【Expert Interview】Prof. Tashun Chou of IKZ: Supporting Gallium Oxide Industrialization by High-Quality Epitaxial Growth

Expert Profile

      He is currently working as an independent researcher in the Gallium Oxide Growth Laboratory at the Leibniz Institute for Crystal Research, Germany. Dr. Tashun Chou received his PhD in chemistry from the Technical University of Berlin, Germany. His research interests include Gallium Oxide film process development, crystal growth theory, application of machine learning to semiconductor material growth, and real-time growth monitoring algorithm development. He has published nearly 30 papers in prestigious journals such as ACS Nano, Applied Physics Letter, Applied Surface Science and APL Materials, as well as several international conference papers and international patents. Received the Young Scientist of the Year Award from the German Association for Crystal Growth, and served as a reviewer for journals such as Advanced Materials, Applied Physics Letter and Journal of Crystal Growth.

AGOA:What made you start studying Gallium Oxide?

      Before my PhD (2020), I actually knew nothing about Gallium Oxide, and before that I was in the OLED field. I officially started my research on Gallium Oxide in 2020 when I joined the Leibniz Institute for Crystal Growth (IKZ) in Berlin, Germany. There were quite a few topics available for me to choose from at the time, but only Gallium Oxide had obvious industrial potential, and IKZ was one of the leaders in this material, so I quickly decided to join the current research group, until now.

AGOA:Please briefly introduce the team, as well as the current research direction, progress, and achievements.

      Our research focuses on Gallium Oxide MOVPE preparation process, especially focus on the growth on the (100) plane. Since 2012, our research group has been exploring how to grow high-quality films on (100) plane, and looking for solutions from the substrate surface engineering and preparation process. Due to the potential of Gallium Oxide in high-power components, we are currently focusing on the growth of high-thickness Gallium Oxide films. In addition to process improvement, we also hope to understand the morphological changes and defect formation during film growth from the mechanism. In addition, we also hope to reduce engineer's execution costs and time costs through machine learning and algorithm development, automated debugging and real-time monitoring of the actual process.

      In terms of film growth, we observed the instability of step-flow morphology. With the increase of thickness and impurity concentration, the film morphology would gradually turn to 3D-island growth and produce a large number of structural defect. This phenomenon can be explained by the well-known non-steady state of Bales-Zwangwil, which is caused by the Schwobel energy barrier on the step edge. By adjusting the effective concentration of Ga adatoms on the surface, this instability can be greatly reduced, so that the step flow of (100) plane can be maintained at different thickths and impurity concentrations, and thus Gallium Oxide film with a µm-level thickness and high electron mobility (>160 cm ² /V) can be used for component fabrication. In addition, we also solved the problem of the formation of suspended particles with the increase of film thickness in the MOVPE process, which greatly improved the process yield.

      In addition to the general material growth, we were forced to work at home for nearly a year due to the epidemic, which prompted us to conduct research on material process using machine learning and develop an optical measurement algorithm that can monitor the growth process in real-time. By training the random forest model, our team systematically analyzed the growth data accumulated in the past ten years, summarized the effects of different parameters and their combinations on the properties of the film, deepened the understanding of the growth system and established an automated quality management system. We then transplanted the machine learning concept to MOVPE's real-time monitoring. Through a combination of hardware and software (machine learning algorithm coupled with Laytec EpiTT optical monitoring equipment), the researchers were able to monitor the homoepitaxy growth rate, morphology changes, and even the impurity concentrations in real time during the growth process, effectively reducing the time and labor cost of the process that characteristics emerge.

AGOA:According to the research direction of your team, what are the current research difficulties that need to be solved? Could you please describe the main research directions you are currently focusing on?

      The growth of (100) plane is a rather complex subject, and the thermodynamic balance of substrate surface treatment and film growth should be considered simultaneously. Especially in the growth of thick films (above µm), the instability of morphology and the inhibition of suspended particle formation are two major problems. Since the (100) plane depends on the chamfer of the substrate surface and step flow morphology to reduce the occurrence of defect, it is extremely critical to control the morphology stability of the (100) plane. At present, there is little understanding of the growth mechanism of Gallium Oxide films in the field, and the control of material growth still depends on experimental experience and observation. Our research group is actively cooperating with other experts with simulation expertise, hoping to understand the microscopic mechanism of the morphology change of Gallium Oxide through Monte Carlo and molecular dynamics. On the basis of (100) plane, we also hope to study the growth of other aspects (such as (001) and (-201)).

      In addition, in the MOVPE process, with the increase of growth thickness, a large number of suspended particles will be produced in the reactor and adhere to the surface of the film, resulting in defect. This is a difficult problem in the field. Our research group hopes to suppress the formation of suspended particles through simulation and engineering means, so that the quality of Gallium Oxide films will not decrease with the increase of thickness, and thus increase the yield and reliability of components.

