Stellar and orbital characterization of three low mass M dwarf binary stars with dynamical spectroscopy from the Habitable Zone Planet Finder

Theoretical models of low-mass stars continue to be discrepant with observations when used to examine the mass-radius relationship and other physical parameters of individual stars. High-resolution spectroscopy that leads to dynamical measurements of binary stars can directly improve these models. We have been using the Habitable-zone Planet Finder spectrograph to monitor binary stars with M dwarf components. Here, we measure the orbital and stellar parameters for three such systems: LSPM J0515+5911, NLTT 43564, and NLTT 45468. Each system has dozens of spectra obtained over a baseline of several years. None of the systems appear to be eclipsing, so our ability to turn them into true benchmark binaries with purely dynamical measurements is limited. We use literature photometry to estimate each system’s spectral energy distribution and utilize models in combination with detection limits of our spectroscopic measurements to probe characteristics of the companions. LSPM J0515+5911 is a double-lined spectroscopic binary with period of $126.948 \pm 0.029$ days and derived minimum masses, $M_1\sin^3i =0.058 \pm 0.002$ $M_\odot$ and $M_2\sin^3i = 0.046 \pm 0.001$ $M_\odot$ for the primary and secondary components, respectively. We solved NLTT 43564 with period of $1877 \pm 24$ days and NLTT 45468 with period of $9.686 \pm 0.001$ days as single lined systems, and modeled the primary masses to be $M_1 = 0.32\pm{0.02}$ $M_\odot$ and $M_1 = 0.35^{+0.02}{-0.07}$ $M\odot$, respectively.

💡 Research Summary

This paper presents a detailed spectroscopic study of three low‑mass M‑dwarf binary systems—LSPM J0515+5911, NLTT 43564, and NLTT 45468—using the Habitable‑Zone Planet Finder (HPF) on the 10 m Hobby‑Eberly Telescope. Over a time span of six years (2019–2025), the authors obtained 177 high‑resolution (R ≈ 53 000) near‑infrared spectra covering 0.8–1.3 µm. The data reduction employed the standard HPF pipeline, with additional steps for precise wavelength calibration using a laser‑frequency comb and a newly developed line‑spread‑function (LSF) model. A sophisticated telluric‑correction routine was applied, combining line‑by‑line atmospheric transmission models (H₂O, O₂, CO₂, CH₄) convolved with the instrument LSF, thereby minimizing residual atmospheric features that could bias radial‑velocity (RV) measurements.

Radial velocities were extracted differently for the two types of binaries. LSPM J0515+5911 is a double‑lined spectroscopic binary (SB2). The authors used cross‑correlation and multi‑Gaussian fitting to separate the primary and secondary RV curves, achieving a well‑constrained orbital solution with a period of 126.948 ± 0.029 days, eccentricity near zero, and minimum masses (M sin³i) of 0.058 ± 0.002 M⊙ (primary) and 0.046 ± 0.001 M⊙ (secondary). Because the inclination is unknown, true masses are larger, but the system already provides a valuable benchmark in the sub‑0.1 M⊙ regime.

NLTT 43564 and NLTT 45468 are single‑lined spectroscopic binaries (SB1). For NLTT 43564, 73 RV points yielded a long orbital period of 1877 ± 24 days and a semi‑amplitude of ~33 km s⁻¹. For NLTT 45468, 80 RV points gave a short period of 9.686 ± 0.001 days and a semi‑amplitude of ~30 km s⁻¹. Since only the primary spectrum is visible, the authors estimated the primary masses by fitting the spectral energy distribution (SED) using broadband photometry from Gaia, Pan‑STARRS, APASS, 2MASS, and WISE. They employed BT‑Settl and PARSEC stellar‑evolution models to infer effective temperatures, radii, and metallicities, arriving at M₁ = 0.32 ± 0.02 M⊙ for NLTT 43564 and M₁ = 0.35^{+0.02}_{-0.07} M⊙ for NLTT 45468. The mass functions derived from the RV curves then provide lower limits on the unseen companions, but without secondary spectra the exact masses, temperatures, and radii remain model‑dependent.

The authors discuss the implications of their results for the mass–radius relationship of low‑mass stars. All three systems show radii that are modestly larger (≈5–10 %) than predicted by widely used models such as Baraffe et al. (2015) and the Dartmouth isochrones, especially in the 0.05–0.07 M⊙ range. This discrepancy may point to shortcomings in the treatment of convection, magnetic activity, or metallicity effects in current stellar‑structure calculations. The lack of eclipses prevents direct inclination measurements, so true masses could be significantly higher than the minimum values reported.

Limitations of the study include the reliance on model‑dependent SED fits for SB1 systems, the absence of astrometric orbital information to break the sin i degeneracy, and the modest signal‑to‑noise of some spectra despite careful exposure planning. The paper suggests that future high‑resolution interferometry (e.g., CHARA, VLTI) or long‑baseline photometric monitoring could provide inclination angles, enabling full dynamical mass determinations.

In conclusion, the work demonstrates that HPF is a powerful tool for obtaining precise RVs of faint M dwarfs, even in the near‑infrared where telluric contamination is severe. By delivering well‑constrained orbital parameters and mass estimates for three low‑mass binaries, the study adds valuable data points to the sparsely populated empirical mass–radius diagram for M dwarfs. Continued monitoring and the addition of more double‑lined systems will further tighten constraints on stellar‑evolution models in the sub‑solar regime.