1. Semiconductor devices

1.1 What is a semiconductor

1.1.1. Overview of related concepts

1.1.1.1. What is a substance

Matter is made up of particles, and different substances are made up of different particles. Molecules, atoms and ions are three kinds of particles that make up matter .

What is a molecule?

A molecule is a particle that retains the chemical properties of a substance . It can exist independently. Molecules are composed of atoms. Molecules can be split into atoms in chemical reactions, and atoms can be recombined into new molecules.

Molecules have a certain size and mass, the same type of molecules have the same properties, and different types of molecules have different properties.

Some non-metal simple substances (such as: O2, H2, P, etc.), gaseous compounds (H2S, SO2, etc.) and organic compounds (alkanes, alkenes, alkynes, aromatics) are composed of molecules.

What is an atom?

Atoms are the smallest particles involved in chemical change . Atoms are the basic particles that form molecules and ions. Atoms can first form molecules and then substances, and atoms can also directly form substances (such as diamond, graphite and rare gases, etc.). Atoms in a substance directly composed of atoms retain the chemical properties of the original substance.

Atoms also have size and mass (mass here is the relative atomic mass, according to international regulations, 1/12 of a C-12 atomic mass is used as a standard), and they are also constantly moving.

What are ions?

Ions are charged atoms or groups of atoms . (Atomic groups refer to atomic groups composed of multiple atoms, which are usually not easy to separate when participating in chemical reactions.) Atoms are not electrically sensitive and generally do not have a stable structure. Due to their own or external effects, atoms lose or gain electrons to reach a stable structure with 8 or 2 electrons in the outermost layer (helium atom) or no electrons (tetraneutrons). Positively charged ones are called cations, and negatively charged ones are called anions.

Substances composed of ions include most salts, strong bases and low-priced metal oxides. Substances made of ions can only form compounds (eg sodium chloride).

extranuclear electrons

The arrangement of electrons outside the nucleus, each layer can arrange at most one electron, and the outermost layer can not exceed 8 electrons. For example, our silicon, atomic number 14, is located at the transition position between metals and non-metals in the periodic table of elements. Its electron distribution outside the nucleus [2, 8, 4], the outermost 4 electrons tend to be a stable structure of 8 electrons. It is not easy to get electrons or lose electrons.

Schematic diagram of Si atomic structure:

Extranuclear electron arrangement: It is mainly used to describe the arrangement of extranuclear electrons of atoms in a stable state (ground state).

For example:

1) The extranuclear electron configuration of the hydrogen atom: $1s^1$ ;

2) The electron arrangement outside the nucleus of the silicon atom: $1s^22s^22p^63s^23p^2$ ;

1.1.1.2. What is a chemical reaction

A chemical reaction is a process in which atoms or electrons are rearranged (converted or transferred) between molecules in contact with each other to form new molecules accompanied by energy changes. The essence is the breaking of old bonds and the formation of new bonds .

In the chemical change of an atom, the nucleus remains unchanged, and only the electrons outside the nucleus change. In atomic physics, once an atom undergoes nuclear fission or nuclear fusion, it is called a nuclear reaction, and this change is not a chemical reaction.

So molecules are the smallest particles that maintain the chemical properties of substances, atoms are the smallest particles in chemical changes, and the chemical properties of substances depend on the properties of molecules. Molecular properties are determined by the internal structure of the molecule, which involves strong interactions and geometric configurations between adjacent atoms in the molecule.

What is a chemical bond?

A chemical bond is a strong mutual binding force between adjacent atoms in a molecule. The basic types of chemical bonds are: ionic bonds, covalent bonds, metallic bonds.

Ionic Bond : An interaction that binds oppositely charged cations and anions (describes the ability of an anion or cation to gain or lose electrons).

Ionic compound : A compound formed by the electrostatic interaction of anions and cations. It must contain ionic bonds and may also contain covalent bonds. Generally speaking, metals are easy to lose electrons, so the compounds formed by metal elements and ammonia $NH_4^+$ (except $AlCl_3$ , ) are generally ionic compounds. $BeCl_2$ Ionic compounds conduct electricity when molten or dissolved in water.

Covalent bond : A strong interaction between atoms formed by sharing electron pairs (describing the strength of non-metallicity).

Covalent compound : A compound in which atoms directly adjacent to each other in a molecule are bonded together by covalent bonds. An atom or molecule containing only covalent bonds (must contain polar bonds). Most of them do not conduct electricity, and only a small part can conduct electricity when dissolved in water.

Metal bond : the chemical bond formed between metal atoms (or ions) and free electrons in metal crystals (describes the metal element, the strength of the metal)

Defining ionic compounds and covalent compounds can be distinguished according to whether they conduct electricity in the molten state, and those that can conduct electricity in the molten state must be ionic compounds .

1.1.2. Definition of semiconductor

Semiconductor refers to a material whose conductivity at room temperature is between that of a conductor and an insulator. It is a material whose conductivity can be controlled, ranging from an insulator to a conductor. Generally, the resistance of semiconductor materials decreases with the increase of temperature, which is opposite to that of metals. Common semiconductor materials: silicon (Si), germanium (Ge), gallium arsenide (GaAs). Silicon is the most widely used .

Note: The narrow definition of semiconductor material resistivity: conductor<semiconductor ( $10^{-5} - 10^{12}$ Ω·cm)<insulator, but more precisely should be defined according to the energy band structure of the material.

1.1.3. Overview of energy band theory

1.1.3.1. What is energy band

Regarding the definition of semiconductors until 1931, AH Wilson successfully explained the differences between metals, insulators and semiconductors based on the energy band theory (an important theory for discussing the state and movement of electrons in crystals).

The electron arrangement outside the nucleus of the atom we introduced earlier is based on the arrangement of electrons outside the nucleus in the ground state. From the energy level level of quantum mechanics, each atom has its own energy level. for example:

Energy level diagram of a hydrogen atom:

The atomic energy level was first proposed from the Bohr model, but the quantum mechanics developed later gave a more profound explanation. The division of atomic energy levels has nothing to do with the number of outer electrons, but is related to the number of nuclear charges.

The energy of the highest energy level of the hydrogen atom is 0 (the energy mentioned here is the potential energy), that is to say, when the electron is infinitely far away from the hydrogen nucleus, when the bound state just ionizes to the free state, and then there are electrons that move freely and are not bound by the nucleus.

