Closed-Loop Matter-Wave Interferometer

• Closed-loop matter-wave interferometers are devices that split, redirect, and recombine matter waves along closed trajectories in various spaces, enabling scalable sensitivity.
• They employ diverse geometries, including Bloch-Bragg-Bloch, standing wave, and guided-wave designs, to maximize enclosed area and control phase accumulation for inertial and gravitational sensing.
• Advanced feedback stabilization and optimized beam-splitting techniques mitigate closure errors and enhance measurement precision, paving the way for quantum-enhanced inertial sensors.

A closed-loop matter-wave interferometer is an interferometric device in which matter-wave amplitudes are split, redirected, and recombined so that the relevant trajectories form a loop in real space-time, momentum-time, phase space, or Hilbert space, thereby enclosing a finite area or returning to the same motional state after a controlled cycle. In the literature considered here, the term also appears in a second, operational sense: an interferometer whose intrinsically periodic phase response is actively stabilized by feedback, so that the physical signal is encoded in control parameters rather than in open-loop fringe amplitude. Across these usages, loop closure is the organizing principle behind sensitivity scaling, common-mode rejection, and the control of recombination fidelity (0903.4192, Amit et al., 2019, Jia et al., 16 Mar 2026).

1. Definitions and forms of loop closure

In light-pulse atom interferometry, loop closure usually denotes a geometry in which the atomic wave packets diverge and reconverge while enclosing a finite area. In Ramsey–Bordé and Mach–Zehnder implementations, this area may be described in space-time or momentum-time, and the enclosed area scales with momentum splitting and interrogation time. A representative formulation is the Bloch-Bragg-Bloch Ramsey–Bordé interferometer, in which the arms are split, accelerated, redirected, and recombined so that the trajectories form a closed loop in momentum-time and the enclosed area becomes scalable through large momentum transfer (0903.4192).

A stricter notion of closure appears in full-loop Stern–Gerlach interferometry. There, the two arms are required to coincide both in phase space and in internal-state subspace at recombination. In the $T^3$-Stern–Gerlach matter-wave interferometer, four magnetic field gradient pulses generate state-dependent forces such that $\mathcal{P}_1(T)=\mathcal{P}_2(T)$ and $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$, so the output depends only on the relative phase $\delta\Phi = \Phi_1(T) - \Phi_2(T)$ (Amit et al., 2019).

A related Hilbert-space notion of closure arises in trapped mesoscopic interferometry. In a levitated nanodiamond scheme with an embedded NV center, the center-of-mass motion follows spin-conditioned phase-space trajectories that return exactly to the initial coherent state after one trap period $t_0 = 2\pi/\omega_z$, leaving only a relative phase $\Delta\phi_{\text{Grav}}$ in the spin sector. The interferometer is therefore closed in center-of-mass Hilbert space even though the readout is purely spin-Ramsey (Scala et al., 2013).

Ring-based orbital-angular-momentum interferometry provides a topological version of the same idea. A localized state in a toroidal trap evolves under $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$, splits by free evolution at $T_{\text{rev}}/2$, and recombines after another $T_{\text{rev}}/2$, so the loop is defined by periodic motion on a closed path rather than by discrete beam splitters (Kiałka et al., 2020).

2. Canonical geometries and atom-optical implementations

One major route to closed-loop interferometry is to enlarge the enclosed area by increasing momentum splitting. In the Bloch-Bragg-Bloch architecture, coherent acceleration by an optical lattice is combined with Bragg diffraction, while both arms remain in the same internal state. Four Bloch-Bragg-Bloch beam splitters in a Ramsey–Bordé geometry produced $15\%$ contrast at $\mathcal{P}_1(T)=\mathcal{P}_2(T)$0 splitting, and single beam splitters reached $\mathcal{P}_1(T)=\mathcal{P}_2(T)$1; the same work discusses prospects for reaching $\mathcal{P}_1(T)=\mathcal{P}_2(T)$2s of $\mathcal{P}_1(T)=\mathcal{P}_2(T)$3 (0903.4192).

