Kieran J. McCarthy

The X-ray energy response of silicon Part A. Theory

Abstract In this, the first part of a two-part study of the interaction of soft X-rays with silicon, motivated by the calibration requirements of CCD imaging spectrometers in astronomy, we describe a Monte Carlo model of X-ray energy loss whose products are the energy- and temperature-dependences of (i) W , the average energy required to create an electron-hole pair, and (ii) the Fano factor F , W and F have invariably been treated as material constants in previous analyses of Si X-ray detector performance. We show that in fact, at constant detector temperature T , W is an increasing function of X-ray energy for E F is predicted to increase slowly with E . The temperature coefficient d W /d T has a calculated value ∼ 1 × 10 −4 K −1 at a typical CCD operating temperature of 170 K. We discuss the practical implications of these results. Finally, we describe our separate calculations of the near-edge variation of CCD quantum detection efficiency arising from silicon K-shell Extended X-ray Absorption Fine Structure (EXAFS).

The X-ray energy response of silicon (B): Measurements

Abstract In this, the second part of a detailed study of the interaction of soft X-rays with silicon, we summarise the results of a large number of experiments on charge coupled devices (CCDs), carried out both in our laboratory and at the Daresbury Synchrotron Radiation Source (SRS). Measurements of the energy variation of the W parameter and of the Fano factor F are in substantial agreement with the predictions of the model developed in Part (A) of the study [G.W. Fraser et al., Nucl. Instr. and Meth. A 350 (1994) 368]. The consequences of using a Gaussian pulse height distribution model in the experimental determination of F are discussed. Variations in X-ray event morphology (i.e. the frequency distribution of single-, two-, three-pixel events) across the silicon K edge are described. Measurements of CCD quantum detection efficiency Q (counts/photon) showing XAFS (X-ray absorption fine structure) modulation in the vicinity of the Si K edge are compared with calculations based upon new, experimentally-determined linear absorption coefficients for Si, SiO 2 and Si 3 N 4 . Finally, the X-ray photoyield from silicon is described, both experimentally and theoretically.

Modelling the X-ray response of charge coupled devices

Abstract Based on the physics of charge generation, diffusion and collection, we have developed a simple three-dimensional Monte Carlo model to predict the X-ray response of deep depletion front-illuminated CCDs. It is shown that the measurable properties of the device (i.e., energy resolution, quantum efficiency and event morphology) can all be reproduced by a simple treatment of charge creation and diffusion within each of the active layers of the device. The simulation reproduces the distinct spectral signatures of X-ray interactions in the depletion layer and the field-free regions. Pulse height spectra can be reproduced to good accuracy and quantum efficiencies and energy resolutions could be predicted to within a few percent over the energy range 500 eV to ∼ 10 keV. Refinements such as the effect of altering electrode voltages have been incorporated into the model.

Proton Damage Effects in EEV Charge Coupled Devices

An examination is conducted of the effects of low-energy protons on CCD performance to evaluate the potential effectiveness of space-borne observational instruments. Degradation is described as a function of incremental dose, irradiation temperature, or proton energy for several device architectures, some of which incorporate design features to minimize signal-charge/trapping-site interaction. Degradation of the charge transfer is studied for very low proton doses, and dark current is found to vary directly with proton dose. Displacement damage in the signal-transfer channels generates charge-trapping sites that have a negative effect on EEV CCD performance. Degradation of charge-transfer performance is shown to be the most significant hindrance to effective CCD operations for X-ray spectroscopic applications.

Event recognition in X-ray CCDs

Abstract The signatures of photon and charged particle interactions in an X-ray CCD have been investigated and techniques for reducing spectral data developed. It is found that the energy-loss spectrum of X-rays which interact in the depletion layer are well described by simple functions and show little evidence of redistributive processes. However, photons which interact in the “so-called” field-free region introduce multi-pixel, line-like features into the spectral data which are shifted in energy and slightly broadened with respect to depletion region events. The efficiency of this process can be surprisingly high, being comparable to the depletion layer efficiency at 10 keV for the devices considered in this study. For astrophysical applications, these artifacts can have energies and intensities comparable to expected non-solar line and edge features. We have developed procedures to identify, quantify and correct for these effects to arbitrary precision. Simple statistical techniques are described to ascertain when, and if, they need to be applied.

Soft X-ray response of charge coupled devices

Abstract The charge coupled device (CCD) is useful as an imaging detector in the X-ray waveband, and when used to detected single photon events can simultaneously provide spatial imaging and energy resolution. This capability is important in the field of X-ray astronomy. With radiation focussed onto the front electrodes, the CCD suffers from reduced response below 1 keV due to absorption in the electrode structure; the same process that limits the blue response in the optical band. Two alternative developments are described that extend the soft X-ray and blue responses of the CCD. Experimental data are presented for these two configurations of CCD and their performance characteristics as X-ray images are compared.

