Study On Extraction Method Of Density Of States In Metal Oxide TFT

Metal oxide thin-film transistors(TFTs) have become increasingly important for active matrix liquid crystal display(AMLCD) and active matrix organic light emitting diode due to their high mobility, good uniformity, transparency, flexibility, and low temperature process compatible with plastic substrate. The density of states(DOS), as an important parameter, determines the electrical characteristics and the long-term reliability in metal oxide TFTs. Therefore, there is an urgent demand to propose a reliable method of extracting the DOS for metal oxide TFTs, which may contribute to understanding the operation mechanism and optimizing the TFT performance by improving the fabrication processes. This may help promote and develop the flat panel display(FPD) industries.This paper aims at understanding the electronic structures of oxide from the perspective of solid-state physics, designing a reliable method of extracting the DOS for metal oxide TFTs based on theory of semiconductor device and analyzing the influences of the DOS on che characteristics of metal oxide TFTs.The subgap trap states of metal oxde are mainly composed of oxygen vacancy deep states, oxygen vacancy shallow donor states, and the tail states of valence band(VB) and conduction band(CB). At present, the metal oxide TFTs are targeted for n-channel TFTs, because the existence of deep, fully-occupied states are screened by the high Fermi level(EF). As a result, the trap states in the vicinity of EC are the most important for on-operation of n-channel TFTs. The proposed method of extracting the DOS of metal oxide TFTs in this paper aims at the trap states in this region. Based on the Poison’s equation and Gauss’ s theorem, the relationship between the induced charge concentration and the gate capacitance is derived, and then an analytical expression for calculating the DOS can be deduced. So we propose this extraction method of DOS for metal oxide TFTs by only using low-frequency capacitance- voltage(C-V) characteristics. The extracted results of metal oxide TFTs indicate that the DOS can be represented by superposition of exponential deep states and exponential tail states. And the density of tail states of metal oxide is 2-3 orders of magnitude smaller than that of a-Si: H TFTs in conduction band tail, which is the reason why metal oxide TFTs have better subthreshold performance. In addition, we found our extracted results are consistent with those from the temperature-dependent field-effect(TDFE) characteristics technique and the differential ideality factor technique(DIFT). And we perform the device simulation incorporating the extracted results by a 2D device simulator ATLAS(Silvaco). It is shown that the the simulation results exhibit good agreement with the experimental transfer and output characteristics of metal oxide TFTs, which verified our proposed extraction method.Based on the extracted DOS, we also calculate the surface potential and establish a physics-based threshold model for metal oxide TFTs. The weak accumulation region and the strong accumulation region are dominated by the tail states carriers and free carriers, respectively, in the channel. In the application of the Lambert W function, the surface potential for amorphous oxide semiconductor thin-film transistors(AOS TFTs) under the subthreshold region is approximated by an asymptotic equation only considering the tail states, while the surface potential under the above-threshold region is approximated by another asymptotic equation only considering the free carriers. And the intersection point between these two asymptotic equations represents the transition from the weak accumulation to the strong accumulation. As a result, the gate volage corresponding to the intersection point is defined as threshold voltage of AOS TFTs. Furthermore, an analytical expression for the threshold voltage is derived from this noval definition. It is verified that the caculated result of surface potential is consistent with that derived from the Cg-Vgs characteristics of AOS TFTs. And the threshold voltage achieved by the proposed physics-based model is agreeable with that extracted by the conventional linear extrapolation method. It is also shown that the free charge per unit area in the channel starts increasing sharply from the threshold voltage point, where the concentration of the free carriers is several times that of the localized carriers. So we think this is a reasonal definition of threshold voltage for metal oxide TFTs.In conclusion, we proposed an extraction method of DOS in metal oxide TFTs using low-frequency C-V characteristics, which only needs small amount of data and can be calculated analytically. We also established a threshold voltage model for metal oxide TFTs. This model is physically meaningful and mathematically convenient. The calculated threshold voltage is consistent with that extracted from the linear extrapolation method, which not only verify our threshold voltage model, but also validate the accuracy of our DOS extraction method for metal oxide TFTs.