HomeThe World We DiscoverGraphene Finally Works as a Semiconductor, Twenty Years Late

Graphene Finally Works as a Semiconductor, Twenty Years Late

Graphene semiconductor breakthrough at Georgia Tech delivers 10x silicon's electron mobility with a usable band gap. Here is what it means for computing beyond silicon.

Youtube video
Share
The World We Discover · Explore this series
January 9, 2024
Updated March 29, 2026
Key Takeaways
  • Epitaxial graphene on silicon carbide is the first graphene semiconductor with a usable band gap
  • Its electron mobility is ten times greater than silicon
  • SiC wafers are already used in foundries, easing adoption

Walter de Heer spent two decades trying to make graphene work as a semiconductor. The physicist at Georgia Tech had watched the material win a Nobel Prize, inspire thousands of papers, and leave every engineer who tried to build a transistor from it empty-handed.

Graphene conducted electricity beautifully. It just could not be switched off.

That changed when his team grew epitaxial graphene on silicon carbide, a specific form that possesses both a usable band gap and electron mobility ten times greater than silicon. The results, published in Nature, represent the first graphene semiconductor that could plausibly enter a chip fabrication line.

Silicon Hit a Wall Before Graphene Was Ready

For six decades, Moore's Law carried computing forward on a simple trick: make transistors smaller. The approach worked until transistors approached atomic scales, where quantum effects began to interfere.

The problem now is thermal. Stacking transistors vertically, the industry's current strategy, generates heat that silicon struggles to dissipate. Chipmakers need a material with fundamentally better electron mobility, not merely a cleverer architecture.

Graphene, a single layer of carbon atoms arranged in a honeycomb lattice, has always been the obvious candidate. Its electrons move through the material with remarkably little resistance.

The catch was fundamental. Without a band gap, graphene could not switch between on and off states. A transistor that cannot turn off is not a transistor.

Band gap

A band gap is the energy barrier that lets a material switch between conducting and not conducting electricity. Silicon has one. Pure graphene does not, which is why it could not function as a semiconductor until de Heer's team engineered one into epitaxial graphene on silicon carbide.

De Heer's Team Built the First Graphene Semiconductor

The approach his group pursued, growing graphene directly on silicon carbide wafers rather than peeling it from graphite, turned out to be the critical variable. The SiC substrate interacts with the graphene lattice in a way that opens a 0.6 electron-volt band gap. That is enough to function as a semiconductor.

Think of it as adding a gate to a river. The electrons still flow with extraordinary speed, but now someone can stop them.

Key figure

5,000+ cm² V⁻¹ s⁻¹

Electron mobility in de Heer's semiconducting epitaxial graphene, more than ten times the mobility of conventional silicon.

The mobility figure matters because it determines how fast electrons travel through a chip. Faster electrons mean faster switching, lower power consumption, and less waste heat. These are precisely the properties silicon can no longer improve at current scales.

De Heer, characteristically understated about a result twenty years in the making, described it plainly: "We now have an extremely robust graphene semiconductor with 10 times the mobility of silicon, and which also has unique properties not available in silicon."

We now have an extremely robust graphene semiconductor with 10 times the mobility of silicon, and which also has unique properties not available in silicon.

Walter de Heer, Georgia Tech

The Lab-to-Foundry Gap Remains Enormous

Promising lab results and a product inside a fabrication plant are separated by years of engineering and billions in capital. Materials science has no shortage of demonstrations that looked convincing on a university bench and then quietly vanished.

SiC wafers are already standard in power electronics. That gives epitaxial graphene one advantage most novel materials lack: a substrate that foundries already know how to handle. Several manufacturers began piloting 8-inch SiC wafer production in 2026, though costs remain considerably higher than mature silicon processes.

Sabine Hossenfelder, the physicist and science communicator, captured the cautious mood well. "This is pretty exciting honestly," she noted. "After 20 years of graphene talk that didn't amount to much it could finally go somewhere."

The emphasis falls on "could." No one has yet demonstrated epitaxial graphene transistors at the density and yield that commercial logic chips demand.

The Next Few Years Will Test the Promise

The path forward is specific and measurable. Foundry trials over the next three to five years should reveal whether SEG transistors can be fabricated at densities competitive with advanced silicon nodes. The cost curve for SiC wafers needs to drop substantially. Current prices make graphene chips impractical for consumer electronics.

More on Materials Science

AI Materials Discovery: 5 Things to Know

The gap between what AI predicts and what chemists can actually make has become the central drama of materials science.

If the material scales, it could extend computing performance gains for decades beyond silicon's physical limits. If it does not, the semiconductor industry will turn to other post-silicon candidates: carbon nanotubes, transition metal dichalcogenides, or architectures that abandon transistor scaling entirely.

De Heer's lab is already working on the next step, integrating SEG into functional transistor arrays. Twenty years ago, Geim and Novoselov peeled graphene from graphite with adhesive tape and opened a new field. De Heer took the longer road, growing the material atom by atom on a wafer.

The question has finally shifted from whether graphene could work to whether anyone can manufacture it.


Sources

Fact Check: Claim-by-Claim Verification Verified

All claims verified against primary sources. Key facts confirmed: 0.6 eV band gap, 10x electron mobility vs silicon, 5,000+ cm² V⁻¹ s⁻¹ mobility figure. Both direct quotes verified. No incorrect or misleading claims found.

1 ACCURATE
Epitaxial graphene on silicon carbide has a 0.6 electron-volt band gap.
2 ACCURATE
Electron mobility is ten times greater than silicon (5,000+ cm² V⁻¹ s⁻¹).
3 ACCURATE
Graphene was discovered in 2004 by Geim and Novoselov.
Well-established fact; Nobel Prize in Physics 2010.
4 ACCURATE
Moore's Law dates to 1965.
Gordon Moore published the observation in Electronics magazine, 1965.
5 ACCURATE
SiC wafers are commercially available and foundries are piloting 8-inch production in 2026.
6 VERIFIED
De Heer quote: "We now have an extremely robust graphene semiconductor with 10 times the mobility of silicon..."
Matches Georgia Tech press release verbatim.
7 VERIFIED
Hossenfelder quote: "This is pretty exciting honestly... After 20 years of graphene talk that didn't amount to much it could finally go somewhere."
Confirmed against YouTube video transcript.
Share
Related Articles
Riemann Hypothesis, Energy Levels and the Endless Hunt for Zeros

A New Scientist video on the Riemann hypothesis is a fine guide to one of mathematics' deepest puzzles. Here is what lies beyond it: the 2024 breakthrough and the stranger...

Quantum Physics Explained: Where Reality Gets Strange

Quantum physics governs atoms, light, and the technology in your pocket. It is also the most counterintuitive framework in all of science. Here is what we know, what we don't,...

Mathematics: The Language That Describes Reality

Mathematics is the language scientists use to describe reality. From prime numbers to infinity, from fractals to unsolved conjectures, here is what makes mathematics so powerful and so strange.

Endless Numbers, Endless Beauty: About Quanta's Infinity Piece

A stunning article from Quanta Magazine walks you through Cantor's diagonal proof for uncountable sets of infinite numbers - and puts the sizes of infinity in context.