J/A+A/697/A39 Spectral classification of young stars using cINN (Kang+, 2025)
Spectral classification of young stars using conditional invertible neural
networks. II. Application to Trumpler 14 in Carina.
Kang D.E., Itrich D., Ksoll V.F., Testi L., Klessen R.S., Molinari S.
<Astron. Astrophys. 697, A39 (2025)>
=2025A&A...697A..39K 2025A&A...697A..39K (SIMBAD/NED BibCode)
ADC_Keywords: Clusters, open ; Stars, pre-main sequence ; Spectroscopy ;
Effective temperatures; Spectral types ; Stars, ages
Keywords: methods: statistical - stars: late-type - stars: pre-main sequence -
HII regions -
open clusters and associations: individual: Trumpler 14 -
open clusters and associations: individual: Carina Nebula Complex
Abstract:
We introduce an updated version of our deep learning tool that
predicts effective temperature, surface gravity, extinction, and
veiling from the optical spectra of young low-mass stars with
intermediate spectral resolution. We determine the stellar parameters
of 2051 stars in Trumpler 14 (Tr14) in the Carina Nebula Complex,
observed with VLT/MUSE. We adopt a conditional invertible neural
network (cINN) architecture to infer the posterior distribution of
stellar parameters and train our cINN on two Phoenix stellar
atmosphere model libraries (Settl and Dusty). Compared to the cINNs
presented in our first study, the updated cINN considers the influence
of the relative flux error on the parameter estimation and predicts an
additional fourth parameter, veiling. We test the prediction
performance of cINN on synthetic test models to quantify the intrinsic
error of the cINN as a function of relative flux error and on 36 class
III template stars to validate the performance on real spectra. We
provide Teff, logg, Av, and veiling values of 2051 stars in Tr14
measured by our cINN as well as stellar ages and masses derived from
the Hertzsprung-Russell diagram based on the measured parameters. Our
parameter estimates generally agree well with those measured by
template fitting. However, for K- and G-type stars, the Teff derived
from template fitting is, on average, 2-3 subclasses hotter than the
cINN estimates, while the corresponding veiling values from template
fitting appear to be underestimated compared to the cINN predictions.
We obtain an average age of 0.7-0.6+3.2Myr for the Tr14 stars. By
examining the impact of veiling on the equivalent width-based
classification, we demonstrate that the main cause of temperature
overestimation for K- and G-type stars in the previous study is that
veiling and effective temperature are not considered simultaneously in
their process. Our cINN performs comparably to the multi-dimensional
template fitting method while being significantly faster and capable
of consistently analysing stars across a wide temperature range
(2600-7000K).
Description:
Catalog of stellar properties of Trumpler 14 sources measured by cINN
based on the VLT/MUSE observations.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 169 2051 Catalouge of low-mass stars in Trumpler 14 with
stellar parameters measured by cINN
sp/* . 2051 Individual spectra in pdf
--------------------------------------------------------------------------------
See also:
J/ApJS/194/10 : NIR properties of YSO in the CCCP (Preibisch+, 2011)
J/A+A/603/A81 : Trumpler 14 and 16 in the Carina nebula (Damiani+, 2017)
J/A+A/658/A198 : CHIPS II. O stars in Trumpler 14 CHIPS-Tr14 (Rainot+, 2022)
J/A+A/685/A100 : Young low-mass stars population in Trumpler 14 (Itrich+, 2024)
Byte-by-byte Description of file: table1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 7 A7 --- ID Star's identification number
9- 10 I2 h RAh Right ascension (J2000)
12- 13 I2 min RAm Right ascension (J2000)
15- 25 F11.8 s RAs Right ascension (J2000)
28 A1 --- DE- Declination sign (J2000)
29- 30 I2 deg DEd Declination (J2000)
32- 33 I2 arcmin DEm Declination (J2000)
35- 45 F11.8 arcsec DEs Declination (J2000)
48- 57 A10 --- Sample [Clean Uncertain] Name of sample group
59- 64 F6.1 % SigMed Median of relative flux error
66- 71 F6.1 K Teff Effective temperature
73- 78 F6.1 K e_Teff Error of Teff
80- 83 F4.2 [cm/s2] logg Logarithm of surface gravity
85- 88 F4.2 [cm/s2] e_logg Error of logG
90- 94 F5.2 mag Av Visual extinction
96- 99 F4.2 mag e_Av Error of Av
101-105 F5.2 --- veilr Veiling at 7500Å
107-110 F4.2 --- e_veilr Error of veilr
112-118 A7 --- Library [Settl Dusty] Phoenix library origin
120-124 A5 --- SpType Spectral type
126-130 F5.2 --- e_SpType [] Left-hand side error of SpType
132-136 F5.2 --- E_SpType [] Right-hand side error of SpType
139-145 F7.4 Lsun LbolJ ? Logarithm of stellar luminosity based on
J-band photometry from HAWK-I or Vista (1)
147-150 F4.2 Msun Mass ? Stellar mass based on PARSEC models (2)
152-157 F6.2 Myr Age ? Stellar age based on PARSEC models (2)
159-169 A11 --- FileName Name of the pdf spectrum file in
subdirectory sp
--------------------------------------------------------------------------------
Note (1): Reference for HAWK-I and VISTA catalogues as:
HAWK-I = Preibisch et al. (2011ApJS..194...10P 2011ApJS..194...10P, Cat. J/ApJS/194/10,
2011A&A...530A..34P 2011A&A...530A..34P)
VISTA = Preibisch et al. (2014A&A...572A.116P 2014A&A...572A.116P, Cat. J/A+A/572/A116)
Note (2): PARSEC evolutionary tracks from Bressan et al., 2012MNRAS.427..127B 2012MNRAS.427..127B
--------------------------------------------------------------------------------
Acknowledgements:
Da Eun Kang, daeun.kang(at)unibo.it
(End) Patricia Vannier [CDS] 31-Mar-2025