2000MNRAS.315..722H

A new Monte Carlo stellar wind radiative-transfer code is presented. The code employs a three-dimensional opacity grid, and fully treats polarization and multiple scattering. Either Mie or Rayleigh scattering phase matrices may be used, and the line-transfer is treated by means of the Solobolev approximation. Variance reduction techniques are employed to increase computational efficiency.

The results of several tests of the code are reported. It is confirmed that no continuum polarization is produced in the spherically symmetric wind case, and that the line profiles computed match those computed using established radiative-transfer codes. The continuum polarization produced by a latitudinally structured low-density wind is found to be in good agreement with that predicted by the single-scattering analytical treatment of Fox, while in the higher density regime the polarizations are consistent with the multiple-scattering code given by Hillier.

Two illustrative applications of the code are described, using the wind parameters of ζ Puppis [O4I(n)f] as the base model. In the first the effect on the line profile of a corotating spiral density enhancement is examined. It is found that the spiral gives line profile variations on the order of 5 per cent, and that it produces an S-wave-like pattern as a function of rotational phase. It is noted that the accelerations described by the spiral wave may mimic those produced by tangentially accelerating wind clumps. The variable polarization produced by the spiral is found to have an amplitude of 0.1 per cent, with two maxima per rotational period in phase with the line emission modulation. The second application investigates the profiles and polarization produced in a clumped wind. Although the parameters of the discrete wind clumps are necessarily arbitrary, it is found that a clumped-wind model reproduces the level of spectroscopic variability found by Eversberg et al. It is shown that the wind emission `bumps' produced in the synthetic spectra often arise from the superposition in velocity space of flux from several spatially discrete wind blobs. Although the two example models may, in combination, reproduce the observed spectroscopic variability of OB supergiants, it appears that the predicted polarimetric variability of these models is too weak to explain the polarimetric observations of OB stars by Hayes and Lupie & Nordsieck.

Finally, a new line polarization effect is described, resulting from line absorption of continuum photons in a rotating wind. The effect has a striking resemblance to the observations of ζ Puppis presented by , and it is possible that the Hα polarization structure observed arises from this effect, rather than by line dilution of a continuum polarization.