Spectral types and star mass

[Evil Dr Ganymede]

I've been getting a bit confused about the relation between spectral type, star mass, and effective temperature (if there is a relation between them at all) for main sequence stars.

I started off assuming that a star with mass W will have a luminosity of X suns, which means it has an effective temperature of Y Kelvin and a spectral type of Z. So if you knew the mass, you could figure out the spectral type.

However, looking at things like the RECONS near star table, I don't think that's true. A given mass yields a given luminosity according to the mass-luminosity relation, but it seems that the effective temperature isn't necessarily linked to these, and that the spectral type is also not necessarily related to the temperature. There are stars on the RECONS list that have the same mass but vary over quite a range of M types.

Is there a logic to how mass links to spectral type? Obviously low mass stars are M V, more massive ones are K, sun-like stars are G, and so on, but I'm after something that links mass more specifically to the 0-9 numbers. Or is the spectral type just determined by the vagaries of the chemistry in the star's atmosphere?

And is there an 'official' table that shows the relation between temperature and spectral type? Or is this relationship again too vague to quantify like that?

[selden]

Don't forget that the mass doesn't change significantly as stars age, but their temperatures and spectral types do. The relationship is rather complicated.

See http://instruct1.cit.cornell.edu/courses/astro101/java/evolve/evolve.htm for a simulation showing how temperature and luminosity change with age for different stellar masses.

[Evil Dr Ganymede]

Don't forget that the mass doesn't change significantly as stars age, but their temperatures and spectral types do. The relationship is rather complicated.

This I knew - I've tabulated the Geneva grids myself, but the way I linked temperature to spectral type doesn't seem to tally with the RECONS data. I use the following relation:

Code: Select all

Type   Temp/K      Type   Temp/K
A9   7850      B9   12400
A8   8200      B8   13800
A7   8550      B7   15200
A6   8900      B6   16600
A5   9250      B5   18000
A4   9600      B4   19400
A3   9950      B3   20800
A2   10300      B2   22200
A1   10650      B1   23600
A0   11000      B0   25000
            
G9   5100      F9   6150
G8   5200      F8   6300
G7   5300      F7   6450
G6   5400      F6   6600
G5   5500      F5   6750
G4   5600      F4   6900
G3   5700      F3   7050
G2   5800      F2   7200
G1   5900      F1   7350
G0   6000      F0   7500
            
M9   2000      K9   3650
M8   2170      K8   3800
M7   2330      K7   3950
M6   2500      K6   4100
M5   2670      K5   4250
M4   2830      K4   4400
M3   3000      K3   4550
M2   3170      K2   4700
M1   3330      K1   4850
M0   3500      K0   5000


But according to the Geneva grid, a 0.3 solar mass star has a surface temperature of 3500K, which makes it an M0 V - so lower mass stars must be M1, M2, M5 etc. But there are many stars on the RECONS table that are more massive than 0.3 solar masses that are M V stars. And M VI (low metallicity) stars are apparently less than about 0.18 solar masses, not 0.3 solar masses.

So it seems that either my conversion between temperature and type is wrong, or the Geneva grids are...

[selden]

A table of temperatures for V, III and I subtypes that includes IR colors is at http://www.jach.hawaii.edu/JACpublic/UKIRT/astronomy/temp.html (United Kingdom Infra-Red Telescope)

Another list of temperatures vs type, by Gerard van Belle of Keck Operations, is at
http://spider.ipac.caltech.edu/staff/gerard/q_refs/sptype_temp.html

It references Dyck et al. (1996), Allen (1976)

[selden]

The Recons site claims to have used the Mv value and the empirical mass-luminosity relations of Henry and McCarthy (1993) and Henry et al (2000) for their mass estimates.

Have you looked at Henry's papers to see how his relationships were determined? I'd be slightly concerned there might be a bias in them, since he also seems to be the author of that Recons page. (See the tag line at the bottom of that page.)

Remember that without a binary companion, a star's mass can't be measured directly, but only estimated from its other characteristics.

[Evil Dr Ganymede]

Thanks Selden. Though there's different temperature schemes for different star sizes?! The first list is probably more useful, but it doesn't seem to say what happens after M6 V - could a M9 V be get as cool as about 2500K? I'll see if I can check those Henry papers too.

Don't suppose someone can actually explain to me what all those V-K, J-H etc numbers actually mean? Are they magnitudes of some sort?

And what does a type of 'm' mean? G 180-060 on the RECONS list (#88) seems to have a type of 'm+'. 'Ve' means it has extra emission lines, yes? And 'Vp' means it has some peculiar spectral lines?

And how does that white dwarf classification system work?

[granthutchison]

And how does that white dwarf classification system work?

