← All detection methods  ·  Exoplanet Archive

Astrometry

Measure the star's position precisely enough and you see it trace a loop.

6 planets found this way → True mass Full three-dimensional orbit
The star's path across the sky is a straight drift with small loops superimposed, one per orbit the drift, if the star were alone what is measured: the same drift, with one loop per orbit

How the measurement works

Radial velocity catches the part of the star's wobble along our line of sight. Astrometry catches the other part — the side-to-side motion across the sky.

Measure the star's position against distant background stars over years, and instead of a straight drift you find a drift with small loops superimposed: one loop per orbit of the unseen planet.

Because the motion is measured in two dimensions rather than one, the orbit's tilt comes out of the fit. That means a true mass, not the "at least" of radial velocity.

This is the oldest proposed method — nineteenth-century astronomers tried it, and a long series of claimed detections through the twentieth century all turned out to be instrumental. Only with Gaia's microarcsecond precision has it begun to deliver.

The arithmetic

Angular size of the wobble
α = (Mp / M*) · (a / d)
a in AU, d in parsecs, result in arcseconds. Unlike radial velocity, this grows with orbital distance — so it favours wide orbits, the opposite bias.

The Sun seen from 10 parsecs (33 light years)

Jupiter's pull (1/1047) × (5.2 / 10) 497 microarcseconds
Earth's pull (1/333000) × (1 / 10) 0.3 microarcseconds
Gaia's per-measurement precision ~20–30 microarcseconds Jupiter: comfortable. Earth: no.

497 microarcseconds is roughly the angle a one-euro coin subtends from 10 kilometres away. That is the easy case.

What it does well
  • Gives true mass and the full 3-D orbit
  • Favours wide orbits, complementing radial velocity's inner bias
  • Gaia is measuring a billion stars, so the yield should grow sharply
Where it cannot help
  • Brutal precision requirements — a century of false claims before the first real ones
  • Needs years of observation to cover an orbit
  • Only six planets in the catalogue so far
Landmark discovery Gaia-4b (2025) — among the first robust astrometric detections, from Gaia's own data.
Instruments Gaia above all; Hipparcos historically, VLTI/GRAVITY for individual systems
Archive name Astrometry — the value in NASA's discoverymethod field

Browse the 6 planets found by this method →

← Back to all eleven methods