Experiment & Discussion

In order to demonstrate an AM-OLED display with the proposed AC driving scheme, we have fabricated a 2.2 inch panel with a resolution of 176 x RGB x 220, by a top-gate poly-Si process. A buffer and an a-Si layer were deposited by PECVD.

Next a XeCl excimer laser was used to crystallize the a-Si layer. After definition of the active island and deposition of the gate insulator, the gate metal was sputtered and patterned. The n-channel TFT S/D and LDD and the p-channel TFT S/D were then doped. Finally, the TFTs were formed after dopant activation, interlayer dielectric deposition, hydrogenation, contact via formation and metallization. Once the TFT process is completed, a hole injection material PEDOT:PSS and a green light-emitting copolymer were spin-coated sequentially onto the ITO anode. Finally, a Ca/Al bi-layer cathode was thermally evaporated through a shadow mask to form the common cathode.

The pixel size is 66 x 198 µm² with an aperture ratio of 25.3%. A 2 mm wide power rail surrounds the active area, and each column of the VDD electrode is connected to this power rail, as shown in Fig. 4-4. The active area was divided into 5 x 5 regions and the brightness of each region was measured using a GmbH Conoscope.

The diameter of the measuring spot size, which covers about 80 pixels, was 2 mm.

The brightness of the top left region (A5) was set to 730 cd/m² as a reference in the

measurement.

Fig. 4-4. The active area of AM-OLED display panel is divided into 5 x 5 regions for brightness measurement and the photograph of Conoscope.

The conventional DC driving scheme, where the parasitic resistance of addressing electrode can induce a voltage drop, shows inferior brightness uniformity to the AC driving scheme. The driving current IOLED of each pixel was calculated by dividing the current measured at the cathode by the number of pixels. Although IOLED

is only 1.6 µA for the DC driving scheme, the voltage drop still causes a significant brightness decrease from the surrounding to the central regions. The lowest brightness was at region (C2), which was found to be only 74.54% of that at the reference region A5, as depicted in Fig. 4-5(a). Since the voltage drop ∆VDD at VDD electrode cannot be measured directly in display operation, ∆VDD can be evaluated approximately by the measured data and the I-V equation of TFT as shown below.

( )

²

(

^'

)

²

% 5 .

74 ⋅

β

⋅ V_DATA −V_DD −V_TH =

β

⋅ V_DATA −V_DD −V_TH

Eq. 4-3

where _{FE OX}

DV

DV C

L

W ⋅ ⋅

=

µ

β

2

The parameters for DC driving scheme are VDATA=8.9V, VDD=12V and VTH=-1V so that the maximum ∆VDD of 0.287V can be calculated.

1

Fig. 4-5. Normalized brightness of AM-OLED display panel for the DC and AC driving schemes. (a) DC driving scheme with 100% duty cycle, I

OLED

=1.6 uA. (b) AC driving scheme with 80% duty cycle, I

OLED

=1.99 uA. (c) AC driving scheme with 40% duty cycle, I

OLED

=3.9 uA. (d) AC driving scheme with 20% duty cycle, I

OLED

=7.6 uA.

In contrast, the AC driving scheme shows effective compensation for brightness uniformity variation. In Fig. 4-5(b), the normalized brightness of all measured regions were well above 91.6%, for a duty cycle of the flashing period of 80% and an IOLED of

1.99 µA. The RC time constant of each data line, which is a critical issue for reducing the programming period, is 100 ns. Even though the programming period is reduced to 20% of the entire frame time, the data voltages still can be programmed accurately into the pixels. When the duty cycle of the flashing period decreases to 40%, so as to possess more programming time, the driving current is increased to 3.9 µA to keep the reference brightness at 730 cd/m². Meanwhile, the AC driving scheme is still capable of maintaining the brightness uniformity higher than 92.4%, Fig. 4-5(c), even though the higher driving current can lead to a significant voltage drop. Although a higher driving current is needed in the AC driving scheme, the treatment of reversed bias voltage can accelerate the recovery from degradation and lead to an improvement in the J-V characteristics and device lifetime of the OLEDs ^[3][4]. In other words, the higher driving current may degrade OLED performance rapidly, however, the AC driving scheme, with proper reversed bias voltage, can alleviate the degradation. Fig.

