release time:2022-03-17Author source:SlkorBrowse:751
author: Gregg H. Jessen
In the May 2002 issue of IEEE SPECTRUM, the Chinese version of Science and Technology Survey, the late Lester F. Hysmans and Umesh K.Mishra talked about a bold technology in the power semiconductor industry at that time: gallium nitride (GaN). They expressed optimistic views on the application prospect of powerful and durable RF amplifiers in the emerging broadband wireless network, radar and power switching applications in power grid. They call gallium nitride devices "by far the most robust transistor".Hysmans and Mishra are right. The wide band gap of gallium nitride (the energy that breaks the bound electrons freely and contributes to conduction) and other properties enable us to make use of this material's ability to withstand high electric fields to manufacture devices with unprecedented performance. Today, gallium nitride is the undisputed champion in the field of solid-state RF power applications. It has been applied in radar and 5G wireless technology, and will soon be popularized in inverters of electric vehicles. You can even buy USB wall chargers based on gallium nitride. They are small and have very high power. But is there anything better than it? Is there a device that can make the RF amplifier more powerful and efficient? Is there a device that can make the volume of power electronic equipment smaller and the electronic equipment used in airplanes and cars lighter and smaller? Can we find conductive materials with larger band gap? Yes, we can. In fact, many materials have larger band gaps, but the uniqueness of quantum mechanics means that almost all these materials cannot be used as semiconductors. However, the transparent conductive oxide gallium oxide (Ga2O3) is a special case. The band gap of this crystal is nearly 5 electron volts. If the gap between gallium nitride (3.4eV) and it is 1 mile, the gap between silicon (1.1eV) and it is like a marathon. And diamond and aluminum nitride have larger band gaps, but they don't have the lucky characteristics of gallium oxide, which helps to manufacture inexpensive but powerful devices. It is not enough for a material to have a wide band gap. All dielectrics and ceramics have wide band gaps, otherwise they won't be used as insulators. Gallium oxide has a set of unique characteristics, and it can play a huge role as a semiconductor for power switching and RF electronic devices. One of its characteristics is that by doping, charge carriers can be added to gallium oxide to make it more conductive. Doping includes adding a certain amount of impurities to the crystal to control the concentration of carriers in the semiconductor. For silicon, ion implantation and annealing can be used to dope phosphorus (to add free electrons) or boron (to subtract free electrons) in the crystal, so that the charge can move freely. For gallium oxide, silicon can be doped in the crystal in the same way to add electrons. If this is done in any other wide band gap oxide, the result may be broken crystals and lattice spots, in which case the charge will be stuck. Gallium oxide can adapt to dopants added by the standard process of "ion implantation" and during epitaxial growth (deposition of extra crystals), so we can borrow various existing commercial lithography and processing technologies. With these methods, it is relatively simple to accurately define the transistor size of tens of nanometers and generate various device topologies. Other wide band gap semiconductor materials do not have this incredibly useful characteristic, even gallium nitride is no exception. Another advantage of gallium oxide is that it is actually easy to manufacture large wafers of gallium oxide crystals as needed. Although there are several types of gallium oxide crystals, the most stable one is β, followed by ε and α. Among them, the comprehensive properties of β -gallium oxide are the most studied, which is mainly due to the pioneering work of Japan National Institute of Materials Science in Tsukuba, Japan and Leibniz Crystal Research Institute in Berlin. What's particularly interesting about β -gallium oxide is that it has good thermal stability, so it can be manufactured using a large number of commercial technologies, including the Czochralski method for manufacturing silicon wafers. In addition, the technology of "edge definition and thin film transistor growth" can also be used to produce β-gallium oxide wafers, and the sapphire window on the bar code scanner used in grocery stores is made in this way. Nowadays, even the highly scalable Bridgman-Stockbarger technique can be used to grow crystals.
The difference between this and other wide bandgap semiconductors cannot be overstated. Except silicon carbide (SiC), all other emerging wide bandgap semiconductors have no large-size semiconductor substrate for growing large crystals. This means that they have to grow in another material disk, and this comes at a price. For example, gallium nitride is usually grown on silicon, silicon carbide or sapphire substrates by complex processes. However, the crystal structure of these substrates is obviously different from that of gallium nitride, and this difference will lead to "lattice mismatch" between the substrates and gallium nitride, resulting in a large number of defects. These defects will bring a series of problems to the produced equipment. Gallium oxide serves as its own substrate, so there is no mismatch, so there is no defect. Novi Jingke Technology Co., Ltd. of Saitama has developed a 150mm beta gallium oxide wafer.
