Microscopic machines :
The microprocessor is a complex beast. The complexity of minimum component costs has increased at a rate of roughly a factor per year.As technology processes, processors shrink in size. The increase in the number of transistors (capacitance) increases the performance of that particular processor. Generally a processor consist of transistors. A transistor is similar to a gate. It stops the flow of current from one side of the gate to the other side. The gate is closed until the transistor is less than a certain 'threshold'. As the voltage crosses the threshold , the gate opens up and lets the current through.Thus a transistor is either ON or OFF depending upon the threshold voltage.
The incredible shrinking through though,throws a wrench to the ideal. As transistors shrink, the wires connecting them shrink and the insulation provided by the transistor gets worse. This can be quoted as an example to a leaking dam with heavy pressure on the bottom. The walls of the dam develops holes and also it gets larger in due course of time, letting more water through.Thus when even the transistor should stop the current, it lets it through (leakage). By this a processor loses as much as 100 watts of current at 45nm.
How can you tame this beast?
Yes you can! Leakage without solutions could cause the processor to leak 100 watts of at 45nm and about 250 watts in total. One part of the solution is the implementation of a technology called silicon on insulator (SOI), pioneered by IBM. Processors are generally built on a substrate of silicon. With SOI , the silicon substrate is replaced by a sandwich of silicon-insulator-silicon. This improves the insulation of the transistor gate. Another part of the solution is using a high-k-material,such as hafnium. Replacing the silicon dioxide with the high-k-material allows increased gate capacitance with the leakage effects.