By Michael Baxter
We think of the engineers, scientists and inventors who change the world as icons. Alexander Graham Bell. Thomas Edison. Albert Einstein – their largest contributions can be recited in just a few words.
But some of them live among us, unnoticed, even though they too made contributions that profoundly impacted everyday life. Russell Dupuis is one of them.
The smartphone you peer into, the LED bulb in your desk lamp, the Blu-Ray player that serves up your favorite film – all are here largely because of Dupuis, a professor in electrical and computer engineering at Georgia Tech.
That’s because an essential component of their manufacturing traces back to a process that Dupuis developed in the late 1970s, a process that ushered in a new breed of mass-produced compound semiconductors. These electronic components – particularly those forged of elements from columns III and V in the periodic table — can operate at extremely high frequencies or emit light with extraordinary efficiency. Today, they’re the working essence of everything from handheld laser pointers to stadium Jumbotrons.
The process is known as metalorganic chemical vapor deposition, or MOCVD, and until Dupuis, no one had figured out how to use it to grow high-quality semiconductors using those III-V elements. Essentially, MOCVD works by combining the atomic elements with molecules of organic gas and flowing the mixture over a hot semiconductor wafer. When repeated, the process grows layer after layer of crystals that can have any number of electrical properties, depending on the elements used.
Dupuis remembers well the autumn day in 1976 he first produced a working III-V semiconductor device using MOCVD. “It was a solar cell,” he recalls. “I had built my own reactor mostly out of spare parts to study the MOCVD process to grow a semiconductor on a gallium arsenide substrate. I took the solar cell outside and connected it to a current meter, and it worked pretty good. Since MOCVD was viable for solar cell technology, I thought it should be good for lasers and LEDs.”
He was right. At the time, Dupuis was a member of the technical staff at Rockwell International, hired to create working devices based on the MOCVD process being explored by Rockwell chemist Hal Manasevit.
“They knew they needed devices to make systems,” he says, “and I sold them on the idea that I could evaluate different materials using MOCVD to make those devices.”
After his initial success, Dupuis built a second reactor and refined the process. He then published a paper on his discovery and presented it at the 1977 Device Research Conference, an annual gathering of industry professionals and academics. But before the presentation, he was approached by a familiar face: Nick Holonyak, a University of Illinois professor who was Dupuis’ mentor.
“He came to my room and said, ‘I see you’ve got an interesting paper – can you build thin layers with MOCVD?’ Dupuis says, laughing. “I said, I can do as many as you need. Nick looked at me like I was crazy and said, ‘I’ve been trying to do this for five years.’”
Holonyak is a history-making engineer in his own right. Mentored by John Bardeen, the inventor of the transistor, he became the first to create a visible light-emitting diode in 1962, a breakthrough that continues to transform electrical lighting. While a senior at Illinois in 1969, Dupuis joined Holonyak’s lab.
Today, at the age of 88, Holonyak continues to operate a lab, and his praise for Dupuis is nothing short of ebullient.
“Russ Dupuis should be known as the person who invented the big process that’s now used to manufacture all the lasers and LEDs,” Holonyak says. “He has all the tricks to handle the complicated gases, the complicated chemistry, the stuff that explodes. I actually call the process ‘Dupuis-MOCVD’ – I hyphenate it.”
Their meet-up at the conference led Holonyak and Dupuis to reunite in the name of electrical engineering. Together, they published a paper after Dupuis demonstrated that MOCVD was superior to another emerging process, molecular beam epitaxy (MBE), in growing high-purity layers for compound semiconductors. In other words, they showed that MOCVD would work even for compound semiconductor devices that required complex structures.
Meanwhile, MOCVD began to take off as an electronics manufacturing process. Today, it remains the most widely used technology for creating thin-film compound semiconductors for electrical devices.
Dupuis left Rockwell in 1979 to join AT&T Bell Labs and later transitioned to academia, joining the faculty at the University of Texas, where he worked for 14 years. In 2002, he inquired about a position in Georgia Tech’s College of Engineering.
“It was a chance to work with really smart graduate students,” he remembers. “Plus, Georgia Tech had a building that was perfect for a clean room setup. I announced I was leaving a year before I actually left Texas, and when I walked in the door at Georgia Tech, the new lab was finished. The support here has been exceptional.”
These days, Dupuis is a Georgia Research Alliance Eminent Scholar and holds the Steve W. Chaddick Endowed Chair in Electro-optics. He continues to explore new combinations of atomic elements to make thin-film compound semiconductors.
And while he may go unrecognized in the local Starbucks, he has not escaped acclaim. In 2003, the White House welcomed him, Nick Holonyak and a third engineer, George Craford, awarding all three the National Medal of Technology. Most recently, in 2015, he was one of five recipients of the National Academy of Engineering’s esteemed Charles Stark Draper Prize – again, honored alongside his mentor, Holonyak.
While appreciative of the honors, Dupuis remains grounded as an engineer, more at ease with labor than with glamour.
“I remember our team in Holonyak’s lab hand-building 36 furnaces in the machine shop to support a project,” says the man whose reactor forged from spare parts ended up making history. “New ideas don’t require the best equipment. So if you’ve got a new idea, get your act together, and with the tools on hand, try it and test it. Because someone somewhere else may get there before you do.”