I started with a ship full of max-level veterans, and I'm just a complete noob.

The next morning, when Wang Shouwu entered the 502 research room, he was carrying a tin file box in his left hand and a sealed ampoule in his right. The silicon seed crystal inside the ampoule was clearly visible under the light.

The ampoule label indicates [111] orientation, purity of 99.9% or higher and chemical polishing treatment, with silicon seed crystals approximately 15 mm in length and very clean surfaces.

When Lao Sun took the ampoule, he handled it gently, first checking the seal and then the crystal end face to confirm that it had not been damaged during transportation before placing it in the cork stand inside the safe.

Wang Shouwu placed the metal file box on the table in the cubicle, but first pulled out a sheet of silicon melt properties.

"The seed crystal can be used directly for the first batch of tests. The rotation parameters need to be calculated first. In the past, when pulling silicon, I preferred slow rotation, but experience can only define a range."

Jiang Ming unfolded his kraft paper notebook and substituted the viscosity of the silicon melt, the impurity diffusion coefficient, and the density into the boundary layer relationship of the rotating disk one by one. Then he estimated the rotating Reynolds number based on the crucible diameter.

Silicon melt has a lower viscosity than germanium and a smaller impurity diffusion coefficient. Changes in the Schmitt number will recompress the concentration boundary layer, and the original safe point of 28 revolutions cannot be directly applied.

Wang Shouwu sat opposite him, using a slide rule to check the order of magnitude of each item. He saw the rotation speed window gradually shrink to twenty to twenty-five revolutions, and he drew a horizontal line next to his record.

"In the few batches I've worked on, the interface tends to become messy after 25 revolutions, while below 18 revolutions it causes the tail section to separate and separate. This range is just right."

"Start with 22 revolutions, keep the pulling speed and temperature gradient as single variables, and only verify the interface stability in the first furnace. Then check the resistivity at the beginning and end in the second furnace."

Jiang Ming wrote down the experimental sequence while simultaneously weighing the difficulty of modifying the single crystal furnace. The melting point of silicon was approaching the limits of the existing equipment, and the crucible, heating element, and observation structure all needed to be recalculated.

Wang Shouwu then opened the tin file box, which contained nearly two years' worth of silicon transistor structure design drawings, diffusion test records, and failure analysis reports. The edges of the pages were already worn and soft from being turned so many times.

The first set of drawings used an alloy junction structure, and the subsequent designs gradually shifted to a diffusion-type NPN. The base width, junction depth, and doping concentration were repeatedly modified, and many margins were marked with test conditions.

The diffusion records range from 1,050 degrees to 1,150 degrees, with six sets of temperature data for phosphorus and boron respectively. The diffusion depth shows a trend, but the leading edge position between repeated batches shows a significant shift.

Wang Shouwu took out two microscopic cross-sectional photographs and placed the two sets of test pieces, taken at the same temperature and time, side by side on the table.

"The difference in this batch is the furnace temperature. When manually controlled, it fluctuates up and down, and the diffusion front moves accordingly. Although the table states the same conditions, the actual thermal history has already diverged."

He tapped the entire record with his pen, his tone carrying a weariness that had been suppressed for a long time.

"If your temperature control system could be used in a diffusion furnace, this data could save at least six months of time."

Jiang Ming flipped to the segmented PID and lookup table compensation circuit of Furnace No. 3, compared it with the temperature zone and gas flow of the diffusion furnace, and quickly formed a migration range in his mind.

The diffusion furnace operates at 1,000 to 1,150 degrees Celsius. The silicon carbide heating element is still in a relatively nonlinear range. The existing lookup table network can reduce the number of nodes, making it easier to improve control accuracy.

The real challenge lies in the diffusion source gas. Temperature determines the rate at which impurities enter the lattice, while flow rate and concentration determine the surface source strength. Any deviation in either of these factors will alter the junction depth and doping gradient.

"The temperature control scheme can be transferred, and the diffusion source gas path must also be recorded in the same set. The flow meter, constant temperature source bottle and furnace temperature and time benchmarks must be unified. Otherwise, even if the temperature is stabilized, the condensation depth will still disperse."

Wang Shouwu pushed over a sketch of a phosphorus diffusion device, and Jiang Ming drew the measuring points along the gas inlet, source bottle, and quartz tube, and then added a bypass stabilization section after the valve.

