Tech giants start by disassembling systems

Chapter 194: A Chronicle of the Industry's Heated Discussion on Weilan's MEMS Technology B

February 9, 2021.

推特。

David Pratt, an assistant professor at the University of Illinois at Urbana-Champaign, posted a long tweet.

Pratt is a 34-year-old researcher in the thermodynamics of microelectromechanical systems (MEMS) and has published over twenty papers in the field. He is an active voice in the academic Twitter community—not a big name, but with considerable influence.

His tweet was titled:

A Chinese startup claims its 300mm microelectromechanical systems (MEMS) achieve an accuracy of ±0.018 degrees. If true, this would be revolutionary. Detailed analysis follows.

The tweet was very long, consisting of twelve parts.

Article 1: I have been following the research of Vilan Microsystems since December. The company claims that, based on the third-order nonlinear extension theory of the thermoelastic coupling model, it has achieved a machining accuracy of ±0.18 degrees for a 300mm microelectromechanical system wafer. Their related paper is currently under review in the journal *Nature Materials*.

Article 2: Industry Reference Comparison: Bosch, a global leader, achieves an accuracy of approximately ±0.030 degrees Celsius in its 250mm wafer manufacturing process. If Vilan's measured data is accurate and valid, it not only improves accuracy by 40% but also achieves high-precision processing of larger wafers. This is not merely an unconventional technological iteration but a paradigm shift in the industry.

Articles three through eight provide a detailed analysis of the theoretical possibilities of the third-order model—Platter deduced a portion of the theoretical framework from publicly available conference abstracts and Villan's patent applications, and offered his own assessment.

Article 9: The core issue is whether this model can be adapted to 400mm wafer fabrication processes. It is understood that Vilan's paper predicts a theoretical accuracy of ±0.015 degrees for 400mm processes. If this prediction is verified, it means that ultra-high precision below 0.02 degrees can be achieved in industrial mass production, which will have a disruptive impact on the fields of autonomous driving sensors, medical microelectromechanical devices, and aerospace flight components.

Article 10: However, there is a critical shortcoming here—currently, no independent third-party institution has verified Vilan's experimental data. Among the 29 industry alliances it leads, none of the leading international microelectromechanical systems (MEMS) manufacturers are involved in endorsing it. Furthermore, the paper in the journal *Nature Materials* is still in the peer review process; before the peer review is completed, all technical achievements have not received authoritative certification.

Article 11: I specifically consulted two industry colleagues who are deeply involved in the field of thermal modeling of microelectromechanical systems. Their views were highly consistent: the mathematical logic of the theoretical model is valid, but the experimental results are too advanced and groundbreaking, and we must wait for the complete experimental data to be released to verify it.

Article 12: In conclusion, if Wei Lan's paper successfully passes the peer review process of *Nature Materials*, it could be one of the most significant technological breakthroughs in the field of microelectromechanical systems (MEMS) in the past decade. Conversely, groundbreaking technical conclusions must be supported by extremely solid empirical evidence; we await the outcome.

Within 48 hours of the tweet being posted, it received 347 retweets, 892 likes, and 156 replies.

This is a topic of considerable discussion within the academic Twitter community for microelectromechanical systems (MEMS).

The responses included supportive voices—"Finally, someone is doing a professional and rigorous technical breakdown, instead of blindly hyping up concepts"—as well as skeptical voices—"A Chinese startup with no industry experience claiming to surpass Bosch in technology? That's absurd."

But the most common voices are different versions of the same sentence:

"Awaiting the official publication of the paper."

"Awaiting the results of the peer review."

"Publicly available test data provides evidence."

……

Pratt's tweet was reposted to the Reddit microelectromechanical systems section, and then spread to the semiconductor and engineering technology sections.

On Reddit, the atmosphere of discussion is more direct—and more pointed—than on Twitter.

"Achieving ±0.018 degree accuracy with a 300mm process? The accuracy has doubled, surpassing Bosch. This is either a once-in-decades breakthrough in the microelectromechanical systems industry, or it's a complete fabrication; there's no middle ground."

