Spectral Classification of Stars

the division of stars into classes according to the stars’ spectra, particularly according to the relative intensities of spectral lines.

The first attempts at spectral classifications were made in the second half of the 19th century by such astronomers as the Italian A. Secchi and the German H. Vogel. The most successful system proved to be the Harvard classification, which was developed at the turn of the 20th century by the American astronomer A. Cannon. The basic criterion in this classification is the intensity of the atomic spectral lines or molecular bands; at the same time, the energy distribution in the continuous spectrum of the star is to some extent taken into account. The Harvard system is based on empirical data and is a temperature classification reflecting differences in the ionization temperatures of the stellar atmospheres and, to some degree, possible differences in the chemical composition of the stars.

The spectral classes are designated by letters and are arranged in the sequence

which corresponds to decreasing temperature; the branches express differences in chemical composition. The transitions between classes are continuous, and each class is subdivided decimally—for example, B0, B1, B2, . . . , B9, A0, . . . . Each successive class or subdivision of a class is said to be later than the preceding class or subdivision. The spectral classes B-M contain 99 percent of all stars. Stars of classes O, R, N, and S are rare. A description of the spectral classes in terms of their principal features follows.

Class O stars have a surface temperature t ≈ 50,000°–30,000°K. This class contains a small number of very hot stars with a strongly developed ultraviolet portion of the spectrum. Lines of ionized helium are characteristic. Lines of neutral helium and of multiply ionized nitrogen, carbon, and silicon appear in the later subdivisions. Stars with broad emission bands due to neutral and ionized helium atoms and ionized atoms of nitrogen, carbon, and oxygen are also encountered. Such stars are called Wolf-Rayet stars and are designated by the letter W.

For class B stars, t ≈ 30,000°–12,000°K. The spectra of such stars are characterized by the presence of lines of neutral helium and ionized oxygen and nitrogen. Hydrogen lines are visible at BO and increase in intensity throughout the class. On the other hand, the helium lines weaken steadily after B and disappear at B9. The K line of ionized calcium and the line of ionized magnesium with wavelength of 4,481 angstroms (Å) appear from B5 onward.

For class A stars, t ≈ ll,500°–7700°K. The spectra of these stars are dominated by hydrogen lines of the Balmer series, which reach their maximum at A0. Hejium is absent. The intensities of the K line and the line at 4,481 Å increase. The line of neutral calcium at 4,227 Å. appears at A2, and lines of neutral iron appear at A5.

For class F stars, t ≈ 7600°–6100°K. The hydrogen lines are still strong, but numerous lines of metals, both ionized and neutral, are also present. The H and K lines of ionized calcium are very strong. Some lines of iron and ionized titanium merge in low-dispersion spectrograms and form what is called the G band (wavelengths from 4,305 Å to 4,315 Å).

For class G stars, t ≈ 6000°–5000°K. The hydrogen lines no longer stand out among the strong lines of the metals and from G5 onward are weaker than certain iron lines. The H and K lines are very strong. The sun is a G2 star.

For class K stars, t ≈ 4900°–3700°K. The H line, the K line, the line at 4,227 A, and the G band reach their greatest development. Absorption bands of titanium oxide molecules first appear at K5. The continuous spectrum is virtually absent in the near-ultraviolet region beyond the K line.

For class M stars, t ≈ 3600°–2600°K. This class contains red stars with band spectra. The titanium oxide bands are very prominent. Among the atomic lines, only the line at 4,227 A. stands out. The H and K lines are scarcely visible. In the spectra of many class M stars, one or more hydrogen emission lines of the Balmer series appear.

For class R stars, t ≈ 5000°–4000°K. In many regards the spectra of this class resemble the spectra of classes G5-K5, but the absorption bands of molecular carbon and cyanogen are very prominent. In R5 stars, the violet portion of the spectrum is very weak at wavelengths less than 4,240 A.

For class N stars, t ≈ 3000°–2000°K. A further intensification is observed of the absorption bands of molecular carbon and cyanogen, which have a sharp boundary on the redward side. The continuous spectrum is very weak at wavelengths less than 4,400 A. This fact accounts for the red color of the stars. Stars of classes R and N are frequently called carbon stars, or C stars.