The atom is most stable when it is in the ground state. When it is in the excited state, it will spontaneously transition to a lower energy level. After one or more transitions to the ground state, energy will be released in the form of photons during the transition.

Band

When atoms form a crystal, due to the periodic arrangement of atoms, as the number of atoms increases, the originally discrete electronic energy levels will become denser and denser, and the dense energy levels will form bands, that is, energy bands.

1.1.3.2, valence band, conduction band and forbidden band

Filled band: When the atom is in the ground state (absolute zero), all its electrons are filled up sequentially from the lowest energy level, and all electrons just fill up an energy band with lower energy, while the energy band with higher energy is completely empty in the ground state (absolute zero), and these filled energy bands are called full bands ;

Valence band: The full band with the highest energy (the outermost layer) is filled with valence electrons, so it is called the valence band;

Forbidden band: The energy band in the crystal is generally band-shaped, and sometimes there is a large energy gap between the band and the band, which is called the forbidden band (or band gap); for example, the forbidden band width of silicon crystal is 1.12eV, and electrons must receive energy greater than this to conduct electricity;

Conduction band: The energy space formed by free electrons, that is, the energy range of free-moving electrons in a solid structure, generally participates in conduction or ionization, and these energy bands are called conduction bands.

For common metal materials, the energy gap between the conduction band and the valence band is very small (the conduction band is full and overlaps almost no forbidden band), and electrons can easily gain energy and jump to the conduction band at room temperature;

For semiconductor materials, the bandgap width between the energy bands is not very large. After photoelectric injection or thermal excitation, some electrons in the valence band will cross the forbidden band and enter the conduction band with higher energy. The directional movement of electrons in the conduction band can detect the passage of current;

For insulators, the band gap between the energy bands is very large, that is, the resistivity is usually high. When the applied voltage is less than the breakdown voltage, no appreciable current flows.

1.1.3.3, direct bandgap and indirect bandgap

The semiconductor absorbs photons to make electrons transition (excite) from the valence band to the conduction band, and the process of forming electron-hole pairs has vertical transitions and non-vertical transitions, which mainly depend on the positions of the top of the conduction band and the top of the valence band in K space.

Direct bandgap: The top of the conduction band and the top of the valence band are at the same position in K space, and electrons only need to absorb (photon) energy to jump to the conduction band.

Indirect bandgap: The different positions of the top of the conduction band and the top of the valence band in K space. According to the law of energy conservation, electrons must not only absorb (photon) energy but also change momentum (phonons) to jump to the conduction band.

General compound semiconductors have direct band gaps such as GaAs, InP, InSb, etc.; while Si and Ge are indirect band gap semiconductors.

1.1.4, semiconductor classification

The basic classification of semiconductors can be divided into: intrinsic semiconductors and impurity semiconductors.

1.1.4.1. Intrinsic semiconductor

Pure semiconductors without other impurities are called intrinsic semiconductors.

The valence electrons in the outermost shell of each atom are not only bound by its own nucleus, but also attracted by neighboring nuclei. Therefore, valence electrons not only move around their own nuclei, but also appear in orbits around neighboring nuclei. Therefore, covalent bonds are formed between pure silicon atoms and silicon atoms and tend to be stable.

But under a certain external force (giving energy [light, heat, etc.]) electrons will also break free, causing the covalent bond to be broken to generate holes (the actual hole does not exist, it is formed by free electrons breaking free), and adjacent valence electrons can fill the hole (this process is called intrinsic excitation). In this way, a certain charge migration occurs in the covalent bond (the direction of hole movement is opposite to that of electrons, and there are two types of carriers in intrinsic semiconductors: negatively charged free electrons and positively charged holes), and the directional movement of electrons forms a current (heat energy is converted into electrical energy).

Note: Holes and current flow in the same direction, while electrons are in the opposite direction. The direction of current flow is defined as the direction of movement of positive charges, opposite to that of electrons.

1.1.4.2, impurity semiconductor - N-type semiconductor

In intrinsic semiconductors, although there are two types of carriers, the conductivity is poor because the carrier concentration is very low. If some impurities can be doped in it, its conductivity will be qualitatively changed (enhance its conductivity).

N -type semiconductor : a (Negative-negative, negative) semiconductor doped with a pentavalent element (such as phosphorus), with 5 electrons in the outermost layer of the element, and one extra electron (negatively charged) can easily break free from the shackles of the nucleus and become a free-moving carrier (there are more free-moving electrons, and the conductivity is also stronger).

In intrinsic semiconductors, the ratio of free electrons to holes is 1:1, while in N-type semiconductors, free electrons are the majority carriers, while holes are minority carriers. At this time, free electrons are called many, and holes are few. In P-type semiconductors, there are more holes (many) than free electrons (minority).

1.1.4.3, impurity semiconductor - P-type semiconductor

P -type semiconductor : A semiconductor doped with a trivalent element (such as boron) is called a P-type semiconductor. There are only three electrons in the outermost layer of the element, which form a covalent bond (4+3) with silicon, and there is one less electron to form a stable structure of 8 electrons.

Whether it is an N-type semiconductor or a P-type semiconductor, it is electrically neutral in itself. This is because the number of protons and the number of electrons outside the nucleus are equal. In (boron-doped) P-type semiconductors, the so-called hole is because the trivalent boron replaces the position of the silicon atom in the crystal lattice. The original covalent bond of Si-Si on the right side is gone, and a position will be vacated. This position is called a hole. Holes are mainly provided by impurities, and free electrons are formed by thermal excitation. Holes are actually equivalent to positively charged particles. In the process of thermal excitation, holes are replenished and holes are generated (breaking of chemical bonds), but it looks like the holes are moving, but it is actually the movement of electrons. So the higher the concentration of holes, the stronger the conductivity.

Similarly, in N-type semiconductors, free electrons are mainly provided by impurities, and holes are formed by thermal excitation. The more impurities doped, the stronger the conductivity.

