Section 7.1.6. Position-sensitive detectors

The detection efficiency of a detector is determined in the first instance by the fraction of the number of incident photons transmitted by any necessary window or inactive layer, multiplied by the fraction usefully absorbed in the active region of the detector. This product, which is often called the absorption efficiency or the quantum efficiency, should be somewhere between 0.5 and 1.0 since the information loss due to incident photons not absorbed in the active region cannot be retrieved by subsequent signal amplification. The useful efficiency is best described by the so-called detective quantum efficiency (DQE), (Rose, 1946; Jones, 1958). For our purposes, this can be defined as where N is the number of quanta incident upon the detector and σ is the standard deviation of the analogue output signal of amplitude S. For a photon counter with an absorption efficiency q, S = qN, , and = q. An analogue detector with a DQE thus behaves like a perfect counter that only detects a fraction of the incident photons.

Under favourable conditions, the DQE of analogue detectors for X-rays is in excess of 0.5, but varies with counting rate and is lower for detectors with a very large dynamic range, as shown below.

The DQE of CCD- and vidicon-based X-ray detectors has been discussed by Stanton, Phillips, Li & Kalata (1992a).

The linearity of a counter depends on the counting losses, which are due to the finite dead-time of the counter and its processing circuits. The counting losses are affected by the time modulation, if any, of the source, as, for example, with storage rings (Arndt, 1978).

Counting losses can affect the behaviour of detectors in two different ways. In most analogue detectors and in counters with parallel read-out, each pixel behaves as an independent detector and the counting loss at any point depends only on the local counting rate. In other devices, such as multiwire proportional chambers with delay-line read-out (see Subsection 7.1.6.2), the whole detector becomes dead after an event anywhere in the detector and what matters is the global counting rate.

Fortunately, the fractional counting loss is the same for all parts of the pattern so that the relative intensities in a stationary pattern are not affected.

The lowest practically measurable intensity is determined by the inherent background or noise of the detector. Some form of discrimination against noise pulses is usually possible with a detector that counts individual photons, but not, of course, with integrating detectors.

The maximum intensity at which a counter can operate is determined by the dead-time. In the case of an integrating or analogue detector with a variable gain, there is a trade-off between maximum intensity and DQE. Such a device can often be regarded as having an output signal with an amplitude S = NV/M that is a noise-free representation of N, the number of photons detected in the integrating period of the device, where V, the maximum signal amplitude, is produced by M photons in this period. M can be varied by altering the gain of the detector. The noise can be regarded as a fixed fraction 1/r of the maximum amplitude V that is added to the signal. Then the DQE will be

This equation shows the importance of having as small a value of 1/r as possible; it also demonstrates that, for a given value of r, M can be increased only at the expense of a reduced DQE. This is valid for X-ray film (Arndt, Gilmore & Wonacott, 1977), for television detectors (Arndt, 1984), for the integrating gas detectors discussed in Subsection 7.1.6.2, and for many semiconductor X-ray detectors.

The spatial resolution of a PSD is determined by the number and size of resolution or picture elements (pixels) along the length or parallel to the edge of the detector. In most diffraction experiments, the size of the pattern can be scaled by altering the distance of the detector from the sample and what is important is the angular resolution of the detector when placed at a distance where it can `see' the entire pattern. We shall see below that linear PSD's can be made with up to 2000 pixels and that area detectors are mostly limited to fewer than 512 × 512 pixels. The sizes of pixels range from about 10 µm for semiconductor PSD's to about 1 mm for most gas-filled detectors.

The useful number of pixels of a detector is determined by its point-spread function (PSF). This is the relative response as a function of distance from the centre of a point image on the detector. PSF's are not necessarily radially symmetrical and may have to be specified in at least two directions at right angles, for example along and perpendicular to the lines of a television raster scan. The width of the PSF at the 50% level determines the amount of detail visible in a directly viewed image. The accuracy of intensity measurements may depend more critically upon the width of the PSF at a lower level, since a weak spot may be immeasurable when sitting on the `tail' of a very intense one. For various physical reasons, the PSF's of all PSD's, including X-ray film, have appreciable tails.

