The Millennium Semiconductor Survival Guide

“How do you know that with the ArF technique you will have more line edge fluctuations and irregularities?

Regarding the ArF technology route, there are IEEE magazines, but the papers published by the engineers of Nikon or Canon only talk about the benefits and not the defects. "Lin Benjian asked in surprise.

Zhou Xin said: "In order to mislead competitors, Nikon and Canon deliberately did not talk about the shortcomings of the ArF technology route.

The purpose is to let competitors also step into this pit.

They have invested a large amount of research and development funds in the ArF technology route, and have not found a way to solve the problem of line edge fluctuation and irregularity caused by the ArF light source, so they think that other competitors cannot solve this problem. "

Lin Benjian nodded silently: "That's right.

That's it. "

Zhou Xin continued: "Isn't it a routine practice in the field of integrated circuits to mislead competitors on the technical route?"

Lin Benjian is still very curious. He knows the flaws of the ArF technical route because he has been engaged in research in this field for more than 20 years and he has a large number of contacts.

Due to the non-disclosure agreement, his friend could not tell him directly, but could only hint to him that there were major flaws in the ArF technical route.

He is just guessing about this, a guess without evidence.

Why is Zhou Xin able to speak so firmly and explain the flaws of the ArF technical route so clearly.

It is necessary to know the laboratories that are qualified to conduct experiments on the ArF technology route, which are the few lithography machine companies.

After taking a drink of water, Zhou Xin said, "Bo Lin, I know what you're thinking, did I bribe Nikon's technicians and obtain technical secrets from them.

In fact, the flaws in the ArF technical route can be easily deduced.

The wavelength of the ArF light source is 193 nm, which is shorter than the 248 nm wavelength of the KrF light source. Short-wavelength light sources are obviously affected by more severe scattering, diffraction, and optical near-field effects during the lithography process, resulting in increased line-edge fluctuations and irregularities during pattern transfer.

At the same time, lithography using ArF light sources is used in more advanced process nodes, which leads to the need to achieve higher resolution on wafers.

As dimensions shrink, pattern edges during lithography are more susceptible to factors such as optical near-field effects, diffraction, and photoresist absorption, resulting in line edge fluctuations and irregularities.

This is an inference based on physical knowledge. "

Lin Benjian applauded, "Wonderful deduction and keen insight.

When Zhengming told me that he had found a gifted student from Yanda, I was still wondering how talented he was that made him so proud.

Your talent in this area is even better than I guessed.

What you just said is two points away.

One point is that the wavelength of the ArF light source is more likely to be absorbed by the photoresist, which will lead to uneven exposure of the photoresist, which will affect the imaging quality. Therefore, it is necessary to design a special photoresist specifically for the ArF light source.

However, the photoresist is specially designed for the ArF light source to reduce absorption, and the new photoresist is likely to further cause line edge fluctuations and irregularities.

In addition, ArF is used to build an advanced process, and multiple exposure technology, phase shift photomask technology and polarized light technology will be used.

These techniques in turn increase line edge fluctuations and irregularities to some extent.

So there's more to the line edge fluctuations than you might guess. "

Zhou Xin asked: "But there are so many difficulties, why are you still optimistic about the ArF route, senior?"

How did Zhou Xin know that Lin Benjian was optimistic about the ArF technology route, because Lin Benjian said it himself when he participated in the interview at the Optics Conference this time.

"Is it a cover for the lithography machine company?"

Lin Benjian smiled: "I have resigned from IBM, and IBM has hardly continued to invest in the field of lithography machines in the past few years.

I don't work for Nikon or Canon, why cover for them.

Of course, I am more optimistic about the ArF technical route from my heart.

What we have just said are the shortcomings of ArF, and these shortcomings are only temporary.

There are also many difficulties in the process of replacing the g-line and i-line with the KrF light source. "

The g-line is a light source with a wavelength of 436nm, and the i-line is a light source with a wavelength of 365nm.

"Compared to g-line and i-line, KrF requires new photoresist and anti-reflection coating materials to adapt to the characteristics of KrF light source.

At the same time, KrF requires higher-performance optical systems and photomask materials. More advanced lenses and optics were required to achieve higher numerical aperture and resolution.

In addition, photomask materials also need to have lower scattering and absorption characteristics.

KrF also needs to optimize the exposure process to improve image quality. At that time, we mainly used double exposure and out-of-focus exposure techniques to reduce lithography errors.

