– 19 –
radio-frequency single-electron refrigerator was proposed. The simplest realization of this device is a nonbiased single-electron box with normal metal (N) island and a superconducting (S) lead, with just one NIS tunnel junction (I = insulator). Such a cyclic refrigerator is expected to remove energy kT at the frequency f of the gate drive. Thus it yields a cooling power of order kTf. We demonstrated experimentally the influence of Coulomb blockade on refrigeration in a single-electron transistor in a device now coined “heat transistor”. Subsequently we discussed a NIS junction subjected to the thermal noise of a hot resistor: we showed that this simple device can act as a Brownian refrigerator. The system provides a particularly illustrative example to discuss “Maxwell’s demon”, yet not violating the second law of thermodynamics.
Fig. 1. Gate-controlled NIS refrigerator.
I = nef
Antti Kemppinen, Matthias Meschke, Mikko Möttönen, Jukka Pekola, Olli-Pentti Saira, and Juha Vartiainen
Collaborators: Dmitri Averin (Stony Brook), Antti Manninen (Mikes), and Yuri Pashkin (NEC)
The work on synchronized electron pumping dealt with three different systems in 2007. Work on the sluice, which is a fully superconducting current pump with combined magnetic flux and gate voltage control, resulted in record high pump current exceeding 1 nA. This promising device suffers, however, from supercurrent leakage, which is a topic of present and near future research. An experiment using the sluice in a closed superconducting loop has also been performed in 2007. It leads to the observation of Berry phase, which is obtained via its relation to the pumped current in a phase coherent configuration. A major discovery of the year was, however, the hybrid SNS turnstile, where accurately positioned current plateaus at I = nef were measured, see Fig. 2. It is a single-electron transistor where charge states are stabilized by the superconducting gap of the leads. This device is extremely simple, and, surprisingly, it had not been realized earlier during the 20 years long quest for a quantum standard of electric current. Yet there is a long way to satisfy the stringent requirements of quantum metrology: in the first experiments the plateaus were determined with precision of 10-3 – 10-2 only, whereas a current standard requires 10-7
Annual Report 2007