Methods of Device Parameter Extraction - Fabrication and Characterization

Chapter 3 Fabrication and Characterization

3.2 Methods of Device Parameter Extraction

In this section, we will introduce the methods of typical parameter extraction such as the threshold voltage (VTH), subthreshold swing (S.S.), field-effect mobility (FE) from the device characteristics.

Several methods are used to determinate the threshold voltage, VTH , which is the most important parameter of the semiconductor devices. The method to determinate the threshold voltage in this thesis is the constant drain current method, the voltage at a specific normalized drain current (NI_D) is taken as the threshold voltage. This technique is adopted in most studies of TFTs. It can give a threshold voltage close to that obtained by the complex linear extrapolation method. Typically, the specific normalized current NID = ID/(W/L) is defined at 10nA for VD operated in linear region and 100nA for V_D operated in saturation region, to extract the threshold voltage of TFTs in most papers.

The subthreshold swing (S.S)is a significant parameter to describe the control ability of gate bias toward drain current and the efficiency of the switch turning on and off. It is defined as the amount of gate voltage required to increase/decrease drain current by one order of magnitude. It should be independent of drain voltage and gate voltage. However, in reality, the subthreshold swing might increase with drain voltage due to the short-channel effects such as charge sharing, avalanche multiplication, and punch through-like effects. It is also related to the gate voltage due to some undesirable factors such as serial resistance and interface state. In this experiment, the sub-threshold swing is defined as 10pA~10nA of drain current range about three orders of magnitude corresponds to the gate voltage range, and then take the inverse of the slope.

The field-effect mobility (FE) is determined from the transconductance gm at

low drain voltage (linear region). The transfer characteristics of α-IGZO TFTs are similar to those of conventional MOSFETs, ignoring any other non-ideal effect and assuming the electric field in the channel is uniform, so the first order I-V relationship in the bulk Si MOSFETs can be applied to the α-IGZO TFTs, which can be where Cox is the gate oxide capacitance per unit area

W is channel width

Therefore, the field-effect mobility can be obtained by

g_m

The mobility value was taken from Equation(3.2.4) with maximum_FE.

Chapter 4 Result & Discuss

4.1 Water Induce Oxygen Adsorption

α-IGZO is one of the ZnO based material which is very sensitive to environment condition especially oxygen, water and other organic material as mentioned in chapter 2. We already know the oxygen tend to absorb on the α-IGZO. Moreover, there are water (H2O), nitrogen, and oxygen in the air, so the ambient effect for α-IGZO TFT is we would like to discuss in this section.

In order to discuss different ambient effect, we use the unpassivation α-IGZO device which means the back channel is exposure to outside. First, we put the sample into the vacuum chamber and then illuminate the α-IGZO TFT at higher temperature to let the oxygen which is already exist on the α-IGZO desorption. Second, nitrogen, oxygen, oxygen through DI water (water-containing oxygen) was fed into the chamber until the chamber pressure reached 0 (vacuum), 1, 10, 100, 150, 610, 760 torr at 303K as shown in Figure 4-1-1. The electrical characterization was carried out using Agilent 4156C Semiconductor Parameter Analyzer in a vacuum chamber system. Figure 4-1-2 to Figure 4-1-4 is the I-V transfer curve (VD=10V) for nitrogen, oxygen, and water-containing oxygen respectively and Figure 4-1-5 is the threshold voltage shift which is extracted by the method which is already mentioned in the chapter 3 for each ambient condition. From the experiment result, we could figure out that as the partial pressure of nitrogen increased, the threshold voltage nearly unchanged. Because nitrogen is inert gas, it does not react with α-IGZO easily which led to threshold voltage (VTH) nearly unchanged. As to oxygen, the larger positive threshold voltage (V_TH) shift occurred when the partial pressure of oxygen increased, owing to oxygen could capture an electron from the conduction band of

α-IGZO and the positive space charge appeared (depletion region). The positive charge cause an upward band bending which means α-IGZO TFT need larger gate voltage to turn on the transistor followed by the equation (2.3.1). In addition, when water is added to certain oxygen gases, the oxygen may be more easily adsorbed in the water-containing environment. Hence, we proposed a model to explain this effect.

According to chapter 2 mentioned, water has two role for α-IGZO TFT. One role is water could donor an electron to α-IGZO TFT. As we know, the equilibrium constant unchanged under the same temperature. As the electron increased, the equation (2.3.1) would occur easily which cause more oxygen absorbed on α-IGZO.