1.1.5. Overview of semiconductor materials

Semiconductor materials can be divided into chemical composition: elemental semiconductors, inorganic compound semiconductors, organic compound semiconductors, etc.;

Semiconductor Classification	distributed	represent
Elemental semiconductor	Distributed in the IIIA to IVA groups of the periodic table of elements; C, P, Se have two forms of insulator and semiconductor; B, Si, Ge, Te are semiconducting; Sn, As, Sb have two forms of semiconductor and metal. Among them, the melting point and boiling point of P are too low, the stable state of As, Sb and Sn is metal, and the semiconductor is unstable state.	Ge, Si
Inorganic compound semiconductor	1）二元化合物半导体材料 IV-IV 族：SiC 和 Ge-Si 合金均具有闪锌矿结构； III-V族：由III族元素Al、Ga、In和V族元素P、As、Sb组成，如：GaAs； II-VI族：由II族元素Zn、Cd、Hg与VI族元素S、Se、Te形成的化合物，是重要的光电材料； I-VII族：I族元素Cu、Ag、Au与VII族元素Cl、Br、I形成的化合物; V-VI族：由V族元素As、Sb、Bi和VI族元素S、Se、Te形成的化合物; 2）三元化合物半导体材料由第II族和第IV族原子取代III-V族中的两个III族原子组成，如ZnSiP2、ZnGeP2、ZnGeAs2、CdGeAs2、CdSnSe2等; 一个I族原子和一个III族原子代替II-VI族中的两个II族原子，如CuGaSe2，AgInTe2，AgTlTe2，CuInSe2，CuAlS2等; 由一个 I 族原子和一个 V 族原子组成，如 Cu3AsSe4、Ag3AsTe4、Cu3SbS4、Ag3SbSe4 等；	SiC 、GaAs
organic compound semiconductor	Naphthalene, anthracene, polyacrylonitrile, phthalocyanine and some aromatic compounds, etc., but have not been used as semiconductor materials

1.1.5.1, the first generation of semiconductor

Rise time: The emergence of germanium (Ge) and integrated circuits in the 1950s was an important step forward for the semiconductor industry. The first integrated circuit manufactured by Jack Kilby of Texas Instruments was manufactured using a piece of germanium semiconductor material as a substrate; while Fairchild replaced germanium with silicon and invented silicon integrated circuits;

By the late 1960s, due to the shortcomings of germanium devices in high temperature resistance and radiation resistance, they were gradually replaced by silicon (Si). On the one hand, silicon reserves are extremely rich; on the other hand, silicon can form silicon dioxide (SiO2) films with good insulation properties, so silicon has become the most widely used semiconductor material;

Representative materials: silicon (Si), germanium (Ge) and other elemental semiconductor materials;

Application scenarios: Due to the narrow bandgap (band gap) of silicon and germanium, and low electron mobility and breakdown voltage, it is mainly used in low-voltage and low-frequency fields, such as: electronic information, new energy, photovoltaics, medium-power transistors, photodetectors and other fields.

1.1.5.2, the second generation semiconductor

Rise time: Since the 1990s, with the rapid development of mobile communications, the rise of information highways based on optical fiber communications and the Internet, the second-generation semiconductor materials represented by gallium arsenide (GaAs) and indium phosphide (InP) began to emerge;

Representative materials: The second-generation semiconductor materials are compound semiconductors, including a variety of III-V compound semiconductors. The most widely used commercial semiconductor devices are gallium arsenide (GaAs), indium phosphide (InP), gallium arsenide phosphide (GaAsP), gallium aluminum arsenide (GaAlAs) and indium gallium phosphide (InGaP), among which gallium arsenide technology is the most mature and widely used;

Application scenarios: Since the electron mobility of compound semiconductors is higher than that of elemental semiconductors, and compound semiconductors have a direct band gap, which is not available in silicon semiconductors, compound semiconductors are mainly used to make high-speed, high-frequency, high-power and light-emitting electronic devices. For example: light emitting diode (LED), laser diode (LD), light receiver (PIN) and solar cells and other products. Gallium arsenide and other second-generation semiconductor materials have the characteristics of high frequency, radiation resistance and high temperature resistance, and are widely used in satellite communication, mobile communication, optical communication, GPS navigation and other fields.

1.1.5.3, the third generation semiconductor

Rise time: Due to the scarcity of gallium (only 0.0015% in the earth's crust) and indium (only 0.001% in the earth's crust), material resources are very scarce, resulting in high prices, arsenic is highly toxic, and will cause serious environmental pollution. The application of the second-generation semiconductor materials has certain limitations.

Since the 21st century, the third-generation semiconductor materials represented by gallium nitride (GaN), silicon carbide (SiC), zinc oxide (ZnO), and diamond have begun to emerge.

Representative materials: The third-generation semiconductor materials are mainly silicon carbide (SiC), gallium nitride (GaN), zinc oxide (ZnO), diamond, aluminum nitride (AlN) as wide bandgap semiconductor materials; among them, silicon carbide (SiC) and gallium nitride (GaN) are more mature;

Application scenarios: The third-generation semiconductor materials have characteristics such as wider bandgap, higher thermal conductivity, higher radiation resistance, and greater electronic saturation drift rate. They can be widely used in high-voltage, high-frequency, high-temperature, and high-reliability fields, including radio frequency communications (5G base stations), radar, satellites, power management, automotive electronics, and industrial power electronics.

Semiconductor material	Forbidden band (eV)	Melting point (K)	Lattice constant (nm)
First Generation: Germanium (Ge)	1.1	1221	0.5658
First Generation: Silicon (Si)	0.7	1687	0.5428
Second Generation: Gallium Arsenide (GaAs)	1.4	1511	0.5635
Third Generation: Silicon Carbide (SiC)	3.05	2826	a:0.3080 c:1.5120
Third Generation: Gallium Nitride (GaN)	3.4	1973	a:0.3190 c:0.5190
The third generation: Diamond C	5.5	greater than 3800	0.3570

What is a lattice constant

It will be in contact with the substrate and epitaxy later.

First of all, a crystal is a substance in which atoms are regularly and periodically repeated in three-dimensional space.

For this reason, in order to describe the structure of the crystal, a concept of spatial lattice is introduced, that is to say, the lattice pattern formed by connecting atoms in this crystal seems to be a regular geometric pattern repeating (generally parallelepiped in space). We call this regularly arranged unit a unit cell. The different shapes of the unit cell determine the different properties of the crystal. The edge length and included angle of the unit cell are called unit cell parameters, and the size of the unit cell can be described by the edge length, which is called the lattice constant. The change of the lattice constant reflects the change of the composition and stress state inside the crystal, so the lattice constant is also called the lattice constant.

