF
state. After the level shift, this charged state can be maintained as the gate voltage goes to -5 V, at
which time the TFT turns off. The main function of the Cs is to maintain the voltage on the pixel
electrode until the next signal voltage is applied.
Liquid crystal must be driven with an alternating current to prevent any deterioration of image quality
resulting from dc stress.
This is usually implemented with a frame-reversal drive method, in which the voltage applied to each
pixel varies from frame to frame. If the LC voltage changes unevenly between frames, the result
would be a 30-Hz flicker.
(One frame period is normally 1/60 of a second.) Other drive methods are available that prevent this
flicker problem.
In an active-matrix panel, the gate and source electrodes are used on a shared basis, but each unit
pixel is individually addressable by selecting the appropriate two contact pads at the ends of the rows
and columns.
By scanning the gate bus-lines sequentially, and by applying signal voltages to all source bus-lines in
a specified sequence, we can address all pixels. One result of all this is that the addressing of an
AMLCD is done line by line.
Virtually all AMLCDs are designed to produce gray levels - intermediate brightness levels between
the brightest white and the darkest black a unit pixel can generate. There can be either a discrete
numbers of levels - such as 8, 16, 64, or 256 - or a continuous gradation of levels, depending on the
LDI.
Generating Colors
The color filter of a TFT LCD TV consists of three primary colors - red (R), green (G), and blue (B) -
which are included on the color-filter substrate.
The elements of this color filter line up one-to-one with the unit pixels on the TFT-array substrate.
Each pixel in a color LCD is subdivided into three subpixels, where one set of RGB subpixels is equal
to one pixel.
(Each subpixel consists of what we've been calling a unit pixel up to this point.)
Because the subpixels are too small to distinguish independently, the RGB elements appear to the
human eye as a mixture of the three colors.
Any color, with some qualifications, can be produced by mixing these three primary colors.
The total number of display colors using an n-bit LDI is given by 23n, because each subpixel can
generate 2n different transmittance levels.
NEXT
Fabricating TFT LCD
We would like to express our appreciation to Samsung Electronics for the preceding information.
The pressure to reduce the manufacturing cost of TFT LCD displays is as constant and intense as it
is in the semiconductor industry. To increase productivity, IC makers continuously reduce the sizes of
c-Si chips and transistors in order to increase the number of chips per wafer.
But this strategy doesn't work for LCDs because the panel sizes users demand most get steadily
larger, not smaller.
Still, by increasing the number of panels produced on a single substrate, the cost of TFT-array
processes can be reduced.
The IC makers' size-reduction strategy doesn't work for direct-view LCDs, but
LCD manufacturers can still reduce the cost of TFT-array processes by
increasing the number of panels produced on a single substrate.
This process requires that the size of the glass substrate be steadily increased so that the number of
LCD panels fabricated upon it can increase.
For more panels to be put on a glass substrate, the substrate size must be
steadily increased - which requires the continual design and construction of
new generations of process equipment.
New generations of process equipment must be continually designed and built to achieve these
increases.
The fabrication processes this equipment must implement will be described below.
We can assume that the display being fabricated is a color TFT LCD that uses an inverse-staggered-
type a-Si TFT as the active-matrix switching element.
The manufacturing process used to fabricate an a-Si TFT array is very similar to those used to
fabricate c-Si semiconductor devices. The various steps, including cleaning, deposition of thin films,
photolithography, and wet and dry etching of the thin films - are alsso very similar. The difference
between the a-Si TFT process and the c-Si semiconductor process is that a semiconductor layer is
deposited onto a glass substrate in the a-Si TFT process, while Si wafers are used as the substrate
in the c-Si semiconductor process. Today, critical issues in the processing of TFT arrays include the
development of a low-resistance gate-bus line, uniform and fine etching, and improved lithographic
accuracy.
TFT-array technologies are aimed at achieving high precision, large aperture ratio, and low power
consumption, in addition to large screen size.
AMLCD manufacturers are also competing to minimize the number of array processes by reducing
the number of photo masks and simplifying the thin-film-formation and etching processes.
In the bottom-gate TFT-array fabrication process, the first layer consists of the gate electrodes and
gate bus-lines, which can have one or two metal layers.
Some storage capacitors can be constructed by using a part of the gate electrode as an electrode of
the storage capacitor - which is called the Cs-on-gate method - while other capacitors are
constructed independent of a gate bus-line.
