Selecting quieter measurement records preserves a simulated time crystal

A rhythm that outlasts each push

A periodically driven system receives the same sequence of pushes repeatedly. A discrete time crystal responds with a rhythm spanning several drive cycles: in the period-doubled case, its observable pattern returns every second cycle. Crucially, the rhythm must resist small imperfections in the drive, rather than appearing only under perfectly tuned conditions.

Repeated driving can also make an interacting quantum system absorb energy and lose its initial order. In a November 2026 arXiv preprint, Pallabi Chatterjee, Tanay Nag and Ranjan Modak ask whether continuous weak measurements can help preserve time-crystalline behavior without introducing disorder. Their evidence comes from theoretical modeling and numerical simulations, not a new experiment.

Their answer has an important qualification: ordinary monitoring fails, while a protocol that restricts fluctuations in the measurement record preserves period doubling over the simulated times.

Following individual measurement histories

The researchers model a chain of quantum spins with interactions extending beyond neighboring sites. Each driving cycle first rotates the spins, approximately flipping them, then lets them interact. Weak measurements monitor each spin’s orientation during this second stage.

Measurement does more than reveal information: it also changes the quantum state. Each possible sequence of measurement outcomes therefore produces a different history of the state, called a quantum trajectory.

The team compares unrestricted measurement records with an idealized limit in which the fluctuating part of each outcome is set to zero. Between these extremes, they introduce a finite acceptance window: only measurement fluctuations inside that window are allowed. This is postselection—conditioning the results on accepted measurement records—not evidence that ordinary measurement automatically removes its own noise.

They track both magnetization and entanglement, the quantum correlations between the two halves of the chain. Initial spins are either all aligned or arranged in an alternating pattern.

Low entanglement is not enough

With unrestricted measurement outcomes, sufficiently strong monitoring suppresses entanglement. Instead of growing roughly in proportion to chain length, the entanglement remains approximately independent of length across the tested sizes, a pattern called area-law scaling.

But the period-doubled response still disappears. The ensemble-averaged state becomes featureless even though individual trajectories can have little entanglement. Thus, in this model, reducing entanglement does not by itself produce a time crystal.

The fluctuation-free limit behaves differently. Strong enough monitoring produces persistent alternating magnetization and a sharp response at half the driving frequency. The response also tolerates imperfect spin flips. The necessary measurement strength depends on the initial spin pattern, with aligned spins developing this behavior under weaker monitoring than alternating spins.

A window between survival and decay

The central bounded-noise calculations use 14 spins and 128 trajectories starting from the alternating pattern. At one tested setting, a narrow dimensionless window of 0.3 preserves a subharmonic response near 0.72, where unity would represent perfect alternating magnetization. The window value is a normalized noise cutoff, not a percentage or a physical detector specification.

At a window of 0.4, the trajectories divide into groups with strong and weak period doubling. As simulations extend toward 5,000 drive cycles, the strongly responding group shrinks while its characteristic response remains near the same value. The authors interpret this as trajectories losing their time-crystalline behavior one by one, rather than all gradually weakening together.

Wider windows destroy the order. Narrow windows preserve it throughout the simulated interval. Changing the interaction range retains this qualitative picture but changes the tolerated window, showing that the numerical thresholds are model-dependent.

What remains unproven

Finite chains and finite observation times cannot establish indefinite stability in an arbitrarily large system. The intermediate-window regime explicitly shows long-lived but finite order; even the narrow-window result should be understood within the simulated sizes and durations.

Postselection also has a cost. Although a finite window accepts more outcomes than selecting one exact record, the probability of accepting an entire experimental run decreases as more spins and longer histories are monitored. The authors discuss trapped ions and superconducting qubits as possible platforms, but demonstrate neither an implementation nor its practical efficiency.

The result identifies a conditional route to robust quantum rhythms: how measurement outcomes are selected matters, not just how strongly the system is watched.