Why diamond could power the future of electronics

A well-dressed gentleman sits in the lobby of an upscale hotel in Washington, D.C., holding a black box full of diamonds.

The man, Adam Khan, isn’t a jewel thief, but he is trying to sell diamonds.

Well, lab-grown diamond films to be accurate. Khan is the founder and chief executive officer of Akhan Semiconductor, and he’s trying to usher in the age of diamond electronics.

Khan and like-minded researchers think diamond and its exceptional properties could replace current electronic stalwarts such as silicon and outperform other advanced materials in high-powered applications. This optimism has been sparked by the ability to grow larger, higher-quality diamond films than ever before. Yet many experts also see several technical hurdles to overcome before commercial electronic devices will capitalize on diamond’s extreme properties. And some people believe these hurdles could take many years to clear.

Khan is not among those people. Akhan launched about three years ago with a technology licensing agreement in place with Argonne National Laboratory. With what Khan calls a “very healthy” early round of venture capital funding, the company is now working to grow this year from 20 to 54 employees. That number should break 200 sometime in 2018, Khan says.

Akhan has its headquarters about an hour north of Chicago, in Gurnee, Ill. The West Coast has Silicon Valley. Khan believes the Midwest will soon be home to the Diamond Prairie.

For the public, diamond is one of the sexiest materials on Earth. Diamond grown in a lab may seem less so, but it’s chemically equivalent to the mined mineral. Researchers can use high-pressure, high-temperature reactors to mimic geological conditions to grow diamonds in the lab. But chemical vapor deposition techniques that grow diamond from gases are becoming more popular with diamond makers.

Growing the carbon allotrope also affords researchers some degree of control over diamond’s dazzling electronic properties and its remarkable ability to handle heat. Semiconducting diamond’s electronic band gap is bigger than both silicon carbide and gallium nitride, two wide-band-gap materials that are making headway commercially. Wide-band-gap materials need more energy to conduct charge than small-band-gap semiconductors, notably silicon.

The allure of wide-band-gap materials, compared with silicon, is that researchers and manufacturers can use less of them to transmit electricity and electronic signals at higher voltages and frequencies.

That makes diamond an attractive alternative for use in commercial electronics as silicon devices slowly reach their upper limits for speed and lower limits for size. But diamond is also uniquely equipped to improve applications in power electronics—the electronics used in sectors such as aerospace, transportation, communications, and power grid development that are designed to operate at high voltages. These applications also run hot, and wide-band-gap devices maintain their functionality at significantly higher temperature than their silicon counterparts.

Akhan is already shipping grown diamond to customers, Khan says, but he’s not disclosing who they are. “But I can say they’re very much household names.”

Such secrecy isn’t unique to Akhan. Element Six, a grown-diamond company owned by the international diamond conglomerate De Beers, has relationships and nondisclosure agreements with clients that buy diamond for electronics.

Element Six sells about $400 million of grown diamond in a year, but diamond for electronics accounts for a small fraction of that, says Thomas Obeloer, a business development manager with the firm. The largest single use for the company’s products is making abrasion-resistant tooling or surfaces that capitalize on diamond’s superior hardness.

That mechanical fortitude comes from the tetrahedral covalent bonds linking carbon atoms in a diamond’s lattice. The vibration of these bonds also allows diamond to quickly dissipate heat. Diamond’s thermal conductivity is about 22 times that of silicon and about five times that of copper, which is cheap and widely used in heat sinks for computers and other electronics. From what Obeloer has seen, clients are primarily interested in using diamond’s heat-spreading ability in their electronics.

Relying on diamond’s passive thermal properties alone may seem like an underutilization of the material, but controlling heat more efficiently will let manufacturers push electronics harder and keep their devices alive longer.

Element Six sells about a dozen different grades of grown diamond—determined largely by a film’s grain size—but roughly speaking, its films cost about $1.00 per mm², according to Obeloer. That’s around three to four times the price of aluminum nitride, a competing advanced heat-spreading material. Element Six’s lowest-grade diamond still has a thermal conductivity several times as great as that of aluminum nitride.

“Diamond is absolutely a premium material,” Obeloer says. As engineers and designers become more acquainted with it, their designs will better capitalize on the advantages it offers. Diamond won’t replace cheap materials like copper in existing electronics; it will reduce the overall cost of creating and operating new devices that aren’t hampered by silicon’s limitations, Obeloer believes. It will take time, but “the opportunities for diamond will flourish,” he says.

Khan agrees, to an extent. He does see thermal conductivity as diamond’s means to get a foot in the electronics industry’s door. But he’s already envisioning active diamond devices that do more than spread heat.

