Observations of Convection in Atype Stars Barry Smalley
Observations of Convection in A-type Stars Barry Smalley Keele University Staffordshire United Kingdom 1
Introduction • Studies of convection from an observers perspective – What effects can we see? – What do observations tell us about convection? • Theoretical predictions – Can we give observers a convection prescription? 2
Mixing-Length Theory • A single bubble of rising gas – Rises a certain length before dispersing • Problems: – Too simple! – No prescription for mixing-length • pick your own value! 3
Turbulent Convection • Canuto & Mazzitelli Model – Using full range of bubble sizes and dispersion lengths – No free parameters! – Implemented in ATLAS 9 by Kupka (1996, ASP 44, 356) 4
Convective Overshooting • Bubbles rise above the convections zone into the stable regions – overshooting – should be present in our models 5
Approximate Overshooting • “[Kurucz] convective models use an overshooting approximation that moves flux higher in the atmosphere above the top of the nominal convection zone. Many people do not like this approximation and want a pure unphysical mixing-length convection instead of an impure unphysical mixinglength convection. ” (http: //kurucz. harvard. edu) 6
Atmospheric Structure OV MLT CM Heiter et al. , 2002, A&A 392, 619 • At Teff = 8000 K CM gives essentially radiative temperature gradient – less convective flux than MLT • Overshooting introduces flux in higher layers 7
Realistic Convection Models • None of the current 1 d models of convection are totally satisfactory – 2 d and 3 d numerical simulations (Freytag) – Improved analytical 1 d treatments (Kupka) How good are 1 d models? 8
Observational Diagnostics • I will discuss the following: – Photometric colours – Flux distributions – Balmer lines – other line profiles • Mostly based on comparison with Kurucz ATLAS 9 models – extensively used – computationally cheap 9
Photometry • Fast and efficient method for determining atmospheric parameters – many calibration grids – especially uvby system • Indices sensitive to Teff, log g and [M/H], as well as convection and microturbulence 10
uvby photometry • Comparison with fundamental stars is in good agreement – uvby photometry is good Teff and log g indicator – CM and MLT are good, but no overshooting Smalley & Kupka, 1997, A&A 293, 446 BUT. . . 11
Bump around 6500 K Smalley & Kupka (1997) • Bump in difference between log g from uvby and that from evolutionary models for Hyades – related to onset of strong surface convection? 12
The m 0 index • Traditionally m 0 index is poorly fitted – combination of varying mixing-length, microturbulence and overshooting might work? 13
Stellar Fluxes • Emergent flux influenced by convection’s effect on atmospheric structure – subtle but measurable effects in optical spectrophotometry – In ultraviolet effects more significant • but severe problems with metal line blanketing – Infrared fluxes less sensitive • Infrared Flux Method (IRFM) 14
Effects on Fluxes @ 8000 K • CM and MLT 0. 5 similar to no convection • MLT with and without overshooting identical 15
Effects on Fluxes @ 7000 K • Flux highly sensitive to value of mixing-length • Overshooting is radically different 16
Spectrophotometry • Current spectrophotometry has insufficient resolution and precision to be really useful • The ASTRA robotic spectrophotometer will provide a huge volume of useful stellar fluxes (see Adelman et al. Poster JP 2) 17
Balmer line profiles • Useful diagnostic – strong in A and F stars • sensitive to Teff • insensitive to log g for late-A and cooler – formed at different depths within atmosphere • probe differing parts of atmospheric structure 18
Balmer profile variations Teff = 7000 K, log g = 4. 0 • Changing the efficiency of convection, by increasing mixing length, has significant effect on computed profile 19
Balmer profile sensitivities • H insensitive to mixing-length • H sensitive to mixing-length • Both lines affected by overshooting – sensitive to temperature and metallicity – surface gravity sensitivity for hotter stars Van’t Veer & Megessier, 1996, A&A 309, 879 20
Fundamental Stars • H and H are in good agreement with fundamental stars – Both CM and MLT (l/H ~ 0. 5 preferred) – no overshooting Smalley et al. , 2002, A&A 395, 601 21
H - H Gardiner et al. , 1999, A&A 347, 876 • Balmer profiles prefer l/H = 0. 5 hotter than 7000 K and l/H = 1. 25 for cooler stars 22
What is Microturbulence? • A free parameter introduced to allow abundances from weak and strong lines to agree? • Small-scale motions within atmosphere added to thermal broadening? • Figment of our imagination caused by incomplete physics in 1 d atmospheres? • Intimately related to convective motions within the atmosphere? 23
Microturbulence Variations Based on Gray et al. 2001, AJ 121, 2159 • Microturbulence varies with Teff – increases with increasing temperature – peaks around mid-A type 24
Line Asymmetries • Line Bisectors – Velocity fields in atmosphere • Rising elements blue shifted • Falling elements red shifted • A-type Stars – small rising columns of hot gas – larger cooler downdrafts – velocities consistent with microturbulence Landstreet, 1998, A&A 338, 1041 Gray’s (1992) Book 25
No need for microturbulence? • Numerical simulations avoid the need for such a free parameter (Asplund et al. , 2000, A&A 359, 729) – de-saturating effects • not turbulent motions • but velocity gradients • No longer a free parameter, but should be constrained when using 1 d models 26
Transition Region • Changing from weak subsurface convection to fully convective. – Observational signatures • e. g. uvby “bump” – Sudden or gradual changes in atmosphere? – Böhm-Vitensse Gap • related to internal structure changes Gray’s Book (1992) 27
Competing Processes • We cannot treat convection and turbulence in isolation – Diffusion – Rotation – Magnetic fields – Metallicity 28
Fundamental Stars • Stars with known properties – reduces number of free parameters when comparing observations to models • Need to extend the number and quality of such stars in the A-F star region – including peculiar stars • Need high-quality fluxes, Balmer-line profiles and high-resolution spectra of these fundamental stars. (see Posters BP 2, IP 1, JP 2, JP 3 and JP 6) 29
A Prescription for Observers? ? g n i t o o h s r e v o • Schematic variation of microturbulence and mixing length with Teff. – The two appear to be intimately linked 30
The Surface of an A Star? Thank you! 31
- Slides: 31