Physicists at Imperial College London and Bard College show that a structured, correlated form of quantum noise can drive a one dimensional quantum system into a robust new phase of matter — a topological phase — whose signature is that particle
A theoretical model from Imperial College London and Bard College (arXiv:2607.07801) suggests that quantum decoherence, long treated as the adversary of topological order, can in fact generate it. The result reframes noise, a problem class quantum-hardware engineers usually try to suppress, as a possible resource whose structure, not just its magnitude, matters.
The paper, 'Topology from Decoherence,' studies a one-dimensional lattice of spinless fermions evolved under a Lindblad master equation, the standard formalism for an open quantum system in which the system of interest continuously interacts with an environment. The authors, based at Imperial's Blackett Laboratory and the Bard College Physics Program, choose the environment deliberately: the jump operators couple the density and current operators on neighboring sites, so the dephasing is correlated across sites and interaction-driven rather than independent.
After averaging over noise realizations, the resulting dynamics settle into a topological phase characterized by a non-trivial winding number, an integer invariant that counts how many times a system property winds around a circle as parameters vary, and accompanied by the non-Hermitian skin effect, a phenomenon in which bulk states pile up at a boundary instead of spreading uniformly in systems described by non-Hermitian operators.
The model also produces a concrete dynamical signature (see PDF): diffusion becomes asymmetric. Noise-averaged transport drifts in one direction at a rate fixed by the winding number, and crossing a topological phase transition reverses the drift. The authors propose that direction reversal as the experimental fingerprint of the new phase.
The mechanism differs from prior non-Hermitian skin-effect work in its interaction origin. Most earlier results either ignore particle-particle interactions or absorb them into effective single-particle physics; the Imperial-Bard derivation requires interacting fermions and a correlated environment. Under postselection, keeping a single stochastic quantum trajectory and discarding the rest, the topological phase collapses, confirming that the effect is genuinely open-system rather than a hidden Hamiltonian in disguise.
The Lindblad formalism is one of the few open-system toolkits where closed-form progress is realistic, and the authors' choice of an exactly solvable hopping term plus a structured jump operator invites paper-and-pencil calculation rather than brute-force numerics.
The paper does not claim decoherence is beneficial, and the result is theoretical. It does suggest a reorientation: when designing qubits, the assumption that less noise is always better may be incomplete. If correlated, structured noise can drive a desired topological phase, then suppressing correlated noise structure, even at the cost of some total noise budget, becomes a coherent design objective. Current topological-qubit efforts focus on noise magnitude and isolation; the paper points at the orthogonal axis of noise correlations.
The result is testable in principle on platforms capable of engineered dissipation, including trapped-ion and superconducting analog simulators. The authors do not claim an experimental demonstration is imminent. The prediction the authors leave for follow-up is a measurement program organized around the directional-diffusion reversal across a topological phase transition.
The submission is a preprint and has not been peer reviewed.