Astronomy › Section M · Celestial distances › Topic M.4
The H-R diagram and cosmic distances
Most stars neither pulsate nor lie close enough for parallax. A careful look at a star's spectrum can still reveal its luminosity, and with it the star's distance, by placing it on the H-R diagram.
When there is no variable star
Cepheids and RR Lyrae stars are superb distance markers, as Variable stars as distance markers showed, but they are rare. Many stars and clusters we want to measure contain none. Every star, though, has a spectrum, and the H-R diagram, met in The H-R diagram, links a star's surface temperature to its luminosity for each kind of star. If a spectrum can place a star precisely on that diagram, we can read off its luminosity and compare it with how bright the star looks to find its distance.
Temperature is not enough
A spectrum's overall pattern of lines gives the spectral type, the familiar sequence from hot O stars to cool M stars and the even cooler L, T and Y objects, which reflects surface temperature, as Spectra of stars and brown dwarfs explained. Spectral type alone does not fix luminosity. A G2 star, the same type as the Sun, could be a main-sequence star of about 1 solar luminosity, a giant of roughly 100, or a supergiant of many thousands. On the H-R diagram they share a column but sit at very different heights, so a second clue is needed.
Recall: the luminosity class
The second clue was met in Using spectra to measure radius, composition, and motion. A giant's thin, low-pressure atmosphere gives narrower lines than a dwarf's of the same type, and the Morgan-Keenan system of 1943 records this as a luminosity class: Ia and Ib supergiants, II bright giants, III giants, IV subgiants and V main-sequence dwarfs. The Sun is G2 V, Arcturus about K1.5 III and Betelgeuse about M1-M2 Ia. Type and class together pin a star to one small patch of the H-R diagram, and so to a fairly narrow range of absolute magnitude.
From class to distance
The method has four steps. Take a spectrum and find the spectral type and luminosity class. Read the star's absolute magnitude \(M\) from a calibrated H-R diagram; recall from The brightness of stars that \(M\) is the apparent magnitude a star would have at 10 parsecs (pc). Measure its apparent magnitude \(m\). Then solve the distance-modulus relation for \(d\) in parsecs:
For historical reasons this is often called spectroscopic parallax, although no parallax angle is measured at all. The name records the parallaxes that calibrate it.
Worked example: a Sun-like star
A star's spectrum matches the Sun's exactly: G2 V. Its apparent magnitude is 12.0. The Sun's absolute magnitude is 4.83. How far away is the star? 1. Same type and class, so assume the same luminosity: \(M = 4.83\). 2. Distance modulus: \(m - M = 12.0 - 4.83 = 7.17\). 3. Solve: \(d = 10^{(7.17 + 5)/5} = 10^{2.434} = 272\) pc, about 886 light-years. 4. Check: a modulus of 5 means 100 pc and a modulus of 10 means 1000 pc, and 7.17 lies between them, as 272 pc does. The answer is only as good as the assumption that the two stars are true twins.
Worked example: why the class matters
An orange K0 star has apparent magnitude 9.2. Compare its distance if it is a giant, K0 III with \(M \approx +0.7\), or a main-sequence star, K0 V with \(M \approx +5.9\). 1. Giant: \(m - M = 8.5\), so \(d = 10^{13.5/5} = 10^{2.70} = 501\) pc. 2. Dwarf: \(m - M = 3.3\), so \(d = 10^{8.3/5} = 10^{1.66} = 45.7\) pc. 3. Compare: \(\frac{501}{45.7} = 11.0\). 4. Interpret: the magnitude difference of 5.2 is a luminosity factor of about 120, and distance goes as the square root of luminosity at fixed brightness, so \(\sqrt{120} \approx 11\). Getting the class wrong gives a distance wrong by a factor of 11.
