Projection x-ray imaging is commonly employed to visualize internal human anatomy in order to diagnose medical conditions and facilitate medical procedures. Modern projection imaging is typically performed using an active matrix, flat panel imager that is comprised of a converter layer overlying a monolithic, large-area pixelated array fabricated using a thin-film amorphous silicon process. The images are formed by converting x-ray photons into electrical signals, and then integrating those signals over a frame time – a method referred to as fluence integration.

Recently, imagers employing a different method for creating x-ray images – referred to as photon counting – have been developed and used clinically to perform mammographic imaging (a form of projection imaging). Photon counting involves measuring the energy of each interacting x-ray photon and digitally recording the number of times photons exceed one or more energy thresholds. Because the imaging information is stored digitally, photon counting imagers are less susceptible to noise than fluence-integrating imagers – potentially improving image quality and/or decreasing the amount of radiation required to acquire an image.

Current photon counting imagers are based on crystalline silicon and have limited detection areas. An imager with larger detection area would allow photon counting to be used in other projection imaging modalities – such as radiography (which produces, for example, chest x-ray images) and fluoroscopy (which is used, for example, to non-invasively insert stents and other medical devices). However, strategies to increase detection area, such as tiling multiple arrays, result in increased imager complexity and/or cost. For this reason, our group has been exploring the possibility of creating photon counting arrays using a different semiconductor material, referred to as low-temperature polycrystalline silicon (poly-Si). This material is fabricated using another thin-film process (which also allows the economic manufacture of monolithic, large-area arrays) and has favorable properties for creating complex, high speed circuits.

Using poly-Si, a set of prototype arrays have been designed and fabricated. The pixels of the arrays have a pitch of 1 mm and consist of four components: an amplifier, a comparator, a clock generator, and a counter. Several circuit variations were created for each component, and circuit simulations were performed in order to determine the energy resolution and count rate capability of each variation of each component.

For the amplifier component, all circuit variations were determined to have an energy resolution of ~10% when presented with a 70 keV input x-ray (a typical x‑ray energy used in diagnostic imaging). This degree of energy resolution is comparable to that reported for photon counting imagers fabricated using crystalline silicon. In addition, while count rates for the amplifier component were determined to be roughly one order of magnitude too low for radiographic and fluoroscopic applications (which require rates on the order of 1 million counts per second per square millimeter [cps/mm2]), a hypothetical poly-Si amplifier circuit variation was identified and determined to have count rate capabilities suitable for these applications (with energy resolution similar to that of the prototype designs). In addition, the count rates for the comparator, clock generator, and counter circuit variations were found to range from 100 to 3000 kcps/mm2. Finally, due to on-going improvements in the poly-Si fabrication process (driven largely by the display industry), future photon counting arrays employing poly-Si could have pixel pitches as small as 250 um – a size suitable for radiographic and fluoroscopic imaging.

Date: Wednesday, January 10, 2018
Time: 10:00 AM
Location: 2210 Lurie Engineering Center (LEC)
Chair: Dr. Larry E. Antonuk