The other role is water could act as defect for α-IGZO, so it could trap electron from α-IGZO to let the total conductance decreased. Base on the two role of water, the total net threshold voltage (VTH) shift for water-containing oxygen environment is larger than only oxygen.

Figure 4-1-1 The measurement system for different ambient condition.

Figure 4-1-2 The I-V transfer curve under different partial pressure of nitrogen.

Figure 4-1-3 The I-V transfer curve under different partial pressure of oxygen.

Figure 4-1-4 The I-V transfer curve under different partial pressure of water-containing oxygen.

Figure 4-1-5 The threshold voltage shift for different ambient.

4.2 Oxygen Adsorption Would Passivate Trap

α-IGZO is one of the ZnO based material which is very sensitive to environment condition especially oxygen, water and other organic material as mentioned in chapter 2. However the 21% of ambient atmosphere is oxygen, once the α-IGZO thin film transistor (TFT) back channel expose to ambient atmosphere that will suffer huge instability for the transfer curve. Therefore we like to discuss the α-IGZO TFT about the oxygen absorption degree at different temperature and the effect of subgap trap.

The device we use is without any passivation layer on α-IGZO TFT which means the back channel is exposed to atmosphere, Hence, we could discuss the relationship between the α-IGZO TFT and oxygen. The measurement procedure of each temperature as follow: First, we illuminate α-IGZO TFT at higher temperature in the vacuum chamber and the pump is always pumping at the same time to clean some part of the oxygen which is already absorbed on the α-IGZO TFT. The electrical characterization was carried out using Agilent 4156C Semiconductor Parameter Analyzer in a vacuum chamber system. Then, oxygen gas was fed into the chamber until the chamber pressure reached 1, 10, 100 and 760 torr as shown in Figure 4-2-1.

Figures 4-2-2 to Figure 4-2-5 show the V_G-I_D transfer curve of α-IGZO TFT at 273 K, 303 K, 333 K and 363 K respectively under the 0 (vacuum), 1, 10, 100, 760 torr oxygen ambient. Note that, the positive threshold voltage (V_TH) shift which was extracted by the method in chapter 3 is not obvious at the lower temperature while the threshold voltage (V_TH) shift is apparent in the higher temperature as shown in Figure 4-2-6. It means that the oxygen of the atmosphere is easier absorbed on the channel of α-IGZO TFT at higher temperature. This would cause positive threshold voltage shift (VTH) of α-IGZO TFT. Because oxygen could capture an electron from α-IGZO and the positive space charge would appear lead to the upward band bending.

In addition, the subthreshold gate swing from Figures 4-2-2 to Figure 4-2-5 for each transfer VG-ID curve was extracted by the method which is already mentioned in the chapter 3 as shown in Figure 4-2-7. Note that, the subthreshold swing does not show any trend with the partial pressure of oxygen increased at each temperature and variation among the range about 0.07 V/decade. The subthreshold swing can be associated with the density of deep bulk state (𝑁_𝐵𝑆) and interface state at the interface between gate insulator and semiconductor layers (𝑁_𝑆𝑆) by the following formula value of the density of state would be dominated by the maximum value among the subgap, this value doesn’t mean that all the density of state are the same in the α-IGZO subgap energy level. Hence we use another method to extract the density of state of α-IGZO under the condition which is oxygen absorbed on α-IGZO at different temperature.

Reference shows the C-V measurement and numerical calculation method in which Kimura et al clamed to estimate the density of state in the subgap of α-IGZO [4.4]. The C-V transfer curve is carried out using Agilent 4284A which the gate of the α-IGZO TFT is connected to Capacitance Measurement High (CMH) and the source and drain is connected to Capacitance Measurement Low (CML). Figure 4-2-8 shows

the measured results of the C-V characteristic of α-IGZO TFT. Here the characteristics at applied frequency of 40 kHz which is the minimum value that we could get the capacitance value. It is observed that capacitance become small as the gate voltage bias below flat band voltage (VFB) because the electrons are depleted in the α-IGZO film which means only the source and drain overlap capacitance detected.

