The ampere (A) will soon be redefined in terms of the elementary charge, and single electron transport (SET) devices are a promising method to count and collect electrons. NIST is testing variations on a pumping configuration in which a voltage applied to a device gate pushes an electron across a junction barrier and onto an "island" made of a microscopic quantum dot. A second gate voltage applied on the other side of the island prompts the electron to pass through a second barrier and into the drain. Repeating the process at precise time intervals provides a measurable stream of individual electrons.
The central problem in using SETs as an electrical current standard comes from the difficulty in achieving a practical level of current simultaneously with the low uncertainties characteristic of other electrical standards. For example, the result with the best uncertainty of a few parts in 108, only produces an output current of a few picoamperes (10-12 A).
Currents large enough to be of practical use as a standard must come from large arrays of SETs, each precisely synchronized and coordinated with the rest. That requires devices fabricated to identical specifications that are extremely stable and not unduly prone to thermal or other stresses. To meet those criteria, the NIST team chose all-silicon construction instead of the more conventional metal/oxide.
They have started pumping measurements in existing devices, which have island sizes in the range of 100 nanometers, and cooled by a dilution refrigerator to ~ 10 mK. The results are not yet at metrological accuracy but do indicate where improvements can be made in the measurement system.
Next generation devices under development include parallelized devices as well as devices with multiplexing features. The devices will have smaller islands and in some cases islands where the size can be tuned electrostatically. To improve the reproducibility of devices and to help eliminate unintentional formation of quantum dots, the project team is also investigating strain in companion tunnel junction devices.
Elsewhere at NIST, scientists are at work on an alternative design for a current standard. It relies on a phenomenon called quantum phase slip (QPS) – a sort of converse complement of the Josephson voltage effect. Whereas a Josephson junction biased with a current and irradiated with microwaves generates step-wise quantized plateaus of voltage, a circuit made of extremely thin superconducting nanowire biased with voltage will conduct only exactly quantized amounts of current when frequencies of a few GHz are applied.
For this to occur, however, the wire needs to be so thin that it behaves as if it were one-dimensional. In practical terms, that’s on the order of 100 square nanometers in cross section. Producing such a wire, and verifying that it is superconducting at extremely low temperature, is a formidable challenge. To date, NIST researchers have succeeded in fabricating tungsten silicide wires that are 20 nm wide and 4 nm thick that are superconducting at ultracold temperatures.
Next steps include characterizing and possibly eliminating thermal and quantum effects that may limit the accuracy of current steps. But the technology offers the possibility of an easily deployable current standard that is traceable to NIST standards, can be manufactured to well-controlled specification, and is inherently simple to operate – requiring only a since continuous-wave microwave drive.