What were all these silicon people doing at a compound semiconductor conference?
CS MANTECH 2015 was the 30th anniversary version of the conference once known as GaAs MANTECH, and which now goes by the more representative name of Compound Semiconductor Manufacturing Technology Conference, the change made along the way to reflect the diverse technology selections being served at recent conferences, including this one, to the more than 400 speakers and attendees who assembled in Scottsdale, AZ, last month.
It’s been said that the GaAs / III-V device manufacturing industry was built on the discarded equipment of silicon fabrication, but that’s probably not the reason why so many silicon people were in the house.
Maybe it’s more a case that the silicon industry, in this More-than-Moore era, knows that silicon is not the single best place to look for photons, for example, as Dr. Milton Feng, The University of Illinois, Champaign – Urbana, joked in his talk on Compound Semiconductor Microelectronics / Photonics Research for the next 30 years – Personal View based on past 30 year’s evolution.
Which means that the silicon world needs to reach into compound semiconductors if silicon is to heterogeneously integrate photonics, something Dr. Feng predicts will be one of the big stories in the next 30 years of microelectronic device fabrication.
Conversely, the compound semiconductor world has been reaching into silicon in order to leverage the cost advantages of using silicon substrates for GaN and SiC applications, for example.
(Even IBM is in on the act – “IBM Demos III-V on Silicon” writes Colin Johnson in the 09 June 2015 EETimes.)
What does the world hear when a Nobel Laureate (2014; Physics), Dr. Hiroshi Amano, Nagoya University, Japan, gives a keynote talk on Lighting the Earth by LEDs, Past, Present and Future Prospects of GaN-Based Blue LEDs?
The world first hears by way of introduction that Professor Amano is a “super nice guy,” and then hears from Professor Amano himself the goal that drove him in his work, which was the thought that “If I could succeed in the fabrication of blue LEDs I could change the world.”
LEDs can bring light to rural communities where reliable electricity supplies are lacking; LEDs, once general illumination is converted to them, will bring significant energy savings to the world and perhaps act as a brake on the deleterious effects of coal-fired electrical power generation; and LEDs, in the form of Deep UV LEDs, can be used for point-of-use water sterilization in those parts of the world were tainted drinking water is a major cause of ill-health and premature death.
Professor Amano, you have succeeded. Domo arigatou gozaimasu.
Dr. Timothy Heidel, Advanced Research Projects Agency – Energy (ARPA-E), and his colleagues are working on Strategies for Wide Bandgap, Inexpensive Transistors for Controlling High-Efficiency Systems. That means making 6” substrate SiC-based devices in a 6” silicon wafer fab, if you were looking for something to do with your old 6” silicon wafer fab.
Dr. Heidel also reported on making GaN devices which are then lifted-off their substrate after processing and integrated with CMOS silicon for additional effect. Heterogeneous integration front and center, as was also the talk by Dr. Daniel Green, DARPA, on Compound Semiconductor Technology for Modern RF Modules: Status and Future Directions.
Regarding heterointegration, Dr. Green asked “What does the aggressive proximity of various devices gain us?” Much, is the answer, about which more will come in a future post.
After giving his insights into Packaging Trends in the Wireless Industry, Dr. Robert Darveaux, Skyworks Solutions, Inc., was asked by Dan Green whether he thought there would be a place for heterogeneous integration in the future. “Yes, and no,” answered Robert; “there’s no need for it in making I/O density improvements, but it could be useful for making size improvements. It’s currently an area of limited exploration [for us].”
And that’s what the silicon world is eying, isn’t it? Implementing any new advanced technology is always going to be a tradeoff between costs and hassles. Now that EUV lithography is years and years behind its original commercial deployment targets, and with the expense of bleeding-edge CMOS designs in 16nm FinFET blowing up, the silicon world is thinking it might not be such a great hassle to employ compound semiconductor materials and devices in this More-than-Moore age. (See Moore’s Law and More Than Moore Are Laws Of Economics: 3D IC At The 25th Annual SEMI ASMC.)
Good thing there are more than 30 years running of GaAs, InP, SiC, and GaN manufacturing expertise just a short reach away.
That’s why all the silicon people were at a compound semiconductor manufacturing conference.
From Pittsburgh, PA, thanks for reading. ~PFW