ABSTRACT

We present a detailed study of AV 83, an O7 Iaf+ supergiant, and AV 69 [OC7.5 III((f))] in the SMC. The stars have similar effective temperatures and luminosities but show very different wind signatures. For our study we have used the non-LTE line-blanketed atmosphere code developed by Hillier and Miller, which explicitly allows for line blanketing by C, N, O, S, Ar, Ne, Fe, and other elements. Our study finds that AV 83 has an effective temperature of approximately 33,000 K and . It has an extended photosphere as a result of a “low” effective surface gravity and a much denser wind than main-sequence O stars. We can match the spectrum only by using a slow velocity law with , a value that is much larger than the values of around 1 predicted by standard radiation wind theory. Further, we show that the Hα emission profile in AV 83 is sensitive to the adopted surface gravity. To fit the spectrum of AV 83, we have considered conventional models in which the wind is smooth and alternate models in which the winds are highly clumped. Both types of winds yield a satisfactory fit to the majority of lines in the observed spectrum; however, strong UV photospheric lines and the P v resonance transitions favor a clumped wind. If clumping is important, it must begin at relatively low velocities (i.e., 30 km s , not 300 km s ). In the smooth wind, the line force is too small to drive the wind. In the clumped wind, the line force is generally sufficient to drive the wind, although there are still some discrepancies around the sonic point. In AV 83, the N abundance is substantially enhanced relative to normal SMC abundances, while both C and O are SMC-like, consistent with the presence of internally processed CNO material at the stellar surface. The N iii multiplet, which is known to be produced by dielectronic recombination, is well reproduced by the models. These lines, and the adjacent C iii multiplet, show a significant sensitivity to surface gravity, as well as the usual sensitivity to abundance and effective temperature. Incoherent electron scattering, occurring within the photosphere, can explain the broad wings seen on these lines. We have modeled the Fe spectrum (Fe iv–Fe vi) in the UV in both AV 83 and AV 69. For stars with an effective temperature around 33,000 K, the Fe iv–to–Fe v line ratios form a useful effective temperature diagnostic and give results consistent with those found from optical and UV line diagnostics. The derived iron abundance, which is sensitive to the adopted microturbulent velocity, is 0.2–0.4 times the solar iron abundance in AV 83, while 0.2 solar gives a good fit for AV 69. The wind of AV 69 is substantially less dense than that of AV 83. Because of the lack of suitable diagnostics, it is impossible to constrain the mass-loss rate and velocity law independently. Its spectrum indicates that it has a similar effective temperature to AV 83 ( K), a substantially higher gravity ( ) than AV 83, and a CNO abundance pattern that has not been influenced by internal CNO processing. We show that the N/C abundance ratio is substantially below solar, in agreement with SMC nebular and stellar abundance studies. The differences between the spectra of AV 83 and AV 69, and between the derived masses and surface abundances, are striking. We have examined possible causes, and only one seems consistent with the observations and our current understanding of massive star evolution. AV 83 was most likely a fast rotator that experienced rotationally enhanced mass loss. The presence of enhanced N but almost normal C and O abundances is a direct indication of rotationally induced mixing. On the other hand, AV 69 is a slow rotator. As part of our analyses, we have systematically examined the influence of the H/He abundance ratio, the mass-loss rate, the velocity law, the Fe abundance, microturbulence, and clumping on the theoretical spectrum. We illustrate which lines provide useful diagnostics and highlight some of the difficulties associated with spectroscopic analyses of O stars. The spectrum of AV 83 shows the presence of photospheric absorption lines, the presence of lines formed at the base of the wind, and numerous wind lines. Since these lines sample the photosphere and the entire wind, extreme O If supergiants, such as AV 83, are ideal candidates to probe conditions in stellar winds and hence further our knowledge of O star winds.