Stellar SpectroscopyNOAO

The Message of Starlight - Introduction

Adapted from the material provided by Research-Based Science Education for Undergraduates (RBSEU), a collaborative program between the University of Alaska Anchorage and Indiana University. RBSEU is funded through a CCLI grant from the National Science Foundation.

The Power of Spectroscopy

Spectroscopy is the analysis of light we observe from an object. It is a measure of the amount of light received at each wavelength. It is a powerful tool in astronomy. In fact, most of what we know in astronomy is a result of spectroscopy: it can reveal the temperature, velocity and composition of an object as well as be used to infer mass, distance and many other pieces of information. Spectroscopy is done at all wavelengths of the electromagnetic spectrum, from radio waves to gamma rays; here we will focus on optical or visible light having wavelengths between 360 and 760 nanometers (nm) - from the deep blue to the far red.

The three basic types of spectra are shown in the Fig. 1: continuous, emission line and absorption line. A continuous spectrum includes all wavelengths of light; that means it shows all the colors of the rainbow (Fig. 1a). It is produced by a dense, opaque object that is hot, either a dense gas (such as the interior of a star) or a liquid or solid (for example, a tungsten filament in a light bulb). In contrast, an emission line spectrum consists of light at only a few wavelengths, i.e., at only a few discrete colors (Fig. 1b). An emission line spectrum can be produced only by a hot, tenuous (low-density) gas. Importantly, the wavelengths of the emission lines depend on the type of gas; e.g., hydrogen gas produces different emission lines than helium. Absorption lines can be best thought of as the opposite of emission lines. While an emission line adds light of a particular wavelength, an absorption line subtracts light of a particular wavelength. Contrary to the production of emission lines, absorption lines are produced by a cool gas. Naturally there must be some light to subtract, so absorption lines can be seen only when superimposed onto a continuum spectrum. Thus, for absorption lines to be seen, cool gas must lie between the viewer and a hot source (Fig. 1c). The cool gas absorbs light from the hot source before it gets to the viewer. Here "hot" and "cool" are relative terms -- the gas must simply be cooler than the continuum source. Also note that a gas emits the same wavelengths of light that it absorbs, and absorbs the same wavelengths of light it emits. Figure 2 gives a slightly different view of the different kinds of spectra.

Fig. 1

Figure 1. The various types of spectra and how they are produced.

Fig. 2

Figure 2. Astronomers plot spectra differently than what is most often seen in textbooks. Spectra are plotted as flux (amount of light received per given area of a detector) as a function of wavelength. The three types of spectra are shown in the top three diagrams of this figure. The bottom spectrum combines the three types, much as they would appear in an actual spectrum of an object.

Since different types of gas produce different patterns and strength of lines, emission and absorption lines are named after the element responsible for the line and the ionization state of the gas. If a gas is heated hot enough its atoms will begin to lose their electrons, either by absorbing photons (particles of light) or by collisions with other particles. When an atom loses one or more of its electrons it is ionized.

Losing electrons changes the wavelengths of the emission and absorption lines produced by the atom, thus it is important to know the ionization state. A Roman numeral suffix indicates the state, where higher numbers indicate higher ionization states; e.g., "Na I" is neutral (non-ionized) sodium, "Ca II" is singly-ionized calcium, etc. In general hotter gases are more highly ionized. Some common lines have special names for historical reasons. Because hydrogen is by far the most common element in the Universe, many of its lines were given special names; e.g., "Ly α" is a strong ultraviolet line which is produced by neutral hydrogen (H I); it is part of the ultraviolet Lyman series of hydrogen lines. In the visible part of the spectrum we find "Hα", "Hβ", "Hγ", etc. These lines are also produced by neutral hydrogen, and are part of we call the Balmer series.

Nomenclature: "Na I" is pronounced "sodium one," "N II" is pronounced "nitrogen two," and "Hα" is pronounced "H alpha" or "hydrogen alpha," etc.

Spectroscopy as an Identification Tool

When looking up at the night sky with thousands of stars overhead it is easy to wonder: How do astronomers know what these points of light are? For example, in the image seen in Fig. 3 there are hundreds of points of light. Most, but not all, are stars within our galaxy; a few are actually galaxies many millions or billions of light years away. Thus, images like this one alone don't tell us much. How then do astronomers know so much about stars? How do they identify them and distinguish them from distant galaxies or other objects? Most often the answer is spectroscopy. In this project, you will study the spectra from a wide range of different types of stars. By analyzing the spectra you will be able not only to figure out an approximate surface temperature (blackbody temperature) but also to classify each star according to its spectral class.

