Kyoto moves the gate underneath, and a 600C silicon carbide JFET stops drifting

What matters in the Kyoto result is where the gate went. In the earlier generation the gate sat above the channel that carries current, and that arrangement made the threshold voltage, the point at which the device turns on, hard to pin down ^[4]. Part of the reason is that implanted atoms travel deeper into silicon carbide than expected ^[5]. Moving the gate beneath the channel reduced that effect, and the reported gap between intended and measured threshold at 673 K came in under 0.1 volts ^[6]. A single switch tolerates a wandering threshold. A circuit with several transistors sharing a bias network does not, which is why this is the part of the work that points at integrated circuits rather than at a hotter discrete part.

The second change concerns what holds the device apart from the material below it. Older designs leaned on the electrical resistance of that underlying material to isolate parts of the transistor, and leakage found its way through as temperature climbed ^[4][7]. The double-well version uses a p-n junction for the job instead, and the team reports this prevented unwanted current passing through the device as temperature increased ^[7].

Then there is the pairing of temperatures, which the announcement does not dwell on. The threshold accuracy figure is quoted at 673 K ^[6]. Continued operation and the on-off ratio above 1,000 are quoted at 873 K ^[8]. Those two temperatures sit 200 K apart ^[1]. Nothing in the account says the threshold stays inside 0.1 volts at the top of the range, and a control loop is designed around where a device switches, not around whether it still switches at all.

An on-off ratio above 1,000 is three decades of separation ^[2], offered as evidence the device still moves clearly between conducting and non-conducting states ^[8]. For one transistor that is a clean measurement. The team's own phrasing keeps the distinction visible: silicon carbide is called already mature as a power-device technology, while the bottom-gate structure is credited with potential for reliable extreme-temperature SiC integrated circuits ^[11].

The temperature headline is the least novel part of this. Silicon carbide circuits above 500C have already been demonstrated, and what stayed out of reach was turning those demonstrations into reliable, practical electronics ^[9]. Six hundred degrees is 100 degrees past that mark ^[3]. The stated prize is removing heavy cooling and thermal protection from systems that run hot ^[12], and reaching it depends on threshold control and isolation, which is exactly what these two structural changes buy.