Robust double Bragg diffraction via detuning control

TL;DR

This paper develops a theoretical and numerical framework for double Bragg diffraction in atom interferometry, introducing detuning control methods that significantly improve robustness against polarization errors and Doppler effects.

Contribution

It presents an effective two-level Hamiltonian model, extends it to five levels for Doppler detuning, and introduces AI-aided detuning control protocols for enhanced robustness.

Findings

Achieves over 99.5% efficiency with linear detuning sweep.

Develops AI-based protocol with 99.92% efficiency for finite momentum width.

Demonstrates robustness against polarization errors up to 10%.

Abstract

We present a theoretical model and numerical optimization of double Bragg diffraction, a widely used technique in atom interferometry. We derive an effective two-level-system Hamiltonian based on the Magnus expansion in the so-called "quasi-Bragg regime", where most Bragg-pulse atom interferometers operate. Furthermore, we extend the theory to a five-level description to account for Doppler detuning. Using these derived effective Hamiltonians, we investigate the impacts of AC-Stark shift and polarization errors on the double Bragg beam-splitter, along with their mitigations through detuning control. Notably, we design a linear detuning sweep that demonstrates robust efficiency exceeding 99.5% against polarization errors up to 8.5%. Moreover, we develop an artificial intelligence-aided optimal detuning control protocol, showcasing enhanced robustness against both polarization errors and Doppler effects. This protocol achieves an average efficiency of 99.92% for samples with a finite momentum width of 0.05$\hbar k_L$ within an extended polarization error range of up to 10%.