Digital flat-panel detector with avalanche gain
Matthew Wronski, Alla Reznik, Giovanni DeCrescenzo, Justin Bimbrahw
Development of a digital flat-panel detector with avalanche gain for low-dose radiography and fluoroscopy
Introduction
Some medical procedures such as cardiac catheterization and angiography are routinely performed using X-ray fluoroscopy. Semiconductor-based digital flat-panel detectors (FPD) are increasingly being used in these procedures; they are replacing traditional X-ray image intensifiers due to their high spatial resolution, compact size and distortion-free imaging. Unfortunately, current state-of-the-art FPD systems suffer from the presence of substantial noise in the readout electronics, particularly in the low clinical X-ray exposure region (0.1 – 1 mR/frame) and as such are not quantum noise-limited. [1,2]
Although significant reductions in electronic noise are unlikely, the noise may be overcome by adding a gain stage, which amplifies the weak image signal before it is read out. Toward this end, certain researchers are developing active pixel readout circuits that incorporate an amplifier at each pixel of the FPD.[3] Such systems, however, are typically difficult to implement in conventional thin film transistor (TFT) manufacturing processes, take up more pixel area and are prone to radiation damage. Alternatively, high-gain photoconductors such as PbI2 or HgI2 may be used [4], but large area, defect-free deposition of these materials is difficult and they suffer from limited charge range.
Our approach consists of using an amorphous selenium (a-Se) photoconductor, which is a well-characterized X-ray image receptor, as a very high sensitivity imager. Biased at a sufficiently high electric field, photo-generated charge in the a-Se undergoes avalanche multiplication and provides the gain required to overcome electronic noise in low-exposure radiography and fluoroscopy applications.
Device operation and experimental methods
The detector structure under investigation is made up of a high-resolution scintillator (structured CsI phosphor), which converts X-rays into light photons that are in turn absorbed in the a-Se photoconductor and generate electron-hole pairs. The a-Se is 15 mm thick and is biased at an electric field in the range 10–120 V/mm using a high voltage power supply. At fields exceeding the avalanche threshold (75 V/mm), photo-generated holes in the a-Se undergo impact ionization and produce additional electron-hole pairs in an avalanche process. Blocking contacts on either end of the a-Se layer limit hole and electron injection from electrodes into the a-Se, thus reducing current leakage through the detector (or dark current). At the output, a TFT array and readout electronics convert the collected photo-generated charge at each pixel into a digital signal, which constitutes the final image.
Here, we characterize the performance of the novel a-Se device (excluding the scintillator and TFT layers). The device is subjected to pulses of light, which mimic the output of a CsI scintillator exposed to very low intensities of X-ray radiation within the clinically relevant exposure region (0.1-1 mR/frame). By varying the high voltage bias applied to the device in the range of 150–1800 V/mm (corresponding to electric fields in the 10–120 V/mm range in the a-Se layer), and measuring the amount of charge generated in the a-Se using an oscilloscope, we obtain the avalanche gain characteristic of the device. The dark current is also measured using an electrometer. Temporal characteristics of the a-Se device are assessed using the time-of-flight (TOF) method, which involves exposing the device to short (1 ns) laser pulses from a 337 nm nitrogen laser and measuring the resulting electric current pulse.
Results and discussion
Our measurements indicate that the a-Se device is capable of maintaining a dark current below 1 nA/mm2, which compares favourably with crystalline silicon-based photoconductors. Gains as high as 104 are possible, which suggests that this device can provide similar sensitivities to photomultiplier tubes.
Avalanche multiplication in a-Se has enabled the development of a high sensitivity broadcasting camera at NHK in Japan with measured avalanche gains as high as 103. In this camera, the photosensitive a-Se layer is enclosed in a vacuum tube and scanned by an electron beam [5]. The a-Se device investigated here is a solid-state alternative, which can provide similar or even larger gains than are possible with the vacuum device. This is an important advantage because it enables the device to be scaled up to as large an area as is necessary for radiographic and fluoroscopic imaging applications, while maintaining the small package size of conventional FPDs.
In previous work, we have determined that an avalanche gain of 20 is required to sustain quantum noise-limited detective quantum efficiency (DQE) at fluoroscopic X-ray exposures in the range 0.1–10 mR/frame. At radiographic exposures in the range 30–3000 mR/frame, a gain of five is enough for optimal operation. Thus, the a-Se device provides much higher gains than what is required for overcoming electronic noise, even at the lowest fluoroscopic and radiographic X-ray exposures. This is encouraging and suggests that the fabrication of robust FPDs with avalanche gain and dark currents below 10 pA/mm2 is feasible. Furthermore, because the avalanche gain is strongly dependent on the applied HV bias, the gain can easily be turned on at low X-ray exposures to maximize sensitivity and turned off at higher exposures to prevent saturation of the detector. This programmable gain feature effectively enables a wide dynamic range, which is crucial for clinical imaging applications in which the exposure at the detector can vary over more than five orders of magnitude.
Our TOF measurements show that the detector response is shorter than 100 ns. Thus, real-time imaging at 30 or 60 frames per second should not present a problem.
These findings are exciting and extend beyond applications in high-sensitivity FPDs: the technology investigated here enables largely scalable and cost-effective solid-state imaging devices for any diagnostic medical application that is concerned with detecting very low amounts of radiation such as positron emission tomography, single photon emission computed tomography or tomosynthesis. There are also numerous other potential applications in other areas including protein crystallography, astronomy, broadcasting and consumer electronics.
Conclusions
We are developing and characterizing a novel solid-state device capable of providing very high avalanche gains and an excellent temporal response. The device, which is based on the amorphous photoconductor a-Se, is scalable (i.e., can be manufactured in large areas), can overcome electronic noise even at the lowest clinical X-ray exposures used in diagnostic imaging and has a low level of dark current. Coupled to a high-resolution X-ray scintillator and TFT array, this device should provide a true solid-state alternative to the X-ray image intensifier that is both robust and cost-effective. This should open the door to dose-efficient flat-panel imaging detectors for radiography and fluoroscopy, as well as a number of other demanding medical imaging applications.
References
- R. B. Benitez, R. Ning and D. Conover, "Comparison measurements of DQE for two flat-panel detectors: Fluoroscopic detector vs. cone beam CT detector," Proc. SPIE 6142, 1 -10 (2006).
- D. C. Hunt, O. Tousignant and J. A. Rowlands, "Evaluation of the imaging properties of an amorphous selenium-based flat-panel detector for digital fluoroscopy,"