1.2. Starting from the diode

A diode is an electronic device made of semiconductor materials (silicon, selenium, germanium, etc.). It has unidirectional conductivity, that is, when a forward voltage is applied to the anode of the diode, the diode conducts. When a reverse voltage is applied to the anode and cathode, the diode is turned off.

Speaking of diodes, we have to mention one of its important structures - the pn junction. The basic semiconductor device is a pn junction connected with conductive terminals. When we focus on the physics of semiconductor operation, we use the term pn junction; when we focus on circuit design, we use the term diode. But they are essentially the same thing.

1.2.1. What is a PN junction

We mentioned earlier: N-type—many carriers are free electrons, and few carriers are holes; P-type—many carriers are holes, and few carriers are free electrons; they are all called carriers .

PN junction : P-type doped region and N-type doped region, after a period of diffusion or drift, there is a dynamic equilibrium region at the junction of the two, that is, a stable space charge region (or barrier region) is formed. This space charge region is the PN junction.

The electric field formed in the space charge region (or barrier region) is called the built-in electric field. For example: the built-in electric field of silicon is about 0.6~0.8V.

1.2.2. What is Diffusion Movement

Diffusion movement: The movement of electrons in the N region to flow to the P region and combine with holes.

The concentration difference of carriers produces multi-carrier diffusion movement. There are more holes and fewer electrons in the P-type region, and more electrons and fewer holes in the N-type region, so that electrons and holes will diffuse from places with high concentration to places with low concentration, so electrons diffuse from N region to P region, while holes diffuse from P region to N region, and finally reach an equilibrium point.

Electrons and holes have opposite charges, and they will recombine (neutralize) during the diffusion process. As a result, the original electrical neutrality in the P and N regions is destroyed, and the P region (which attracts electrons from the N region) loses holes and leaves negatively charged ions (negatively charged on the left), and the N region loses electrons to leave positively charged ions (positively charged on the right), so (recombination of electrons and holes) forms a space charge region (or barrier region). Just like the battery in our daily life, a large amount of positive and negative charges have accumulated at both ends of the battery (positive (high potential) -> negative (low potential), this potential difference is what we call voltage ) .

Diffusion motion widens the space charge region. You can understand that the more electrons and holes are recombined in the adjacent area during the diffusion movement, the more negatively and positively charged ions on the left and right sides, the greater the potential difference, and the wider the space charge area.

1.2.3. What is drift movement

Drift movement: the movement of a small amount of electrons in the P region to the N region.

After the space charge region is formed, due to the interaction between positive and negative charges, an electric field is formed in the space charge region, and its direction is from the positively charged N region to the negatively charged P region. Since the electric field is formed inside the semiconductor after the diffusion of carriers, it is called the built-in electric field. Because the direction of the built-in electric field is the same as the diffusion direction of electrons and opposite to the diffusion direction of holes, it prevents the diffusion movement of carriers.

The stronger the built-in electric field, the stronger the drift motion, and the drift motion thins the space charge region. That is to say, he will push the electrons originally in the P region back to the N region, trying to prevent the combination of electrons and holes, so the space charge region will become thinner.

1.2.4, forward bias and reverse bias

As we mentioned earlier, the two kinds of motion of the PN junction (drift motion and diffusion motion), when these two motions finally stabilize, there is no current. When we connect a power supply at both ends of the PN junction, the potential difference at this time will form two types, one is forward bias and the other is reverse bias. This is also the principle why diodes have unidirectional conduction characteristics (forward conduction, reverse cutoff).

forward bias

When the positive pole of the power supply is connected to the terminal of the P region, and the negative pole is connected to the terminal of the N pole, it is called PN junction forward bias, referred to as PN junction forward bias .

When the PN junction is forward biased, the direction of the external electric field is opposite to that of the built-in electric field (high potential -> low potential, and the current is the direction of positive charge movement). This makes the holes in the P region offset a part of the negative space charge in the space charge region (the negative charge in the P region is getting weaker and weaker), and the free electrons in the N region enter the space charge region to offset a part of the positive space charge (the positive charge in the N region is getting weaker). As a result, the space charge region is narrowed and the built-in electric field is weakened. The weakening of the built-in electric field enhances the diffusion movement of the majority of carriers. When the electrons and holes diffuse completely freely, the built-in electric field disappears, forming a stable current, and the PN junction is forward-conducting.

reverse bias

When the positive pole of the power supply is connected to the terminal of the N region, and the negative pole is connected to the terminal of the P pole, it is called PN junction reverse bias, referred to as PN junction reverse bias .

When the PN junction is reverse-biased, the applied electric field is in the same direction as the internal electric field in the space charge region, which will also lead to the destruction of the equilibrium state of diffusion and drift movement. The N region absorbs electrons, and the P region absorbs holes, making the space charge region gradually widen and the internal electric field strengthened, making it difficult for the majority carrier to diffuse. However, due to the constant and small number of minority carriers at room temperature, the reverse current is extremely small. The small current indicates that the reverse resistance of the PN junction is very high. Generally, it can be considered that the reverse biased PN junction is not conductive and is basically in an off state.

Note: The forward conduction may cause the diode to burn out due to excessive current, and the reverse cut-off will reverse breakdown when the breakdown voltage is reached.

1.2.5. Preparation process of PN junction

1.2.5.1. Substrate

what is substrate

A substrate is a clean single-crystal flake manufactured from a semiconductor material with specific crystal planes and appropriate electrical, optical, and mechanical properties for growing epitaxial layers.

Why do you need a substrate

There are mainly the following points:

1) grow

Some semiconductors with complex structures are not easy to grow into single crystals under normal circumstances, because there is no good attachment point, and they cannot nucleate, let alone grow into crystals. Therefore, a corresponding lattice is required as the attachment point.

2) support

A film with a thickness of several microns or even several nanometers must be attached to the substrate so that it is not easy to break and damage.

3) Participate in conduction

Many substrates are semiconductors, such as silicon, which form heterojunctions with functional materials and participate in the realization of device functions.

Substrate material

The main material of the substrate is monocrystalline silicon, which is the second most abundant element in the earth's crust and has huge reserves (basically in the form of silicon dioxide (sand) or silicate).

At present, the raw material for the preparation of polysilicon is quartz sand (reduction).