The spatial resolution of a detector is affected by parallax: when an X-ray beam is absorbed in a thick planar detector at an angle to the normal, the width of the resultant image is smeared out exponentially and its centroid is shifted by an amount . For 8 keV X-rays incident at 45° on a xenon-filled counter, for example, this shift is about 4 mm for a filling pressure of 1 atm and 0.4 mm for a filling pressure of 10 atm. These figures illustrate the desirability of high-pressure xenon (Fig. 7.1.6.2 ) for gas-ionization detectors intended for wide-angle diffraction patterns.

Figure 7.1.6.2| top | pdf | Absorption of 8 keV and 17 keV photons in argon and xenon as a function of pressure in atm × column length in mm.

All PSD's show long-range and pixel-to-pixel variations of response to larger or smaller extents. These can be corrected, in general by means of a look-up table, during data processing, but the measurements necessary for the calibration are often time-consuming. The output signals of many analogue detectors contain fixed-pattern noise that is synchronous with the read-out clock. This noise is usually removed during data processing, which in any case requires the subtraction of the background pattern.

In most detectors, there is some spatial distortion of the image. Again, the necessary calibration procedure may be time-consuming. Distortions cause point-to-point variations in pixel size, which produce response variations additional to those from other causes.

Corrections for spatial distortion and for non-uniformity of response have been discussed by Thomas (1989, 1990) and by Stanton, Phillips, Li & Kalata (1992b).

The amplitude of the signal due to a single photon is usually a function of the photon energy. The variance in this amplitude, or the full width at half-maximum (FWHM) of the pulse-height spectrum, for a monoenergetic input, depends on the statistics of the detection process. A sharp pulse-height-distribution (PHD) curve may permit simultaneous multi-wavelength measurements with a suitable counter, or at least afford a reduction of the background by pulse-height discrimination. In an analogue detector, the variance in the primary amplitude affects the DQE (Arndt & Gilmore, 1979).

Many investigations, such as low-angle or fibre diffraction studies on biological materials carried out with synchrotron radiation (Boulin, Dainton, Dorrington, Elsner, Gabriel, Bordas & Koch, 1982; Huxley & Faruqi, 1983), involve the time variation of a diffraction pattern. For very short time slots, only counters can be employed; the incoming pulses must then be gated and stored in an appropriate memory (Faruqi & Bond, 1982).

Stores used for these experiments are described as histogramming memories; the contents of a given storage location are incremented by one whenever the corresponding address, which represents the position and the time of arrival of a photon, appears on the address bus (Hendricks, Seeger, Scheer & Suehiro, 1982; Hughes & Sumner, 1981).

Stability of the performance of a detector is of paramount importance. Most position-sensitive detectors are used in connection with microcomputers, which make calibration and corrections for spatial distortion, non-uniformity of response, and lack of linearity relatively easy, provided that these distortions remain constant. The long-term stability of many detectors, notably of semiconductor devices, is affected by radiation damage produced by prolonged exposure to intense irradiation.

The size and weight of the detector determine the ease with which the detector can be moved relative to the sample and thus the extent to which the diffractometer can be adapted to varying resolution and collimation conditions. Some detectors cannot be moved easily (Xuong, Ng, Freer, Hamlin, Neilsen & Vernon, 1978), or need very heavily engineered rotational and translational displacement devices (Phizackerley, Cork & Merritt, 1986). Others, such as spherical drift chamber multiwire proportional chambers, are designed for use at a fixed distance from the sample and may only be swung about the latter but not translated (Kahn, Fourme, Bosshard, Caudron, Santiard & Charpak, 1982).

In single-crystal diffraction patterns, the Bragg reflections may be visualized as diverging from the X-ray source while the background – fluorescence X-rays, scatter by amorphous material on the specimen crystal and its mount – diverges from the sample. Consequently, the highest reflection-to-background ratio is achieved by using a large detector at a large distance from the specimen. Of all the detectors discussed here, the image plate can most readily and economically be used to cover a large area; its present popularity is chiefly due to this property (Sakabe, 1991).

There is fairly general agreement on the specification of an ideal X-ray area detector. It should be at least 250 mm in diameter, contain at least 1000 × 1000 pixels, have a large dynamic range and a high detective quantum efficiency for photon energies up to 20 keV and be capable of being read-out rapidly.

Many suggestions have been made for improving the performance of existing detectors. It has become apparent, however, that the development of ideal, or even better, X-ray detectors is extremely expensive and, therefore, that their development and installation can be undertaken only in central national or international laboratories such as storage-ring synchrotron-radiation laboratories.

