Top-Emission and Bottom-Emission

With the creation of a 13-in. color active-matrix organic electroluminescent (OEL) display, Sony (Tokyo, Japan) has demonstrated its proprietary top emission adaptive current (TAC) drive technology [2-24], which allows the fabrication of high-quality displays measuring over 10 in. The 800 x 600-pixel OEL display is presented as the largest of its type, presaging the use of OEL technology in a wider variety of applications beyond portable and compact devices.

Traditional polysilicon TFT drive systems often induce nonuniformity of luminance between the individual pixels, which makes it difficult to create a large-scale OLED display.

TAC technology expands on the two-transistor driver circuit by using four transistors, of which two are paired to offset pixel variation for uniform luminance over the entire screen. A detail description for circuit is introduced in next chapter.

In addition, unlike the bottom-emission structure of traditional OLED displays where light is partly blocked by the TFT structure as shown in Fig. 2-17(a), the top-emission structure as shown in Fig. 2-17(b) emits light through the opposite side of the display matrix without any interference. In addition to increasing luminance, the pixels may be fabricated smaller to achieve higher resolutions.

(a) Bottom Emission Structure

(b) Top Emission Structure

Fig. 2-17 Bottom and top emission structure

Chapter3

AMOLED Review for Driving Design

3.1 Category

In AM addressing scheme, electronic switches are used with OLED for each pixel providing the means to retain the video information on a storage capacitor during the complete frame time.

An additional active component, known as drive transistor is needed to provide the OLED with drive current at each pixel. The channel material of the TFTs can be Poly-silicon or amorphous-Silicon or even organic molecules. AM OLED pixels turn on and off more than three times faster than the speed of conventional motion picture film – making these displays ideal for fluid, full-motion video.

The AM addressing mechanisms are broadly categorized in two categories

Analog 　

Digital 　

In the Digital driving methods [3-1] [3-2], the pixel circuit drives OLED in digital mode using constant current source i.e. OLED is either on or off. Grayscales are generated using either Area ratio grayscale (ARG) method or Time ratio grayscale (TRG) method or combination of both. In Area ratio grayscale (ARG) generation method, the pixel circuit is repeated for every bit required. In Time ratio grayscale (TRG) generation method, pulse width modulated current source is used. In the analog methods [3-1][3-2], each OLED is driven by controlled current source of varying current depending upon the brightness required. The controlled current source can be programmed by using either a current source or a data voltage source and should provide desired current throughout the frame time corresponding to the gray- level.

we will present the general concept of the AMOLED technique. Transistor technologies including a-Si:H, poly-Si, crystalline Si, and organics will be reviewed with respect to application to AMOLED displays. Then we will introduce some traditional pixel structures and driving schemes.

3.2 Transistor Technologies

polysilicon and amorphous silicon(a-Si) are known as semiconductors of TFTs used in active matrix organic light emitting diode panels. However, only expensive polysilicon backplanes were used for OLED production, because of their greater electron mobility. But researchers have started creating OLED structures on cheaper amorphous silicon substrates.

In fact, IDTech—a joint venture between Chi Mei Optoelectronics and IBM—has demonstrated a 20-inch full-color OLED panel on an amorphous silicon substrate. In next, transistor technologies including a-Si:H, poly-Si, crystalline Si, and organics will be reviewed with respect to application to AMOLED displays.

3.2.1 Poly Silicon

Polysilicon is recently the active matrix technology of choice for direct view AMOLED displays. This is due to the high drive currents and the long term stability of the devices compared to amorphous silicon transistors. The field effect mobility is around 50~400 cm²/Vs and the on-off ratio is about 10⁶ [3-3] [3-4]. It is suitable for pixel as well as driver integration on glass. Sanyo/Kodak demonstrated a stunning display at the SID’00 and Sony announced a new display at SID’01 Symposium. Thin film transistor technologies have large variations in output characteristics from device to device. These non-uniformities are due to the nature of the poly silicon material growth and the drift of the transistor characteristics during operation.

These variations make it difficult to produce uniform current sources across many pixels.

Since the pixel brightness is proportional to the current, any variation in the threshold voltage and mobility of the current drive transistor shows up as a variation in pixel brightness, creating a “salt and pepper” effect across the image. This effect is prevalent in poly silicon AMOLED displays. To avoid the objectionable effect, a lot of ideas are whether the driving method or the circuit attempts to correct for drive transistor variations.