Polycrystalline silicon to monocrystalline silicon ingots are usually prepared by the Czochralski method (pulled through a furnace), so the monocrystalline silicon we usually see looks like this.

The single crystal material undergoes mechanical processing, chemical treatment (for example: doping to enhance electrical conductivity), surface polishing and quality inspection to obtain a single crystal polished sheet that meets certain standards. The purpose of polishing is to further remove the remaining damaged layer on the processed surface. The polished sheet can be directly used to make devices, and can also be used as a substrate material for epitaxy.

In addition to silicon (Si), substrate materials include sapphire ( $Al_2O_3$ ), silicon carbide (SiC), silicon nitride (SiN), aluminum nitride (AlN), gallium nitride (GaN), gallium arsenide (GaAs), zinc oxide (ZnO) and other materials.

A story about the substrate material:

Silicon substrate has the advantages of low cost, large area, high quality, good electrical and thermal conductivity, easy integration, etc., and its preparation process is relatively mature; (in 2016, Nanchang University and Jiang Fengyi's team from Lattice Optoelectronics invented "GaN-based blue light-emitting diode with high light efficiency on silicon substrate" and won the first prize of the 2015 National Science and Technology Award)

Sapphire has excellent optical properties, mechanical properties and chemical stability, high strength (second only to diamond), high hardness and erosion resistance. As a substrate material, sapphire has the advantages of stable chemical properties at high temperature (2000°C), difficult absorption of visible light, and low price. (In 1993, Isamu Akasaki and others broke through the core technology of preparing high-efficiency GaN-based blue LEDs on sapphire substrates. In 2014, they won the Nobel Prize in Physics for the invention of "high-efficiency blue light-emitting diodes".)

Silicon carbide has excellent thermal, mechanical, chemical and electrical properties, and is one of the best materials for making high-temperature, high-frequency, and high-power electronic devices. The thermal conductivity of silicon carbide substrates is more than 10 times higher than that of sapphire, but silicon carbide substrates have a strong absorption effect on ultraviolet light, so the cost is relatively high;

Aluminum nitride single crystal substrate has good thermal conductivity and is an ideal substrate material for the preparation of high current, high power, LED chips and deep ultraviolet lasers;

GaN single crystal substrate has good electrical and thermal conductivity, but the price of GaN substrate is still too high;

Substrate selection

Of course, the choice of substrate material mainly considers the following factors:

1) Structural matching of substrate and epitaxial film

The crystal structure of the epitaxial material and the substrate material is the same or similar, the lattice constant mismatch is small, the crystallization performance is good, and the defect density is low;

2) The thermal expansion coefficient of the substrate and the epitaxial film match

The matching of the thermal expansion coefficient is very important. The large difference in the thermal expansion coefficient between the epitaxial film and the substrate material may not only reduce the quality of the epitaxial film, but also cause damage to the device due to heat during the working process of the device;

3) The chemical stability of the substrate and the epitaxial film match

The substrate material must have good chemical stability, and it is not easy to decompose and corrode in the temperature and atmosphere of epitaxial growth, and the quality of the epitaxial film cannot be reduced due to the chemical reaction with the epitaxial film;

4) The difficulty and cost of material preparation:

Considering the needs of industrialization development, the preparation requirements of substrate materials should be simple and the cost should not be high.

1.2.5.2. Oxidation

The role of the oxidation process is to form a protective film on the surface of the wafer. It protects the wafer from chemical impurities, prevents leakage currents into circuits, prevents diffusion during ion implantation, and prevents wafers from slipping during etching.

1.2.5.3. Photolithography

Every chip needs to be engraved by a lithography machine at the beginning of its birth, and the lithography machine is also the core machine in the chip manufacturing process. Photolithography technology is a process technology that uses the principle of photochemical reaction and chemical and physical etching methods to transfer the pattern on the mask plate to the wafer. The principle of lithography originated from photolithography in printing technology, which is to process and form micro graphics on a plane.

According to the exposure light source, lithography technology is mainly divided into optical lithography and particle beam lithography (common particle beam lithography mainly includes X-ray, electron beam and ion beam lithography, etc.), among which optical lithography is currently the most important lithography technology.

Specifically, photolithography is divided into gluing, masking, exposure and development.

1) Photoresist

According to the difference of photoresist, photoresist can be divided into two types: positive resist and negative resist.

Positive photoresist: it will decompose and disappear after being exposed to light, leaving a pattern in the unexposed area;

Negative glue: it will polymerize after being exposed to light and make the graphics of the light-receiving part appear;

2) mask

Mask is referred to as a mask for short. Generally, the mask carries a design drawing, and light passes through the mask to project the design pattern onto the photoresist.

3) Exposure

Circuit printing is accomplished by controlling light exposure.

4) Development

The developer is sprayed on the wafer to remove the photoresist in the area not covered by the pattern, so that the printed circuit pattern can be revealed.

Stories about lithography machines:

The main manufacturers of lithography machines are: ASML (Netherlands), Nikon (Japan), Canon (Canon) Japan, Panlin Semiconductor, ABM, MYCRO, SUSS, etc. Among them, ASML’s new EUV lithography machine is the most advanced (7-22 nanometers).

In 1984, Philips and ASMI each invested 2.1 million to establish ASML;

In 1990, ASML broke away from Philips and called it an independent company. In 1995, ASML officially went public, and then acquired a number of American lithography technology companies;

In 2006, ASML released its first market-leading product, the TWINSCAN system using immersion lithography;

In 2000, a large amount of money was invested in the research and development of extreme ultraviolet light (EUV) technology. Later, due to the difficulty and high investment, ASML once wanted to give up, but faced with the strong attraction of EUV, Intel, TSMC, and Samsung couldn’t sit still, and decided to invest in ASML (the three competitors jointly acquired 23% of ASML’s shares, of which Intel accounted for 15%) to support the development of EUV lithography machines.

After the EUV lithography machine was developed, the stock price of ASML rose sharply, so Intel, Samsung, and TSMC sold a large number of ASML shares. The largest shareholder who took over became the American Capital International Group, holding 15.81% of the shares.

At present, the mainstream lithography machine in my country is still the fourth-generation DUV lithography machine with a wavelength of 193 nanometers, also known as ArF excimer (hydrogen fluoride) laser. When the wavelength of the light source is developed to 157 nanometers, due to technical limitations, it has encountered a bottleneck that is difficult to break through.

