Suzhou Delta Semiconductor Technology Co., Ltd

Abstract:

Downsizing may come to the end of 14 nm is inflection point

At present do manufacturers of chip facing difficult decisions to make. Following the advanced process? Or "no"? In fact today to see who the most advanced technology nodes, by Intel first breakthrough, then it design the AASP chip 10% follow it. 10% is not a random number, but said the 10% products to achieve mass production, to meet the needs of the market. In 130nm many manufacturers chasing, who are almost ready to fall behind. But to 90/65nm find ASSP products has been a 2 year lag, process capability depends largely on the foundry. To 32nm discovery and Intel out of the product of the time has been 4 years gap. Because the foundry point of interest (sweetspot), or AASP and ASIC interest point total after the process is the most advanced in the world, that is, the consumer electronics market, including mobile phone etc. the demand will not process adopts the most advanced at the outset, only when the market demand up, will consider to get more the benefit of using the most advanced process.

To measure the effect of reduced size using the average cost of each gate to consider, to 22 nm may be reversed, and began to rise. The industry began to question, the size will continue to shrink? It seems that Moore's law near the end key not necessarily in the technology can achieve, but more important from the economic aspect, considering the size of investment to narrow the need, and its rate of return in where?

As all life microfilm method may extend optics in a semiconductor manufacturing process is very important, although whether EUV or EB electron beam, due to the wave length is very attractive, but to achieve industrial production level, still need time. The so-called "distant water can not solve the thirst". So the current 193nm immersion, plus two times (or even 4) patterning technology and DFM technology to realize the microfilm is a realistic solution. According to the current practice, the method has been able to achieve the level of 22-20nm, but can not break through 14nm, it is hard to predict, so 14nm could be a turning point.

But recently there are good news, according to equipment applied materials company said in its Q2 financial meeting, Intel has been ready for 14nm equipment orders.

In addition, the recent Intel, TSMC and other semiconductor technology that is 7Nm, 5nm, believe not flicker, because from the wavelength 14nm and BEUV EUV is 6.8nm, the realization of 7Nm in a laboratory environment, 5nm, is entirely possible. But there are two aspects must be considered, one is due to the optical lithography technology has come to an end, must use EUV or EB electron beam generation microfilm technology, on the other hand is that these methods are feasible in theory, but the realization of industrialization production Shang Youzhu key technology and supporting materials need to solve. Therefore the technology of the future must take consideration of cost and return on investment in the first place, the economic feasibility of more key.