Nowadays there are processes in which the channel length of a transistor can be made 45nm or even 32nm narrow. This miniaturization evolution has been going on for some time. But will she ever stop in its current form? When making ICs smaller and smaller, one will eventually bump into a physical limit that certain parts can no longer be made smaller, because then one will have to build structures that are smaller than an atom. So one day a physical limit will be reached. So I wonder whether this ultimate physical limit, in which building blocks are no bigger than an atom, is also a functional limit. Which is to say whether it would actually be possible to make functioning chips, with such dimensions, or would everything go awry? Thanks for the effort.
Answer
Dear Niels
The race of ever-increasing miniaturization has been going on for many years. Gordon Moore predicted this exponential evolution as early as 1965. Soon his prediction was increasingly regarded as a concrete goal that the semiconductor industry set itself.
Where chips were made in 10 μm technology (10000 nm) in the early 1970s, we are now, as you write yourself, at 45 nm or even 32 nm. The International Technology Roadmap for Semiconductors, which indicates what we can expect in terms of technology, states that a feature size of 9 nm can be reached by 2024. So for now, that’s the limit of what the industry thinks it can handle.
But before that happens, many technological barriers have to be overcome. Today, these technical problems are mainly in the production process. We are currently bumping into the limits of classical lithography, among other things, and solutions are being sought in the direction of double patterning.
In 11 nm technology, the gate length of the transistors becomes only 5 to 6 nm. In the lattice structure of Silicon this corresponds to about 25 atoms. The physical limit for transistors as we know them will then be approximately reached. After all, you will always need several atoms for a transistor. Problems also arise at those short distances, for example due to quantum tunneling.
So for transistors the limit is about ten atoms. You cannot make a transistor with a single atom.
But there is still life beyond the transistor … In the early 1980s, the idea of ​​the quantum computer that calculates with quantum bits arose. By superimposing, a quantum bit always represents a linear combination of the 0 and the 1 state. Quite recently it has been possible to put these originally purely theoretical concepts into practice. For example, there is an algorithm for prime factorization using a quantum computer, which was successfully used at IBM in 2001: 15 = 3 x 5.
I hope this answers at least part of your question.
Kind regards
Peter
Answered by
dr. ir. Peter Bertels
http://www.ugent.be
.