NewsCompany newsIndustry News

2019

03/25

Graphene is expected to replace silicon, bringing the semiconductor industry the next spring

Semiconductor companies find themselves in a difficult position. For decades, conventional innovation in silicon has kept the industry profitable and making impressive performance improvements. Recently, it has become increasingly difficult for companies to derive more value from silicon. This dilemma has led companies to wonder what materials to replace silicon and when. Taking graphene as an example, it is known as a miracle material with the same or beyond silicon performance potential. However, the commercialization of this material may take 25 years, requiring substantial investment in R&D and capital costs before it can be put into production. Because so much of the money is currently allocated to silicon, executives must determine the right time to shift their focus to the next substance, even if the effect is not necessarily guaranteed.

MetInfo enterprise content manager system | MetInfo CMS

The road from breakthrough discovery to transformational industrial applications may be long and circuitous. Usually, the first major discovery that is likely to follow is decades of development, improvement and experimentation. Even so, there is no guarantee of success. Laboratories around the world are full of once promising technologies, but these technologies have never been commercially available in the market. This precedent leaves executives uncertain about when and where to invest in emerging innovation. For every company betting on emerging digital technology, dozens of competitors have completely missed the trend, and they have to catch up. Time will tell us whether Kodak's recent entry into the Bitcoin industry is an isolated initiative or a long-term strategy with foresight.

Semiconductor companies find themselves in a difficult position. For decades, conventional innovation in silicon has kept the industry profitable and making impressive performance improvements. Recently, it has become increasingly difficult for companies to derive more value from silicon. This dilemma has led companies to wonder what materials to replace silicon and when. Taking graphene as an example, it is known as a miracle material with the same or beyond silicon performance potential. However, the commercialization of this material may take 25 years, requiring substantial investment in R&D and capital costs before it can be put into production. Because so much of the money is currently allocated to silicon, executives must determine the right time to shift their focus to the next substance, even if the effect is not necessarily guaranteed.

The challenge goes far beyond graphene: as semiconductor companies seek to identify and harness the next wave of innovation, executives must take different approaches. A broader perspective is needed to understand how seemingly different development materials can create new business models and applications. Semiconductor executives should use this perspective to develop a long-term strategy to extract value from existing materials and technologies while monitoring new and innovative technologies. This thinking mode will enable enterprises to cope with both known and unknown challenges in the next few years.

Silicon: Difficult steps?

Silicon, the main material used in semiconductor industry, has always been in accordance with Moore's law in history, bringing unprecedented progress. Advanced analytical technology, augmented reality technology, self driving cars, digital technology and Internet of things are all the most advanced technology innovation technologies in modern life, and all of them require higher performance of silicon materials. However, there are serious doubts about the future of silicon and its ability to support innovation: the following three trends can show this.

Slow performance improvements lead to pricing pressures

Silicon has provided a stage for designers and engineers to show their skills, and its performance has been continuously improved for decades. Looking at the data from the 1970s, we can see that these performance improvements are exponential. However, this rate has slowed down significantly in recent years. The processing power of personal computers has stabilized and the performance of smartphone processors has slowed down - in short, silicon is entering a recession (see Table 1). These trends mean that companies that build competitive advantage on the basis of continuous innovation begin to lose their leading position as other companies develop.

MetInfo enterprise content manager system | MetInfo CMS

Increasing Funds and R&D Costs

As semiconductor companies move to the next generation of wafer plants, their costs will continue to rise. In order to achieve performance improvement, we estimate that companies must increase capital expenditure by up to 40% (requirements for new equipment) and R&D expenditure by 150% to achieve the same throughput (see Table 2). The main reason for the increase in capital costs is production equipment, which has increased by about $2 billion since the industry's transition to multiple schemas. Not surprisingly, manufacturers of integrated devices have rapidly increased their R&D investments in leading node technologies.

MetInfo enterprise content manager system | MetInfo CMS

Physical Limitation of Silicon

In addition to the commercial challenges, it is uncertain whether silicon performance will continue to improve, because innovation has reached the physical limitations of materials. For example, the length of the node is approaching the width of the conduction channel, which will severely limit the performance: silicon transistors will stop working due to small-dimensional quantum effects such as tunneling effect, leakage and thermal problems. Limitations in lithography, instrumentation and nanostructure manufacturing will also hinder progress.

