In the semiconductor industry, especially in the field of integrated circuits, wafers can be seen everywhere.
Wafer is a thin, circular high-purity silicon wafer, which can be fabricated into a variety of circuit components, making it an integrated circuit product with specific electrical functions.
When it comes to integrated circuits (IC), or physical objects, or pictures, you've seen them, as shown in the following figure:
At present, the dense components are neatly placed on a single crystal silicon material, which are regular and square. It can be seen that wafers are still cut into squares in practical applications.
So here comes the question: why do silicon wafers become round? Why is it "wafer" instead of "crystal square"?
There are two reasons to explain this problem: on the one hand, it seems that it is determined by genes; on the other hand, it is caused by environment.
"Gene Decision" - Growth Method Causes
To answer this question, first of all, we need to say a reason that has a history of about 100 years.
In 1918, Czochralski, a former Soviet scientist, established a method of crystal growth, CZ method for short.
The fabrication of silicon wafers can be summarized into three basic steps: silicon refining and purification, single crystal silicon growth and wafer forming.
The first is the purification and smelting of silica sand. In this stage, polycrystalline silicon is obtained mainly through a series of measures such as dissolution, purification and distillation.
Next is the growth process of monocrystalline silicon. Monocrystalline silicon is grown from silicon melt. High purity polysilicon is placed in a quartz crucible and heated at high temperature in a protective atmosphere to melt it. With a small seed crystal rising slowly from the rotating melt, large diameter single crystal silicon ingots can be drawn vertically.
The last step is wafer forming. Monocrystalline silicon ingots are usually cylindrical, with diameters ranging from 3 inches to more than 10 inches. After the silicon ingot is sliced and polished, a single crystal silicon wafer, or wafer, is obtained.
Its growth process is decomposed as shown in the following figure.
At present, Czochralski method is the most commonly used method for wafer growth. Besides Czochralski method, zone melting method is also commonly used.
Zone melting method, Fz method for short. In 1939, W. G. Pfann, who worked in Bell Laboratory, first came up with the idea of "regional homogeneity". Later, with the help of Henry Schuler and Dan Dorsey, high purity germanium and silicon single crystals were grown and patented.
This method uses thermal energy to produce a melting zone at one end of the semiconductor polycrystalline rod to recrystallize it into a single crystal. The melting zone moves slowly along a certain direction to the other end of the bar, and then through the whole bar, the polycrystalline bar grows into a single crystal bar. The zone melting method also needs seed crystals, and the final columnar single crystal ingot has the same crystal orientation as seed crystals.
Zone melting method is divided into two kinds: horizontal zone melting method and vertical suspension zone melting method. The former is mainly used for the purification and crystal growth of germanium, GaAs and other materials, while the latter is mainly used for silicon.
Why is there a difference between horizontal and vertical lengths? This is because of the high melting point, active chemical properties and vulnerability to foreign contamination of silicon, so it is difficult to find suitable utensils to serve, and the natural horizontal zone melting method can not be used for the growth of silicon.
The biggest difference between zone melting method and Czochralski method is that zone melting method generally does not use crucibles, introduces fewer impurities, and the content of impurities in growing materials is less.
In a word, the single crystal silicon rod is cylindrical, and the single crystal silicon wafer obtained by this method is naturally circular. That's what's shown below.——
Do you know how the sharp ends are made? Bingo, the tip on the left is the seed crystal, and the tip on the right is the crystal rod that grows to the end. When it comes out of the molten state, it is caused by complex hydrodynamic principles and the rapid solidification of molten silicon.
"Environmental Decision" - Technology Development Causes
In fact, besides the fact that the wafer growth method determines that the "wafer" is circular, there are three other determinants as follows:
2. Some people have calculated that comparing the circle with the square with the diameter of Lmm and the square with the side length of Lmm, considering that the edge of 5 to 8 mm is not available in the wafer manufacturing process, we can calculate that the ratio of square wasted area is higher than that of circular shape. Therefore, the circle is a two-dimensional figure with large surface area when the circumference is equal. It can make full use of raw materials in processing, and can separate the most chips on a wafer.
3. In the actual manufacturing process, the circular object is more convenient for production operation. Arbitrary axisymmetry of a circle is an inevitable requirement for wafer fabrication. It can be imagined that a uniform photoresist coating can be obtained on the surface of a circular wafer by spin coating. In fact, it is the only way to uniformly coat photoresist at present. But what about wafers of other shapes? It's impossible or very difficult, and it's expensive if you can.
4. As a cylindrical single crystal silicon ingot, it is more convenient to transport, and effectively avoids collision damage and material loss caused by broken edges and corners during transportation.
Also, please recall carefully, do you see that the wafer is completely round? No! See the following picture:
After the silicon rod is made, a flat angle, called Flat, is cut on the silicon rod below 200 mm. In order to reduce waste, only a small circular opening, called Notch, is cut on a silicon rod of 200 mm or more. Finally, the wafers are sliced as follows.
So, what's the use of this small gap? It's easy to think about location. In this way, the orientation of each wafer can be determined, and it is not easy to make mistakes in processing.
So why is the "wafer" square? The answer is simple. It's a legacy of history and a technical limitation. And there's really no need to make square silicon wafers.
However, there are really square silicon wafers in existence ~Look at the figure below ~~~
That's the silicon wafer used in solar cells.
So, the question arises, why is there four corners missing in the last picture???