3.2.2 Amorphous Silicon

Hydrogenated amorphous silicon (a-Si:H) TFT is of great interest because of the low cost and the large amorphous silicon manufacturing base[3-5]. Amorphous silicon deposition process tends to generate uniform initial threshold voltage, but the threshold voltage drift significantly over time. Besides, a-Si:H TFT has much smaller field mobility (<1cm²/Vsec) than poly-Si TFT (50~400 cm²/Vsec) [3-3] [3-6]. Therefore, the OLED driving current provided by an a-Si:H circuit will be much smaller than a polysilicon circuit. However, recent progress [3-3] [3-7] [3-8] in OLED efficiency make it possible to achieve a high brightness with much smaller current. For a typical OLED emission efficiency of 5.0 cd/A and a pixel

area of 2.0 x 10⁴ um², a current of 4.0 x 10^-7 A is need to achieved a pixel brightness of 100 cd/m². Table 3-1 shows the comparison of the saturation currents of a-Si:H and poly-Si TFTs.

item

material a-Si:H TFT Poly-Si TFT

Typical Mobility(cm²/Vsec) 0.5 50

TFT W/L Ratio 10:01 10:01

Saturation Current(A) 2×10^-6 2×10^-4

Table 3-1 Comparison of the saturation currents of a-Si:H and poly-Si TFTs

3.2.3 Organic Transistor

A new technology just starting to enter the display area is organic thin-film transistors (OTFTs). The technology [3-9] uses organic films to make up the active regions of the transistor. Displays and circuits can be made on flexible plastic substrates, because the entire fabrication process takes place at temperatures<100 . Using this process, analog and digital℃ circuits can be built on plastic substrate. Mobility on the order of 1 cm²/Vsec has been achieved with low threshold voltages, providing a performance level comparable to amorphous silicon transistors. It is hoped that greater stability will be achieved through refinements in materials and processes, making for a more robust approach.

3.2.4 Crystalline Silicon

Another approach to produce uniform OLED pixels is to implement the active matrix in a crystalline silicon technology where the transistor variations are small enough to produce uniform displays without correction. The process of silicon is limited to wafers, so techniques are suitable for microdisplay application. For example, eMagin is pursuing small area displays on silicon that can be used in head-mounted systems. Besides, due to the nature of higher mobility, crystalline silicon can operate at higher frequency. As shown in Fig. 3-1, it is very easy to integrated scan and data driver on a AMOLED chip if the crystalline silicon is used.

Parallel driving method is also applied to compensate the lower mobility for amorphous silicon and polysilicon, but it will induce higher hardware cost.

Row(Scan Line) Driver

Fig. 3-1 Relation of operation frequency and field effect mobility for different technology

3.3 Analog Driving

Variations in threshold voltage and mobility depending upon implementation may add to luminance variations. Consequently, special pixel driver circuits are designed to overcome such difficulties in order to get thee good image quality. An important distinction between different pixel circuits is whether the data being written to the pixel is current or voltage.

Voltage driving works by applying image signals to pixels as voltage data, which are converted to current data by pixel drive transistors. Converted current data flow to pixel OLED devices, which then emit light. Current driving works by applying image signals to pixels as current data, which are retained by pixel capacitors. Retained current data flows from pixel drive transistors to the OLED devices, which then emit light. In next, the different circuits and driving methods are reviewed.

3.3.1 Voltage Programming Pixel

3.3.1.1 Voltage Programmed with 4T2C Pixel Structure

The problem of variation of TFT threshold voltages was first tackled by Dawson et. al in 1998, using a four-TFT configuration [3-10][3-11]. An improved AMOLED pixel was designed which eliminates the effect of the transistor threshold voltage variation. The schematic for this pixel is shown in Fig. 3-2. The advantage of the four-transistor pixel is that it uses an auto-zero cycle to reference the data against the transistor threshold voltage eliminating the effects of the transistor threshold voltage variation.

Fig. 3-2 Conventional voltage programming pixel with threshold compensation

The functionality of TFT M1 through TFT M4 is identical to that in the current data pixel circuit . However, the driving method is more complex in that TFT M2’s threshold voltage must be established on the capacitors before writing data. With illuminate line low, the scan line is brought low with Vsup (Voltage on supply line) on the data line. Then, AZ is set low turning M3 on. A voltage across capacitor is developed that M2 forces to conduct. The illuminate line is brought high turning M4 off to isolate the circuit from the OLED. With M3 on, a voltage that is proportional to the threshold voltage of M2 is developed across both capacitors. The AZ input is brought high turning M3 off. The data voltage for luminance is the presented on the data lines.