"Being able to enter the peer review process of Nature Materials is a positive sign in itself. This journal will not waste space reviewing worthless research."

Does anyone know who the project leader is? It's Zhou Zhiyuan, but I can't find any information about his academic credentials. Is he a legitimate research professional?

"He is indeed a legitimate researcher. I found his information on a scientific research platform. His research results are solid. Although he is not the top in the industry, his qualifications are compliant. However, the first author of this groundbreaking result is a doctoral student named Su Chen. I had never heard of his research before."

"A groundbreaking paper that has the potential to revolutionize the industry and has been submitted to Nature Materials, with a doctoral student as the first author? Either they are exceptionally bold and possess hard-core achievements, or they are reckless and impulsive, engaging in empty theoretical discussions."

"Accurate data demonstrates courage; falsified data demonstrates recklessness."

……

Su Chen did not check Twitter or Reddit.

He was looking at the ninth set of experimental data sent by Zhang Li.

Nine sets of data. The S-shaped temperature gradient transition is present in every set. The deviation gradually decreased from an initial 4% to 5%, and is now stable at around 2.8%.

Zhang Li included a note in his latest experimental report:

"Group 9 (Vilan 400 Ultimate Type - 009): The first set of data after the second optimization of the temperature control scheme. The deviation in the first to second segment region is 1.6%, and the deviation in the second to third segment region is 2.8%. The S-shaped turning point is clear. Note: The deviation in the second to third segment region may have approached the temperature control accuracy limit of the existing equipment (±0.3 degrees Celsius). Further reduction of deviation requires the use of higher precision temperature control equipment. It is recommended that the current parameters be kept unchanged in the subsequent six sets of experiments to ensure data consistency."

After reviewing the report, Su Chen replied to Zhang Li: "Agreed. Keep the parameters unchanged going forward. Data consistency is the priority."

Then he opened a new document—he was organizing the data analysis framework for the 400mm edge verification experiment.

If all fifteen sets of data are finally available, he needs to do the following:

Statistical analysis: mean, standard deviation, and confidence interval of fifteen data sets.

Model comparison: The experimental data are quantitatively compared with the prediction curve of the third-order model, and the goodness of fit is calculated.

Feature analysis: the location and magnitude of the S-shaped turning point, and the deviation from the model prediction.

Error source analysis: Defining and separating systematic errors (temperature control accuracy) from random errors (environmental fluctuations).

Conclusion: Whether the data from the edge region supports the 400 mm extrapolation prediction conclusion of the third-order model.

Su Chen had been thinking about this framework for a long time—since the very first day of the experiment. But he hadn't started writing it down yet because he didn't have enough data.

The nine sets of data have now accumulated sufficient information. The S-shaped turn is real. The model's predicted direction is correct.

The remaining six sets of data are to make the statistical conclusions more robust.

Su Chen wrote the first line in the document:

"400mm Edge Validation Experiment - Data Analysis Report (Draft)"

Then he wrote a passage below—a passage not for anyone else to see, but only for himself:

"This data is not just a bargaining chip for dealing with reviewers. It is the first experimental signal of the third-order model on a 400-millimeter scale. The S-shaped temperature gradient transition is a direct manifestation of the third-order correction term—a feature that does not exist in any existing second-order models. If all fifteen sets of data confirm the S-shaped transition, this in itself is an independent scientific discovery."

Defense is the best form of offense.

Guard: Prepare supplementary supporting materials for the reviewers.

The objective is to gather empirical data for the second core paper.

He looked at the four characters "defense is the best offense" and nodded.

Then continue to refine the data analysis framework.

……

February 11th.

The IEEE International Conference on Microelectromechanical Systems 2021 will be held online in March – marking the second consecutive year that the conference has been moved online due to the pandemic.

During the pre-conference discussions, a panel discussion titled "Frontier Research Directions in Thermal Modeling of Microelectromechanical Systems" unexpectedly attracted attention from the industry.

