Automated Fitting of High-Resolution Spectra of HAeBe stars Improving fundamental parameters Jason Grunhut Queen’s University/RMC
Motivation Common ways to determine temperature Common ways to determine temperature Photometry Photometry SED SED Problems Problems Extinction/emission and calibrations Extinction/emission and calibrations Many corrections necessary Many corrections necessary Take advantage of high-res ESPaDOnS spectra Take advantage of high-res ESPaDOnS spectra Minimal corrections required Minimal corrections required
Full ESPaDOnS Spectral Range K synthetic model
Spectrum variation with temperature from nearest Kurucz models (±500 K)
Automated Fitting of Spectra Search through a pre-defined grid of synthetic spectra. Search through a pre-defined grid of synthetic spectra Angstroms Angstroms Solar abundances. Solar abundances. Most current VALD line list. Most current VALD line list. Micro-turbulent velocity of 2 km/s. Micro-turbulent velocity of 2 km/s. No macro-turbulence. No macro-turbulence. Models computed using synth3 Models computed using synth3 Grid from K, log(g) from Grid from K, log(g) from K resolution up to K, 200 K resolution from then up. 100 K resolution up to K, 200 K resolution from then up.
How Program Works Radial velocity is first determined based on suggested model. Radial velocity is first determined based on suggested model. Projected rotational velocity is fit for each model in the specified range (computed using slightly modified s3dIV code). Projected rotational velocity is fit for each model in the specified range (computed using slightly modified s3dIV code). Model with minimum chi-square represents best fit. Model with minimum chi-square represents best fit. Radial velocity is fit for a final time for best model. Radial velocity is fit for a final time for best model.
Theoretical Results for K synthetic model with vsini of 40 km/s CLEAR MINIMUM EXISTS
Chi-Square Map HD Using chi-square map, can estimate uncertainties. Using 3 parameter fitting space, chi- square difference of 21.1 represents a formal 99.99% confidence level. closest model has greater than 2300 chi-square difference
Theoretical Results Investigated SNR vsini varying Fe abundance random noise to log(gf) values micro/macro turbulence binaries normalization conclusion other than binaries, for reasonable variations, ~ K uncertainties
Results Name Metallic Lines (Teff, Log(g)) Reduced χ 2 Hγ (Teff, Log(g)) Hβ (Teff, Log(g)) LiteratureValue HD , , , HD , , , HD , , , HD , , , 4.0 ~9500 HD , , ± 1005 HD , , , /- 198 HD , , , ± 420 HD , , , ± 358 HD , , , ± 3650
HD B7IV Classification Best Fit K Log(g)=4.0 vsini=20 km/s Literature Results ~12300 K
HD 17081
HD Balmer Fits Best Fit: K, Log(g)=3.5Best Fit: K, Log(g)=4.0
HD A0e+sh Classification Best Fit K Log(g)=4.5 vsini=108 km/s Literature Results ~8700 (+410,- 198) K
HD 34282
HD Balmer Fits Best Fit: 9800 K, Log(g)=4.5Best Fit: K, Log(g)=4.5
HD A8e Classification Best Fit 7900 K Log(g)=4.0 vsini=52 km/s Literature Results ~7700 K
HD 36112
HD Balmer Fits Best Fit: 8000 K, Log(g)=5.0Best Fit: 8100 K, Log(g)=5.0
HD A3pshe+ Classification Best Fit 8200 K Log(g)=3.5 vsini=95 km/s Literature Results 8700 K 9250 K, Log(g)=3.5
HD 31648
Difficult Stars: BF Ori A5II-IIIe var Best Fit ~7500 Log(g)~4.0 vsini~53 km/s Literature Results 6750
BF Ori
HR DIAGRAM: New Temperatures
HR Diagram: New Temperatures and Distances
HR Diagram: New Temperatures and Computed Photometry
FUTURE WORK Automated fitting for all field HAeBe stars with ESPaDOnS observations. Automated fitting for all field HAeBe stars with ESPaDOnS observations. Use improved temperatures to improve mass and age estimates. Use improved temperatures to improve mass and age estimates. Use Bayesian statistical approach to improving luminosities. Use Bayesian statistical approach to improving luminosities. Major Issues Abundances for chemically peculiar stars. Abundances for chemically peculiar stars. Micro/macro turbulence. Micro/macro turbulence. Systematic normalization issues. Systematic normalization issues.
THE END
Balmer Line Normalization: HD36112
Balmer Line Normalization: HD139614
Balmer Line Normalization: Comparison between ESPaDOnS and FORS1 HD 36112
Uncertainty vs SNR For K synthetic model with 40 km/s vsini.
15000 K synthetic model with 40 km/s vsini
Difficult stars: HD 31293
HD 31293