X-Ray Array Detectors
By Orren Tench
In The Beginning
In the 1980s, Germanium detectors began to show great promise in X-Ray applications. This came about as a result of innovations in detector geometry first suggested privately by John Howes from AEA Harwell and pioneered by Canberra Industries, and by innovations later on window electrodes. The latter work by Canberra radically changed the prevailing view that (because of hot-electron escape) Germanium detectors naturally produced extremely asymmetric peaks below about 3 keV. The Canberra LEGe Detector, with its small rear electrode, demonstrated that germanium detectors provide intrinsically better resolution than do silicon detectors (simply due to statistical considerations- more charge carriers per keV). The notion that germanium detectors naturally produced distorted peaks at low energy was proven erroneous with the advent of the Canberra Ultra-LEGe Detector and its proprietary window electrode.
Canberra was not in a position to take full advantage of these new and exciting technologies because the company was fully focused on nuclear radiation detection and the myriad applications in this field. We had long since decided to focus on nuclear and the once promising X-Ray product line had been abandoned by the early 1970s. We did take the opportunity to license some of the technology to an X-Ray company and we began to supply such detectors to OEMs but our own sales of these detectors remained at a modest level.
In the mid 1980s we were approached by a young scientist, Dr. Stephen Cramer, from Schlumberger, who was doing EXAFS (Extended- X-Ray Absorption Fine Structure) experiments at The Brookhaven National Laboratory. Dr. Cramer convinced us that the EXAFS application desperately needed detectors with multiple elements in order to achieve decent count rates so important with dilute samples. Furthermore he had a budget for a new “Array” Detector and so we proposed to make a nine-element LEGe Detector which we knew would be risky because getting nine Ge X-Ray detectors all working at once in close proximity had never been done at Canberra and probably never elsewhere. Knowing little about the business prospects we gave Dr. Cramer a price, and to our dismay, he told us he had enough money for four additional elements, bringing the total to 13. This is how the 13 Element Ge Array Detector came to be, and despite a measure of “triskaideckaphobia” (fear of the number 13) at Canberra, we built and delivered this detector in about 10 weeks time.
The deployment of this detector caused quite a sensation within the Synchrotron community and soon we had orders for similar detectors from SSRL, Daresbury and many other light sources in the US, Europe, and Japan. Some customers needed low energy operation which our Ultra-LEGe detector provided with or without windowless cryostats. Eventually the original pulsed optical reset preamplifiers gave way to electronic reset preamps (now known at I-TRPs), and the demand for larger arrays led to the production of detectors with as many as 32 elements.
Some users in the Synchrotron Community were not completely satisfied with the largest of the Canberra Discrete Array Detectors demanding even more channels. For two reasons this is somewhat impractical. First of all, getting more than 32 discrete elements all working together in a single array is a technological challenge. Second, the solid angle subtended by larger arrays becomes so large that the outermost elements are marginally effective when the detector is operated in close proximity to the target.
Enter Mirion Lingolsheim
Fortunately the Mirion (formerly EURISYS Mesures) Specialty Detector Facility in Lingolsheim, France, possesses technology which provides a means of segmenting electrodes on germanium detectors. This segmentation technology is the basis for our Monolithic Ge Array Detectors, in which multiple pixels are formed within a single Ge substrate. With pixel sizes of 5mm by 5 mm, 100 channels can be packed within an area of about 50 mm by 50 mm. Larger pixels can also be formed with a commensurate increase of it’s overall area. Equally importantly, the production of monolithic detectors involves only a single slice of germanium, and this makes it practical to produce larger arrays.
In 1998, Dr. Oyanagi, from Spring 8 in Japan ordered the first 100 Pixel Detector from Lingolsheim. Since then Lingolsheim has produced several more 100 Pixel Detectors as well as detectors with fewer Pixels, 36 being a popular size.
As with Discrete Array Detectors performance has been improved over the years. One of the newest innovations involves built-in synchronous reset circuitry, and to make this more useful in some applications, the ability of the user to disable specific channels which incur extremely high flux (and thus would paralyze the system if left in operation). This feature, called SAFE, can be activated automatically by a pre-programmed count rate limit or it can be selected on a channel-by-channel basis by the operator.
The monolithic detectors usually require synchronous reset of preamps and internal collimation to reduce charge sharing and thus improved peak-to-background. The data acquisition rate of the Mirion 100 Pixel Detector cannot be matched by any other detector on the market however.
Mirion’s product offering today contains both monolithic array detectors having 3 up to 100 pixels or discrete array detectors of up to 32 channels.