Stellar spectroscopy before 1860

Fraunhofer first observed stellar spectra in 1814. Using his 2.5-cm aperture theodolite telescope, he found three broad stripes in the spectrum of Sirius [1]. Nine years later, with his 10-cm refractor he described the lines he saw in Sirius, Castor, Pollux, Capella, Betelgeuse and Procyon [2]. The main result from this work was that stars have dark absorption lines in their spectra, yet that the lines present differ from star to star. Sirius, for example, with its three strong lines, was quite dissimilar to sunlight, while Betelgeuse displayed countless lines in its spectrum, some of which corresponded in position to the solar lines (see Chapter 2).

It is perhaps remarkable that the first pioneer to explore line spectra of any source at all systematically should have included stellar spectra in his observations. After Fraunhofer, no significant work was undertaken in stellar spectroscopy for 40 years. It is also surprising that these decades that saw so much activity in solar and laboratory spectroscopy should have seen practically no continuation of the spectroscopic work on stars that Fraunhofer had initiated.

Fraunhofer's 1823 paper describes his objective prism mounted on the 10-cm telescope. One of the few references to stellar spectroscopic observations in the intervening four decades came in 1838 from the Scottish-born German astronomer Johann (John) von Lamont (1805–79), who was then director of the Royal Observatory in Munich [3]. Lamont set up Fraunhofer's apparatus again, and observed spectra of some of the brightest stars.

Introduction to peculiar stars

The study of peculiar stellar spectra has always played an important role in stellar spectroscopy, ever since emissionline stars such as γ Cas, T CrB and the Wolf–Rayet stars were first recognized during the 1860s as being unusual (see Chapter 4).

Thirteen types of peculiar star are discussed in this chapter. In general, the peculiarities are either the presence of emission lines in the spectra, or the presence of absorption lines of abnormal strength or profile. These peculiarities often aggravate attempts to classify stellar spectra in a two-dimensional scheme such as the MK system, described in the previous chapter. This is especially so if the peculiarities are due to photospheric abundance abnormalities which result in either unusually strong or weak absorption lines.

The list of types of star having peculiar spectra is certainly not exhaustive; emphasis is given to discussing those in which the spectral peculiarities predominate or were the first observed. Other peculiarities, such as intrinsic light variability or those arising from the star being a binary, may also be present in many of these types. However a full discussion of the peculiarities in the spectra of all known types of variable or binary star is omitted for want of space.

Carlyle Beals and the Wolf – Rayet classification

By 1928 the problem of how to classify the absorption-line spectra of O stars had been essentially settled by the IAU Commission 29.

Edward Pickering at Harvard College Observatory

So far only occasional reference has been made to the work in stellar spectroscopy at Harvard College Observatory in Cambridge, Massachusetts. However, the developments that took place there from 1885 are so important that they merit a separate chapter. The action took place over a period of four decades from this date, and five actors filled the leading roles. Four of these were women. The part played by Professor Edward C. Pickering in the development of Harvard stellar spectroscopy was, however, the most significant.

Pickering (Fig. 5.1) came from a prominent New England family, and his brother (William Pickering (1858–1938)) was also a physicist and astronomer of some note (he was an assistant professor in astronomy at Harvard from 1887). The fact that Edward Pickering was appointed to a chair as Professor of Physics at the Massachusetts Institute of Technology (MIT) at the age of only 22 shows that his scientific abilities were already manifest in comparative youth. His research at MIT was mainly in the field of optics, although astronomy was also an interest, as he took part in solar eclipse expeditions in both 1869 and 1870.

After nine years as an MIT professor, Pickering was appointed director of Harvard College Observatory, where he took up his duties in February 1877. Apparently the appointment of a physicist provoked some criticism, as several able astronomers were also candidates.

Introduction

The successful analysis of the absorption lines in stellar spectra to obtain the abundances of the different elements became a feasibility for stars of a range of spectral types from the early 1940s. In this chapter the development of the subject from about 1940 to the end of the twentieth century is reviewed. In addition, the developments in stellar atmosphere theory, in instrumental techniques and in laboratory data which made abundance analyses possible are discussed here, from the position reached at the end of Chapter 7.

Several important preliminary problems first had to be settled before substantial progress was possible. Those relating to instrumental techniques and laboratory data are deferred until Section 10.8, and we treat here the problems that arose in the theory which was necessary to interpret stellar spectra. First, the whole issue of stellar temperature measurements from the continuous spectra, or spectral flux gradients, was in a mess in the 1920s. This problem was gradually sorted out during the 1930s decade once it was realized that stars do not really radiate like black bodies. This is the subject of Section 10.2.