The X-Ray CCDs Developed for the Joint European X-Ray Telescope

The charge coupled devices (CCDs) developed for the Joint European X-ray Telescope (JET-X) are described in detail. A history of the development program and device performance is given. We present results from a comprehensive study to characterize the x-ray response of the flight model focal plane detectors. The goal of the program is to calibrate the efficiency, energy resolution, gain, etc. down to a precision of ∼1%. Final calibration data sets will be based on combinations of measurements and calculations. For example, the CCD quantum efficiency will be composed of discrete line measurements made at the University of Leicester test facility and calculation and synchrotron measurements from the Daresbury Synchrotron Radiation Source (SRS). The absolute normalizations will be provided by x-ray long beam pipe measurements at the Max Planck Institut fur Extraterrestrische Physik (MPE) Panter test facility in Munich. Using the available data, it is shown that it is possible to calibrate the quantum efficiency, the FWHM energy resolution, and the system gain of the flight devices to better than 1%.

Energy deposition in X-ray CCDs and charged particle discrimination

Abstract We report the results of investigation into ionization energy loss in X-ray CCDs and the ability of fast protons to masquerade as X-ray events by their energy and spatial signatures. The study was made to explore background contamination in X-ray detectors intended to operate in the mixed radiation environment of space. It was carried out using the ionization energy loss code of Hall [1] which includes atomic electron binding effects. Three devices were considered in our study; a front illuminated 35 μm deep depletion device developed for the Joint European X-ray Telescope, and two CCDs baselined for ESAs XMM mission — a PN device of active depth 280 μm and a back-illuminated MOS device of active depth 90 μm. It was found that for the JET-X device, 4.9% of minimally ionizing protons deposit less than 10 keV in the active volume of the CCD. Whilst the frequency distribution of such events versus energy deposited increases up to a peak energy of ∼9 keV, the number having a morphology consistent with X-ray events simultaneously decreases. Thus, when coupled with the requirement that events should extend no more than 2 pixels spatially, the fraction of masquerading events drops to 0.016%. The analysis was also carried out for the two XMM CCDs. For the PN device the fraction of events depositing less than 15 keV is miniscule whereas almost 4% of events were recorded for the MOS device. When coupled with an event size requirement of ≤2 pixels, the fraction of events falls to 1 × 10 −9 . A surprising result of the present work is that the ability of a CCD to discriminate against minimally ionizing particles is dependent on pixel size. For a JET-X like device this varies by a factor 10 for pixel sizes ranging from 13 to 35 μm on a side. In fact, the optimum pixel size for both the JET-X and XMM MOS devices was found to be near 40 μm. We conclude that given the energy deposition and morphological distinctions between photon and particle events, CCDs used in X-ray astronomy can be optimally designed to provide simultaneously the highest signal-to-noise ratios for X-rays and rejection ratios for charged particles.

Measured and modelled CCD response to protons in the energy range 50 to 300 MeV

Abstract A deep depletion X-ray sensitive CCD was exposed to monoenergetic, collimated protons of energies 50, 100, 200 and 300 MeV at the proton irradiation facility (PIF) at the Paul Scherrer Institute. Measurements were made at angles-of-incidence 0°, 20° and 45° to the detector normal. The energy-loss spectra and spatial distribution of the collected charge were derived and compared to theoretical predictions which take into account the CCD geometry and the generation and diffusion of charge within the active device. These data were then used to determine the efficiency with which protons could be discriminated from 0.3–10 keV X-rays. By applying energy deposition and size requirements (in terms of the number of reporting pixels) on each event, it was found that X-rays, single proton and multiple proton events could all be separated unambiguously although with varying efficiencies. For energy depositions in the range 0.3 to 7 keV, 100% particle discrimination can be achieved for all sizes of X-ray event. Above 7 keV, the rejection efficiency can still be maintained at 100%, if one is prepared to reject between 0 and 39% of the X-rays depending on incident proton energy, by the selective use of size and energy cuts. It was found that for this CCD the dominant form of masquerading events (i.e., events which cannot be discriminated against) arise from proton interactions in the readout register. Since the register constitutes only 1% of the active area, the number of contaminating events is a small fraction of the total incident flux. Approximately 99.9% of charged particles can still be rejected on size and energy grounds whilst still accepting 100% of 7 to 10 keV X-rays.

Measurements and simulations of x-ray quantum efficiency and energy resolution of large-area CCDs between 0.3 and 10 keV

Large area CCD arrays, 770 X 1024 pixels, with 27 micrometers square pixels and packaged in matching pairs in a ruggedized focal plane assembly have been developed for the JET-X instrument for the Russian Spectrum-RG spacecraft. The devices achieve low noise readout (<EQ 5 electrons rms), low dark current (< 1 electron/pixel/frame-integration time) and high charge transfer efficiency (> 0.99997) enabling the device to combine high quantum efficiency and good energy resolution over the operating range of JET-X. New results from flight prototype JET-X CCDs will be presented which reveal the detail of charge spreading behavior in the device and the consequent effect on quantum efficiency, energy resolution and background rejection. Theoretical modelling and simulation of these processes are used to analyze the experimental results.