This chapter describes the spectroscopic techniques that are used in most near-UV, visual, and IR spectrometers. Spectroscopy in this range is particularly important for astrophysics, and instruments for these wavelengths can be constructed using conventional optical elements, such as lenses, prisms, gratings, and normal-incidence mirrors. These instruments share many common features.

Commercially Available Spectrometers

Because optical spectroscopy plays an important role in many different branches of science, medicine, and industry, many manufacturers of optical instrumentation offer commercially produced spectrometers of different types. Most of these devices are not suited for the low light levels from astronomical sources. However, in addition to some instruments that have been designed specifically for astronomical applications (mainly for amateurs), there are commercial “low light level” general-purpose spectrometers on the market, which can be (and are) operated successfully for low-resolution spectroscopy at small telescopes. These commercially available instruments have the advantage of being complete systems, which include high-quality CCD or IR detectors. In many cases they can be conveniently connected to the USB port of a computer for instrument control and data readout. Usually the light from a telescope must be fed to the spectrometer using an optical fiber (see Section 4.5), but some of these spectrometers can be attached directly to a telescope focus. Observers interested in commercially available spectrometers will have no difficulties finding the addresses of potential suppliers by searching for optical spectrometers in the Internet.

So far, we have discussed two basic types of spectroscopic techniques. At high photon energies, the observations were based on the detection of individual photons. Their frequency was determined either by measuring their energy or by measuring their wavelength by means of optical effects. At low (radio) frequencies, the electromagnetic waves were directly recorded, and their frequency distribution was derived using electronic methods. As noted in Chapter 3 (Equation 3.42), under thermal equilibrium conditions, the detection of individual photons requires photon energies hv > kT. Thus, depending on the detector temperature, the transition between photon detection and radio-astronomical methods is expected to take place at FIR or submillimeter wavelengths. In practice, there exists a significant overlap of frequencies at which both types of detection methods can be used, and for which the preferred technique depends on the specific scientific objective. Moreover, in the submillimeter range it is sometimes of advantage to use bolometers, which record light indirectly by measuring the heat that is produced when photons are absorbed. Because of the choice of methods, special techniques have been developed for astronomical observations at these wavelengths, and sometimes combinations of radio and optical techniques are employed. A good example of the diversity of methods used in the FIR/submillimeter range are the three spectroscopic instruments of the Herschel Space Observatory (see Figure 9.4), which (as will be described later) use three different techniques.

The purpose of this chapter is to give a brief introduction to the special methods of the FIR/submillimeter range and to discuss their relative advantages and drawbacks for practical observations.

Although the grating instruments discussed in the preceding chapter dominate present-day astronomical spectroscopy at optical wavelengths, currently there are several other techniques in use at ground-based and space-based observatories. This chapter outlines the principles, the present applications, and the potential of three of these alternative methods.

Fabry-Perot Techniques

Fabry-Perot (FP) devices are based on the interference of light rays that are multiply reflected between partially transmissive mirrors. In astronomy, the main applications of the FP technique are special spectrometers and narrowband interference filters. FP spectroscopy was developed in the final years of the nineteenth century by the French physicists Charles Fabry and Alfred Perot. The technique was soon applied to astronomy. Alfred Perot himself used this method for measuring motions in the solar atmosphere.

The basic principle of the FP technique is outlined in Figure 5.1. The heart of any FP device is a pair of partially transmissive mirrors, separated by the distance l. In the context of FP devices, such a mirror pair is called an etalon (from the French designation of a standard length). In Figure 5.1, the light rays from the telescope are assumed to enter the space between the mirrors at angles γ to the mirror normals from the left (A). As indicated in the figure, the partially transmissive mirrors result in multiple reflections at the two mirror planes. At each reflection a fraction of the light passes through the mirror, while another fraction is returned into the space between the mirrors.

Radio astronomers typically use coherent receivers, which record and analyze the electromagnetic radiation directly. As was pointed out in Section 1.4, this facilitates spectroscopy at radio wavelengths. Making use of electronic frequency filters or of natural resonance effects, receivers can be built that are sensitive to a narrow frequency range only. If such receivers contain electronic components for which the parameters can be varied, the receivers can be “tuned” to specific frequencies. In this case, spectra can be be obtained by tuning narrowband receivers within a certain wavelength range, or by combining receivers that are tuned to different frequencies. As described in Section 1.4, the first radio spectra were obtained in this way, and commercially available radio spectrometers still use this technique. However, if spectra are obtained by tuning a receiver, the different frequencies are measured sequentially and the duty cycle for a given frequency is inversely proportional to the spectral resolution. Therefore, such spectrometers are inefficient and not well suited for the low radiation levels of faint astronomical radio sources. To study the very low signal levels from cosmic sources, more efficient methods must be used, which allow us to record many frequencies simultaneously. Some of these methods are described in the following sections.

