"Japan's rubber, China's soil, reverse engineering of EUV"

In recent days, there has been considerable discussion about China's Manhattan Project — the reverse engineering of ASML's lithography machines. Everyone is quite excited, as it is a significant breakthrough.

But if we look at it from an engineering and technical perspective, setting aside ideology and examining it closely, we will find that there is a gap between reality and ideals, and it's not small.

Because if we break the problem down to the engineering level, we will find that what determines whether a technical route can run for a long time is often not the hottest media topics, but some seemingly inconspicuous details.

This article does not discuss the progress of reverse photolithography machines or other details, but rather focuses on a recent less prominent news item: Japan's limited export ban on photoresists to China.

In advanced processes, the photoresist determines how wide the process window is, whether random defects can be suppressed, and whether the yield tail will go out of control.
In other words, there is EUV, but there is no photoresist, and it is impossible to improve the yield of advanced processes. Without yield, costs cannot decrease; if costs do not decrease, there will be no orders, at least no international orders.
This is also why photoresist is often more covert than the equipment, but more lethal in engineering.

Not many amateur scientists understand the difficulties of Japanese photoresist, and they habitually think of it as just a consumable in the photolithography process. But the reality is much more complex.
High-end photoresists from Japan are very difficult to replicate, as this involves not only certain chemical structures but also depends more importantly on the entire production and control process, which is linear and cannot be quickly iterated.
From ultra-pure raw material control, selection of polymerization reaction pathways, management of molecular weight distribution, to impurity statistics, batch consistency, and long-term aging behavior, this is a highly engineered and long-evolving system.

It is built up slowly from decades of failed samples.
Accumulated a large number of patents, but more importantly, many key judgments cannot be written into papers and are also difficult to fully incorporate into patents.
They exist in the engineer's intuition about whether "this batch of materials can go on the production line," in the production line's experience-based judgments of abnormalities, in the decades of process parameters and failure data of the enterprise, and in the gradual improvements to processes and controls over time.

This is the true meaning of "Japan's glue."
It's not a bottle of product, but a complete set of long-term operational material industrial capabilities.

China has a very interesting but often overlooked reference in this regard:
The refining and processing capacity of rare earth elements.

Among these, what is truly difficult to replicate is not the resources themselves, but the technological system that separates, purifies, and stabilizes complex minerals to a state suitable for engineering use.
This is a process of trial and error that has gone through countless failures, consuming a large amount of resources and generating a lot of pollution.

The West does have rare earth resources, but the real challenge is to turn the "earth" into industrial materials that can be supplied on a large scale, in a controllable manner, and for the long term.
That is also a set of highly engineered capabilities accumulated over a long period.
This is also why China's "rare earths" can be used to choke the necks of Europe and America.

Interestingly, one of the main suppliers of photoresists in Japan, Shin-Etsu Chemical, is also one of the few manufacturers in Europe, America, and Japan with the capability to refine heavy rare earths. (We will elaborate on why Shin-Etsu Chemical can produce both photoresists and rare earths in a separate article next time.)

Similarly, Japan's photoresist can also choke China's neck.
The supply disruption of photoresist in Japan affects stability and yield.
The photoresist from Japan is an absolute monopoly in the mass production of EUV lithography machines. In other words, even with EUV technology, without Japanese photoresist, it is still impossible to produce chips at 5nm and below.

Even in the current 7nm production process in China, although EUV is not used and instead 193nm ArF DUV with multiple exposures is utilized, the yield is still limited by Japan's high-end photoresists:
Non-critical layers, domestic rubber can now be used stably.
The secondary key layer can be mixed with domestically produced and imported products.
The key layer that truly determines success or failure still relies heavily on Japan's high-end ArF photoresist.
Because multiple exposures magnify any slight instability.

Once the supply of high-end photoresists is cut off, the already low yield of 7nm will further decline, and costs will rise further.

Why is EUV more dependent on Japanese photoresists? Because only mature photoresists can suppress photon statistical noise and random defects.
The 7nm key layer has truly achieved controllable domestic substitution, but it may still require multiple R&D cycles, not to mention the EUV photoresist needed for 5nm.

One of the development cycles generally takes 3-5 years, as it must include at least 5 stages that are almost impossible to parallel and are very time-consuming:
Basic Formula Exploration
Laboratory → Pilot Scale
Device integration (Scanner + Track)
Production Line Verification (Wafer Level)
Long-term stability verification

The parts of the device may be explicit, but the materials and processes are implicit; machines can be disassembled and replicated, but the time required for materials and industry cannot be compressed through reverse engineering.

Photoresists from Japan, like China's heavy rare earths, are the kind of materials whose true value can only be realized on the production line, at the tail end of the yield curve, during stable operations without incidents for several consecutive years.

If reverse EUV can perhaps "open the door", then the photoresist determines whether one can walk this path for a long time.

And the most brutal thing about semiconductors is —
One success means nothing; only a continuous period of several years without issues can be considered a success.

And time, even in the era of AI, is the only thing that cannot be reverse-engineered.