Masataka Higashiwaki of Japan National Institute of Information and Communication Technology (NICT, located in Tokyo) was the first person to discover the potential of β -gallium oxide in power switches. In 2012, his team reported the first single crystal β -gallium oxide transistor, which shocked the whole semiconductor device industry. This is a device called "Metal Semiconductor Field Effect Transistor". How good is it? Breakdown voltage is one of the key indicators of power transistors. When reaching this critical point, the semiconductor's ability to prevent current flow will collapse. The breakdown voltage of the pioneering transistor researched by Tohoku is more than 250 volts. In contrast, it took nearly 20 years for gallium nitride to reach this level.
ery high Ec is very important for an ideal power switching transistor. Ideally, the device will instantly switch between two states: always conducting and conducting without resistance; Always disconnected, in a completely nonconductive state. These two impossible extremes mean two completely different device geometries. For the off state, there needs to be a thicker material area between the source and drain of the transistor to prevent conduction and large voltage. For the conducting state, an infinitely thin area is needed to make it have no resistance.
However, is gallium oxide useful in faster switching power supply applications? Ec is also important here, which may bring great advantages to gallium oxide. At higher frequencies, such as 100 kHz to 1 MHz, compared with the on or off state, the switching time of the device will increase proportionally. The loss in the switching process is equal to the product of the resistance of the device and the accumulated charge on the transistor gate when the switch is switched. Mathematically, this means that the loss is proportional to the square of the critical electric field strength, not to the cube (just like at low frequency).
In 2015, when measuring the Ec of power switches, we also speculated that gallium oxide might achieve similar success in RF circuits by allowing higher electric fields in smaller devices. However, at that time, we lacked a key piece of information, that is, there was no public data about the functional relationship between electron velocity and electric field in materials.
So, what are the disadvantages of gallium oxide? The achilles heel of this material lies in its poor thermal conductivity. In fact, among all semiconductors that can be used for RF amplification or power switching, its thermal conductivity is the worst. The thermal conductivity of gallium oxide is only 1/60 of diamond, 1/10 of silicon carbide (the base of high-performance radio-frequency gallium nitride) and about 1/5 of silicon. (Interestingly, it is comparable to gallium arsenide, the main RF material. Low thermal conductivity means that the heat generated in the transistor may stay, which may greatly limit the lifetime of the device.
Another more basic problem is that we can only make gallium oxide conduct electrons, but not holes. No one has ever made a good P-type conductor with gallium oxide. In addition, it is frustrating that the basic electronic characteristics of this material make it hopeless in this respect. In particular, the valence band part of the energy band structure of this material does not have the shape of hole conduction. Therefore, even if there is a dopant that can make the acceptor at the correct energy level, the generated holes will trap themselves before it helps conduction. When the theory is so consistent with the data, it is difficult to find a solution to this problem.
Although this weakness does bring more challenges, it is not an obstacle. Many so-called devices limited to most operators have also achieved commercial success, such as USB-C wall charger. The research stage of gallium oxide device technology has just begun to reach a critical scale. We are planning the application space of fast switches, multi-kV power transistors and RF devices. Nowadays, new kilovolt devices are often developed. Radio frequency transistors with critical dimensions of tens of nanometers are coming out soon. I very much hope that with the development of this technology, we can realize the device topology that was previously impossible in any other material.
Of course, we will break some things (mainly dielectrics) on the way to development, but this is the definition of disruptive technology. We exchange what we know for potential performance, but at present, the performance potential of gallium oxide is far greater than its problems.
Disclaimer: This article is reproduced from "EETOP". This article only represents the author's personal views, and does not represent the views of Sacco Micro and the industry. It is only for reprinting and sharing to support the protection of intellectual property rights. Please indicate the original source and author when reprinting. If there is any infringement, please contact us to delete it.
Company Tel: +86-0755-83044319
QQ: 3518641314 Manager Li
QQ: 332496225 Manager Qiu
Address: Room 809, Block C, Zhantao Technology Building, No.1079 Minzhi Avenue, Longhua New District, Shenzhen
links exchange: SiteMap
Copyright ©2015-2022 Shenzhen SlkorMicro Semicon Co.,Ltd. All rights reserved
Hall Effect Sensor
Get product information