Chapter 526 We workers have power
The workpiece table heats the silicon wafer again, makes a hard film on the workpiece, strengthens the photoresist that has been retained, and further removes the unnecessary part of the glue and other residues things.
After the hardening film is completed, the silicon wafer is restarted from scratch and the second photolithography process begins, using the same standard mask.
After the development is completed, the photoresist for secondary photolithography is cured, the excess part is removed, and the silicon wafer is sent to the microscope.
As soon as the clear microscopic image of the silicon wafer surface popped up, Ma Juan exclaimed with joy.
"Brother, yes, yes, you can see it, the accuracy is so high!"
Ma Juan shouted while looking at the microscope.
Gao Zhendong had expected that the overlaying accuracy would be high. The main factor that determines the overlaying accuracy is the workpiece table, and a group of masters from the No. 3 rolling mill showed their power by positioning the workpiece table. The accuracy is 1.5μm.
The optical system pattern distortion of the Northeast Optics Institute is less than 1 μm. Combining the two, the overlay accuracy is absolutely the same.
But it is estimated that no one dares to completely let go before seeing the real results.
Hearing Ma Juan's words, Gao Zhendong laughed loudly: "Come on, let me take a look."
Ma Juan stepped away from the microscope, and Gao Zhendong put his eyes outside the eyepiece of the microscope. He saw his own eyelashes, but he obviously didn't have the time to observe whether his eyelashes looked good, and was eager to look into the field of view.
In the microscope, the traces of the two photolithographs are very obvious. One is superimposed on the other, and the degree of overlap is very high.
The traces of the two photolithography shows a small part on both sides. The single-sided width of this part is basically an overlay error. At a rough look, it is even less than half of the entire line width. Less than.
"Okay! Great! Ma Juan, please select some typical points to cover the entire silicon wafer and make a more comprehensive measurement and comparison immediately."
Another experimenter came up and together with Ma Juan, they found points, measured and recorded.
The theory of relativity of time is vividly reflected here. Ma Juan and others feel that time passes very quickly, while Gao Zhendong and others who are waiting for the result feel that time passes too slowly.
After a fast and slow period of time passed, the experimenters sat down in front of the computer and began to enter the data into the computer. For this simple statistical analysis, they typed the data directly into the source code.
The C code for processing the algorithm has been written long ago, and now they are just filling in data into an array.
The good thing about this approach is that the processing algorithm can be modified at any time as needed.
The keyboard clicked, and the mood of the people on site calmed down. Judging from the numbers they just measured, the results were pretty good.
As the last data input was completed, Ma Juan started the compile-run command.
The algorithm was too simple and did not cause any trouble in debugging. Soon, a number popped up on the screen.
Anyone who has compiled temporary programs knows how to make it as simple as possible. This number does not even have any units.
"1.97".
In fact, one decimal place is enough. The unit that is not written out is μm.
2μm!
Although he had expected it, Gao Zhendong still laughed out loud when this number popped up.
“Hahaha...Okay! Great!”
This overlay accuracy has exceeded current needs. The number of PMOS overlays is small. In fact, it is not strictly speaking, 10μm Under the resolution, the overlay accuracy of 5μm is sufficient.
Now, Gao Zhendong can directly announce that the line width of our first-generation PMOS process integrated circuit is 10μm.
This line width does not sound very good, but for example, you can see its usefulness. The prototype of DJS-59, the golden C8008, its process is 10μm-PMOS! On that small chip, the number of transistors is 6000+.
The first 4Kbit (512Byte) DRAM, the process is 8μm-NMOS. Although it is NMOS, it does not mean that only NMOS can be used for DRAM. The biggest difference between PMOS and NMOS is that the switching direction is opposite, and the switching speed of NMOS is about 2.5 times faster than that of PMOS.
But this is 1961, and PMOS is slower but it doesn’t mean it cannot be used.
It’s a pity that when those two chips appeared, Gao Zhendong was not able to lock in the ability to redeem, but it didn’t matter. It’s still the same sentence. In the early days of the industry, the main focus was on a simple one that could be used, and the core difficulty of the design was Not big.
The transistor circuit of DJS-59 is actually basically the same as that of C8008. What Gao Zhendong needs to do now is to restore this structure to the photolithography mask.
However, this is a laborious job, and Gao Zhendong cannot do it alone. It is a waste of time. This matter falls on some senior students of Beijing University of Technology.
Originally, Gao Zhendong’s consideration was not to be too optimistic about the photolithography machine. The initial estimate was about 50μm. Later, after seeing the great efforts of the masters of the No. 3 Rolling Plant and Northeast Optics Institute, he made three SZ61XXX series chip heads. Changed to 20μm.