Until the invention of immersion technology (refraction of light after entering the immersion liquid) and the improvement of projection lens technology, the lithography technology has changed from 130 nanometers in 2003 to 22 nanometers now, and this technology has encountered a bottleneck again.

We know that the wavelength of light is at the nanometer level. If you want to improve the manufacturing process, you need shorter wavelengths. The wavelength from red light to purple light is getting shorter and shorter, so you have to change the light source. The wavelength of extreme ultraviolet light is 13.5 nanometers. Before, extreme ultraviolet light was not used because it consumes too much power. Extreme ultraviolet light is easily affected by the environment (air will absorb ultraviolet light, so it has to be carried out in a vacuum), and the lens will absorb ultraviolet light, so the focusing lens can only be changed to a reflector, but the reflection energy of extreme ultraviolet light is lost. Larger, so the electro-optical conversion efficiency of extreme ultraviolet light is extremely low, and the power consumption is extremely high.

1.2.5.4, etching, doping

After obtaining the lithographic pattern, the next step of processing, such as etching, doping or thin film deposition, can be carried out.

Etching can erode away the parts not protected by photoresist. It is generally used to dig grooves on the wafer. It is usually divided into dry etching and wet etching. The former mainly uses plasma bombardment, and the latter generally uses solvent immersion to dissolve. After the etching is completed, the residual photoresist is removed to obtain the desired groove pattern.

In order to change the electrical properties of semiconductors and form structures such as PN junctions, resistors, and ohmic contacts on the wafer, we also need to dope specific impurities into specific regions. The most important doping method is ion implantation, which directly implants impurity ions with high energy into the semiconductor substrate, and can precisely control the depth and concentration of doping.

1.2.5.5. Thin film deposition and epitaxy

The chip is a 3D structure formed by stacking a series of active and passive circuit components. The film deposition is to alternately stack insulating dielectric films such as SiO2 and SiN or metal conductive films such as Al and Cu on the surface of the wafer through physical/chemical methods. There are many techniques for depositing thin films, such as chemical vapor deposition and physical vapor deposition.

Common thin films are mainly divided into three categories: semiconductor, dielectric, and metal/metal compound thin films.

1) Semiconductor film

Mainly used to prepare source/drain channel regions, single crystal epitaxial layers and MOS gates, etc.

2) Dielectric film

Mainly used for shallow trench isolation, gate oxide layer, side wall, barrier layer, dielectric layer before the metal layer, intermetallic dielectric layer, etch stop layer, barrier layer, anti-reflection layer, passivation layer, etc. in the rear segment, and can also be used for hard masks.

3) Metal and metal compound films

Metal thin films are mainly used for metal gates, metal layers, pads, and metal compound films are mainly used for barrier layers, hard masks, etc. For example, metal thin films include Al, Cu, etc., which have good electrical conductivity and are used to make electrodes, wires, and superconducting devices.

epitaxy

Epitaxy : refers to the process of growing a new single crystal on a single crystal substrate that has been carefully processed by cutting, grinding, polishing, etc. The new single crystal can be the same material as the substrate, or it can be a different material (homoepitaxial or heterogeneous epitaxy).

For the traditional silicon semiconductor industry chain, making devices on silicon wafers (especially high frequency and high power) cannot achieve the requirements of high breakdown voltage in the collector area, small series resistance, and small saturation voltage drop. The development of epitaxial technology has successfully solved this difficulty. Solution: A high-resistivity epitaxial layer is grown on an extremely low-resistance silicon substrate, and the device is fabricated on the epitaxial layer, so that the high-resistivity epitaxial layer ensures a high breakdown voltage of the tube, while the low-resistance substrate reduces the resistance of the substrate and the saturation voltage drop, thus solving the contradiction between the two.

1.3. Transistor and its working principle

1.3.1, what is a triode

Semiconductor triode, also known as "crystal triode" or "transistor", is a semiconductor device that controls current and has the function of current amplification. There are two PN junctions that influence each other inside, forming an NPN (a P-type semiconductor sandwiched between two N-type semiconductors) or a PNP structure. As shown below:

The difference between the doping degree of the triode and the ordinary diode is:

1) The emitter region is highly doped so that electrons injected into the base region from the emitter region form a fairly high electron concentration gradient in the base region when the emitter junction is forward biased;

2) The base area is designed to be very thin and the doping concentration is the lowest, so that only a small part of the electrons injected into the base area recombine with many sub-holes to form a base current, and the continuous holes that recombine with the base area electrons require the base to provide current to maintain;

3) The area of the collector area is large and the doping is low: so that the high-concentration electrons in the base area diffuse into the collector area to form a collector current;

The structure of the triode:

We define three semiconductors as triodes in turn: emitter, base, and collector;

The first PN junction is called the emitter junction, and the second PN junction is called the collector junction;

Each of the three areas of the triode leads to a pin, in order:

1) Emitter e (Emitter): emits electrons

2) Base b (Base): control electronics

3) Collector c (Collector): collects electrons

1.3.2, the working principle of the triode

In fact, the triode has three different working states, which are cut-off state, amplification state, and saturation state. How did these three states come about?

Let's look at the cutoff status first :

For the emitter, when the power supply is connected $V_{bb}$ (the potential of point b is higher than that of point e), and the external voltage at this time $V_{bb}$ is too small to reach the gate voltage value of the emitter junction, the PN junction on the left is cut off, that is, there is no carrier (electron) injection in the emitter region to the base region, which is almost 0;

For the collector, when the point is connected to the power supply $V_{cc}$ (the potential of point c is higher than that of point b), the collector junction is in a reverse bias state, $I_C$ which is almost 0. The entire triode is in the cut-off state.

Next, let's analyze the zoom state :

For the emitter, when the power supply is connected $V_{bb}$ (the potential at point b is higher than that at point e) and the emitter junction is forward biased (that is, the first PN junction on the left is turned on), electrons enter the base region from the emitter region to form a current. Some of these electrons are intercepted by the holes in the base region, and a small amount of electrons will be sucked away by the positive electrode of the first power supply to form a current, while the other part will accumulate on the right side of the base region, waiting to be sucked away by the collector $IN A}$ region $I_{BN}$ .

