The 2015 gravitational wave discovery was made in a quiet patch of space. A new framework, published in Physical Review Letters, pins down what such a signal really means when the universe around it is expanding and lumpy, an answer that will shape how LIGO, the planned LISA…
The 2015 detection of gravitational waves felt like a closing chapter. Two black holes spiraling into each other 1.3 billion light-years away sent a ripple through detectors in Louisiana and Washington, and a new kind of astronomy began. A decade on, the same kind of signal has been caught by LIGO, Virgo, and KAGRA dozens of times. The picture of what a gravitational wave is seemed settled.
It is not. Hidden inside the standard treatment is an assumption so rarely examined that it has slipped past a generation of theorists: that the universe around the wave is quiet, nearly empty, and still. That assumption holds for two black holes merging in deep space. It frays the moment the wave is part of a cosmos that is expanding, lumpy, and shaped by gravity from everything it contains. Every interpretation of a LIGO signal already includes a quiet patch of spacetime built into the math. Every future measurement planned for the space-based LISA detector, and for pulsar timing arrays listening to the ticking of distant neutron stars, will sit inside an expanding universe that the standard definition cannot fully describe.
A team at Leibniz University Hannover has now written down, for what appears to be the first time, what a gravitational-wave detector actually measures in that case. Their paper, "Observable Gravitational Wave Strain at Second Order," by Guillem Domènech, Shi Pi, and Ao Wang, appeared in Physical Review Letters on 3 June 2026. Domènech is at the university's Institute of Theoretical Physics.
The framework is built on two freely falling test masses, modeled as atomic clocks, connected by a beam of light. The observable is concrete: the shift in light travel time and frequency between the two clocks. From that, the authors derive what a detector would register in an expanding, fluctuating cosmos, carried out to second order in the size of those fluctuations. The point is not to build a new instrument. It is to fix the meaning of the signal one.
"We calculate these quantities exactly within an expanding spacetime and distinctly isolate what is genuinely measurable from effects that rely on the mathematical description," Domènech told SciTechDaily, which reported the result. "This ensures that theoretical predictions for future experiments are rigorous and reliable."
The distinction matters because every model of the universe comes with a choice of coordinates, and a gravitational wave is, strictly, the part of the spacetime signal that is not just an artifact of how one chose to draw the map. In a quiet, isolated patch of spacetime, the wave and the artifact separate cleanly. In an expanding cosmos, they do not, and prior treatments have leaned on assumptions about the cosmic background that mix the two. The new work builds a detector-based, coordinate-independent definition that does not require that separation up front. In the limit where the surrounding cosmos fades out, it reduces to the familiar interferometer signals LIGO already records.
That last point is why the result reaches beyond theory. The same definition will be used to read signals that the LIGO-style detectors cannot reach: primordial gravitational waves from inflation, phase transitions in the early universe, and the slow background hum that pulsar timing arrays are already seeing hints of. LISA, the European Space Agency's planned space-based detector scheduled for the 2030s, will sit in a gravitational environment shaped by the whole cosmos. Reading its data correctly depends on a definition that holds when the universe is in motion.
What changes now is not the hardware. It is the standard against which future data will be compared. A signal once defined by a quiet patch of spacetime has a rigorous counterpart in a living one, and the first careful derivation of that counterpart is in print.