Preliminary remark. The numbers that are given in this appendix should not be considered as reference values. Although the present authors have tried to provide numbers that are as close as possible to what is estimated to be the best values, they do not guarantee either that they are the best possible, or that all are mutually consistent. In several cases, authorities that publish values of astronomical constants do not agree, and the choice of one or the other is necessarily subjective. Furthermore, as time goes on, better values will become available. For these reasons, we do not associate uncertainties to values, and one should consider these lists as providing orders of magnitude of the parameters and not as a basis for accurate and dependable calculations.

IAU system of astronomical constants, best estimates

SI units

The units meter (m), kilogram (kg), and second (s) are the units of length, mass and time in the International System of Units (SI).

Astronomical units

The astronomical unit of time is a time interval of one day (D) of 86 400 seconds. An interval of 36 525 days is one Julian century.

The astronomical unit of mass is the mass of the Sun (S).

We now turn to thermal detectors, the second major class of detector listed in Chapter 1. Unlike all detector types described so far, these devices do not detect photons by the direct excitation of charge carriers. They instead absorb the photons and convert their energy to heat, which is detected by a very sensitive thermometer. The energy that the photons deposit is important to this process; the wavelength is irrelevant, that is, the detector responds identically to signals at any wavelength so long as the number of photons in the signal is adjusted to keep the absorbed energy the same. Thus, the wavelength dependence of responsivity is flat and as broad as the photon-absorbing material will allow. Because the absorber is decoupled from the detection process, it can be optimized fully, and quantum efficiencies can be as high as 90–100%. Bolometers based on semiconductor or superconductor temperature sensors are the most highly developed form of thermal detector for low light levels and are the detector of choice for many applications, especially in the submillimeter spectral range. They are also used as energy-resolving X-ray detectors. For the highest possible performance, such detectors need to be cooled to below 1 K. Bolometers manufactured by etching miniature structures in silicon and silicon nitride provide new possibilities for very high performance with large pixels and also for detector arrays.

The general principles derived in Chapter 10 are equally valid for submillimeter- and millimeter-wave receivers. Performance attributes that limit the general usefulness of infrared heterodyne receivers, such as limited bandpass and diffraction-limited throughput, cease to be serious limitations as the wavelength of operation increases. Heterodyne receivers are therefore the preferred approach for high-resolution spectroscopy in the submillimeter spectral region, and their usefulness is expanded as the wavelength increases into the millimeter regime. At wavelengths longer than a few millimeters, they are used to the exclusion of all other kinds of detectors.

Basic operation

The operational principles of heterodyne receivers were described in Section 10.1, and the operation of the components that follow the mixer in a submillimeteror millimeter-wave receiver is essentially identical to the systems discussed in Chapter 10. Such components can be used for amplification, frequency conversion, and detection. Often, much of the expense in a heterodyne receiver system is in the “backend” spectrometer (for example, filter bank or autocorrelator) and in other equipment that processes the IF signal. Because these items can be identical from one system to another, it is common to use a single set of them as back ends with many different receiver “front ends” that together can operate over a broad range of signal frequencies.

This book provides a comprehensive overview of the important technologies for photon detection from the millimeter-wave through the ultraviolet spectral regions. The reader should gain a good understanding of the similarities and contrasts, the strengths and weaknesses of the multitude of approaches that have been developed over a century of effort to improve our ability to sense photons. The emphasis is always upon the methods of operation and physical limits to detector performance. Brief mention is sometimes made of the currently achieved performance levels, but only to place the broader physical principles in a practical context.

Writing is a process of successive approximations toward poorly defined goals. A second edition not only brings a book up to date, it also allows reconsideration of the goals and permits a new series of approximations toward them. Specific goals for this edition are to:

• Provide a bridge from general physics into the methods used for photon detection;

• Guide readers into more detailed and technical treatments of individual topics;

• Give a broad overview of the subject;

• Make the book accessible to the widest possible audience.

Based on the extensive survey of the literature that accompanied preparation of this edition, these goals have led to a unique book.