The aim of this chapter is to provide an introduction to the technical background of the different types of radio spectrometers at a level that an observer needs to select the optimal instrument for a given task and to assess the potentials and the limitations of the different methods.

The purpose of this book is to provide an introduction to present-day astronomical spectroscopy. Thus, this chapter on the historical development will be restricted to a brief outline of selected milestones that provided the basis for the contemporary techniques and that are helpful for an understanding of the present terminologies and conventions. The reader interested in more details of the historic evolution of astronomical spectroscopy may find an extensive treatment of this topic in two excellent books by John Hearnshaw (1986, 2009). Additional information can be found in older standard works on astronomical spectroscopy, which were published by Hiltner (1964), Carleton (1976), and Meeks (1976). Apart from (still up-to-date) historical sections, these books provide extensive descriptions of methods that have been used in the past, before they were replaced by the more efficient contemporary techniques.

Early Pioneers

Astronomy is known for its long history. Accurate quantitative measurements of stellar positions and motions were already carried out millennia ago. On the other hand, spectroscopy is a relatively new scientific tool. It became important for astronomical research only during the past 200 years. The late discovery of spectroscopy may have been due to the scarcity of natural phenomena in which light is decomposed into its different colors. Moreover, for a long time the known natural spectral effects were not (or not correctly) understood. A prominent example is the rainbow. Reports of rainbows and thoughts about their origin are found in the oldest known written texts, and in most parts of the world almost everybody alive has seen this phenomenon.

This chapter covers astronomical spectroscopy at wavelengths between the UV and the gamma rays, which corresponds to a range of photon energies between about 4 eV and about 1014 eV. Naturally, different techniques are used at the extreme ends of this vast frequency range. However, the methods change continuously with the photon energy, and there exist extended regions in which the different methods overlap. Moreover, some of the basic problems and solutions are common throughout this range. In the literature, the high photon-energy range is often divided into the near ultraviolet (NUV, wavelengths about 200–380 nm), far ultraviolet (FUV, 100–200 nm), extreme ultraviolet (EUV, 10–100 nm), soft X-rays (0.5–10 nm), hard X-rays (2.5 pm–0.5 nm), and gamma rays (<2.5 pm). For convenience, these subdivisions will also be used in this text.

Energetic photons can interact with matter in many different ways. Among the relevant physical processes are the ionization of matter, the photoeffect, Compton scattering, and (at energies > 1 MeV) electron–positron pair production. All these processes can absorb over broad continuum bands. Because of this high absorption probability of energetic photons, optical techniques, which are based on the refraction and reflection of light, either cannot be used at all or require special layouts. The effective absorption cross section decreases again for the very high photon energies of hard X-rays and gamma rays. However, at these high energies, the radiation penetrates the normally reflecting materials, and the refractive index of all materials is uniformly very close to 1.

Astronomical observing projects generally can be divided into three phases: (1) the preparation, (2) the actual observations, and (3) the reduction of the collected data.

At most large ground-based telescopes and at all space facilities, the actual observations are executed automatically by means of a dedicated computer program. The program points the telescope, configures the spectrometer, starts the individual observing steps, and stores the resulting data. The observer initiates the observations by entering the observing parameters (such as the target coordinates, the desired instrument configuration, and the integration times) into a data file, which is uploaded to the computer prior to the beginning of the observations. Only at smaller telescopes, during tests, or at seldom-used instruments are the spectrometers operated manually by the astronomer or by an operator.

All observations result in raw spectra, which, in addition to scientific information, contain artifacts caused by the instruments, by backgrounds, and (in the case of ground-based observations) by the Earth's atmosphere. Therefore, in a final (reduction) step, these artifacts must be removed, and the data must be properly calibrated.

This chapter describes mainly how spectroscopic observations are prepared and how the raw data are reduced. As the execution phase follows in a straightforward way from the preparatory work, less space is devoted to the phase of the actual observations.

Planning and Preparing Observing Runs

General Considerations

The planning phase is the most important part of a successful observing run.

Like all fields of science, astronomy is developing rapidly. This also applies to the methods and techniques used in astronomical spectroscopy. It seems safe to predict that many of the instruments, methods, and procedures that have been discussed here will become obsolete during the next decades; and many techniques that are still experimental or unknown at present may become the standard tools of the future. Some new technical opportunities are expected to become available soon. Others will take longer. Some new methods, which are discussed at present, may never become practical. Nevertheless, when planning new scientific programs and future instruments, such new developments and opportunities must be taken into account.

