Nowadays many people know what a CPU is, but only few of them can describe how to build millions of microscopic transistors and place them in order to make them work together.
From sand to “ingots”…

Sand is a very common material on Earth’s surface and it is composed by 25% of Siliceous.
The first thing to do is to separate siliceous from sand. This purification process requires several passages to grant a high level of purity (the maximum accepted is 1 atom per million).Purified Siliceous is then melted to obtain a pure Siliceous crystal called “ingot” that weights 100 kg and is 99,9999% pure.

The ingot is then cut in many thin discs called “wafers”, which are thoroughly cleaned in order to eliminate every defect.
Generally modern CPUs are obtained from 300 mm diameter wafers.

A blue photoresist liquid is then distributed on their surface in order to prepare them to the next step. During this phase, the wafer is constantly turned around itself in order to grant an uniform liquid distribution (as we can see in the picture).

The wafer is then exposed to UV radiations, which are partially filtered by a mask (as we can see in the picture). The parts which have been exposed to UV ray become soluble. Through the mask, it is possible to give a precise shape to the Siliceous and through a lent it is possible to obtain a smaller shape on the wafer than the one on the mask.
In fact, the shape on the mask is bigger than the real one on the wafer; this procedure enables to create microscopic transistors from bigger molds.
From “wafers” to transistors…

In the picture we can see a transistor’s enlarged image. A transistor works as an interrupter, as it is able to control the electric current inside itself.
Nowadays transistors are so small that it’s possible to place 30 millions on the head of a pin.
The areas exposed to UV rays are now dissolved and eliminated using a specific chemical solvent. This is the first step to the CPU building. Those which haven’t been exposed to UV rays are now covered by a photoresist, which protects the ones to be preserved. The photoresist film is eventually removed and the outcome is similar to the one in the picture.

Subsequently, a new photoresist film is applied and the transistor is exposed again to UV rays and to a new washing phase. This step is called “ion doping”, because it consists on the ionized particles exposition, which modifies chemical properties of siliceous. This is fundamental to create the necessary properties for CPUs.
The next step consists of an ion bombing on the exposed areas of the wafer and it’s called “ion implantation” (ions are shot at 300.000 km/h). After that ions are planted in siliceous altering its chemical properties.
Photoresist layer is removed after ions bombing and the exposed material (the green one) now contains new atoms.

Now the transistor is almost finished and three holes are applied on the isolation layer (magenta) on the transistor. These holes will be covered by copper, which is fundamental to link each transistor with the other ones.

Copper ions are settled on the transistor (the “elettroplanting”). The wafer is then plunged in a CuSO4 solution, which is subjected to an electric field. So, copper ions move from the positive terminal (anode) to the negative one (cathode).
The outcome is a thin copper layer on the wafer and the excess material is removed.

Several metal layers compose “think wires” between transistors, which are extremely important for CPU structure and functions.
If we watch a CPU by naked eye, we’ll probably observe that it’s very thin, but if we watch it through a microscope we acknowledge that it has 20 transistor’s circuit layers.
From transistors to “dies”…

If its quality has been granted, the wafer is cut in different unities called “die”. Properly functioning dies are taken and the other ones are tossed.
Now the die is placed in its case with a heatsink and...

a new CPU is born!
0 comments:
Post a Comment