3D non-LTE Ca II line formation in metal-poor FGK stars. I. Abundance corrections, radial velocity corrections, and synthetic spectra

Title: 3D non-LTE Ca II line formation in metal-poor FGK stars. I. Abundance corrections, radial velocity corrections, and synthetic spectra
Authors: Cis Lagae, Anish M. Amarsi, Karin Lind

Abstract: The Ca II resonance doublet (HK) and the near-infrared triplet (CaT) are among the strongest features in stellar spectra of FGK-type stars. These spectral lines remain prominent down to extremely low metallicities and are thus useful for providing stellar parameters via ionisation balance and as radial velocity diagnostics. However, the majority of studies that model these lines in late-type stars still rely on one dimensional (1D) hydrostatic model atmospheres and the assumption of local thermodynamic equilibrium (LTE). We present 3D non-LTE radiative transfer calculations of the CaT and HK lines in an extended grid of 3D model atmospheres of metal-poor FGK-type. We investigate the impact of 3D non-LTE effects on abundances, line bisectors and radial velocities. We used a subset of 3D model atmospheres from the recently published STAGGER-grid to synthesize spectra in 3D (non-)LTE. For comparison, similar calculations were performed in 1D (non-)LTE using models from the MARCS grid. Abundance corrections for the CaT lines relative to 1D LTE range from +0.1 to -1.0 dex, with more severe corrections for strong lines in giants. With fixed line strength, the abundance corrections become more negative with increasing effective temperature and decreasing surface gravity. Radial velocity corrections relative to 1D LTE based on cross-correlation of the whole line profile range from -0.2 km/s to +1.5 km/s, with more severe corrections where the CaT lines are strongest. The corrections are even more severe if the line core alone is used to infer the radial velocity. The line strengths and shapes, and consequently the abundance and radial velocity corrections, are strongly affected by the chosen radiative transfer assumption, 1/3D (non)-LTE. We release grids of theoretical spectra that can be used to improve the accuracy of stellar spectroscopic analyses based on the Ca II triplet lines. 

----------------------------------------------------------------------------------------------------------------------

We provide all the synthetic normalized flux spectra together with equivalent widths and abundance corrections only for the CaT lines. All of the output files contain python dictionaries stored in hdf5 format. Each model is defined by an unique identifier, allowing easy access of its properties across the different output files. This identifier is the first dictionnary key.

An example identifier is the following: 't65g45m30_z295(a100)'
    t65  -> Teff   = 6500 K
    g45  -> logg   = 4.5
    m30  -> [Fe/H] = -3.0   
    z295 -> A(Ca)  = 2.95   #Calcium abundance
    a200 -> microtubulent velocity = 2.00 km/s   #This is only added in the EW1D.h5 file

An example on how to read the files using python, which also clarifies the dictionnary structure, is given in example.py.

We do not provide any interpolation routines to interpolate inside the grid. We refer to Canocchi et al. (2024, DOI: 10.1051/0004-6361/202451972) who provided interpolation routines for the flux spectrum, for a similar grid.
     
--------------------------------------------------------------------------------
The flux grid is set on a 6-dimensional grid with the following dimensions:
Number of models (~50) x number of abundances (9) x 1D,3D,LTE,NLTE (4) x microturbulence (0 or 3) x number of lines (5) x number of wavelength points (~900)

The corrections and equivalent widths are only computed for the Ca Triplet lines, with the following dimensions:
Number of models (~50) x number of abundances (9) x 1D,3D,LTE,NLTE (4) x microturbulence (0 or 3) x number of lines (3) 

--------------------------------------------------------------------------------
For completeness, the vacuum and air wavelengths used to compute the reduced equivalent widths are the following:
wl0_8497_vac = 8500.358770604962 #Å
wl0_8542_vac = 8544.434860987318 #Å
wl0_8662_vac = 8664.521664166832 #Å

wl0_8497_air = 8498.023772239989 #Å
wl0_8542_air = 8542.087945880607 #Å
wl0_8662_air = 8662.142276340805 #Å

--------------------------------------------------------------------------------
File Summary:
--------------------------------------------------------------------------------
 FileName                Explanations
--------------------------------------------------------------------------------
ReadMe                   This file
wl.5                           Wavelength grid corresponding to the grid of fluxes
fluxes.h5                 Grid of synthetic normalized fluxes 
EW1D.h5                 Grid of 1D equivalent widths
EW3D.h5                 Grid of 3D equivalent widths
corrections.h5       Grid of abundance corrections and equivalent widths

--------------------------------------------------------------------------------
Detailed description of each file:

Each file contains a pandas.dataframe type object (similar to a python dictionary).
Each key is a string.

