Since Prof. Liang Jianbo joined Osaka City University in 2012 (in 2020, Osaka City University and Osaka Prefectural University merged and changed its name to Osaka Public University, Japan).
From 2013 to 2016, Liang Jianbo's research team successfully fabricated high-efficiency and low-cost InGaP/GaAs/Si tandem solar cells through the normal temperature bonding of dissimilar materials, and since 2018, Liang Jianbo's research team has made a lot of work and achievements in the field of normal temperature bonding of diamond and GaN.
Presently. In recent years, he has presided over a number of R&D projects, including 12 national key R&D projects funded by the Japan Society for the Promotion of Science (JSPC), the National Agency for Advanced Energy and Industrial Technology (NEDO), and the Japan Science and Technology Agency (JST), as well as corporate collaborative R&D projects.
At the same time, it has carried out a number of projects and technical cooperation with more than 10 well-known scientific research institutes at home and abroad, and promoted international scientific research exchanges. As the first author, corresponding author or mentor student, he has published a number of publications in internationally renowned journals such as "Adv. Mater." He has published more than 150 papers, applied for 12 patents, and authored 8 monographs. He has presented at 41 international conferences and has been invited to give presentations at 14 international conferences. He has won many awards for best publication at international conferences, and has been recognized for many awards such as the Outstanding Research Award from Yoichiro Minami (Nobel Laureate in Physics) at Osaka City University and the Outstanding Reviewer Award for prestigious journals. He currently serves on the editorial boards of the journals "Functional Diamond" and "Science Talks", as well as "Adv. Mater." ，"ACS Appl. Mater. Inter." ，"ACS Nano Lett." , reviewer of 13 international journals such as "Appl. Phys. Lett".
"Bonding of diamond and GaN at room temperature" is one of his main research directions in recent years.
。 然而，晶片直接键合技术要求键合材料具有非常高的表面平整度。 随后，该团队对大面积键合、界面热传导特性评估、直接与金刚石键合的GaN层晶体管的试制、实用散热演示等研究进行详细开发，并取得系列成果。 （Fabrication of GaN/Diamond Heterointerface and Interfacial Chemical Bonding State for Highly Efficient Device Design, Adv. Mater. 33, 2104564 （2021））
Optical image of the GaN/Diamond bonded sample and cross-sectional diagram of the bonded sample
The thickness of the SiC layer formed after heat treatment at 1000 °C increases slightly, because SiC is generated from free Si and C in the SiC layer, and the amorphous SiC becomes polycrystalline after heat treatment. No voids were observed at the bonding interface even after high temperature treatment, indicating that the diamond-SiC bonding interface has excellent thermal stability. The results show that the deposition of SiC layer can reduce the requirement for diamond surface roughness and promote the room-temperature bonding of polycrystalline diamond and semiconductor materials.
相关的成果以“Room-temperature bonding of GaN and Diamond via a SiC layer”为题，发表在《Functional Diamond》杂志上。
The research team found that the technology more than tripled the heat dissipation performance compared to the same shape transistors fabricated on SiC substrates. To maximize the high thermal conductivity of diamond, the researchers integrated a 3C-SiC layer, a cubic polymorphic SiC, between GaN and diamond.
The bonding interface exhibits exceptional robustness and is able to withstand a wide range of device manufacturing processes. Spalling of the 3C-SiC/diamond bonding interface was not observed even after annealing at 1100°C, which is essential for high-quality GaN crystal growth on diamond. The thermal boundary conductivity at the 3C SiC/diamond interface is measured at 119 W/m2∙K, which greatly reduces the thermal resistance of the interface and improves heat dissipation.
相关研究成果以“High Thermal Stability and Low Thermal Resistance of Large Area GaN/3C-SiC/Diamond Junctions for Practical Device Processes”为题发表于《Small》。
"This new technology has the potential to significantly reduce CO2 emissions and potentially revolutionize the development of power and RF electronics through improved thermal management capabilities. ”
GaN HEMTs enable power electronics systems to move toward higher efficiency and power density.
So, what is the practical significance of the diamond-gallium nitride bonding they are studying? This has to mention the significance of the existence of GaN devices.