On the contrary, owing to the electrons are accumulated at the channel of the α-IGZO film, the capacitance become larger (overlap capacitance and channel capacitance) as gate voltage over flat band voltage

The trap densities were extracted using the following algorithm. First, the surface potential (ФS) is calculated from the C-V characteristic. By applying Q = CV to the gate insulator, differentiating it by gate voltage, and transforming it, equations (4.2.2) and (4.2.3) are acquired. Moreover, by integrating equation (4.2.3) of gate voltage, equation (4.2.4) is acquired. Here, C_i is the geometrical capacitance of the gate insulator of channel region, Cg is the measurement capacitance in the C-V characteristic subtracts the source and drain overlap capacitance which means the minimum value of the C-V characteristic and VFB is the flatband voltage, which

Surface potential is a function of gate voltage. Figure 4-2-9 shows the calculated result of the surface potential for α-IGZO TFT form C-V measurement. It is found that surface potentialis saturated at about 1 eV for high VG-VFB which indicates that the electrons are accumulated at the surface of the α-IGZO film and that conduction

band edge approach the Fermi level as shown in Figure 4-2-10. Owing to the electrons are increased rapidly when the gate voltage larger than the threshold voltage, all the incensement of the gate bias above threshold voltage crossed over the gate insulator. Hence, the surface potential is no longer bending.

Next, we want to calculate the charge density in the channel layer by applying boundary condition and Poisson equation. Equation (4.2.5) is the boundary condition which means that the electric displacement is continuous between the vertical directions of gate insulator and semiconductor interface. Here, 𝜕∅ 𝜕 ⁄ is the surface potential gradient in the channel layer, 𝜖 is the dielectric constant of the gate insulator, 𝜖 is the dielectric constant of channel layer, and ti is the thickness of the determine the potential at each position alone the α-IGZO bulk direction. We would assume the charge is nearly the same in any position in the α-IGZO alone the α-IGZO bulk direction and the thickness of the α-IGZO is small so the integrate of equation (4.2.6) of each position is equation (4.2.7). Here, the d is the thickness of the α-IGZO.

Note that, the gradient of potential in the α-IGZO at any position is the same so we could calculate from equation (4.2.5) and (4.2.7) to get the equation (4.2.8) which means the average charge density among the α-IGZO for each gate bias voltage (i.e.

each surface band bending).

Figure 4-2-11 shows the calculation result of the channel charge as a function of gate bias voltage. Note that the charge density ρis saturate about 10^-2 coulomb/cm³ which means the number of electron is 10¹⁷ #/cm³ when the gate voltage is 2V, this value consist with previous study result. On the contrary, there are tail state near the conduction band is about 10¹⁷~10¹⁸ cm^-3 and deep state 10²⁰ cm^-3 which is 2.3 eV below the conduction band [4.5]. However the accumulation electron is only 10¹⁷

#/cm³ which means the relatively holes by gate bias is approximate this amount.

When applied positive gate voltage Fermi level could be pinned near the conduction band owing to not too much electron could not fill this amount of density of state. On the other hand, when applied negative gate voltage Fermi level could be pinned at 2.3 eV below the conduction band due to the deep occupied state which is already full of electrons and it is not easy to exclude them.

The charge density ρcontain localized trapped electron, free electrons… and so on. The localized trapped electron only related to the density of state according to the equation (4.2.9). Here, the N is density of state of α-IGZO subgap. So if differentiating the charge density ρby the potential among the subgap of α-IGZO, the density of state will be obtained as shown in Figure 4-2-12. Such C-V characteristic measurement range is among the subgap of α-IGZO where the interval of the red dash line as shown in Figure 4-2-13 so the density of state which we

measurement by the method which is mention above. Note that, the density of state nearly unchanged when the oxygen absorbed on the α-IGZO under 273K and 303K while the density of state decrease when the oxygen absorbed on the α-IGZO under 333K and 363K. This means at higher temperature not only oxygen absorption easier but also oxygen could passivation the trap among the α-IGZO lead to the density of state which we calculated decreased.

To further confirm the oxygen absorbed on the α-IGZO could compensate trap of α-IGZO. We take two conditions into consideration. One is the α-IGZO TFT at 363 K under vacuum chamber; the other is also at 363 K under vacuum chamber but after the oxygen absorbed on the α-IGZO. The reason for the chamber temperature at 363 K is that according to the previous experiment the oxygen absorbed efficiently consider the electron trapping effect. Here, we do not consider the negative bias side because it is not easy for α-IGZO TFT to inverse hole. From Figure 4-2-22, we observe that the flat band voltage (V_FB) shift successively to more positive voltage for each hysteresis loop during the return sweep. On the other hand, the device which the oxygen is already absorbed has lower flat band voltage (V_FB) shift than the device which is without oxygen absorbed. The flat band voltage (VFB) shift result from three possible cases that electrons which are from α-IGZO TFT channel could be trapped:

(1) in the bulk or back channel of α-IGZO TFT. (2) at the channel and dielectric interface. (3) injected into the dielectric. Due to the same device, the possible reason that could cause flat band voltage shift must not be electrons trapped at the channel

and dielectric interface or injected into the dielectric. Hence, the only one possible reason for the oxygen absorbed device which has lower flat band voltage (VFB) shift under back to back sweep is the trap of α-IGZO decreased. As we sweep on the reverse loop, the Fermi level is from near conduction band to away conduction band.

