Reply to Comment on ‘Giant suppression of shot-noise in double barrier resonant diode: a signature of coherent transport’

Shot noise suppression below 1/2 of the full Poisson value in double barrier resonant diodes is confirmed to be a signature of coherent rather than sequential tunnelling transport. We reply to the arguments of the previous comment which dispute the above claim. We anticipate the development of a rigorous theory that improves previous approaches without contradicting the essential findings we recently reported (Aleshkin et al 2003 Semicond. Sci. Technol. 18 L35). In a recent letter [1] we showed that suppression of shot-noise in resonant diodes below 1/2 of full shot-noise value 2qI , with q the unit charge and I the steady current, is a signature of coherent against sequential tunnelling transport regime. The proof stems on the fact that the theoretical approach we developed for the coherent tunnelling regime predicts under suitable bias and temperature conditions a Fano factor below 1/2 while the standard sequential tunnelling regime [2, 3] never allows for a value of the Fano factor below 1/2. These findings were in agreement with experiments. In their comment, Blanter and Büttiker (BB) contest our theoretical discussion on the basis of the results they obtained on the same subject [4, 5] and which never allow for a Fano factor drop below 1/2. To this purpose the comment reports several arguments on which we reply in order of appearance. In our opinion, the sequential tunnelling regime cannot be interpreted as the semiclassical limit of the coherent tunnelling regime when considering current noise. The microscopic noise source in each regime are of a different nature, even if both correctly recover the Nyquist relation under equilibrium, because the sequential and coherent tunnelling are different mechanism for electron transfer in the resonant tunnelling diode. In the coherent transport there is no scattering during tunnelling while scattering in the quantum well is essential for sequential tunnelling. Moreover, the increase in the number of channels does not introduce scattering. Consider the simple case when the Coulomb interaction is neglected. Then, in the sequential tunnelling regime the fluctuations are described by the differential rates controlling the relaxation of carrier number fluctuations inside the two terminal device through the contacts [3]. By contrast, in the coherent tunnelling regime the fluctuations are described by the transparency of the whole device through the partition noise mechanism. These two descriptions are inherently different and provide in general different results. Therefore, we do not agree on the expected equivalence of phase coherent and sequential results and on the estimated quantum corrections reported in the comment. Concerning the details of calculations which lead to our equation (8) we note the following. We consider high voltage regime, v > EF . However, to our knowledge neither explicitly nor implicitly we assumed E0 EF , and thus we believe that equation (8) is rigorously obtained. We add that a similar expression for the Fano factor was obtained in [6, 7] and 0268-1242/04/050665+02$30.00 © 2004 IOP Publishing Ltd Printed in the UK 665