Subject: Emma's Mdots From: David Cohen Date: 11/15/11 10:27 PM To: Jon Sundqvist CC: David Cohen , Maurice Leutenegger Jon, I've gone back and reanalyzed Emma's taustar values to derive Mdots, using a wind opacity model based on Asplund 2009 solar abundances. I used the Rstar and vinf values in Table 1 of her paper. This is essentially what I did in late August, but this time I didn't stupidly misinterpret the format of the uncertainties in the data files she sent me. Du-oh! So, maybe when you get a chance, you could remake your Emma-Vink and Vink-Halpha(unclumped) plot. A couple of specific comments about the attached file: - I didn't include HD 93129A (since we've published it already) or tau CMa or 15 Mon (because of the very poor S/N in their Chandra spectra). - Now that I'm doing it right, my values agree with those in Emma's paper to within 10 or 20%, the differences being plausibly due to the slightly different wind opacity models we each used. - In parentheses after the Mdot is the 68% confidence range. - epsilon Ori and HD 150136 are the only objects for which accounting for resonance scattering (simply by removing those lines most likely subject to the effect) makes a difference. Those res. sc. Mdots are also listed in the file. (And should be used, I think, in your plot.) - xi Per has a not-insignificant radial velocity and I've refit all the lines in that spectrum, including a shift of +57 km/s (which makes the line profile blue shifts effectively bigger and thus raises the taustar values and ultimately the mass-loss rate). Details are here: http://astro.swarthmore.edu/~cohen/projects/emma/www/xiPer/ The upshot (!) is that there's a marginally significant increase in the derived mass-loss rate. We should use that. But also (and unrelatedly) there appears to be a divergent set of Rstar values for this star in the literature. On the webpage with my new fit results, I use Haser's large value, but Emma used Repolust's small value. I include both results in the attached summary. Don't know which one we should use. What do you think? As I mentioned last night, settling on our best stellar and wind parameter values for each star would be good to do soon. David emma_Mdot_check_summary.txt Using Emma's tau_star values and Aspl 2009 solar abundance based opacities 9 Sgr N = 7 Mdot = 3.72e-7 (2.80:3.70e-7) Emma's paper has 4.07e-7 (for solar) chi2 = 3.3 HD 150136 N = 7 Mdot = 9.2e-8 (5.4:11.3e-8) Emma's paper has 8.3e-8 (for solar) chi2 = 8.8 When we exclude the lines subject to resonance scattering we find Mdot = 2.28e-7 (1.35:3.26e-7) Cyg OB2 8A N = 4 Mdot = 7.96e-7 (2.88:too high for my grid) Emma's paper has 8.85e-7 (for solar) chi2 = 2.96 xiPer N = 7 Mdot = 1.52e-7 (1.00:2.02e-7) Emma's paper has 1.61e-7 (for solar) chi2 = 5.35 NOTE: Emma's paper has Rstar = 14 in the table (from Repolust et al. 2004) but Haser has 25.5, which will of course raise the mass-loss rate accordingly. (Note further that I used 25.5 for the comparison between the unshifted model and the model that accounts for the RV of +57 km/s.) AND: If I use my results from assuming a shift (and including Si XIII but not Ne IX) I get 1.80e-7 (1.33:2.29e-7) (directly comparable to Emma's 1.52e-7). iota Ori N = 7 Mdot = 3.2e-10 (rather unconstrained - upper limit a few e-9) - Emma's paper has 4e-10 (for solar) chi2 = 1.0 zeta Oph N = 8 Mdot = 1.48e-9 (<2e-10:4.29e-9) Emma's paper has 2e-9 (for solar) chi2 = 4.66 delta Ori N = 8 Mdot = 4.4e-8 (2.1:6.9e-8) Emma's paper has 5.4e-8 (for solar) chi2 = 6.31 zeta Ori N = 9 Mdot = 2.16e-7 (1.76:2.54e-7) Emma's paper has 2.34e-7 (for solar) chi2 = 13.3 epsilon Ori N = 9 Mdot = 2.08e-7 (1.39:2.74e-7) Emma's paper has 2.25e-7 (for solar) chi2 = 14.1 NOTE: This is the one star where censoring some data due to resonance scattering makes a significant difference (6.36e-7 (4.71:8.10e-7) with chi2 = 1.1) Attachments: emma_Mdot_check_summary.txt 1.6 KB