This would cause some electrons trapped in the α-IGZO trap lead to the higher flat band voltage (VFB) as show in Figure 4-2-23. Base on this model, the lower flat band voltage shift means the lower trap of the oxygen absorbed α-IGZO. Therefore, we could conclude that the oxygen absorbed on the α-IGZO could compensate trap of α-IGZO.

Figure 4-2-1 The schematic of oxygen molecular fed into the measurement.

Figure 4-2-2 The I-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 273K.

Figure 4-2-3 The I-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 303K.

Figure 4-2-4 The I-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 333K.

Figure 4-2-5 The I-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 363K.

Figure 4-2-6 The threshold voltage (V_TH) shift under different oxygen partial pressure at different temperature.

Figure 4-2-7 The subthreshold swing (S.S.) shift under different oxygen partial pressure at different temperature.

Figure 4-2-8 The C-V transfer curve of α-IGZO TFT.

Figure 4-2-9 The surface potential as a function of gate voltage.

Figure 4-2-10 (a) The flat band condition (b) The maximum surface potential as gate voltage larger than threshold voltage (VTH) of α-IGZO.

Figure 4-2-11 The maximum surface potential as a gate voltage larger than threshold voltage.

Figure 4-2-12 The density of state of of α-IGZO TFT.

Figure 4-2-13 The C-V detects range among two red dash lines. [4.8]

Figure 4-2-14 The C-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 273K.

Figure 4-2-15 The C-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 303K.

Figure 4-2-16 The C-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 333K.

Figure 4-2-17 The C-V transfer curve for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 363K.

Figure 4-2-18 The density of state for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 273K.

Figure 4-2-19 The density of state for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 303K.

Figure 4-2-20 The density of state for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 333K.

Figure 4-2-21 The density of state for different partial pressure of the oxygen molecular which absorbed on α-IGZO at 363K.

Figure 4-2-22 The flat band voltage shift for back to back sweep under vacuum and vacuum which is oxygen already absorbed.

Figure 4-2-23 The schematic band diagram of flat band voltage (VFB) shift because of charge trapping in the a-IGZO.

4.3 Negative Illumination Bias Stress Instability (NBIS)

Thin film transistors (TFTs), served as driving devices and switching devices for the active matrix liquid crystal display (AM-LCD) and the active matrix organic light emitting diodes (AM-OLED), have been universal applied. Since most of the proposed uses of the TFT will expose the TFT to a backlight or ambient light during operation, the black matrix (BM) integrated in the pixels is usually utilized, which can obstruct the formation of the photo-leakage current. The aperture ratio is the ratio between the transparent area, excluding a pixel’s wiring area and transistor area (usually hidden by black matrix), and the whole pixel area. As the open area ratio gets higher, the light penetrates more efficiently, which results in the brighter images for the display, with the same magnitude of back-light source. Several technologies of the wiring and element design improved the aperture ratio by reducing the black matrix area, thus increasing the projector’s luminosity. Hence, if the integrated black matrix can be fully removed from the pixel of flat panel display, the highest aperture ratio might be achieved. The photo-leakage current of the α-IGZO TFTs obtained from our experimental results exhibited no significant variation, while a remarkable negative shift of threshold voltage is observed. The reason of this finding has been explained by several models as mentioned in Chapter 2. However, theα-IGZO TFTs, used as switching devices, usually working at negative gate bias side for most of the time to ensure that the transistor is under the “off-state”. For application, if the device did not display at fully off-state condition, it would possibly cause the image persistence or bright dot. Hence, the instability of negative bias under illumination for theα-IGZO TFTs is a emergent issue we need to discuss in this report [4.6].

在文檔中銦鎵摻雜之非晶態氧化鋅薄膜電晶體在環境與照光偏壓下穩定度之研究 (頁 41-0)