The time-local master equation for a driven boson system interacting with a boson environment is derived by way of a time-local Heisenberg-Langevin equation. Extension to the driven qubit fails-except for weak excitation-due to the lost linearity of the system-environment interaction. We show that a reported time-local master equation for the driven qubit is incorrect. As a corollary to our demonstration, we also uncover odd asymptotic behavior in the "repackaged" time-local dynamics of a system driven to a far-from-equilibrium steady state: the density operator becomes steady while time-dependent coefficients oscillate (with periodic singularities) forever. Issues of formal exactness aside, the current move from given quantum systems-an atom or radiation field in a cavity-to engineered systems, drives a more pragmatic interest in the theory of non-Markovian open quantum systems. While generalizations of input-output theory to the non-Markovian regime have been considered [5][6][7], more commonly the Schrödinger picture is adopted, where, after the work of Hu et al.
Treatments in the[8], time-local master equations are derived [9][10][11][12][13][14][15][16]. Of particular interest in this paper are the time-local master equation for a driven boson system in interaction with a boson environment (first derived in [11]) and the equation for spontaneous emission from a two-state system, or qubit [9][10][11][12]. They invite a conflation: a time-local master equation for the driven qubit.The driven qubit is an important example, considering its role in quantum information science and the numerous physical realizations. The question of a time-local master equation is an old one. It is raised in Sec. IVA of Ref. [11], where, after first noting obstacles to its derivation, the author mentions an equation reported in a preprint [18], following with: "Such a relatively simple result seems to be inconsistent with the conclusion reached above about the difficulty of dealing with the two level atom problem, and is hence an issue that requires further consideration." The noted equation is absent from the published version of [18] (Ref. [10]), which can be seen to endorse the call for "further consideration."The "relatively simple result" substitutes qubit raising and lowering operators for the creation and annihilation operators in the time-local master equation for the driven boson system. Having appeared first in [18], it reappears in a recent work of Shen et al. [17], along with a derivation from Feynman-Vernon influence functional theory which invokes a coherent state representation in Grassmannian variables for the qubit state. We show in this paper that the derived time-local master equation is incorrect; it is not influence functional theory and coherent-state path integrals that yield a successful derivation, but linearity, which for the driven qubit is lost. While in spontaneous emission [9][10][11][12] linearity is effectively retained-due to the one-quantum truncation-for the driven qubit, multiphoton scattering (...