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

After the fifth batch of directional-cut germanium wafers was divided into two groups, Jiang Ming marked the double diffusion sequence of boron first and then phosphorus on the process card, and set the target structure as NPN. Only boron diffusion was carried out in the first round.

The silicon stencil process of Yuanlei-2 was completed at around 1100 degrees Celsius. The thermally oxidized silicon film served as a diffusion mask, and the window boundary remained clear even after high temperature.

The working conditions for germanium needed to be re-established. Jiang Ming set the diffusion temperature of the first batch of boron at 620 degrees Celsius and kept it at that temperature for two hours to extend the time and compensate for the difference in junction depth caused by the decrease in diffusion temperature.

A thin layer of germanium oxide is first formed on the surface of the germanium wafer. Then, the local film layer is removed according to the design window. The width of the window, the edge position and the wafer number are all verified under a magnifying glass and photographed for record.

Da Liu checked the nitrogen cylinder and pressure reducing valve, replaced the diffusion furnace tubes for the specified time, and Lao Sun loaded the test pieces into the quartz boat. After confirming that the germanium pieces were separated from each other, he pushed them into the preheating zone.

Fang Xudong connected the furnace temperature, gas flow rate, and propulsion position to the same record sheet. After the temperature rise curve exceeded 500 degrees Celsius, he recorded the film surface observation results at each fixed temperature zone.

When the temperature stabilizes at 620 degrees, the germanium plate inside the quartz tube retains its original outline, and no change is visible from the window edge when viewed from the outside. The heat preservation timer then begins.

After two hours, Lao Sun first moved the quartz boat to the cooling section. Only after the temperature dropped to the range for taking the sample did he open the end and, through gloves, move the sample to the clean tray.

When the first germanium wafer was placed under the working light, the thin film on its surface was covered with fine cracks, the edge of the diffusion window was curling outwards, and a large portion of the film layer near the edge of the wafer had peeled off.

Fang Xudong examined each piece with a magnifying glass. The originally neat window boundaries were pushed outward by the peeling area, and several cracks extended along the surface to the adjacent area.

"The peeling boundary coincides with the edge of the window, and the opening is the first to break." He pressed the pre-diffusion photo next to the germanium sheet and pointed out the starting point of the peeling film with a fine needle.

Jiang Ming sent the test piece to the four-probe stage and measured it point by point along the center, edge and outer side of the design window. The surface resistivity changed continuously over a short distance, and the distribution far exceeded the allowable range of the device process.

Subsequent junction area inspections revealed that as boron entered the germanium body through the window, it also diffused to both sides through the exfoliation area, with the actual boundary deviating from the designed position by about ten micrometers.

The base region width of a dual-diffusion triode depends on the difference in the depth of the two diffusion junctions. If the first junction deviates by ten micrometers, the structure will deviate from the design value even if the subsequent phosphorus diffusion is completed as planned.

Jiang Ming circled the two parameters, 620 degrees and two hours, on the process card, and then aligned the location of the oxide film crack with the window diagram. The actual degree of peeling exceeded the previous estimate based on the data.

"Reduce the temperature to 580 degrees Celsius, keep it for two hours, then do another batch, only checking the film layer and lateral diffusion." He took samples from the same section from the spare section and reprocessed the surface and window.

After the second batch of test pieces entered the diffusion furnace, the nitrogen flow rate, quartz boat position, and heating/cooling rates were all kept the same as the first batch. The only active variable was the temperature decrease of forty degrees.

When the wafer was taken out, the number of cracks on the film surface did decrease, and the area of ​​peeling off the edge of the wafer also decreased. However, there were still some thin flakes that were raised near the window, and the germanium substrate was exposed in some areas.

Four-probe measurements show that the lateral diffusion narrows as the temperature decreases, and the window boundary still forms an irregular resistivity band, which is still far from meeting the requirements of a double-diffused device.

Fang Xudong plotted the two batches of results on the same cross-sectional diagram. The diffusion boundary corresponding to 620 degrees spread outwards, while the curve at 580 degrees was slightly narrower but retained the same membrane failure morphology.

Jiang Ming consulted the thermodynamic data of materials and listed the stability ranges of germanium oxide and silicon oxide side by side. Germanium oxide is prone to decomposition and volatilization under the current diffusion conditions, which leads to the destruction of the film continuity.

Lowering the temperature can only reduce spalling, while continuing to lower the temperature will prolong the diffusion time, making it more difficult to control the impurity concentration and junction depth. The thermal oxidation of germanium has lost its engineering value.

After reviewing the two batches of test wafers, Wu Hanzhang put the magnifying glass back on the table and asked, "The oxide film on the silicon wafer can protect the window, but the germanium wafer can't. Where can we find alternative materials?"

Jiang Ming listed three paths on the blackboard: silicon nitride thin film, silicon dioxide film deposition, and photoresist. Each path was followed by the current equipment conditions and material sources.

Silicon nitride requires ammonia, a silicon source, and controllable high-temperature reaction conditions. Currently, there is a lack of film-forming equipment for 502, and the purity and safe handling of process gases also need to be addressed separately.

Depositing silicon dioxide can bypass the thermal oxidation process on the germanium surface, but vacuum evaporation or sputtering equipment needs to be redesigned, and film adhesion, pinholes, and thermal stress still need to be tested one by one.

Photoresist is suitable for forming tiny windows, but there is currently a lack of stable industrial materials in China. Its resistance to temperature and residual contamination also need to be tested in practice, making it difficult to directly connect to the diffusion furnace in the short term.

Old Sun picked up a piece of peeling germanium sheet with tweezers, looked at the thin film curling up at the edges, and asked, "Electron tubes can be sealed with wax and glass glaze, can we also apply a layer here to block impurities?"

"The wax will carbonize at this temperature, and the residual carbon will contaminate the germanium surface. The thermal expansion of glass glaze and germanium is much different, so it is easy for the surface to crack when heated."

After Jiang Ming finished speaking, he kept the words "glass glaze" and added "low-temperature glass film" next to it. Old Sun's approach of directly using diffusion would not work, but the material direction was worth keeping as an alternative.

Fang Xudong compiled two batches of diffusion curves, film surface photos, and junction offset data into a process briefing, which was sent to Wang Shouwu through internal exchange channels to inquire about the semiconductor research group's masking test experience.

The reply was sent to 502 the next day, in which Wang Shouwu mentioned that his research group had also encountered similar problems.

They tried evaporative deposition of silicon oxide films, which showed better temperature resistance than thermal germanium oxide, but adhesion and pinholes still hampered the window boundaries.

At the end of the letter was a handwritten note: "Germanium's surface is probably inherently less troublesome than silicon's."

After reading the letter, Wu Hanzhang handed it to Jiang Ming and casually put the two batches of waste test pieces into the control box. The lids of the boxes were labeled "620 degrees peeling" and "580 degrees partial peeling".

On the same day, the Institute of Physics forwarded an internal technical newsletter with a summary of recent semiconductor research printed on the cover.

Wu Hanzhang opened the booklet, his gaze lingering on one page for a long time, before pushing the thin sheet of paper in front of Jiang Ming.

His index finger pressed heavily on a signature: Xie Xide.

The next topic concerns the measurement of the surface state density of germanium. The abstract discusses surface treatment, interface defects, and electrical properties, which act like a precise crosshair, directly targeting the cracks, peeling, and leakage problems that the 502 diffusion sample is suffering from.