Sky at Night Magazine

[david48]

I wonder, why don't more stars show colours. They ought to, because the colour of a star, obviously comes from how hot it is.

If it's not very hot, it will have a red colour. Hotter stars will be orange, then yellow, then as they get hotter, going up to green, then blue - then violet, when really hot. Following the well-known colour sequence: Red - Orange - Yellow - Green - Blue - Violet.

So shouldn't there be many different-coloured stars in the night-sky? But if you look up at the stars, that doesn't seem to be the case. The stars mostly lack colour. Obviously, there are a few exceptions. Betelgeuse looks reddish, Arcturus slightly orange, Capella a bit yellow, and Vega has a distinct bluish tinge. But the rest of the stars look uniformly and depressingly white.

Why all this pallid monochrome - where are all the coloured stars?

[arthur dent]

No, I'm afradid that you are not correct when you state the following:

"If it's not very hot, it will have a red colour. Hotter stars will be orange, then yellow, then as they get hotter, going up to green, then blue - then violet, when really hot. Following the well-known colour sequence: Red - Orange - Yellow - Green - Blue - Violet."

As the temperature of the emitting body increases, the energy output also increases and the wavelength of the maximum of the curve moves to shorter wavelengths. This is shown in the following curve - known as "Wein's Displacement Curve". This is for objects emitting 'Black Body Radiation' (which stars do).



As you can see from the graph, for a low temperature body (2000K) most of the radiated energy (the vast majority of the area under the curve) is below the visible (shaded) region. So, this object is radiating infra-red but you can't see it very well (only a dull-red glow at best).

As the temperature increases, the object radiates across the visible spectrum (eg the 3200K curve) - so is emitting white light but again, the vast majority of the radiation is emitted in the infra-red region. A star with this temperature would display a distinctly red-ish colour - eg Betelgeuse with a surface temperature of 3500ºC (3227K).

So, a cold(-ish) object radiates in the infra-red (and you can't really see it). Cold-ish in this context means below around 650ºC (like electric hotplates that are "on" but not "glowing")! Hotter objects glow red hot (giving a dull red glow at around 700ºC). As the emitting object gets hotter, the colour changes to orange, then yellow but then (effectively) to white. This is because of the skew in the curve. So, you don't get any green stars!

Our Sun appears to be white although the peak energy is radiated in the yellow-green part of the spectrum (the Sun's classification is spectral class G2 - yellow dwarf) as the surface temperature is around 5778K (5500ºC).

Below is a spectrograph of the light from our own Sun. As you can see there is a large proportion of red, yellow and blue light with a bit of green and orange, giving the Sun an overall white appearance.



VERY hot stars emit radiation mostly in the ultra-violet region, but there is still a huge amount of energy emitted across the visible region, so they still appear as a white star. They don't disappear because although the most energy is at the short-wavelength (UV part of the em spectrum) the shape of the curve means it is still emitting in the visible. For very cool stars, if the curve is so far to the right that there is very little energy emitted in the visible part of the spectrum and what is is at the red end of the sepectrum.

The spectrograph of a very hot star would have an enormous amount of blue light, hence the blue tinge to some stars - eg Vega - whose surface temperature is 9600K, Rigel - whose surface temperature is around 11,000K and Eta Carinae - whose surface temperature is around 40,000K.

Hope this answers your question.

Art

[david48]

Art, thanks for your excellent and informative reply, which I've read through carefully several times. The graph is especially interesting. It shows the low-temperature curves - 2,000K to 2,400K - quickly leaving the graph's "Visible Region". Without making much impression on it.

This gets at a point that puzzles me - why we don't see more red stars. After all, aren't most stars in the galaxy supposed to be fairly low-temperature, red dwarfs. Radiating in long-wave dull-red light. If this kind of of light is emitted by the great majority of stars, then shouldn't red stars, so to speak, "dominate" the sky. Yet this obviously isn't the case - most stars don't look red. The graph you provided might explain this.

This leads on to a second point, which puzzles me - the lack of blue stars. Blue light is definitely shown by Vega. And you say, by Eta Carinae, though I haven't seen that star. Rigel doesn't look very blue - it may be a contrast effect caused by comparing it with the other bright star in Orion, ie Betelgeuse. Anyway, there seem to be very few really blue stars. Whereas the graph might lead us to expect more.