A second route uses standing light waves as all atom-optical elements. In a planar Bragg interferometer, an atomic beam is split into two arms separated by angle $\mathcal{P}_1(T)=\mathcal{P}_2(T)$4, reflected by two middle standing waves, and recombined by a final standing wave, creating a diamond-like closed region with area

$\mathcal{P}_1(T)=\mathcal{P}_2(T)$5

This is a closed-path matter-wave interferometer in the usual Mach–Zehnder sense, but implemented entirely with standing light waves rather than material gratings (Gao et al., 2011).

Guided-wave realizations replace ballistic free fall with confinement. A matter-wave analog of a fiber-optic gyroscope confines Bose-condensed $\mathcal{P}_1(T)=\mathcal{P}_2(T)$6 in a tight optical waveguide, uses Bragg pulses to split, reflect, and recombine the packets, and moves the waveguide transversely so that the trajectories enclose an effective area of $\mathcal{P}_1(T)=\mathcal{P}_2(T)$7, with an average fringe contrast of $\mathcal{P}_1(T)=\mathcal{P}_2(T)$8 and underlying contrast up to $\mathcal{P}_1(T)=\mathcal{P}_2(T)$9 (Krzyzanowska et al., 2022). A TOP-trap Sagnac interferometer pushes the closed-loop idea further by guiding four packets around one or more full circular orbits, with an effective Sagnac area of $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$0 per orbit and interference with $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$1 visibility after two orbits at a total interrogation time of $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$2 s (Beydler et al., 2023).

Integrated trapped interferometers realize closure without extended free propagation. On an atom chip, a full Mach–Zehnder sequence was implemented with trapped Bose–Einstein condensates in a double-well potential, where splitting, differential phase imprinting, and non-adiabatic recombination all occur inside a single guided structure (Berrada et al., 2013). In waveguide geometries with attractive interactions, bright-solitonic interferometers use Bragg pulses as beam splitters and mirrors while the trap forces the two arms to return to the barrier region, so the loop is closed by dipole oscillation rather than by geometric circulation (McDonald et al., 2014, Polo et al., 2013).

3. Phase accumulation, enclosed area, and scaling laws

For inertial sensing, the central quantity is the enclosed area. In scalable light-pulse interferometers, the area obeys

$\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$3

so increasing $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$4 directly increases sensitivity (0903.4192). In guided Sagnac interferometers, rotation produces

$\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$5

and in the multi-orbit TOP-trap geometry

$\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$6

making explicit the linear scaling with orbit number $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$7 and orbit radius $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$8 (Krzyzanowska et al., 2022, Beydler et al., 2023).

Closed loops need not exhibit the usual $\mathcal{Z}_1(T)=\mathcal{Z}_2(T)$9 scaling. In the full-loop Stern–Gerlach interferometer, the phase for $\delta\Phi = \Phi_1(T) - \Phi_2(T)$0 is

$\delta\Phi = \Phi_1(T) - \Phi_2(T)$1

and for $\delta\Phi = \Phi_1(T) - \Phi_2(T)$2 it reduces to a pure cubic scaling with total interferometer time, $\delta\Phi = \Phi_1(T) - \Phi_2(T)$3 (Amit et al., 2019). In ring OAM interferometry, the revival time

$\delta\Phi = \Phi_1(T) - \Phi_2(T)$4

sets the effective beam-splitting and recombination times, while a gauge flux rotates the revival by

$\delta\Phi = \Phi_1(T) - \Phi_2(T)$5

so the interferometer measures geometric phases tied directly to the loop topology (Kiałka et al., 2020).

