Why are monocrystalline silicon products generally cylindrical?

Someone asked me some time ago, what is a transistor? How is it made? This made it difficult for me to answer. Because the question to be answered is somewhat complicated, today we will talk about what is a transistor and how it is made.

A transistor is a general term for components such as a diode, a triode, and a field effect transistor. It is a component used to process circuit signals. Of course, there is no transistor in the circuit light, and other components must be matched to complete certain circuit functions. Let me talk about the manufacturing process of transistors:

semiconductor

The so-called semiconductor, it refers to a substance whose electrical conductivity is between the conductor and the insulator. But in reality, the semiconductors we need are not simply semi-objects that are less conductive. A semi-substance like a conductive substance in an insulator cannot be used as a transistor. The semiconductors used to make transistors and integrated circuits are mainly carbon-based elements. For example: silicon and germanium. Silicon and germanium are one of the most important materials for making semiconductors. In addition, gallium arsenide, gallium phosphide, cadmium sulfide, zinc sulfide, and some metal compounds are all semiconductors. But the most versatile is silicon semiconductors. Today, silicon semiconductors are taken as an example to discuss the principle and manufacturing process of transistors.

Carbon cluster element

The carbon cluster element belongs to the fourth main group element in the periodic table, and the outermost layer of the family element has four electrons. According to research, the maximum number of electrons in the outermost layer of the atom is 8. When the number of outermost electrons is 8, the atom reaches the most stable state. The number of outermost electrons of the eighth main group element in the periodic table is eight, so the elements of the eighth main family are all ambiguous elements. And if the number of electrons in the outermost layer of the atom is less than four, then the atom is more likely to lose electrons and reach a stable structure. And the fewer the number of electrons, the easier it is to lose the outermost electrons. If the number of electrons in the outermost layer of an atom is greater than five, then the atom is easy to get electrons to reach a stable structure. Moreover, the more the number of outermost electrons (5-7), the more electrons are obtained to reach a stable structure. The carbon cluster element is one of the most special elements, and its outermost electron number is four. Between these two atoms, a structure in which eight electrons are common to two atoms is used. This structure, we call it a shared pair of electrons. This is also a relatively stable structure.

Smelting of silicon

Silicon is one of the most important semiconductor materials. Our commonly used transistors and various integrated circuits are basically made of silicon as the main material. The CPU is no exception. Silicon is one of the most abundant elements in the world. Its content in the earth's crust is 26.30%, second only to oxygen (48.60%, also considered to be 48.06%). The cement, clay, sand, stone, quartz, etc. we usually see are all silicon compounds.

Although the silicon compound is under the sole of the foot, it is not used to smelt the single crystal silicon. The reason is very simple. On the one hand, the amount of single crystal silicon we need is not very large. We don't need to use all the soil to smelt. On the other hand, there are many other impurities contained in the soil, and the process of smelting single crystal silicon with mud is complicated and difficult to purify. Therefore, quartz sand and quartz are generally used in the industry to smelt single crystal silicon. The white translucent stone we usually see by the river is quartz. The main component of quartz is silica.

The smelting of silicon is divided into smelting and refining: the smelting method uses a reduction method. Quartz, petroleum coke and bituminous coal are mainly placed in the furnace for smelting using a high-power electric furnace. The reaction equation: SiO2 + 2C → Si + 2CO↑, wherein the source of carbon is mainly petroleum coke and bituminous coal. The specific smelting method is not introduced here, and interested friends can refer to the related articles.

Silicon refining: Silicon produced by the reduction method is generally not of high purity and can only be used for general industrial silicon. In order to achieve silicon-grade high-purity silicon, further purification is required. The single crystal silicon currently used in integrated circuits is mainly a chemical method, such as the Siemens method (trichlorosilane reduction method), which is one of the main methods. In this method, hydrochloric acid (HCl) is reacted with ground crude silicon at a high temperature to form SiHCI3, and then the formed SiHCI3 is chemically purified to finally reach electronic grade polycrystalline silicon. The purification principle of silicon is not very complicated, but since the purity of silicon at the electronic level is extremely high, the purity of the single crystal currently required is 99.999999999% to 99.99999999999%. Therefore, the entire purification process of silicon is very complicated.

