Hot Cools First: Quantum Mpemba Effect Without Bath Engineering ✛
Context
The classical Mpemba effect, where hot water freezes faster than cold, has been known since antiquity, but a rigorous explanation emerged only recently within Markovian dynamics. In the quantum domain, the effect has been predicted theoretically and observed in trapped ions, but there the relaxation proceeded through artificially engineered decay channels rather than natural thermalization. Our work joins a line of research seeking the Mpemba effect in natural processes: neighboring studies explore it in quantum simulations, random circuits, and disordered systems. We turn to nuclear spins in a liquid, a system that thermalizes on its own without external intervention, and show that the effect arises naturally.
Methods
The experiment was performed on a two-spin system of protons in 2-chloroacrylonitrile dissolved in dimethyl sulfoxide, using a 500 MHz NMR spectrometer at room temperature. The dominant relaxation mechanism is the dipolar interaction between nuclear spins, described by the GKSL equation with a Liouvillian that has a block structure (the Redfield 'kite' structure). The authors theoretically derived that under certain conditions, the population dynamics reduces to independent relaxation of two qubits with different effective bath temperatures. They then prepared two initial states: one closer to thermal equilibrium, the other farther, but with zero overlap with the slowest decay mode. After preparation, the states were left to evolve without external pulses for up to 29 seconds, significantly exceeding the longitudinal relaxation time T1 = 5.7 s. Relaxation was monitored by measuring the trace distance to the thermal state.
Results
Experimentally confirmed: the state initially farther from equilibrium ($ρ_f$) reaches the thermal state faster than the closer one ($ρ_n$). For θ = 70° and 60°, the distance curves cross, showing the Mpemba effect. Measurement of overlaps with decay modes showed that $ρ_n$ overlaps with both the slow (λ1) and fast (λ3) modes, while $ρ_f$ overlaps only with the fast mode. This explains the acceleration: $ρ_f$ decays through the fast channel (η±(2)), bypassing the slow one (η(0)). Furthermore, a 'genuine quantum Mpemba effect' is demonstrated: the state with higher free energy ($ρ_f$) reaches equilibrium faster than the state with lower free energy ($ρ_n$). Theoretical predictions agree with the data for bath correlation time $τ_c$ = 2.1 ps and dipolar coupling constant b = 5.903 kHz.
Implications
This work shows that the quantum Mpemba effect is not an artifact of specially engineered systems but a natural phenomenon arising in ordinary thermal relaxation. Unlike neighboring studies where the effect was achieved through engineered decay channels or non-standard baths, here relaxation proceeds through the standard dipolar mechanism in a liquid. This settles a long-standing question of whether the effect can occur in natural thermalization and moves it from the realm of curiosities to fundamental properties of open quantum systems.
Future development
Understanding the conditions for the Mpemba effect in natural relaxation opens the way to purposeful use of this phenomenon. One can look for similar effects in more complex spin systems, including solid-state qubits, where relaxation is often a limiting factor. It may be possible to develop methods for accelerated initialization of quantum states using natural relaxation channels without additional control.
Impact
The results are important for quantum technologies, especially for qubit initialization and reset, where reducing relaxation time directly decreases idle time between computational cycles. They also affect fundamental non-equilibrium statistical mechanics, refining the limits of applicability of the Markovian description.
Next steps
The next step could be to investigate the Mpemba effect in systems with many spins and in more complex baths, for example with non-Markovian dynamics. It would also be interesting to test whether the effect can be used to speed up cooling of quantum simulators.
Key open problems
This work is directly related to the unresolved problem of describing relaxation in open quantum systems: how anomalously fast paths to equilibrium arise and whether they can be predicted from the structure of the Liouvillian. It also touches on the question of how general the Markovian approximation is for real physical systems.
The knowledge machine's view
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The paper experimentally demonstrates the quantum Mpemba effect in the natural thermal relaxation of nuclear spins in a liquid, without engineering the bath. The authors used dipolar relaxation in a two-spin system, derived the conditions for the effect to occur, and prepared states with different overlaps with the decay modes. Measurements showed that a state initially farther from equilibrium reaches thermal equilibrium faster, which is explained by its lack of overlap with the slow mode.
Where its strength is
The work shows that the quantum Mpemba effect can arise under natural conditions without specially designing relaxation channels, making it more universal and potentially applicable for speeding up the initialization of quantum devices. A strong point is the clear link between the structure of the initial state and the relaxation rate, providing a practical criterion for choosing states.
Where it could go next
- It might be interesting to examine how the Mpemba effect manifests in systems with a more complex level structure or with controlled interaction with the bath. Nearby is work on the charge palette of a quantum bath, which investigates how different decoherence channels affect dynamics.based on: Conductor of Heat Death: The Charge Palette of a Quantum Bath arXiv:2505.02040v2
- One could also study the connection between the Mpemba effect and entanglement or other quantum correlations. There is a paper in the archive that considers the slowdown of entanglement propagation, which might offer new insights into the role of correlations in accelerated relaxation.based on: Entanglement Takes the Slow Lane arXiv:2608.07465v1
Worth trying
- Test whether the Mpemba effect persists when the temperature or solvent viscosity is changed, which would allow assessing its robustness to bath parameters.
- Use state control methods to create initial states with a desired overlap with decay modes and test the theory's predictions over a wider range.
- Apply a similar analysis to systems with a larger number of spins to determine how the effect scales and whether it can be used to speed up the initialization of multi-qubit registers.
Why this may matter
If the Mpemba effect is indeed a general property of natural thermal relaxation, it could lead to practical methods for accelerating the equilibration of quantum systems, which is important for initializing quantum computers and other devices. Furthermore, understanding the mechanism could help in managing decoherence in open quantum systems.
Nearby works in our archive
- Quantum Mpemba Effect: Controlled Coolingsimilarity 0.58
- Quantum Leap to Calmsimilarity 0.58
- Hot overtakes warm in a quantum networksimilarity 0.57
- Hotter Faster? It's All in the Wallsimilarity 0.57
- Entanglement Takes the Slow Lanesimilarity 0.57
- Conductor of Heat Death: The Charge Palette of a Quantum Bathsimilarity 0.57
The archive is sparse around this work: the closest review is only 0.59, against a usual 0.64. It looks like few have drilled nearby.
The classical Mpemba effect is named after Tanzanian schoolboy Erasto Mpemba, who in 1963 noticed that hot ice cream freezes faster than cold. His observation was long dismissed as an experimental error, but now the effect has a rigorous quantum generalization.
In Kurt Vonnegut's novel 'Cat's Cradle', there is ice-nine, which makes water freeze at room temperature. The Mpemba effect is a milder but real version of paradoxical water behavior, only in the quantum world.
Key numbers
- relaxation time T1: 5.7 s
- observation time: 29 s
- NMR frequency: 500 MHz
- dipolar coupling constant b: 5.903 kHz
- bath correlation time τ_c: 2.1 ps