Let me think about it, by the way, when we were developing the KrF light source, we had to consider the control, because KrF achieved higher resolution, so it required stricter process control requirements.

More precise control of parameters such as photoresist coating thickness, exposure dose, and development process is required to ensure photolithography imaging quality and yield.

We overcame so many difficulties, and finally the KrF light source replaced the g-line and i-line and became the most advanced process light source today.

Similarly, short-wavelength light sources will inevitably replace long-wavelength light sources in the future, which is inevitable for technological progress.

The difficulties we have just discussed are temporary difficulties. "

ASML, Nikon and Canon just postponed their investment in ArF, but did not stop their investment in ArF technical route.

Later, ASML realized corner overtaking, also because they chose the correct technical route. The ArF light source uses water as the medium, and a layer of water is applied on the photoresist.

The medium refractive index of water is 1.44, 193 nanometers ÷ 1.44 ≈ 134 nanometers, and the wavelength of the ArF light source is further reduced.

ASML was able to kill Nikon and Canon and monopolize the field of lithography machines. They made the right choice at two key nodes of the technical route.

When they should be stable, they are more stable than Nikon, thus gaining a firm foothold in the lithography machine market.

When they should be radical, they were more radical than Nikon, thus occupying the high-end market, eating up all the high profits, and maintaining their technological leadership in the follow-up.

After the breakthrough of the ArF technology route, the wavelength of ArF is 193nm, and he can manufacture chips of up to 65nm.

If you want to manufacture chips with a process below 40nm, you need to find a light source below 160nm.

Companies such as Nikon and Canon chose 157nm F2 lasers, but ASML found a light source as a medium and did it in one step.

Of course, it is not as simple as adding a layer of water.

"I have always believed that there is a solution to the ArF technical route.

First of all, it is necessary to design the corresponding photoresist specifically for the ArF light source, secondly, to develop anti-reflection coating, and then to improve the optical system of the lithography machine, and to find Zeiss to customize more advanced lenses.

To achieve higher numerical aperture and resolution.

Simultaneously optimize during exposure. I designed a suite of resolution enhancement techniques specifically for ArF light sources.

Includes assist features, off-axis illumination, double exposure techniques, etc. to improve pattern transfer quality and increase resolution. "Lin Benjian was very excited, it was rare to meet a colleague who knew so much about ArF technology.

Others who knew this thing signed a non-disclosure agreement, and Lin Benjian had a very unhappy chat with them.

What Huaguo is short of is not only the lithography machine, but the entire high-end manufacturing industry. Even if a lithography machine can be produced, the optical lens can still get stuck in your neck.

However, Zeiss is a German company and is not listed. American's long-arm jurisdiction cannot control Zeiss.

After listening to Zhou Xin, he felt that the other party is worthy of being a master, and he has a complete plan before joining TSMC.

The same thing, some people talk about it on paper, and some people win it thousands of miles away.

Huaguo's values ​​are result-oriented, and Zhou Xin knows Lin Benjian's future achievements, not to mention that what Lin Benjian said is absolutely dry.

After discussing with Lin Benjian about the future of the ArF technology route, Zhou Xin asked: "Lin Bo, I plan to set up a lithography machine company, and I need someone who really knows how to help me.

Are you willing to help me do things? The position I give you is subject to research and development.

The first batch of research and development funds is one billion US dollars, and I will give you 20% of the shares, but you need to sign a non-competition agreement after joining, and our lithography machine company will be located in Huaguo. "

After hearing the first half of the sentence, Lin Benjian wanted to agree immediately, one billion US dollars of research and development funds, and 20% of the shares.

What kind of trust is this, just with one billion dollars in research and development funds, this company is worth at least one billion dollars.

Just give you 200 million yuan.

After listening to the second half of the sentence, his heart sank: "Newman, I know you are from Huaguo, and you hope to help Huaguo develop in the semiconductor field.

I am also of Chinese descent. Under the premise of ability and opportunity, I also hope to contribute to the development and progress of China's semiconductor industry.

However, the research and development of lithography machines is very dependent on the cooperation with chip manufacturers.

There are no mature chip manufacturers in Huaguo. We put our R\u0026D base in Huaguo, so it is difficult to make progress.

Regardless of whether I agree to your conditions or not, I suggest you put it in Wanwan or American.

Wanwan has TSMC, and American has Intel's chip foundry.

