← All detection methods  ·  Exoplanet Archive

Transit timing variations (TTV)

A known planet arrives early or late, because something else is pulling on it.

29 planets found this way → Mass of both planets Orbital resonance Presence of non-transiting planets
A second planet's pull makes the transits arrive early and late in a repeating pattern predicted times (strict clock) how late or early each one was the pattern repeats — something else is tugging

How the measurement works

A lone transiting planet is a metronome: the dip arrives at the same interval, forever. Deviations mean something is interfering.

A second planet tugs on the first, speeding it up on one part of the orbit and slowing it on another. The transits then drift early and late by seconds to hours, in a pattern that repeats on its own longer cycle.

The pattern's shape and period encode the unseen planet's mass and orbit. Crucially, the second planet does not need to transit at all — this method sees planets whose orbits never cross the star from our angle.

The effect is strongest when the two orbits are in resonance, such as one planet completing exactly two orbits for the other's three: the tugs then line up repeatedly instead of averaging away.

The arithmetic

What is measured
Δt = tobserved − tpredicted
Simply how late or early each transit was against a fixed-period prediction. Plotted against time, this is an O−C diagram.
Rough size near resonance
Δt ∼ P · (Mperturber / M*) · f(resonance)
Amplified enormously near resonance — minutes rather than milliseconds — which is why almost every TTV detection involves resonant pairs.

Kepler-9, the first clear case

Two transiting giants periods ~19 and ~39 days close to a 2:1 resonance
Timing drift tens of minutes, growing then reversing over a ~months-long cycle
What it gave true masses of both planets from timing alone, no spectrograph

The value here is the true mass, not a minimum: the geometry is already known from the transits, so there is no sin i left unresolved.

What it does well
  • Reveals planets that never transit and are too light to find by wobble
  • Gives true masses, not lower bounds
  • Free — it reuses transit data already collected
Where it cannot help
  • Needs a known transiting planet to begin with
  • Effectively requires near-resonant orbits for a measurable signal
  • Solutions are often ambiguous — several configurations fit the same timings
  • Needs years of transits to build the pattern
Landmark discovery Kepler-9 b and c (2010) — the first planets with masses measured purely from transit timing.
Instruments Kepler above all; TESS and ground-based follow-up
Archive name Transit Timing Variations — the value in NASA's discoverymethod field

Browse the 29 planets found by this method →

← Back to all eleven methods