Shot noise of photon-excited electron-hole pairs in open quantum dots

Abstract

We investigate shot noise of photon-excited electron-hole pairs in open multi-terminal, multichannel chaotic dots. Coulomb interactions in the dot are treated self-consistently giving a gaugeinvariant expression for the finite frequency correlations. The Coulomb interactions decrease the noise, the strong interaction limit coincides with the non-interacting adiabatic limit. Inelastic scattering and dephasing in the dot are described by voltage and dephasing probe models respectively. We find that dephasing leaves the noise invariant, but inelastic scattering decreases correlations eventually down to zero. 73.21.La, 73.50.Td Investigation of noise 1-3 generally provides information not available from current measurements, such as effective charge and quantum statistical properties of the carriers. Most of the work has focused on noise in the presence of stationary (dc) applied voltages. Of interest here are shot noise measurements in photon-assisted transport.4 Initial experiments investigated noise in the presence of both ac-and dc-potentials.2 However, recently Reydellet et al.3 reported shot noise experiments in the presence of ac-potentials only. There is no dccurrent linear in voltage. Reydellet et al.3 subject one of the contacts of a two-terminal conductor, a quantum point contact, to rf-radiation. Good agreement was found between experiment and theory. 5,6The purpose of our work is to investigate a generic conductor with one or several contacts subject to acpotentials and to investigate the effect of Coulomb interactions, dephasing and inelastic scattering on the photonassisted noise. We consider a multi-terminal chaotic quantum dot 7,8 connected to electronic reservoirs via quantum point contacts α = 1, 2, .., M with a large number of channels (see Fig. 1). The reservoirs are subject to oscillating potentials V α (t) = V α cos(ωt + φ α ) at the same frequency ω, but with arbitrary phase φ α (this is in contrast to the widely investigated case of shape modulating potentials applied to gates 9 ). In the absence of a dc-bias, the elementary excitations generated by acpotentials can be understood as electron-hole (e−h) pairs rather than single electrons.10 The interpretation of the noise in terms of e − h pairs generated in the contacts of the sample provides an intuitively appealing picture 11 of the resulting shot-noise. In the many-channel regime the noise correlator S λµ between the contacts λ and µ fluctuates weakly from sample to sample, and we consider only the mesoscopically averaged shot-noise S λµ .The effect of inelastic scattering and/or dephasing on photon-assisted noise has, to the best of our knowledge, not been addressed in the literature. For dc-biased chaotic cavities, it was shown that the ensemble averaged noise is insensitive to dephasing. 12,13 It was expected 14 and demonstrated 15,16 that this is true not only for the conductance and shot noise but also for higher order cumulants. To investigate inelastic scattering and dephasing we employ here two models. In the first, inelastic ...

Photon-assisted electron-hole shot noise in multiterminal conductors

“…When screening is taken into account the frequency must even be comparable to the inverse charge relaxation time. 29 As a consequence of these conditions, the correlations of ee and hh, as well as eh and he correlations, are identical. Neglecting corrections of the order of (eV /hω) 4 , we find the correlations between electron and hole currents:…”

Elementary charge-transfer processes in mesoscopic conductors

“…37 The photon-assisted noise at small driving amplitudes is due to an electron-hole pair which is created by ac drive ͑with a low probability per voltage cycle͒ and injected toward the scatterer. [38][39][40] However, to obtain the elementary processes, the knowledge of full charge-transfer statistics is needed.…”

Semiclassical approach to the ac conductance of chaotic cavities

“…[10][11][12] This experimental progress has since triggered renewed theoretical interest in time-dependent mesoscopic transport. [13][14][15][16][17][18] One way to tackle the ac-transport problem is to start from linear-response theory for a given potential distribution of the sample. [19][20][21] This involves the difficulty that, in principle, the potential distribution and more precisely its link to the screening is unknown.…”