Carbon, not silicon, might run the next generation of power electronics — and Chemical & Engineering News is putting diamond back on the semiconductor roadmap. The outlet's feature carries the headline framing that the long-promised "diamond semiconductor" idea is once again moving from lab curiosity toward real device work.

Diamond's pitch is brutally physical. It can hold off far higher voltages than silicon, and it sheds heat roughly five times faster than the copper heat sinks bolted onto today's CPUs and EV inverters. That combination — high breakdown voltage plus high thermal conductivity — is what makes it interesting for the unglamorous end of the chip market: power electronics, RF amplifiers, and the switches inside grid-scale converters.

Why now?

Silicon carbide (SiC) and gallium nitride (GaN) already grabbed the slot silicon couldn't fill in high-voltage power chips, and both are now shipping in volume. Diamond, in theory, sits one rung above them on what engineers call the Baliga figure of merit — the rough yardstick for how efficiently a material switches power. The C&EN piece frames diamond as the next candidate in that escalation, not a replacement for logic chips.

What still has to work

The catch has always been the same. Growing a usable diamond wafer costs orders of magnitude more than pulling a silicon ingot, and the best films still come out small. Researchers have spent decades chasing chemical vapor deposition reactors that can lay down single-crystal diamond at wafer scale, with mixed results. Until that bottleneck moves, diamond stays a specialty play.

Who would actually use it

If the economics ever bend, the buyers are predictable: defense radar, satellite power systems, and the high-frequency RF front ends where heat is the enemy. None of those markets move silicon's volume needle. All of them will pay a premium for a part that simply does not fail when it gets hot.

The bottom line

Diamond will not show up in a smartphone next quarter. C&EN's framing positions it as the material waiting in the wings behind SiC and GaN, with wafer-scale growth as the single number that will decide whether it ever leaves the lab. Watch that line item, not the headlines.