After chemical purification, the purity of silicon has been solved, but this is not a single crystal silicon. Some special methods are needed to make silicon (polysilicon and amorphous silicon) into single crystal silicon. Currently used methods are Czochralski (CZ), suspension zone melting (FZ) and epitaxy. Due to space limitations, I will not introduce it here.

The finished products of single crystal silicon are generally cylindrical, and the cylindrical single crystal silicon is sliced ​​and is the raw material for manufacturing integrated circuits and transistors. Also, since the shape of single crystal silicon is circular, it is also called a wafer.

The reason why the wafer is made in a cylindrical shape is because the single crystal silicon is "pulled" in a molten state by using polysilicon or non-oriented silicon. The drawn single crystal silicon naturally forms a cylindrical shape due to the tension of the material itself, just like letting a drop of water let it fall freely, no matter what shape it was before it fell, but it will change after a period of time in the air. Rounded into a sphere. Polycrystals are also the same in the process of drawing into single crystal silicon, which naturally becomes cylindrical during the drawing process. Although the monocrystalline silicon is drawn into a square shape, the utilization rate may be higher, but there is currently no good way. In fact, the most important reason, I think there is no need for this, because the wafer scrap can also be used to make other products.

Penetration principle

To clarify the principles of diodes and transistors, let's do an experiment first. Drop a drop of ink into a bottle of water, even if we don't stir it anymore, after a period of time, the whole bottle of water will turn into the color of the ink. This is because in the fluid, the high concentration of the substance always diffuses and diffuses toward the lower concentration of the substance. In water, the concentration of ink is low, and the concentration of water is high, so water penetrates into the ink. In the ink, the concentration of water is a low concentration, and the concentration of the ink is a high concentration, so the ink penetrates into the water.

Doping

The sliced ​​single crystal silicon can be integrated into the integrated circuit and the triode without being connected to the wire, but it has to go through many processes to be finally completed. The first process is doping. Maybe you have to wonder if I am wrong? It’s hard to purify silicon to the purest element in the world. Now it’s going to be mixed with impurities? I am not wrong, this is indeed the case. In order to make diodes and transistors, doping is absolutely necessary. Now let's see how the diodes and transistors are made.

PN junction

It has been said that the single crystal silicon after slicing needs to be doped with impurities, but the impurities are not equivalent to garbage, and the impurities are not impurities before the purification of the single crystal silicon. Single crystal silicon for diode and triode integrated circuits is doped with high purity elemental phosphorus and high purity elemental boron. The doping method is currently commonly used in the diffusion method, and the amount of doping is also strictly controlled, and it cannot be more or less. Otherwise it will affect the performance of the component and even become a waste product.

After a single crystal silicon is doped with phosphorus (of course, other pentavalent elements such as arsenic may be doped), an N-type semiconductor is formed. Since phosphorus is pentavalent, that is, the outermost layer of the phosphorus atom has five electrons. These five electrons form a covalent bond with the four electrons at the outermost layer of the adjacent silicon atom, and there is one more electron, and the combined electron and the phosphorus atom and the silicon atom are much weaker. Therefore, this extra electron becomes a "free electron" that is relatively easy to move. As a result, the conductivity of the semiconductor is greatly enhanced. We use the same method to incorporate a trivalent boron atom into another single crystal silicon. The semiconductor doped with boron atoms is a P-type semiconductor. Since the outermost layer of the boron atom in the P-type semiconductor has only three electrons, the three electrons form one less electron with the adjacent silicon atom. Thus, this atom forms a hole, and this hole easily gets an electron from other atoms. But the atoms that lose electrons form holes, so that these holes become a "free hole" like free electrons.

diode

Combining the upper P-type semiconductor and the N-type semiconductor forms a PN junction, and the PN junction is added with a lead (ohmic contact) to form a diode.

