What colour are stars really?

Title: Digital color codes of stars

Authors: Jan-Vincent Harre, René Heller

First Author’s Institution: Institüt fur Astrophysik, Georg-August-Universität Göttingen, Friedrich-Hund-Platz 1, 37077 Göttingen, Germany

Status: Accepted for publication in Astronomische Nachtrichten

What is a colour anyway?

A bunch of photons – particles of light – leave the surface of the Sun. They traverse the distance between the Sun and the Earth in about 8 minutes. After a bumpy ride through the atmosphere they finally hit your eyeball. In the back of your eye, they stimulate the photoreceptive cone cells. The wavelength of the photons will be interpreted by your brain as a colour, and you will see the Sun as being white-ish (probably followed by a spike of pain telling you to look away).

Colour is a “life” thing. The eyes of human beings can perceive particular wavelengths of light, which we perceive as the visible spectrum, and we associate colours with those wavelengths to tell them apart. The reason we, and most other life on earth, can interpret this visible spectrum in particular (and not say, UV or radio waves) is because the Sun emits most strongly in these wavelengths; we have evolved to see Sunlight.

Imagine a theoretical object that can perfectly emit light across all wavelengths. This so-called “black-body” emits light in a spectrum that is determined by its temperature alone. At hotter temperatures a more intense spectrum with more ‘blues’. At lower temperatures, a more shallow spectrum with more ‘reds’ (the human body emits infrared radiation, for example). The emission spectrum of stars is often approximated as a black-body, and as so the colour of stars is often assumed to be that of the peak of a black-body of the temperature of the star.

However, a black-body spectrum does not perfectly describe how a star would appear to us, let alone incorporate effects of absorption from elements in stars, reducing the intensity at specific wavelengths. Today’s authors ask the question: “how can we find out exactly what stars look like to the human eye”?

Figure 1: A synthetic spectrum for a 2500 K, solar-metallicity red dwarf. The black line shows the black-body spectrum for a star of this temperature, and the orange shows the observed spectrum. Because the star’s spectrum is much less intense than the black-body spectrum, we can see that there is a lot of absorption of wavelengths occurring here. The blue, red and green lines show the different CMFs for the human eye. The inset shows the rest of the star’s long, high-wavelength spectrum.

Spectra versus eyes

The authors generated stellar model spectra (using the PHOENIX library) for stars ranging in temperatures from 2300 K to 15,000 K, at metallicities of [Fe/H] = (0, -1, -2), to account for how absorption at different metal abundances changed the spectra. They also generated spectra up to 55,000 K, using a different model (TLUSTY) that accounted for the ‘looser’ outer layers of such massive stars. To mimic the reception of the human eye, they used established colour matching functions (CMFs), which model the reception of the human eye in Red, Green and Blue (RGB) wavelengths, and map this to hexadecimal colour codes. Figure 1 above shows these CMFs over the spectrum of a 2500 K stars — a red dwarf.

Figure 2: The colours of model spectra, as perceived by the human eye, shown on a hexadecimal colour wheel. Left (a): main sequence stars ranging from stellar types M (low temperature) to types O (high temperature). Right (b): a comparison of the colours derived from synthetic spectra (filled circles) to those calculated from black-bodies at the same temperature (open circles), connected by lines. 

Different than expected

The authors calculated the colours for both the synthetic spectra and black-bodies at the same temperature, and found that overall they differed significantly. This is because the absorption occurring in the star’s atmosphere contributes significantly to the final colour, absorbing more of certain wavelengths than of others. Figure 2 above shows the trend of stellar type (ranging from massive O stars to small M dwarfs), and how these differ from black-bodies at the same temperature.

This has interesting consequences for stars that we typically associate with particular colours. They found that stars at high temperatures (~8000 K) were slightly bluer than their black-bodies (because of less absorption of red wavelengths). Conversely cooler ‘red’ dwarfs (~3000 K) were actually more like ‘orange’ dwarfs, with more absorption of red wavelengths occurring than in the black-body. Especially ‘white’ dwarfs, which we tend to think of as white, go through a whole temperature, and thus colour, range as they cool. White dwarfs have relatively small atmospheres and very little absorption, and the cooler stars will actually look orange, not white!

How subjective are these colours?

Just to make sure these colours were as objective as possible, the authors checked some cool phenomena. This included measuring the colour in the case of broadened absorption lines in the cases of extremely fast-spinning stars, and the absorption in the Earth’s atmosphere. In all cases, they found these effects to only cause a 1% change in the RGB colour codes, which would be imperceptible to most humans. Metallicity on the other hand does provide a perceptible difference (You can compare them yourself! Input the Hex codes from Tables 2 and 4 in the paper here).

The results of today’s paper, while seemingly small, have a large significance. For the first time we can look at an image on our screens and know with certainty that it represents a star exactly as we would see it (see Figure 3), connecting us directly to distant stars in the sky. The exact way we perceive colour, and therefore stars, is unique to life raised on Earth by the light of the Sun. Knowing the colour of stars therefore tells us — a little — about our own place in the Universe.

Figure 3: A selection of main sequence stars of different stellar types. The darkening around the edges is due to an effect called ‘limb darkening’, which accounts for the difference in brightness of the star at the edges, as seen by us. For reference, the Sun is a G-type star. The authors note that they did not account for brightness— an O2 size star would be a billion times more luminous than an M9 star. 

About Oliver Hall

Oliver just started a postdoc at ESA ESTEC in the Netherlands, after completing a PhD at the University of Birmingham, UK. His research focuses on asteroseismology, the study of stellar pulsations, and what it can tell us about stellar populations. When not doing research he enjoys playing piano, walking, and not moving from the sofa all weekend with a good book, show, or game.

Leave a Reply