Now that there is a solid hope for 10μm, it is not a happy life. Even if you have to teach students for half a year and then draw pictures for a year or two, it is still worth it!
And with these foundations, there are many things that can be done, such as the favorite of automation people-single chip microcomputer.
However, whether it is a CPU-based DJS-59 or a microcontroller, it still takes time, and you have to wait for at least two things.
——The true maturity of integrated circuit technology and the complete design of chips. Especially the microcontroller, which looks similar to a general-purpose CPU, but under the current computing speed, it is best to use another structure-the Harvard structure.
Compared to the von Neumann structure, this structure is more efficient.
These are all things for later. The top priority now is to put the photolithography machine into the integrated circuit process, mature the entire integrated circuit process as soon as possible, and at least produce a certain number of basic digital logic chips first. , let everyone use it.
Although this thing is basically completely handmade, the output is not low in this era. Of course, it is not comparable to the fully automatic photolithography production lines of later generations.
Based on a rough calculation using a 5mm*5mm silicon wafer area for a logic circuit chip, a 130mm wafer can produce 500 chips.
Although the area of 5*5 is small, in fact, under the 10μm process, it can almost produce a full series of digital logic chips similar to the 74. Most digital logic chips consume the most silicon area. In fact, it is the crimping point where its circuit leads.
If CPUs and microcontrollers are still a little far away, then for most commonly used digital logic circuits, at this stage, the main restriction is how fast the chip is designed. After all, The family of digital logic circuits is really quite large.
Gao Zhendong had already made arrangements for this matter.
Wearing protective clothing and being in a clean room, it was not convenient for Gao Zhendong to announce that the prototype was successful.
Looking at the anxious comrades of the lithography machine research team outside the glass window, Gao Zhendong, who couldn't bear their long wait, had an idea, picked up a pen and paper, and wrote three big characters.
“Successful!”
Gao Zhendong picked up the piece of paper and showed it in a circle to all the comrades in the clean room and outside.
If the comrades in the clean room had already anticipated the outcome and seemed calmer, then the group of comrades outside the window who were in the unknown due to poor information cheered when they saw these three words.
"It's done!"
"It finally worked!"
"This thing is very important, you see Chief Engineer Gao is so worried about it."
" No, I guess this is at least another science and technology award.”
“I didn’t expect that I would be able to engage in science after a lifetime of hard work.”
"I didn't expect it. It's been going on for a lifetime, and now I still don't believe that I am involved in this thing."
"You are just too small-minded. Didn't you listen to what I said above?" , we have a lot of strength.”
“Yes, you are big-hearted, so why are your hands shaking? You have been an old friend for decades, and you don’t know who you are.”
< br>The comrades outside were talking and laughing. I don’t know who started it, and slowly sang a song in unison.
"We workers have power, hey, we workers have power..."
"..."
"The transformation has changed the world, hey!"
"..."
"For liberation! Hey~hey!" Hey~Hey!”
“……”
There are men and women singing, old and young, some holding hammers, some holding books, some standing next to machines, some in front of computers...
Simple, bright, sonorous and powerful The singing sounded in the photolithography machine trial production workshop, and the sound shook the walls.
As the last "Hey!" fell, all the comrades stood together, looked at Gao Zhendong, and applauded enthusiastically.
Gao Zhendong's face was full of joy. When everyone calmed down a little, he signaled the test personnel to shut down their phones and finish the work, then walked into the cubicle and took off their protective clothing.
After walking out of the clean room, Gao Zhendong arranged for a comrade to go and arrange a meal. The rules cannot be broken. Even if he can only eat chaffy vegetables, he still has to have a meal, let alone now It is much better than eating glutinous rice bran.
Then he turned to Ma Juan and said: "Ma Juan, please continue to do a few more sets of tests to see how the stability is. Then you and Master Yi will prepare. This equipment will need to be transferred next. Your entire team will not only prepare for the transfer. In addition to field preparations, preparations must be made. The core backbone will go to the relevant brother factories to support for a period of time.”
The next thing is naturally to move the equipment to the 1274 factory, where, as a Key nodes are directly integrated into the entire integrated circuit process.
During this process, key members of the research team, regardless of manufacturing, installation, debugging, or use, have to settle in the 1274 factory to deal with various environmental problems and emergencies until the photolithography machine is installed in the 1274 factory. Only when the comrades in the factory are skilled at using it can they be withdrawn.
The two nodded in agreement, and Gao Zhendong looked at the comrades: "Well, although everyone knows it, I still want to officially announce it, haha, the GK-61 proximity 10μm lithography machine has been successful. Good work, comrades! Now let’s enjoy the joy of success. Everyone knows the old rules at noon.”
(End of this chapter)