There are several methods of deriving the position of the detected photon that are applicable to both linear and area detectors.

In the gas-filled detectors that we have considered so far, the electric field is cylindrically symmetrical in the immediate vicinity of the wire or wires near which gas multiplication takes place and the maximum count rate is limited ultimately by the electrostatic shielding effect of the ion sheath owing to previous X-ray photons. In parallel-plate chambers, the electrodes are in the form of very fine electro-formed grids: With this structure, the pulse shape is quite different; the very sharp initial part, due to the rapidly moving electrons, can be separated, at the expense of a loss of signal amplitude, from the slow component due to the positive ions; in addition, the shielding effect is much less pronounced. Accordingly, counting rates up to at least 10¹¹ s⁻¹ m⁻² are possible with parallel-plate PSD's (Stümpel, Sanford & Goddard, 1973; Peisert, 1982; Hendrix, 1984).

For the very highest counting rates, it is necessary to abandon all methods in which individual X-ray photons are counted and instead to measure the ionization current produced by the incident X-rays on either cathode or anode. Fig. 7.1.6.6 shows the principles of a cathode read-out linear PSD. The cathode is divided into strips, each of which is connected to a capacitor and to an input terminal of a CMOS analogue multiplexer. The charge accumulated on each capacitor in a given time period is transferred to a charge-sensitive amplifier when the associated channel is selected by an addressing signal. The output voltage of the amplifier is digitized by means of an analogue-to-digital converter. The complete pattern is scanned by incrementing the addresses sequentially: The resolution is that of the strip spacing (∼0.5 mm) and the principle can be extended to two dimensions (Hasegawa, Mochiki & Sekiguchi, 1981; Mochiki, Hasegawa, Sekiguchi & Yoshioka, 1981; Mochiki, 1984; Mochiki & Hasegawa, 1985). Global count rates in excess of 10⁹ s⁻¹ are possible with this method. Lewis (1994) has published a comprehensive survey of the present status and the future potentialities of gas-filled position-sensitive detectors.

Figure 7.1.6.6| top | pdf | Integrating LPSD. From Mochiki (1984); courtesy of K. Mochiki.

For X-ray diffraction applications, the main disadvantage of semiconductor devices is that the universal trend in their manufacture is in the direction of miniaturization, leading to a very small pixel size. Thus, imaging devices with up to 2000 × 2000 pixels have been produced but the pixel size is typically ∼10 µm for CCD's and ∼20 µm for PDA's; in most X-ray diffraction applications, it would be difficult to scale down sample and source sizes to a point where the pattern size is appropriate to a semiconductor imager, even though a 2000 × 2000 CCD with 27 × 27 µm pixels is available.

Semiconductor point counters cooled to 80 K are characterized by their very high energy resolution (better than 2% in the 8 keV region). Some counting PSD's potentially have a similar energy resolution. Two-dimensional X-ray-sensitive CCD's for X-ray astronomy research have been used as photon counters (Walton, Stern, Catura & Culhane, 1985; Lumb, Chowanietz & Wells, 1985) but they can only be used at very low counting rates. In integrating devices, the energy discrimination is lost. Silicon detectors for visible-light applications are made with depletion depths of the order of 10 µm. For the detection of 8 keV photons with more than 90% efficiency, depletion depths of 165 µm are necessary and these can be produced only from very high resistivity material (Howes & Morgan, 1979). Moreover, in commercial visible-light imagers, the depletion region is covered by circuitry or by an inactive layer that constitutes an absorbing window for X-ray detection. For the detection of X-rays or electrons, it is, therefore, customary to thin the device and to illuminate it from the back (see, for example, Meyer-Ilse, Wilhelm & Guttmann, 1993).

One-dimensional X-ray detectors utilizing PDA's, such as those made by the Reticon Corporation, have found a number of applications, especially in dispersive X-ray absorption spectroscopy (EXAFS) (Jucha, Bonin, Dartyge, Flank, Fontaine & Raoux, 1984).

A particular problem with silicon detectors is the damage caused by the incidence of X-rays or of energetic electrons. The effects can be minimized by masking all but the active part of their device and by operating it at low temperatures (Jucha et al., 1984).

The use of the room-temperature semiconductor mercuric iodide in place of silicon seems promising (Patt, Delduca, Dolin & Ortale, 1986).