Also - are there really no green stars. Green lies between yellow and blue. Yellowish stars exist, like Capella. And bluish ones, like Vega. So shouldn't there be greenish stars in between?

[arthur dent]

Hi David

I am puzzled by the following in your post: "Anyway, there seem to be very few really blue stars. Whereas the graph might lead us to expect more". Why do you think the curve would lead you to expect more blue stars? It isn't showing you the different proportions of stars, merely the spread of radiation produced by objects of differing temperature.

As to why, with around 75% of all stars being spectral type M, we don't see most stars in the sky as red - *shrugs*

Anyway, to answer your question about green stars...

Looking at the curve again:



Looking at the bottom curve (for 2,000K). For low temperature stars there is practically no (well very very little) radiation being emitted at the high-energy, blue end of the spectrum. The graph rises gently towards the red-end of the spectrum and continues to rise well into the infra-red. So these low-temperature stars are (at best) red in colour (if you can see them).

If you extrapolate the dotted line to higher temperatures and imagine what the curve will look like, at around 4,000-4,200K - the peak of the curve will be on the lower boundary of the visible spectrum. Again, not so much violet and blue light, but far more towards the orange/red part of the spectrum - so the star would again have a reddish-cast (and would probably look orange).

At even higher temperatures, the dotted line still climbs and moves towards the left but by the time the peak is in the middle of the visible band, you have a fair amount of violet and blue, a bit more green and a fair amount of yellow, orange and red - in other words you have a complete spectrum and so the star appears white. This is shown in the picture below - erroneously stating that the surface temperature of the Sun is 6,000K - it is in fact around 5,780 K (5,505 °C).

Continuing the extrapolation of the dotted line - ever higher and still moving to the left at higher temperatures. Imagine what the curve looks like at very high temperatures, where the peak is at the violet end of the spectrum. The right-hand side of the curve drops steeply, not as steeply as the left-hand side but steep enough. Therefore you have a preponderance of blue/violet light - so the star looks blue.

There are no green stars due to the fact that when the peak of the curve is sitting squarely in the middle of the visible spectrum (as it is with the Sun which peaks in the yellow-green part of the spectrum), you have a lot of blue/violet radiation to one side and a lot of yellow/orange/red radiation the other. The distribution is fairly even across the visible spectrum - even though it is centred around green - and because we have a full mix of colours, we see the star as white.

This curve may help to explain the above (for temps up to 6,000K):



So, in summary:

Very cool stars radiate in the infra-red, we can't see them in the visible = brown dwarfs

Cool (M-type) stars radiate in the red and orange. These are red giants (Antares and Betelgeuse) and Red Dwarf stars (Proxima Centauri) with a surface temerature around 3,500K (3,200ºC). M-type stars are the most common with around three-quarters of all stars being type-M. Orange Dwarfs (K-type) such as Alpha Centauri B and Epsilon Indi. Roughly 1 star in 8 is a type K star (12.5%), surface temperatures are below 5,200K.

A slightly hotter star (Yellow Dwarf, G-type) has a surface temperature aroung 5,200-6,200K and the Sun is a prime example along with Alpha Centauri A, Capella and Tau Ceti. The spectral classification of the Sun is actually G2 (each class - O, B, A, F, G, K, M - being split into 10 smaller sub-classes). Around 1 star in every 12 is a G-type star like the Sun.

Hotter than this we have the F and A-type. F-type stars are yellow-white (examples being Canopus, Dubhe B, Polaris and Procyon) with a surface temperature of around 6,000-7,500K. Around 1 star in 30 is a type-F star. A-type stars are white (examples being Sirius, Deneb, Altair and Vega) and have a surface temperature between 7,500K and 10,000K. Only about 1 star in 160 of the main-sequence stars are spectral type A.

Even hotter are B-type stars with a surface temperature of between 10,000K and 33,000K (examples being the brighter stars of the Pleiades, Algol A, Rigel, Spica and VV Cephei B). These stars are bluish-white in colour. Only about 1 star in 800 of the main-sequence stars are B-type stars.

