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

The next morning, the quartz fragments in the tray were pushed into the pre-processing stage.

After being acid-washed, cleaned with deionized water, and dried, this batch of high-purity scraps was loaded one by one into the evaporation crucibles of the modified vacuum chamber, while Da Liu stood by and pre-adjusted the output of his self-made electron beam power supply.

At the same time, the first batch of germanium sheets was taken out of the concentrated hydrochloric acid.

Fang Xudong completed the cleaning according to the countdown given by Xie Xide. The routine drying process was skipped, and the surface water was blown away by clean nitrogen. The sample was then sent into the vacuum chamber to wait for further processing.

Once the equipment has reached the low vacuum working zone, the sample stage steadily rises to 400 degrees Celsius and remains there for ten minutes.

Fang Xudong stared intently at the panel, recording the temperature, pressure, and entry interval into the new process card one by one.

After annealing, the sample was left in the vacuum chamber to cool naturally. The pumping system continued to press down to 1.0 x 10 = -4 Torr, and Da Liu immediately turned the electron beam toward the evaporation crucible containing quartz fragments.

The electron beam power made by Da Liu was too low, and the quartz surface first showed local softening, and the evaporation rate could not be increased for a long time. The interference colors of the first three pieces changed significantly from the center to the edge.

Jiang Ming turned off the electron beam and waited for the high-voltage circuit to discharge before instructing Lao Sun to adjust the position of the crucible and sample holder and move the evaporation source to the center below the turntable.

"Increasing the power further will compromise electron beam stability. We should first shorten the source-plate distance, then extend the evaporation time, prioritizing film uniformity."

During the second round of deposition, the quartz fragments were heated in batches, the sample holder rotated at a low speed, and Fang Xudong recorded the changes in reflectivity through the observation window and then estimated the film thickness using optical interferometry.

The adjusted silicon oxide film is close to 0.2 micrometers, the interference color difference between the center and the edge is reduced, and subsequent samples also fall into the same interpretation range.

The first three samples were compiled separately as a control for uneven deposition, and the remaining samples were transferred to the diffusion process. The boron diffusion window was pierced through the silicon oxide layer with a needle, and then kept at 600 degrees Celsius for two hours.

The second phosphorus diffusion was achieved by opening a window on the other side, with the temperature dropping to 560 degrees Celsius and the diffusion time extended to three hours. The furnace temperature and nitrogen flow rate continued to be recorded using the same standard.

After the quartz boat exited the cooling section, Jiang Ming first inspected the film surface under the work light. The silicon oxide layer remained continuous, and the thin flakes that had previously rolled up along the cracks were no longer visible on the outside of the window.

Fang Xudong measured the surface resistivity point by point along the center, edge and masked area of ​​the window on the four-probe stage. The transition zone of surface resistivity narrowed significantly, and the masked area did not show the lateral diffusion of the first two batches.

One set of photos was sent to the Institute of Physics that day. Xie Xide remeasured the photoconductivity spectrum using the original method. The returned results showed that the surface state density decreased by about 50% compared to the original polished sheet.

When Jiang Ming received the curve, he marked the 30% contribution from pickling and annealing alone alongside the 50% contribution from deposition. The protective film did indeed lock in some of the treatment effect while continuing to reduce interface defects.

The new film is still far from an ideal passivation layer. Pinholes, adhesion, and post-diffusion stress all require long-term testing, but it has already secured the first batch of dual diffusion windows.

After completing the junction area inspection, Lao Sun used a miniature bench vise to fix the wafers and performed gold wire wedge soldering on each one. The range of motion of his wrist comes from years of electron tube lead wire technology, and the landing point is controlled neatly.

Da Liu was in charge of packaging and lead continuity checks, while Fang Xudong connected the chip number, crystal orientation angle, film thickness estimate, and two diffusion conditions to the device card.

A total of 27 germanium double-diffused NPN transistors were finally packaged. Three samples with uneven deposition and three samples with destructive cross sections were not included in this batch of finished products.

The DC test started with the common-emitter connection. Fang Xudong changed the base current one by one, recorded the collector current, leakage current and breakdown data, and then calculated the current gain.

The common-emitter current gain of the 27 tubes is concentrated between 35 and 50, the leakage current outside the window is significantly reduced compared to the old test piece, and the basic amplifier specifications have met the standards.

