EUV Lithography's Materialization Gap | Generated by AI
Question: Is a photolithography machine really that insurmountable, and can’t we master it just by studying all publicly available chip data and electronic circuitry?
Answer:
You’ve touched on a very fascinating debate in the tech world. Liang Wenfeng’s perspective highlights a fundamental truth about software and general engineering: with open-source communities and enough time, most logical and architectural “wheels” can indeed be reinvented.
However, the reason a photolithography machine (specifically EUV—Extreme Ultraviolet) is considered the ultimate gatekeeper of modern tech isn’t a failure of intelligence, learning, or data access. It is an apex challenge of extreme physics, global supply chains, and precision manufacturing. Here is a breakdown of why studying all the available literature still leaves a massive execution gap, and how companies like Huawei are adapting.
1. The “Information” vs. “Materialization” Gap
Knowing how something works conceptually is entirely different from being able to manufacture it. Even if you have the complete blueprints of an ASML EUV machine, you cannot build it by studying electronic circuits alone.
- The Mirror Challenge: EUV light is absorbed by almost everything, including air and regular glass. Therefore, the machine uses specialized mirrors to bounce the light. These mirrors must be so perfectly smooth that if you scaled a single mirror to the size of the Earth, the largest bump on it would be less than a millimeter high.
- The Light Source: To create EUV light, a tiny droplet of molten tin is dropped in a vacuum and blasted by a high-power laser twice—50,000 times per second. This creates a plasma that emits the correct wavelength.
This is no longer just an electronic circuit problem; it is a materials science and atomic-scale engineering problem.
2. The Global Supply Chain Monopoly
No single country or company builds a lithography machine alone. An ASML EUV machine is the culmination of Western industrial history, containing over 100,000 highly specialized components:
- The Optics: Developed exclusively by Zeiss (Germany).
- The Lasers: Developed by Trumpf (Germany).
- The Light Source: Developed by Cymer (USA).
Even if a company learns the theories perfectly, it cannot easily duplicate the hundreds of hidden, proprietary manufacturing processes (know-how) that these individual suppliers have perfected over 50 years.
3. De-Verticalization: Huawei’s Smart Adaptations
Because building an identical “wheel” (like an ASML EUV machine) from scratch takes immense time, Huawei has masterfully shifted the battlefield. This aligns perfectly with your observation about their learning agility and global research footprint:
- 3D Stacking and Advanced Packaging: If you cannot make the transistors smaller using advanced EUV lithography, you can stack older, more mature chips vertically (3D folding/stacking) or connect them via advanced packaging (Chiplets). This achieves the performance of a 3nm or 5nm chip using 7nm or 14nm equipment.
- The Turkish Professor Connection (Erdal Arikan): Your mention of the Turkish professor refers to Polar Codes for 5G. Huawei recognized the academic genius of Arikan’s theoretical paper on Polar Codes when the rest of the world ignored it. They poured massive resources into turning that mathematical theory into a global 5G standard.
Summary
Liang Wenfeng is right in spirit: given infinite time and resources, humanity can solve any engineering problem. It is “just a matter of time.”
However, the photolithography machine is “牛叉” (incredible) because it sits at the absolute limit of what physical matter allows us to do on Earth. Huawei’s current success doesn’t come from instantly cloning ASML; it comes from their brilliant ability to look at the global landscape, find alternative mathematical/structural paths (like 3D stacking and Polar Codes), and utilize their worldwide labs to bypass the bottleneck entirely.