To be useful, the output signal from any of the detectors we have discussed must be processed by external electronics. Conventional electronics, however, are not very well suited for an infinitesimal current emerging from a device with virtually infinite impedance. Nonetheless, highly optimized circuit elements have been developed to receive this type of signal and amplify it. Most of these devices are based on very high input impedance, low-noise amplifiers that can be built with field effect transistors (FETs). FETs are used in a variety of circuits that are constructed to give the desired frequency response and to accommodate the electrical properties and operating temperature of the detector, among other considerations. In the most sensitive circuits, signals of only a few electrons can be sensed reliably.

Building blocks

There are two basic kinds of transistor out of which amplifiers for the detector outputs could be built: bipolar junction transistors (BJTs) and field effect transistors (FETs). BJTs are generally unsuitable for directly receiving the signal from the detectors we have been discussing because they have relatively modest input impedances. FETs are used to build first-stage electronics for virtually all high-sensitivity detectors.

There are two basic classes of FET: the junction field-effect transistor (JFET) and the metal—oxide—semiconductor field-effect transistor (MOSFET).

In Chapter 1, we listed three fundamental types of photon detector, and a variety of characteristics that would help define the applications for which a detector is suited. Since then, we have introduced a vast profusion of detectors with a chaotic variety of performance characteristics. We will now return to the basic detector characteristics to examine them in the light of the potential applications of the detectors we have discussed. In comparing different detector systems, a useful figure of merit is the speed, that is, the inverse of the time required for a system to make a given measurement. In selecting a detector system, the general considerations discussed below can be combined with the measurement requirements for a given situation and the characteristics of competing detector systems to estimate relative speeds of these systems, leading to selection of an optimum approach.

Quantum efficiency and noise

Two regimes must be distinguished in discussing the effect of quantum efficiency in choosing a detector: (1) photon noise limited and (2) all other cases. In the first regime, the speed of any incoherent detector is proportional to the detective quantum efficiency. Consequently, detectors with very high DQEs such as bolometers and photodiodes are favored for photon-noise-limited applications.

Coherent receivers are the third and last general category of detector listed in Section 1.2. These devices mix the electromagnetic field of the incoming photons with a local oscillating field to produce a signal at the difference, or beat, frequency. Unlike the output from the incoherent detectors discussed so far, this signal directly encodes the spectrum of the incoming signal over a range of input frequencies and also retains information about the phase of the incoming wavefront. As a result, these receivers are easily adapted for spectroscopy; in addition, their outputs can be combined to reconstruct the incoming wavefront, making interferometry between different receivers possible. A dramatic application of this latter capability is the use of intercontinental baseline radio interferometers to achieve milliarcsecond resolution in astronomy. Very weak signals can be detected in part because mixing the local field with the incoming photons allows the signal to be amplified in some situations. However, the main advantage of the technique is that the signals are downconverted to frequencies where extremely low noise electronics can be used to amplify them. Coherent receivers monopolize radio applications; they are not used as widely at infrared or optical frequencies because of their narrow spectral bandwidths, small fields of view, inability to be constructed in simple, large-format spatial arrays, and fundamental noise limitations that arise from the Heisenberg uncertainty principle (the “quantum limit”).

Intrinsic photoconductors are the most basic kind of electronic detector. They function by absorbing a photon whose energy is greater than that of the bandgap energy of the semiconductor material. The energy of the photon breaks a bond and lifts an electron into the conduction band, creating an electron/hole pair that can migrate through the material and conduct a measurable electric current. Detectors operating on this principle can be made in large arrays, and they have good uniformity and quantum efficiency. They are the basic component of CCDs (charge coupled devices), which are widely used two-dimensional, low light level, electronic detectors in the visible and very near infrared, also used to detect X-ray and ultraviolet photons. Photoconductors made of semiconductor compounds with small bandgaps are also useful as high-speed detectors over the 1–25 μm range.

These devices also illustrate in a general way the operation of all electronic photodetectors. All such detectors have a region with few free charge carriers and hence high resistance; an electric field is maintained across this region. Photons are absorbed in semiconductor material and produce free charge carriers, which are driven across the high-resistance region by the field. Detection occurs by sensing the resulting current.

