A large fraction of massive stars evolve in interacting binary systems , which dramatically modifies the outcome of stellar evolution .
We investigated the properties of blue supergiants in binary systems and whether they are suitable for extragalactic distance determinations using the flux-weighted gravity luminosity relationship ( FGLR ) .
This is a relationship between the absolute bolometric magnitude M _ { bol } and the spectroscopically determined flux-weighted gravity g _ { F } = g / T ^ { 4 } _ { eff } , where g is the surface gravity and T _ { eff } is the effective temperature .
We computed a grid of binary stellar evolution models with MESA and use the v2.1 BPASS models to examine whether they are compatible with the relatively small scatter shown by the observed relationship .
Our models have initial primary masses of 9 – 30 M _ { \odot } , initial orbital periods of 10 – 2511 days , mass ratio q = 0.9 , and metallicity Z = 0.02 .
We find that the majority of primary stars that produce blue supergiant stages are consistent with the observed FGLR , with a small offset towards brighter bolometric magnitudes .
In between 1 % – 24 % of cases , binary evolution may produce blue supergiants after a mass transfer episode , that lie below the observed FGLR .
A very small number of such stars have been found in extragalactic FGLR studies , suggesting that they may have evolved through binary interaction .
Some models with shorter periods could resemble blue hypergiants and luminous blue variables .
We used CMFGEN radiative transfer models to investigate the effects of unresolved secondaries on diagnostics for T _ { eff } and g , and the biases on the determination of interstellar reddening and M _ { bol } .
We find that the effects are small and within the observed scatter , but could lead to a small overestimate of the luminosity , of T _ { eff } and of g for extreme cases .
We conclude that the observed flux-weighted gravity luminosity relationship can , in principle , be well reproduced by close binary evolution models .
We outline directions for future work , including rotation and binary population synthesis techniques .