4-6 shows the normalized brightness at region C2 versus the duty cycle in which the

duty cycle of 100% represents the DC driving scheme. The experimental results show that the normalized brightness is higher than 91.6% when the AM-OLED panel operates in the AC driving scheme with various flashing duty cycles. In contrast, once the panel is driven by the DC driving scheme, the brightness drastically decreases to 74.5%. Nonetheless, the parasitic capacitance of TSW between the storage node and the gate of TSW causes the voltage drop at the storage node to be reduced so that

∆VDATA is smaller than ∆VDD. Besides, the thickness of spin-coated polymer film has a slight variation from the central to the outer areas. Consequently, the brightness uniformity cannot be compensated completely and a remaining nonuniformity of about 8% can be observed.

The AC driving scheme changes the voltage at cathode dynamically and will induce additional power consumption. The power consumption for the DC driving scheme are 2.3W and the power consumption for the AC driving scheme with varied flashing periods is 2.33W, as shown in Fig. 4-6.

20 40 60 80 100

70 75 80 85 90 95 100

1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0

I_OLED=7.6uA 3.9uA 2.63uA 1.99uA

Normalized brightness at C2 (%)

Duty cycle of flashing period(%)

1.6uA

Power consumption (W)

Fig. 4-6. Normalized brightness at region C2 as a function of flashing duty cycle.

Although an additional transient power is generated by the AC driving scheme, the total power consumption just increases 40mW. Since the OLED consumes static power of 2.3W during operation, the transient power induced by charging and discharging the capacitance of cathode only accounts for 1.74% increase of the entire power consumption. Therefore, the AC driving scheme will not significantly increase the power consumption even though the higher driving current is required. Two demo photographs of AM-OLED with conventional DC and proposed AC voltage driving schemes are shown in Fig. 4-7. The obvious luminance gradient can be observed in AM-OLED with DC driving scheme. The bottom area is close to the input pads so that the influence of resistance of addressing wire is uncritical. However, when the

driving currents propagate to a far away pixels in upper area, the large parasitic resistance can generate significant voltage drop and results in an luminance decay. In contrary, the AC driving scheme substituting for DC driving can effectively compensate the resistance induced voltage drop and is superior to improve the luminance uniformity, as shown in Fig. 4-7(b).

(a) (b)

Fig. 4-7. The color photographs of voltage-driven AM-OLED deposited with green copolymer. (a) Conventional DC and (b) proposed AC driving schemes.

The AC driving scheme divides the whole frame time into a programming and a flashing periods hence comparatively decreases the scan time TSCAN of each scan line.

Display size and resolution that the AC driving scheme is applicable are evaluated and shown in Fig. 4-8. The RC delay of a scan line can be calculated by using the product the total scan line resistance and the total scan line capacitance to ground. The product of scan line RC is a conservative estimate of the RC delay compared to the actual distributed RC delay.

0 500 1000 1500 2000 2500

Fig. 4-8. The RC delay and the required T

SCAN

as a function of the number of scan lines.

The solid lines shown in Fig. 4-8 denote the RC delay as a function of the number of scan lines for a range of display diagonals from 1.8” to 30”. The dash lines are the required TSCAN calculated with corresponding number of scan lines and varies flashing duty cycle. It is explicit that increasing the number of scan lines and reducing the duty cycle will result in the lower TSCAN. In other words, the AC driving scheme may affect the data programming accuracy when the display resolution and size increase. From small (1.8”) to moderate (12.1”) display size, the RC delay is of less than about 1 µsec, and that a high resolution display with AC driving scheme shows good capability to improve the brightness uniformity. However, as the display size becomes larger than 20”, the influence of RC delay is inevitable. For example, the duty cycle cannot lower than 60% for a 20” display with QUXGA resolution (3200 x 2400), otherwise, the RC delay can cause the data programming error. Consequently, the most suitable display resolution, size, and duty cycle that the AC driving scheme is used for can be evaluated according to Fig. 4-8.