The X-ray diffraction applications of directly sensitive semi-conductor PSD's are likely to remain limited. A previous conversion to visible light or to electrons offers the possibility of an optical or electron-optical demagnification onto the imager, as well as of avoiding some of the other problems discussed above (Deckman & Gruner, 1986).

Standard light-sensitive semiconductor imaging devices can be used for X-ray detection if the X-rays are first converted to light by means of a phosphor. One 8 keV X-ray photon produces several hundred light photons in a good phosphor (see Table 7.1.6.2). Because of the low noise levels possible with cooled CCD's (∼10 electrons r.m.s.), only 1 to 3% of these photons need to reach the device to produce a perfect X-ray detector. Unfortunately, even this is possible only with optics that do not demagnify to any considerable extent (see Subsection 7.1.6.5) and one is thus restricted to a small detector. However, CCDs are now available that can be butted along two or three edges and these make possible the construction of `tiled' detectors that contain four or six of such CCD chips (Burke, Mountain, Harrison, Bautz, Doty, Ricker & Daniels, 1991; Fordham, Bellis, Bone & Norton, 1991; Allinson, 1994). Individual channels can be read simultaneously (Hopf & Rodricks, 1994), thus making possible a relatively rapid read out of a large number of pixels.

In image intensifiers, an output image is produced on a phosphor screen by electrons with an energy of a few keV. A promising device with direct positional read out consists of a demagnifying intensifier in which the phosphor is replaced by a CCD. Various intensifiers have been described in which the electrons reaching the CCD have an energy of several keV so that the electron–hole generation is amplified (EBS process) (Lowrance, Zucchino, Renda & Long, 1979; Lemonnier, Richard, Piaget, Petit & Vittot, 1985). High-energy electrons are liable to cause radiation damage and similar precautions as for X-rays, such as thinning and back-illumination, are necessary. Experiments have been reported with less-damaging low-energy (200 eV) electrons produced in a microchannel plate (MCP) image intensifier (Dereniak, Roehrig, Salcido, Pommerrenig, Simms & Abrahams, 1985). Both approaches look promising but it is questionable whether a device with an X-ray phosphor input and of adequate size and resolution at an affordable price will emerge in the near future.

The incoming X-rays are converted to light in a phosphor that is coupled to the first photocathode of the system. Both polycrystalline and monocrystalline phosphors are used for X-ray detection. The former give a higher light output but have a limited resolution; the latter tend to have a poorer light-conversion efficiency but have the best resolution. The most useful phosphors are shown in Table 7.1.6.2.

Many attempts have been made to improve the spatial resolution of phosphor screens by constructing them in the form of scintillating fibres that are optically isolated from one another so that the scintillation does not spread. This can be achieved by growing columnar scintillating crystals (Oba, Ito, Yamaguchi & Tanaka, 1988), by intagliating polycrystalline phosphors (Fouassier, Duchenois, Dietz, Guillemet & Lemonnier, 1988) and by using arrays of scintillating fibres (Bigler, Polack & Lowenthal, 1986; Ikhlef & Skowronek, 1993, 1994).

Image intensifiers designed for radiography with relatively hard radiation usually have an X-ray-transparent window – which may be up to 300 mm in diameter – and an internal phosphor–photocathode sandwich deposited on an X-ray-transparent substrate. Problems of compatibility of phosphor and photocathode have restricted the phosphor used, but CsI works well with multialkali photocathodes. Moy and his collaborators have constructed a large-diameter television detector in which the image intensifierhas been modified by using beryllium for the window and for the sandwich substrate; this intensifier is coupled to a slow-scan CCD camera (Moy, 1994).

Each incident X-ray photon should give rise to several photoelectrons from the first photocathode in order to achieve a high DQE (Arndt & Gilmore, 1979). The best photocathodes have a yield of about 0.2 electrons per light photon; only fibre-optics coupling between the phosphor and the photo-cathode can give an adequate light-collection efficiency (in excess of 80% for 1:1 imaging). With demagnifying fibre optics, the light loss is considerable for purely geometrical reasons. Fibre-optic cones, in which each individual glass fibre is conical, are available for magnification or demagnification up to about 5:1. It should be noted that image intensifiers and TV tubes with electrostatic focusing are normally made with fibre-optics face-plates and that some CCD's are also available with fibre-optics windows.

