Wanwan is not mainland China, it has not been blocked by technology from Europe and the United States. There is no generation difference between TSMC's chip manufacturing equipment and Intel's chip manufacturing equipment.

The reason why the process is a whole generation worse is because the latest copper interconnection process has not yet made breakthroughs. Of course, there are other reasons. The entire chip process has thousands of processes, and each chip manufacturer has its own specialty technology. In this regard, Intel is definitely more than better than the new TSMC.

Before the copper interconnection process, everyone had an aluminum interconnection process, but when the aluminum interconnection process reached the 0.5 micron process, it basically reached a technical bottleneck and could never be broken through again.

It is very complicated to explain this matter clearly. In short, the properties of the aluminum material itself determine the bandwidth and clock frequency of the aluminum wire.

Everyone knows that the resistance of aluminum is 2.8μΩ.cm, while the resistance of copper is 1.7μΩ.cm, which makes the copper conductors thinner and narrower, and correspondingly increases the capacity of transistors per unit area.

This is a technical background for the 0.5 micron process process to 0.35 micron process process.

The copper interconnection process was invented by ibm, but they are now only able to achieve laboratory verification and are not really practical.

Because of this, ibm has been selling their immature copper interconnection technology to other wafer companies in 1994.

It was also sold to TSMC, but at that time, TSMC may have been due to shortage of funds and did not pay for it. Instead, it was Taiwan UMC, which is also a wafer foundry company, invested in it.

In the 1990s, TSMC's process technology was one level lower than Intel because of its lack of technical strength.

This made them only able to produce some low-end chips with very meager profits. In short, TSMC was still struggling at this time and had not yet soared to the sky.

The principle of copper interconnection technology is very simple. It seems that it is as simple as replacing aluminum with copper, but it is not that easy.

The biggest problem is that copper does not stick to transistors, which is this simple problem that stumps all wafer companies.

You should know that the inside of the chip is very, very precise. Inside that we cannot see in our human eyes, if you use a thousand times electron microscope to view it, the transistors and metal circuits inside are interlaced and densely packed.

Especially metal circuits are not as simple as one layer, but as many as two or three layers.

In such a precise and narrow space, it is really difficult to solve the problem of the copper wire not sticking to the transistor.

Without the experience of previous generations, scientists can only experiment with various bonding materials one by one by casting a wide net.

Liu Meijuan is a time traveler and an insider in the industry in integrated circuits. Of course, she knows very well. TSMC will also solve this problem in 199.

In fact, it is very simple. It is necessary to first coat a thin layer of seed crystal material, that is, single crystal copper. Single crystal copper has good adhesion to sio2.

Of course, things are not that simple. This also involves technical issues between the dielectric layer and the barrier layer. It is necessary to ensure the adhesion of the copper wire, and to prevent copper ions from diffusing into the transistor, affecting the semiconductor performance of the transistor.

This technology will not make a breakthrough until August 1997, and Liu Meijuan was wondering whether to tell TSMC the key to the copper interconnection process in advance.

Logically speaking, it is nothing to tell them about the technology that others will break through in three years. In this way, it will solve the manufacturing process problem of Dolphin Technology's 10 million transistor-level chip, which can be put into production quickly, which is beneficial to both parties.

But Liu Meijuan just couldn't get through this hurdle in her heart. Why should she help TSMC for free?

If TSMC is a domestic company or a patriotic company, then that's nothing. If you help, it's just helping. Liu Meijuan won't seek any compensation.

But the problem is not!

Of course, Liu Meijuan was unhappy to help, so she told Wang Yong exactly what she thought.

Before he got out, Liu Meijuan ordered Wang Yong: "Honey, you ask the business department to purchase some single crystal copper wire and let Hu Weiwu's team do the experiment. Just do it my method, just verify the feasibility, and apply for international and domestic patents for me.

This material was invented by Atsumi Ono from Nino University of Technology, and it may only be available if Ninona is a dumb.

We took the patent and then asked TSMC to manufacture it. This new copper interconnection process is only allowed to be used on our chips.

If TSMC wants to use it, you can only purchase a patent license from us.

There is no way. Currently, only Wanwan has two professional chip foundry companies. If I do, based on my temper, I don’t plan to go to Wanwan.”

It’s not that Liu Meijuan is so stingy. Until 2018, Wanwan was still clamoring for independence. What’s even more annoying is that there are far more independent people than unificationists. It can be seen that these Wanwan people are already determined not to return to the embrace of their motherland.

Liu Meijuan has no good feelings for people in Wanwan, so she simply applied for a patent herself, regardless of what others think. How did a company without chip manufacturing technology invent the copper interconnection process?

In fact, with excellent materials such as single crystal copper, it is far more than just the copper interconnection process. It can also be used as copper bonding wires.

Today's chips are still using gold wire bonding. Not only is it expensive, but more importantly, because the gold wire is too soft, its stretch length and fineness have certain limitations, which are far inferior to single crystal copper bonding wire.

There is also the current chip process technology that is relatively low. One of the reasons is the gold wire bonding process. It can only fit two to three layers at most. In the future, with the improvement of process technology, the fitting will soar to eight and ten layers. This is not a single crystal copper bonding wire.

You may not know that the chip has thousands of pins, all of which are bonded with gold wire, and the length reaches an astonishing 1.4 kilometers, which is a huge expense.

If you use single crystal copper wire, you can save a lot of gold.

Now everyone knows why some people used to collect used mobile phones! They used it to refine the gold in the chip.

The preparation process of single crystal copper wire is actually not particularly complicated, it uses continuous hot casting and stretching molding.

Later domestic enterprises have been able to manufacture them. Liu Meijuan has read relevant process and technical information, and she knows how to do it.

But now Dolphin Technology does not have a decent laboratory, and she can't do it even without her hands.

Another raw material for this single crystal copper must be high-purity copper, which is currently not available in China.

Everything is still early, Dolphin Technology does not have its own wafer fab, so it doesn't make sense to build this kind of material.

In the afternoon, Wang Yong went to work and arranged what Liu Meijuan ordered, and he devoted himself to reviewing the documents.

Since it is a technology that ibm has verified, Liu Meijuan has made it clear again, and of course there is no problem with the experiment.

Then Wang Yong instructed the Legal Affairs Department to step up the application for patents. This is time-consuming and the earlier the better. It must not delay the time to go to Wanwan.

In this way, Dolphin Technology inadvertently obtained another key patent.
Chapter completed!