arXiv:2609.21355
Weighted Berry curvature and global geometry of mixed quantum states
Dominik Kuczyński, Erik Sjöqvist
Abstract
We examine the geometric interpretation of the quantum geometric tensor proposed in [Phys. Rev. B {\bf 110}, 035404 (2024)] for mixed quantum states, focusing on its imaginary part, which is proportional to a weighted sum of the Berry curvatures of the eigenstates of the density operator. While the real part naturally decomposes into Fisher--Rao and weighted Fubini--Study contributions, we show that the imaginary part does not, in general, coincide with the curvature of a connection on a globally defined U(1) line bundle whose holonomy yields a mixed-state geometric phase. Using a two-level system as an explicit example, we demonstrate that its surface integral depends on the choice of surface bounded by the same closed path, with ambiguities that are not integer multiples of $2π$.
AI Summary
This paper asks whether the imaginary part of a proposed mixed-state quantum geometric tensor can define the curvature of a globally consistent $U(1)$ connection, with holonomy interpretable as a geometric phase. The tensor’s real part separates into Fisher–Rao and eigenvalue-weighted Fubini–Study contributions; its imaginary part is proportional to $\Omega=\sum_n\lambda_n F^{(n)}$, the eigenvalue-weighted sum of eigenstate Berry curvatures. Two obstructions are identified. For parameter-dependent eigenvalues, $d\Omega=\sum_n d\lambda_n\wedge F^{(n)}$ is generally nonzero, violating the closure condition for an Abelian curvature. Even with constant eigenvalues, closure does not ensure the integral periods required for a global line-bundle connection. This is demonstrated for $H=\mathbf b\cdot\boldsymbol\sigma$, whose eigenstate bundles over the direction sphere have Chern numbers $\mp1$. Integrating $\Omega$ over two oriented hemispheres sharing the same boundary gives a difference $2\pi(\lambda_- -\lambda_+)$. For full-rank, non-degenerate states, $0<|\lambda_- -\lambda_+|<1$, so the difference is not an integer multiple of $2\pi$. The corresponding weighted Chern integral is therefore noninteger. The conclusion is that this gauge-invariant two-form does not generally define a path-dependent geometric phase through surface integration. The analysis assumes non-degenerate nonzero eigenvalues and does not exclude special cases admitting a connection; the pure-state limit recovers the usual quantized ambiguity. It distinguishes the weighted-curvature construction from both the interferometric phase, which uses an argument of weighted phase factors, and Uhlmann–Bures geometry, rather than ruling out mixed-state geometric phases generally.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.