Recent updates / problems / where we stand about subsub HOT 7 CLOSED

Here is relevant text from Mathieu et al. 2003 regarding the temperature and logg from optical spectra, with bold emphasis mine:

Evidently, the primary star dominates the light at 5187 A˚ . As noted above, the CfA spectra indicate an effective temperature of Teff = 5000 K and a gravity of log g = 3.5. New analyses of the higher signal-to-noise spectra of van den Berg et al. (1999) yield similar measures. Comparison of the V i �6251.83/Fe i 6252.57 line ratio with the spectral analyses of Gray (Gray 1989; Gray & Johanson 1991) yield a Teff of 5150 K for a dwarf and 4900 K for a giant. Alternatively, in Figure 5 we use the spectral diagnostic Is as defined by Malyuto & Schmidt-Kaler (1997) to derive a spectral type of G8–K0. Such a spectral classification indicates effective temperatures similar to the spectral line results (SchmidtKaler 1982; Bessell & Brett 1988; Bessell, Castelli, & Plez 1998). Detailed analyses of the optical spectra obtained by van den Berg et al. (1999) indicate that the primary of S1063 is neither a giant nor a dwarf. We have taken two approaches to determining the surface gravity. First, we have compared the gravity-sensitive Mg b lines with Kurucz spectra. In Figure 6, we show the spectrum of S1063 and Kurucz spectra at Teff = 5000 K over a range of log g. Based on the Mg b triplet line shapes, S1063 is neither a giant (log g = 2.5–3.0) nor a dwarf (log g = 4.5–5.0). Spectral fitting to the Mg i b lines (with Teff a free parameter) results in Teff = 5000 K, log g = 3.5, and v sin i = 8 km s1, corroborating the results obtained from the lower signal-to-noise CfA spectra. However, the referee has correctly noted several disparities of relative line depths in other wavelength regions between the observed spectrum and the models, for example, comparing the Fe i 5187.9 A˚ and Ti ii/Ca i 5188.8 A˚ lines. More detailed spectroscopic analyses and modeling may prove fruitful.

Additional information from Emily via email:

As you’ll see, they use a few different spectroscopic methods but don’t quote any errors. They do give an error for S1113 of +/- 150 K, which they determine in the same way, so its probably similar. That said, the 5000 K temperature is pretty hard to reconcile with the photometry so I don’t know how much we want to trust it. Even with very large spot filling factors, the photosphere would have to be more like 4800 K for the SEDs to make sense at all.

So, I think a fair but not overly restrictive Teff prior is 4900+/-200 K.

from subsub.

Also, apparently there are discrepant lines also in the optical. WEIRD.

from subsub.

The work directory is incomplete.

gully at login2.maverick.tacc.utexas.edu in ~/science/subsub/sf on master [!?]
$ du -hs *
120M	m115

gully at login2.maverick.tacc.utexas.edu in /work/03342/gully/maverick/science/subsub/sf on master [?]
$ du -hs *
51M	m115
4.0K	README.md

gully at gigayear in ~/GitHub/subsub/sf on master [!?]
$ du -hs *
4.0K	README.md
  0B	m110
  0B	m113
124M	m115

from subsub.

You know what, let's just overhaul everything-- I'm going to run my automatic batch scripts to make config files and reduced hdf5 data for ALL orders. I'm borrowing code from my gorey project.

from subsub.

We looked at the spectrum and decided to force the prior on hot Teff to be equal to the optical-measured value (within the uncertainty). We also noticed discrepant lines (present in model but not data and vice versa). These are a mystery. More orders could help.

from subsub.

Update: Currently running order m=115

from subsub.

(with corrected flux treatment for mixture models).

from subsub.