How the diode works

When a P-type semiconductor is combined with an N-type semiconductor (see Fig. 1), since there are many holes in the P-type semiconductor, there are many free electrons in the N-type semiconductor. When they are combined, the concentration of electrons in the N-type semiconductor is high. According to the principle of permeation, electrons in the N-type semiconductor diffuse into the P-type semiconductor. The result of the diffusion is that the originally uncharged N-type semiconductor is positively charged, while the originally uncharged P-type semiconductor is negatively charged (see Figure 2). As a result of charging, an electric field is formed between the PN junctions. It is precisely because of the existence of this electric field that the expansion of electrons is slowing down (same-sex, reciprocal). This diffusion movement will stop after a period of time, of course, it is not completely stopped, but a small amount of electrons continue to spread. This is because the edge of the P-semiconductor combined with the N-type semiconductor is affected by thermal motion and various rays (including various visible rays), so that a small amount of electrons inside the P-semiconductor are accelerated and then escape from the original position. In the electric field. The electrons entering the electric field are subjected to an electric field to move the electrons toward the N-type semiconductor. We call this electron motion a drift motion, and the result of the drift motion causes the diffusion motion to continue. When the diffusion motion of electrons and the drift of electrons reach a dynamic balance, they are in a stable state. Since the drift motion is a minority carrier in the P-type semiconductor, the current formed by the drift is small. What is a minority carrier? The so-called minority carrier means that if a current is mainly formed by electron motion in this semiconductor (such as an N-type semiconductor), then the electron in the semiconductor is the majority carrier fluid, and the hole is the minority carrier, and vice versa. anti.

Unidirectional conductivity of the diode

The diode fabricated by the above method has unidirectional conductivity, which allows current to flow in only one direction through the diode current. Let us analyze the unidirectional conduction principle of the diode

Refer to (Fig. 2) when the P-type semiconductor is connected to the positive terminal of the power supply after the load is connected, and the N-type semiconductor is connected to the negative terminal of the power supply. The electrons in the negative electrode of the power source flow to the N-type semiconductor under the action of the power source and recombine with the positive ions in the N-type semiconductor. Similarly, the electrons in the P-type semiconductor flow to the positive electrode of the power supply under the action of the power source, and are combined with the positive ions inside the power source. When the electrons and ions inside the semiconductor PN junction are recombined, the concentration of holes and electrons inside the semiconductor PN junction increases. As a result of the increase in concentration, the diffusion motion continues, and the half body conductance is in a conducting state.

When the N-type semiconductor is connected to the load and connected to the positive pole of the power supply, the P-type semiconductor is connected to the load and then connected to the negative pole of the power supply. What is the situation? In a P-type semiconductor, an electron passing electrode of a power source becomes a co-bond with an atomic type of a P-type semiconductor. The N-type semiconductor also forms a co-bond due to the loss of electrons, which is equivalent to the thickening of the entire PN junction. A thickened PN junction has a blocking effect on the flow of electrons. At normal voltage, the electrons of the power supply cannot pass through the thickened PN junction. Therefore, it can be said that the PN junction at this time is non-conducting. However, due to the influence of thermal motion rays, etc., the PN junction still has a very small current. This current is the reverse current, and the reverse current is generally small, which can be ignored under normal conditions.

Transistor

The structure of the triode is actually equivalent to two back-to-back diodes (pictured), but with two back-to-back diodes

It cannot be used as a triode. This is because the internal structure of the triode is different from the two back-to-back diodes. Let's take a look at how the triode works:

For the triode to work properly, its circuit must be properly connected. The figure above is the basic principle diagram of a triode amplifier circuit. In the figure, if the EB voltage is 0, the triode is in the off state. Because, at this time, the PN junction of the C zone and the B zone are in a reverse bias state (the same principle as the above two poles). When the EB is applied with a suitable voltage, since the PN junctions of the E and B regions are forward biased, the PN junction is in an on state. The electrons in the E region of the PN junction after conduction are continuously diffused to the B region. Since the triode is made thin during the manufacture, the electrons diffused into the B region are easily diffused to the edge of the C region. The electrons that diffuse to the edge of the C region form a current under the action of the C-pole power source, and the EC pole of the triode is in a conducting state. In addition, since the contact surface of the BE junction and the CE junction is large, the current formed in the CE region is also large. This is the amplification principle of the triode. The above is the NPN type triode. In addition to the NPN type triode, there are also PNP type triodes. Their principles are the same and will not be repeated here.

Field effect transistor

Field effect transistors are also classified as transistors, except that the field effect transistors are slightly different from the NPN type and PNP type transistors described above. The NPN type and PNP type transistors are bipolar transistors, and the field effect transistors are unipolar transistors. Due to the length of the relationship, I will not say more here.