Photography is based on chemical changes that are initiated by the creation of a conduction electron when a photon is absorbed in certain kinds of semiconductor. These changes are amplified by chemical processing until a visible image of the illumination pattern has been produced. Compared with other modern detectors, the quantum efficiency achievable with photography is low, about 1–5%. Photography also suffers from comparatively poor linearity and, in some applications, limited dynamic range. It remains, however, the leader in pixel quantity; an 8 by 10 inch (200 × 250 mm) plate can have 1011-1012 grains, providing some 109 potential picture elements. In addition, photographic materials are inexpensive, provide efficient information storage, and, if treated appropriately, are stable for long periods of time. For these reasons, photography has often been the best detection method in the X-ray, ultraviolet, visible, and very near infrared spectral regions, in spite of the performance improvements attained by individual pixels in electronic detectors. Despite increasing application of electronic detectors, photography remains in wide use. In addition, as virtually the only detector array type for many years, a thorough description of photography is warranted to understand many historical measurements.

Basic operation

Photography was invented and developed through a long series of experiments before solid state physics was understood (as described by Newhall 1983).

A photodiode is based on a junction between two oppositely doped zones in a sample of semiconductor. These adjacent zones create a region depleted of charge carriers, producing a high impedance. In silicon and germanium, this arrangement permits construction of detectors that operate at high sensitivity even at room temperature. In semiconductors whose bandgaps permit intrinsic operation in the 1–15 μm region, a junction is often necessary to achieve good performance at any temperature. Because these detectors operate through intrinsic rather than extrinsic absorption, they can achieve high quantum efficiency in small volumes. However, high performance photodiodes are not available at wavelengths longer than about 15 μm because of the lack of high-quality intrinsic semiconductors with extremely small bandgaps. Standard techniques of semiconductor device fabrication allow photodiodes to be constructed in arrays with many thousands, even millions, of pixels. Photodiodes are usually the detectors of choice for 1–6 μm and are often useful not only at longer infrared wavelengths but also in the visible and near ultraviolet.

Other detectors use different types of junctions. Schottky diodes are based on the interface between a semiconductor and a metal. Quantum wells are the foundation for another class of detector — they are formed when thin layers of different, but similar, crystals are grown on top of one another.

As described in the preceding chapter, the spectral response of intrinsic photoconductors is limited to photons that have energies equal to or exceeding the bandgap energy of the detector material. For the high-quality semiconductors silicon and germanium, these energies correspond to maximum wavelengths of 1.1 μm and 1.8 μm, respectively. A number of semiconductor compounds have smaller bandgaps. They are often used as intrinsic photoconductors with rapid time response; however, they are generally unable to achieve the extremely high impedances needed to reduce Johnson noise adequately for very low light levels. In addition, the performance of intrinsic photoconductors degrades rapidly as the wavelength extends beyond about 15 μm due, for example, to poor stability of the materials, difficulties in achieving high uniformity in material properties, and problems in making good electrical contacts to them. Detector operation far into the infrared must therefore be based on some other mechanism. The addition of impurities to a semiconductor allows conductivity to be induced by freeing the impurity-based charge carriers. This extrinsic process can be triggered by lower energies than are required for intrinsic photoconductivity, enabling response at long infrared wavelengths. Because the lower excitation energies also allow for large thermally excited dark currents, these detectors must be operated at low temperatures.

Photoemission refers to a process in which the absorption of a photon by a sample of material results in ejection of an electron. If the electron can be captured, this process can be used to detect light. Photoemissive detectors use electric or magnetic fields or both to accelerate the ejected electron into an amplifier. At the amplifier output, the photon stream can be detected as a current or even as an individual particle. These detectors are capable of very high time resolution (up to 10-9 s) even with sensitive areas several centimeters in diameter. They can also provide excellent spatial resolution either with electronic readouts or by displaying amplified versions of the input light pattern on their output screens. They have moderately good quantum efficiencies of 10–40% in the visible and near infrared; in some cases, significantly higher values apply in the ultraviolet. They are unmatched in sensitivity at room temperature or with modest cooling, leading to many important applications. In addition, they provide unequalled performance in the ultraviolet. They can be readily manufactured with 106 or more pixels. If the photon arrival rate is low enough that they can distinguish individual photons, the detectors are extremely linear.

General description

A photoemissive detector is basically a vacuum tube analog of a photodiode; in fact, the simplest form of such a detector is a vacuum photodiode, illustrated in Figure 7.1.