A portion of data voltage is coupled onto the capacitor C2 by capacitor C1. The scan line is brought high followed by a low on the SW input turning M4 on. The OLED current is, to a first

order, proportional to the square of the C2’s coupled data voltage. The step of the above operation is shown in Fig. 3.3.

Fig. 3-3 Timing diagram for 4T2C operation

3.3.1.2 Voltage Programmed with 5T1C Pixel Structure

Fig. 3-2 shows the conventional pixel structure and timing diagram of that structure which can compensate threshold voltage variation and the degradation of supply voltage. But is needs 3 control lines and complex driving signals for data line and control line. Moreover the data line must be alternated to supply voltage level with every row (scan) line time to store a threshold voltage.

A novel pixel structure having 5TFTs, 1 capacitor and 1 control line is shown in Fig. 3-4 [3-12] with a timing diagram for operation. M2, M3, M4 and M5 are switching TFTs, and M1 is driving TFT, which supplies constant current to OLED for a frame time. Vsus produces a constant voltage which is lower than the programmed data to sustain the gate node voltage of M1 for a frame time. The gate node voltage of M1, the right side of CST, increases to

‘VDD-VTH,M1’ by diode connected M1 and the left side of CST is set to programmed data voltage during the current scan line time. After the row line time M4 and M5 are turned on and the potential of both side of CST decrease as a level of ‘VDATA-VSUS. From that time OLED current flow by driving TFT M1, as follow equation (3-1).

Data

)

Fig. 3-4 Novel voltage programming pixel with threshold compensation

It has only one control line, scan line, for pixel operation and can use a whole row line time to program the data. The OLED current is controlled only by the VDATA and VSUS so that it has good immunity against the degradation of panel supply voltage. The step of the operation is shown in Fig. 3-5.

Data Line

Scan Line

Fig. 3-5 Timing diagram for 5T1C operation

3.3.2 Current Programming Pixel

3.3.2.1 Current Programmed with Current-Copy Pixel

To avoid the non-uniformity problem induced by the variation of the threshold voltage and mobility of the TFTs, various pixel structures for current programmed driving scheme have been proposed[3-13] [3-14] [3-15]. Rather than the voltage signal to be programmed in the voltage programmed pixel structures, current signal on the data line is to be programmed into the pixel in these structures. In programming state, the scan line in Fig. 3-6 is activated, and the data line sinks for the specific TFT M2. This TFT is diode-connected by the supplementary switching TFT M3. Meanwhile, Vgs for the specific current through the TFT in saturation region is self-adjusted and stored in the storage capacitor (Cs). In the next state while the scan line is deactivated, often called the reproduction state, the configuration in the pixel is changed by the supplementary switching TFTs so that the stored Vgs on the driving TFT M2 will reproduce the current for the OLED. In this driving scheme, OLED current can be programmed and reproduced precisely, regardless the variation of the threshold voltage or mobility.

Fig. 3-6 Current programmed with current-copy pixel

3.3.2.2 Current Programmed with Current Mirror Pixel

However, though the foregoing current programmed pixel structures seem immune to the process variation at first glance, it takes a vary long time to charge the data line in the programming state due to the current needed is vary small (nA~uA). Even though the current programming method can be applied to achieve excellent image quality, its panel driving speed is too slow to implement high resolution displays.

Other current programmed circuits have been proposed, Fig. 3-7 show the scheme [3-16]

proposed by Sony Corporation. To set the OLED current, the scan line input is brought low while pulling or sinking the data current out of the data line. TFT M1 and TFT M2 are turned on by the scan line. TFT M1 operates as a data switch allowing current to flow from the pixel circuit into the data line. TFT M2 functions as a switch changing operation of TFT M3 to that of a diode, which allows adjustment of the capacitor voltage for matching TFT M3’s drain to source current to Idata. M4 drives the OLED based upon the capacitor voltage. The width of TFT M3 is large than the width of TFT M4. By this way, we can adjust the aspect ratio of the TFT M3, M4 to achieve the current mirror ratio needed. Thus, a larger current will change the data line during the programming state, and the current driving OLED is still small. This allows the data current to be larger in order to set the pixel current within a scan time for low luminance levels.

