Table of Contents
Fetching ...

Correlation-Powered Work: Equivalence in Peak Yield, Differences in Robustness

Karl Svozil

TL;DR

The paper investigates how different system–environment correlation laws—classical, quantum, and hypothetical stronger-than-quantum—act as thermodynamic resources. Using a generalized second law that incorporates mutual information, it shows that the maximum work extractable per correlated binary system is universally bounded by $k_B T \ln 2$, determined by the initial mutual information; however, the ease of accessing this resource under misalignment differs. Classical correlations degrade linearly with misalignment, quantum correlations degrade quadratically, and stronger-than-quantum correlations would be nearly alignment-insensitive, creating an operational hierarchy captured by an energetic CHSH parameter $\mathcal{S}_W$. A concrete Szilard-engine demonstration and the mutual-information currency bridge thermodynamics and information theory, suggesting that information can function as a physical fuel for microscopic energy resources.

Abstract

Initial system-environment correlations are a thermodynamic resource, enabling work extraction via their erasure. We compare the work potential of classical, quantum, and hypothetical stronger-than-quantum correlations as a function of measurement misalignment. While all models can yield a peak extractable work of k_B T ln 2, corresponding to a mutual information of ln 2, their value as a resource differs critically in its robustness. The classical resource is fragile, decaying linearly with misalignment, whereas the quantum resource is robust, decaying only quadratically. Thus, the degree of nonlocality maps not to the maximum energetic value of a correlation, but to its operational robustness as a thermodynamic fuel.

Correlation-Powered Work: Equivalence in Peak Yield, Differences in Robustness

TL;DR

The paper investigates how different system–environment correlation laws—classical, quantum, and hypothetical stronger-than-quantum—act as thermodynamic resources. Using a generalized second law that incorporates mutual information, it shows that the maximum work extractable per correlated binary system is universally bounded by , determined by the initial mutual information; however, the ease of accessing this resource under misalignment differs. Classical correlations degrade linearly with misalignment, quantum correlations degrade quadratically, and stronger-than-quantum correlations would be nearly alignment-insensitive, creating an operational hierarchy captured by an energetic CHSH parameter . A concrete Szilard-engine demonstration and the mutual-information currency bridge thermodynamics and information theory, suggesting that information can function as a physical fuel for microscopic energy resources.

Abstract

Initial system-environment correlations are a thermodynamic resource, enabling work extraction via their erasure. We compare the work potential of classical, quantum, and hypothetical stronger-than-quantum correlations as a function of measurement misalignment. While all models can yield a peak extractable work of k_B T ln 2, corresponding to a mutual information of ln 2, their value as a resource differs critically in its robustness. The classical resource is fragile, decaying linearly with misalignment, whereas the quantum resource is robust, decaying only quadratically. Thus, the degree of nonlocality maps not to the maximum energetic value of a correlation, but to its operational robustness as a thermodynamic fuel.

Paper Structure

This paper contains 15 sections, 15 equations, 2 figures.

Figures (2)

  • Figure 1: A comparison of the three correlation functions $E(\theta)$ versus the relative measurement angle $\theta$. The quantum (solid red) and classical (dashed green) correlations only achieve perfect anti-correlation ($E=-1$) at $\theta=0$. The hypothetical stronger-than-quantum correlation (dotted blue) is perfect for any angle $\theta \neq \pi/2$.
  • Figure 2: Mutual information as a function of relative measurement angle $\theta$ for the three correlation laws discussed in the text: Quantum ($I_q$, solid red), Classical ($I_c$, dashed green), and stronger-than-quantum ($I_s$, dotted blue). While classical and quantum correlations only provide the maximal information resource of $\ln 2$ nats at perfect alignment ($\theta=0$) or anti-alignment ($\theta=\pi$), the hypothetical stronger-than-quantum correlation yields this maximum for all angles except the single point $\theta=\pi/2$.