button cell battery cr1620.Research on GaN-based materials and devices on Si substrate

　　As a new wide-bandgap semiconductor material, GaN has always been a hot spot in international compound semiconductor research. GaN is a direct bandgap material that can form a ternary or quaternary solid solution alloy (AlGaN, InGaN, AlInGaN) with continuously variable compositions with InN and AlN. The corresponding wavelengths cover the range from red light to near-ultraviolet light, and have It has excellent properties such as good chemical stability and thermal stability, so it has great application prospects in the field of optoelectronics. Secondly, compared with other materials such as Si and GaAs, GaN material has a greater electron migration speed under high electric field intensity, making it also has high application value in microelectronic devices. In the past ten years, wide-bandgap semiconductor materials and devices represented by GaN have developed rapidly, which has played a huge role in promoting the development and application of information science and technology. It is known as the first generation of semiconductors represented by Si and GaAs. Represents the third generation of semiconductors after the second generation of semiconductors.

　　It only took more than twenty years from the first GaN light-emitting diode reported by Pankove in 1971 to the GaN-based blue laser developed by Nakamura. In recent years, research and development on GaN-based materials and devices have been greatly accelerated. Since it is extremely difficult to grow large-size bulk single crystals of GaN, all mature devices are now based on sapphire or SiC heterogeneous substrates. However, from the perspective of lattice matching and electrical and thermal conductivity characteristics, sapphire is not yet an ideal heteroepitaxial substrate. Although the lattice mismatch between SiC substrate and GaN is smaller than that of sapphire substrate, its processing is difficult and expensive. Price also limits further application development of this substrate. Compared with the above two substrates, in addition to the larger lattice mismatch and thermal mismatch, the Si substrate is more in line with the requirements for GaN material growth in other aspects, such as low cost, large size, high quality, conductivity, etc. And the development of Si substrate GaN-based materials and devices will further promote the integration of GaN-based devices and traditional Si-based device processes, and is considered the most promising GaN substrate material. However, because people have focused considerable attention on finding substrates with smaller lattice mismatch in the past, the use of Si substrates has not attracted much interest. With many technological and conceptual breakthroughs, Si substrates The growth of GaN-based materials has increasingly become the focus of attention. Nanchang University in my country was the first to break through the welding and stripping technology of silicon-based GaNLED epitaxial wafers and new substrates. It used the LP-MOCVD system to successfully grow high-quality InGaNMQW blue LED epitaxial wafers on Si (111) substrates, with X-ray twin crystal symmetry. And the half-maximum width of the asymmetric rocking curve has reached the level of sapphire substrate GaNLED on the market.

　　2Epitaxial growth technology

　　The epitaxial technologies that realize the growth of GaN-based materials mainly include metal-organic chemical vapor deposition (MOCVD), molecular beam epitaxy (MBE), hydride vapor phase epitaxy (HVPE), etc.

　　2.1MOCVD

　　MOCVD is a non-equilibrium growth technology that relies on the source gas transfer process and the subsequent thermal cracking reaction of Group III alkyl compounds and Group V hydrides. Both composition and growth rate are determined by gas flows of various compositions and precisely controlled source flow rates. An important feature of MOCVD is that the temperature of the reaction tube wall is much lower than the temperature of the internally heated substrate, which reduces the reaction consumption of the heat tube wall. The growth rate of the MOCVD method is moderate and the film thickness can be controlled more accurately. It is especially suitable for the large-scale industrial production of LEDs and LDs. It has become the most used method with the highest quality growth materials and devices. EMCORE in the United States, AIXTRON in Germany and Thomas Swan in the UK have all developed Group III nitride MOCVD (LP-MOCVD) equipment for industrial production.

　　2.2MBE

　　MBE directly uses Ga molecular beam as the Ga source and NH3 as the N source to react on the substrate surface to generate GaN. This method can achieve the growth of GaN at a lower temperature, potentially reducing the volatilization of N, thereby reducing the background electron concentration. The growth reaction process is simple, and the structure, composition and film thickness of the growth surface can be characterized in real time or accurately monitored. The growth temperature is low and the uniformity is good. Since the growth rate of this method is slow, the film thickness can be precisely controlled, and it is particularly suitable for the growth of materials with ultra-thin layer structures such as quantum wells and superlattices. However, for devices with thicker epitaxial layers, such as LEDs and LDs, the growth It takes a long time and cannot meet the requirements of mass production. Moreover, when plasma-assisted method is used, measures must be taken to avoid damage to the film by high-energy ions.

　　2.3HVPE

　　GaN single crystal thin films were first prepared using this growth technology. Hydride vapor phase epitaxy technology is a chemical vapor phase transport technology. Compared with traditional physical vapor phase transport technology, it can provide a very high growth rate (more than 100 μm per hour) and grow very thick in a short time. GaN film, thereby reducing the impact of thermal mismatch and lattice mismatch on material properties. The stripping technology can be used to separate the obtained thick film with low dislocation density from the substrate, thus becoming a substitute for bulk single crystal GaN wafers. For substrates that use other methods for homoepitaxial growth. The disadvantage of HVPE is that it is difficult to accurately control the film thickness, and the reaction gas is corrosive to the equipment, which affects the improvement of the purity of GaN materials.