AGOA:What are the effects of choosing different substrate materials for the epitaxial growth of Gallium Oxide in your study?

      My research focuses on homoepitaxy growth, with particular emphasis on preparation process development in all aspects. Because of the difference of surface structure, the influence of different surface morphology on the quality of the film is different. In the process of growth, the process parameters need to be adjusted according to different aspects to ensure that the ideal surface topography and electronic properties can be achieved. For example, compared with (100) plane, the growth window of (010) plane is quite wide. If the growth parameters of (010) plane are directly applied, it will be found that the film on (100) plane cannot form ideal step flow morphology in most parameter intervals.

AGOA:Please briefly introduce several mainstream epitaxy methods of Gallium Oxide at present, what are the advantages and disadvantages of each of these methods?

      At present, there are three main epitaxy methods (limited to Beta phase) : MBE, HVPE, and MOVPE(MOCVD).

      MBE has long been the preferred growth method for new materials because of its ability to achieve single-layer growth control and relatively pure (ultra-high vacuum) environment. The early studies of Gallium Oxide films were mostly grown by MBE method. One of the disadvantages of MBE method is the thickness limitation of the film (surface Ga accumulation makes it difficult to reach more than µm) and the inability of continuous impurity accumulation (which has been solved in recent years). In addition, the electron mobility of the films published in the paper is always lower than that of the other two epitaxy methods.

      The HVPE method has previously been widely used for GaN growth and is known for its extremely high growth rate (>10 µm/hr). In the application of Gallium Oxide, it is also commonly used for the growth of µm thick films, in addition, because the precursor does not contain carbon (halide is the main), it is believed that it can greatly reduce the carbon residue in the film and improve the electronic performance. However, the output of single-machine epitaxial wafer at present is not large, and the thickness and impurity uniformity control capability of large-size epitaxial wafer still need to be improved.

      MOVPE method is the mainstream way of growing compound semiconductors in the industry, also has been a lot of the team for the growth of Gallium Oxide film. The current mobility record for Gallium Oxide (~200 cm²/V) is from the epitaxial wafer by the MOVPE method. However, the precursor in the process contains a large number of carbon groups, which may leave a large carbon residue in the film and reduce the electronic properties.

AGOA:In your opinion, how does the thickness of Gallium Oxide film affect the performance of optical, high power and other devices?

      As the target application of Gallium Oxide is in high power components, research interest in the field has gradually shifted to vertical component structures. Therefore, in order for the component to withstand higher voltages, the thickness of the drift layer (the thicker the better) is critical. At present, the epitaxial research groups and manufacturers both domestic and international focus on the growth of high-quality thick films (more than 10 µm).

AGOA:Which epitaxy process do you think is more suitable for the development of Gallium Oxide industrialization, and what are your reasons?

      Under the premise of the same film quality, the growth rate and single machine yield are the key to industrialization. Under this requirement, both HVPE method and MOVPE method have great potential. Both are known for their high growth rates, but the HVPE method is still unable to compare with the MOVPE method in the short term in terms of single machine production. However, in recent years, equipment manufacturers have increased the research and development of Gallium Oxide growth machines, and it is believed that HVPE method and MOVPE are expected to become the mainstream Gallium Oxide epitaxy process in the near future.

AGOA:The quality of substrate will have a great impact on the growth of epitaxy, so could you please give us some advice on the growth and processing of substrates?

      At present, the number of defects of Gallium Oxide substrate needs to be reduced. Taking EFG method as an example, recent studies have shown that many nanopipe defect will be generated due to the residual bubbles inside the crystal during substrate growth. Such defect will penetrate the film (especially at (010) plane) with epitaxial growth, and many Hillock-like structures will be formed on the surface of the film. Thus, the electron mobility of the film and the yield of component preparation are greatly reduced. The improvement of substrate process is a key factor for the take-off of Gallium Oxide (next are size and price) and it still needs to be improved by domestic and international manufacturer partners.

AGOA:What kind of cooperation between colleges and enterprises is more conducive to the industrialization of Gallium Oxide?

      In order to establish more effective communication channels between colleges and enterprises, many research and development problems of enterprises should be solved through the research group of colleges, and the focus of enterprises should be on mass production process and quality improvement. On the other hand, we should also encourage more "scientists entrepreneurship", so that excellent scientific research results can be realized, but also strengthen the practice of scientific research projects.

AGOA:What kind of services do you expect alliance to provide?

      I hope that the alliance can become a broad and efficient communication platform, with more conferences held to enable academic and industrial professionals to quickly exchange achievements and difficulties, and jointly promote the development of Gallium Oxide technology.