Where lens optics must be employed, it is best to use two infinity-corrected objectives of the same diameter, but not necessarily of the same focal length, back to back. In practice, the best light-collection efficiency that can be achieved at a demagnification of M is about . It is for this reason that there are limitations on the maximum size of image that can be projected onto a CCD without the use of an image intensifier (but see Naday, Westbrook, Westbrook, Travis, Stanton, Phillips, O'Mara & Xie, 1994; Koch, 1994; Allinson, 1994). The function of this intensifier is to match a relatively large diameter X-ray phosphor to a small-size read-out device. As a result of the high sensitivity of CCD's, especially of slow-scan CCD's, a low photon gain in the intensifier and a low light-coupling efficiency in the coupling between intensifier and read-out device are quite adequate. It is, however, essential to couple the X-ray phosphor as efficiently as possible to the photocathode.

Roehrig et al. (1989) have described a design in which an image intensifier with 150 mm diameter input and output face plates is coupled by means of six demagnifying fibre-optic cones to six CCD's. Allinson (1994) has examined the need for image intensification and has shown that it is possible to construct a 150 mm square detector that has a performance approaching that of an ideal detector without using an image intensifier. Different methods of light coupling and format alteration have been discussed by Deckman & Gruner (1986).

In an image intensifier, the photoelectrons from a photocathode are made to produce a visible intensified image on an output phosphor. In so-called first-generation tubes, the intensification is produced by subjecting the electrons to accelerating voltages of up to about 15 keV: the number of visible-light photons at the output per keV of electron energy is about 80. The photon gain of the devices is typically about 100; there may be a brightness gain factor of M² if the electrostatic electron-optical system produces a demagnification of M. Standard first-generation image intensifiers of this type are made with input field diameters up to 80 mm, they always have fibre-optics input face plates on which X-ray phosphor may be deposited, they are stable and robust, and have a good resolution of better than 100 µm at the input. Their low gain requires the use of a low-light-level TV camera tube in the next stage (Arndt & Gilmore, 1979) or of two or more intensifiers of this type in tandem (Kalata, 1982) or of an intrinsically more sensitive slow-scan read out (Eikenberry, Gruner & Lowrance, 1986).

For military and civilian night-vision applications, first-generation image intensifiers have largely been replaced by devices embodying one or two microchannel plates (MCP's) that produce an electron gain of up to 1000 per stage (see, for example, Emberson & Holmshaw, 1972; Garfield, Wilson, Goodson & Butler, 1976). Commercial second- and third-generation intensifiers (Pollehn, 1985) are less suitable for quantitative scientific purposes than the first-generation devices: their GaAs photocathodes are less well matched to most X-ray phosphors, the gain of MCP's decreases with time, and the tubes have a slightly lower resolution than diode types of comparable diameter. Most third-generation intensifiers have plain rather than fibre-optics face plates and none appear to be available with a diameter greater than 50 mm (Airey & Morgan, 1985). Nevertheless, these high-gain intensifiers do make it possible to construct relatively cheap moderate-performance X-ray detectors using standard-sensitivity TV pick-up devices, including CCD's (Dalglish, James & Tubbenhauer, 1984), instead of the low-light-level camera tubes necessary with a lower pre-amplification.

An intensifier can, in principle, employ a variety of read-out methods, e.g. by substituting a resistive disc anode, a coded anode or a CCD for the output phosphor. However, the only way of employing standard modules is to couple them to a TV pick-up device.

Vacuum-tube television cameras have been largely replaced by semiconductor devices but of the former the preferred tube for use in an X-ray detector is still the silicon-intensifier-target (SIT) tube (Santilli & Conger, 1972). It has an adequate resolution for images up to 512 × 512 pixels and a linear transfer function (unity gamma) and its sensitivity is well matched for use with a single-stage image intensifier with a gain of 100. When cooled, this tube can be used for long exposures in an integrated slow read-out mode (Milch, Gruner & Reynolds, 1982).

When better high-gain image intensifiers become available, the preferred choice may be tubes like the Saticon (Goto, Isozaki, Shidara, Maruyama, Hirai & Fujita, 1974) whose diode gun gives them a superior resolution (Izosaki, Kumada, Okude, Oguso & Goto, 1981) and which have superior `lag' or `sticking' performance, that is a short `memory' for previous high-intensity patterns to which they have been exposed (Shidara, Tanioka, Hirai & Nonaka, 1985).