For class S stars, t ≈ 3000°–2000°K. These stars are similar to M and N stars with respect to the energy distribution in the continuous spectrum but differ from M and N stars in the presence of zirconium oxide bands and of less prominent bands of yttrium oxide and lanthanum oxide (yttrium and lanthanum are very rare on the earth). The hydrogen lines often take the form of emission lines, as in class M. Titanium oxide bands also appear in classes R,N,andS.

A small number of stars have spectra that do not fit into the sequence described above or that have some special characteristic. Such stars are indicated by the letter “p.” Alternatively, a more precise specification may be given through the use of other letters. The letters “e” denotes the presence of emission lines, which are encountered especially often in the spectra of B and M stars (for example, B2e). Nebulous lines are indicated by “n” (for example, A3n). The letter “s” stands for sharp lines (for example, A3s). Particularly narrow and deep absorption lines are indicated by “c” (for example, cA2). The letter “k” indicates the presence of pronounced lines of interstellar calcium in the spectrum (for example, BOk).

Stars often exhibit changes in their spectral class. For example, emission lines (the e characteristic) sometimes appear and then disappear in the spectra of class B stars. Changes in the luminosity of intrinsically variable stars are accompanied by changes in the stars’ spectral class. The spectrum of a nova undergoes very complicated transformations after the star reaches maximum brightness. The spectra of planetary nebulas have emission lines without a continuous spectrum and are designated by the letter “p.” Complex spectra are sometimes encountered in which the characteristics of two or even three spectral classes are mixed. Examples of designations that may be used for these spectra are G0A2 and G0 + A2. Such spectra often belong to close binary stars.

The use of more precise methods, including spectrophotometric techniques, permitted stars of greater or lesser luminosity to be distinguished within each spectral class. The supergiant stars were found to have narrow, deep absorption lines (the c characteristic). Owing to the low gas pressure in the atmospheres of the giant stars, ionization occurs more easily than in the dwarf stars. At the same temperature, therefore, the lines of ionized atoms are stronger than the lines of neutral atoms in the giants and are weaker than the neutral lines in the dwarfs. The hydrogen lines of the Balmer series are very sensitive to the Stark effect. Because of the high density of electrons in the atmosphere of a dwarf star, the Balmer lines are broadened in the spectra of dwarfs. By contrast, the Balmer lines are very narrow in the spectra of giants. Along with other criteria, these criteria regarding spectral lines permitted the spectra of giants to be distinguished from the spectra of dwarfs. (Giant and dwarf stars of the same spectral class are differentiated by the use of the letters g and d, respectively, before the letter indicating the class.) The same criteria were subsequently used in determining the absolute magnitude of stars on the basis of their spectra.

Knowledge of the absolute magnitudes of stars made possible the determination of the stars’ spectroscopic parallaxes and permitted the development of a two-dimensional spectral classification in which stars are classed not only by temperature but also by absolute magnitude. The most detailed two-dimensional classification was worked out between 1940 and 1943 at the Yerkes Observatory in the USA. The Yerkes system uses, together with the old letter designation for the spectral class, Roman numerals to indicate the luminosity class. These luminosity classes are as follows: Ia, the brightest supergiants; lb, less luminous supergiants; II, bright giants; III, normal giants; IV, subgiants; and V, main-sequence stars. The numerals VI and VII are occasionally used to indicate the subdwarfs (sd) and white dwarfs (wd), respectively. The establishment of a star’s spectral class in the Yerkes system provides a broad characterization of the physical properties of the star’s surface layers. On the basis of these data, the characteristics of the star as a whole, including its internal regions, can be established by theoretical means. The two-dimensional classification of stellar spectra has many advantages over the one-dimensional classification. It is difficult, however, to apply the two-dimensional system to faint stars, whose spectra are usually photographed with an objective prism. Criteria for the two-dimensional classification of faint stars have been developed at the Crimean and Abastumani observatories of the USSR.