MedAppl. - How a FPD works - Banner
Sensing is Life
As a global leader in optical solutions we offer a unique product and technology portfolio for sensing, illumination and visualization:
from prime-quality light emitters and optical components to micro-modules, light sensors, ICs and related software.
MedAppl. - How an FPD works - standard article

The pixel array 

The sensitive part of the Flat Panel Detector (FPD) is the pixel array. This is a rectangular or square area with dimensions ranging from a few centimeters, for applications such as dental radiography, up to several tens of centimeters, as required by chest imaging, for example.

The array contains thousands of pixels: each is a square the sides of which are tens or hundreds of micrometers long, depending on the required spatial resolution.

Every time an X-ray image is taken, the pixel array receives a very short shot of irradiation. The pixels collect this irradiation and store it until it is read out. Each pixel consists of a photodiode and a switch. The photodiode is the fundamental component that generates electrical charge from the impinging X-rays. It can be implemented using either direct or indirect conversion, as explained here

The other element of each pixel is the switch, which often consists of a Thin-Film Transistor (TFT) – a technology derived from the display industry. Alternatively, switches can be implemented with Indium Gallium Zinc Oxide (IGZO) – the latest development, which offers better performance than TFTs, and which is expected to become more common in FPDs in the coming years.


How to read out a Flat Panel Detector (FPD) 

An FPD contains a grid of control and signal traces. Each control trace connects to the gates of all pixel switches along a row, while each signal trace connects to the output nodes of all switches along a column. In each pixel, the input node of the switch connects to the photodiode. Therefore, the switch allows the photodiode to be connected to its signal trace.

When a gate control line is electrically enabled, all its switches turn on. This connects the photodiodes along that row of pixels to their signal traces. The electrical charge originating from the impinging X-rays can then flow into the readout nodes. Switching on one row at a time, the system reads out all rows one by one. At the end of the process, the FPD has read out the whole pixel array. The pixel array is then ready to acquire another X-ray image.

In order to implement this readout scheme, the FPD features two important Integrated Circuits (IC): ReadOut ICs (ROIC) and gate driver ICs

ReadOut ICs are distributed along a side of the pixel array; for the largest FPDs, they can be on two opposite sides, in order to speed up the readout sequence. Their inputs are the signal traces. Typically, a ROIC is a multi-channel device, serving 256 lines. Each channel features an analog front-end that collects and amplifies the charge from the readout line; an integrated Analog-to-Digital Converter (ADC) then generates a digital output word corresponding to the energy released by the impinging X-rays in the pixel, and ready for further processing by a computer.

Gate driver ICs are located along the side of the pixel array perpendicular to that of the ROICs. They drive the control lines and can therefore enable or disable all the switches along each row of pixels. As TFT switches require relatively large voltages to turn on, gate drivers must be able to handle such high voltages; furthermore, they have to toggle quickly between different voltage levels and to produce low-noise outputs that do not perturb the operation of ROICs.


Performance parameters 

The quality of the final image, and therefore of the medical diagnosis, depends on several parameters of FPDs. The performance of both ROICs and gate drivers can have an impact on most of these parameters. Therefore, it is vital to choose ICs specifically designed and manufactured for high performance. The most important performance parameters are:

  • Sensitivity of the detector to a given energy of the X-rays. This includes the quantum efficiency of the photodiodes and, in the case of indirect conversion, the efficiency of the scintillator
  • Spatial resolution, which depends also on the pixel size. This is a challenge also for the development of ROICs and gate drivers, because the design of their channels needs to be small enough to match the required pitch.
  • Readout time, especially for dynamic applications where a high number of frames (readout of the full pixel matrix) per second is required. The ICs have to be fast enough to cope with these speeds. The time it takes to convert a line during the readout phase is referred to as line time.
  • Image noise. The higher the noise, the lower the capability to discern details and the higher the dose of X-rays that the patient receives in order to produce clear images. Noise added by the ICs must be a small fraction of the intrinsic noise of the detector.
  • Power consumption. This has an impact not only on electricity costs and on the environment, but also on the useful operating time of portable, battery-powered FPDs, not to mention the size and cost of their batteries.

Find out more about the digital X-ray sensor AS5850B