Diagrams often use red for hot and blue for cold. In the night sky the relationship is different: reddish stars usually have cooler surface layers than bluish stars. Why does colour offer a clue to temperature—and why does a green-region peak in the Sun’s spectrum not make it green?
A star’s intrinsic colour depends mainly on the temperature of the layers from which its visible light escapes. Hotter stars emit relatively more short-wavelength light, making them look whiter or bluish; cooler stars look reddish or orange. But the colour we or a camera observe also depends on intervening matter, the instrument, and viewing conditions. OpenStax: colours of stars.
A star emits a mixture, not one colour
Imagine a glowing piece of metal. At a lower temperature it may glow a dull red; as it gets hotter, its light becomes brighter and whiter. This is thermal radiation: its distribution across wavelengths depends on temperature. The analogy explains hot matter’s glow; a star is not a metal ball or an ordinary flame.
A spectrum shows how much radiation falls at different wavelengths. In the visible range, shorter wavelengths are associated with blue and violet light, longer ones with red. Beyond that range lie ultraviolet and infrared radiation, which our eyes cannot see.
A useful first model is a blackbody—an ideal absorber and thermal emitter. “Black” does not mean that a hot object cannot shine. Its spectrum is broad: at every temperature, it emits at many wavelengths. Heating changes their proportions and shifts the spectral maximum toward shorter wavelengths. Real stellar atmospheres add absorption lines and other departures from a smooth curve. OpenStax: thermal spectra.
Colour concerns the photosphere, not the core temperature
Most of an ordinary star’s visible light escapes from its photosphere, its visible surface layers. This is not a solid surface. Temperature changes with depth inside a star, so its colour cannot directly reveal the core temperature.
Astronomers use effective temperature for a general description: the temperature of a blackbody that would emit the same total energy flux per unit area. It is a useful stellar property, rather than a claim that every part of the photosphere has the same temperature. For the Sun it is about 5800 K; K stands for kelvins. OpenStax: the Sun’s photosphere and effective temperature, NASA: blackbodies and temperature.
The three panels below do not show measured spectra of particular stars. They are model blackbodies: cooler, near the Sun’s temperature, and hotter. Find the highlighted visible band. At 3000 K the curve rises across it toward longer wavelengths; at 10,000 K it falls. The balance of blue and red therefore changes, although neither part vanishes.
The spectrum stays broad at every temperature
Compare the peak position and the curve’s slope within the visible band, not the heights of different panels.
3,000 K
Cooler photosphere
Relative spectral density Bλ
Wavelength λ, nm
Highlighted band: 400–700 nm
Model peak: 966 nm
Within the visible band, red is relatively stronger than blue.
5,800 K
Near the Sun’s temperature
Relative spectral density Bλ
Wavelength λ, nm
Highlighted band: 400–700 nm
Model peak: 500 nm
There is plenty of light on both the blue and red sides of the peak.
10,000 K
Hotter photosphere
Relative spectral density Bλ
Wavelength λ, nm
Highlighted band: 400–700 nm
Model peak: 290 nm
Within the visible band, blue is relatively stronger than red.
Each curve has separately been scaled to the same maximum. Equal heights in these panels do not mean equal brightness or luminosity. This graph compares shape; a star’s total power also depends on the size of its emitting surface.
Branch · With formulasOptional · you can skip this
Where the peak falls, and what it means
For a blackbody, Wien’s law relates temperature T in kelvins to the peak wavelength λ_(max). Here it applies specifically to energy spectral density per unit wavelength, as in our graph:
λ_(max) = (b)/(T).
Use the rounded constant:
b ≈ 2.898 × (10)^(6) nm ⋅ K.
Separate substitutions give:
T = 3000 K,; λ_(max) ≈ 966 nm.
T = 5800 K,; λ_(max) ≈ 500 nm.
T = 10 000 K,; λ_(max) ≈ 290 nm.
The first peak is infrared, the second visible, and the third ultraviolet. A star with an ultraviolet peak is still visible because its broad spectrum includes visible light. NIST: Wien’s displacement constant.
Plotting spectral density per unit frequency, rather than per unit wavelength, gives a different numerical peak location. These are different ways of dividing the spectrum into intervals. Wien’s law here neither identifies a single “star colour” nor describes human vision.
A peak near green does not make the Sun green
The solar-like model peaks near 500 nm. Yet the curve is broad around its maximum: the Sun emits plenty of blue, green, and red light together. Our vision responds to the mixture across the visible range instead of choosing the graph’s highest point. Without Earth’s atmospheric effects, sunlight would appear white. The label “yellow dwarf” does not mean the Sun emits only yellow light. NASA/ESA SOHO: the Sun’s colour.
Likewise, a reddish star does not emit only red photons, nor a bluish star only blue ones. These names describe the tint of a broad mixture. A rainbow scale mapping wavelength to colour cannot, by itself, turn a thermal-spectrum peak into the correct stellar colour.
Light travels through matter; images undergo processing
Imagine a hot star behind a dust cloud. Interstellar dust usually weakens short-wavelength visible light more strongly, absorbing it or scattering it out of the beam. A different balance of light reaches us, making the star look redder, even though its photosphere has not cooled. Reddening does not mean that the dust added red photons. OpenStax: interstellar reddening.
Earth’s atmosphere also changes the light. Near the horizon it travels through more air; scattering can remove more of the blue component. The same star need not therefore have the same apparent tint high in the sky and close to the horizon. A related mechanism helped us explain the daytime sky’s colour.
A photograph adds detector sensitivity, filters, and the assignment of colours to channels. Ultraviolet images of the Sun, for example, may be displayed in green or blue to distinguish bands and features. Those colours communicate data, but do not show what an eye would see: the radiation itself is invisible to us. NASA/ESA SOHO: colours in ultraviolet images.
Our eyes are not cameras with fixed settings either. Colour differences are harder to distinguish in faint light: different photoreceptors work under different conditions. Faint stars can therefore look nearly colourless even when their spectra differ. National Eye Institute: vision at different light levels.
Branch · A closer lookOptional · you can skip this
How astronomers turn colour into a measurement
A telescope can measure light flux through several known filters. Comparing a bluer and a redder band describes the spectrum more reliably than saying “bluish” or reading a screen pixel’s colour. An astronomical colour index expresses such a comparison as the difference in magnitudes in selected bands.
Even this is an observed colour that dust may have changed. Temperature estimates account for reddening and compare observations with stellar-atmosphere models; spectral lines supply further constraints. Chemical composition and atmospheric properties matter too, so one colour without context is not an exact universal thermometer. OpenStax: measurable colours, dust and reddened hot stars.
Test the conclusion: two images show the same star in red and blue. Does that prove its temperature changed? No. First establish the bands, filters, colour mapping, and observing conditions. Only comparable measurements let us investigate a physical change.
Stellar colour carries temperature information because heating changes the balance within a broad spectrum. Reading that information requires separating the star’s emission from dust, atmosphere, and display choices. Next in the course we will examine how temperature and size together relate to a star’s luminosity.