ETH Zurich’s LIFE project: How it could search for alien life

One of astronomy’s biggest questions may be answered by a faint heat signal: radiation from a small planet orbiting a nearby star. The ETH Zurich-led LIFE initiative aims to capture such a signal and split it into a spectrum. Researchers could then investigate whether the planet has an atmosphere, whether liquid water might be possible and whether several chemical clues together point to biological activity.

That capability does not exist yet. LIFE—short for “Large Interferometer for Exoplanets”—is an international mission concept and technology programme, not a mission that the European Space Agency (ESA) has already selected or funded. The distinction matters: the science case is concrete and key components are being tested, but construction and launch have not been approved.

Five spacecraft would form one virtual telescope

LIFE’s baseline design uses four free-flying collector telescopes separated by tens to hundreds of metres. A fifth spacecraft at the centre would combine their light. Working as a formation, the five satellites would provide the resolving power of a much larger virtual telescope.

The observatory is designed for the mid-infrared, roughly 4 to 18.5 micrometres. Temperate planets emit much of their detectable thermal radiation at these wavelengths, while molecules leave characteristic absorption features. A spectrum could therefore reveal information about a planet’s temperature, surface conditions and atmospheric composition.

The main targets would be nearby rocky worlds roughly comparable in size and temperature to Earth. “Earth-like” does not automatically mean habitable—and certainly not inhabited. LIFE would first measure the physical and chemical conditions on those planets.

The key technique: cancelling the star’s light

The central problem is contrast. Even a nearby Earth-sized planet is extraordinarily faint beside its star. LIFE therefore relies on nulling interferometry. In the combiner spacecraft, the light waves collected by the four telescopes would be shifted so that the on-axis starlight largely cancels itself through destructive interference. Light from a planet at a slight angular offset would arrive with a different phase and remain partly detectable.

Put simply, the instrument would create an artificial blind spot at the position of the bright star. Only then could the much weaker thermal signal from a planet emerge. Achieving this requires the spacecraft to hold a precise formation and the optical path lengths to remain stable to tiny fractions of a wavelength.

A test with Earth data provides a reality check

A team from ETH Zurich and the University of Zurich tested whether the proposed method could recognise a habitable world by using the only inhabited planet we know: Earth. For a 2024 study in The Astrophysical Journal, the researchers used real measurements from NASA’s Aqua Earth-observation satellite. They converted the data into an averaged infrared spectrum resembling what LIFE might receive from a “second Earth” about 30 light-years away.

Their analysis recovered carbon dioxide, water, ozone and methane, along with surface conditions compatible with liquid water. The result was robust across different viewing geometries. The test also exposed a limitation: seasonal changes were difficult to detect at the simulated data quality.

This does not prove that LIFE will be built, nor that every inhabited world would be recognised unambiguously. It does show that simulated LIFE observations based on real Earth data can retrieve essential properties of our planet.

NICE shows both the progress and the remaining gap

The next question is whether the required starlight suppression can be achieved with sufficient stability and sensitivity. ETH Zurich is addressing this with the NICE test bench—the “Nulling Interferometry Cryogenic Experiment”. Its long-term operating target is 15 kelvin, about minus 258 degrees Celsius. Cooling the instrument reduces its own unwanted thermal radiation.

ETH’s current project page reports an instantaneous null depth of about 2 × 10⁻⁵ and a one-minute average of roughly 5 × 10⁻⁵. The smaller the number, the more completely the starlight is suppressed. The target is below 10⁻⁵, with optical path-length stability better than one nanometre.