--------------------------------------------------------------------------------
 FileName                Keys
--------------------------------------------------------------------------------
wl.5                     wl_H, wl_K, wl_8497, wl_8542, wl_8662

fluxes.h5                HK, cat
                         -> model identifier (example: t65g45m40_z270)
                            -> 3D, 1D100, 1D150, 1D200
                               -> H, K OR 8497, 8542, 8662
                                  -> IF 3D: mean_nflux_nlte, mean_nflux_lte; ELIF 1D: nflux_nlte, nflux_lte 
       
                            
EW1D.h5  
                         -> model identifier (example: t65g45m40_z270a150)
                            -> Teff_target, logg, feh, ACa, CaFe, vmic, 
                               W8497nlte, W8542nlte, W8662nlte, W8497lte, W8542lte, W8662lte

EW3D.h5  
                         -> model identifier (example: t65g45m40_z270)
                            -> Teff_target, Teff_real, Teff_sigma, logg, feh, ACa, CaFe,
                               W8497nlte, W8542nlte, W8662nlte, W8497lte, W8542lte, W8662lte,
                               new, pec_outer

corrections.h5           
                         -> model identifier (example: t65g45m40_z270)
                            -> Teff_target, Teff_real, Teff_sigma, logg, feh, ACa, CaFe,
                               Acorr8497_vm10, Acorr8497_vm15, Acorr8497_vm20, 
                               Acorr8542_vm10, Acorr8542_vm15, Acorr8542_vm20, 
                               Acorr8662_vm10, Acorr8662_vm15, Acorr8662_vm20
                               new, pec_outer

--------------------------------------------------------------------------------
Detailed description of each keyword:

NAME                      Description                            UNIT
Teff_target            Effective temperature of the 1D model                Kelvin
Teff_real                 Average effective temperature of the 3D model            Kelvin
Teff_sigma             Standard deviation on the 3D model Teff                Kelvin
logg                          Surface gravity                            dex, log10(cgs)
feh                            Metallicity [Fe/H]                        dex    
ACa                           Calcium abundance A(Ca) = [Ca/H] + A(Ca)_solar
CaFe                         Calcium-to-iron ratio [Ca/Fe]
W8497nlte             non-LTE equivalent width of the CaT 8497 line            Angstrom
W8542nlte             non-LTE equivalent width of the CaT 8542 line            Angstrom
W8662nlte             non-LTE equivalent width of the CaT 8662 line            Angstrom
W8497lte                LTE equivalent width of the CaT 8497 line            Angstrom
W8542lte                LTE equivalent width of the CaT 8542 line            Angstrom
W8662lte                LTE equivalent width of the CaT 8662 line            Angstrom
new                           True if 3D model is from the new Stagger grid (Rodriguez Diaz et al. 2024, DOI: 10.1051/0004-6361/202348480)
pec_outer               True if model is flagged according to Section 3.2 in Rodriguez Diaz et al. (2024)

mean_nflux_nlte             3D spatially and temporally averaged normalized flux in non-LTE    
mean_nflux_lte               3D spatially and temporally averaged normalized flux in LTE
nflux_nlte                          Normalized flux in 1D non-LTE
nflux_lte                             Normalized flux in 1D LTE

Acorr8497_vm10   3D non-LTE vs 1D LTE abundance correction for the 8497 line and microturbulence of 1.0 km/s
Acorr8497_vm15   "                                                 8497                             1.5 km/s     
Acorr8497_vm20   "                                                 8497                             2.0 km/s     
Acorr8542_vm10   "                                                 8542                             1.0 km/s     
Acorr8542_vm15   "                                                 8542                             1.5 km/s     
Acorr8542_vm20   "                                                 8542                             2.0 km/s     
Acorr8662_vm10   "                                                 8662                             1.0 km/s     
Acorr8662_vm15   "                                                 8662                             1.5 km/s     
Acorr8662_vm20   "                                                 8662                             2.0 km/s