Clusters and main-sequence fitting
Star clusters make the method far more reliable. All members lie at about the same distance, so plotting their apparent magnitudes against colour reproduces the shape of the H-R diagram, shifted downwards by the cluster's distance modulus. Sliding the cluster's main sequence up until it lines up with a calibrated main sequence gives that shift directly, averaged over hundreds of stars instead of one. For the Pleiades, the shift is about 5.6 magnitudes, giving \(d = 10^{(5.6 + 5)/5} \approx 132\) pc, close to Gaia's parallax distance of about 136 pc, roughly 440 light-years.
The real world is messy
Real stars scatter around the ideal lines on the H-R diagram, because they differ in age, composition and rotation, and many are unresolved binaries. A single-star distance from this method is often uncertain by 10 to 25%, sometimes more: a 20% error on 5,000 light-years means anywhere from 4,000 to 6,000 light-years. Dust adds a systematic error. It dims starlight, so the extinction \(A\), in magnitudes, must be included: \(m - M = 5\log_{10} d - 5 + A\). Ignoring just 1 magnitude of extinction makes a star seem \(10^{0.2} = 1.58\) times farther than it is, a problem taken up in Cosmic dust.
How magnitude errors become distance errors
Because the distance modulus is logarithmic, an error in the assumed absolute magnitude becomes a fixed percentage error in distance. Solving the relation shows that an error of \(\Delta M\) magnitudes multiplies the distance by \(10^{\Delta M/5}\). Stars of one type and class typically scatter by a few tenths of a magnitude about the calibrated value: \(10^{0.3/5} = 1.15\) and \(10^{0.5/5} = 1.26\), the source of the 10 to 25% quoted above. Averaging many stars at one distance, such as the members of a cluster, shrinks the random part of the error roughly as \(1/\sqrt{N}\): 100 stars cut a 25% scatter to about 2.5%. An error in the calibration itself never averages away.
The cosmic distance ladder
Each method rests on the one below it, forming the cosmic distance ladder. Parallax, pure geometry, is the foundation. Parallaxes of nearby stars calibrate the H-R diagram, which gives distances to clusters, which in turn helped calibrate Cepheids and RR Lyrae stars; today Gaia also measures many of those directly. With Gaia, parallax reaches thousands of light-years, the H-R diagram method extends to giants in the outer halo a few hundred thousand light-years out, RR Lyrae stars to the nearest galaxies a few million light-years away, and Cepheids beyond 100 million light-years. An error in a low rung spreads to every rung above it.
Common mistakes to avoid
First, using spectral type alone; without the luminosity class a G2 star's luminosity is uncertain by a factor of a thousand or more. Second, thinking spectroscopic parallax measures a parallax angle; it compares luminosity with brightness. Third, getting the sign of the distance modulus backwards: a larger \(m - M\) means a greater distance. Fourth, forgetting dust, which makes stars look farther away. Fifth, treating one star's result as exact when the method's scatter is often 10 to 25%.
Key takeaways
A spectrum gives a star's spectral type (temperature) and, from line widths that depend on atmospheric pressure, its luminosity class: Ia, Ib, II, III, IV or V. Together they fix the star's position on the H-R diagram and so its absolute magnitude. The distance then follows from m − M = 5 log₁₀ d − 5, or with dust, plus A. This is called spectroscopic parallax, though no angle is measured. Main-sequence fitting does the same for whole clusters, such as the Pleiades at about 136 pc. Parallax anchors every rung of the cosmic distance ladder.
Flashcards
Flashcard
Find luminosity from the spectrum, then distance from brightness
Test your understanding
Question 1 of 5
Why is a star's spectral type alone not enough to find its luminosity?
Question 2 of 5
A main-sequence star has absolute magnitude +3.0 and apparent magnitude 13.0. Ignoring dust, how far away is it?
Question 3 of 5
A star's spectrum is classified as K1 III. What does the III tell you?
Question 4 of 5
An astronomer wrongly classifies a K0 III giant as a K0 V star and uses the H-R diagram method. What happens to the distance?
Question 5 of 5
What ultimately calibrates the luminosities used in the H-R diagram method?
