Reversing time to probe quantum dynamics
In the 22 October issue of Nature, researchers from Google Quantum AI and collaborators report the observation of constructive interference at the edge of quantum ergodicity, a regime probed by manipulating the dynamics of a quantum system in ways that resemble running its evolution backwards. The result offers a new window on one of the deepest questions in quantum physics: how isolated systems forget, or refuse to forget, their past.
One of the deepest questions in quantum physics is about memory: what an isolated quantum system remembers of its own history as it evolves. The standard account says it remembers very little; given time, a system of many particles tends toward equilibrium, and the details of how it began wash out, a process quantum physicists call thermalisation. But the standard account has edges, and at those edges the memory persists, and the behaviour of a system that refuses to forget is one of the most informative things quantum mechanics can show us. A paper in the 22 October issue of Nature reports how Google Quantum AI and its collaborators probed one of those edges, using manipulations that effectively run a quantum system’s evolution in reverse.
The idea behind the experiments is a delicate one. On a quantum processor, the evolution of a system can be engineered, and, within limits, it can also be undone: a sequence of operations can be followed by its reverse, so that a system that had been heading toward thermal equilibrium is returned, in a controlled way, toward the order it began with. The reversal is not literal time travel; it is a laboratory tool, a way of studying how the system behaves when its normal arrow of dynamics is perturbed. By combining forward and reversed evolution, the team could probe the regime at the edge of ergodicity, the boundary between systems that thermalise completely and systems that hold on to structure.
The result, the paper reports, is the observation of constructive interference at that edge, a quantum signature that appears when the system’s evolution is manipulated in this way. The finding gives experiment a purchase on a regime that theory had described but could not easily reach: a quantitative look at how order persists, and how it is finally lost, in a controlled quantum system.
The significance of the work extends in two directions. For fundamental physics, it is a new experimental window on thermalisation, one of the oldest problems in the statistical mechanics of quantum systems, the question of how the quantum world produces the irreversible behaviour of the everyday world. For the engineering of quantum computers, the same techniques matter practically: understanding how a processor’s qubits lose coherence, and how their dynamics can be manipulated and reversed, is directly relevant to building machines that must preserve fragile quantum states long enough to compute with them.
The paper is part of a pattern that has become the field’s signature: the quantum processor as a laboratory instrument, used not only to compute but to explore the physics that the processors themselves embody. The experiments sit at an unusual intersection, where the machine built to run algorithms becomes the apparatus for studying the theory the machine is based on, and the same controls that make computation possible, the gates, the measurements, the error correction, become the tools of the inquiry. Reversing a system’s evolution in a controlled apparatus, watching the interference at the edge of chaos, is a demonstration of how far the control of quantum systems has come, and of how much the machines being built are teaching the physics that built them. The system remembers, when pushed; and the remembering, measured, is the result.