IC manufacturing rainy silicon-based IC is not more Moore's Law Lei Chi

In the field of ICs, the manufacturing process determines whether the final design can be achieved. Essentially, it is the industry in which manufacturing determines the design. In this field, the most skilled use of this method is a vendor like IBM or Intel. However, due to the IC industry's competition, the pace of IC manufacturing is already approaching the limit of materials, and the barrier between circuits is almost at the atom level. Some problems that would not have arisen before have gradually emerged. For example, if semiconductors need to be mixed, they need to control specific quantities, and they must not allow free combination of impurities as before. This will greatly increase the difficulty of IC manufacturing; the insulating layer of semiconductors will also be due to the lithography chemicals themselves. The degree of unevenness will directly affect whether or not the IC can operate normally. Dr. Chen Ziqiang, who is an IBM academician, pointed out that when the width of the interlayer between circuits may be 320 angstroms, the pits therein may be 80 angstroms. That means that, in some extreme places, the width of the compartment is only 160 Angstroms. When is the current material still available?

Silicon-based ICs do not leap over Moore's Law Although this is something that will happen in the next fifteen years, scientists who have focused on cutting-edge scientific research have been unable to wait. Not only is the material limitation, the process below 22nm cannot make the circuit run stably, but also because the transistor is integrated too much, making the power of the processor reach an astonishing point.

We can see that in the past few years, we have not really reduced the power consumption due to the improvement of the production process. Instead, we have increased the frequency and the number of transistors. This is the current IC frontier. The institute needs urgently solved problems. At the same time, because the insulation layer distance of the CMOS transistor is greatly reduced, the reduced threshold will be lower than 1V, which means that the IC is greatly increased in the probability of interference, and the supply current will be on the order of a few hundred amps. It is indeed quite Deaf people.

However, Moore's Law has not been broken because of this, in its proposed nearly 40 years, the semiconductor industry has been careful to follow Moore's Law, do not go over the mine. According to IBM Research Institute ThomasJ. Dr. Chen Ziqiang, Vice President of Science and Technology at Watson Research Center, said that in the foreseeable future, Moore's Law will not be broken. Although people in the IC field have lost interest in frequency chasing, the multi-core layout will also give Moore's Law a glorious glory.

Multi-core is the only way to comply with Moore's Law. Now, Intel's 65nm quad-core processor is already like Chen Bing million, and TI's 65nm process has already entered the factory. However, we have found that in the field of IC, more people compete with themselves and lack support for the industries they serve. This proves that IC is over-competitive.

From a technical point of view, the IC field does not have to rely on the whip to run faster than anyone else. It still needs to pursue faster and pushes itself to the top of Liangshan. The bottleneck in the IC field is also followed.

Do we need such advanced technology now? For example, reducing the operating voltage does increase the frequency and reduce the power consumption under the same performance. However, the pursuit of performance has led to a significant increase in the degree of chip integration, which has led to a significant increase in power consumption and a gradual sensitivity to external interference. In some complex environments, it is less useful.

In addition, the substantial increase in technology, this progressive technological achievements have not been fully inherited in the product, but make the threshold to enter the IC field, and increase the pressure on the output. In order to feed such a factory, a large number of designs and ideas must be made into chips to make the industry chain alive.

So what are the tricks to solve this problem? Dr. Chen introduced that there are mainly three aspects of research on alternative materials, logical innovation and system innovation.