Research shelf / Materials & engineering / Diamond batteries
Written Updated
Nuclear diamond batteries, scaled from kilowatts to gigawatts on paper
The first carbon-14 diamond battery was demonstrated in December 2024 and produces about 15 joules per day per gram. That is enough for a sensor and nothing else. This paper asks what changes to isotope, architecture and conversion physics would be needed to move up nine orders of magnitude, and is explicit that the answer is a research programme rather than a product.
Theory or design only. No in-house measurement.
Four design series extrapolating the 2024 Bristol/UKAEA carbon-14 diamond cell toward utility scale, with the 400,000-tonne spent-fuel inventory as the feedstock argument.
The taxonomy runs across four series. Series A targets kilowatts with multi-isotope hybrid cores — americium-241 as a 5.5 MeV alpha emitter for power density, paired with carbon-14 as a 5,700-year beta baseline for longevity, in a 3D nanostructured diamond matrix that maximises capture area and minimises self-absorption. Series B addresses megawatts by hybridising radioisotope thermal conversion with direct betavoltaics.
Series C pushes into tens of megawatts and beyond using controlled subcritical neutron physics, driven by curium-244 — itself a waste product — with the safety case resting on the system being always below critical mass, so that losing the driver terminates the reaction passively. Series D is the frontier tier: NV-centre quantum efficiency enhancement, diamond quantum dots for tuned capture cross-sections, and radioluminescent photonic conversion.
The motivating argument is materials, not physics. Over 400,000 tonnes of spent nuclear fuel exists globally; strontium-90 and caesium-137 are among its most abundant and energetic constituents and cannot sustain a chain reaction under any circumstance. Extracting them for power would reduce the radiotoxicity and heat load of what has to go into geological disposal.
Every number, and what stands behind it
A claim is only worth the evidence attached to it. Each row below carries its basis: measured on the author’s own hardware, derived from the construction, measured on synthetic data, projected from literature, or simply cited.
| Claim | Figure | Basis | Context |
|---|---|---|---|
| Current C-14 cell output | 15 J/day per gram | Cited | Bristol / UKAEA, December 2024 |
| C-14 beta decay energy | 156 keV max | Cited | Insufficient to break diamond covalent bonds |
| Am-241 alpha decay energy | 5.5 MeV | Cited | ~100× the energy output per equivalent source mass |
| Standard betavoltaic efficiency | 0.1 – 5% (6 – 8% high-end) | Cited | Current device literature |
| Record SiC / perovskite efficiency | >21% | Cited | Recent literature — the headroom argument |
| Current RTG efficiency | 6.6% | Cited | Against a Series B target of 50 – 65% |
| Global spent-fuel inventory | >400,000 tonnes | Cited | The feedstock premise |
| Development horizon | 15-year phased roadmap | Projected | Author’s proposed programme |
Measured — author-run experiment on the stated setup. Synthetic — measured, but on synthetic rather than real data. Derived — follows from the stated construction or proof. Projected — paper-stated projection, not an author-run benchmark. Cited — taken from external literature.
How it works
- Multi-isotope hybrid cores. Alpha emitter for power density paired with a long-half-life beta emitter for baseline output as the alpha decays.
- 3D nanostructured diamond matrix. Maximised capture surface and minimised source-to-junction distance, attacking self-absorption losses.
- Thermal–betavoltaic hybridisation. Thermoelectric conversion layered with direct betavoltaic conversion in one device.
- Subcritical neutron driver. Curium-244 external driver with the reaction terminating passively if the driver is removed.
What it does not do
Taken from the folder’s own README. Nothing here has been softened.
- The paper’s own disclaimer: all four series are hypothetical, none is experimentally validated, and every power figure is a theoretical upper bound.
- The gap between 15 J/day/g demonstrated and gigawatt-scale proposed is roughly nine orders of magnitude.
- Series B’s 50–65% target sits against a demonstrated RTG efficiency of 6.6%. The thermal-management problem that closes that gap is named, not solved.
- Alpha emitters damage semiconductor converters. Diamond’s radiation hardness is the mitigation, but device lifetime under alpha flux is not quantified.
- Nothing here addresses the regulatory reality of distributing americium-241 or curium-244 outside a licensed facility.
Free under AGPL-3.0+ for almost everyone
Personal use, charities, education and organisations under AUD 50,000 a year pay nothing. A tiered commercial licence covers everyone else.