Muon-Catalyzed Fusion
A heavy cousin of the electron pulls hydrogen atoms close enough to fuse — at room temperature, hundreds of times, before it dies.
Catalyses fusion of deuterium and tritium nuclei at cryogenic temperatures by replacing one of the orbital electrons with a negative muon. Because the muon is 207× heavier than the electron, the resulting muonic molecule (dtµ) has nuclei held only ~500 fm apart — close enough that quantum tunneling produces fusion in picoseconds. The muon is released after each fusion and recycles into another molecule until it either decays (lifetime 2.2 µs) or is captured by an alpha particle ('α-sticking'), which limits the cycle count.
(1) A pulsed proton accelerator (typically a synchrotron or cyclotron, several hundred MeV) hits a target to produce pions; the pions decay to negative muons. (2) Negative muons are stopped in a cryogenic deuterium-tritium mixture, where they replace electrons in atoms via Coulomb capture. (3) A µ⁻d or µ⁻t atom collides with a regular molecule and forms a muonic molecule dtµ. The muonic molecule's internuclear separation is so small (~500 fm) that the d and t nuclei tunnel through their Coulomb barrier and fuse to ⁴He + n + 17.6 MeV in picoseconds. (4) The muon is freed and can catalyse the next fusion. The cycle stops when (a) the muon decays (2.2 µs intrinsic lifetime) or (b) the freed muon binds to an outgoing α (probability ω_s ≈ 0.5–1 % per fusion), known as α-sticking. Best measured cycle counts reach ~150 fusions per muon. To break even commercially, ~300 fusions per muon would be needed at current accelerator efficiencies — and this number has not been reached.
In plain terms
Two hydrogen atoms normally float past each other politely because their electron clouds keep them apart. Now imagine swapping each atom's electron for a much heavier 'muon': the cloud collapses inward and the two nuclei are suddenly close enough to fuse, releasing a burst of energy. Then the muon, a temporary visitor, falls apart in two microseconds. Each muon can do this trick about 150 times before disappearing — and it costs more energy to make a muon than the fusions return.
Also called: μCF, cold fusion (Frank–Sakharov sense), muon-induced fusion.
Why it matters
- A genuine, repeatable fusion reaction at cryogenic temperatures — confirmed in many labs since 1957.
- Provides a clean source of ~14 MeV neutrons for materials testing and medical isotope production.
- Tests muon-atomic and muon-molecular physics in a regime no other technique can access.
- Anchors the discussion that 'low-temperature fusion' is real, while showing precisely why it is not energy-positive.
Where it lives
- RIKEN-RAL muon facility (UK)
- PSI muon-catalyzed fusion experiments (Switzerland)
- TRIUMF and J-PARC pulsed-muon beamlines
Key numbers
- Muon mass: ~207 m_e (Sets the dtµ size)
- Muon lifetime: 2.197 µs (Intrinsic decay)
- Cycles per muon (record): ~150 (RAL, RIKEN data)
- Energy per fusion: 17.6 MeV (d + t → ⁴He + n)
- α-sticking probability: 0.5–1 % (Limits cycle count)
Surprising facts
- A muon can catalyse roughly 150 fusion events before decaying — a single particle running a tiny industrial cycle 150 times in two microseconds.
- The d–t separation in a dtµ molecule is about 500 fm, smaller than 1/200 of a normal hydrogen molecular bond — close enough for nuclei to tunnel through the Coulomb barrier in picoseconds.
- Muon-catalyzed fusion was predicted by F. C. Frank in 1947 and observed experimentally by Alvarez et al. in 1956 — well before high-temperature tokamak fusion was even attempted.
- α-sticking — the muon binding to a freed alpha particle — is the single physical effect blocking energy break-even, and it appears to be a fundamental, not engineering, limit.
Open questions
- Pressure-dependent fusion-rate plateau: Cycle rates measured at PSI and RIKEN-RAL plateau at deuterium-tritium pressures higher than naive resonant-formation models predict. Subtle three-body effects in the µCF kinetics contribute, but the full quantitative picture is not closed.
- α-sticking lower bound: Despite decades of measurement, the lowest reproducibly measured α-sticking probability remains around 0.4–0.5 %, well above some early theoretical hopes for sub-0.1 % values. This appears to be a genuine, possibly fundamental, ceiling on cycles per muon.
A short history
F. C. Frank predicted the catalysis effect in 1947; Andrei Sakharov independently outlined its theory shortly thereafter. Luis Alvarez and collaborators at Berkeley unexpectedly observed muon-catalysed pd fusion in a hydrogen bubble chamber in 1956. Vesman (1967) explained the resonant formation of muonic molecules at low temperature, which boosts the catalysis rate. Major experimental campaigns at LAMPF (Los Alamos), PSI, RIKEN-RAL and TRIUMF in the 1980s–2000s mapped cycle rates and α-sticking. The phenomenon is settled physics; the open question is purely whether commercial energy break-even is reachable, which remains speculative.
Sources
- Alvarez et al., Phys. Rev. (1957) — observation of muon-catalysed fusion (https://journals.aps.org/pr/abstract/10.1103/PhysRev.105.1127)
- Vesman, JETP Lett. (1967) — resonant µ-molecular formation (http://jetpletters.ru/ps/1716/article_26113.shtml)
- Jones, Nature (1986) — review of muon-catalysed fusion prospects (https://www.nature.com/articles/321127a0)
- Nagamine, RIKEN-RAL muon-catalysed-fusion programme (https://www.riken.jp/en/research/labs/rms/muon_phys_basic/)