A novel flat-panel imager for medical x-rays
The last decade has seen rapid development and clinical adoption of active-matrix flat-panel imagers (AMFPIs) for diagnostic x-ray imaging. AMFPIs provide better image quality than traditional screen films and computed radiography,1,2 but further improvement is desirable, especially in spatial resolution and low-dose performance. Existing AMFPIs use a two-dimensional array of thin-film transistors to read out a charge image generated by an x-ray image sensor. In ‘direct conversion’ this sensor is an x-ray photoconductor, while in ‘indirect conversion’ the sensor is a scintillator coupled with discrete photodiodes. These methods have two main limitations. First, in low-dose applications such as fluoroscopy (real-time x-ray imaging), electronic noise degrades the imaging performance behind dense tissues. Second, the smallest pixel size currently used for digital mammography, 70μm, may not be adequate in some cases. For example, characterizing the shape of microcalcifications has been shown to be compromised with pixel size del=100 μm, while del=50μm can preserve the needed information.3
To improve the detector low-dose performance with high resolution, we are investigating a new, indirect-conversion flat-panel imager with avalanche gain and a field-emitter array (FEA), which is referred to as SAPHIRE (scintillator avalanche photoconductor with high-resolution emitter readout).
The concept of SAPHIRE is shown in Figure 1. It consists of a needle-structured cesium iodide (CsI) scintillator, optically coupled (for example, through fiber optics) to a uniform thin layer of amorphous selenium (a-Se) photoconductor, with thickness dSe ~ 4−25 μm. The selenium layer is operated in avalanche-multiplication mode, and is called HARP (high avalanche rushing amorphous photoconductor).4
The visible photons emitted from the scintillator pass through a transparent indium tin oxide (ITO) bias electrode to generate electron-hole pairs near the top of the HARP layer. Applying a positive voltage to the ITO causes holes to move toward the bottom (free) surface of the HARP. On the way, they experience avalanche multiplication under an electric field strength ESe>90 V/μm,5,6 which is an order of magnitude higher than is typically used in direct-conversion a-Se x-ray detectors.
The holes form an amplified charge image at the bottom surface of the HARP layer, which is read out with electron beams generated by a two-dimensional FEA, which is placed at a short distance, for example, 1–2mm, below the scintillator-HARP structure. The FEA pixels are addressed by orthogonal gate and base lines.
We envision dividing the ITO electrode into N stripes, each connected to a separate charge amplifier, as shown in Figure 2. Using this scheme, N pixels can be turned on simultaneously for parallel readout, increasing the readout speed by a factor of N. The parallel-beam readout also decreases electronic noise by reducing the load capacitance at the input of the amplifiers. Our calculation shows that N=128 is sufficient to ensure real-time readout of 30 frames per second and electronic noise less than 1000 electrons.
We investigated the spatial resolution and noise properties of SAPHIRE through modeling and experiments. By adjusting the potential applied to the ITO bias electrode, the avalanche gain of the HARP layer can be varied to accommodate the wide dynamic range required for x-ray imaging. At low exposures used in fluoroscopy, avalanche multiplication will be activated by applying ESe=100 V/μm. With dSe=15μm, x-ray-quantum-noise-limited performance can be obtained down to 0.1μR per frame, which is the lowest detector exposure encountered in fluoroscopy. The spatial resolution of the FEA readout is determined by the lateral spread of the electron beam emitted by the FEA.7 With proper electron-optical focusing, this spread can be confined to the pixel size, which ranged from 20 to 90μm for the prototype sensors used in our investigation.8,9
Future work will focus on developing engineering methods to maintain good spatial resolution and uniform avalanche gain over a large area SAPHIRE detector.
We gratefully acknowledge financial support from the National Institutes of Health (R01 EB002655) and Army Breast Cancer Research Program (W81XWH0410554).
Wei Zhao is an associate professor of radiology. Her research interest is developing image sensors for medical diagnosis. She has investigated different approaches for making large-area x-ray flat-panel detectors. Her recent research work also includes detector and system optimization for breast tomosynthesis.
