Melting Temporal Limits with Quantum Dynamics
In Salvador Dalí’s most iconic painting, The Persistence of Memory, a surrealist landscape draped with melting clocks seems to challenge our sense of time as something rigid and eternal. Quantum mechanics evokes a similar unease: It reveals a world in which multiple possibilities coexist, and only observation solidifies outcomes. Meshing this quantum world with the more familiar classical world remains one of the most enduring challenges in physics. A key tool in this quest is the Leggett-Garg inequality (LGI), which was designed to test whether a system’s evolution over time can be described by classical realism [1, 2]. Quantum systems are expected to violate the LGI but only up to a theoretical limit: a peak in the strength of correlations known as the “temporal Tsirelson bound” [3]. Now Arijit Chatterjee at the Indian Institute of Science Education and Research and his colleagues show that this limit can be surpassed by a quantum system evolving under a superposition of operations, such as a spin that undergoes a rotation and a flip at the same time. Their work reveals that not only quantum states but also quantum dynamics can form a coherent superposition that enhances nonclassical temporal correlations [4].
For decades, experiments have shown that spatially separated particles can share correlations that no classical model can explain [5, 6]. Such experiments are inspired by John Stewart Bell’s now-famous inequality, with which he sought to distinguish quantum mechanics from an alternative understanding of the physical world, known as local realism [7]. The LGI extends these experiments to the temporal domain, probing correlations between measurements performed on a single system at different times. If a system always possesses definite properties independent of observation—a principle known as macroscopic realism—then the LGI must hold. Quantum mechanics, however, allows violations of this inequality, introducing correlations that reveal the fundamentally contextual nature of temporal evolution. Several experiments have confirmed that the LGI can be violated [8]. Oscillations in a tiny superconducting circuit, for example, were shown to affect the dynamics of the circuit at different time points—introducing correlations that couldn’t be explained classically [9].
These violations were long thought to obey the temporal Tsirelson bound, a quantum-mechanical limit on the strength of the correlations that appear. This bound has served as a benchmark for the “quantumness” of temporal behavior, just as an analogous spatial Tsirelson bound constrains Bell-type experiments. Yet recent theoretical and experimental progress has started to erode physicists’ confidence in the rigidity of this temporal bound. Studies of gain–loss dynamics, non-Hermitian systems, and higher-dimensional controls have revealed that quantum evolution can harbor untapped potential, stemming from the fact that multiple paths can coexist and interfere in nonclassical ways [10]. Chatterjee and collaborators have now pushed this frontier further, offering perhaps the most direct evidence that quantum dynamics can yield stronger-than-expected temporal correlations.
The central idea is both simple and profound. Ordinarily, an LGI protocol involves measuring a quantum system at different time points as it undergoes a unitary transformation, which is a time-reversible operation, such as the evolution induced by the application of an electric field. Chatterjee and colleagues instead let a qubit evolve under a coherent superposition of two distinct unitary transformations. Conceptually, the system follows two dynamical paths at once—extending the superposition principle from describing what a system is to describing how it evolves.
This approach turns the notion of time evolution into a property that can be controlled. By coherently combining different dynamical paths, the system explores an enlarged space of temporal correlations that ordinary evolution cannot reach. The researchers found that these superposed transformations lead to LGI violations that surpass the conventional temporal Tsirelson bound, revealing a previously unexplored regime in which quantum dynamics interfere constructively to amplify nonclassical behavior. Moreover, the strength of the effect grows continuously with the degree of superposition and with the contrast between the underlying dynamical directions, offering a way to sculpt temporal correlations.
To demonstrate this phenomenon experimentally, Chatterjee and collaborators turned to nuclear magnetic resonance (NMR) techniques, in which quantum control is both precise and flexible. The researchers chose as their target the 13C-dibromofluoromethane (DBFM) molecule, whose three nuclear spins can store quantum information. One of the spins acted as a qubit that evolved under two unitary transformations. A second spin (called the ancilla) acted as a quantum control, coherently determining which transformation the qubit underwent (Fig. 1). Maintaining coherence between both possibilities effectively created a superposition of dynamical paths—a kind of “quantum timeline interference.”
Through a series of NMR measurements, the researchers tracked how this composite evolution influenced temporal correlations, and they observed clear violations of the LGI that exceeded the expected quantum limit by more than 3 times the experimental uncertainty. Remarkably, these enhanced correlations also proved more resilient to environmental noise. This resilience arises because the composite evolution redistributes how the system explores its state space, and this modified exploration reduces the system’s exposure to dephasing processes that normally diminish quantum signatures. The result is not just a conceptual advance, but it also might have a practical advantage: Superposed dynamics could make temporal quantum correlations longer lived and more controllable than previously thought.
The implications of these findings are twofold. First, they deepen our foundational understanding of quantum temporal correlations, showing that the Tsirelson bound is not as rigid as once believed, at least when one incorporates a superposition of dynamical paths. This suggests that the limits of quantum correlations are context dependent—determined not only by the structure of quantum states but also in the way that quantum operations are performed.
Second, the work introduces an exciting new resource for quantum technologies. On the basis of their ability to stretch the temporal Tsirelson bound, superpositions of unitary transformations could potentially be engineered to mitigate decoherence, enhance temporal signal-to-noise ratios, or generate novel forms of quantum control. They could be used in quantum simulations to synthesize complex dynamics from simpler components or in quantum error-mitigation protocols that exploit dynamical interference to preserve the coherence of qubits.
By revealing that time evolution can itself exist in a superposition, this study invites us to rethink one of quantum theory’s most familiar concepts. It suggests that quantum dynamics can bend and merge, much like Dalí’s melting clocks. Whether this new dynamical degree of freedom can be scaled to larger systems or harnessed for computation and metrology remains to be seen, but the message is clear: The quantum boundary between “state” and “evolution” is more porous than we thought.
References
- A. J. Leggett and A. Garg, “Quantum mechanics versus macroscopic realism: Is the flux there when nobody looks?” Phys. Rev. Lett. 54, 857 (1985).
- A. J. Leggett, “Realism and the physical world,” Rep. Prog. Phys. 71, 022001 (2008).
- A. Palacios-Laloy et al., “Experimental violation of a Bell’s inequality in time with weak measurement,” Nat. Phys. 6, 442 (2010).
- A. Chatterjee et al., “Extreme violations of Leggett-Garg inequalities for a system evolving under superposition of unitaries,” Phys. Rev. Lett. 135, 220202 (2025).
- J. S. Bell, “On the Einstein Podolsky Rosen paradox,” Physics 1, 195 (1964).
- S. Kochen and E. Specker, “The problem of hidden variables in quantum mechanics,” J. Math. Mech. 17, 59 (1967).
- D. Garisto, “Quantum milestones, 1964: John Stewart Bell quietly rings in new era of quantum theory,” Physics 18, 53 (2025).
- C. Emary et al., “Leggett–Garg inequalities,” Rep. Prog. Phys. 77, 016001 (2013).
- A. Palacios-Laloy et al., “Experimental violation of a Bell’s inequality in time with weak measurement,” Nat. Phys. 6, 442 (2010).
- P. Lu et al., “Experimental demonstration of enhanced violations of Leggett-Garg inequalities in a PT-symmetric trapped-ion qubit,” Phys. Rev. A 109, 042205 (2024).