3. Observations and Data Reduction
All spectra used in this study were acquired between 2017
June and 2020 March using the CHIRON echelle spectrograph at
the SMARTS 1.5 m telescope described in Tokovinin et al.
(2013)and Paredes et al. (2021). We utilized CHIRON’s slicer
mode to attain spectra with resolution R=80,000, with each
spectrum split into 59 orders and covering a wavelength range of
4150–8800 Å. Each observation consisted of a single exposure of
900 s (stars with V<10.5)or 1200 s (V>10.5). This was done
to ensure that we attained an S/N greater than 30 for all spectra.
Each observation was followed by a single ThAr lamp exposure
of 0.4 s that was used for wavelength calibration. Included with
each night’s observations were two sets of calibration frames
integral to the data reduction process; one set is taken each
afternoon before nighttime operations begin, and another after all
observations have been completed for the night.
All CHIRON data are reduced using a customized data
reduction pipeline described in Tokovinin et al. (2013), with
additional details specific to our program in Paredes et al.
(2021). The pipeline is currently run by members of the
RECONS team, with CHIRON spectra reduced and distributed
to dozens of research teams for over 1000 nights by the end of
2021. Briefly, each spectrum is bias-corrected and flat-fielded
using quartz lamp calibrations to remove electronic readout noise
and to correct for individual pixel sensitivities. After removing
cosmic rays from the spectrum, profile order extraction is
performed using an extraction algorithm based on the REDUCE
package by Piskunov & Valenti (2002). Finally, each spectrum
with extracted orders is matched with its closest ThAr calibration
frame to obtain the sampled wavelength solution.
Once the basic pipeline reductions are done, the S/N per
pixel is calculated between 6717 and 6720 Åusing the method
described in Tokovinin et al. (2013). Spectra with an S/N less
than 30 are omitted from the present analysis because their EW
measurements are often unreliable for the spectral features of
interest. Each order is then trimmed at the edges to eliminate
poor signal portions of the orders, and the spectra are
normalized and flattened using a MATLAB script. Spectra
are then shifted twice: (1)first by applying a barycentric
velocity correction, and (2)then shifted to zero velocity relative
to the Sun. Spectral analysis of specific lines was then carried
out using the methods listed below.
4. Spectral Analysis
4.1. Line Selection
Four spectral features have been selected to create a rubric to
evaluate ages and activity for K stars—Hαat 6563 Å,theNa
I
doublet at 5890 and 5896 Å,LiIat 6708 Å,andCaII at 8452 Å.
Plots of the spectral regions containing the four selected lines for
all 42 stars are shown in Figures 2and 3.
The most direct information about stellar age for K dwarfs can
be gleaned from a Li Iresonance line at 6707.8 Å. Depletion of
lithium in late-type dwarfs has been well documented in previous
studies, such as Soderblom & Jones (1993), White et al. (2007),
López-Santiago et al. (2010), and Binks & Jeffries (2014).Itis
proposed that lithium is destroyed as a young K star settles onto
the main sequence through a process similar to the proton–proton
chain reaction, with the end product being two helium atoms and
the release of energy (Soderblom 2010). The decrease in the EW
of the Li Iline has been associated with increased age and is used
in our study of K dwarfs as the most direct age marker
(Soderblom et al. 2014). However, using the Li Ifeature is
somewhat limiting because the majority of lithium is depleted
within the first 200 Myr for dwarfs of type late G through early
M(Soderblom et al. 2014).ThetrendintheLi
Iλ6707.8 line
strength is also temperature dependent, fading faster for cool,
late-type dwarfs (K8V–M9V), as shown by Riedel et al. (2017),
who find an absence of Li Ifeatures in M-dwarf members of
moving associations with age estimates of only 50 Myr.
To enhance our efforts to estimate ages, we also consider
spectral lines resulting from stellar activity. Increased activity
in the chromospheres of late-type stars has been linked with age
since the publication of Skumanich (1972)five decades ago, so
a comprehensive literature search was done to find spectral
lines that might be used as activity, and presumably age,
markers for K dwarfs. Candidate lines need to be located within
CHIRON’s spectral range of 4150–8800 Å, and therefore the
popular Ca II H and K lines at 3968 and 3934 Åthat trace
chromospheric activity are excluded. For this study, we have
identified the Hαline at 6563 Åand one line of the Ca II
infrared triplet at 8542 Åas activity tracers; both lines exhibit
core emission or filled-in profiles when a K dwarf’s chromo-
sphere is active (Montes & Martin 1998). The other two Ca II
infrared triplet lines at 8498 and 8662 Åare omitted because
the orders produced by CHIRON’s slicer mode are truncated at
longer wavelengths and miss both lines.
Surface gravity diagnostic lines have also been proposed as
age markers for K dwarfs (Soderblom 2010). The idea is that
younger stars with ages <100 Myr are still contracting and
have bloated atmospheres compared to their older counterparts
already on the main sequence. Thus, the younger star’s larger
radius at the same mass results in a lower surface gravity, and
this can be revealed via relatively narrower spectral lines. In
effect, increased opacity in a fully contracted main-sequence
star leads to more atomic collisions and interactions in its
atmosphere, resulting in a wider absorption feature with
broader wings—this process is called pressure broadening.
The Na Idoublet lines at 5889.95 and 5895.92 Åare very
sensitive to pressure broadening. An increase in the EW of the
Na Idoublet feature (EW[Na ID])is therefore theorized to
accompany an increase in age (Soderblom 2010).
4.2. Equivalent Width Measurements
To carry out the spectroscopic analysis of age and activity
for the sample stars, we measured the EWs of both Na Ilines at
5889.95 and 5895.92 Å, the Hαline at 6563 Å, the unresolved
Li doublet at 6707.8 Å, and the Ca II infrared triplet line at
8542 Å, using the SPLAT-VO software, which is distributed by
the Starlink Project (Škoda et al. 2014). We compared SPLAT-
VO to other comparable methods for measuring EWs,
including the SPLOT package in IRAF, specutils using Python,
and VOSpec from the European Space Agency. All methods
resulted in similar EWs for a test sample of K dwarfs, and we
decided to use SPLAT-VO owing to its user-friendly interface.
All 42 K dwarfs were analyzed, including 35 from the
benchmark sample and the 7 RVV stars. A 20 Åwindow was
created for the Na Idoublet feature, and a 10 Åwindow was
created around the centers of the Li, Hα, and Ca II lines, with
pseudocontinuum fits made across each window using the
normalized data before carrying out the EW measurements.
SPLAT-VO uses the ABLINE technique to fit a Gaussian,
4
The Astronomical Journal, 164:174 (20pp), 2022 November Hubbard-James et al.