International
Tables for
Crystallography
Volume F
Crystallography of biological macromolecules
Edited by E. Arnold, D. M. Himmel and M. G. Rossmann

International Tables for Crystallography (2012). Vol. F, ch. 7.1, pp. 179-181   | 1 | 2 |

Section 7.1.3. Characteristics of different detector approaches

S. M. Gruner,a* E. F. Eikenberryb and M. W. Tatea

aDepartment of Physics, 162 Clark Hall, Cornell University, Ithaca, NY 14853–2501, USA, and bSwiss Light Source, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland
Correspondence e-mail:  smg26@cornell.edu

7.1.3. Characteristics of different detector approaches

| top | pdf |

7.1.3.1. Point versus linear versus area detection

| top | pdf |

A point detector may be based on a scintillating crystal or a gas-filled counter, with the sensitive area defined by slits or a pinhole mask. The spatial resolution of such a detector can be made arbitrarily fine at the expense of data collection rate. Point detectors can have very high accuracy if the background is removed by energy discrimination. They find application in powder diffractometry and small-molecule crystallography, in which the reflections are widely dispersed, thereby simplifying measurement of individual reflections. Clearly, specimen and source stability are important for such work.

Throughput can be greatly increased by area detection, which is often required for macromolecular crystallography or investigations of unstable specimens. Typical area detectors, such as film, storage phosphors and charge-coupled devices (CCDs), are described below.

7.1.3.2. Counting and integrating detectors

| top | pdf |

Detectors can be broadly divided into photon counters and photon integrators. Photon counters have the advantage that some designs permit energy discrimination, allowing them to reject inelastically scattered radiation, thereby improving the signal-to-noise ratio. However, photon-counting detectors always have a count-rate limitation, above which they begin to miss events, or even become unresponsive (the time during which a detector misses events is known as dead time). Prototype systems have demonstrated linear count rates greater than 106 photon s−1. Fabrication difficulties have limited the commercial availability of photon-counting detectors with large areas, high spatial resolution and high count rates. The count rate is a particular concern at modern synchrotron sources, which are capable of generating diffraction that delivers two or more photons to a pixel during one bunch time, an instantaneous count rate greater that 1010 photons per second per pixel. Integrating detectors are more typically used in situations where very high event rates are expected.

In contrast, integrating detectors have no inherent count-rate limitation, though at very high fluxes several sources of non­linearity can theoretically become important, such as nonlinearity in the phosphor used to convert the X-ray image to a visible image. Integrating detectors, however, do not discriminate energy, and they have noise that increases with integration time. Nonetheless, film, image-plate and CCD integrating detectors are currently commercially available and in widespread use.

7.1.3.2.1. Photon-counting detectors

| top | pdf |

Commonly used photon counters include scintillator/photomultiplier combinations, gas-filled counters and reverse-biased semi-conductor detectors.

Scintillator/photomultipliers usually consist of a relatively thick crystal of a scintillator coupled to a high-gain photomultiplier tube. These detectors are generally designed to serve as point photon counters with moderate energy resolution. In order to perform this function, several constraints must be met:

  • (1) The scintillator crystal must be thick enough to have almost unity stopping power.

  • (2) It is necessary to collect as many of the converted visible photons as possible, so an optically clean scintillator crystal is used in a reflective housing to direct as many photons as possible toward the phototube.

  • (3) The scintillator must emit its light quickly, so as to minimize dead time, and be efficient, so as to emit much light. NaI:Tl, CsI:Na and CsI:Tl meet these constraints. NaI is more commonly used, but CsI may be preferred at higher X-ray energies because of its higher stopping power. Both materials are hygroscopic and are usually encased in hermetically sealed capsules with beryllium windows.

  • (4) The phototube is usually operated in its linear region for energy discrimination.

Scintillator/phototube combinations are relatively trouble-free and often have near-unity DQE. Their main limitations are count rates well below 106 photon s−1 and the lack of spatial resolution. Even so, such detectors are still preferred in many applications where the data are effectively zero- or one-dimensional.

