Title:Propagation of initial uncertainties to Arthurs-Kelly inequality

Abstract:When the interaction Hamiltonian is $\widehat{H}_I = \kappa \left(\hat{x}_3 \hat{p}_1 + \hat{p}_3 \hat{p_2} \right)$ for large $\kappa$ and the initial quantum state is $\Psi_{in} (x_1, x_2, x_3) = \prod_{j=1}^3 \phi_j (x_j)$, the joint measurement of complementary variables $x_3$ and $p_3$ at $t = 1 / \kappa$ induces the uncertainty $\Delta_{\hat{x}_1} (t = 1 / \kappa) \Delta_{ \hat{x}_2} (t = 1 / \kappa) \geq \frac{1}{2} (\Delta_{\hat{x}_1} \Delta_{\hat{p}_1} + \Delta_{\hat{x}_2} \Delta_{\hat{p}_2} ) + \Delta_{\hat{x}_3} \Delta_{\hat{p}_3}$ , where the standard deviations in the right hand are the deviations at $t=0$. This is a generalized version of the well-known Arthurs-Keller inequality $\Delta_{\hat{x}_1} (t = 1 / \kappa) \Delta_{ \hat{x}_2} (t = 1 / \kappa) \geq 1$ arising when all $\phi_j (x_j)$ are the minimal uncertainty Gaussian states. If the initial probe state is entangled, it is shown that the generalized Arthurs-Kelly inequality can be violated. We show the violation explicitly by introducing a special example.