Helium Abundances from tess Ensemble Asteroseismology

TESS Sci Con 2021

Joel Ong, et al.

August 2021 | Slides at http://hyad.es/talks

Motivation and Background

The Big Question

How are the initial helium abundances of stars
related to their metallicities?


M_0 = 1.0 M_\odot; Z_0 = 0.019

T_\text{MS} = 8.6\ \mathrm{Gyr}

T_\text{MS} = 13\ \mathrm{Gyr}


Y_0 = 0.27; Z_0 = 0.019

grey lines: models from previous slide

Power spectra of MDI dopplergrams

Individual mode frequencies exhibit
detailed structure on an échelle diagram.

\underbrace{\color{red}\Delta\nu, \nu_\text{max}}_{\mathclap{\text{Global properties only}}}, \overbrace{\color{blue}\left\{\nu_{nl}\right\}}^{\mathclap{\text{interior structure}}}

The Helium Signature

The Helium Signature in p-modes

Second frequency differences/\mu\mathrm{Hz}

“Acoustic glitches” as indicator of surface helium abundances

Main-sequence Kepler sample consistent with broad range of \Delta Y / \Delta Z.

Mixed Modes


Main Sequence
Red Giant

The Helium Signature in Mixed Modes

only l=1, pure p-modes: subtle changes

The Helium Signature in Mixed Modes

only l=1, mixed modes: pretty sensitive

TESS Targets

Proposed targets

Final Modelling Targets

(some missing metallicities)

Stellar Modelling

  • Inputs:
    • Classical constraints (L, [\mathrm{Fe/H}], T_\text{eff})
    • Detailed seismic constraints (individual mode frequencies \nu_{nl})
  • Inference:
    • Grids of evolutionary models (MESA; YREC)
    • Computation of mixed-mode frequencies (GYRE; ADIPLS)
    • Different inference schemes (glitches vs. avoided crossings)
  • Outputs:
    • Initial mass fractions Y_0 and Z_0

Preliminary Results

Avoided crossings appear to constrain Y_0
better than glitches.

(main-sequence Kepler stars, for comparison)

(TESS subgiant sample, two ways)

Suspiciously small/asymmetric
posterior errors?

Caustics in posterior distribution:
low-metallicity results prior-dominated
(we’re already fixing this)

(including Kepler/K2 subgiants)

Main-sequence results analysed differently — possible methodological offset?


  • Inference of Y_0 from mixed-mode pattern is feasible, yielding stronger constraints on Y_0 than from pure p-modes
  • Constraints may be placed on \Delta Y / \Delta Z given large enough subgiant sample (so far only have processed Southern CVZ)
    • Bigger sample from Northern CVZ and 20-second data may help
  • May be a need for a homogenous reanalysis of all available asteroseismic targets (including Kepler/K2, MS and RG stars)

Backup Slides