4.4 Summary

An effective AC driving scheme and the corresponding pixel circuit were designed to improve the brightness uniformity of AM-OLED. By means of an alternating polarity of the cathode voltage and the corresponding pixel circuit, the voltage drop at the VDD electrode caused by the intrinsic parasitic resistance of the addressing wire can be compensated. The experimental results demonstrate that the proposed AC driving scheme can achieve the uniform brightness of higher than 91.6%

with various flashing duty cycle. Furthermore, this AC driving scheme shows exceptional competence for the high resolution displays from small to moderate diagonal sizes.

Reference

1. Y. He, R. Hattori, and J. Kanicki, “Current-Source a–Si:H Thin-Film Transistor Circuit for Active-Matrix Organic Light-Emitting Displays,” IEEE Electron Device Lett., vol. 21, no. 12, pp. 590-592, Dec. 2000.

2. Y. He, R. Hattori, and J. Kanicki, “Improved A-Si:H TFT pixel electrode circuits for active-matrix organic light emitting displays,” IEEE Tran. Electron Device, vol. 48, no. 7, pp.

1322-1325, Jul. 2001.

3. D. Zou, M. Yahiro and T. Tsutsui, “Improvement of Current-Voltage Characteristics in Organic Light Emitting Diodes by Application of Reversed-Bias Voltage,” Jpn. J. Appl. Phys., vol. 37, pp. L1405-L1408, 1998.

4. X. Liu, W. Li, J. Yu, J. Peng, Y. Zhao, G. Sun, X. Zhao, Y. Yu and G. Zhong, “Effect of Duty Ratio of Driving Voltage on the Forming Process in Aging of Organic Electroluminescent Device,” Jpn. J. Appl. Phys. vol.37, pp. 6633-6635, 1998.

Chapter 5

Current Driven AMOLED with Fully Integrated Driver

5.1 Introduction

Driving OLED uniformly with TFT is more challenging than driving liquid crystal due to OLED current-dependent luminance. Since the threshold voltage variation and mobility degradation become serious problems for the 2-T pixel with analog voltage driving scheme and result in luminance non-uniformity, digital driving schemes have been proposed to improve the luminance uniformity ^[1][2]. In the digital driving scheme, the switching TFT samples a digital voltage signal and the driving TFT controls the OLED in digital mode, i.e. OLED is either on or off. Grayscales are generated by either Area Ratio Gray scale (ARG) or Time Ratio Gray scale (TRG) schemes or by the combination of both. In ARG scheme, the pixel circuit consists of plural sub-pixels and the number of sub-pixel is corresponding to the required bits of gray scale, as shown in Fig. 5-1. The gray scale is acquired by selecting the number of the on-state sub-pixels that is the ratio of the light emitting portion. In TRG scheme, pulse width modulated is used instead of sub-pixels. Similar to ARG, the frame time in TRG is divided into plural sub-frames which is weighted by binary.

The digital voltage signal with large swing can effectively eliminate the influence of threshold voltage variation and maintain the luminance uniformity of the pixels. However, both the ARG and TRG divide either the pixel area or the frame time into several sub-pixels and sub-frames associated with the gray scales. Thus, not only the number of electrodes becomes larger but also the operation frequency is much

higher than the conventional analog driving AM-OLED panel. In other words, the peripheral driver and driving scheme are too complicated to be implemented.

Therefore the digital driving schemes are only suitable for low resolution and low gray scale applications.