The organizer of this special session is Professor Heinrich Vogt of the Technical University of Munich—one of the leading scholars in the field of microelectromechanical systems thermodynamics. Vogt is 62 years old this year and is about to retire, but he continues to delve into the forefront of academia and maintain a high frequency of research output.

In the preview email for the special feature, Vogt wrote:

"This year's symposium will focus on the latest research progress in nonlinear thermoelastic coupling models. Among them, the research results of the third-order extended classical theoretical framework recently proposed by Weilan Microsystems in Suzhou, China, are the core focus of this symposium. Although the results have not yet been formally published, they have extremely high potential for industry innovation and are worthy of in-depth discussion by the academic community."

This statement caused quite a stir in the academic community of microelectromechanical systems.

Vogt is a heavyweight authority in the industry. His initiative in mentioning Wei Lan's unpublished work in a symposium at the IEEE Microelectromechanical Systems (MEMS) top conference is itself a high level of professional recognition.

The news reached the Suzhou Weilan team.

Zhou Zhiyuan forwarded the preview email to Su Chen on WeChat, adding the sentence: "Vogt mentioned our research."

After reading the email, Su Chen replied, "How did he find out about our submission results?"

"Nature Materials' peer review process is strictly confidential, but the information about the paper submission is not confidential. We have previously published research abstracts at academic conferences, and related patent applications are also publicly available. For an industry authority of Vogt's caliber, understanding the latest research developments is a piece of cake."

Is this a good thing or a bad thing?

"That's good." Zhou Zhiyuan replied very briefly. "Fogg wouldn't endorse worthless research. The fact that he mentioned our results is enough to prove that he recognizes the academic value and industry potential of this research direction."

Su Chen flipped through the calendar. The IEEE Microelectromechanical Systems 2021 Online Conference was scheduled for mid-March. At that time, the team's paper would most likely still be in the peer review stage, or it might have already received the first round of peer review comments.

Regardless of the outcome, Vogt's symposium will draw the attention of more scholars in the field of microelectromechanical systems (MEMS) worldwide to this third-order theoretical model.

This is a double-edged sword.

High visibility means that once a paper is successfully published, it can quickly gain academic recognition and industry resources.

High attention also means that if a paper fails the peer review, it will face questioning and pressure from the entire academic community online.

Su Chen didn't dwell on the matter too much and replied to Zhou Zhiyuan, "Understood. We just need to focus on doing our research well and gathering solid data."

He then continued to check the tenth set of experimental data sent by Zhang Li.

……

February 13th.

domestic.

The technical controversy surrounding Vilan has spread from discussions within the niche academic circle on Zhihu to the mainstream industry.

The catalyst for this spread was an in-depth industry article published in Semiconductor Industry Observer.

Semiconductor Industry Observer is a highly influential and authoritative professional media outlet in China's semiconductor field. This article is authored by Professor He Wentao of the School of Microelectronics at Fudan University—a renowned scholar in the field of microelectromechanical systems (MEMS) in China, who has published over eighty SCI papers and serves on the program committees of several international academic conferences.

The article title is:

From the Vilan Microsystems Third-Order Model, we can see the paradigm shift in the microelectromechanical systems industry.

The full text is about 5,000 words and is divided into four main sections.

The first part provides an in-depth technical analysis of the third-order model. Based on publicly available research information, He Wentao reconstructs the core theoretical framework of the third-order model and provides a professional assessment:

"From a theoretical perspective, Wei Lan's proposed third-order nonlinear extension theory is mathematically and logically consistent and has a complete system. The traditional second-order thermoelastic coupling correction model has good adaptability in 150-200 mm wafer processes, but its accuracy is a significant weakness in large-size wafer processing, which is a recognized industry pain point in the entire microelectromechanical systems academic community. Wei Lan's third-order optimization approach precisely addresses the core challenges in the industry, and its research direction is entirely correct."

If their claimed ±0.018 degree measured accuracy is indeed valid—I emphasize the word "if"—this is not merely a simple process upgrade, but a fundamental theoretical paradigm shift in the field of microelectromechanical systems (MEMS). This means that the entire process system and production standards built upon second-order models over the past two decades will need to be thoroughly reviewed and iteratively optimized.

