Near-IR CO and CN in classical Cepheids

TL;DR

This paper tackles how CNO processing and metallicity influence near-IR molecular features in Cepheids by employing time-series, medium-resolution spectra of 12 Galactic Cepheids to measure CN and CO across pulsation phases and by fitting LTE atmospheres to derive $T_{eff}$, $log g$, and $v_{mic}$. It finds that CN is more sensitive to dredge-up than CO, while CO strength depends on both metallicity and pulsation-driven temperature variation, and identifies ET Vul as a possible merger-channel Cepheid. The authors connect CO variations to the mid-IR period-colour-metallicity relation, showing that disentangling temperature variation is essential for metallicity inferences from mid-IR colours in extragalactic Cepheids. The study demonstrates that near-IR CO and CN measurements are practical abundance probes at moderate resolution, informing metallicity calibrations in the cosmic distance ladder and anticipating JWST/NIRSpec Cepheid studies.

Abstract

We present medium resolution near-infrared spectral measurements of the carbon monoxide (CO) and the cyano radical (CN) features in 12 Galactic classical Cepheids. The pulsation periods of our sample range from 5.5 to 69 days, and the stars studied each had five or more near-IR spectral observations. The CO and CN measurements were used to probe CNO abundances of these stars, and elemental abundance values from the literature were used to identify the trends of [C/N] and [O/N] with CN and CO. To put these measurements in context, we performed stellar atmosphere fitting to obtain estimates of stellar parameters, with a primary focus on effective temperature. Our measurements and temperature estimates show that CN is significantly affected by dredge-up of processed material. We provide discussion as to the potential nature of the recently confirmed classical Cepheid, ET~Vul, and connect our near-infrared CO measurements to the mid-infrared period-colour-metallicity relation.

Figures (23)

• Figure 1: Examples of the indices in Table \ref{['tab:bandmeasurements']}, with the two Paschen line index regions are on the left and CN and CO regions on the right. These example spectra come from observations of BM Per, with the left panels at phase $\phi = 0.01$ and the right panels at a cooler phase, $\phi = 0.88$. Note: these examples have been normalized for illustration purposes, while the index measurements were performed without normalization.
• Figure 2: Measurements of CO, CN, and two Paschen lines compared with the TESS light curve and $(J-K)$ colour curve from Monson_2011 for classical Cepheid, BM Per. The gloesspersson04 curves are shown through the $(J-K)$ data and our index measurements.
• Figure 3: Spectral fitting via KORG for two observations of IY Cyg. The gray lines are the observational data. The thicker lines overlaid are the fitting windows adapted from inno2019, and the thin lines represent the synthesized spectra for the best-fit parameters. Residuals (offset from zero) are shown beneath both, where the shaded region represents $\pm 0.05$ or $\pm 5\%$.
• Figure 4: Phased CO and CN measurements for each star in the program, organized from top to bottom by period length (top-left longest, bottom-right shortest). Red circles represent CO and blue triangles, CN. gloess curves are shown following the same colour scheme, and continue through $1.0<\phi<2.0$ without our measurements for visualization purposes.
• Figure 5: The $\mathrm{CO_{(2-0)}}$ (left panels) and $\mathrm{CN_{(0-0)}}$ (right panels) vs. $\mathrm{(CN_{(0-0)}}-\mathrm{CO_{(2-0)}})$. The filled observations colours represent stars with [C/N] (left) and [O/N] (right) derived by luck18. Open circles are for stars without published N abundance. Mean errors are shown in the lower left corners of the top panels. The bottom panels are the same as the top but only for five stars: BM Per, CP Cep, GY Sge, IY Cyg, and S Vul. Lines represent weighted linear fits through the data for each star.