Present-day major astronomical instrumentation projects typically take years or decades to complete. Some large projects are in progress at present, or in advanced planning stages. In these cases, the future scientific opportunities can be assessed with a fair amount of probability. There are other projects and new ideas whose chances of realization are less certain. Moreover, there may be different opinions concerning the potential of some of these ideas. Therefore, the following sections are necessarily more subjective than the earlier chapters of this book.

Scientific Drivers

Although astronomers do not always fully agree on priorities, a look into the recent relevant publications and reports (such as the 2010 Decadal Report of the U.S. National Academy of Sciences, the justification of NASA's Origins program, and the ASTRONET Science Vision document produced under the auspices of the European Commission) shows a surprising degree of agreement concerning the most important current research topics in astronomy.

As pointed out in the preface, this book is devoted to the observational and technical aspects of astronomical spectroscopy. Thus, a detailed discussion of the physical analysis and the use of astronomical spectra is outside the scope of this work. On the other hand, some details described in the following chapters can be understood only from the special requirements of astronomical applications. Moreover, many astronomical spectrometers have been designed for specific tasks (although they often turned out to be most useful for other applications, and for work on objects that had not yet been discovered when the instruments were planned). Even spectrometers that are designed to cover a large range of scientific problems usually work best for certain tasks and are less efficient, or totally unsuited, for specific other applications. Therefore, in the following sections, some important applications of astronomical spectroscopy are summarized briefly to provide guidelines for the technical aspects that are described in the later chapters.

Spectral Classification

In many natural sciences, the first step toward an understanding of objects or phenomena has been sorting them into classes according to their observed properties. In astronomy, the first classification occurred when early observers of the night sky started to distinguish between planets and stars on the basis of their different motions. As noted in Chapter 1, during his first spectral observations of stars, Joseph Fraunhofer found significant differences between individual stellar spectra. A first crude classification of the stellar spectra (with four classes) was devised in 1860 by G. B. Donati (1826–1873).

This chapter is devoted to the physical effects and methods that are available to measure electromagnetic flux as a function of its frequency or its wavelength. The content of the present chapter will form the basis for the following chapters, in which the technical realization and the practical use of these methods will be described.

In the literature, the term light is often reserved for radiation at visual or optical wavelengths. However, there exists no qualitative difference between the basic properties of electromagnetic waves of different frequencies. Therefore, in the following text, the shorter term “light” will be used for all types of electromagnetic radiation.

Electromagnetic Radiation

That light is composed of electromagnetic waves has been known since the nineteenth century. Along their paths these waves produce variations of an electric and a magnetic field, which are periodic in time and space. A complete treatment of the theory of electromagnetic waves can be found in the textbooks on electrodynamics and on optics. (For a concise introduction to the subject see, e.g., Chapter 2 of Wilson et al. (2009).) In the present chapter, some of the optical effects and important relations that are essential for practical spectroscopy will be summarized briefly.

Monochromatic Plane Waves

We first consider the special case of a monochromatic (or single-frequency) plane light wave propagating in a nonconducting medium in the positive x direction. Because the electric and magnetic fields are coupled, a light wave can be characterized by either its electric or its magnetic component. In the following text we discuss the electric field only.

With the exception of a few objects that have been successfully identified as sources of highly energetic charged particles or of neutrinos, all our knowledge about the universe outside the inner solar system is based on the analysis of electromagnetic radiation. Some valuable information has been derived by measuring the flux, the time variations, or the polarization of astronomical radiation sources. By far the most important tool for investigating cosmic objects, however, has been the analysis of their energy distributions and of their line spectra. There are obvious reasons for this predominance of spectroscopic methods in modern astronomy. First, spectra contain a particularly large amount of physical information. If properly analyzed, spectra allow us to determine the chemical composition, local physical conditions, kinematics, and presence and strength of local physical fields. Second, apart from the cosmological redshift and the reduced observed total flux of faraway objects, spectra are independent of the distance, making spectroscopy a particularly valuable remote-sensing tool. Finally, there exists a well-developed theory of the formation of continua and line spectra.

The gathering of information on distant objects by means of spectral observations requires several steps. First, suitable instruments must be designed that allow us to measure the spectra of the faint astronomical sources. Then, these instruments must be employed to obtain spectra of optimal quality. Finally, the spectra must be analyzed and physical information on the observed objects must be extracted.