If the independent Cs lines are constructed simultaneously with the gate bus-lines using the same
metal layer, there is no difference in the fabrication process between the Cs-on-gate method and the
independent Cs bus-line method.
The processing of an a-Si TFT array is complex.
This flowchart outlines the processes for making an a-Si TFT array using a
bottom-gate TFT structure and an independent storage capacitor.
After constructing gate and storage-capacitor electrodes with 2000-3000A of a metal such as
aluminum, chromium, tantalum, or tungsten, a triple layer of silicon nitride and amorphous silicon is
deposited by using plasma-enhanced chemical-vapor deposition (PECVD).
In the etch-back type of TFT structure, the triple layer consists of 4000A of SiNx, 2000A of a-Si, and
500 A OF n+a-si, which is deposited over the gate electrode in a continuous process, i.e., a process
without a vacuum break.
For the etch-stopper type of TFT structure, 4000A of SiNx, 500A OF a-Si, and 2000A of n+a-si are
deposited.
Let us look at the etch-back TFT fabrication process in more detail.
TFT Fabrication
After defining the a-Si area by using photolithography and plasma dry etching, an ITO layer is
deposited with a thickness of about 500A via sputtering.
Then, the pixel electrodes are patterned. About 2000A of metal is sputter deposited, while data bus-
lines and TFT electrodes are patterned by photolithography.
Then the ohmic contact layer (n+a-Si) at the channel region is etched by dry etching using the source
and drain electrodes as an etch-protect mask.
Finally, a protective 2500A SiNx layer is deposited by PECVD and contact windows are opened.
The etch-stopper TFT structure requires one more process step - a chemical vapor deposition (CVD)
- than does the etch-back TFT structure.
For etch-stopper TFT fabrication, a n+a-Si layer is deposited separately after the top insulator of
triple-layer (SiNx/a-Si/SiNx) is patterned.
The a-Si area is patterned and the n+a-Si layer at the top of etch-stopper is removed. The source
and drain electrodes are formed using about 2000A of metal; then, about 500A of ITO is sputter
deposited, and pixel electrodes are patterned.
A SiNx protective layer is then deposited by PECVD and, finally, the contact windows are opened.
Color filters (CFs) can be made with either dyes or pigments, utilizing coloring method such as
dyeing, diffusion, electro-deposition, and printing.
Color filters (CFs) can be made with either dyes or pigments, and can be
further divided by coloring method.
Among the many combinations of configuration and types of CF fabrication methods, the color-resist
method with stripe-type RGB arrangement is currently the most popular.
Between the blocks of color in the CF is a black matrix (BM) made of an opaque metal, such as
chromium, which shields the a-Si TFTs from stray light and prevents light leakage between pixels.
A double layer of Cr and CrOx is used to minimize reflection from the BM.
The sputter-deposited BM film is patterned using photolithography.
For reduced cost and reflectivity, black resin made by diffusing C and Ti in photo resist - can be used
as a BM material.
In the color-resist method, the primary color-filter patterns are formed by using a photolithography
technique.
The color-resist is negative and made by diffusing pigment in a UV-curing resin, such as an acryl-
epoxy resin, and by dissolving the resin in a solvent.
A red colored resist is spin-coated onto a glass substrate on which a BM has previously been formed.
The red pattern is then formed by exposing the red resist through a mask and developing it.
The process is repeated using the same mask with a shifted mask-align technique for green- and
blue-colored resins.
A protective film is then applied, and 1500A of ITO for the TFT array's common electrode is sputter-
deposited to finish the color filter.
The TFT-array and color-filter substrates are made into an LCD panel by assembling the two
substrates together with a sealant, while the cell gap is maintained by spacers.
The TFT-array and color-filter substrates are made into an LCD panel by
assembling them with a sealant.
The assembly is begun by printing a polyimide alignment film on a cleaned TFT-array, and then
rubbing the surface of the film with a piece of cloth wound on a roller, which orients the polyimide
molecules in one direction.
Similarly, alignment film is applied to the color-filter substrate, and this substrate is also rubbed.
After the rubbing process, a sealant is applied to the periphery of the TFT-array substrate. To form
electrical connections from the common electrodes on the color-filter substrate to the TFT array, the
TFT-array substrate is coated with a conducting paste around the periphery.
At the same time, spacers to control the cell gap are sprayed onto the color-filter substrate. (In some
cases, spacers are sprayed on to the TFT-array substrate, and a sealant is applied to the color-filter
substrate.)