对于集电极来说，当点接入电源 $V_{cc}$ （c点的电位比b点高），集电结就处于反偏状态时内电场增强（此时集电极电势高于基集），造成多数载流子扩散运动较难进行，同时加强了少数载流子的漂移运动（P型——少子为自由电子，N型——少子为空穴），虽然是反偏状态，但是在基区的电子是作为少数的载流子的，此时的内电场的增强不仅没有阻碍电子，反而给电子助力了，并且由于设计的浓度差和基区非常薄的缘故（电子在基区还没来得及复合），所以更多的电子被第二个电源拉进了集电区（此时占主导地位的是注入到基区的电子而不是 $V_{cc}$ ），形成电流 $I_{CN}$ ，同时因为内建电场的增强，致使集电结的空穴向左移动，形成了反向电流(漏电流) $I_{CBO}$ ，这种接法很奇妙。

When the emitter junction is forward-biased and the collector junction is reverse-biased, the electrons entering the base region are divided into two paths, that is, there must be a certain shunt relationship between the electrons $I_{B}$ and the electrons. $I_{C}$ When we constantly change $V_1$ the voltage $I_B$ will also change at this time (according to Ohm's law: $I=\frac{U}{R}$ ), and Δ $I_c$ will also change. $I_B$ The change itself is small, and a slight Δ $I_b$ corresponds to a larger Δ $I_c$ (due to concentration differences, etc.). Usually, the ratio of the change amount Δ of the collector current to the change amount Δ $I_c$ of the base current $I_b$ is called the common emitter current magnification factor (current magnification factor) and is represented by the symbol β (β = Δ $I_c$ / Δ $I_b$ ). This is the principle of three-stage tube amplification. From the above description, we can see that the change of the input voltage is to change the input current through it, and then control the change of the output voltage through the transmission of the input current, which is why the triode is called a current control device.

Finally analyze the saturation state:

What is saturation? The collector current increases with the increase of the base current. When the collector current increases to a certain extent, and then increases the base current, the collector current no longer increases. This phenomenon is called saturation. From the perspective of electron movement, when the electrons from the base region are too much to be consumed by the collector region, the entire P region can also be regarded as an N-type semiconductor, and they become a whole semiconductor, and the collector junction disappears theoretically.

From the perspective of circuit analysis, it means that when $V_{bb}$ it increases to a certain extent, it remains $V_{cc}$ unchanged. With the increase of Ic, due to the voltage division effect of the series resistance, the voltage at point c decreases gradually (the maximum value of Ic is I, c=Vcc/Rc). The triode doesn't have to do that either.

1.3.3 Three working states of the triode

Three working states of the triode:

1.4, MOS field effect transistor and its working principle

1.4.1, the origin of MOS

In 1831, Faraday, a British physicist and chemist , discovered electromagnetic induction, so some people called this year the year of the electrical era;

In 1833, Faraday discovered a material: it is different from other metals in that its resistance decreases as the temperature rises. At that time, it was only absolutely peculiar and did not arouse great sparks. This is also the first physical property of semiconductors discovered in the dark;

In 1874, German physicist Ferdinand Braun noticed that the conductivity of sulfide is related to the direction of the applied voltage, which is the rectification effect of semiconductors;

Between 1930 and 1940, the exploration of semiconductors continued;

In July 1945, that is, after World War II, Bell Laboratories in the United States established a research department for solid-state physics. In order to produce transistors, it has been facing many difficulties. Shockley began to use semiconductors to develop a new electronic amplifier. The clever but indifferent Shockley chose to work at home, allowing colleagues Walter Bratton and John Bardeen to freely carry out experiments at Bell Laboratories in New Jersey;

In November 1947, an accidental discovery by Bratton gave theoretical physicist Bardeen a crucial new understanding of the behavior of electric currents on semiconductor surfaces. Bratton cobbled together an amplifier out of plastic, gold leaf, and the semiconductor germanium, and tested it. This amplifier is successful, and it can control a huge current with a small input voltage. This is the first transistor in history;

In 1949, Martin M. "John" Atalla entered Bell Laboratories, where he focused on research on the surface properties of semiconductors. He was born in Port Said, Egypt, and received a Bachelor of Science degree from Cairo University in Egypt; a Master degree in Mechanical Engineering from Purdue University in the United States in 1947, and a Ph.D. degree in 1949;

In 1950, Bell Laboratories used the Czochralski method (Czochralski, rotating around the lifting axis at a certain speed) to successfully prepare single crystal germanium (Ge), which was also the earliest semiconductor material used in discrete devices;

In 1951, Western Electric Corporation (Western ElectrIC) began to produce commercial point-contact germanium transistors;

In 1952, companies such as Western Electric, Raytheon, and American Radio produced commercial bipolar transistors;

In 1955, Shockley returned to his hometown of Santa Clara Valley (later "Silicon Valley"), founded Shockley Semiconductor Laboratory, and served as director;

In 1956, Shockley, Bardeen, and Bratton won the Nobel Prize in Physics for the invention of the point-contact germanium transistor, and Shockley was even known as the "father of the transistor";

In 1959, Dawon.Kahng (Cohen, Korean, also known as Jiang Dayuan ) received a doctorate in electrical engineering from Ohio State University in the United States, joined Bell Labs, and worked under Atala at that time. Their experiment successfully overcomes and penetrates the surface state that hinders the formation of the electric field, and successfully enters the semiconductor material. This is the first successful insulated gate field effect transistor (FET) in the global semiconductor science community. Later, Bell Labs applied for a patent for this;

In 1962, Bell Laboratories officially announced that Cohen, Atala and others had successfully developed the metal-oxide semiconductor field effect transistor (MOSFET). Compared with bipolar transistors, it has low power and simple manufacturing process, so it is called the basic component of later large-scale integrated circuits. This has also become one of the most important milestones in the history of semiconductor development.

1.4.2, what is a MOS field effect transistor

What is a Unipolar Transistor

Field Effect Transistor, referred to as FET (Field Effect Transistor), is a voltage-controlled device that controls the magnitude of the output current by the electric field effect generated by the input voltage. When it works, only one type of carrier (majority carrier) participates in conduction, so it is also called a unipolar transistor.

Note: The triode is a bipolar device, and both carriers participate in conduction.

Its main features:

1) The input resistance is high, up to $10^7$ ~ $10^{15}$ Ω, and the insulated gate field effect transistor can be as high as $10^{15}$ Ω, which is much greater than the input resistance of the triode.