Fang Xudong sat at the side table taking notes, writing down everything from the rotation speed of the silicon single crystal to the diffusion gas path. More than a dozen pages quickly filled the corner of the table. The dinner bell rang once, but he only went out to get a few steamed buns and a pot of hot water.

When Wu Hanzhang entered, Wang Shouwu was talking about a batch of failed boron diffusion samples, with temperature curves and micrographs spread across half the table, while Jiang Ming was rearranging the experimental matrix next to him.

"Are you planning to spread dinner into the oven too?"

Old Sun picked up the conversation from the workbench, handed over the steamed buns on an enamel tray, and the tension in the room eased a little.

Wang Shouwu ate half a steamed bun and then placed the last stack of failure analysis reports next to the file box.

"These data are all kept in 502. After the silicon single crystal is pulled out, the device-level wafers are given priority for diffusion comparison. The joint number follows from the material batch all the way to the transistor test."

Jiang Ming looked at Wu Hanzhang. This promise would take up the first batch of qualified silicon wafers, but in return, it would provide the semiconductor research group with two years of device data and complete testing capabilities.

Wu Hanzhang did not rush to speak. He flipped through the catalog that Wang Shouwu had brought, and then looked at the division of labor between single crystal and diffusion that Fang Xudong had noted down, before pulling the docking documents in front of him.

"The first batch of device-level slices will be prioritized for joint comparison, while 502 will retain the same batch of samples for independent retesting, with both parties sharing the original data and failure records."

After listening, Wang Shouwu signed his name in the semiconductor research group column, and Wu Hanzhang then stamped it with the "502" seal, the red mark falling at the boundary of the responsibilities of the two parties.

The relationship shifted from verbal collaboration to a system where materials, records, and test pieces were all linked together, meaning the success or failure of an experiment would be recorded in both files simultaneously.

After dinner, Wang Shouwu took out another blank application form and wrote the project name at the top.

"The 502 and Semiconductor Research Groups are working together to tackle key challenges in silicon single crystal growth and device fabrication, and have applied for funding from the Chinese Academy of Sciences' Semiconductor Materials Research Project."

When he mentioned the word "unity," his pen was still pressed against the paper, while Wu Hanzhang wrote the word down separately in his worn-out notebook.

In the past, Wang Shouwu often collaborated and cooperated, with each side guarding a section of the technical boundary. Now, he has personally handed over all the dissemination records and put the funding, materials and results into the same project.

Looking at the application form, Jiang Ming felt even more pressure. If the silicon single crystal pulling process failed, 502 would not only be responsible for its own pre-research progress, but would also hinder the semiconductor research group's device development path.

"The application materials were compiled by Fang Xudong, including the records of six runs of No. 3 furnace, the derivation of single crystal parameters, and the modification plan for the diffusion furnace. The original data were kept on file by each party."

Wang Shouwu nodded, put away his pen, and walked to the other side of the research room. The old medical X-ray tube was installed in a lead shield, and next to it was a precision angle turntable processed by Lao Sun.

He looked through the speckled index book, then tested the backlash of the turntable locking mechanism with his hand, and as he turned around, his gaze fell on Jiang Ming.

"What you did with discarded hospital tubes, I couldn't do in two years. I'm not admitting defeat, I'm just being defiant."

After he finished speaking, he picked up his empty briefcase and walked out of the research room. Jiang Ming, who was about to hand over the Laue cross-validation form, had to put it back on the corner of the table.

After Fang Xudong returned from dropping off the people, he bound up the records for the day and wrote a line on the last page of his notebook: "From today onwards, the joint research team will be tracked using the same set of numbers."

Wu Hanzhang put the application form and the documents into the safe, and glanced at the metal file box one last time before closing the door.

"He's almost one of us."

Jiang Ming reached out and fastened the lid of the box. The metal clasp made a soft click. The style of this box was almost identical to the one left behind by Lei Zhenbang.

The two metal boxes are now placed on the upper and lower levels of the safe. One contains the technological legacy left by the previous generation of engineers, and the other contains all the successes and failures of silicon devices over the past two years.

Jiang Ming's hand stopped by the cabinet door, his gaze falling on the [111] silicon seed crystal. The amount to be received in the next batch was far more than that of a single crystal.