After Wu Hanzhang finished reading the first statistical table, he pushed the cigarette box back into the drawer and asked Da Liu to preheat the simple resonant cavity test stand.

"Amplifying DC power only gives you the idea of ​​'standing up'; you only know how fast it can go after sweeping the frequency range."

Jiang Ming used a Soviet-made tube with known parameters to calibrate the test bench, and then connected the first domestically produced germanium triode to the fixture, sweeping from the low-frequency end to the high-frequency end step by step.

The gain curve begins to fall after 100 MHz, and the cutoff point is near 130 MHz. The second and third ones also stop in adjacent intervals.

After Fang Xudong replaced 27 tubes in a row, the frequency points on the blackboard were finally squeezed between 120 MHz and 150 MHz, which was more concentrated than expected.

Based on the germanium electron mobility of 3,900 square centimeters per volt-second and a base region width of 0.5 micrometers, the theoretical cutoff frequency should be close to 210 MHz.

The actual results were systematically lower than expected by about 30%. This concentrated distribution actually made Jiang Ming feel troubled. Random process variance usually widens the range of points, but right now it's as if the entire batch of devices is carrying a delay together.

Fang Xudong first reconnected the Soviet-made tube to the test bench. The resonant point still matched the original calibration value. He also checked the sweep frequency harness, the parasitic parameters of the fixture, and the time base of the recorder item by item.

Wu Hanzhang turned around from in front of the blackboard and pointed to the diffusion process card, saying, "First look at the base region. The design is 0.5 micrometers, but low-temperature long-term diffusion may have widened it."

Jiang Ming selected three tubes that were close to the upper, middle, and lower positions of the frequency distribution and handed them to Fang Xudong for destructive cutting. Then, selective etching solution was used to reveal the position of the PN junction.

Under the microscope, the boundaries of the two junction regions gradually became visible, and the base region width of the three tubes fell between 0.6 and 0.7 micrometers, which was indeed higher than the original design value.

Jiang Ming substituted the measured width back into the transit time estimation. He found that the widening of the base region could only explain about 10% to 15% of the attenuation, and the remaining difference was still pressed onto the frequency curve.

Old Sun took the statistics paper to the worktable, first looking at the DC gain of the twenty-seven tubes, and then at the black line that stopped near 130 MHz.

"The gain is sufficient, but the frequency is insufficient. Is this transistor a success or a failure?"

Jiang Ming was flipping through the photoconductivity spectrum sent by Xie Xide when he stopped what he was doing. The edges of the pages were warmed by the lamp, but he first recalculated the 24 gigahertz pre-research index.

The devices in front of us prove that the surface treatment, deposition masking and dual diffusion routes are feasible, and we can conduct preliminary research on the required preamplifier tubes with a cutoff frequency of at least one gigahertz.

He wrote on the notepad that frequency is fate, gain is fate, both are fate, and then pushed the paper to the empty seat opposite the worktable, which was the seat Wang Shouwu often sat in during discussions.

The kraft paper notebook then listed four sources: the base region width exceeding the design has already explained part of it, and the relatively gentle base region doping gradient may also weaken the internal drift electric field.

The third term refers to minority carrier recombination caused by the interface state. Carriers are repeatedly captured and released by surface traps, which will continue to lengthen the transit time.

The last factor applies to the germanium material itself; a narrower bandgap increases thermally generated carriers and leakage current, and limits the combined boundaries of breakdown, electric field, and high-frequency operation.

The first three items can still be improved through the process, but the fourth item is a parameter of the crystal itself. Jiang Ming only drew a circle for the time being and did not write any words in the conclusion column.

In the next office, Wu Hanzhang opened the page on the target frequency of the 24 gigahertz pre-research program.

The current frequency of 130 MHz is still some distance from one-seventh of the required level. Even if subsequent processes double the frequency, the gap from 1 GHz will still be significant.

He held a red pencil in his hand, his gaze lingering on the word "Ge" on the cover for a long time, the tip of the pencil hovering in mid-air, and finally chose to leave it blank.

Just as Fang Xudong had finished redrawing the twenty-seven frequency points, the urgent ringing of the secure telephone shattered the tranquility of the laboratory.

The Institute of Physics called to inform Wang Shouwu that he had read the preliminary briefing and would go to Room 502 the following morning to verify the original data.