Type II (core-collapse) supernovae occur shortly after star formation and enrich a galaxy with $\alpha$ elements such as O, C, NE, Mg, Ca and Si. On the other hand, Type Ia supernovae occur on a more delayed timescale and produce mostly Fe and heavier elements and negligible amounts of $\alpha$ elements. There is great interest in constraining the [$\alpha$/Fe] ratio of galaxies since that tells you something about the rate at which stars formed in the galaxy as a function of cosmic age.

I have usually seen studies rely on measuring abundances of Mg, Si and Ca relative to Fe, but not oxygen. For reference, the Wikipedia page for the alpha process says

As for oxygen, some authors consider it an alpha element, while others do not. Oxygen is surely an alpha element in low-metallicity Population II stars stars. It is produced in Type II supernovas and its enhancement is well correlated with an enhancement of other alpha process elements.

Why would oxygen not be an $\alpha$ element in high-metallicity environments even though it is produced by Type II supernovae?

Do studies prefer [Mg/Fe], [Ca/Fe] and [Si/Fe] as a proxy for [$\alpha$/Fe] instead of [O/Fe] for an astrophysical reason, or just because the spectral features of Mg, Ca and Si are easier to measure than O?

  • 1
    $\begingroup$ Re "Type Ia supernovae occur on a more delayed timescale and produce mostly Fe and heavier elements" (my emphasis): Shouldn't "produce mostly" be "also produce" (or similar)? From Type Ia supernova (my emphasis): "Near the time of maximal luminosity, the spectrum contains lines of intermediate-mass elements from oxygen to calcium; these are the main constituents of the outer layers of the star.". Perhaps quantify the ratio between the elements? $\endgroup$ Jul 22, 2021 at 13:44

1 Answer 1


Two things.

  1. The abundance of oxygen is a difficult thing to measure in optical spectra - much harder than Mg, Ca, and Si. So these latter are usually used to represent "the alpha elements". There is a strong OI triplet at 777 nm and a much weaker OI forbidden line at 630 nm. But these often give contrary results because they are blended with other lines and suffer from uncertain NLTE effects (e.g. Ting et al. 2018). Mg, Ca and Si in contrast have many easily measured and well-understood absorption lines in stellar spectra.

  2. Oxygen isn't only produced in massive stars and spewed out in supernovae (though that is probably the major cause of early galactic enrichment - e.g. Meyer et al. 2008). Oxygen is also formed in the He-burning zones of lower mass stars via the capture of a He nucleus by a $^{12}$C nucleus. Helium burning eventually happens in all stars that have left the main sequence in our galaxy. In particular, asymptotic giant branch stars of mass $2<M/M_\odot<8$ will never explode as supernovae, yet are much more numerous than high-mass stars and can distribute some of the products of their internal He-burning into the interstellar medium via deep convective mixing, pulsations and winds - the so-called third dredge-up (which brings up more carbon than oxygen actually). But this will happen later and in more metal-rich environments because the lifetimes of lower-mass stars are much longer. Thus even though oxygen is produced by alpha capture it may not behave like a typical "alpha element" in that it may show continuing enrichment beyond that produced by an initial burst of massive star formation in metal-poor populations. The oxygen abundance of the interstellar medium (and the stars born there-in) could continue to rise more slowly thanks to the contributions of numerous longer-lived AGB stars, not just the rarer supernovae. This complication is absent for Mg, Ca and Si because these are not produced in lower mass AGB stars.

Having said that, my understanding is that most authors treat oxygen as an alpha element and the contribution from AGB stars is not sufficient to stop [O/Fe] behaving very much like [Mg/Fe], [Si/Fe] and [Ca/Fe] in stellar populations (albeit with more noise, because it is difficult to measure). The plot below is from a well-cited paper by Bensby et al. (2014), and clearly shows that [O/Fe] behaves like the other alpha elements, although there is perhaps signs of a continuing decline for [Fe/H]>0 that is weaker in the other alpha elements.

Alpha/Fe vs [Fe/H] from Bensby et al. (2014)

  • 1
    $\begingroup$ Wow, thank you for the super clear and fast response! Regarding your first point, for measuring stellar oxygen abundance, I get that we are limited by the rarity and noise of oxygen absorption lines. But what about measuring oxygen abundance for the interstellar gas? It seems that is routinely done because you can easily detect oxygen ion emission lines in the optical (3727AA, 5007AA, etc.). Are there difficulties with constraining and interpreting [alpha/Fe] of interstellar gas given so many prominent gas emission lines? $\endgroup$ Jul 22, 2021 at 2:00
  • 2
    $\begingroup$ @quantumflash I think measuring Fe in the ISM is difficult. But what would you be trying to show? The ISM is well-mixed so you don't see how [O/Fe] has evolved; you see it as it is now. The point about measuring [O/Fe] in stars, is that it is a fossil record of the ISM when they were born. $\endgroup$
    – ProfRob
    Jul 22, 2021 at 6:52
  • 1
    $\begingroup$ Amusingly, I have just received the "Stellar Astrophysics" bronze tag badge for this! $\endgroup$
    – ProfRob
    Jul 22, 2021 at 6:54

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .