6.1 Fabrication

The potential for single-electronics is no longer a question of physics but of fabrication. Therefore most of the work, to lead single-electronics to a success, has to be done in process technology. Is it possible to mass produce structures with nanometer feature sizes? Granular films made from metals or semiconductors provide feature sizes down to one nanometer. Table 6.1 gives the required feature size and achievable integration density for operation at 4.2 K, 77 K, and 300 K. To operate a circuit at the temperature of liquid helium is only acceptable for the largest and highest-performance computers, where the cost of the cryogenics is only a small fraction of the overall machine cost. The availability of small self-contained closed-cycle liquid nitrogen refrigerators could make single-electronic integrated circuits possible for high performance workstations. However, the goal is to develop fabrication techniques for circuits operating at room temperature. Room temperature operable single-electron devices could have, due to their ultimate low power consumption, a high impact on mobile electronic equipment, such as notebook computers.

 

Table 6.1: Mandatory feature sizes and achievable integration densities for single-electron circuits which work at 4.2 K, 77 K or 300 K.

temperature	thermal	feature size	feature size	integration
	energy	for Coulomb	for Coulomb	density
		oscillations	blockade	gates/ $\mbox{cm}^{2}$
4.2 K (liquid helium)	0.36 meV	15 nm	6.0 nm	109
77 K (liquid nitrogen)	6.64 meV	4 nm	1.6 nm	$1.5\cdot 10^{10}$
300 K (room temperature)	25.9 meV	2 nm	0.8 nm	$6\cdot 10^{10}$

 

Much more important than room temperature operation is the independence to random background charge, since it can destroy device behavior at any temperature. The random background charge dependence can be addressed either on the fabrication level, by searching for impurity free fabrication techniques so that no background charges exist, or on the circuit level, by looking for circuits which can cope with random background charges. For example circuits which build on Coulomb oscillations instead of the Coulomb blockade are independent to random background charges, since oscillations exist independent of background charges. Despite first proposals and solutions to this problem much more research has to be done in this direction. Ultimately, the solution of the random background charge issue will decide over the future of single-electron integrated circuits.