In these two places, you have the opportunity to fully experiment with your product in a production environment. "

ASML has R\u0026D centers all over the world.

Zhou Xin said: "Zhang Rujing of Shida Semiconductor is preparing to return to China to establish a chip foundry.

Their site selection, funding and team are already in place.

In our preparatory stage and laboratory research and development stage, we do not need to cooperate with foundries during this period.

Our progress will be slower than Zhang Rujing's. When our laboratory products are developed, they can be sent to them for verification.

We only had the opportunity to enter Zhang Rujing's chip foundry when it was just starting.

Whether it is TSMC, Intel or Samsung, it is difficult for them to give a newly established lithography machine company a chance.

Introduce our laboratory products into their production process.

On the contrary, in my opinion, Huaguo is the opportunity. We need to use Huaguo's foundry to make a name for ourselves before we have the opportunity to cooperate with other chip foundries.

This is also how the countryside surrounds the city. "

Why Huaguo's lithography machine is a laboratory product, because it has not been fully verified in the production environment.

Therefore, ASML's embargo on Huaguo is a good thing in a sense. If there is no embargo, the lithography machine independently developed by Huaguo will never have the opportunity to enter the production process.

Lin Benjian said: "How many people are there in the lithography machine company you want to establish?"

Zhou Xin said: "You are the first, you are responsible for setting up the R\u0026D team, and you are the R\u0026D supervisor.

I will give you complete trust and autonomy.

I will only give my suggestions and thoughts on the direction of research and development. "

Lin Benjian was a little speechless after hearing this. To put it bluntly, there is only money but no one.

"I need time to think.

Newman Since you invited me, I don't mind being more direct.

For a lithography machine company, one billion US dollars may only be an investment in early research and development. If we can't break even earlier, we can't manufacture a lithography machine that can enter the production environment.

In the future, you will need to continuously invest in research and development expenses, or directly dissolve the company.

Therefore, 1 billion US dollars is just an admission ticket, and relying on the product to achieve breakeven means that the lithography machine company has survived.

Secondly, even if we set up a company, there are still many unavoidable patents that need to be purchased from companies such as Nikon, ASML, IBM, and Texas Instruments to purchase patent usage rights.

It is impossible for us to develop all technologies by ourselves.

Of course, if they want to use the technologies we develop in the future, they will also need to pay us patent royalties.

In addition, your salary is very high, and this is the first time I have received such a high salary offer.

From the perspective of treatment, I am definitely satisfied.

Linchuang, which I founded, has been around for almost eight years, and its valuation is not even 50 million US dollars.

It's just that I'm in charge of it all by myself, and I'm not sure I can do well.

In fact, I don’t have much experience in management. I have only managed a R\u0026D team of a few people, and managed the R\u0026D team of a lithography machine company. I have never done management work of this scale.

I am responsible for the team building of all R\u0026D personnel of a lithography machine company from scratch, and I have never done it.

I'm worried that I won't do well. "

Zhou Xin is not worried. After joining TSMC, Lin Benjian became the highest vice president of TSMC.

For a company of TSMC's size, the vice president has to manage many more people than the lithography machine company they want to establish.

Zhou Xin said: "I fully believe that you have this ability.

In fact, it is almost the same to manage a few people and dozens of people. You only need to manage the supervisor well.

In addition, I will help you find a good enough assistant, who will be responsible for the formation and management of the company with you. "

Lin Benjian asked curiously: "Who? Do I know him? Zhengming?

Do you persuade him to leave Berkeley and start a business with you?

Zhengming is better than me, and Silue Semiconductor is more valuable than Linchuang, and its valuation is much higher.

Zhengming probably doesn’t have the idea of ​​starting another company yet.”

Zhou Xin shook his head: "It's not Professor Hu, I don't know if you recognize him or not.

He is from Wanwan, and he used to do analog IC design in a semiconductor company in California.

Do you know Guan Jianying? "

Lin Benjian said: "I seem to have a little impression, but I am definitely not familiar with it. At most, I have seen it."

Yes, according to what Zhou Xin just learned, Lao Guan was kicked out.

This is to make up for yesterday, and today it is updated by 1w, yesterday's 4k plus today's 6k.

At first, I wanted to play badly in the middle, but I finally gritted my teeth and persisted.

Even on weekends, writing 10,000 words is still tiring.

Finally ask for a monthly pass!