Non-Unitary Quantum Butterfly Effect: Theory and Experiment

In the quantum world, information behaves in ways that seem to defy our everyday intuition. When a system is isolated and evolves according to the strict rules of quantum mechanics, information does not vanish; it merely spreads out, becoming scrambled across the entire system like a drop of ink dispersing in a glass of water. This process, known as scrambling, is central to understanding how complex quantum systems, from black holes to superconductors, process information. Scientists have long used a specific mathematical tool, called an out-of-time-ordered correlator, to track this scrambling. Think of it as a way to measure how much a small change in one part of a system affects a distant part later on. In a perfectly isolated system, this effect travels outward at a fixed speed, creating a clear boundary of influence, much like ripples spreading from a stone dropped in a pond.

However, real-world quantum computers are not perfectly isolated. They are constantly interacting with their environment, and researchers are increasingly using measurements—looking at the state of the system—as a tool to control and correct errors. This introduces a new element: the act of measuring a quantum system disturbs it, potentially erasing the very information scientists are trying to track. The question arises: if we keep peeking at a quantum system while it evolves, does the information still scramble? Does the ripple of influence still travel, or does the act of looking stop it in its tracks? This is the core mystery that a team of physicists at the University of Maryland and the National Institute of Standards and Technology set out to solve.

The researchers investigated how the continuous act of measuring a quantum system changes the way information spreads. They focused on a specific type of quantum circuit, a sequence of operations that evolve a group of quantum bits, or qubits. In their theoretical model, they built a circuit where random quantum operations were interspersed with local measurements. These measurements were not just passive observations; they were active interventions that collapsed the quantum state at specific points. The team wanted to see how these measurements affected the "butterfly effect" of the system—the sensitivity of the future state to a tiny change in the past. In a standard, unmeasured system, a small change at one location eventually ripples out to affect the entire system. The researchers asked: what happens to this ripple if we keep measuring the system as it evolves?

To answer this, the team developed a new theoretical framework to calculate the behavior of these measurements. They found that the act of measuring does indeed change the story. In a system without measurements, the information spreads out in a sharp, predictable front, leaving a persistent memory of the initial disturbance behind it. But when measurements are introduced, this memory begins to fade. The researchers discovered that the measurements act like a form of dynamical amnesia. Even though the disturbance continues to travel outward, the signature of that disturbance, which the scientists track using their correlator tool, starts to decay. The system progressively forgets the initial push. This decay is not random; it follows a specific pattern. In systems where measurements are sparse, the signal does not just fade away slowly; it follows a Gaussian shape, a bell curve, meaning it rises to a peak and then falls off in a very predictable way as time goes on.

The team did not stop at theory. They built a computer simulation to verify their calculations, using a method that allowed them to track the behavior of thousands of different possible circuit outcomes. These simulations confirmed that their mathematical predictions were correct. The signal decayed exponentially over long periods, but for the early and middle stages of the process, the Gaussian shape held true. To take this a step further, they moved from simulation to reality. They ran their experiments on a real quantum computer built by Quantinuum, which uses trapped ions as its qubits. This machine is known for its high precision, but like all quantum hardware, it is not perfect and contains its own background noise.

Despite the presence of this hardware noise, the experiment successfully resolved the difference between the ideal, unmeasured case and the measured case. The researchers observed that in the unmeasured scenario, the signal spread out in a clear, ballistic front, reaching a stable level as it filled the system. When they introduced measurements, the behavior changed exactly as their theory predicted. The signal still spread, but its strength diminished over time within the region where it had spread. The measurements effectively erased the trace of the disturbance, separating the physical spread of the influence from the persistence of the information about that influence. The experiment showed that even with the imperfections of real hardware, the unique signature of this "non-unitary" effect was detectable.

The findings have significant implications for the future of quantum computing. As engineers build larger and more complex quantum machines, they rely heavily on mid-circuit measurements to correct errors and run advanced algorithms. Understanding how these measurements affect the flow of information is crucial. The study provides a way to quantify exactly how much a measurement disturbs the rest of the system. It offers a new benchmark for hardware: by measuring how quickly the signal decays, engineers can determine the cumulative effect of noise and measurement errors on the system's dynamics. This is particularly important for architectures that use measurement as a computational resource, such as measurement-based neural networks, where the "memory" of the system is defined by how far and how long a measurement's influence lasts.

The researchers also noted that their method is remarkably simple and efficient. Unlike previous approaches to measuring these effects, which often required extra, auxiliary qubits or complex interferometric setups, their protocol uses only the qubits already present in the system. This makes it a practical tool that can be deployed on current and future hardware without needing additional resources. The team demonstrated that this approach works even on systems with eight qubits, a size that is small enough to be simulated by classical computers but large enough to show the essential physics. They believe this method can be scaled up to much larger systems, potentially those too large for any classical computer to simulate, opening a window into the behavior of monitored many-body systems that were previously inaccessible.

Ultimately, this work bridges the gap between abstract quantum theory and the practical realities of building quantum computers. It shows that while measurement is a powerful tool for control, it comes with a cost: the gradual erasure of information. The study characterizes this cost with precision, providing a clear picture of how the quantum world remembers and forgets. By resolving the behavior of these systems in both simulation and on real hardware, the team has provided a foundational understanding of how information flows in the presence of observation, a key step toward mastering the next generation of quantum technologies.