#	Pagina
attuale pagina	/open-h2020/projects/217370/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - CoolHEMT (Next Generation GaN Power Amplifiers)

Teaser

Summary

The CoolHEMT project is designed to commercialize SweGaNâ€™s unique gallium nitride (GaN) on silicon carbide (SiC) structure for high frequency devices. SweGaNs technology to growth GaN on SiC enables an extremely thin structure (called QuanFINEÂ®) to be grown which gives several advantages such as higher power density, lower dispersion, and better thermal management. These advantages are critical for a robust and high-performance communication network. In the project, high electron mobility transistor (HEMT) devices will be developed and tested in systems with customers.

During the first year, SweGaN has:
â€¢ Optimized the epitaxial stack and benchmarked devices made on the QuanFINE material against the best commercial material on the market today. The QuanFINE devices manifests a 50% lower dispersion that the devices made on commercial material.
â€¢ A production tool has been designed and installed and material from this tool is being qualified with customers.
â€¢ Devices have been sent to customers for evaluation and feedback has been given back to SweGaN for further optimization of the processing and epitaxial stack.
â€¢ Transistor modelling has commenced and MMIC designs have started.
â€¢ We have presented at several conferences and workshops.
â€¢ We have hired two industrial PhD students.
â€¢ One new patent application has been filed and one trademark has been registered.
â€¢ Sales have increased both within and outside of Europe. Particularly Japan and Taiwan are showing great interest in SweGaNâ€™s material.
â€¢ As added benefit, we have also started to look at the power market due to the outstanding ability of the QuanFINE to handle high electric fields.

Mobile data is growing at an exponential rate requiring completely new technologies to be developed to meet the associated societal and environmental challenges. Gallium nitride (GaN) has been identified as a suitable material to replace silicon (Si) in base stations due to its higher efficiency and linearity. Unfortunately, the technology used today is compromised already at the material level which does not make full use of the advantages GaN can offer. The CoolHEMT project is designed to explore, develop, and commercialize a new structure, QuanFINE, developed by SweGaN which uses ten times less material, generates higher power density, can be used at higher bias, and provides higher efficiency. Used correctly, the QuanFINE can greatly reduce the power consumption in base stations while providing a greater bandwidth. The overall objectives are to optimize the QuanFINE structure, start developing the market and selling the material, develop high electron mobility transistor (HEMT) devices, and generate transistor models and MMIC design which can be tested by end users.

Work performed

The work during the first year has been focused on improving the performance of the QuanFINE, set up a pilot production process for growing epitaxial layers and processing HEMT transistors, and commercializing the material. Several improvements have been developed on the epitaxial stack resulting in a large reduction in dispersion as measured by the knee walkout. However, new unforeseen issues in the supply chain have been encountered. Previously semi insulating (SI) silicon carbide (SiC) substrates of high quality could be obtained in Europe, however this is no longer possible which led to the need to purchase from USA and China. The quality of the substrates has not been satisfactory and required us to modify our process. The price of the SI SiC substrates is furthermore increasing which required us to start developing our own substrate manufacture out of necessity. The project has hence been amended to include substrate development. Another aspect we discovered during the project was that the QuanFINE defied previous assumptions and, despite it only being a few 100 nm thick, could block extremely high voltages which led to a slight adjustment of our focus within the project. The commercialization has gone well in terms of creating interest around our material. We have more customers and most customers are repeat customers as well. Fiscal year 2019 ended September 30th was a record year for SweGaN in terms of sales to customers. Most interesting is the Asian market where we have sold material to several customers in Japan. We could have sold significant amounts of material to Taiwan, but we are now still waiting for the answer for our applications from the Swedish Inspectorate for Strategic Products (ISP), which due to internal timings, has delayed this opportunity.

Final results

The QuanFINE outperforms the best material in the world and has done so directly from start. However, as mentioned in the previous section, work is ongoing to improve the performance. Parallel processing with the best material that can be purchased today shows that QuanFINE gives about 10% higher power density, about 30% (still uncertain) lower thermal resistance, and about 10% lower knee walkout. The improved version (version 2) of the QuanFINE gives 50% lower knee walkout. The dispersion is something we continue to look carefully at and a version 3 will soon be launched with even better performance. The impact this has is that more efficient base stations can be built and thanks to the high thermal dissipation, it will be easier to realize massive MIMO which is an essential technology for an energy efficient 5g network; A major challenge to realize massive MIMO is the cooling. The high efficiency and low dispersion will also reduce the energy consumption in the base stations. What has come as a surprise for us is the very high voltage blocking capability the material manifests. There is a lot of hype around GaN power devices and enormous amounts of money is pumped into researching GaN on Si which is proving to be very unreliable and hard to push beyond 600 V. The higher the voltage, the thicker the layer needs to be which is creating huge difficulties for GaN-on-Si. The QuanFINE, by contrast, is extremely thin and should only be able to block 30 V but when tested, it blocked 1.5 kV. This measurement has been repeated several times by different groups and blocking voltages in excess of 2 kV have been demonstrated. The impressive thing is not so much the peak voltage but the peak electric field. The higher the peak electric field can be, the closer the contacts can be placed on the HEMT structure and the lower the on resistance and higher the frequency. The maximum field measured so far is close to 2 MV/cm which should be compared to the best GaN on Si which only reaches around 0,7 MV/cm. With a contact spacing of only 5 micrometer, the QuanFINE can hence block close to 1 KV. This leads to one order of magnitude lower on resistance compared to GaN on Si or SiC on SiC power devices. The impact this has is that an electric vehicle can for instance get around 20% longer range and charging the battery can be done faster. The substantially higher switching frequency can greatly increase the power density of the system making laptop and phone chargers much smaller than today and more efficient at the same time.