next up previous
Next: The canonical distribution Up: Incoherent tunnelling through two Previous: Incoherent tunnelling through two

Generalised master equation

 

When it is assumed that the tunnelling rates are small compared to the Coulomb energy and the average level spacing, then use of the master equation is justified [13]. In order to take account of the discrete nature of the energy-levels in the dot, the master equation method for dots with a continuous energy level spectrum [6] has to be generalised. The tunnelling rates not only depend on the number of electrons in the dot, but also on their configuration, i.e. how the electrons are distributed over the available energy levels. When there is a high relaxation rate, then the electrons will revert to their local equilibrium distribution between tunnelling events. This is likely to be the case for small tunnelling barriers which cause the electrons to spend a long time in the dot. Since this distribution will depend only on the temperature, the state of the dot may yet again be described simply in terms of the number of electrons present in the dot.

Let P(p,N) be the probability that the system is in a state which is characterised by N electrons in the dot which are distributed over the energy levels according to a configuration p. Define tex2html_wrap_inline1116 as the tunnelling rate coefficient corresponding to the transition tex2html_wrap_inline1118 by an electron tunnelling through the left barrier. tex2html_wrap_inline1120 refers to the rate that corresponds to the reverse process. The tunnelling rates through the right barrier are defined similarly. Finally, there are the relaxation rates tex2html_wrap_inline1122 which account for the intra-dot transitions at a fixed electron occupation N. The rate of change of the probability of occurrence of the state tex2html_wrap_inline1126 is thus given by
eqnarray34
If the system consists of several dots in series, then the above formalism needs to be extended and it is necessary to define the tunnelling rates between the dots. Define tex2html_wrap_inline1128 as the tunnelling rate corresponding to the transition tex2html_wrap_inline1130 and tex2html_wrap_inline1132 where the subscript i labels the site of the dot. The evolution of the multi-particle state occupation probabilities for a dot not neighboured by a reservoir is given by
eqnarray57
For both the single and the multiple dot system the steady state current can be written as
equation91

When it is assumed that the relaxation rates are large compared to the tunnelling rates, then the number of electrons in the dots is the only significant variable, since knowledge of this quantity allows one to deduce the probabilities of the various electronic configurations. Hence the formalism is greatly simplified. However, the tunnelling rates between states with a different occupation number have to be redefined to include the weighted sum over all possible tunnelling paths. Define tex2html_wrap_inline1136 and tex2html_wrap_inline1138 as the tunnelling rates for an electron tunnelling into and out of the dot through the left barrier with N other electrons present in the dot.
eqnarray101
where P(p|N) is the conditional probability that the system is in configuration p given that there are N electrons in the dot. Since the master equation takes only single electron tunnelling events into consideration, the only contributions that need to be taken into account are those where p and p' differ by the occupation of one energy level. Therefore the double sum over p and p' can be replaced by a single sum over the single particle energy levels k.
eqnarray111
Here P(k|N) is the conditional probability that level k is occupied given that there are N electrons in the dot. For inter-dot tunnelling events the rates are re-expressed in terms of a sum over the tunnelling paths from all levels k in dot i to all levels l in dot tex2html_wrap_inline1170
 equation117

For the case of negligible energy level broadening the probability tex2html_wrap_inline1172 of finding an electron on level k of dot i given tex2html_wrap_inline1178 electrons can be determined from the Boltzmann distribution.
  eqnarray128
where tex2html_wrap_inline1180 and tex2html_wrap_inline1182 is the energy required to add an electron at an energy level k when tex2html_wrap_inline1186 electrons are already present in the dot. tex2html_wrap_inline1188 is the partition function for tex2html_wrap_inline1178 electrons in dot i, tex2html_wrap_inline1194 is the conditional partition function for tex2html_wrap_inline1178 electrons given that level k is unoccupied. When the conditional probabilities are substituted in equation 8, one obtains
 equation136
Tunnelling events between dots will in general not preserve energy, since it is unlikely that energy levels will line up. The high relaxation rate suggests that all tunnelling events can be considered inelastic. Hence
equation150
It follows from the two preceding equations that the tunnelling rates are inelastic in the total free energy of the system. This is also true for tunnelling between dot and reservoir. This implies that the following condition holds true at zero bias for tex2html_wrap_inline1200 dots for all possible occupation numbers tex2html_wrap_inline1178.
equation159


next up previous
Next: The canonical distribution Up: Incoherent tunnelling through two Previous: Incoherent tunnelling through two

Angus MacKinnon - Aonghus Mac Fhionghuin Fri Nov 8 18:10:54 gmt 1996