Scan line Data line

1-1

Data line 1-2

Data line 1-3

Data line 1-4 Sub-pixel

pixel

(a) (b)

Fig. 5-1. Digital driving schemes: (a) area ratio gray scale and (b) time ratio gray scale.

Several VTH-compensation AM-OLED pixel electrode circuits using voltage signals have been proposed to be compatible with conventional AM-LCD driving system [ 3 ][ 4 ][ 5 ]. In addressing period, more than two operations are performed sequentially to address each pixel to couple the data voltage onto the capacitor in which the VTH of driving TFT has been memorized. In addition to the original data and scan lines, more electrodes and higher operation speed are needed to ensure the correct pixel functions, consequently increasing the complexity of the driving system.

As the panel size and resolution increase, the programming time will not be sufficient to accurately set the threshold voltage and write the data voltage. Furthermore, the voltage-type VTH-compensation pixel electrode circuits are incapable of compensating mobility degradation that may occur sooner or later according to the different value of

OLED current.

In order to ensure the luminance uniformity, a current driving schemes has been proposed ^[6][7]. In the current driving scheme, the design of pixel electrode circuit is based on the switch current (SI) memory cell which must be driven by external current signal instead of the conventional voltage signal. Since the OLED is a current driven device, the pixel electrode circuit should not only ensure the matching of OLED current between addressing and non-addressing phase, but also compensate the variations in OLED and TFT characteristics e.g. threshold voltage variations and aging effects.

In previous works, the pixel electrode circuits are fabricated by a-Si:H TFT technology because of the low-cost fabrication steps and compatibility with AM-LCD industry ^[6][7]. Since the driving TFT should provide a continuous current over a large portion of the frame time to efficiently drive the OLED for desired luminance levels.

The pixel area limits the number of TFTs and their geometric widths, which are directly proportional to TFT transconductance. As a result, the a-Si:H TFT technology is limited due to the low transconductance as discussed in Chap. 1. The mobility µ of poly-Si TFT can be of one to two orders of magnitude higher than that of a-Si:H TFT.

As a consequence, poly-Si TFT widths can be smaller, with possibilities of allowing for more TFTs in the pixel area for additional error correction. In addition, while a-Si:H TFT is able to produce steady-state currents suitable for reasonable brightness levels, the terminal voltages at each electrode are fairly large. Due to the high terminal voltages resulting from low mobility the two transistors in series between the power supply and the OLED in the pixel circuit significantly increase the power dissipation

[6]. As a result, the a-Si:H TFT may not be suitable for large-sized displays having high luminance. Besides, due to high gate-to-source and gate-to-drain overlap

capacitances, the ON/OFF switching of a-Si:H TFT can create large voltage offsets and affect the accuracy of stored signal in CST. Thus, poly-Si TFT possessing low gate overlap capacitances is more attractive than a-Si:H TFT. The lack of p-channel TFT in a-Si:H TFT technology decreases the design flexibility so that more control signals are needed for the current-driven pixel electrode circuit. Furthermore, unlike poly-Si technology where the current data drivers can be designed and integrated into the display panel, a-Si backplanes must incorporate c-Si data drivers but now the commercial current-type drivers are not available. A current-type driver for various display sizes, formats, pixel designs, and brightness levels are more complex and costly than conventional voltage-type drivers.

5.2 Design of current-driven pixel with poly-Si TFT

With high mobility, high transconductance and implementation capability of p-channel device, poly-Si TFT technology is beneficial for current-driven AM-OLED to compensate the threshold voltage and mobility variations due to non-uniform grain size and time-related electrical stress. The schematic of pixel electrode circuit based on SI memory as well as the timing diagram are shown in Fig. 5-2. The operation of the pixel electrode circuit operates can be defined into programming and reproduction periods. In the programming period, the scan line signal is HIGH to turn on T1 and T2, then the input data current IDATA flows through the transistors T1, T4, and OLED to the ground. Since the drain and gate electrodes of T4 are connected together by T2, the gate voltage of T4, VG4, is automatically charged to a suitable value for the flow of this current and then stored in storage capacitor CST, as shown from Eq. 5-1 to Eq. 5-3 where µFE, COX, W and L are field-effect mobility, capacitance of gate oxide per meter square, width and length of TFT.

