Guéret, Blanc, Germann, and Rothuizen reply

Abstract

Gueret, Blanc, Germann, and Rothuizen Reply: In the Comment [1] on our Letter [2], Beton, Eaves, and Main claim that our "observations are more consistent with an increase in the potential in the quantum well due to the electrostatic effect of the gate voltage than with a gatecontrolled quantum confinement effect," that ''the peaks in the conductance are more likely to be due to transitions between states localized by random variations in the lateral potential," and that ''the experiment described shows only an overall shift of the conductance peaks with increasing negative gate bias." The authors of the Comment conclude that our observations are to be interpreted as transitions through localized (donor) states without gate confinement, and refer to their work with other authors (Dellow et ai [3]) for this recently proposed model.Although a model based on transitions through localized states has (in our opinion as well) its own range of validity, it is certainly not the universal explanation for features in conductance or liV) data. In particular, the presence of potential fluctuations is not a sufficient condition to preclude gate confinement and its observation. Channel size is quite important in this respect.It may therefore be necessary to begin the discussion by stressing a significant difference between our work [2] and that of Dellow et al. [3], namely, the difference in the lateral device and channel sizes reported.The device reported in [3] has a designed (lithographic) size of 1 ^m, and can be squeezed with a gate voltage of -5 V to a channel diameter estimated from the data in [3] to be about 300 nm. With such large sizes (and large also compared to the spatial scale of potential fluctuations), there is virtually no hope of observing gate confinement. The device behavior is in this case most likely dominated by the local nonuniformities of the potential landscape, as indeed reported in [3].In contrast, our work [2] deals with designed (lithographic) sizes in the 0.4-to 0.3-pm range, which can further be squeezed to channel sizes smaller than 100 nm with a gate voltage of -2 to -3 V. This is a spatial scale which is now approaching that of possible potential fluctuations. Moreover it yields measurable confinement energies of the order of a few meV (about 10 times larger than they would be in the device reported in [3]).In terms of both confinement energies and channel size, our work [2] deals therefore with quite a different regime from that reported in [3]. And although nonuniformities certainly play a role in our case as well (as noted in our Letter [2]), we find that the prominent features observed and their dependence on channel squeezing (particularly the appearance and gradual evolution with gate voltage of the 1-nA current plateau) provide good evidence for the proposed gate confinement effect. Local fluctuations of the potential are significant in our case insofar as they affect the mode ladder in the dot, which makes a direct and detailed quantitative comparison with idealized models difficul...

Electron Transport in Quantum Dots

“…To date there only have been a few reports of how these difficulties have been solved. These include transport through two-terminal dots [133][134][135][136][137], transport through threeterminal gated dots [130,131,[138][139][140][141][142], and capacitance measurements on twoterminal gated dots [15,50].…”

Electron transport through double quantum dots

“…Increasing the bias voltage lifts the blockade. The current shows a stepwise increase: each time when an additional level enters the bias window Ϫ͉e͉V, an extra transport channel is opened and the current increases (Gué ret et al, 1992;Johnson et al, 1992;Su et al, 1992;Foxman et al, 1993). Hence the voltage spacing of the current steps directly reflects the energy spacing of the levels.…”

Characteristics of electron transport through vertical double-barrier quantum-dot structures: Effects of symmetric and asymmetric variations of the lateral confinement potentials

“…The lateral confinements are created by, e.g., etching the structures to columns, 5,14,15,20 ion-beam implantation, 3 and the use of a surface metal gate. 9,11,13,18,19,21 The method of using a surface metal gate is of particular interest, due to the possibilities of tuning the lateral confinement with a voltage applied to the gate and thus of manipulations of the electronic structure in the quantum confined region and the transport properties of the devices. Very recently, an embedded grating metal-gate technique was developed for fabrication of laterally-confined DBRT structures.…”