Listening to cool stars: When observations challenge theory
Stars may look like perfectly still points of light in the night sky, but they are anything but quiet. Much like musical instruments, they continuously vibrate, producing a rich spectrum of oscillations that carries information about their internal structure. By studying these oscillations, a technique known as asteroseismology, we can measure fundamental stellar properties such as mass, radius, and age with remarkable precision.
But what happens when the oscillations become so faint that we can barely detect them? And what if their measured amplitudes do not agree with our best theoretical predictions?
These are the questions we explore in our latest paper.
From ε Indi to two new stellar targets
For a couple of decades now, asteroseismology has benefited enormously from space missions such as Kepler and TESS, which measure the tiny brightness variations caused by stellar oscillations. However, these observations become increasingly challenging for K dwarfs — stars that are smaller, cooler, and less luminous than our Sun. Their oscillations are simply too faint to be readily detected through changes in brightness.
Fortunately, there is another way to "listen" to stars. As their surfaces oscillate, they move toward and away from us, producing tiny shifts in the wavelength of the light they emit. These shifts can be measured using the Doppler effect.
In 2024, our team demonstrated the power of this approach by detecting extremely faint oscillations in ε Indi, a nearby K dwarf, using ESPRESSO, an exceptionally precise spectrograph installed on the European Southern Observatory's Very Large Telescope in Chile. The measured oscillations had an amplitude of just 2.6 centimeters per second — approximately 16% of the solar value.
That breakthrough opened the door to studying stellar oscillations in a previously difficult-to-access regime. It also raised an intriguing question: Do oscillations become systematically weaker than expected as we move toward cooler stars?
Our latest study takes the next step. We used ESPRESSO to observe two additional nearby K dwarfs, GJ 570 A and GJ 783 A, over five and six nights, respectively, collecting thousands of measurements of their line-of-sight velocities.
We detected oscillations with characteristic periods of just 3–4 minutes and amplitudes of only a few centimeters per second. We also identified 6 individual oscillation modes in GJ 570 A and 24 in GJ 783 A, allowing us to estimate both stars' masses, radii, and ages.
A surprising change in the rules
The most intriguing result emerged when we combined our measurements with previous observations, including those of ε Indi.
For many years, we have known that stellar oscillation amplitudes depend on the ratio of a star's luminosity to its mass. Our observations provide further evidence that this relationship changes significantly as we move from Sun-like stars to cooler K dwarfs.
For warmer stars, oscillation amplitudes approximately scale with the luminosity-to-mass ratio raised to the power 0.7. For cooler stars, however, the dependence becomes much steeper, with an exponent closer to 1.5.
This transition occurs near the boundary between G- and K-type stars, at an effective temperature of approximately 5300 K. It means that oscillations become weaker much more rapidly toward cooler, lower-luminosity stars than a simple extrapolation from Sun-like stars would suggest.
When theory and observations disagree
Why does this happen?
Stellar oscillations are excited and damped by turbulent convection — the continuous movement of hot gas rising toward the surface and cooler gas sinking back into the stellar interior.
To investigate the underlying physics, our team performed three-dimensional computer simulations of convection in the outer layers of K dwarfs. These simulations calculated how energy is transferred into the oscillations and how it is dissipated.
Surprisingly, the simulations failed to reproduce the steep decline in oscillation amplitudes revealed by the observations. Instead of the observed exponent of approximately 1.5, they predicted a much shallower value of about 0.9.
This discrepancy suggests that something important may be missing from our current description of how convection interacts with stellar oscillations.
One possible ingredient is magnetic activity, which is known to suppress oscillations and was not included in our simulations. Although we find evidence that more magnetically active stars tend to oscillate less strongly, magnetic activity alone does not clearly explain the observed change in scaling. Other modeling limitations may also contribute.
Why does this matter?
Beyond revealing a gap in our theoretical understanding, these findings have practical implications.
Accurately predicting oscillation amplitudes is essential when planning future asteroseismic observations, including those of ESA's upcoming PLATO space mission. If we overestimate how strongly cool stars oscillate, we may also overestimate how many of them can be successfully studied.
More broadly, our results highlight the growing scientific opportunities offered by instruments originally developed to search for Earth-like exoplanets. Their extraordinary precision is now allowing us to explore stellar oscillations that were previously beyond our reach.
Comments