We use continuous, stochastic quantum trajectories within a framework of quantum state diffusion (QSD) to describe alternating measurements of two noncommuting observables. Projective measurement of an observable completely destroys memory of the outcome of a previous measurement of the conjugate observable. In contrast, measurement under QSD is not projective and it is possible to vary the rate at which information about previous measurement outcomes is lost by changing the strength of measurement. We apply our methods to a spin 1/2 system and a spin 1 system undergoing alternating measurements of the S z and S x spin observables. Performing strong S z measurements and weak S x measurements on the spin 1 system, we demonstrate return to the same eigenstate of S z to a degree beyond that expected from projective measurements and the Born rule. Such a memory effect appears to be greater for return to the ± 1 eigenstates than the 0 eigenstate. Furthermore, the spin 1 system follows a measurement cascade process where an initial superposition of the three eigenstates of the observable evolves into a superposition of just two, before finally collapsing into a single eigenstate, giving rise to a distinctive pattern of evolution of the spin components. Published by the American Physical Society 2024
The information-theoretic formulation of quantum measurement uncertainty \nrelations (MURs), based on the notion of relative entropy between measurement \nprobabilities, is extended to the set of all the spin components for a generic \nspin
s. For a physical class of approximate joint measurements of the spin \ncomponents, we define the device information loss as the maximum loss of \ninformation per observable occurring in approximating the ideal incompatible \ncomponents with the joint measurement at hand. By optimizing on the measuring \ndevice, we define the notion of minimum information loss. By using this \nnotions, we show how to give a significant formulation of state independent \nMURs in the case of infinitely many target observables. The same construction \nworks as well for finitely many observables, and we study the related MURs for \ntwo and three orthogonal spin components. The minimum information loss plays \nalso the role of measure of incompatibility and in this respect it allows us to \ncompare quantitatively the incompatibility of various sets of spin observables, \nwith different number of involved components and different values of
s.