Generally, semiconductors with a bandgap width greater than 2 eV are called wide bandgap semiconductors, also known as third-generation semiconductors.
As a third-generation semiconductor material, gallium nitride (GaN) has excellent material characteristics, such as large band gap, high breakdown field strength, and high electron saturation drift rate. Due to the presence of high-density two-dimensional electron gas (2DEG) at the AlGaN/GaN heterojunction interface, GaN HEMT has the advantages of high electron mobility, high temperature resistance, high pressure resistance, and strong radiation resistance, and can obtain higher operating capacity with less power consumption. These characteristics enable electrical and electronic systems to move towards higher efficiency and power density.
The heat dissipation problem is one of the main technical bottlenecks restricting the further development and wide application of GaN-based power devices.
In high-power operation, the output power of GaN devices is limited by self-heating, and the power density is often only 8-10 W·mm-1. In addition, the performance and reliability of GaN is related to the temperature on the channel and the Joule heating effect. In particular, in recent years, with the continuous improvement and improvement of the design and process of GaN microwave power devices, their theoretical output power has become higher and higher (4 GHz, ~40 W/mm), the frequency has become larger and smaller, and their reliability and stability have been seriously challenged. Therefore, due to the limitations of traditional packaging heat dissipation technology, this problem cannot be solved, and it is necessary to improve the heat transfer capacity of GaN devices from the proximal junction hot region, so it has become an important direction to further promote the development of GaN-based devices by exploring the issue of efficient heat dissipation of GaN-based devices.
Diamond has gradually become the first choice for heat sink materials for GaN devices.
GaN power devices are generally prepared on Si substrates, and the low thermal conductivity of the original substrate (Si: 150 W·mK-1) cannot meet the requirements of device heat dissipation, resulting in serious degradation of device performance, which greatly limits the application of GaN-based power devices. Substrates such as SiC and diamond integrated into GaN can improve thermal management. This makes it possible to reduce the operating temperature of the device. For diamond substrates with higher thermal conductivity, the thermal conductivity is 14 times higher than that of silicon, and the electric field resistance is 30 times that of silicon. High thermal conductivity allows heat to diffuse. The band gap of diamond is 5.47 ev, the breakdown field strength is 10 mv/cm, the electron mobility is 2200 cm2, and the thermal conductivity is about 21 w/cm·K. In addition, compared with other similar structures, the thermal resistance reduction rate of GaN HEMT on diamond substrate is the most significant compared with SiC, and diamond has gradually become the preferred heat sink material for GaN devices.
Various substrate materials and common properties of GaN
How to integrate diamond with GaN?
The use of diamond can effectively improve the heat dissipation capacity of the near-junction region of GaN devices, reduce the peak temperature, and greatly improve the reliability of the device. However, how to integrate diamond with GaN has become a difficult point. The integrated methods of diamond and GaN that have been reported so far can be divided into three main categories:
Due to the large lattice mismatch between the two materials and the difference in thermal expansion coefficient, it is quite difficult to grow high-quality GaN on diamond, and the GaN after growing diamond is easy to form high-density dislocations or even rupture.
The growth of diamond on GaN often needs to be carried out at 800°C or even higher temperatures, and the high-temperature process is easy to cause wafer warping and cracking, and a dielectric layer needs to be deposited before depositing diamond, which will cause poor quality of diamond nucleation and low thermal conductivity of the nucleation layer, which makes the interface thermal resistance high.
The bonding of diamond and GaN is a parallel process, and the GaN layer and diamond substrate can be prepared separately, and the low-temperature bonding technology can avoid the difference in lattice mismatch and thermal expansion coefficient caused by high-temperature growth, and there is no need to consider the low thermal conductivity of the nucleation layer. Choosing a better quality diamond with higher thermal conductivity as the bonding material can maximize the heat dissipation capacity.
