The quantum interference effect is suppressed by dephasing processes. The dephasing in mesoscopic systems at low temperatures is mainly caused by electron–electron Coulomb interactions. Usually, the magnitude of the suppression is characterized by the dephasing time . It has been shown that its T-dependence changes according to the effective dimensionality of a system. Using nano-fabrication technology, we can prepare a mesoscopic cavity which is connected to reservoirs through a pair of thin leads. A mesoscopic cavity with a weak opening can be regarded as a quasi-zero-dimensional system, and the quantum transport properties in such a system are qualitatively different from those in higher dimensional systems. We expect that the dephasing nature is also qualitatively different. We hereafter focus on the case where the classical electron motion is diffusive or chaotic. If the width of the leads is reduced, the dwelling time d for electrons becomes very long. Due to the confining effect, there arises a nearly uniform diffusion mode with the lowest eigenvalue 0 1⁄4 h = d. This mode dominantly contributes the quantum interference corrections to conductance when temperature T is sufficiently low. The present author has studied the Tdependence of for the mode by a self-consistent renormalization approach (SCRA), and found an interesting crossover behavior of . It has been shown that behaves as 1 / T when T 0=kB. If T is decreased below 0=kB, the T-dependence of crosses over to 1 / T and the dephasing effect becomes nearly negligible. In this short note, we reconsider the dephasing problem in an open mesoscopic cavity by adopting a test-particle approach. We explicitly calculate the dephasing factor for the weak-localization correction arising from a timereversed pair of paths with duration . In the long-time regime of h ðkBTÞ , the dephasing factor obeys a simple exponential law expð = 1Þ, as usually expected. However, much slower decay expð ð = 2ÞÞ is observed in the short-time regime of h ðkBTÞ . Here, 1 and 2 are characteristic time scales and satisfy h 1 1 , h 1 2 kBT . It is shown that in the case of T 0=kB is determined by the slow decay expð ð = 2ÞÞ. This is the reason why the dephasing effect becomes nearly negligible in the lowtemperature regime. On the other hand, the exponential law expð = 1Þ determines in the high-temperature regime. The resulting is in agreement with that of the SCRA apart from numerical factors. We set h 1⁄4 kB 1⁄4 1 in the following. We adapt the test-particle approach to our problem. Let us consider a time-reversed pair of paths with duration . We introduce a classical path x1ðtÞ which starts from x1ð0Þ and ends at x1ð Þ, and express its time-reversed path as x2ðtÞ. Let us assume that wave packets 1ðr; tÞ and 2ðr; tÞ follow the classical paths x1ðtÞ and x2ðtÞ, respectively. The test particle interacts with a bath of environment electrons. We introduce the density operator WenvðtÞ for the environment and the state vectors j1i and j2i which correspond to 1ðr; tÞ and 2ðr; tÞ, respectively. The test-particle state is described by a linear combination of j1i and j2i. We assume that the environment at t 1⁄4 0 is in thermal equilibrium with respect to the Hamiltonian Henv for the environment electrons including the Coulomb interactions between them. The density operator at t 1⁄4 0 is given by Wð0Þ 1⁄4 WS Wenvð0Þ, where WS 1⁄4 P i;j1⁄41;2 jiiwijhjj. The interference amplitude is Fð Þ 1⁄4 h1j trenvfWð Þgj2i. The dephasing is characterized by the dephasing factor ð Þ Fð Þ=Fð0Þ. In terms of the density of environment electrons ðrÞ and its average value 0, we define ðrÞ as ðrÞ 1⁄4 ðrÞ 0. The Coulomb interaction of the test particle with the environment electrons is expressed as
Read full abstract