Differential clock interferometry in the Imperial AION system
February 01, 2025
We are proud to present our latest work, demonstrating noise rejection in a differential atom interferometer using the strontium-87 clock transition. This is an important step towards the long-baseline atom interferometer AION needs to search for gravitational waves and dark matter, validating the principle of laser noise rejection between two separated interferometers.
Fig. 2. An illustration of the detector's sensitivity to gravitational waves. In the moments before the
final π/2 beam splitter pulse, the two atom interferometers can be treated as freely-
falling atomic clocks (a) accruing phase at a rate ω0. The pulse halts this accrual of phase for the lower
cloud, resulting in an accrual of differential phase (b, d) that continues until the pulse reaches the
second cloud (c). In the proper frame of the bottom cloud (as pictured), the atoms are tidally displaced
by a transient gravitational wave. This has the effect of delaying (or hastening) this second interaction,
imparting (at leading order) a detectable differential phase. Crucially, any
phase noise due to the laser pulse itself is strongly suppressed in the differential measurement since
it impacts both interferometers equally. The mechanism for sensitivity to dark matter (not pictured)
is similar, but results from modulation of ω0 instead of L.
This figure is reproduced from arXiv:2504.09158.
Fig. 4 (a) Clock atom interferometry fringes in the top and bottom atom clouds with a fixed Stark
shift applied to the top cloud. The lower (upper) plot shows the fringes with (without) the
addition of artificial laser noise (see Methods for details). (b) A correlation plot—or Lissajous figure—
of the top and bottom atom interferometer signals, with (blue) and without (red) added laser noise.
Despite the entirely obscured interferometer fringes, the differential measurement rejects laser noise,
recovering a clean ellipse. (c) Overlapping Allan deviations of the differential phase calculated from 112 independent
ellipse fits to consecutive bins of 100 shots each, comprising a total of 11 200 interferometer samples
taken over 18.6 h. Using absorption spectroscopy (see Methods), we measure 1280(130) in the top trap
and 1600(160) atoms in the bottom trap, forming the basis of our estimate of the Standard Quantum
Limit (SQL, black). (c inset) The standard deviations σ of the differential phase measurements for
the Low Laser Noise (LLN) and High Laser Noise (HLN) datasets are consistent with the SQL.
This figure is reproduced from arXiv:2504.09158.
