arXiv:2610.17334
Generating Superposition of Transverse Ising Ground States by Quantum Interaction
Koushik Das, Sayan Mondal, Ujjwal Sen
Abstract
We investigate superpositions of quantum Ising ground states generated through a quantum interaction, in which the nearest-neighbor coupling strength is promoted to an operator acting on a central auxiliary spin. We prepare the central spin in a superposition, which allows the chain to evolve coherently under different effective interaction strengths. The states are prepared by adiabatic evolution of the joint system comprising the central spin and the many-body spin chain, followed by a measurement on the central spin and post-selection of a specific outcome. It is possible to prepare superpositions of transverse-field quantum Ising ground states belonging to the three magnetic phases, namely, the paramagnetic, ferromagnetic, and anti-ferromagnetic phases. For a spin-1/2 central system, it is possible to prepare superpositions of ground states belonging to any two phases, while for the spin-1 case, the superpositions may contain three ground states, each belonging to one of the three phases. We analytically derive expressions for the magnetization, spin correlations, and entanglement of such states. The magnetization and spin correlations show distinct features in different superposition sectors. The nearest-neighbor entanglement is largest where the transverse field and Ising interaction compete, and its behavior differs from that of the magnetic observables. We also derive the quantum Fisher information relevant to estimating the Ising interaction strength. And it exhibits multiple peaks near the critical points of the constituent ground states, with the maximum value showing Heisenberg scaling, for the parameters and system sizes examined.
AI Summary
The paper studies how to prepare coherent superpositions of transverse-field Ising ground states belonging to different magnetic phases, and how their magnetic, entanglement, and parameter-estimation properties depend on the components. The Ising coupling is promoted to an auxiliary-spin operator, $\hat J=J\mathbb I+\Delta(t)\hat Z_C$. Starting with the auxiliary spin in a superposition, an adiabatic ramp from $\Delta=0$ to $\delta$ produces branches with couplings $J+m\delta$. Measuring and post-selecting the auxiliary spin yields a superposition of the corresponding ground states. A spin-$1/2$ auxiliary system permits two components, including ferromagnetic–antiferromagnetic combinations; spin $1$ permits three components spanning ferromagnetic, paramagnetic, and antiferromagnetic phases. The post-selection probability is expressed through ground-state overlaps. Jordan–Wigner methods and generalized Wick contractions give analytic expressions for magnetization, nearest-neighbor correlations, concurrence, and coupling-estimation quantum Fisher information (QFI). Parameter-space calculations, mainly at $L=100$, show concurrence concentrated where field and interaction compete, rather than where magnetic correlations are largest. Additional QFI peaks occur near the constituent critical points, $J+m\delta=\pm h$. For one examined parameter choice, the maximum QFI scales consistently with $L^2$ over $L=100$–$1000$. The generalized geometric entanglement measure is reduced to $(1-M_z)/2$ using single-site bipartitions, whose optimality is checked numerically only for $L=4,6,8$. The QFI scaling is likewise restricted to the parameters and sizes examined, and its precision interpretation assumes a fixed number of successful post-selected preparations. Whether interphase coherence provides an operational advantage remains open.
Error found: The formula GGM=(1-M_z)/2 fails for allowed two-component states: at J=0 and sufficiently large delta/h, an odd/even bipartition has a larger Schmidt coefficient than any single-site bipartition. The reported fixed-coefficient QFI omits the generally J-dependent relative phases accumulated during the stated adiabatic preparation, so it is not automatically the QFI of that protocol's output.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 12 pages, 8 figures