fictionpage Chapter 212 The Last Straw
Chapter 212 The Last Straw
Chapter 246 The Last Straw
May 23rd.
Eight o'clock in the morning.
Beijing.
Su Chen saw a PDF forwarded by Zhou Zhiyuan at his workstation.
This is Bertoli's first editorial in the European Industrial Semiconductor Review: "The Distance Between Mathematical Models and Industrial Robustness".
He opened it and watched it for a few minutes.
The article is written more shrewdly than He Wentao described. Bertoli did not directly attack the third-order model itself. He attacked the gap between the "laboratory-level mathematical model" and the "real production line environment."
He listed three specific industry issues:
First, there is the impact of regional water quality differences on the etching process. The composition of trace impurities in pure water varies in Asia, Europe, and North America, and these impurities can have a small but cumulative effect on MEMS etching.
Secondly, the impact of summer and winter temperature differences on equipment calibration. Even in a clean room with constant temperature, temperature differences before the equipment enters the constant temperature workshop can cause calibration drift.
Third, there are differences in the micro-stress distribution between different wafer batches. Even wafers of the same specification from the same supplier may have significantly different micro-stress distributions between different batches.
Bertoli concluded: "While the mathematical models perform well in the laboratory and small-scale pilot production stages, whether they can maintain the same accuracy in large-scale industrial production is an open question. It may be premature to promote such models as industrial infrastructure before this question is rigorously verified."
Su Chen did not react immediately after reading it.
He opened his work log and created a new document.
Title: Solutions to Robustness Concerns
The document states only one line: "These issues were considered during the simulation system design phase."
Then he listed three points:
I. Regional Water Quality Differences – The simulation system already includes a “regional chemical environment correction layer.” The required input parameters are the conductivity, pH value, and content of major trace elements in local pure water. These parameters are already routinely monitored in most MEMS factories.
II. Summer-Winter Temperature Difference – The simulation system already includes “Equipment Thermal History Correction”, and the required input is the temperature profile of the equipment over the past 72 hours. This is a standard log item.
III. Wafer Batch Variation – The simulation system already includes a “batch micro-stress correction model,” and the required input is the stress test data for each batch. This is a standard QC indicator.
Su Chen sent these three messages to Lin Wei.
He didn't say, "This is a response to Bertoli's proposal." He simply sent it out.
After receiving the message, Lin Wei replied, "Put these three points in the second annex of the Bosch cooperation framework. Have them include this in the cooperation announcement."
Su Chen replied with an "okay".
He closed the Bertolli editorial.
He didn't write any rebuttal. He didn't issue any statements to journalists. He simply listed the capabilities of the simulation system and then included that list as an annex to the Bosch collaboration framework.
He is responding to an industry-level question in an industry-level manner.
Instead of using a debate.
……
May 25th.
3 PM.
Munich.
He Wentao published a follow-up article in Semiconductor Industry Observer.
Title: [Bertoli's "Robustness Question": Why this is Italy and France's last card].
The article is structured very calmly. It first reiterates Bertolli's core points from two industrial magazines, and then provides an objective timeline:
Bertoli's strategy over the past six months has gone through four phases:
Phase One: The Deceleration Battle. Starting in the second half of last year, STMicroelectronics spearheaded a series of academic challenges to the third-order model. This phase essentially ended after the NM acceptance letter was issued.
Phase Two: Supply Chain Pressure. From March to April of this year, STMicroelectronics exerted pressure on some of Vilan's upstream suppliers through contract leverage. This phase essentially ended after Vilan initiated the acquisition of Qirui Precision Test's equity and the commissioning of its packaging line.
Phase Three: The Alliance Concept. On May 19th, Bertoli proposed the concept of a "Third-Order Model Industrialization Alliance" to Stein, attempting to reshape the agenda. This phase substantially concluded after a breakthrough in direct communication between Bosch and Villeroy & Boch.
Phase Four: Robust Questioning. This is the current phase, and it's the only strategy Italy and France are currently using.
In his article, He Wentao did not determine whether Bertolli's robustness argument was right or wrong. He simply clearly marked the position of this argument—it was the "last strategic option."
Within two hours of the article being published, it had been forwarded over 10,000 times.
A user on Zhihu commented: "I've worked in Italy and France for eleven years. To be honest, Berlusconi is constantly retreating, but each time he retreats to a smaller, but harder-to-refute, position. This robust challenge is indeed a 'small position'—but it is also indeed difficult to refute quickly."
This comment received 1,800 likes.
Lin Wei saw this comment in her office.
She took a screenshot and forwarded it to Su Chen.
He added, "Bertoli is no novice. He knows what he's doing."
Su Chen replied, "But he also knows he's lost. He's just giving himself time to gracefully exit the game."
After reading it, Lin Wei remained silent for a while.
She recalled her first encounter with Bertoli. It was last September, on the sidelines of a European MEMS standards conference. Bertoli was fifty-eight years old then, two years away from retirement. He appeared composed at the conference, discussing the future direction of next-generation MEMS standards.
He once told Lin Wei, "Madame Lin, in this industry, the final question is never about technology. It's about who writes the future textbooks."
That sentence stuck with her for a long time.
Now she understood Bertolli's feelings when he said those words. He knew his name wouldn't be in the future textbooks. He was trying his best to delay the finalization of the textbooks.
