Image of a fly taken at the diamond light source. A dead fly was placed in the x-ray and the transmitted x-rays were converted into optical photons using a scintillator and then focused onto an EM-ccd

High Speed EM-CCD camera system

Many X-ray imaging detectors, from medical diagnosis to the examination of energy levels in atoms, are based on the same technology as found in digital cameras. Most X-rays pass straight through these Charge Coupled Devices (CCDs) and CMOS detectors and it is therefore necessary to use a thicker detector (e.g. CMOS hybrid) or a scintillator. However, current methods produce detector systems with limited performance: CMOS-hybrid technology requires intricate features limiting the minimum pixel size and spatial resolution, whilst the readout speed of CCD-based systems is limited by higher noise levels.

Over recent years an STFC funded concept study was completedon a novel photon-counting detector. The Electron-Multiplying (EM) CCD was designed for low light level imaging such as night-time surveillance or night-vision. The EMCCD differs from the standard CCD through the addition of a "gain register". By multiplying the signal by thousands, the effective read noise of the device can be reduced to the sub-electron level, allowing operation at very high speeds. If a scintillator is coupled to an EM-CCD then this low effective noise allows analysis of single photon interactions (photon counting), providing higher resolution imaging and energy discrimination.

The small area (8mm x 8mm) scintillator-coupled EM-CCD operated at 2fps, limiting potential applications. Further developments are required to transfer this technology and expertise to the marketplace. 

We aim to produce a large area, high speed X-ray detector module, making use of the now commercially available highspeed electronics developed from the STFC funded 'Lucky Imaging' at the Institute of Astronomy, Cambridge. By coupling a fibre-optic taper to a larger area EM-CCD, an increase in area of over 44 times is possible. It is envisaged that a series of modules will then be formed into an array, creating a much larger system. We also aim to test the suitability of the scintillating fibre-optic (SFO). Whilst the SFO has largely been ignored for use with a CCD due to lower light output, it has a highly structured form, minimising the signal spread. With the EM-CCD's ability to apply gain to the signal, it is expected that a high-resolution integrating system may be produced. 

In comparison to previous detectors, the expected performance of the new module will give a higher resolution, faster speed (increasing beamline throughput), higher effective dynamic range through higher maximum flux before saturation and higher detection efficiency, higher signal to noise and operation at higher temperatures. The projected specifications of the module will provide these substantial benefits to users, including allowing higher throughput in the beamline facilities and shutter-less performance, providing the high speed of the low resolution 'pixel detectors' and the high resolution of the low speed CCD systems. 

Through the production of a proof of concept prototype module, not only will this technology be opened up to the marketplace, but the range of applications for the EM-CCD will be dramatically expanded, opening new markets for this device. 

This detector is aimed towards applications at synchrotron facilities such as macromolecular crystallographysurface diffraction or small-angle scattering techniqueshigh energy X-ray diffraction andphase-contrast imaging. Applications in medical imaging may also be envisaged for a larger array of modules.  To transfer this technology and expertise to the marketplace, we propose to build a proof of concept module with the support of e2v technologies, a leading designer, developer and manufacturer of high performance imaging sensors.

QE plot of a back-illuminated CCD showing the difference between an uncoated version and devices with the thinnest and thickest filters used on the RGS on XMM-Newton

Thin filter study

To maximise the performance of an X-ray instrument you need a detector with a high Quantum Efficiency (QE) in an environment with a low light background.  Unfortunately, the environment that X-ray detectors operate in have a large amount of stray-light that will cause an increase in the detector background.

To counter this increase in background, the detector can be coated with a thin optical filter of aluminium coupled with an insulator (typically magnesium fluoride or silicon dioxide).  This study is designed to measure the effect that the thickness of the aluminium and insulator has on the overall detector QE.  A thicker filter will reduce the optical stray-light background detected compared to the thinner one, but it will also cause a drop in QE of the instrument.

By evaluating the effect that different filters have on stay-light and QE, better models can be developed to aid decisions on filter thickness in future X-ray missions.

So far I have been able to test two EM-CCDs (e2v CCD97) using the PTB beamlines at BESSY II to get a baseline performance for detectors with no filters over the 50 eV to 1.9 keV energy range.  Later in the year I hope to have thin filter covered EM-CCDs to compare with these results.

The Reflection Grating Spectrometer (RGS) on XMM-Newton