The ability to harness correlations between quantum systems represents a promising pathway towards new technologies, and researchers are now investigating how effectively these correlations can perform work. Karl Svozil from the Institute for Theoretical Physics, TU Wien, leads a team that demonstrates that while classical, quantum, and even stronger-than-quantum correlations all possess the potential to yield the same peak amount of extractable work, their usefulness as a resource differs dramatically. The team shows that classical correlations are easily disrupted by measurement inaccuracies, losing effectiveness quickly, whereas quantum correlations exhibit significantly greater resilience, decaying much more slowly. This research establishes that the true value of non-locality lies not in the maximum energy attainable from a correlation, but in its operational robustness as a fuel for computation and other applications.
Classical correlations prove fragile, decaying linearly with misalignment, whereas quantum correlations exhibit resilience, decaying only quadratically. Consequently, the degree of nonlocality maps not to the maximum energetic value of a correlation, but to its operational robustness as a thermodynamic fuel. In classical thermodynamics, an isolated system is defined by boundaries impermeable to matter and energy. For macroscopic, uncorrelated systems, this definition proves adequate, and the second law dictates an irreversible evolution towards maximal entropy.
Correlation Robustness to Measurement Imperfections
The study investigated how effectively different types of correlations, classical, quantum, and a hypothetical stronger-than-quantum form, can be harnessed as a thermodynamic resource, specifically examining their resilience to measurement inaccuracies. Researchers focused on the extractable work potential from these correlations, assessing how much energy could be obtained when measurements are slightly misaligned. The core of the work involved comparing these correlation types by simulating measurement imperfections and quantifying the resulting loss of extractable work, revealing a clear hierarchy of robustness. To achieve this, scientists developed a method for evaluating the “fragility” of each correlation type, modelling how its work potential degrades with increasing misalignment between measurement settings.
For the classical model, the team demonstrated a linear decay in work potential with misalignment, meaning even small errors significantly reduce energy extraction. In contrast, the quantum model exhibited a quadratic decay, demonstrating substantially greater resilience to measurement errors. Further extending this analysis, the researchers posited a hypothetical correlation exceeding even quantum limits, demonstrating perfect resilience, maintaining maximum work potential regardless of measurement angle. This was achieved through theoretical modelling, defining a correlation function that remains constant except for a single point of discontinuity.
To quantify this hierarchy, the team introduced an “energetic CHSH parameter,” adapting a standard measure of non-locality to a thermodynamic context. This involved substituting correlation functions with extractable work potential, allowing for a direct comparison of the three models under identical measurement conditions. The calculations were performed using specific angles to maximize the quantum value and reveal the energetic differences. The resulting analysis confirmed that while all three models can theoretically yield the same peak work potential, their practical utility is dictated by their robustness to measurement errors, with the stronger-than-quantum model offering the highest resilience.
Erasure of Correlations Yields Extractable Work
This work demonstrates that initial system-environment correlations represent a valuable resource, enabling work extraction through their erasure. Scientists compared the work potential of classical, quantum, and hypothetical stronger-than-quantum correlations as a function of measurement misalignment, finding that all models can theoretically yield a peak extractable work of kBT ln 2, corresponding to a mutual information of ln 2. However, the robustness of these correlations as a resource differs significantly. Further analysis revealed that the thermodynamic value of these correlations is determined not by their degree of nonlocality, but by their mutual information.
Scientists defined mutual information as Iθ(A : E) = ln 2 −H(A|E), where H(A|E) represents the conditional entropy. Measurements confirmed that perfect correlation or anti-correlation yields the maximal mutual information of ln 2. However, the team discovered significant operational differences in accessing this resource. Classical correlations proved brittle, decaying linearly with misalignment, while quantum correlations exhibited a quadratic decay, demonstrating greater robustness. Specifically, the quantum resource showed a deviation of approximately Iq(δθ) ≈ln 2 −O(δθ2) for a small misalignment δθ, highlighting its resilience to experimental noise.
Correlation Robustness Dictates Work Extraction Potential
This research demonstrates that correlations between a system and its environment represent a genuine thermodynamic resource, capable of enabling work extraction even when energy and matter exchanges are blocked. The team rigorously compared classical, quantum, and hypothetical stronger-than-quantum correlations, revealing that all models can yield a maximum extractable work equivalent to kBT ln 2, corresponding to a mutual information of ln 2. However, the robustness of these correlations as a resource differs significantly; classical correlations degrade linearly with misalignment, while quantum correlations exhibit a more resilient, quadratic decay. These findings clarify that the value of a correlation lies not in the magnitude of extractable work, but in its operational robustness, suggesting quantum correlations offer a distinct advantage in maintaining efficiency under imperfect conditions. By establishing mutual information as a unifying measure of work, the researchers reconcile thermodynamics and information theory at microscopic scales, demonstrating that information, like energy, must be considered a physical fuel.
👉 More information
🗞 Correlation-Powered Work: Equivalence in Peak Yield, Differences in Robustness
🧠 ArXiv: https://arxiv.org/abs/2511.03679