The modern system for white dwarfs starts with a letter D, then adds letters indicating the dominant components of the spectrum: A = hydrogen, B = neutral helium, C = continuous spectrum, O = ionized helium, Z = metal lines, Q = carbon. Other letters are added to indicate additional properties, such as P (polarized), X (unidentified features), H (magnetic), V (variable). Then there's a number indicating the temperature, equal to 50400/T.

So a star with spectrum DBAQ4 has He I, H and C components in its spectrum (with diminishing strength in that order), and a temperature of around 12600K.

(But there's also an old classification that tries to shoehorn the wd spectra into a main-sequence analogue DO, DB, DA, DF, DG, DK, DM format ... so if all the wd spectra in your catalogue are listed with single-letter codes and no numbers, you're seeing old data.)

Grant

[Evil Dr Ganymede]

Thanks, Grant. That clears that up a bit (I was only familiar with the old system).

[granthutchison]

Ah. Just turned this one up while searching for your more recent RECONS post, Consty - I now recall intending to come back with more information.
Spectral classes given in lower case letters, like your "m" example, are deduced from the colour of the star, rather than an observed spectrum. The trailing "+" or "-" indicates that the spectral subclass is likely to be late or early, respectively.
Curiosity made me check back with RECONS for your "V-K, J-H" letters, but I'm not spotting them. Have they disappeared since you posted?

Grant

[Spaceman Spiff]

Just a quick reply, Evil Dr Ganymede.

The V-K, J-H numbers: yes, they are magnitudes, but actually differences between magnitudes in different filtered bands. V is 'visual', and magnitude taken through a V filter gives magnitudes that are supposed to match as closely as possible that seen by the human eye. B is 'blue' and comes from older black and white photo emulsions (which are more blue senesitive than our eyes). The H, I J, K series are filters for IR. It's no co-incidence that you are looking at V-K, J-H when asking about M type stars...

Magnitude differences between two filters are a quick way of measuring many stars' colour temperatures in one go.

I'll have to explain this and the relation to black body temperature, spectral type, and try and answer your spectrum-mass conundrum tomorrow.

Spiff.

[Malenfant]

A table of temperatures for V, III and I subtypes that includes IR colors is at http://www.jach.hawaii.edu/JACpublic/UKIRT/astronomy/temp.html (United Kingdom Infra-Red Telescope)

Sorry for the necropost, but I feel I should point out that this page is now at:
http://www.jach.hawaii.edu/UKIRT/astron ... /temp.html

Kinda useful

[Malenfant]

Actually, I do have a new vaguely related question...

Are there any official definitions for:

1) how big a star has to be before it's classed as a supergiant (size I) instead of a Giant (size III).

2) how luminous a star has to be before it's classed as a Bright Giant (size II) instead of a Giant (size III)

Come to think of it, now I do some research on this I notice that what I've always thought of as "Sizes" are actually usually referred to as "Luminosity Classes" nowadays... but their labels are obviously size related - i.e. "size IV" stars are Subgiants, "size III" stars are Giants, etc. But I've still never actually seen any actual strictly defined definitions of these classifications.

What I want to know is whether a large star with say 1000 Sols luminosity would be classed as a Giant, a Bright Giant, or even a Supergiant. I know it's generally defined by spectral lines, but surely there's got to be some correlation with size and luminosity that can be used to draw dividing lines?!

[selden]

"An atlas of stellar spectra, with an outline of spectral classification" by W.W. Morgan, P. C. Keenan and E. Kellman, 1943

http://adsabs.harvard.edu/abs/1943QB881.M6.......

The MKK classifications are still used today.

[High Dark Templar]

If you want only to "play" with one parameteres and see the efect on others, use the excel file in :

http://usuarios.lycos.es/darktemplars/Index.html

[Malenfant]

"An atlas of stellar spectra, with an outline of spectral classification" by W.W. Morgan, P. C. Keenan and E. Kellman, 1943

http://adsabs.harvard.edu/abs/1943QB881.M6.......

The MKK classifications are still used today.

Couldn't read that there, but I did find the full article here:
http://nedwww.ipac.caltech.edu/level5/A ... tents.html

I'll wade through it and see if the luminosity classes are actually defined anywhere... I don't actually see anything in the document that defines the classes except for this:

The luminosity classes are designated by Roman numerals; stars of class I are the supergiants, while those of class V are, in general, the main sequence. In the case of the B stars the main sequence is defined by stars of classes IV and V. For the stars of types F-K, class IV represents the subgiants and class III the normal giants. Stars of class II are intermediate in luminosity between the supergiants and ordinary giants.

I guess I may be out of luck here. For now I guess I'll assume that Bright Giants have luminosity between 1000 and 9999 Sols, and Supergiants have luminosity above 10000 Sol.

BTW, I have written up a little Fortran program that will tell you the spectral type of a star (using the temperatures found at the http://www.jach.hawaii.edu/UKIRT/astron ... /temp.html link posted earlier) - anyone interested in that?