The Courage Not to Fear Failure Leads to a Successful Innovation (Part I)

[The Power to Create the Future Vol. 6] A Kyoto Prize 2017 laureate, Dr. Takashi Mimura, the inventor of HEMT

Ultra-Fast HEMT Transistors Support the Foundation of This ICT Society

Have you ever heard of the High-Electron-Mobility Transistor (HEMT)?

It is an ultra-fast transistor (semiconductor device) that was invented by Dr. Takashi Mimura at Fujitsu Laboratories Ltd. (currently an honorary fellow) in 1979 and is now crucial to devices used in this modern-day ICT society, including cell phones, car navigation systems, automotive radars and satellite dishes.

Recently, Dr. Mimura was selected as a laureate of the 33rd Kyoto Prize, which is regarded as a stepping stone to the Nobel Prize, and his HEMT is attracting much attention.

This article traces the historic innovation of HEMT, which was invented at Fujitsu about 40 years ago. Dr. Mimura talked about behind-the-scenes secrets and hints regarding this innovation.

A Double-Layered Structure That Separates Electrons and Donors Results in the Greatest Increase in Speed

Dr. Mimura used diagrams to illustrate how HEMTs work. He also explained the meanings of some difficult terms.

"Semiconductors are a type of material that conducts electricity under specific conditions and function in a variety of ways depending on the material devices and their combination. My invention, HEMT, is a type of transistor that amplifies current. Conventional transistors had the problem of inevitable slow operation that resulted from obstacles (donors: dopant atoms) existing on the electronic path as a necessary evil. In contrast, HEMTs use a double-layered structure that spatially separates electrons and donors, which is to say having an expressway exclusively for electrons. This allows electrons to move quickly without being hindered by obstacles, making it possible to amplify weak high-frequency signals."

"Columbus' Egg" Idea Surprises Researchers Around the World

This is a notebook on which Dr. Mimura noted experiment data for papers. Every detail was noted by hand.

The principle of operation of HEMTs is "extremely easy to understand for anyone with basic knowledge of semiconductors." This "Columbus' egg" idea had an impact on researchers around the world at that time.

"The reason why I was the only one that could reach the idea of HEMT is attributable to my experience in the attempted development of the gallium-arsenide metal-oxide-semiconductor field-effect transistor (MOSFET)*. The development resulted in failure, which in turn led to the creation of HEMT."

*Development of GaAs MOSFET :
An attempt to create a MOSFET, or a type of transistor with a thin layer of silicon dioxide on the silicon substrate, using gallium-arsenide instead of silicon, in anticipation of achieving higher speed and performance.

Failure Makes Dr. Mimura More Receptive to Various Types of Information, Allows Him to Find Hints for Development

What does this mean exactly: "failure creates innovation?"

"I had worked on MOSFET for two years but I could never succeed in accumulating electrons; I really didn't have an idea regarding practical applications. When I felt that I ran out of ideas and should give up the research, I was strongly shocked to learn of the technology known as the modulation doped superlattice structure*, which provided hints for the HEMT's structure.
Even though this technology was from an unfamiliar field for me, the study of MOSFETs made me always think about accumulating electrons and be very receptive to related information. I think this is why I could find this technology."

It was soon after finding the modulation doped superlattice structure that Dr. Mimura came up with the idea of HEMT. According to him, the idea popped into his head in the middle of chatting with a researcher.
"My mind became clear when I presented at an academic meeting all the findings of the MOSFET study, thinking that I had done everything I could do and should conduct another study. Then the idea of HEMTs swept through my cleared mind, so to speak. Ideas may pop up when the mind is free, not when feeling pressure."

*Modulation doped superlattice structure :
It was discovered that stacking dozens of gallium-arsenide (GaAs) and aluminum-gallium-arsenide (AlGaAs) layers alternately and moving electrons in parallel causes the electrons to accumulate on the non-doped GaAs layers. Bell Labs in the US published a paper on this in 1979.

Dr. Mimura Starts Secret Research and Development Project with In-House Volunteers

Dr. Mimura took detailed notes during meetings with the development team members. He always included the date.

After several weeks, Dr. Mimura drew up clear conceptual diagrams of HEMT and formed a development team with in-house researchers, whom he invited to join. He needed cooperative volunteers to continue his research because HEMTs were not approved as a formal research theme at the company.

"The development of HEMTs required an advanced crystallization device called MBE for sharpening the junction of gallium-arsenide (GaAs) and aluminum-gallium-arsenide (AlGaAs). Since our department did not have one, I asked the MBE group at another department to lend theirs to me. Then Dr. Hiyamizu and two other researchers became interested in my research and the four of us formed a HEMT development team. Although the research was unofficial and not funded and we had limited time to work on it, the morale of the team was very high."

They Successfully Create a Prototype in Just Four Months--a Sense of Crisis Causes Team to be Highly Motivated

This picture shows the team members creating a HEMT prototype using an MBE (crystal growth device). From left to right: Dr. Ishikawa, Dr. Hiyamizu and Dr. Mimura.

Dr. Mimura quickly succeeded in creating a HEMT prototype and proved its practicality in just four months after forming the development team. According to him, there was a sense of crisis shared by the team members behind this speedy development.

"At the time, arguments arose about the real value of MBEs both in our company and in the industry due to the lack of devices suitable for the advanced crystallization unit and due to its high price. That is why the researchers, including Dr. Hiyamizu, felt a sense of crisis. Since I gave up developing MOSFET, I myself definitely wanted to make the HEMT project a success. In addition, I learned that there were other researchers in the world trying to do the same as we did and felt a sense of urgency not to let other rivals proceed before us. The desire to create the first HEMT led to a level of motivation, which accelerated the research. We shared joy when we succeeded in obtaining experiment data for the first time. I remember that I felt relieved, thinking that I could repay the members for their cooperation."

The next objective after the successful development of HEMTs was commercialization. The second part of the interview includes the path to commercializing HEMTs, future prospects and words of encouragement for the next generation.

Takashi Mimura,
Honorable Fellow
at Fujitsu Laboratories Ltd.
1944 Born in Osaka Prefecture.
1970 Joins Fujitsu Limited. Comes in contact with transistors for the first time.
1975 Transfers to Fujitsu Laboratories Ltd. Invents HEMT and conducts research and development on its mass production.

"It seems that my entire research career is a series of twists and turns dominated by randomness. I believe that randomness created surprises and excitement, which motivated me, a humble engineer, to successfully create what had not existed until then."
Dr. Mimura is now engaged in a joint research project with National Institute of Information and Communications Technology to obtain the best performance out of HEMTs.