Osaka Team: GaN On Diamond, diamond is GaN's best friend
GaN-on-Diamond shows promise as a next-generation semiconductor material because both materials have a wide bandgap width that enables high electrical conductivity and high thermal conductivity of diamond, positioning them as superior heat dissipation substrates. Researchers have tried to create a GaN-on-Diamond structure by combining GaN and diamond with some form of transition or adhesion layer, but in both cases, a technology that can directly integrate diamond and gallium nitride is needed. However, due to the large differences in their crystal structures and lattice constants, it is extremely difficult to grow diamond directly on GaN or grow GaN on diamond.
"Therefore, there is a need for a technology that can directly integrate Diamond and GaN," explains Professor Liang Jianbo.
Fusing two components together without any intermediate layers, known as wafer direct bonding, is one way to solve this mismatch. However, in order to produce sufficiently high bond strength, many direct bonding methods require the structure to be heated to extremely high temperatures in a process known as post-annealing. This often leads to cracks in bonded samples of different materials due to thermal expansion mismatch – GaN-Diamond structures cannot survive the extremely high temperatures experienced during manufacturing.
In 2021, Prof. Jianbo Liang's research team successfully fabricated various interfaces with Diamond at room temperature using surface-activated bonding (SAB), all of which exhibited high thermal stability and excellent practicability. It is understood that the SAB method establishes a highly strong bond between different materials at room temperature by cleaning and activating the bonding surfaces to react when they come into contact with each other by atomic cleaning.
Since the chemical properties of GaN are completely different from the materials used by the Osaka research team in the past, in order to characterize the residual stresses in GaN at the heterogeneous interface, they used micro-Raman spectroscopy, transmission electron microscopy (TEM) and energy-dispersive X-ray spectroscopy to reveal the nanostructure and atomic behavior of the heterogeneous interface, electron energy loss spectroscopy (EELS) showing the chemical bonding state of the carbon atoms at the heterogeneous interface, and tested the thermal stability of the heterogeneous interface at 700 degrees Celsius in N2 gas ambient pressure," This is necessary for the GaN-based power device manufacturing process," said Professor Liang Jianbo in a past interview.
"The results show that an intermediate layer of about 5.3 nm is formed at the heterogeneous interface, which is a mixture of amorphous carbon and Diamond, in which Ga and N atoms are distributed. As the team increased the annealing temperature, we noticed a decrease in the thickness of the middle layer due to the direct conversion of amorphous carbon to diamond. After annealing at 1000 degrees Celsius, the layer is reduced to 1.5 nm, which indicates that the middle layer can be completely removed by optimizing the annealing process," explains Professor Liang Jianbo. "Although the compressive strength figures at heterogeneous interfaces increase with the annealing temperature, they do not match the compressive strength of the GaN structure on the diamond formed by crystal growth. Since no spalling was observed at the heterojunction after annealing at 1000 degrees Celsius, these results suggest that the GaN/diamond heterojunction is able to withstand the rigors of the manufacturing process, and the temperature rise of the GaN transistor is suppressed by a factor of four. ”
In the latest research progress, Prof. Jianbo Liang's team has successfully transferred the AlGaN/GaN/3C-SiC layer from silicon to a large diamond substrate, and fabricated GaN high electron mobility transistors (HEMTs) on diamond. Notably, the 3C-SiC/diamond bonding interface did not peel off after annealing at a high temperature of 1100°C, which is essential for high-quality growth of GaN crystals on diamond.
The AlGaN/GaN/3C-SiC layer bonded to the diamond undergoes tensile stress, which is released as the annealing temperature increases. Compared to GaN HEMTs on silicon and SiC substrates, GaN HEMTs on diamond substrates exhibit the highest drain currents and lowest surface temperatures. In addition, the thermal resistance of GaN HEMTs on diamond substrates is less than half that of SiC and about a quarter of that of Si. These results show that
Large-scale commercialization, or not far away
Based on the industry's long-term R&D activities, the functional applications related to diamond semiconductors have begun to be gradually put into practical use. However, it still takes a long time to truly popularize the application of diamond in the semiconductor field.
Source: DT Semiconductor
*Disclaimer: This article was originally written by the author. The content of the article is the author's personal point of view, and the wide bandgap semiconductor technology innovation alliance is reprinted only to convey a different point of view, and does not mean that the alliance agrees with or supports the view, if you have any objections, please contact us.