JD

JIADONG LI 李佳东

Astronomer · Max Planck Institute for Astronomy

Reading starlight.
Revealing
hidden stars.

A point of light can hide another star—and a story about how our Galaxy came to be. I use stellar spectra, physics, and machine learning to uncover both.

Explore my research
ONE BINARY · LIGHT AND MOTION
Binary SED Two stars, one spectrum · relative flux
Blackbody continuum spectra for two stars and their sum on a fixed flux scale 3506008501100 nm
PrimaryCompanionCombined light
Kepler orbit Sky projection · AU
Two stars orbiting their common centre of mass, projected onto the sky 1 AU
Astrometric wobble Light centroid · mas
Motion of the unresolved light centroid around the centre of mass at a distance of 100 parsecs 2 mas
Radial velocity km/s · positive = receding
Radial velocities of both stars over one orbital period, with markers at the current phase 000.51.0Orbital phase
PrimaryCompanionLight centroid+ Centre of mass

Try a face-on orbit: radial velocities vanish. Try equal masses: equal light cancels the photocentre wobble.

Model assumptions

Keplerian two-body model: M₁ = 1 M☉, a = 1 AU (relative orbit), d = 100 pc, ω = 45°, Ω = 30°. Period follows P² = a³/(M₁ + M₂). Uniform time steps; 12 s per orbit. RVs are individual stellar velocities, with zero systemic velocity. Astrometry shows the light centroid at 550 nm, excluding proper motion and parallax. Blackbody continua: T₁ = 5,800 K, T₂/T₁ = q⁰·⁵, R₂/R₁ = q⁰·⁸; fixed flux scale, no line spectrum or eclipses. Star symbols are not to scale. Photocentre geometry ↗

FROM INDIVIDUAL STARS TO THE MILKY WAYFour questions I work on

THE SCIENCE

Small points of light.
Big questions.

01 / HIDDEN COMPANIONS

Is that really
one star?

Two stars can look like a single source. Their combined light carries clues to the companion we cannot resolve.

I use Gaia spectra to search for these hidden pairs, from main-sequence companions to white dwarfs. Finding them helps us understand stellar evolution and interpret the stars we see.

14 millionmain-sequence binary candidates identified from Gaia XP spectra
Read the binary-star study Explore white dwarf companions ↗
EVIDENCE IN THE LIGHT
Colour–magnitude diagrams from the Gaia main-sequence binary study
Searching for unresolved binaries in Gaia's view of the stellar population. Li et al. · A&A, 2025

02 / THE BIRTH MASSES OF STARS

Does nature always
make stars the same way?

How many small stars form for every massive one? This balance—the stellar initial mass function—shapes how we infer the mass and history of galaxies.

Using stellar surveys, I study how that balance changes with chemical composition and time. Our work finds evidence that the recipe for making stars is not universal.

Read the IMF study Follow the story to metal-poor stars ↗
A CHANGING STELLAR RECIPE
Measured IMF slope variation with metallicity and stellar age
The distribution of stellar birth masses carries information about the conditions of star formation. Li et al. · Nature, 2023

03 / THE PHYSICS OF STARLIGHT

How much can
light tell us?

A stellar spectrum encodes temperature, chemical composition, and the conditions in a star's atmosphere.

I develop models that connect those physical conditions to the light we observe, combining stellar physics with machine learning—from differentiable atmospheres to stellar labels across large surveys.

Explore differentiable atmospheres Explore AspGap and Galactic archaeology ↗
CONNECTING PHYSICS AND SPECTRA
Synthetic spectra from Kurucz-a1 and Kurucz-ATLAS compared with an observed spectrum
Comparing synthetic spectra from two atmosphere models with an observed stellar spectrum. Li et al. · 2025

04 / THE INITIAL–FINAL MASS RELATION

What does a star
leave behind?

The initial–final mass relation connects a star's birth mass to the white dwarf it leaves behind. It tells us how much mass stars return to their surroundings, and helps us read the history of stellar populations.

Using 2,543 Gaia wide binaries, we investigate how chemical composition and past mergers affect this relation. Allowing for mergers changes the inferred IFMR; our model constrains metallicity-dependent shifts without establishing a clear overall trend.

View the study

submitted

FROM BIRTH MASS TO STELLAR REMNANT
Initial–final mass relations across metallicity bins, with posterior uncertainty bands from the merger-aware model
The inferred initial–final mass relation across metallicity bins, allowing for a merger contribution. Shaded bands show posterior uncertainty. Li et al. · submitted · Figure 7, left panel

Figures use a dark display palette. Select a figure to view its original colours.

PAPERS & MANUSCRIPTS

First-author papers.

All publications ↗
2026 · submitted

Nature vs. nurture in wide binary white dwarf masses: mergers in hierarchical triples and metallicity

2026 · ApJL

Variations in the Milky Way's Stellar Mass Function at [Fe/H] < −1

 2025 · A&A

Millions of Main-Sequence Binary Stars from Gaia BP/RP Spectra

 2025 · Workshop

Differentiable Stellar Atmospheres with Physics-Informed Neural Networks

 2025 · ApJS

Identification of 30,000 White Dwarf–Main Sequence Binary Candidates from Gaia DR3 BP/RP (XP) Low-resolution Spectra

 2024 · ApJS

AspGap: Augmented Stellar Parameters and Abundances for 37 Million Red Giant Branch Stars from Gaia XP Low-resolution Spectra

 2023 · Nature

Stellar initial mass function varies with metallicity and time

 2021 · ApJS

Stellar Parameterization of LAMOST M Dwarf Stars
Portrait of Jiadong Li

BEHIND THE RESEARCH

Jiadong Li 李佳东

I am a postdoctoral researcher at the Max Planck Institute for Astronomy in Heidelberg.

I studied astronomy at Beijing Normal University and completed my PhD at the National Astronomical Observatories of China. Along the way, I was a visiting researcher at the Flatiron Institute in New York.

I am interested in what large stellar surveys can tell us about the lives of stars and the history of the Milky Way.

Get in touch ↗GitHub ↗ORCID ↗CV ↗