Gravitational-wave proposals emphasize yet another scaling regime. In the standing-light-wave interferometer, the high-frequency response is dominated by terms proportional to $\delta\Phi = \Phi_1(T) - \Phi_2(T)$6, that is, to $\delta\Phi = \Phi_1(T) - \Phi_2(T)$7, rather than to $\delta\Phi = \Phi_1(T) - \Phi_2(T)$8 itself; the leading contribution is

$\delta\Phi = \Phi_1(T) - \Phi_2(T)$9

This contrasts with optical-displacement sensors and motivates closed-path matter-wave architectures for the $t_0 = 2\pi/\omega_z$0–$t_0 = 2\pi/\omega_z$1 Hz band (Gao et al., 2011). A single-photon interferometer based on the $t_0 = 2\pi/\omega_z$2 optical clock transition of $t_0 = 2\pi/\omega_z$3 reaches the gravimeter and gravity-gradiometer regimes without reduction of interferometric contrast up to $t_0 = 2\pi/\omega_z$4 ms, limited by geometric constraints of the apparatus (Hu et al., 2017).

4. Closure errors, interactions, and interpretive pitfalls

Because closed-loop interferometers amplify inertial phase, they also amplify sensitivity to vibrations, field inhomogeneity, and imperfect recombination. A scalable large-momentum example addresses this by running simultaneous conjugate interferometers and extracting the differential phase from an ellipse built from normalized output populations $t_0 = 2\pi/\omega_z$5 and $t_0 = 2\pi/\omega_z$6; common-mode vibrations move points around the ellipse without changing its eccentricity (0903.4192). Guided-wave Sagnac devices face analogous issues in the trap: residual anisotropy $t_0 = 2\pi/\omega_z$7, cross-term $t_0 = 2\pi/\omega_z$8, and anharmonic coefficients $t_0 = 2\pi/\omega_z$9 deform trajectories and limit multi-orbit visibility unless tilt and symmetry are tuned in situ (Beydler et al., 2023).

Strong interactions create a separate closure problem. In trapped $\Delta\phi_{\text{Grav}}$0 Feshbach-molecule interferometers, interactions shift the motional energies, reduce the Ramsey frequency, add a phase during the lattice pulses, and induce dephasing through nonuniformity; in the Michelson-type realization, coherence is observed in the presence of significant interaction, however coherence degrades with increasing interaction strength (Li et al., 2024). In solitonic interferometers the same nonlinearity can instead suppress dispersion. A bright-solitonic Mach–Zehnder in a horizontal optical waveguide showed that tuning the $\Delta\phi_{\text{Grav}}$1-wave scattering length to a small negative value significantly increased fringe visibility even compared with a non-interacting cloud, precisely because the attractive interaction balances inherent matter-wave dispersion (McDonald et al., 2014).

Massive Stern–Gerlach interferometers make the distinction between dephasing and imperfect loop closure especially explicit. For linear spin-independent noise, spin contrast loss does not depend on the initial thermal state of the matter wave function; the noise generates only a random phase. For linear spin-dependent noise, by contrast, spin contrast loss does depend on the initial thermal occupation of the quantum state because the two spin branches acquire a displacement mismatch in phase space (Zhou et al., 17 Mar 2025).

A recurring interpretive pitfall concerns closed-loop electronic interferometers. In a gate-defined GaAs/AlGaAs Aharonov–Bohm device with very small SOI, the observed beating of $\Delta\phi_{\text{Grav}}$2 oscillations, the $\Delta\phi_{\text{Grav}}$3 nodes, and multiple $\Delta\phi_{\text{Grav}}$4 peaks in the Fourier spectrum were interpreted without resorting to the SOI effect, by the existence of two-dimensional multiple longitudinal modes in a single transverse subband (0806.1595). The broader implication is that complex beating in a closed-loop interferometer is not, by itself, evidence for Berry-phase pickup or strong spin-orbit interaction.

5. Feedback-closed operation and quantum measurement

A distinct contemporary meaning of “closed-loop matter-wave interferometer” is feedback-stabilized operation beyond the half-fringe limit. In a dual-channel continuous $\Delta\phi_{\text{Grav}}$5 atomic beam interferometer, two spatially separated Raman Mach–Zehnder interferometers produce phases $\Delta\phi_{\text{Grav}}$6 and $\Delta\phi_{\text{Grav}}$7, and the combinations

$\Delta\phi_{\text{Grav}}$8

diagonalize acceleration and rotation. Independent feedback on $\Delta\phi_{\text{Grav}}$9 and $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$0 then locks the acceleration and rotation channels, with estimators

$\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$1

This realizes unambiguous measurements up to $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$2 in rotation and $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$3 in acceleration while maintaining high fringe contrast, corresponding to nearly two orders-of-magnitude extension beyond the conventional half-fringe limit, with long-term stability of $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$4 for rotation and $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$5 for acceleration at $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$6 (Jia et al., 16 Mar 2026).