Secondly, for the analysis both of the flux gradients and of spectral lines, some sort of model was necessary to describe the structure of stellar atmospheres. The development of model atmospheres from McCrea's first work in 1931 is therefore vital to the overall discussion.

Some early theories of stellar evolution

At the end of the nineteenth century the two main branches of stellar spectroscopy were spectral classification and radial-velocity measurements. The latter department was still in its relative infancy, but classification, thanks mainly to the energy of Pickering at Harvard, was a major activity. Classification had become closely related to theories of stellar evolution and these two aspects could hardly be disentangled in, for example, the classification devised by Lockyer [1], which involved first rising and then falling temperatures of stars over their life cycles, and which had some theoretical support from the work of Jonathan Homer Lane (1819–80) and August Ritter on the gravitational collapse of gaseous spheres.

The classification schemes used by Antonia Maury and Annie Cannon were also implicitly evolutionary theories, but with the direction of evolution being from the ‘earlier’ to ‘later’ spectral types. Rival theories to Lockyer's were then proposed for stellar evolution in the early years of the nineteenth century, with the Harvard spectral types as their basis, notably by Sir Arthur Schuster (1851–1934) in Great Britain [2] and by George Ellery Hale (1868–1938) in the United States [3]. Schuster's scheme involved gravitational collapse and cooling of gaseous masses on the socalled Kelvin–Helmholtz timescale.

Early history of the Doppler effect

An important chapter in the history of astronomical spectroscopy opened on 25 May 1842. On this day Christian Doppler (1803–53) (Fig. 6.1), the professor of mathematics at the University of Prague (then part of Austria), delivered a lecture to the Royal Bohemian Scientific Society entitled ‘Concerning the coloured light of double stars and of some other heavenly bodies’ [1]. By analogy both with sound and waves in the sea, Doppler maintained that light waves undergo a change in frequency of oscillation, and hence of colour, either when the luminous source or the observer is in motion relative to the aether (whose existence was at that time supposed necessary for the transport of light waves). He gave formulae for the frequency change ∆ν when either the source or observer were in motion, and these amounted to a statement of the now familiar equation ∆ν/ν0 = V/c. Here V is the relative speed in the line of sight, c is the speed of light, and ν 0 is the light wave's frequency for sources at rest.

Doppler then made two incorrect assumptions: first, that the radiation from stars was largely confined to the visual region of the spectrum, and secondly that the space motions of the stars were frequently a significant fraction of the speed of light. As a consequence, stars normally appearing white are seen instead as strongly coloured, either violet or red, depending on their approach towards or recession from Earth.

Isaac newton and the composition of sun light

The story of solar, and hence also of astronomical spectroscopy, began in 1666 when the young Isaac Newton (1642–1726) wrote these famous words:

The quotation is from Newton's first paper, which he sent in 1672 to the Royal Society. As he himself notes, the phenomenon was already well known and views on the nature of colour by Descartes, Grimaldi, Hooke and others had already been published. The key feature distinguishing Newton's repeat of the experiment was probably his large distance from the prism to the screen or far wall [2]. Newton tells us he used a wall 22 feet from the prism, so allowing sufficient space for the colours to separate out and to give a clear spectrum. ‘Comparing the length of this coloured spectrum with its breadth, I found it about five times greater; a disproportion so extravagant, that it excited me to a more than ordinary curiosity of examining, from whence it might proceed’ [1].

The first International Astronomical Union meeting in Rome, May 1922

Solar Union in Bonn in 1913 Frank Schlesinger had been able to conclude, as a result of the questionnaire [1] to 28 prominent spectroscopists from the Committee on the Classification of Stellar Spectra, that ‘… the preference for the Draper classification is nearly unanimous, but… the general feeling among investigators is opposed at the present time of any system as a permanent one’ [2]. (The ISU questionnaire of 1911 is discussed in Section 5.10.)

In practice 1913 represents the point when the Harvard system was universally adopted. Nine years later, at the first meeting of the newly formed International Astronomical Union (IAU) in Rome, in May 1922, the earlier temporary acceptance of this system was formally and unanimously approved as permanent. The chairman of the Spectral Classification Committee was then Walter Adams. By this time six volumes of the Henry Draper Catalogue had been published and the classification of nearly a quarter of a million stars had been completed by Annie Cannon six years previously. The adoption of the Harvard system was therefore no longer an issue: the Adams report prescribed that ‘the Draper Classification or “Harvard System” … should be the basis on which any further extensions should be built. Classification on other and different systems should be abandoned permanently’ [3], although it was conceded that in ‘cases of great uncertainty Secchi's types may be employed’.