Reverse-biased semiconductor detectors are designed to have a thick depletion zone in which charge can be efficiently collected and conveyed to an amplifier. X-rays that stop in the depletion zone produce electron–hole pairs; these are separated by the depletion zone field and the electrons are swept to the input of a low-noise amplifier. Single-photon counting can be readily achieved, even for low-energy X-rays, especially if the detector is cooled to minimize thermally generated charge. These detectors are typically fabricated as silicon diodes, but germanium and gallium arsenide are also used (Hall, 1995[link]). Until recently, these devices were generally configured as point detectors or strip detectors consisting of a linear array of narrow sensitive regions, forming a one-dimensional detector (Ludewigt et al., 1994[link]). Two-dimensional arrays of square pixels are being developed, e.g. see the description of pixel array detectors below. In another device, the silicon drift detector, potentials are arranged in the silicon to funnel signals from a large area to a low-noise collection point (Rehak et al., 1986[link]). Such devices are being developed for both linear and area applications.

By increasing the electric field strength in an appropriately designed p-n junction in silicon, avalanche multiplication of the X-ray-induced electrons can be obtained as they move toward the anode where they are collected. This gives rise to a very high linear signal gain, with high speed and low noise. Arrays of such avalanche photodiodes as large as [8 \times 8] elements, each [{1 \times 1\;\hbox{mm}}] square, have been fabricated (Gramsch et al., 1994[link]; Farrell et al., 1994[link]).

Gas discharge (wire) counters make use of the ionization produced when an X-ray is stopped in the high-atomic-number gas, usually xenon, that fills the detector. A strong electric field between a fine anode wire and a cathode plane accelerates the products of the primary ionization to produce an ionizing multiplication (either a proportional or an avalanche discharge, depending on the field strength) that is detected as a charge pulse on one or both of the electrodes. The discharge is quenched by the presence of a few per cent of a second gas, e.g. methane or carbon dioxide. Gas discharge detectors have been configured in zero-, one- and two-dimensional versions and continue to be widely used in some applications. The venerable Geiger counter is in this class and is used for radiation monitoring and beam alignment in home laboratories. Properly designed gas discharge counters have very low noise, but the quantum efficiency depends critically on design, gas and X-ray energy.

Linear wire detectors have been used to record small-angle X-ray scattering. The localization of the X-ray event along the length of the detector is often performed by measuring the difference in arrival time of the charge pulses at the two ends of one of the electrodes (Barbosa et al., 1989[link]). The pulses are stretched to permit this measurement. One design uses a resistive anode wire to perform this function, whereas others configure the cathode plane as a delay line. Various two-dimensional arrangements of crossed planes of wires, broadly classified as multiwire proportional counters (MWPCs), have been widely used in crystallography, and some types have been commercially successful (Hamlin et al., 1981[link]; Blum et al., 1987[link]).

The design of MWPC area detectors has had difficulty keeping up with improvements in X-ray sources, particularly the high fluxes available at storage rings, and the shift toward use of higher-energy X-rays. The electric discharge at the heart of the technology has an inherent dead time associated with it. Added to this inherent dead time are the pulse propagation and processing times which limit the counting rate for a given wire. Thus, MWPCs are subject to a severe count-rate limitation. A second limitation of MWPCs has been their large pixel size and the relatively small number of pixels across the detector face, as well as parallax effects. These problems have been addressed by changes in the detector geometry (e.g. spherical drift chambers; Charpak, 1982[link]), by microfabrication on glass substrates of the wires comprising the back plane of the detector, and by dividing the active area into small zones, each of which is read out independently. Robustness of MWPCs has also been a problem.

The dead time can be reduced by reducing the thickness of the detector. However, reducing the detector thickness reduces the X-ray stopping power. Increasing the gas pressure not only improves the quantum efficiency, but also helps to reduce the dead time further. Unfortunately, high gas pressure complicates the design of the front window of the detector. Despite these problems, two-dimensional gas-detector prototype modules with 200 µm square pixels have been constructed that are expected to have a local linear count-rate limit of 7 MHz mm−2 and a quantum efficiency above 80% at energies used in crystallography (see Sarvestani et al., 1998[link]).