Part Two provides an in-depth analysis of the industry value of ±0.018 degree precision. He Wentao dedicates an entire page to explaining the core significance of this data in simple terms:

What exactly does a machining accuracy of ±0.018 degrees mean? Let me explain with a simple analogy: If Bosch's current accuracy of ±0.030 degrees is a "qualified and usable industrial ruler," then ±0.018 degrees is an "ultra-precise scientific research-grade measuring tool." The difference between the two is far more than a simple 40% numerical difference. In industrial applications, this breakthrough in accuracy will significantly improve the operational reliability of autonomous driving sensors, the machining precision of medical micro-devices, and the operational stability of aerospace attitude control systems.

If this level of precision can be achieved in stable mass production, Vilan will become the industry benchmark in the field of precision for global microelectromechanical systems.

More importantly, the theoretical prediction of ±0.015 degrees for the 400mm process proposed in their paper, once experimentally verified, will completely reshape the development landscape of the microelectromechanical systems (MEMS) industry and usher in a new era of high-precision industrialization.

Whether you're an industry professional or a general reader, you'll be amazed by the groundbreaking value of this data when you get to this point.

However, He Wentao's article did not simply praise the work; the third part accurately poured cold water on the criticism and rationally analyzed the risks.

"But we must remain absolutely rational and view this technological breakthrough with caution."

First, Vilan's paper is still in the peer review stage of *Nature Materials* and has not yet passed authoritative peer review or been formally published. Before formal publication, all technical data and research conclusions are merely the company's "self-claims," ​​not industry-recognized established facts. *Nature Materials* has a rejection rate of over 90%, so successful submission does not guarantee publication.

Secondly, the process evolution from 300mm to 400mm wafers is not a simple linear upgrade. As wafer size increases, the thermal complexity increases non-linearly—a core principle summarized from decades of industrialization practice in the microelectromechanical systems (MEMS) industry. Even if a third-order model is adapted and accurate in 300mm processes, it cannot be directly derived to be suitable for 400mm ultra-large wafer processes. There is a significant gap between theoretical predictions and industrial measurements.

The history of the microelectromechanical systems (MEMS) industry is replete with similar cautionary tales. In 2008, a Japanese team proposed a novel nonlinear correction algorithm, achieving extremely accurate data in a laboratory setting, but ultimately failed in industrialization verification—the ideal laboratory conditions differed significantly from the complex mass production environment of a factory. In 2012, an American team made a breakthrough in the thermal theory of MEMS, publishing their findings in *Physical Review Letters*, but to date, there have been no practical applications—the process conditions required by their theoretical model cannot be met by existing mass production lines.

The gap between theoretical breakthroughs and industrial application is the harshest reality in the microelectromechanical systems (MEMS) industry.

Third, although the industry alliance led by Vilan has 29 members, none of the leading international microelectromechanical systems (MEMS) manufacturers are involved. The lack of independent testing and verification from top-tier manufacturers is the biggest weakness in the credibility of the current third-order model.

I don't deny that Vilan's research has disruptive potential. However, before the paper passes peer review and receives independent third-party verification data, the industry should maintain cautious optimism; the key word is "cautious."

"If Wei Lan's third-order theoretical model is fully verified, it will become the most core and significant theoretical breakthrough in the field of microelectromechanical systems in the past decade. However, the word 'if' represents a great deal of uncertainty."

My advice to the Weilan team is: avoid being overly aggressive. Prioritize solidifying the stability of the 300mm process, ensuring smooth publication of the paper, and conducting thorough independent third-party verification. Breakthroughs in the 400mm process can be achieved gradually and steadily.

Ultimately, for a Chinese startup less than three years old to claim that its technology has completely surpassed that of industry giant Bosch requires extremely comprehensive and irrefutable empirical evidence. At present, however, the core supporting materials are insufficient.

Everything awaits the final draft of the paper.