Pictures without a pause

Researchers at Cornell University have developed an X-ray camera capable of capturing a succession of images that would normally be hidden to optical cameras.

The first experiment using the camera – dubbed the Cornell Pixel Array Detector (PAD) – has reportedly captured a moving image of shock waves from diesel fuel as it emerges at supersonic speeds from an automobile engine fuel injector.

The X-ray imaging was able to penetrate the fog of aerosol droplets formed by the fuel as it cycles through the injector within a thousandth of a second. In a series of images, the camera depicted the shock wave created by the fuel; a phenomenon never before observed or measured, according to the camera’s principal developer, Sol Gruner, professor of physics at Cornell.

The development of the camera is said to take advantage of recent advances in semiconductor technology to build a bonded integrated circuit that incorporates a silicon layer, which converts X-rays into electrical signals, and a second layer of electronics.

The chip used by the camera to capture the diesel fuel shock wave had a density of 9,200 pixels. The X-ray camera will be used in its final form at Cornell and at the DOE’s Argonne laboratory for a number of major experiments never before possible. ‘The applications we are targeting are unique,’ said Gruner. They include the X-ray imaging of fractures in materials at the instant they break.

The complexity of the task facing the development of the PAD was finding a way to capture a sequence of X-ray images without any pause. The imaging devices widely used in research, silicon electronic light sensors called charge-coupled devices (CCDs), capture images well enough, but each pixel first has to release its electronic signals before capturing another image, requiring a pause of about one-tenth of a second between exposures.

Instead, Gruner’s group has built a detector on which the processing electronics, instead of being on the chip, are built into every pixel, even though each pixel is no more than 150 microns across.

The X-ray signal enters the pixel, goes into an amplifier where it is turned into an electrical signal and is shuttled to a storage area. Thus each pixel is able to capture a ‘slice of time,’ store it and then capture the next ‘slice.’ In this way each pixel, operating in parallel, can capture eight images very rapidly before releasing its signals.

For the shock wave experiment, eight X-ray images were taken, stored, downloaded and then eight more were captured. In this way a composite, moving image of the shock wave, a sum of many images, was able to capture the explosive release of the fuel.

The PAD chips are said to require ‘solder bump-bonding’ of two layers of silicon: the first, a specialised layer containing the X-ray conversion pixels; the second, the electronics processing layer. The bonding joins the two layers by laying down ‘solder bumps,’ each one joining a pixel to the processing layer. In the chip used for the shock wave experiment, that meant laying down 9,200 solder bumps.

The latest chips designed by Gruner’s group contain almost 45,000 pixels. Experiments are under way to bump-bond four chips together to build larger area detectors with as many as 726,000 pixels.