The quantum-measurement limit of such operation can be framed by linear quantum measurement theory. In that formulation, the optical field is the probe, the atoms are the detector, and the measured population difference contains signal, imprecision, and back-action terms. The resulting Standard Quantum Limit for the phase-estimation error is obtained from a balance of atom shot noise, optical shot noise, and light-mediated back-action; the optimized scaling is

$\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$7

This establishes a direct parallel with laser interferometer gravitational-wave detectors and makes clear that feedback reshapes the transfer functions but does not remove the imprecision–back-action tradeoff (Ma et al., 2019).

6. Applications, limitations, and research directions

The application space of closed-loop matter-wave interferometry spans inertial sensing, gravitational-wave detection, surface science, interaction metrology, and quantum-enhanced sensing. Large-momentum Ramsey–Bordé systems are explicitly connected to gravitational wave sensors, Lense–Thirring measurements, equivalence-principle tests with $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$8, atom neutrality tests, and precision measurements sensitive to supersymmetry (0903.4192). Standing-light-wave interferometers and clock-transition interferometers are discussed as routes toward gravitational-wave detection in frequency ranges not covered by existing optical detectors (Gao et al., 2011, Hu et al., 2017). Full-loop Stern–Gerlach devices are presented as high-precision surface probes at very close distances, while ring OAM interferometers are proposed for gauge-field sensing and scattering-length measurements (Amit et al., 2019, Kiałka et al., 2020). Integrated atom-chip Mach–Zehnder systems exploit nonlinearity to generate non-classical states having reduced number fluctuations inside the interferometer and thereby pave the way for integrated quantum-enhanced matter-wave sensors (Berrada et al., 2013).

The main practical limitations remain highly platform-specific. Scalable light-pulse systems are limited by frequency-ramp resolution and wavefront distortions in the Bloch sequences (0903.4192). Guided-wave gyroscopes are presently limited mainly by phase noise caused by mechanical vibrations of the optical components, even though the interferometer phase noise falls with averaging time $\mathsf{H} = \mathsf{L}_z^2/(2mR^2)$9 as $T_{\text{rev}}/2$0 for $T_{\text{rev}}/2$1 up to $T_{\text{rev}}/2$2 seconds (Krzyzanowska et al., 2022). Multi-orbit TOP-trap Sagnac devices are limited by residual tilt, trap asymmetry, and anharmonicity (Beydler et al., 2023). Trapped molecular interferometers are limited by dephasing, collisional effects, and coherence degradation at strong interaction (Li et al., 2024). In chip-based trapped interferometers, the non-adiabatic recombiner still falls short of the $T_{\text{rev}}/2$3 contrast predicted by simulation, while the recombination-based phase readout currently remains noisier than time-of-flight fringe analysis (Berrada et al., 2013).

The literature explicitly discusses routes toward larger enclosed areas and more robust loop closure: $T_{\text{rev}}/2$4s to $T_{\text{rev}}/2$5s of $T_{\text{rev}}/2$6 in large-momentum beam splitters, multiple guided orbits in compact Sagnac geometries, and arrays or hybrid light–matter interferometers in gravitational-wave concepts (0903.4192, Beydler et al., 2023, Gao et al., 2011). Taken together, this suggests a convergence of three design principles: scalable momentum separation, recirculating guided trajectories, and feedback-stabilized readout. In that regime, “closed-loop matter-wave interferometer” no longer denotes a single architecture, but a family of devices in which geometric closure, phase-space closure, and control-loop closure are engineered simultaneously.