(a) (b)

Fig. 5-2. (a) Schematic diagram of 4-T current-driven pixel electrode circuit. (b) Timing chart of pixel operation. ( (1) programming and (2) reproduction period )

( )

L C W V

V

I_{DATA 4 GS4 TH} ² _{4 FE OX} ⁴ 2

,

β

1

µ

β

− =

=

Eq. 5-1

Therefore the overdrive voltage is

4

4 _{TH DATA}

β

GS V I

V − =

Eq. 5-2

Since the source voltage of T4 is equivalent to the OLED voltage VOLED, the T4 gate voltage is

OLED TH

DATA

G I V V

V ₄ =

β

₄ + +

Eq. 5-3

The voltage VG4 is variable and can be different from pixel to pixel and from time to time, according to the desired current flow and the magnitude of the VTH shifts of both OLED and T4 of the addressed pixel. If an OLED and/or T4 threshold voltage variation appears, VG4 will always be set to ensure the desired OLED current corresponding the precise data current supplied by external driver. In this manner, the OLED current can be maintained to the set-value, no matter how large the VTH shift is for the driving TFT and the OLED.

When the pixel electrode circuit is deselected and the scan line signal is LOW,

both the T1 and T2 are OFF and the pixel operates in reproduction period. At the same time, the T3 is ON due to the opposite polarity of the charge carriers, allowing the current flowing from T3 to T4. Because the gate voltage tracks the threshold voltage of T4, the effect of VTH variation is practically cancelled in this circuit and the constant current ensures a minimum variation of the gray scales. Consequently the threshold voltage shifts of TFTs in this circuit will not have a major impact on the output current and display luminance.

An important issue of this current-driven pixel is a mismatch between the input diving current and output OLED current. Ideally the relationship between these two currents should be linear, however, there are a number of factors resulted in non-linearity in this transfer characteristic and limit the useful current range and thereby also the gray scales. It should be noted that transistor T4 must operate in the saturation mode either during the programming or reproduction periods. In other words, nonlinearity between the IDATA and IOLED will appear as the T4 operates in the undesirable non-linear mode. This is understandable because only in the saturation mode, the transistor can act as a current source in which the current is only dependent on the gate-to-source voltage. In practical, the T1 and T3 are not ideal switches, hence the IDATA and IOLED passing them can generate voltage drops that drive T4 to operate in linear mode. In programming period, a voltage drop between drain and source electrodes of T1 results from the IDATA. The T1 source voltage TS1 equivalent to T4 drain voltage TD4 is smaller than T1 drain voltage TD1. Because no current will pass through T2, the turn-on resistance is like an ideal switch, the T4 gate voltage VG4 is equal to T1 drain voltage VD1. Based on above reasons, the VD4 can be a bit lower than VG4 and T4 may probably work in linear region, consequently, resulting in output current deviation. In order the ensure T4 working in saturation region in programming

period, the overdrive voltage should be defined as following.

The turn-on voltage of scan line is usually set to Vdd which is high enough to make sure the T1 and T3 work in deep linear region. The data current passing through T1 is

( )

Taking account of Eq. 5-5, the non-ideal effect of T1 can be derived as following equation.

It is evident that during the programming period, the size of T1 should be deliberated designed based on Eq. 5-9 to prevent the T4 operation in linear region.

In reproduction period, the non-ideal switch T3 possessing a certain quantity of turn-on resistance can result in a voltage drop as an IOLED passing through it. The

In reproduction period, the non-ideal switch T3 possessing a certain quantity of turn-on resistance can result in a voltage drop as an IOLED passing through it. The

在文檔中 多晶矽薄膜電晶體於主動矩陣式有機發光二極體顯示器之應用 (頁 78-0)