2) Low noise, good thermal stability, simple process, easy integration and small size, easy to control device characteristics.

What is a MOS field effect tube

MOS field effect transistor is the abbreviation of MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor), that is, metal oxide semiconductor field effect transistor, which belongs to the insulated gate type of field effect transistor. Therefore, MOS transistors are sometimes called insulated gate field effect transistors . MOSFETs use the magnitude of the gate-source voltage to change the amount of charge induced on the surface of the semiconductor to control the magnitude of the drain current. In general electronic circuits, MOS tubes are usually used in amplifying circuits or switching circuits.

Classification of MOS field effect tubes

JFET (junction type) is usually called "ON device" and is a depletion-type tool with low drain resistance. It uses the control of the reverse voltage of the PN junction on the thickness of the depletion layer to change the width of the conductive channel, thereby controlling the magnitude of the drain current. It allows a smaller input impedance than MOSFET, because MOSFET is embedded with an insulator, so the leakage current is smaller, and there are not many practical applications.

MOSFETs are often referred to as "OFF devices" and can operate in both depletion and enhancement modes and have high drain resistance. Due to the different doping of the substrate, it can be divided into N-channel (N-type MOS transistor-NMOS) and P-channel field effect transistor (P-type MOS transistor-PMOS).

Gate (gate electrode) : gate-gate, translated into Chinese as grid, electrode-electrode, plays a controlling role.

Source (source) : source-resources, translated into Chinese as source, an electrode that acts as a collector.

Drain (drain) : drain-discharge, the Chinese translation is drain, the electrode that acts as an emitter.

1.4.3, the working principle of MOS transistor

When VGS=0 V, there are two back-to-back diodes between the drain and source, and applying a voltage between D and S will not form a current between D and S;

When a voltage is applied to the gate, if 0<< $V_{GS}$ threshold $V_{GS}(th)$ ( $V_{GS}(th)$ -turn-on voltage), through the action of the electric field formed between the gate and the substrate, the multiple sub-holes in the P-type semiconductor near the bottom of the gate will be repelled downward, and a thin layer of negative ion depletion layer will appear; at the same time, the minority sub-holes will be attracted to move to the surface layer, but the number is limited, not enough to form a conductive channel and communicate with the drain and source, so it is still not enough to form a drain current;

With $V_{GS}$ the increase of , when $V_{GS}$ > $V_{GS}(th)$ , since the gate voltage at this time is relatively strong, more electrons are gathered in the surface layer of the P-type semiconductor near the bottom of the gate, which can form a channel (equivalent to resistance, the larger the channel, the smaller the resistance and the stronger the conductivity), connecting the drain and source $V_{GS}$ . At the same time, due to the positive voltage applied to the drain at this time, the current from the drain to the source can be formed, and the MOS transistor is turned on. As the voltage continues to $V_{GS}$ increase, $I_D$ it also increases;

And the size of will also affect the width of the right channel, and there are also three working states in the NMOS tube:

1) When $V_{GS}$ > $V_{GS}(th)$ and $V_{DS}<V_{GS}-V_{GS}(th)$ , it is in a non-saturated state, and is controlled by and $I_D$ at the same time ; $V_{GS}$ $V_{DS}$

2) When $V_{GS}$ > $V_{GS}(th)$ and $V_{DS}>V_{GS}-V_{GS}(th)$ , it is in a saturated state, that is to say, when $V_{GD}$ the decompression gradually approaches $V_{GS}(th)$ , the right channel is pinched off at this time, and it $I_D$ is only $V_{DS}$ controlled after the pinch off;

3) When $V_{GS}$ < $V_{GS}(th)$ , the channel is not formed and is in a cut-off state;

That is:

1.5, CMOS inverter

1.5.1, what is a CMOS inverter

Inverter (Inverter), also known as NOT gate, is a logic gate that realizes logic in digital logic, and is one of the most basic gate circuits. Since NAND logic can realize arbitrary logic operations (greatly simplifying complex operations), NAND gates are the most widely used logic gate circuits.

We have been exposed to TTL inverters in digital logic circuits. TTL circuits use bipolar transistors (triodes) as switching elements, so they are also called bipolar integrated circuits. Bipolar digital integrated circuits are devices that use two different polarity carriers, electrons and holes, for electrical conduction. It has the advantages of high speed (fast switching speed) and strong driving capability, but its power consumption is relatively large and its integration level is relatively low.

The CMOS circuit is composed of insulating field effect transistors. Since there is only one carrier, it is a unipolar transistor integrated circuit. Its main advantages are high input impedance, low power consumption (such as static power consumption is almost zero), strong anti-interference ability and suitable for large-scale integration.

CMOS inverter

The CMOS inverter consists of a P-channel enhancement MOS transistor and an N-channel enhancement MOS transistor connected in series. Usually the P-channel tube is used as the load tube, and the N-channel tube is used as the input tube.

It can be clearly seen that the TTL type circuit is obviously much more complicated.

1.5.2. Working principle of CMOS inverter

From the above COMS inverter circuit diagram, we can see that the inverter consists of a load transistor (PMOS) and an input transistor (NMOS) connected in series. The turn-on voltage of the two MOS transistors $V_{GSth-p}\leqslant0$ , $V_{GSth-n}>0$ , is required to ensure normal operation $V_{DD}>|V_{GSth-p}|+V_{GSth-n}$ .

If $V_1$ the input is low level (such as 0V), the load tube is turned on, the input tube is cut off, and the output $V_0$ voltage is close to $V_{DD}$ ;

If $V_1$ the input is high level, the input tube is turned on, the load tube is turned off, and the output $V_0$ voltage is close to 0V;

1.5.3. Main characteristics of CMOS inverters

Voltage Transfer Characteristics

The voltage transfer characteristic curve of a CMOS inverter can be divided into five working areas.

The working area we need is the stable state area. When it is stable, the CMOS inverter works in the I and V areas, and there is always a MOS transistor in the off state, and the leakage current flowing is extremely small.

Current Transfer Characteristics

According to the above analysis, we can know that when the CMOS inverter works in Zone III, since the load tube and the input tube are in a saturated state, a large current will be generated, and the current in other cases is extremely small, so the power consumption is extremely low.

Digital integrated circuits (on)