The two substrates are then assembled after the sealant is pre-hardened.
The sealant is then hardened completely with heat and pressure.
Then, the assembled substrates are scribed using a diamond wheel and separated into individual
cells, and the empty cells are filled with liquid crystal material by vacuum injection.
Finally, a sealing agent is used to seal the cell, and the polarizers are applied to both cell surfaces
after a visual function test.
Although critical for producing panels with the desired characteristics and price, the details of the
manufacturing process for AMLCD panels are often of less immediate interest to the OEM
purchasers of displays than are the details of the module assembly process.
This is so because it is the physical and electrical characteristics of the module that OEMs must deal
with when integrating the display into products for end users.
The process flow for assembling a module using the tape-automated-bonding (TAB) method is
The first decision to make is whether you want to use TAB at all, or whether you would prefer the
other basic way of applying the LDI chips needed to drive the TFT panel.
In the TAB method, the LDI chip is attached to a tape-carrier package (TCP), and the TCPs are then
connected to the TFT-array substrate.
Anisotropic conducting film (ACF) is applied to the contact pads, where the stripe-shaped contact
leads are formed as a group. The TCPs are then aligned and subjected to pressure-bonding.
The drive-circuit components, such as the timing controller, EMI filters, op amps, chip capacitors, and
resistors, are mounted onto a multi-layered PCB using a surface-mount technology (SMT).
A soldering method is usually employed to connect the gate and control PCBs to the other end of the
TCP leads, but in some cases ACF bonding can be used instead.
Sometimes, to minimize bezel size, the drive-circuit unit is set to the back side of the LCD module by
using bent TCPs.
Alternatively, one can use the chip-on-glass (COG) method, in which LDI chips are mounted directly
on the TFT-array substrate.
The choice of COG or TAB is determined by the peripheral area available and the limitations on
bezel size for the display.
After testing the electrical functions, only the good LCD panels are subjected to the final assembly
process, in which a backlight unit and a metal bezel are attached to compete the LCD module.
There are many structures for thin-film transistors (TFTs), with the first major distinction among them
being planar CMOS structures vs. staggered amorphous-silicon (a-Si) structures.
The a-Si TFTs are further divided into staggered and inverse-staggered types.
In the inverse-staggered type, the ohmic layer (n+ a-Si) in the channel region can either be etched
directly (the etch-back method) or etched by forming a protective film on the a-Si thin film (the etch-
stopper method).
Each method has its own set of advantages and disadvantages. The inverse-staggered structure
offers a relatively simple fabrication process and an electron mobility that is about 30 percent larger
than that of the staggered type. These advantages have resulted in the bottom-gate TFT structure
becoming more widely adopted in TFT-LCD design, despite the fact that it's technically an upside-
down structure.
Because a-Si has photoelectric characteristics, the a-Si TFT must be shielded from incident light .
The a-Si layer must also be as thin as possible to minimize the generation of photo-induced current,
which can cause the TFT to malfunction.
In the top-gate structure, a light-shield layer must first be formed at the region of the TFT channel
The formation of this light shield may cause an extra process step. In bottom-gate TFTs, on the other
hand, a gate electrode is first formed at the TFT channel region, where it also serves as a light-shield
layer.
The operational characteristics of a TFT are determined by the sizes of its electrodes, the W/L ratio,
and the overlap between the gate electrode and the source-drain .
The parasitic capacitances resulting from the overlap of electrodes can not be avoided in staggered
TFT structures, but the parasitic effects must be minimized to maximize the LCD's performance.
To reduce the overlap between the electrodes, a self-align process is often implemented .
It turns out that the characteristics of the a-Si TFTs used in AMLCDs are very similar to the
characteristics of the MOSFETs in semiconductor devices.
When a TFT panel is operated under real-world conditions, the gate voltage is set at either 20 V for
switch-on, or at -5 V for switch-off. Under these operating conditions, the a-Si TFT is a good
switching device with an on/off current ratio larger than 106.
The performance of the TFT also depends on fabrication process parameters, such as electron
mobility and thickness of the gate insulators. If we wish to increase the current gain of the TFT for
better pixel-switching performance, and the process parameters are fixed, the only thing we can do is
increase the W/L ratio. But doing this is not without a significant trade-off: The larger W/L results in a
lower aperture ratio - less of the pixel's area is transparent to light when the pixel is ON - so the
display's brightness and contrast are reduced.
