A new study reveals that quantum entanglement, a fragile resource crucial for quantum computing and secure communication, can withstand local thermalization processes that were thought to destroy it. This finding challenges conventional wisdom in quantum thermodynamics and suggests that shared classical randomness can protect quantum correlations, even as systems heat up locally. The research, conducted by Chung-Yun Hsieh, Matteo Lostaglio, and Antonio Acín from ICFO – Institut de Ciències Fotòniques, demonstrates that entanglement preservation is possible at all nonzero temperatures, with implications for quantum error correction and heat management in quantum devices.
The key finding is that entanglement preserving local thermalizations (EPLTs) exist universally for finite-energy systems at any nonzero temperature. The researchers proved that two spatially separated agents, each thermalizing their subsystem to a predefined thermal state using only local operations and shared classical randomness, can still output an entangled state for some inputs. This means that even after local heating processes that bring each subsystem to thermal equilibrium, quantum entanglement between them can survive, contrary to the expectation that local thermalization always destroys such correlations.
The methodology involves constructing explicit protocols based on local operations and shared randomness (LOSR). One protocol, denoted E_γ, combines a twirling operation—which applies coordinated random local unitaries—with a probabilistic mixing step that ensures local thermalization. The twirling operation preserves the overlap with maximally entangled states, while the mixing adjusts local states to match thermal marginals. The researchers showed that for any local dimension d and thermal states τ_A and τ_B, setting γ = dP_min (where P_min is the smallest eigenvalue among τ_A and τ_B) guarantees that E_γ is an EPLT. This protocol was analyzed using entanglement criteria like the positive partial transpose (PPT) condition, confirming that outputs remain entangled for inputs with sufficient initial entanglement.
Results analysis from the paper shows that for two-qubit systems with identical thermal states not equal to the ground state, the output of E_γ is entangled if and only if the input's overlap with the Bell state |Φ⁺⟩ exceeds 1/2. The researchers quantified preserved entanglement using the fully entangled fraction (FEF), proving that F_max[E_γ(ρ_AB)] ≥ ⟨Φ⁺|ρ_AB|Φ⁺⟩ for all inputs. This bound indicates that the protocol maintains a significant level of entanglement, with higher initial entanglement leading to stronger preserved correlations. The study also extended these findings to multipartite systems, showing that genuine multipartite entanglement, such as in GHZ states, can survive local thermalizations under similar conditions.
In practical terms, this discovery matters because it redefines how we understand quantum resource resilience in noisy environments. For everyday readers, it means that future quantum technologies—like quantum computers or secure networks—could be more robust against thermal noise, which is a common source of decoherence. The thermodynamic interpretation suggests that classically correlated heat baths can speed up local thermalization while preserving entanglement, potentially leading to more efficient quantum thermal machines or error mitigation strategies in quantum devices.
Limitations noted in the paper include the exclusion of zero-temperature cases with non-degenerate ground states, where no entanglement can be preserved because local thermal states are pure and uncorrelated. Additionally, the protocols rely on shared randomness, which may pose implementation challenges in real-world scenarios. The researchers also highlight that the exact twirling operation requires infinite time in theory, though finite approximations achieve exponential precision, and the speed-up mechanism depends on parameters like thermalization time scales relative to unitary operation times.
Original Source
Read the complete research paper
About the Author
Guilherme A.
Former dentist (MD) from Brazil, 41 years old, husband, and AI enthusiast. In 2020, he transitioned from a decade-long career in dentistry to pursue his passion for technology, entrepreneurship, and helping others grow.
Connect on LinkedIn