How is the CPU made; Although the way the CPU works seems magical, it is actually the result of decades of intelligent efforts by engineers. The smaller the transistors, which are the building blocks of any microchip, to microscopic scales, the more complex they become.
How is the CPU made
Photolithography
Transistors are now so small that manufacturers cannot make them using conventional methods. Although older but accurate devices and even 3D printers can produce extremely sophisticated chips, they usually work at the micrometer level and of course with high precision. One micrometer is about thirty-thousandths of an inch. These devices are not suitable for the nanometer scales on which today’s chips are made.
Photolithography solves this problem by eliminating the need for the precise movement of complex machines. Photolithography uses light to engrave an image on a chip instead of making a chip with the precise movement of machines. This device works like an old ceiling projector that you may have encountered in the classroom.
But unlike a projector that magnifies small images, it reduces the stencil of large images by a certain scale.
This image will be engraved on a high-precision silicon wafer in a controlled laboratory, as any dust stain on the wafer can cost thousands of dollars.
The wafer is covered with a material that is resistant to light. The material responds to light and washes out, creating an etch of the CPU that can be filled with copper or doped to form a transistor.
This process is repeated many times after that. Building a CPU is just like making plastic layers with a 3D printer. The main difference is in the scale of the CPU components compared to a plastic layer.
Issues and problems related to nanoscale photolithography
Transistors can shrink too much smaller scales, but they may not work well. Nanoscale transistor technology faces many physics problems. Transistors are supposed to stop the flow of electricity when they are off, but these semiconductors are getting smaller every year and electrons can pass through them. This phenomenon is called quantum tunneling and is a huge problem for silicon engineers.
Another problem is the shortcomings of chip-making machines. Even photolithography, despite its high accuracy, is not 100% accurate. Just like a projector that may provide a blurry image, a reduced photolithographic image may have similar problems. It is not entirely clear how accurate the images given will be when reduced to photolithography.
At present, foundries are trying to produce lasers much shorter than humans can perceive by using lasers in a vacuum chamber and using intense ultraviolet light to reduce this effect. But as the size of the transistors gets smaller, this problem will continue.
These defects can sometimes be reduced with a process called binning. If defects in the transistors interfere with the processor core, that core is deactivated and the chip is sold as a lower core chip. In fact, most cheaper processors are manufactured using the same design, but their cores are disabled and sold at a lower price.
If the defect reaches the cache or some other necessary component, the chip manufacturer may have to recycle that chip. Of course, some of these defective chips will enter the market, which will have less performance due to defects in cache or other components. Chips based on newer architectures such as 7nm and 10nm have higher failure rates and are therefore more expensive than older chips.
CPU packaging
Packaging the CPU to reach the consumer is more than just putting it in a box with some polyester. When the processor is finished, this part is still not suitable for use by the user unless it can be connected to the appropriate port. The packaging process is how a thin silicone mold is attached to a PCB, which most people think of as a CPU.
This process requires a lot of precision, but the precision required is not as high as the previous steps. The processor mold is mounted on a silicon plate and electrical connections are made at all pins that make contact with the motherboard. Modern processors can have thousands of pins. For example, AMD’s advanced processors have 4,094 pins.
Because the CPU generates a lot of heat and must also be protected from the front, an integrated heat dissipator is installed at the top. This heat dissipator causes the processor to contact the mold and transfers heat to a cooler that is installed at the top of the processor (such as a fan). Thermal paste is used for better heat transfer from the processor to the heatsink.
Of course, the thermal paste performance is not good enough, which has led researchers to look for better ways to transfer heat from the processor out.
When CPU production is complete, it can be packaged in a variety of boxes. Despite the complexity of making these chips, it is surprising that most CPUs cost just a few hundred dollars.
Frequently Asked Questions:
How does thermal paste transfer heat?
The surface of the processor and the heatsink have very small holes. If the processor and the heatsink come into contact normally and without the use of thermal paste, these pores will be filled with air. Because the air has a very low thermal conductivity, heat transfer from the processor to the heatsink is not good. The thermal conductivity of thermal pastes is high and by covering the processor surface of this type of pastes, heat transfer from the processor to the heatsink is done much faster.
What is Dop?
Doping means the entry of impurities into a semiconductor crystal to modify its structure. Two of the most important materials are dopable silicon boron (3 electrons capacity = 3 valence) and phosphorus (5 electrons capacity = 5 valence). Other materials are aluminum, indium (trivalent) and arsenic, antimony (trivalent).
Dupont is integrated into the structure of a semiconductor crystal lattice and defines the number of external electrons of the doping type. Elements with a capacity of 3 electrons are used for doping type Pi. The conductivity of a deliberately doped silicon crystal can be increased by a factor of 10.