“What we’re unfurling today is actually diamond for optoelectronic and electromechanical systems,” he says. Akhan is also working toward an all-diamond integrated circuit or microchip. So why is Khan optimistic that now is the time for diamond-based electronics?

The answer starts with chemical vapor deposition. Diamond grown by CVD started taking off in the 1980s, which is when it caught Richard Garard’s interest. Garard is CEO of Microwave Enterprises, a company that makes CVD systems to grow diamond.

To grow CVD diamond, researchers flow hydrogen gas and a carbon-rich gas, usually methane, into a chamber heated to hundreds of degrees Celsius. Garard’s systems turn the gas into plasma, liberating carbon atoms and allowing them to grow onto seed crystals on a substrate. The CVD conditions also create atomic hydrogen, which ensures the carbon grows into diamond by etching away any graphite that tries to form.

As CVD techniques and technology has improved, diamond growers have been able to create larger and higher-quality films, making research-grade diamond more accessible. The availability of such diamond was one of the bottlenecks that prevented scientists from studying it.

“Diamond’s always been a logical material to think about” for making electronics, Garard says. “But until recently it’s been unrealistic.”

Further signaling that now is an important time for lab-made diamond, Garard helped form the International Grown Diamond Association earlier this year. The association is a group of about 20 companies worldwide that grow diamond, sell it, or sell machines that grow it.

Nevertheless, Garard believes that diamond electronic devices are still about a decade out, he says. Pallavi Madakasira, an analyst with the market research firm Lux Research, agrees.

The firm doesn’t see the electronics industry adopting diamond within the next decade, she says. “It’s much more expensive than silicon carbide and gallium nitride, which are good enough to do the job today.”

Lux projects that silicon carbide and gallium nitride will claim a $3 billion stake of a $20 billion power electronics market in 2024. Silicon will make up the remainder. Diamond didn’t make Lux’s list. “We don’t see diamond moving the needle,” Madakasira says.

Researchers in academia, however, are a little more enthusiastic about diamond, says Robert J. Nemanich, a physicist at Arizona State University, who has also worked with silicon carbide and gallium nitride. “The properties of diamond are extreme and unusual,” he says. “It really is an ideal material for power electronics.”

Two main challenges stand in the way of diamond’s transition from the lab to commercial electronics, Nemanich says. One challenge is growing larger diamond wafers. His team currently grows diamond on 3- by 3-mm substrates. For comparison, commercial silicon wafers can have diameters greater than 300 mm.

The second challenge is doping, the replacement of a handful of diamond’s carbon atoms with ones that either donate or accept electrons. Doping is required to make diamond semiconducting.

It’s easy to dope diamond with electron-accepting atoms such as boron, but doing the same with electron-donating atoms remains a challenge. Nitrogen, carbon’s neighbor on the periodic table, seems like a sensible donor with its five valence electrons. But nitrogen impurities distort diamond’s lattice, which can trap the extra charge carriers instead of donating them, Nemanich explains.

Phosphorus could be a better choice, he says. With funding from the Department of Energy’s Advanced Research Projects Agency-Energy, Nemanich and his colleagues recently found a way to dope diamond with phosphorus. They demonstrated that they could use this technique to make bipolar diodes, the precursors to transistors, capable of withstanding a whopping 1,000 V. The wrinkle is that they had to use an unorthodox crystal orientation for their diamond, Nemanich says. The team is currently developing a new approach to dope more commonly used crystals.

Working with Anirudha V. Sumant at Argonne, Akhan has found ways to make doped diamond at large scales. The team developed a process for growing diamond on large silicon substrates at about 400 °C, a low enough temperature to dovetail with existing semiconductor processing techniques. “We really don’t want to see all the intellectual and financial investments that have been made into silicon go to waste,” Khan says.

And the team has turned to ion implantation to force electron donors, usually phosphorus atoms, into its diamond. Further breakthroughs are still needed to bring large-area growth and doping to single-crystal diamond, Sumant says. Compared with the diamond that Akhan specializes in growing, which has nanoscopic crystal grains, single-crystal diamond is better suited for power electronics. But that nanocrystalline material still earned Khan and Sumant an R&D 100 Award in 2013. Khan thinks that material has potential in low-voltage consumer electronics.

Although it’s unclear when diamond’s age will dawn, Akhan is optimistic about its plan to supplant silicon by growing diamond directly on it.

“I see it as diamond-on-silicon building toward a true diamond age where you have an all-diamond integrated circuit,” Khan tells C&EN at the D.C. hotel. “Something like a diamond Pentium processor, really, within the five-year horizon.”