Fig. 3

Figure 3. A small section of the sky can present much more than just stars.


The Continuum Spectrum and Wien's Law

Stars can simply be thought of as hot balls of gas and plasma in space. Their interiors are opaque, very hot and dense, and is surrounded by a thin outer layer of cooler, low-density gas. This extremely thin layer of cooler gas is what we call the star's "atmosphere." Because the interior of a star is opaque, hot, and dense, it produces a continuous spectrum also called it blackbody continuum.

The spectral shape of a blackbody continuum depends on the temperature of the object, as shown in Fig. 4. Interestingly, the shape of the continuum does not depend on the star's composition of elements. The spectra of hot stars (>10,000 K) peak at blue wavelengths, giving them a bluish color. The spectra of cool stars (< 4000 K) peak at red wavelengths, giving them a reddish color. Stars like the sun (~6000 K) peak at greenish-yellow wavelengths, giving them a yellowish color. Cooler objects, such as planets and people, also produce a blackbody continuum, but due to their low temperatures (~300 K) the peak of their spectral continuum is in the infrared. The relationship between an object's temperature and the peak of its spectrum with the wavelength expressed in nanometers (10-9 m) is given by Wien's Law:

Wien's Law eqn. 1

Where T is the temperature of the object in Kelvin and λpeak is the peak wavelength of the continuum, measured in nanometers (this is important to remember).

Fig. 4

Figure 4. Examples of blackbody spectra for temperatures of 5500 K down to 3500 K. Note that the peaks of the spectra occur at shorter and shorter wavelengths as the objects get hotter and hotter.

Nomenclature: Astronomers use the Greek letter λ (pronounced lambda) as a symbol for wavelength. The unit is usually nanometers (or meters) although occasionally the defunct Ångstrom is used.

Spectral Absorption and Emission Lines

A star's continuum spectrum is useful for determining the temperature of the surface of the star, but most of what is known about stars is determined from the many spectral lines seen in their spectrum. Figure 5 gives an example of a spectrum from a star. Here one can see the overall continuum shape plus absorption lines.

A close inspection of a star's spectrum usually will reveal many absorption lines, and for some stars, emission lines as well. These spectral lines can be used to determine an incredible amount of information about the star, including its temperature, composition, size, velocity relative to the Sun, if it has an orbiting companion, and age, along with many other properties. Most of what we know about stars has been determined by the study of their spectral lines.

star spectrum
Figure 5. An actual stellar spectrum.

Spectral Classification

At the end of the 19th century astronomy underwent a revolution with the invention of the objective prism and photographic plates. For the first time astronomers were able to record and analyze the spectra of stars. Spectroscopy revealed that stars show a wide range of different types of spectra, but at the time it was not known why. Astronomers at the Harvard College Observatory obtained spectra for over 20,000 stars in hopes of understanding how each star was related to the others. Annie Jump Cannon, Fig. 6, developed a scheme in which each star was classified based upon the strength of the hydrogen absorption lines in its spectrum. "A-class" stars were those stars which had the strongest hydrogen absorption lines; "B-class" stars had slightly weaker lines, etc. Originally the classification scheme went from A to Q, but over time some of the stars were reclassified and some categories were removed. Through the work of Indian astronomer Meghnad Saha and others it was realized that a primary difference between stars was their temperature, and so the classification scheme was reorganized into "OBAFGKM" based upon temperature, from the hot O stars to the cool M (and most recently L and T) stars.

Several mnemonics have been created to remember this confusing sequence, the most common one being "Oh Be A Fine (Girl/Guy/Geek) Kiss Me Like This." A couple more: "Overt Binary Affairs Form Great Keplerian Marriages." "Only Bad Astronomers Forget Generally Known Mnemonics."

Absorption lines are produced by the thin outer layers of the atmosphere of a star. The density of this gas is too low to produce a continuum spectrum. What kinds of spectral lines are seen strongly depend on the star's temperature. For example, helium is very difficult to ionize, so spectral lines formed by ionized helium (He II) appear in only the hottest stars, the O stars. B stars are hot enough to energize their helium, but are not hot enough to ionize it. Thus B stars have He I lines but do not have He II lines, and A stars do not have any helium lines, not even those of neutral He.

Annie Jump Cannon

Figure 6 . Annie Jump Cannon was one of several women who developed the stellar classification scheme at Harvard College Observatory.

