We have already heard from the leading manufacturers of semiconductor products that they intend in the coming years to master the 7-nm and 5-nm technology, with many preparing to introduce at this stage for lithography equipment with superhard ultraviolet (EUV). In an interview with the resource EE Times researcher ARM Yerik Greg (Greg Yeric) shared his thoughts on the difficulties to be faced processor developers in an effort to extend the so-called "Moore's Law".
First, he argues that the 5-nm process technology market participants have a sufficient degree of confidence in the ability of its master. As for the 3-nm process, then everything is a bit more complicated. There are various technological projects to shrink the size of the elements to 3 nm, but no one can guarantee that the existing proposal can be successfully put into practice. The problem is not even in the ability of engineers to create smaller and smaller transistors - in this respect, as the time, there are a lot of talented ideas. It is much more difficult to reduce the size of conductors and connecting them to increase their packing density. The physical interaction of neighboring elements already significantly reduces the possibility of increasing the speed of the transistors in the transition to a new process technology level. Yerik argues that talk about the growth performance standard of 25% by changing the process technology is no longer necessary. Moreover, developers can face the fact that the speed of the transistors even worsen in the pursuit of benefits by cost.
Already within the 7-nm process technology a number of these problems call into question the value of the speed growth. According to Greg Yerika, the frequency potential of 7-nm products will be severely limited. EUV-lithography should be introduced as part of a 5-nm technology, a convinced ARM representative . It will be necessary to develop new types of memory cells.