A theoretical team has proposed a way to trigger and precisely control the imaginary-time Mpemba effect in quantum spin systems using spatially modulated local rotations. The distinctive feature of the protocol is that it leaves the initial bipartite entanglement spectrum and correlation magnitudes unchanged, so the only difference between states is the redistribution of spectral weights. That allows an unambiguous isolation of what drives the anomalous relaxation in which a "hotter" state converges to equilibrium before a "colder" one.

The Mpemba effect, known for decades in classical thermodynamics, describes how a system farther from equilibrium can relax faster than one initially closer to it. Its imaginary-time version, relevant to quantum ground-state preparation algorithms, has proved difficult to control cleanly. The authors describe the applied twists within a compact-boson theory, in which the rotations act as mode selectors: higher-harmonic twists exclusively populate high-energy standing waves, structurally bypassing the slow-decaying infrared modes that slow convergence.

The work derives exact analytical crossing times for the relaxation reversals and shows that deliberately reintroducing the slow-mode population systematically delays the effect. The mechanism is validated beyond fixed-point theory through calculations on interacting XXZ chains and ladders. In addition, a twelve-qubit simulation of variational quantum imaginary-time evolution reproduces the reversals and their controlled delay, offering a concrete route to implement the protocol in realistic quantum circuits.

Beyond the fundamental curiosity, the proposal establishes an exact and strictly controllable starting point for studying anomalous non-equilibrium relaxation and harnessing it in quantum ground-state preparation, a routine step in quantum simulation and variational algorithms. The validation in twelve-qubit simulations suggests experimental verification could come in moderate-size spin or superconducting processors, although the work presented is theoretical and numerical in nature.