This dissertation is primarily concerned with the characterization of the pixel diodes for a flat-panel detector. Photodiode signal with no incident radiation, which is known as leakage current, is an important parameter. Leakage current will reduce the signal capacity available for incident x-ray and could result in additional noise. It is therefore desirable to minimize the magnitude of the leakage current. In addition to the leakage current, the degree to which radiation affects current must necessarily be investigated. In this dissertation leakage current and radiation damage of a-Si:H P-I-N diodes are measured and analyzed.
The new P-I-N diodes, base on the ion shower doping process instead of the conventional PECVD method, are fabricated and characterized. Comparison between ion shower and PECVD diodes is discussed. This method is chosen for two reasons. First, diodes made by conventional PECVD showed poor uniformity of leakage current in a large-area substrate. Among the components of the leakage current, the injection current seems to be the major cause of the poor uniformity. Second, when we fabricate a TFT array only for an Active Matrix Liquid Crystal Display (AMLCD) application, no boron doping is required. However, diode fabrication requires boron doping, which may contaminate the PECVD chamber and affects the TFT deposition process. Since the ion shower doping is performed in another process chamber, it doesn`t result in boron contamination of the PECVD chamber and it may provide high reproducibility of the photodiode leakage current. Also, a reduction of bad pixels is expected.
The experiments are performed to investigate the radiation damage to the leakage current of a-Si:H diode made by PECVD and ion shower doping methods, and to study the effect of a copper plate used in portal imaging. Basically the damage effects seem to be same for two diodes. Leakage current as a function of absorbed dose shows that there exists a threshold dose. However this threshold dose shifts whether a copper plated is used or not. Monte Carlo simulation tells that the copper plate used in portal imaging generates more energetic secondary electrons, which transfer more energy to the diode as they travel through. Transient currents measured at high dose and high negative bias present that the defect creation is higher near p-i junction than bulk i-layer when a copper plate is not used. It is considered that an ion shower diode is weaker than a PECVD diode when radiation damage occurs near interface. Since the p-n junction property is not as good as a PECVD diode, the leakage current ncrease induced by radiation can be higher. However if we used the copper plate as in the portal imaging, the damage to the leakage current of an ion shower diode is identical with a PECVD diode
Finally, a prototype flat-panel detector is constructed to investigate the practical difficulties in making the detector feasible and reliable, and also to study the imaging performance of the detector. The detector consists of two 36 × 21.5 ㎠ half panels which contain 2560 x 1536 pixels, respectively. The pixel pitch is 139 μm and the fill factor is 57 %. The detector uses an ion shower doped P-I-N diodes, and shows good uniformity with 2.5% dark current variation. Its performance are characterized in terms of MTF (Modulation Transfer Function), NPS (Noise Power Spectrum) and DQE (Detective Quantum Efficiency). The presampling MTF is found to be found to be 0.51 and 0.26 at 1 lp/mm and 2 lp/mm. The measured DQE at 1 lp/mm and 2 lp/mm are 0.35 and 0.17.
