stellar parameters

[jpdeleon]

• without spectroscopic priors from IRD, parameters are generally consistent whether we add optical photometry or not, i.e. plx+NIR vs plx+NIR+optical

  • fig1

• adding spectroscopic priors favors a young age solution (this is clearly seen in EEP panel in Figs. 3 & 4) and yields huge discrepancy in parameters, especially radius where the star appears anomalously inflated without optical photometry:

  • fig2

Note that EEP<202 is PMS stage. See Table 2.

• the above discrepancy is minimized when optical photometry is added, whether or not we add spectroscopic priors:
  • fig3
  • fig4

Having independent stellar parameter determinations from IRTF/SpeX spectrum can help add credence to those parameters derived from IRD. But unless we have evidence that the star is young, I am not inclined to use the spectroscopic priors in isochrones fit.

The figures are generated using this notebook.

[jpdeleon]

I ran astroARIADNE with 3 setups: run1. without spec priors run2. with IRD spec priors run3. with IRTF spec priors

In each run, I ran 4 out of 6 available stellar models:

• phoenix
• btsettl (identical to btnextgen & btcond for stars with Teff > 4000 K)
• kurucz
• ck04 (Kurucz and Castelli & Kurucz are known to work poorly on stars with Teff < 4000 K)

Looking at corner.pdf in each directory,

• run1 shows high Teff, and low Rs for a M dwarf.
• run2 shows reasonable Teff and Rs, consistent with those from tfop, but Teff posterior peak is at the edge (suggesting my Teff prior is biased against M dwarfs)
• almost the same as run2

I will check the priors and re-run 2 & 3. Or run setup # 4 where spec priors is combination of IRD and IRTF.

[jpdeleon]

The problem with stellar parameters having young solution has been solved. The problem originated from erroneous logg input being too low (<1), when the correct value should be 4.95 derived from mass and radius. I re-ran isochrones and ariadne using the weighted mean of IRD and IRTF values:

IRD	IRTF	weighted mean
Teff	3116, 114	3207, 99	3161, 75
logg	4.95, 0.01	4.959, 0.0627	4.954, 0.0099
Fe/H	-0.02, 0.35	0.3325, 0.0883	0.1563, 0.0856

You can check the updated results using the links above. I will add the results of each method in overleaf and let's decide which one to adapt in the paper.

[jpdeleon]

Using Mann's code, I got a mass of 0.2537+/-0.0064 Msun regardless of the Fe/H values I used.

edit: I realized that the relation is only valid for 4<MK<11; TOI1696 has Kmag=11.3.

edit2: MK range above is absolute magnitude and Kmag is apparent. TOI 1696's MK is actually 7.3, so the relation holds.

I also checked the probability of the posterior and it is well within the specified range (see line 137)

The problem now is that the mass uncertainty seems very small at ~2.5%.

[jpdeleon]

I have updated the stellar parameters based on the new stacked IRD spectra. The main difference is the new [Fe/H] value has a larger mean and a smaller uncertainty compared to the IRD single spectra results.

IRD	IRTF	weighted mean
Teff	3156, 117	3207, 99	3159, 63
logg	4.95, 0.01	4.959, 0.0627	4.9562, 0.0531
Fe/H	0.38, 0.24	0.3325, 0.0883	0.3382, 0.0829

I re-ran isochrones using the weighted means of IRTF and stacked IRD spectroscopic priors above. Generally, the resulting stellar parameters are consistent whether we add optical photometry or not (to the NIR and parallax inputs). Although, we should adopt the results of the latter case (see folder) so that uncertainties are not unrealistically small.

radius: 0.30-0.01+0.01 mass: 0.29-0.01+0.01 Teff: 3151-34+40 logg: 4.94-0.01+0.01 feh: 0.26-0.04+0.04

Without spectroscopic priors (see fig), we encounter a young age solution (pre-main sequence star) which we already ruled out.

[jpdeleon]

Issue: The Av posterior peaks at 0.3 but the extinction map expected for the star with distance modulus of 4 mag (=5*log10(64.9 pc)-5) is 0.

[jpdeleon]

The issue is fixed by setting maxAV=0.01. See star.ini.