7.1.3.2.2. Integrating detectors

| top | pdf |

X-ray film was the first area detector and has a long history of important contributions to the solution of structures and to X-ray imaging (Arndt et al., 1977[link]). In many applications, film has effectively been displaced because of its relative insensitivity caused by the high level of background fog, its multistep processing leading to long delays before digital data are obtained and its nonlinearity. However, X-ray film is still a superior integrator for long exposures, and lithographic film has much higher spatial resolution than any other area detector. Polaroid film in an X-ray cassette is an excellent diagnostic tool for beam alignment problems. And, in all cases, film is inexpensive.

Storage phosphors, also called image plates, are probably the most widely used area X-ray detector for crystallography at present, particularly in laboratories with conventional X-ray sources (Amemiya et al., 1988[link]; Eikenberry et al., 1992[link]). These sheets of material are a much improved functional replacement for X-ray film in many applications, including medical radiography and autoradiography in biological research. Storage phosphors are made from a BaFBr:Eu or other photostimulable phosphor coated on a suitable backing. These phosphors have the property that absorbed X-ray energy can be trapped in long-lived states within the phosphor grains, and that this energy can later be released as blue fluorescence upon photostimulation with red light. Grain-size distribution, grain orientation, binder choice and coating thickness are important parameters in the commercial preparation of the sheets.

The exposed phosphor sheet is raster-scanned with a finely focused red laser and the resulting photostimulated emission is recorded by a photomultiplier. The result is a digital image of the X-ray intensity distribution. The scanning can be done either on-line in self-contained systems or off-line in a separate scanning instrument. Scanning typically requires several minutes. In advanced scanners, this is reduced to several tens of seconds, where the limit is set by the time constant of the photostimulated emission process, which in turn determines the minimum time the laser should dwell on each pixel. The off-line scanner, preferred at synchrotron sources, permits a new exposure to be made while scanning is performed. Self-contained systems offer the advantages of simplified operation and the possibility of calibrating the detector, since only a single sheet of material is used in a mechanically stable setup.

Storage phosphors are typically read out with 100 µm square pixels, resulting in [2000 \times 2500] pixel images for the common size of sheet. Larger formats are available. The wide PSF of storage phosphors makes the effective pixel size considerably larger than the nominal value. Some scanners permit smaller pixels, but this is of limited utility because there is a readout noise component associated with each pixel and too small a pixel harms the signal-to-noise ratio without improving data. The most critical component of the storage-phosphor system is the mirror assembly that gathers the photostimulated emission during readout. There are very few emitted photons for each stored X-ray (at the photon energies used for diffraction), and only a small fraction of these are detected in the photomultiplier (in some cases less than one per stored X-ray). This step in the detection process is critical in maintaining high quantum efficiency. Although image plates have an inherently wide dynamic range, the practical value is always limited by the scanner analogue-to-digital converter.

Television detectors. Numerous integrating detector designs based on television sensor technologies have been published and, in one case, produced commercially (Milch et al., 1982[link]; Arndt, 1991[link]). These detectors span a wide range of design complexity and performance. The primary element is a phosphor screen, which converts the incident X-ray pattern to a light image that is directly or indirectly coupled to the sensor, such as a Vidicon or CCD. Many of the designs employ image intensifiers to raise the signal strength of the visible image above the noise of the sensor system. Some designs employ cameras operated at video rates, with frames accumulated in the attached computer or on videotape. Other designs use cooled cameras operated in a slow-scan mode, which greatly reduces noise. The X-ray exposure is integrated in the camera, then read out once at the end of the integration period.

Most of these systems would be classified as complex, but several of them are working reliably today in laboratories with conventional X-ray sources. The image intensifiers improve the DQE of the systems, while sacrificing dynamic range and image sharpness. In addition, intensifiers are sensitive to magnetic fields, requiring great care in their use if proper detector calibration is to be maintained.