Why can graphene change the current dilemma?

The industry is experimenting with several exotic new materials, including silicone, germanium and black phosphorus, but graphene is touted as the most promising material (see Table 3).

MetInfo enterprise content manager system | MetInfo CMS

In 2004, two researchers at the University of Manchester, UK, discovered an atom-thick graphene, which sparked speculation that it might become an advanced alternative to silicon. The properties of graphene make companies in all walks of life salivate: It is estimated that graphene is about 250 times more mobile than silicon, and its flexibility and other characteristics make it an ideal choice for a range of applications ranging from battery technology to touch screen optoelectronic products. Recent patents, academic papers and research papers have demonstrated the widespread interest in graphene.

Nevertheless, the use of graphene is still very difficult. So what hinders the adoption of graphene? We have identified four limitations, two technical limitations and two industrial limitations. In technology, bandgap engineering remains a major obstacle: without bandgap, graphene switches cannot be turned off. Over the past decade, researchers have focused on solving this problem, but have not yet solved it. In addition, graphene manufacturing must produce high quality crystals and be compatible with existing complementary metal oxide semiconductor (CMOS) devices. In industry, wafer factories need to invest a lot of money, but most of the resources of semiconductor companies are related to the current wafer factory improvement plan. In addition, silicon already has an integrated value chain, but it will take billions of dollars to re-create a value chain for graphene.

Given these uncertainties, we expect that graphene adoption and market growth will be divided into three stages: the reinforcing agent stage, the replacement silicon stage and the revolutionary electronic stage (see table 4).

MetInfo enterprise content manager system | MetInfo CMS

In the near future, we hope that graphene can be used as a silicon reinforcer, because graphene protective layer can be used to improve the reliability and performance of interconnection. At present, 14 nano tantalum nitride metal barriers are used on copper interconnects to prevent diffusion into silicon. In the gap less than 10 nanometers, diffusion becomes the main cause of equipment failure. Graphene barriers have several advantages over other substitutes such as ruthenium and cobalt, including better protection, only 1/8 of their size and 30% faster interconnection speed.

There are two main reasons why graphene can not be widely used at present. The requirements of graphene transfer and coating process need to be fully developed and integrated in the manufacturing process. In addition, the cost of graphene must be greatly reduced in order to achieve large-scale commercial production. We expect graphene to be a viable alternative to silicon in at least five to ten years to solve these problems.

In the next 10 to 25 years, graphene will replace silicon as the main material of semiconductors, provided that researchers can find ways to overcome their band gap limitations. Even so, graphene will take advantage of the advantages of its technology in applications (such as high speed, low loss, small scale and flexibility) to be more suitable for electronic applications than other materials (see Table 5). Our analysis calculates that the total addressable market value of graphene in data processing, wireless communications and consumer electronics products will be $190 billion.

MetInfo enterprise content manager system | MetInfo CMS

Overall, optimistic forecasts show that the market value of graphene semiconductors is expected to reach about $70 billion by 2030.

How should leading semiconductor companies develop?

History shows that some technologies take a long time to commercialize, but once they enter the market, they can quickly change industries. In our experience, companies that have used extensive networks to discover the next revolutionary technology are often more able to withstand industry disruption.

The prospects for graphene are offset by the serious technical and commercial challenges discussed, which may hinder the use of graphene as a silicon substitute. Therefore, in assessing the real potential of graphene, semiconductor industry executives should use structured innovation methods to assess their choices. Innovation X-ray consists of 10 major problems in three categories (innovation strategy, technology interruption and innovation practice) (see Table 6). Solving these problems can help business leaders to better understand their organizational capabilities while pursuing innovation, and support exploring different solutions with or without graphene. As a result, a strategy has been developed to prepare enterprises for subversive, technology-driven industry changes.

MetInfo enterprise content manager system | MetInfo CMS

After a long and fruitful development of silicon, executives began to think about what to replace silicon and provide a similar innovation curve. The properties of graphene have inspired people's imagination, but so far its physical limitations have prevented it from being named the heir of silicon. Recent technological innovation histories show that developments are likely to change rapidly - so executives should see graphene as a powerful competitor. Whatever the end result, semiconductor companies can position themselves by adopting a thinking model focused on structural innovation to deal with disruptive technologies and make themselves stand out in the competition.