O-type stars are textremely hot with surface temperatures above 33,000K. These stars are also the most massive stars - the Blue Supergiants and they appear, well, blue. Examples are Zeta Orionis, Zeta Puppis, Lambda Orionis, Delta Orionis and Theta Orionis C. Interstingly, these stars output most of their energy in the ultraviolet range. They are the rarest of all main-sequence stars with only about 1 star in three million being O-type stars. Another interesting fact is that these stars shine with a power output of over a million times that our own Sun!

There is a further spectral class - type W - with surface temperatures in excess of 50,000K.

This Wiki article should help. There is a table under the heading "Harvard Spectral Classification" which shows the stellar classes (O, B, A, F, G, K, M) and their associated colours. A mnemonic to remember these (in decreasing-temperature order) is: Wow! Oh Be A Fine Girl Kiss Me Right Now Sweetie!

http://en.wikipedia.org/wiki/Stellar_classification

I hope this is of some interest and answers your question as to why there are no green stars.

Art

[david48]

Art, thanks for your long, well-written, and information-packed post. It overwhelmed me! I've read and studied it, many times over the last few days. Desperately trying to formulate a sensible response. But I've had to give up. Your grasp of astronomy and physics, is at a far higher level than I can attain.

So I'm just retreating to a very basic level - which is this - shouldn't we see more red stars in the night sky, than we actually do?

[arthur dent]

Hi David,

I have to agree with you. I don't know why, but I think the same - since most stars are M-class (approx 75%) you'd think that we'd see more of 'em. Perhaps the majority are so far away that their apparent magnitude makes them too dim to see - we only see nearby younger and brighter stars - *Art shrugs*

This isn't a convincing argument for me though as when I look at Hubble images (not the deep field images but images of our Milky Way), you certainly don't see 3/4 of the stars a red colour - the vast majority appear white, for example see: http://hubblesite.org/gallery/album/star

Here's a typical picture taken by Hubble:

Nowhere near 75% of these stars appear red. However in this second Hubble image: it has to be said that the majority are probably yellow orange or red. What do you think?

Art

[david48]

Thanks Arthur. I'm not sure how far the colours can be relied on, in images like these. The images must have gone through a lot of processing, between first capture of light by the camera-lens, and final "print-out". Perhaps the colours are an artefact of the processing, or the result of a decision made by a human as to which colours "should" appear.

I'm reminded of the first images produced by the Viking-1 Mars lander, back in 1976. When first publicly released, they showed the Martian sky as blue. This was later amended, to show the sky as pink. Nowadays, it seems to be regarded as brownish.

This demonstrates the uncertainty of relying on instruments, and image-processing, when it comes to colours. The stars nearly all look white, when we use our own eyes. How do we know that the supposed predominance of red stars, isn't some kind of instrumental - or theory-generated - artefact?

[Astro Chav]

Some of the lack of colour is to do with how our eyes perform in low light conditions. Colour is more pronounced in photographs, although most cameras give an approximation (the dreaded Bayer matrix) by having pixels sensitive to red, green and blue light. To be pedantic, we should really use a monochrome camera sensor and keep using loads of narrowband filters at several different wavelengths but we know that this is very labourious and expensive.

Yes you are right about colour processing. I posted an April Fool joke a couple of years ago about the Moon turning blue by using processing tools.

In fact, I deliberately use false colour on many objects to make features more visible. I use green light for "white light" solar photography and red, orange or yellow for hydrogen alpha light. Apart from the blue Moon joke, I never do this to mislead.

[lupuvictor]

What our eyes perceives as a single star is actually a binary star: Sirius A and Sirius B A1V (A) / DA2 (B). These two stars are visible in the video below. Distance that separates the two stars, varies between 8.1 and 31.5 AU.
Telescope Celestron C8" Newtonian, super plossl 20mm, 2x Barlow

http://lupuvictor.blogspot.com/2012/04/sirius-brightest-star-in-sky-video.html

[arthur dent]

Sorry lupuvictor,

I'm puzzled as to the question that your post is answering. Could you explain please?

Art