Because the old 1996-era PDF files were starting to show their age, I reformatted my dissertation into a new version in 2020, and added a few footnotes throughout the document to refer to more recent follow-on research. The new PDF is here.
This thesis was defended on August 2, 1996 at the University of Delaware. This work is Copyright (c) Steven R. Cranmer, 1996. The author grants the right for individual copies of these files to be made for personal use, provided they remain unmodified, original authorship is retained, and they not be used in any widely-distributed or commercial publication.
For posterity, the old PDFs for individual chapters are still listed below. Following that is the text of the abstract.
| Title and Approval Pages, Contents, Abstract | [PDF] |
| Chapter 1: Introduction | [PDF] |
| Chapter 2: The Theory of Radiatively Driven Stellar Winds | [PDF] |
| Chapter 3: Synthetic Observational Diagnostics | [PDF] |
| Chapter 4: Rapid Stellar Rotation: Centrifugal Effects | [PDF] |
| Chapter 5: Rapid Stellar Rotation: Wind Compressed Disks | [PDF] |
| Chapter 6: Dynamical Models of Corotating Wind Structure | [PDF] |
| Chapter 7: Pulsations, Waves, and Discontinuities in Stellar Winds | [PDF] |
| Chapter 8: Summary and Conclusions | [PDF] |
| References | [PDF] |
Title:
Dynamical Models of Winds from Rotating Hot Stars
Thesis Abstract:
The hottest and most massive stars (spectral types O, B, Wolf-Rayet) have strong stellar winds that are believed to be driven by line scattering of the star's continuum radiation field. The atmospheres and winds of many hot stars exhibit the effects of rapid rotation, pulsation, and possibly surface magnetic fields, inferred from observations of ultraviolet spectral lines and polarization. The complex time variability in these observations is not yet well understood. The purpose of this dissertation is to model the dynamics of winds around rotating hot stars and synthesize theoretical observational diagnostics to compare with actual data.
Before dealing with rotation, however, we derive the theory of radiative driving of stellar winds, and uncover several new useful aspects of the theory for spherical, nonrotating stars. The presence of limb darkening of the stellar radiation is found to be able to increase the mass flux by 10-15% over standard models assuming a uniformly-bright star, and the wind's asymptotic terminal velocity should decrease by the same amount. We also introduce a new approximation method for estimating the terminal velocity, which is both conceptually simpler and more physically transparent than existing approximation algorithms. Finally, from theoretical line profile modeling we find that observational determinations of the terminal speed may be underestimated by several hundred km/s if unsaturated P Cygni lines are used.
Rotation affects a star by introducing centrifugal and Coriolis forces, decreasing the effective gravity and making the star oblate. This in turn redistributes the emerging radiative flux to preferentially heat the stellar poles, an effect known as gravity darkening. Although previous models have computed the increase in equatorial mass flux due to the lower effective gravity there, none have incorporated gravity darkening. We find that the brighter (darker) flux from the poles (equator) has a much stronger impact on the mass flux, increasing (decreasing) the mass loss and local wind density. This, in addition to the existence of nonradial radiation forces from a rotating star, which tend to point latitudinally away from the equator and azimuthally opposite the rotation, produces a net poleward deflection of wind streamlines. This is contrary to the "wind compressed disk" model of Bjorkman and Cassinelli, and also seems incompatible with observational inferences of equatorial density enhancements in some systems. This work is ongoing, and we are endeavoring to include all the relevant physics in hydrodynamical simulations.
We also dynamically model spectral-line time variability by inducing corotating nonaxisymmetric structure in the equatorial plane of a hot-star wind. By varying the radiation force over localized "star spots," the wind develops fast and slow streams which collide to form corotating interaction regions (CIRs) similar to those in the solar wind. We synthesize P Cygni type line profiles for a stationary observer, and find that "discrete absorption components" (DACs) accelerate slowly through the profiles as complex nonlinear structures rotate in front of the star. We also examine the photospheric origin of such variability, in a preliminary manner, by deriving the theory of stellar pulsations, waves, and discontinuities. Although most observed low-order pulsation modes are evanescently damped in the photosphere, we find that the presence of an accelerating wind can allow waves of all frequencies to propagate radially. We thus make a first attempt at outlining the possible "photospheric connection" between interior and wind variability that observations are beginning to confirm.