Metal Pipe: A Broadly-Applicable Stellar Abundance Pipeline Using Isochronal Parameters

TL;DR

Metal Pipe addresses the need for homogeneous stellar abundance catalogs across FGK and M dwarfs by integrating photometric isochrone parameters with high-resolution spectra in a MOOG-based abundance pipeline. It uses Gaia-based parallaxes and MIST isochrones to derive $T_{ m eff}$, $\log{(g)}$, $M_*$, $R_*$, and $L_*$, while iteratively fitting abundances for C, O, Na, Mg, Al, Si, S, Ca, Ti, and Fe with line-by-line weighted $\chi^2$ minimization and NLTE corrections for select elements. Benchmarking against 503 HIRES stars shows RMS scatters of about $100\,\mathrm{K}$ in $T_{ m eff}$, $\sim0.10$ dex in $\log{(g)}$ and $\sim0.10$ dex in abundances, indicating the method’s reliability within LTE uncertainties and its suitability for building a detailed abundance catalog. The authors plan to extend the approach to late-K and M dwarfs, expand line lists to the near-IR, refine line-broadening physics, and create a large, uniformly analyzed catalog to uncover chemical-architecture correlations with planetary systems.

Abstract

Characterizing exoplanet host stars at a population level requires a method of homogeneously characterizing stellar properties across all spectral types. To this end, we have developed Metal Pipe, a new code for determining stellar parameters and abundances, which is designed for use across a wider range of spectral types than many commonly used codes. It combines the widely-used package MOOG with photometric stellar parameters, a user-supplied high-resolution spectrum, and a newly curated list of spectral lines. Metal Pipe outputs values for $T_{\rm{eff}}$, $\log{(g)}$, $M_*$, $R_*$, and $L_*$ from isochrones, and abundances of C, O, Na, Mg, Al, Si, S, Ca, Ti, and Fe from MOOG. In this paper, we describe the Metal Pipe algorithm and tests of Metal Pipe against previous abundance measurements on archival HIRES spectra of 503 F, G, and K type stars. We find RMS scatters of ~100 K in $T_{\rm{eff}}$, ~0.10 dex in $\log{(g)}$, and ~0.10 dex for all measured abundances. These values are comparable to estimated measurement uncertainties, verifying Metal Pipe for continued use in building a detailed abundance catalog. Future papers in this series will test Metal Pipe's applicability to late K and M dwarf stars, and provide other improvements.

Figures (19)

• Figure 1: Histogram of abundances from three Ti lines as outlined in Section \ref{['linelistcuration']}. Line 'A' returns highly variable abundance readings (note the broad distribution of abundance measurements). Line 'B' achieves consistent abundance readings, but with a severe systematic offset from the median (i.e. this line consistently underestimates the Ti abundance), making it unsuitable for analysis. Line 'C' is ideal for our work, achieving abundance values within 0.1 dex of the median in roughly 90% of stars.
• Figure 2: An illustration of our spectrum normalization process. Top: A 1-D flat-fielded HIRES spectrum of 51 Peg where flux has been scaled by an arbitrary constant. The red line shows the computed continuum after the median and maximum filters have been applied. The yellow line shows the result after the alpha ball and convex hull have been applied, and the blue line shows the final smoothed continuum. Bottom: The final continuum has been divided out. The blue line is set to 1 at all wavelengths.
• Figure 3: A flow chart of the Metal Pipe algorithm. Blue rectangles represent the execution of a process or function. Yellow diamonds represent a conditional check (e.g. an "if" statement). Purple rhomboids represent inputs to the code.
• Figure 4: An automatically generated diagnostic plot, illustrating stellar parameter derivation for 51 Peg. Bottom left: Normalized probability grid from $\chi^2_v$ values of the synthetic photometry. Top left: Photometric SED of 51 Peg ("Stellar Photometry"), best-fit synthetic photometry from the isochrone grid (black), and other synthetic photometry values sampled from the probability grid (grey). Right: Histograms of the sampled values for stellar age, effective temperature, and surface gravity.
• Figure 5: Line fits from Metal Pipe showing the best fit to three Fe lines in 51 Peg. Top Row: The grey line represents the observed line as seen through the spectrograph, and the red line represents the synthetic best fit to the data. The shaded regions around the lines represent $1\sigma$ flux error bars: either intrinsic flux uncertainty in the observed spectrum (grey) or the added uncertainty to the synthetic spectrum as a function of line depth (red). Bottom Row: Weight of each data point in the $\chi^2_\nu$ fit process. Higher weights correspond to regions where the flux is more sensitive to a change in the Fe abundance.