The next hurdle is not hotter plasmas, but smarter eyes watching them.
Researchers in California are betting on a striking answer. Lab-grown diamond, cut thin and wired with micrometric precision, is being turned into a radiation-hard sensor that can live where others fail. A new €478,000 funding boost, part of a wider state and federal push on fusion, aims to prove these detectors can measure what happens at the edge of a miniature star without burning out.
Why diamond is suddenly the sensor to beat
Silicon built the modern detector world. It times particle hits with finesse and tracks them with micron accuracy. Inside a fusion device, it dies fast. Neutrons smash its crystal lattice. Heat and high fields push it beyond safe operation.
Diamond brings a tougher toolkit. It has a wide bandgap, a colossal thermal conductivity, and a breakdown field that laughs at intense electric gradients. Charge moves fast through it. Most of all, diamond shrugs off radiation that would cripple other substrates. That is why physicists at the Santa Cruz Institute for Particle Physics (SCIPP) have pivoted from silicon to carbon crystal for the most punishing jobs.
| Property | Silicon | Detector-grade diamond |
|---|---|---|
| Bandgap (room temp) | ~1.12 eV | ~5.45 eV |
| Thermal conductivity | ~150 W/m·K | ~2000 W/m·K |
| Breakdown field | ~0.3 MV/cm | ~10 MV/cm |
| Radiation tolerance | Moderate | High |
| Timing potential (LGAD class) | ~20–50 ps (mature) | Tens to hundreds of ps (emerging) |
The project aims to qualify diamond as the only sensor material that can operate in the immediate blast of a fusion core, where heat, neutrons and fields peak together.
A €478,000 push to build fusion-grade eyes
SCIPP has secured €478,000 (about $555,000) to turn diamond into a Low-Gain Avalanche Diode (LGAD) that can time nuclear events with exquisite precision. LGADs are thin, fast detectors with a small internal gain layer. In silicon they power the new generation of time-of-flight trackers at particle colliders. In diamond, they could sit millimetres from a fusion plasma and keep working.
The team is partnering with Advent Diamond, one of only a few firms that can grow detector-grade diamond on wafers and pattern it for high-voltage operation. Early prototypes suggest an order-of-magnitude improvement in radiation endurance compared with conventional devices, while keeping timing sharp enough to resolve sub-nanosecond bursts.
Funding is flowing because fusion is no longer a distant promise. Since 2022, ignition shots and new public-private programs have moved the field into deployment mode.
Fusion is getting real — and it needs better diagnostics
When a tokamak or stellarator lights up, the plasma can exceed 150 million °C. Deuterium and tritium nuclei fuse, releasing 14 MeV neutrons and a storm of charged particles. Stability hangs on fast feedback. Sensors must see disruptions as they form, track neutron flux profiles, and detail burn conditions in microseconds.
Most detectors are buried behind shielding. That keeps them alive, but blurs the picture and adds delays. Diamond detectors could move closer to the action and still survive. The payoff is cleaner data for control algorithms and safer operation in long pulses.
➡️ It looks like cheesy pasta but hides a surprising green heart, lunch ready in 5 minutes guaranteed
➡️ Bed bugs: the one move to make immediately if you find even a single one at home
➡️ Inheritance : the new law arriving in February reshapes rules for heirs
- What the new sensors will measure: local neutron rates and spectra near the first wall.
- Fast bursts from edge-localized modes and other instabilities.
- Burn uniformity and ash buildup during sustained operation.
- Hard radiation doses and transients that stress materials and magnets.
The California plan behind the lab work
California has put together a package to push fusion from physics to engineering. Several UC campuses share an €8 million, three-year effort to build a full “toolbox” for industrial-scale monitoring and control. SCIPP is leading on sensors. UC San Diego is stress-testing materials in extreme shots. UCLA and UC Irvine are tackling plasma and transport modelling. This autumn brought another €5 million from Sacramento to extend the runway.
That momentum sits on top of national moves. The U.S. Department of Energy has launched fusion hubs. Private capital has piled more than €10 billion into startups chasing both magnetic and inertial paths. A demonstration plant in California is pencilled in for the 2040s, assuming diagnostics, materials and regulatory pathways line up.
Inside the tech: what a diamond LGAD really does
LGADs add a lightly doped gain layer that multiplies the signal without saturating. In silicon they deliver timing in the tens of picoseconds for minimum-ionizing particles. Diamond changes the equation. The material’s high breakdown field allows aggressive biasing. Its fast charge carriers and low noise floor promise crisp signals even under brutal radiation loads.
The goal here is not just survival. It is precise time stamping in harsh conditions. That means stable gain at high fields, low leakage at elevated temperatures, and metallization that resists sputtering and graphitization under neutron bombardment. It also means packaging that routes signals out of strong magnetic fields without distortion.
Only a handful of companies can make large, uniform, detector-grade diamond. Supply chain depth will shape how fast these sensors reach reactors.
What could go wrong, and how the team is tackling it
Radiation damage still accumulates in diamond. Neutrons create defect centers that trap charge and slowly reduce gain. The project maps that damage versus dose and temperature, then tunes operation points to stretch lifetime. Uniformity across a wafer is another headache. Grown diamond can have variable defect densities. That pushes engineers toward tight quality control and smaller die sizes for critical positions.
Electrode adhesion at high temperature is nontrivial. Metal stacks must endure thermal cycling and intense fields without delamination. Finally, cost matters. Detector-grade diamond is expensive. The near-term use case is high-value positions close to the plasma, with conventional sensors handling the periphery.
Why this matters beyond fusion halls
The same rugged sensing could ride on spacecraft. High-radiation orbits, like the Jovian system or Earth’s Van Allen belts, punish electronics. Diamond detectors can log particle storms and protect payloads. Prototype packages also make sense for nuclear safeguards, advanced fission, and high-energy flash radiography where dose rates spike.
Signals to watch over the next two years
- Timing resolution at dose: Can diamond LGADs hold to tens of picoseconds after heavy neutron exposure?
- Bias stability: Do devices run at multi-megavolt-per-centimetre fields without breakdown in reactor-like conditions?
- Wafer yield: Are uniform, 4–6 inch detector-grade substrates feasible at scale?
- Integration tests: Performance inside a tokamak port with real electromagnetic noise and thermal gradients.
Extra context for readers who want the nuts and bolts
LGAD stands for Low-Gain Avalanche Diode. It is a solid-state sensor with a designed-in electric field peak that multiplies charge carriers by a factor of roughly 5–20. That small gain cleans up timing without the wild noise of full avalanche photodiodes. In fusion, timing matters because instabilities grow fast. If a detector can tag neutrons and charged fragments with sub-nanosecond accuracy, control systems can trigger mitigation before a disruption blossoms.
Non-ionizing energy loss (NIEL) is the damage channel that haunts sensors in neutron fields. It displaces atoms from lattice sites, creating traps. Materials with high displacement energy and strong bonds, like diamond, rack up damage more slowly. Even so, calibration drifts. Teams build in online calibration using internal pulsing and reference sources to keep sensitivity steady during long pulses.
For a concrete picture, imagine a diamond LGAD tile a few tens of microns thick, bonded to a ceramic carrier, biased at high voltage, and placed just behind the first wall of a tokamak. A 14 MeV neutron knocks out a recoil carbon nucleus. The recoil deposits charge. The gain layer amplifies it. Fast electronics time-stamp the hit and stream it into a model predictive controller that shapes magnetic fields in real time. Fewer surprises, fewer quenches, more reliable runs.
Originally posted 2026-03-09 02:19:00.