Room: Room 207
Purpose: Imagers capable of photon counting offer several advantages compared to conventional active-matrix, flat-panel imagers used for diagnostic x-ray imaging - such as decreased noise and the ability to acquire spectral information. Photon counting imagers (employing narrow arrays of pixels based on crystalline silicon) are used for mammographic imaging and fan-beam CT. Such arrays are not practical for large-area projection imaging (such as radiography) or dynamic imaging (such as fluoroscopy). For this reason, our group has created the first, prototype photon counting arrays based on polycrystalline silicon (poly-Si) - a thin-film material capable of economic manufacture of monolithic, large-area devices. Empirical evaluations have been conducted to investigate the performance of these circuits.
Methods: The pixels of the prototype arrays consist of a 4-component circuit architecture (an amplifier, a comparator, a clock-generator, and a counter). Initial empirical studies were focused on the amplifier since it is the first component in the signal chain and also one of the most important components. The gain and count rate of the amplifier was determined using test input pulses corresponding to diagnostic x-rays.
Results: Early empirical results indicate that the amplifier (with circuit complexity an order of magnitude greater than any previously reported poly-Si imaging circuits) exhibits operational behavior in line with expectations. For the range of operating conditions examined, the amplifier gain ranged from ~100 to 400 - sufficient to allow correct pixel circuit operation down to x-ray energies of ~20 keV - and count rate was up to 1000 cps.
Conclusion: These empirical results represent the first measurements of poly-Si amplifier circuits designed for photon counting. The results are encouraging and are similar to those obtained from circuit simulations. Further improvement in performance is expected through design modifications guided by simulation and empirical investigations. This research was supported by NIH grant R01-EB000558.