To maintain a constant voltage on a charged pixel over the entire frame cycle, a storage capacitor
(Cs) is fabricated at each pixel. A large Cs can improve the voltage holding ratio of the pixel and
reduce the kickback voltage, with resulting improvements in contrast and flicker, but a large Cs
results in a lower aperture ratio and higher TFT load.
The storage capacitor can be formed by using either an independent storage-capacitor electrode or
part of the gate bus-line as a storage-capacitor electrode (Cs-on-gate method)
The advantages of the Cs-on-gate method are that it eliminates the need for modification in the
fabrication process; it minimizes the number of processes; and it produces a larger aperture ratio
than does the independent Cs method. But few things are free in TFT-LCD design. The trade-off with
the Cs-on-gate method is an increase in the RC time constant of the gate bus-line, which reduces the
TFT switching performance.
This RC delay problem can have serious effects on the appearance of the display.
The solution lies in fabricating the gate bus-line with a low-resistance material such as aluminum (Al).
The requirement that the gate bus-line must have a small RC time delay is particularly important for
larger and higher-resolution LCDs. If the widths of the signal bus-lines are increased to reduce
resistance, the aperture ratio of the pixels is reduced, so the preferred approach is to use a low-
resistance material for the bus-lines. For this, Al offers advantage over other metals, such as Cr, W,
and Ta.
But, in the bottom-gate TFT process, the gate electrodes are first fabricated on the glass substrate
and then subjected to high-temperature processes and various chemical etches. So, to use Al as a
gate-electrode material, the Al gate electrodes must be protected from damage produced by hillock
formation.
A thin film of an aluminum oxide (Al2O3), formed by anodic oxidation of the Al surface at room
temperature, can protect the Al electrodes from the problems associated with hillock formation.
Double-metal or clad structures over the Al electrodes - using a relatively stable material such as Cr,
Ta, or W - can also be used to protect the Al electrodes. The trade-off is that these approaches
require an additional process. Recently, Al alloy (such as Al-Nd), which can suppress hillock
formation, has been used as a gate-electrode material to eliminate the additional process.
Aperture Ratio
As implied previously, another important design consideration is maximizing the aperture ratio of the
pixel. In the unit cell, TFT electrodes, storage-capacitor electrodes, signal bus-lines, and the black-
matrix material constitute opaque areas.
The combined areas of these elements, along with the area of the pixel aperture through which light
can pass, determine the aperture ratio of the pixel. The aperture ratio is given by the area of the pixel
aperture divided by the total pixel area (aperture area plus the area of the opaque elements). To
increase the aperture ratio as much as possible, the size of the opaque elements must be made as
small as possible, while maintaining a design that maximizes the size of the pixel-electrode area.
Unfortunately, one can only go so far in reducing the opaque areas before degrading image quality
and yield. As shown in Fig. 12, the light-shield area on the color-filter substrate must be extended to
block the light leaking through the gap between the data-line and the pixel ITO. To do this in
conventional TFT-LCD cell structures, while simultaneously providing an adequate plate-alignment
margin, significantly reduces the aperture.
But far higher aperture ratios can be achieved by switching from a conventional structure to the BM-
on-Array structure, regardless of the accuracy of the plate alignment. The aperture ratio of this cell
structure is not determined by the BM opening at the color filter substrate, but by the BM-on-Array,
which can be formed with a very high positioning accuracy.
Even when the greatest care is taken and sophisticated quality-management procedures are applied,
it is not possible to make the TFT-array fabrication process so perfect that it produces only
completely defect-free arrays.
To improve the production yield in the fabrication process, redundancy design, repairable design,
and fault-tolerant designs are often used. Dual-bus-line design or double-metal structure can help
recover from problems of line breakage. Dummy-repair-line design can save the defective panel from
data-bus-line open failures. While these redundant-design techniques can effectively improve
fabrication yield, in some cases they can also reduce the aperture ratio.
The TFT-array must be protected from electrostatic discharge (ESD), which can be generated in the
fabrication processes such as during the rubbing of the alignment layer and spin-drying. Design
approaches for protecting the TFT-array against ESD include bus-line shorting and ESD protection
circuits.
Temperature / Humidity
Environmental Consideration
It is recommended to use the product in a clean place and to exercise caution to ensure it is
not affected by dust or liquids, etc.
Handing
Usage
PREVIOUS MAIN
TFT Device Design Learn About TFT LCD