In very hot stars (> 10,000 K) most of the hydrogen gas in the star's atmosphere will be ionized. Since an ionized hydrogen atom has no electron - it is just a single proton - it cannot produce any spectral lines. Thus the hydrogen lines are weak in O stars, as seen in Fig. 7. B, A, and F stars are within the right range of temperatures to energize their hydrogen gas without ionizing it. Thus the hydrogen "Balmer" lines are very strong in these stars. At lower temperatures the hydrogen gas isn't as easily excited, thus the Balmer lines aren't as strong in G and K stars, and are barely present at all in M stars.

hot O star
Figure 7. Spectrum of an O star.

Metals are easier to ionize than hydrogen and helium and therefore don't require such high temperatures, thus spectral lines from ionized metals (e.g., Fe II, Mg II, etc.) are common in stars of moderate temperatures (roughly 5000 to 9000 K). Metals produce many more spectral lines than hydrogen and helium because they have more electrons. In general the cooler the star the more "metal" lines it will have. Ca II lines at the wavelengths of 393.3 and 396.8 nm (known as the "Calcium K and H " lines respectively) are a particularly strong set of lines seen in cooler stars. In F stars and stars cooler than F, the Ca II lines are stronger than the Balmer lines. In the cool G and K stars lines from ionized metals are less abundant and lines from neutral metals are more common. In the very cool M stars, their atmospheres are cool enough to have molecules which produce wide absorption "bands" that are much wider than the atomic spectral lines discussed above. These absorption bands radically alter the shape of the continuum, to the point where it is not even clear what the continuum really looks like. Table 1 lists the spectral properties - colors and surface temperatures - of the different classes. Table 2 lists the spectral line characteristics to look for when determining the spectral class of a star.

Spectral Lines Characteristics
Table 2.

Wikipedia information

Spectral Class

Bright stars
Table 1.

Nomenclature: Curiously, astronomers refer to any element heavier than helium, even all the rest of the inert gases, as "metals." This has driven more than one chemist crazy.

Hint: The most important spectral lines to look for are the hydrogen Balmer lines, the Ca II 393.3 and 396.8 nm lines, the Na I 589.3 nm line, and the G-band 430.0 nm line.


The "strength" of a line is a measure of how much light is absorbed by the line; that is, how big and deep of a dip it has. A strong line will absorb as much as 25% or more of the flux. Weaker lines will absorb less. The strength of a line is easy to estimate: Measure the flux density (how much light the detector is receiving per unit area per wavelength) at the lowest point of the absorption line (fline), then estimate the flux density of the continuum level (fcontinuum) at a point a little ways away from the line, either towards the red or the blue part of the spectrum. The strength of the line (see below for examples) is given by:

eqn. 2 eqn. 2

The strengths of the lines listed in Table 2 are relative to the different classes. For example, the hydrogen lines are present but relatively weak in spectral class O stars. The hydrogen lines become stronger in class B, and are the strongest in class A and F stars. They become weaker in G stars and weakest by spectral class K. In M stars hydrogen lines will either appear as very weak absorption or occasionally as emission lines (if the star is flaring, something M stars tend to do often). See Fig. 8 for colorful depictions of spectra for the spectral classes.

To reflect the fact that there is a continuous range of temperatures among stars, the classes must have subdivisions. Each class is divided into 10 groups, with larger numbers having lower temperatures. For example, an A0 star ("A-zero") lies at the hot end of the A class, with a temperature of 9500 K, while A9 is at the cool end near 7000 K. An A9 star is therefore more like an F0 star than an A0 even though an A9 and F0 are technically different classes. Similarly, a K3 star is hotter than a K7 star, and so on.

Note! The cooler the star, the more "metal" lines present. Thus, the overall spectra for hot (OBA) stars are relatively smooth (except
for the hydrogen lines). The spectra for cooler (FGK) stars become more distorted by metal lines the cooler the star. The M stars have spectra so dominated by TiO molecular bands and metal lines that no continuum is visible.


Figure 8 . Examples of stellar spectra for each spectral class.

41 Cygni


In Fig. 9, 5 lines have been identified as "strong," 4 as "weak," and 3 as "medium" strength. The weak lines are those that are barely visible.

If we pick the 2nd strong line from the left, Ca H, we can estimate its strength by our rough estimate of where the continuum would be around it. The strength is

eqn. 2

or ~ 0.7 (70%). The 1st weak line has a strength of about 0.03 (3%).

See if you can come close to these numbers. An exact match is not necessary.



Figure 9
. A spectrum of 41 Cygni with sample lines of various strengths indicated.