　　3 Growth Difficulties and Solutions

　　3.1 Main difficulties

　　First, the lattice mismatch of GaN epitaxially on Si substrate is 17%. The lattice mismatch during the growth process will introduce a large number of dislocations. Secondly, the large thermal expansion coefficient difference between the Si substrate and GaN results in a large thermal mismatch. As can be seen from Table 1, the thermal expansion coefficient of Si is 3.59×10-6K-1, while the thermal expansion coefficient of GaN is 5.59×10-6K-1. There is a big difference between the two, which causes the epitaxial layer to withstand the temperature during the cooling process after high-temperature growth. Due to the large tensile stress, since the thickness of the epitaxial layer is much smaller than the thickness of the substrate, cracks will occur in the epitaxial layer.

　　Another issue is polarity. Since the bonds formed between Si atoms are pure covalent bonds, it is a non-polar semiconductor, while the bonds between GaN atoms are polar bonds, which is a polar semiconductor. The polar/nonpolar heterojunction interface has many physical properties that are different from traditional heterojunction devices, so interface atoms, electronic structure, lattice mismatch, interface charge and dipole moment, band order, transport characteristics, etc. There is a big difference, which is also an issue that must be recognized when studying GaN-based materials and devices on Si substrates.

　　Finally, the diffusion of Si atoms on the Si substrate is also an important issue. The diffusion of Si atoms is intensified during the high-temperature growth process, resulting in a certain amount of Si atoms in the epitaxial layer. These Si atoms are prone to react with ammonia in the growth atmosphere. reaction, and an amorphous SixNy film is formed on the surface of the substrate, reducing the crystal quality of the epitaxial layer. In addition, Ga atoms can also diffuse to the surface of the Si substrate to cause a strong chemical reaction, which will cause melting back of the substrate, destroy the interface, and reduce the crystal quality of the epitaxial layer.

　　3.2 Means of solving problems

　　3.2.1 Selection of buffer layer

　　The introduction of buffer layer technology is an effective means to solve the problems of lattice mismatch, Si diffusion and polarity when growing GaN on Si substrate. It can also relieve the stress in the film to a certain extent. ZnO, 3C-SiC, AlN and AlAs have all been tried as buffer layers between the GaN epitaxial layer and the Si substrate. Among them, AlN has the best result. This is because the AlN buffer layer has better properties on the Si substrate. The wettability can effectively reduce the interface energy, allowing GaN to enter a two-dimensional growth mode from the beginning, which partially alleviates the difficulty of Si-based GaN growth. In addition, AlN can also grow in the same reaction chamber as GaN, and the Al-N bond is formed Better than Si-N bond, inhibiting the formation of SiNx to a certain extent. At present, various research groups have obtained high-quality GaN epitaxial films by optimizing the growth conditions of the AlN buffer layer, such as growth temperature, thickness, V/III ratio, reaction chamber pressure, etc. However, because the GaN material grown under buffer layer technology still has a high defect density, which will affect important technical indicators such as luminous intensity, working life and reverse characteristics of the light-emitting device, people have developed many new technologies on this basis. Buffer layer technology to obtain higher quality GaN single crystal materials.

　　3.2.2 Solving the problem of micro cracks

　　Due to the large thermal mismatch between the Si substrate and the GaN epitaxial layer, the thickness of the crack-free GaN epitaxial layer obtained simply by using the buffer layer is relatively limited (about 1 mm). The micro-crack problem has become the most important factor affecting GaN epitaxial growth on Si substrates. As the main obstacle, various countries have successively carried out research to solve the problem of micro-cracking.

　　Min-HoKimn et al. used the method of gradient composition AlGaN buffer layer, which gradually changes the composition of Al and Ga between the AlN buffer layer and the GaN epitaxial layer, so that there is a gradual transition, which can be used during the growth process. The difference in lattice constants between AlN and GaN creates compressive stress, which partially compensates for the tensile stress formed during the cooling process, thereby effectively reducing the dislocation and crack density in the epitaxial layer. M.Seon et al. proposed the method of superlattice buffer layer, that is, growing the superlattice buffer layer directly on the Si substrate, and then growing the GaN epitaxial layer. In this way, the superlattice layer can alleviate the friction between the substrate and the epitaxial layer. Stress can prevent Si diffusion from the substrate. But the problem is that it is difficult to grow the superlattice layer directly on the Si substrate, so the role of the superlattice buffer layer is weakened. EricFELtin et al. used 10 cycles of AlN/GaN superlattice as an insertion layer between the GaN epitaxial layer and the AlN buffer layer, and grew thicker GaN crystals (0.9-2.5mm) without cracks. In addition to generating additional compressive stress, the use of a superlattice structure can also effectively filter dislocations, especially threading dislocations, and significantly improve the crystal quality of the epitaxial layer. Y. Honda et al. used the Selected Area Epitaxy (SAG) method, taking advantage of the selectivity of GaN growth on the dielectric hidden film and substrate, to limit the growth of the GaN epitaxial layer in the area without the hidden film, forming discrete windows, thereby releasing Tensile stress in the entire epitaxial layer, resulting in a high-quality epitaxial film with a half-maximum width of 388 arcs measured by XRD twins (0004). Dadgar et al. used low-temperature AlN insertion layer technology for the growth of Si-based GaN. This method can effectively control the stress of the GaN epitaxial layer and reduce the dislocation density. Research on the stress compensation effect shows that because the AlN insertion layer is very thin, the amorphous nature of low-temperature deposition will become crystalline when the GaN epitaxial layer is grown at high temperature. The compressive stress introduced during this relaxation process can effectively compensate for the thermal mismatch. of tensile stress. The German A.Krost group used the low-temperature AlN insertion layer method to achieve MOCVD growth of a 3mm microcrack-free GaN layer. The XRD half-maximum width is 400arcs. Recently, they have obtained a 7mm microcrack-free GaN layer, which meets the needs of almost all device applications.