Fig. 3-7 Current programmed with current mirror pixel

3.4 Digital Driving

A full color display is more and more important in a life application. There are two types of driving method for Fig. 2-11. The analog driving and digital driving methods are well known.

For analog driving, the intensity of each color is controlled by varying the amplitude of voltage applied to the sub-pixel during a picture frame.

The tendency, seems to be in favor of digital driving. For instance, Seiko-Epson use what they call area ration grey scale (ARG) in conjunction with time ration gray scale (TRG) [3-1][3-2].In digital driving, each EL element is connected to a switch. The stored voltage range must exceed the worst-case threshold voltage to operate correctly. By pushing the driving transistor into its linear operating region, the pixel circuit operates in a digital mode, being fully on or off. The method can reduce the threshold voltage sensitivity of display image.

3.4.1 Area Ratio Grayscale

Fig. 3-8 show the ARG method, the number of grayscales is proportional to the total number of sub-pixels of equal size to connect to the supplied voltage. In this driving scheme, the idea is to divide an elementary pixel in sub-pixel and to have driving TFT in each sub-pixel working either in the ‘on’ or the ‘off’ state. In other words, no intermediate voltage values are used on the gate of the driving TFTs. For example, if the elementary pixel is divided into two sub-pixels, four grey levels (2²) can be obtained. However, to use this gray-scale control method is not expected to attain high resolution, since the decrease of a pixel pitch may be difficult. In addition, it is hard to fabricate the TFT-OLEDs with large gray scale level because further increase of the number of sub-pixels may be difficult.

Fig. 3-8 Sub-pixel arrangement in monochrome display using area-ratio technique

3.4.2 Time Ratio Grayscale

This is implemented by dividing each picture frame into sub-frames with serial scan.

During a sub-frame, all pixels are addressed – lit pixels are addressed by a specific voltage and then the display voltage is applied to the entire screen lighting those during the lighting time. As Fig. 3-9 [3-17][3-18], each sub-frame has a weighting ranging from 1 time unit to 32 time units for a typical six sub-frame arrangement. Time Unit depends on size and the number of pixels on the screen. This is a purely digital time-ration control mechanism, which is a key advantage as it eliminates any unnecessary digital to analog conversions, making the OLED technology ideal for the all-digital age. The OLED is lighting for different amounts of time to obtain different brightness level, and then obtain the different gray scales.

Fig. 3-9 A schema for time ratio gray-scale

Chapter4

Design of AMOLED Driver

4.1 Active Matrix Driving

Fig. 4-1 show the schema for the active matrix addressed OLED (AMOLED) , including TFT pixel array which are composed of organic light emitting diodes (OLED), scan driver, data divers and regulator. Pixel could be any structure as shown in Chapter 2 and Chapter 3, the design of driver must be varied according to the pixel structure. If the pixel size and the best luminance are identical, then maximum aperture-ratio [4-1] is required. The aperture-ratio result from the number of TFT. The large TFT dimensions tend to limit the available pixel area for bottom emission. Therefore, to use the minimal number of TFT for AMOLED cell is a goal.

As shown in Fig. 2-11, it owns the least number of TFT in all pixel structure, and has a max. aperture-ratio comparing other pixel structure. The pixel structure by the voltage programmed driving scheme contains one switching transistor (TFT M1), one driving transistor (TFT M2) and one storage capacitor ( Cs ). In this circuit, TFT M2 is a PFET device connected in a common source arrangement, play the role of a current source to supply constant current to the OLED. The Vgs of the driving transistor for a specific output current will be programmed in the capacitor through the switching transistor (TFT M1).

Unfortunately, this structure does not compensate the variation of Vt and mbility, the former two variables would lead to variation in illumination strength.

TFT is always made from both amorphous and ploy-silicon. Poly-silicon has much higher mobility, reducing the size of the drive TFT. The improving mobility of TFT is necessary to make high resolution and high performance AMOLED display which integrate driver circuits on the display panel. 100MHz shift register [4-2] was reported. Poly-silicon is required for system on panel (SOP). For poly-silicon display, both mobility and threshold-voltage vary randomly across the plate from pixel to pixel. As shown in Fig. 4-2 [4-3], this chart shows the theoretical display brightness non-uniformity for a display with threshold-voltage of 3V.

From the chart, we know that variation in Vt are more serious than variation in mobility (Up)

From the chart, we know that variation in Vt are more serious than variation in mobility (Up)