Indian Physicists Surpass Tsirelson’s Bound—But Only in Time

November 20, 2026

A team led by Arijit Chatterjee has advanced a qubit into a coherent superposition of two distinct unitary evolutions, and achieved violations of the Leggett-Garg inequality that exceed the temporal Tsirelson bound. The work is published in Physical Review Letters (volume 135, article 220202, 2025). It does not touch the spatial Tsirelson bound associated with Bell tests, the most widely known bound among the general public. And a rebuttal is already circulating.

Two bounds share the same family name. Bell and Leggett-Garg formulated inequalities that bound correlations in space and in time, and their quantum limits are called the Tsirelson bound and the temporal Tsirelson bound. The Indian experiment concerns only the latter.

Two bounds, one name

The original bound dates back to 1980. It caps the quantum violation of the CHSH inequality at 2√2. It has been approached very closely in laboratory settings: a photon-pair experiment measured S = 2.82759 ± 0.00051, i.e., a gap of 0.00084 from the bound.

The Leggett-Garg inequality operates on a different plane. It tests whether a system behaves classically by comparing correlations of a single system measured at successive times rather than correlations between distant particles. Under unitary dynamics and projective measurements, its violation is limited by the temporal Tsirelson bound.

It is this bound that is the target of the present work.

Calling that the “ultimate limit of quantum physics” has been breached is therefore an extrapolation. The spatial Bell bound remains intact, and the result does not speak to it. Early work carried out at the Indian Statistical Institute in Calcutta showed that a quantum violation of the Tsirelson bound would imply signaling, i.e., communication, which explains the attention given to any attempt to cross it.

What the experiment actually accomplished

The team comprises Arijit Chatterjee, H. S. Karthik, T. S. Mahesh and A. R. Usha Devi. Chatterjee is affiliated with the Indian Institute of Science Education and Research in Pune. The manuscript circulated as a preprint on November 4, 2024, a year before journal publication.

The underlying idea is simple: the larger the share of superposition between the two unitary evolutions, the larger the Leggett-Garg violation beyond the bound, and the effect grows with the degree of superposition. The authors do not merely cross the threshold by a hair; the enhancement scales with the superposition parameter, as the figures in the paper illustrate from several angles.

To realize this, they used nuclear magnetic resonance (NMR), a platform where quantum control is both precise and flexible. An NMR register allowed the superposition of unitaries with the help of an auxiliary qubit. One of the spins acted as the qubit and evolved under two unitary transformations.

Measurements align with the theoretical prediction: the observed violation greatly surpasses the temporal bound.

Noise robustness: the useful point, not an overstatement

The potential practical relevance lies in decoherence, the number-one enemy of quantum computers. In the presence of noise, the two-unitary superposition extends the duration over which the Leggett-Garg inequality remains violated, indicating a higher resilience of the nonclassical nature to environmental disturbances.

That does not, however, justify proclaiming quantum computers that are “much more robust than anything today.” The article concerns the persistence of a violation on an NMR qubit, not a processor or a calculation error rate. No algorithmic performance metrics appear in the abstract.

The value lies elsewhere. A violation that endures longer despite noise is easier to observe and can serve as a witness of nonclassical behavior, making it a useful diagnostic tool for experiments involving qubit control.

A rebuttal published as early as late 2025

The journal’s own commentary presents the result with enthusiasm: the dynamically superposed evolutions produced stronger temporal correlations, surpassing the theoretical limits. Other groups have explored related avenues, including a PT-symmetric trapped-ion qubit mentioned in the same commentary.

But the community is not unanimous. A preprint posted in response argues that the proposed evolution can be understood within a conventional framework, without superposed evolutions, and that the apparent violation arises from measured quantities that conflict with the assumptions behind Leggett-Garg. It does, however, acknowledge that the experimental results support the claim—the debate centers on interpretation rather than raw data.

This preprint had not been peer-reviewed at the time of writing. It does not settle the matter. But it raises the central question: has the bound truly been violated, or has the definition of what is being measured changed?

For a cautious reader: an NMR experiment published in a peer-reviewed journal reports a crossing of the temporal Tsirelson bound, with the effect increasing with the degree of superposition. This crossing is contested on its interpretation; the spatial Bell bound remains untouched, and the practical implications for quantum computing remain to be demonstrated at this stage.

One item to watch: the authors’ response to the critical preprint. If they insist that their protocol adheres to Leggett-Garg’s hypotheses, the debate will hinge on the precise definition of measurements—a discussion where the experiment alone cannot provide a definitive answer.

Article prepared with the help of AI · Our method

Sindre Halvorsen

I write about space exploration, frontier science and the technologies that are quietly shaping the future. From Norway, I follow the missions, discoveries and ideas that connect life on Earth with what lies beyond it. My goal is to make complex subjects clear, useful and worth paying attention to.