However, the textbook has already been finalized.
The Su-Zhou model—this name began appearing in English-speaking MEMS discussions. It's not a research paper; it's a term accepted throughout the industry.
Lin Wei took out a pen from the table.
She wrote down a time on a piece of white paper.
2021 5 Month 25 Day.
She wrote the following sentence:
"Bertoli moves to the fourth stage. Robustness is questionable."
Then he wrote another sentence:
"This is the last straw."
She put the paper at the very back of the drawer, along with other time points accumulated over the past six months.
These papers will one day be compiled into an industrial history book.
but not now.
There are ten days left.
……
May 26th.
Nine o'clock in the evening.
Beijing.
Su Chen finished the last section of his second paper at his workstation.
The title of the second paper is "Multi-platform Validation and Material Extension Protocol for the Third-Order Nonlinear MEMS Model".
The full text is eighty-two pages long, shorter than the first article, but its content is more industry-oriented.
The second paper answered three questions:
1. How reproducible is the third-order model on different experimental platforms? Answer: The deviation is within five percentage points on all five validated platforms.
How are second- and third-order models applied to new material systems? Answer: Through the Materials Extension Protocol (in collaboration with Kosuke Yamamoto). The current version covers three material families.
3. How reliable is the third-order model in engineering, as the core engine of the production line design simulation system? Answer: In Bosch's double-blind verification test, the median deviation between the prediction and the actual production line model was 0.3 percentage points.
The third answer was added by Su Chen in the last section. He originally only planned to write the first and second answers. But Bosch's 0.3% deviation result made him decide to add this section.
This is not about showing off. It's about setting a benchmark for future researchers.
Su Chen saved the entire text and sent it to Zhou Zhiyuan.
Twenty minutes later, Zhou Zhiyuan replied with a message:
"Old Su, should we keep that 0.3 figure in the third answer? Once it's published, our competitors all over the world will know what we did to Bosch."
Su Chen thought for a moment.
"Keep it. We don't need to hide it."
Why?
"Because Bosch's collaboration announcement will be released next week. At that time, the entire industry will know about this number. We're writing it in the paper first so that our competitors can start digesting it before the official announcement."
Zhou Zhiyuan remained silent for a while.
"Okay. I agree."
Su Chen uploaded the full text to the arXiv pre-release queue.
According to the plan, the preprint will be available online simultaneously with the official release of NM on June 1st.
On that day, MEMS researchers around the world will read this paper.
Including Bertoli.
……
May 28th.
morning.
Shanghai.
Lin Wei received a new email from Bosch.
Attached to the email was a document—meeting minutes of Bosch's internal technical committee.
The minutes were drafted by Graff. There was a passage in them that Lin Wei read three times:
"Based on the verification result delivered by Vialan on May 22, with a median deviation of 0.3%, the Internal Technical Committee unanimously recommends accelerating the framework discussion. Furthermore, the Committee proposes that Bosch's 400mm pilot line be considered as the first joint verification platform for the simulation system. This proposal would be subject to commercial terms agreed in the framework discussion."
400mm pilot production line. Serving as a joint verification platform.
This suggestion came from Mayer, Bosch's chief scientist. His influence within Bosch was second only to Stein.
Lin Wei read this passage three times.
This means that Bosch is not only prepared to accept technical cooperation, but also to allow Vilan's simulation system to be directly integrated into Bosch's 400mm pilot production line. This is a proposal that goes far beyond conventional licensing cooperation.
She picked up her phone and sent a message to Su Chen.
"Bosch has made an internal decision. They have proposed using the 400mm pilot production line as a joint verification platform."
Su Chen replied three minutes later:
"Meyer's."
Lin Wei was taken aback. "How did you know?"
"He noted down three terms at the last meeting: infrastructure, mathematical layer, and MEMS-specific. These three terms don't describe a product; they describe a platform. He's the only one who could think of using a simulation system as a platform."
After reading it, Lin Wei put her phone down.
She sat in the chair and leaned back for a while.
This project started in November and seven months have passed since then.
Seven months ago, the third-order model was just a handwritten draft.
It is now about to enter Bosch's 400mm pilot production line as the "core engine of the joint verification platform".
This is an industry position that far exceeds that of licensing and cooperation.
Lin Wei picked up the pen on the table and wrote a line on a piece of white paper:
"May 28th. Bosch provided a 400mm joint verification platform."
Then she wrote another line:
"Bertoli's robustness questioning is now irrelevant."
She put the paper in the drawer.
Outside the window, the afternoon sun was shining brightly in late May.
Pedestrians hurried along the street below the Villan Building. No one knew what had happened inside that building five minutes earlier.
But ten years later, all MEMS design textbooks will mention this day.
Lin Wei closed her laptop. She walked to the window.
She stood by the window for a while.
Then she turned around, went back to the table, and picked up the phone.
"Please have Jiang Mingyuan come to my office. We need to readjust our pricing for the cooperation framework."
If Bosch proactively proposes to use the 400mm production line as a joint verification platform, then the original starting price of 28 million euros for licensing fees would be too low.
It should now start at 40 million euros.
Even higher.
She picked up the draft cooperation framework from the table.
They began to recalculate the accounts line by line.