Considerable enhancement to the television-type detector is made possible by the low-noise imaging capabilities of CCDs, described in Chapter 7.2[link] . In this case, high DQE can be maintained without intensification when the CCD is cooled and read with slow-scan electronics. As such, these detectors are much more robust and have improved imaging qualities.

References

Amemiya, Y., Matsushita, T., Nakagawa, A., Satow, Y., Miyahara, J. & Chikawa, J.-I. (1988). Design and performance of an imaging plate system for X-ray diffraction study. Nucl. Instrum. Methods Phys. Res. A, 266, 645–653.
Arndt, U. W. (1991). Second-generation X-ray television area detectors. Nucl. Instrum. Methods Phys. Res. A, 310, 395–397.
Arndt, U. W., Gilmore, D. J. & Wonacott, A. J. (1977). X-ray film. In The Rotation Method in Crystallography, edited by U. W. Arndt & A. J. Wonacott, pp. 207–218. Amsterdam: North-Holland Publishing Co.
Barbosa, A. F., Gabriel, A. & Craievich, A. (1989). An X-ray gas position-sensitive detector – construction and characterization. Rev. Sci. Instrum. 60, 2315–2317.
Blum, M., Metcalf, P., Harrison, S. C. & Wiley, D. C. (1987). A system for collection and on-line integration of X-ray diffraction data from a multiwire area detector. J. Appl. Cryst. 20, 235–242.
Charpak, G. (1982). Parallax-free, high-accuracy gaseous detectors for X-ray and VUV localization. Nucl. Instrum. Methods, 201, 181–192.
Eikenberry, E. F., Tate, M. W., Bilderback, D. H. & Gruner, S. M. (1992). X-ray detectors: comparison of film, storage phosphors and CCD detectors. Inst. Phys. Conf. Ser. 121, 273–280.
Farrell, R., Vanderpuye, K., Cirignano, L., Squillante, M. R. & Entine, G. (1994). Radiation detection performance of very high-gain avalanche photodiodes. Nucl. Instrum. Methods Phys. Res. A, 353, 176–179.
Gramsch, E., Szawlowski, M., Zhang, S. & Madden, M. (1994). Fast, high-density avalanche photodiode-array. IEEE Trans. Nucl. Sci. 41, 762–766.
Hall, G. (1995). Silicon pixel detectors for X-ray diffraction studies at synchrotron sources. Q. Rev. Biophys. 28, 1–32.
Hamlin, R., Cork, C., Howard, A., Nielsen, C., Vernon, W., Matthews, D. & Xuong, N. H. (1981). Characteristics of a flat multiwire area detector for protein crystallography. J. Appl. Cryst. 14, 85–93.
Ludewigt, B., Jaklevic, J., Kipnis, I., Rossington, C. & Spieler, H. (1994). A high-rate, low-noise, X-ray silicon strip detector system. IEEE Trans. Nucl. Sci. 41, 1037–1041.
Milch, J. R., Gruner, S. M. & Reynolds, G. T. (1982). Area detectors capable of recording X-ray diffraction patterns at high count rates. Nucl. Instrum. Methods, 201, 43–52.
Rehak, P., Walton, J., Gatti, E., Longoni, A., Sanpietro, M., Kemmer, J., Dietl, H., Holl, P., Klanner, R., Lutz, G., Wylie, A. & Becker, H. (1986). Progress in semiconductor drift detectors. Nucl. Instrum. Methods Phys. Res. B, 248, 367–378.
Sarvestani, A., Besch, H. J., Junk, M., Meissner, W., Pavel, N., Sauer, N., Stiehler, R., Walenta, A. H. & Menk, R. H. (1998). Gas amplifying hole structures with resistive position encoding: a new concept for a high rate imaging pixel detector. Nucl. Instrum. Methods Phys. Res. A, 419, 444–451.








































to end of page
to top of page