　　Currently, the use of low-temperature AlN insertion layers is considered to be the simplest and most effective method to solve micro-cracks in GaN-based materials on Si substrates. In the actual growth process, several methods can also be used together.

　　Progress of GaN-based devices on 4Si substrate

　　With the gradual development of GaN material growth on Si substrate, various device applications have also been developed, the most important of which is GaN-on-Si substrate LED. In 1998, IBM reported the first use of MBE-grown materials to produce Si substrate ultraviolet and purple GaN light-emitting diodes with wavelengths of 360nm and 420nm respectively. It can be seen from the I-V characteristics of the 300mm×300mm device that its forward operating voltage is very high, greater than 12V, which may be related to p-type doping and p-type contact. In 1999, EMCORE used MOCVD technology to selectively epitaxially grow InGaN/GaN multiple quantum well LED structures on Si substrates. The electrofluorescence spectrum wavelength was 465nm, the half-maximum width was 40nm, the threshold voltage was 4V, and the operating voltage at 20mA was 8V. In 2000, M. Asif Khan and others from the University of South Carolina reported new progress in the selective epitaxial growth of InGaN/GaN multiple quantum well structure LEDs on Si substrates. The peak wavelength of the LED structure is 465nm, the half-maximum width is 40nm, and the forward direction The resistance is 4 times higher than that of the same structure on a sapphire substrate, and the light output intensity is 5 times stronger than that of an LED on a sapphire substrate. Although the results are not ideal, it is a major breakthrough in terms of being suitable for low-cost and large-scale manufacturing. People have seen the dawn of commercialization of Si substrate LEDs. In the same year, Nagoya Institute of Technology in Japan used normal pressure MOCVD technology to grow Si substrate LEDs. The peak wavelength of LED light fluorescence was 430nm and the half-maximum width was 18nm (reported in 2002 to achieve optical power output of 20mW). It can be seen from the I-V characteristics of the LED structure It is found that its threshold voltage is 3V, which is already equivalent to the sapphire substrate LED structure. Recently, A. Krost and others from Germany have made major breakthroughs in the research of high-brightness InGaN/GaN multiple quantum well LEDs. In 2002, they reported the use of low-temperature AlN insertion layers and SiNx insertion layers to achieve LED output powers of 152mW and 400mW. , pioneering high-brightness Si substrate LEDs. In addition, in terms of LD, in 2002, the joint laboratory of Aachen, Minsk and Aixtron realized the optical pump emission of a Si substrate InGaN/GaN multi-quantum well blue laser for the first time. The laser has a wavelength of 447nm and a maximum operating temperature of 420K.

　　In addition to progress in Si substrate GaN-based materials in light-emitting devices, researchers such as E. Calleja of Spain and the University of California, Santa Barbara have also made great progress in field effect tubes and detectors. It is particularly worth mentioning that the American company Nitronex recently realized a high electron mobility transistor (HEMT) on a 100mmSi substrate. They used MOCVD epitaxial technology to grow low-defect density GaN, and the defect density was reduced to 105cm-3. The device is two-dimensional at room temperature. The electron gas mobility is greater than 1600cm2/V·s.

　　The growth technology and device manufacturing process of GaN-based materials on Si substrates have achieved substantial progress and breakthroughs in commercial applications. However, the physical mechanism of the growth process of GaN-based materials on Si substrates is still unclear, and many defects in GaN require further study. In addition, the contact resistance of ohmic contact is relatively large, which affects the performance of the device. It is believed that a series of optoelectronic devices based on the wide bandgap semiconductor GaN will occupy a pivotal position in the semiconductor industry.

Read recommendations:

LR20

What are the characteristics of decomposing ternary lithium batteries?5kwh energy storage 48v lifepo

Graphene battery technology principle.lithium battery energy storage Processing

36v 7.5ah lithium ion battery pack

3.7v 3000mah 18650 battery