arXiv:2610.20151
Directional effects of the Unruh effect on steering harvesting
Shu-Min Wu, Rui-Yang Xu, Zhihong Liu
Abstract
We investigate directional quantum steering harvesting by two uniformly accelerated non-identical Unruh-DeWitt detectors coupled to a massless scalar field in the Minkowski vacuum. We show that the Unruh effect does not simply degrade harvested steering, but can reshape its directionality in a strongly parameter-dependent manner. At low detector energy scales, acceleration predominantly suppresses steering, whereas at higher energy scales, moderate acceleration can enhance steering and even activate an initially absent steering direction. Most strikingly, within a finite acceleration window, the Unruh effect enhances steering in one direction while simultaneously suppressing it in the opposite direction, revealing an intrinsically directional influence of relativistic acceleration. Detector non-identicality provides an additional control mechanism: a moderate energy-gap mismatch can enhance or activate one steering direction while suppressing the reverse one, whereas sufficiently large energy-gap mismatch suppresses the harvested steering. Moreover, the maximal steering asymmetry approximately marks the transition from two-way to one-way steering. Our results establish relativistic acceleration and detector non-identicality as means of controlling the strength and directionality of vacuum-harvested quantum correlations.
AI Summary
The paper asks whether acceleration can change the directionality of quantum steering extracted from vacuum correlations, rather than simply suppressing it. It considers two initially ground-state, non-identical Unruh–DeWitt detectors coupled to a massless scalar field in the $(3+1)$-dimensional Minkowski vacuum. Both follow uniformly accelerated trajectories with equal proper acceleration and a fixed longitudinal displacement in Minkowski coordinates. Their interactions have Gaussian switching of width $\sigma$, and their energy gaps satisfy $\Omega_B\geq\Omega_A$. The detector state is calculated perturbatively through order $\lambda^2$. Directional steering indicators are constructed using sufficient steering criteria based on entanglement of auxiliary two-qubit states, then evaluated numerically versus separation, acceleration, and gap mismatch. At $L/\sigma=0.5$, acceleration monotonically suppresses both indicators for $\Omega_A\sigma=0.5$, whereas moderate acceleration can enhance them for $\Omega_A\sigma=1.2$. Their maxima occur at different accelerations, producing an interval where one indicator increases while the other decreases. Acceleration can also activate the initially undetected reverse direction, yielding an indicated one-way–two-way–one-way–no-way sequence. Moderate gap mismatch can enhance or activate $A\rightarrow B$ steering while suppressing $B\rightarrow A$; sufficiently large mismatch suppresses both. Increasing separation suppresses both indicators, and maximal indicator asymmetry approximately coincides with loss of the weaker direction. The stated scope is this detector model, perturbative order, and numerical parameter study. Because the steering criteria are sufficient rather than necessary, vanishing indicators do not establish absence of steering; the reported directional regimes should be read with that qualification.
Central claim may not hold: The paper interprets failure of sufficient steering witnesses as absence of steering, without proving unsteerability, so its one-way/no-way classifications and activation claims are not established. The reported enhancement and suppression concern witness margins that are not established to quantify the amount of steering. The supplied source contains no plotted numerical results and omits parameters needed to reproduce the claimed finite acceleration window. The nonlinear steering analysis lacks higher-order error control despite using a second-order density-matrix approximation whose omitted double-excitation population is necessary for positivity when X is nonzero.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 21 pages, 3 figures