One step closer to ‘Graphene Valley’

With help from Derek Robertson

Silicon has long reigned as the material of choice for the microchips that power everything in the digital age, from AI to military drones — so much so that “silicon” is almost a synonym for tech itself.

It’s anyone’s guess how long that will last. Silicon chips have been bumping against the limits of miniaturization for years, dividing chip makers on whether Moore’s law, the longstanding assumption that transistors will steadily get smaller and computers more powerful, is already dead.

But the global semiconductor industry is still under just as much pressure to produce ever more powerful chips, and keep up the pace of technological progress. It’s no surprise there’s been a growing frenzied interest in novel 2D materials that can pick up where silicon leaves off. They promise advantages over silicon — higher electrical conductivity, better temperature control, the ability to use less energy — that could, possibly, keep that decades-long momentum going.

Enter graphene. This month, researchers at Georgia Tech and China’s Tianjin University made a breakthrough in one of the top contenders for a silicon alternative — graphene, the one-atom-thick form of carbon that won a Nobel Prize in 2010. Georgia Tech physicist Walt de Heer and his team created the world’s first functional graphene-based semiconductor, marking what he dubbed a “Wright brothers moment” for the next-generation materials that could make up the electronic devices of the future.

“The silicon industry is completely reliant on the fact that you can make incredibly large single crystals,” de Heer told DFD, referring to the first step of microchip creation. “We’re at the level of basically having a crystal that you can start the entire [graphene] industry with.”

His team grew a layer of graphene on silicon carbide wafers, and testing showed the material acted as a semiconductor that beats silicon in terms of mobility, a measure of how easily electrons can move through. With ten times greater electron mobility at room temperature, it had the kind of performance that might one day enable faster computer chips.

“This is the first indication that it’s possible to make graphene a semiconductor, and it’s possible to use this type of graphene in making devices that could lead to new transistors and chips,” observed Mauricio Terrones, a materials science professor at Penn State. “It comes down to the basic fundamental challenge, which is ‘can you make uniform, large-area 2D material that is highly crystalline without imperfections?’”

De Heer said his discovery cleared the main technical barriers that have been holding graphene back as a semiconductor material. It unlocked a working “band gap” for graphene, a property the material needs to turn electronic switches on and off, but lacks naturally. Earlier attempts to engineer a bandgap in graphene compromised its intrinsic electronic properties. The recent success, though, came with preserving the material’s high electrical conductivity, ensuring quality crystals, and making it compatible with existing microelectronics manufacturing methods.

Graphene isn’t the first wonder material to be touted as a silicon alternative — the list includes gallium arsenide, boron nitride, and even an industrial lubricant called molybdenum disulfide. Some have turned out to be duds; others are in testing.

For graphene, it’s early days: The lab work still needs to be replicated and the material’s performance validated in a small manufacturing setting, said Gaurav Batra, Ayna.AI CEO and former co-lead of McKinsey’s Americas advanced electronics practice. It’s not clear yet whether graphene will replace silicon outright or mainly be used to enhance silicon’s capabilities — but he calls graphene “the most promising” potential large-scale alternative for silicon on the horizon.

“A CTO or head of R&D for a semiconductor company, I think they will be taking things much more seriously now in terms of investment,” Batra said of the research breakthrough. “This will definitely push graphene up in their priority list.”

Over the long term, what would it mean for silicon to be replaced? The modern semiconductor supply chain has been honed around a single bulk source material. Despite the hunger for faster and more efficient chips, Batra expects it will be a long haul before Silicon Valley can truly move away from its namesake element: “That infrastructure, which basically is billions of dollars in equipment, processing know-how, expertise being developed — all that getting replaced is going to take a while.”

Initially, Batra sees graphene semiconductors gaining traction in select high-end niche applications, such as machine learning, where consumers are willing to pay more for higher processing speed and lower power consumption. As hype spreads to more applications, he said costs will fall, paving the way for broader commercialization.

Penn State’s Terrones set a timeline of five years before the first product if R&D investments and an army of talent mobilize around tackling graphene’s commercial challenges, otherwise both he and de Heer reckon it will take about 10 to 15 years for any transformative industry applications.

For now, research funding is hugely skewed toward the legacy material. The CHIPS and Science Act boosted funding for federal research agencies like the National Science Foundation, some of which will support programs furthering advancements in new semiconductor materials. But awards for such programs total millions of dollars, compared to the billions set aside to build new semiconductor fabs continuing the silicon-to-chip tradition.

As Terrones said, with understatement: “It’s a little bit of a longer vision.”

the robot in your office

A new study from the International Monetary Fund warns that 40 percent of workers’ jobs will be affected in some way by artificial intelligence.

POLITICO’s Pieter Haeck wrote on the report yesterday, which specifies that “mature” economies like those of the European Union and United States are especially vulnerable. At the same time, with their high levels of education, these economies are also in a better place to reap AI’s benefits. Conversely, countries that are less economically developed are less vulnerable to AI but worse prepared to benefit from it.

“In advanced economies, about 60 percent of jobs are exposed to AI, due to [the] prevalence of cognitive-task-oriented jobs,” the report specifies. “Of these, about half may be negatively affected by AI, while the rest could benefit from enhanced productivity through AI integration.”

So, good news… -ish? The IMF has also published a “global preparedness index,” which specifies the U.S., Denmark, and Singapore as the countries best prepared for AI to impact the workforce. — Derek Robertson

california crackdown

A California lawmaker is proposing a bill to criminalize the production of child sexual abuse materials by artificial intelligence.

This morning’s POLITICO California Playbook had the scoop on the bill from Democratic state Assemblymember Marc Berman, which builds on a bill signed into law last year in the state that increased legal accountability for social media platforms that host such material.

Berman’s bill would target the creators and distributors of such content, not the AI platforms that generated it, with the lawmaker warning that “You could argue that every AI-generated image actually victimizes thousands of real children… Because they are a part of the formula that goes into creating that AI-generated image.” — Derek Robertson

Tweet of the Day

THE FUTURE IN 5 LINKS

Stay in touch with the whole team: Ben Schreckinger ([email protected]); Derek Robertson ([email protected]); Mohar Chatterjee ([email protected]); Steve Heuser ([email protected]); Nate Robson ([email protected]); Daniella Cheslow ([email protected]); and Christine Mui ([email protected]).

If you’ve had this newsletter forwarded to you, you can sign up and read our mission statement at the links provided.

CORRECTION: The Friday, January 12 edition of Digital Future Daily was updated to reflect Liz Hyman’s correct answers to the first two questions, which were originally and erroneously swapped.