The elementary components of transistors have kept shrinking, now they have reached 10nm. These are the “gates” a junction point between two layers of material that separates them and makes the switching between conductive and insulating. Based on theoretical and experimental studies it is impossible to have gates smaller than 5 nm. Below this size electrons jump from one layer to the next short-circuiting them.
Now a team of researchers at Berkeley have demonstrated a working gate at 1nm, five times smaller than what was supposed to be the absolute limit. This has been made possible by using as layers molybdenum disulfide joined by a gate made with a carbon nanotube no bigger than a molecule of glucose.
In practice the created a transistor made by a 2D (two dimension) layers, just one atom thick, and 1D (one dimension) gate.
The use of molybdenum disulfide is crucial because it allows a better control of electrons, thus allowing a much smaller gate.
Don’t get over excited though. The Moore’s law has already ceased to work in 2014 when the decreasing size of transistors on a chip was no longer matched by a decreasing cost per transistor. As we move down from 14nm to 10 nm the cost per transistor increases.
The result obtained at Berkeley is amazing but does not address this crucial issue: getting both smaller AND cheaper. Besides, it is a demonstration prototype of a single transistors when we need a manufacturing process able to cram billion of them in a single chip. Yet it is a proof the word “impossible” has to be used cautiously when we deal with technology evolution….