Time-sampling measurements are used in this paper to build time dependent LTPS TFT current model. The device model that considers bias and time dependent threshold voltage (V th ) shift and mobility degradation is implemented in Eldo through GUDM for simulating a pixel circuit as an indicator of panel performance. Author Keywordstime-sampling measurement, LTPS, V th shift, mobility degradation, Eldo, GUDM, transient current. Objective and BackgroundThe performance of a low-temperature poly-silicon (LTPS) thin film transistor (TFT) is usually judged by its threshold voltage (V th ), mobility (), sub-threshold swing (SS) and on-off current ratio (I on /I off ) [1] . In addition to the trap concentration, the activation energy of the traps should be noticed too [11]. Nonetheless, the activation energy of the traps is bias dependent [11]. Therefore, the factors for threshold voltage shift include the gate and drain bias can be expressed as ∆V th = ∆V th (V gs , V ds , Time, Temp). Our goal is to extract the time and bias dependency factor of V th shift from I d -V g and time sampling measure method. The objective of building this model is to enable designers to simulate the image retention time in AMOLED pixel circuit for panel performance estimation. ResultsA P-type LTPS TFT is fabricated with 120nm gate insulator layer and channel dimension of 30μm in length and 3 μm wide. The electrical characteristics were measured in ambient temperature (25℃).When studying the image retention behavior, the chessboard pattern shown in Figure 1 is usually used to observe the image residual phenomenon which is the direct evidence of image retention and the criterion of the panels' quality check. The panel is controlled to display chessboard image for a certain length of time and then switch to pure gray image to observe how long the residual images of previous chessboard image remains. For the pixels, the image change consists of two types of operations, which are white (L255) to gray (L128) and black (L0) to gray (L128). Two panels with different image retention time are selected as the target for examining the effectiveness of the proposed model extraction method. One of them has residual image that lasts 10 seconds, and that of the other lasts 120 seconds. On each panel, one TEG is selected to represent the I-V characteristics of all TFT in the panel. In the following paragraphs, the origins of the residual image will be analyzed and an analytical approach of deriving and extracting the model parameters with transient current will be presented. For an AMOLED panel, the luminance is from the OLED device whose current is controlled by the driving TFT. Therefore, we will focus on the variations of the LTPS TFT because the behavior variations of OLED are within milliseconds and is far from the scale of retention time observed [12]. To mimic the stability of the driving TFT current on a