SCIENTIFIC CORRESPONDENCE

model. PERONI PAOLO Cold fusion: what's going on? Via 0. Regnoli 10, SIR-A significant point which is not temporal (hourly, daily) variation of Rome 00152, Italy widely known, and may therefore be over­ the cosmic-ray-induced neutron fluxes looked in neutron measurements of cold require that this background be carefully SIR-The only recognized mechanism for fusion rates, is the possibility of contami­ accounted for. nuclear fusion at ambient temperature is nation by cosmic-ray-generated neutrons; Comparable neutron fluxes can be that induced by negative muon binding of these should be taken into account in the generated by accelerators, isotope sources precursor hydrogen isotope molecules as design and interpretation of experiments. and nuclear reactors, even at considerable first described by Frank 1 with estimates of The cosmic-ray-induced neutron back­ distances; these contaminants of neutron fusion rates by Sakharov' and experi­ ground arises primarily from extra-solar measurements must also be reckoned mental observations by Alvarez' some 30 protons with energies above a few GeV, with. Obvious means for suppressing years ago. It is now known that more than which can penetrate the Earth's atmos­ these backgrounds are time-gating of the 100 fusion events may be induced by a phere and the Sun's and Earth's magnetic source, monitoring spurious sources with single muon during its lifetime of 2.2 x w-• 1 fields • Primaries and secondaries reach­ a second detector operated simultane­ seconds in a mixture of liquid deuterium ing the surface include neutrons and other ously with the detector(s) near the source and tritium'. If this number could be energetic particles which produce neu­ under investigation, or going under­ increased by a factor of 1,000, a break-even trons in the atmosphere and the first few ground-350 g em_, (two or three metres fusion reactor could be the result'. metres of the surface by spallation re­ of earth or concrete on all sides) should It is tempting to interpret the recent actions. While the spectrum contains reduce the cosmic-ray neutrons by a factor claims in terms of this process. It should neutrons up to energies comparable to of about 10. be remembered that more than 70 per cent incident particle energies, a major com­ JOHN M. CARPENTER of the cosmic-ray flux at the Earth's ponent is due to evaporation of neutrons Argonne National Laboratory, surface consists of positive and negative from struck nuclei; at birth these have 9700 South Cass Avenue, muons. There are about 200 m·' s· 1 with a energies in the range 1-3 MeV, and Argonne, Illinois 60439-4814, USA stopping rate in an absorber of some 5 1 1 • appear as a knee or shoulder on an other­ 1. Hayakawa, S. CosmicRayPhysics\'I'Jiley, New York, 1969). 2x10- g- s- The flux may be twice as wise continuous energy distribution. great at the altitude of Salt Lake City', There is also a substantial component con­ Dr Carpenter, a referee of the paper by which would be equivalent to more than sisting of 'thermal' neutrons, which have Jones et al. on page 737, provided this two muons a minute in an absorbing slowed down in the environment to a comment at our invitation. volume of about a litre. poorly equilibrated thermal distribution The following are extracts from the sub­ The rate at which fusion events could below energies of about 0.1 eV as well as stantial numbers of letters from readers occur is limited by the rate of formation of 'epithermal' neutrons whose distribution offering explanations of the two series of muon-bound molecules, which is itself a is roughly inversely proportional to the cold fusion experiments which have been sensitive function of parameters on the energy in the range 1 eV to 1 MeV. At generally reported. atomic scale, whence its dependence on energies above a few MeV, the spectrum Editor, Nature. resonance effects' and temperature'. tails off rapidly; this cascade component Once a muon is captured, the resulting contains about 10% of the total. The flux SIR-From the newspaper accounts, the muonic molecule is two orders of magni­ of each of the low-energy components is very small flux of neutrons generated tude smaller than the typical lattice spacing of the order of 10-' neutrons em-' s- 1 at sea during the experiment of Fleischmann in solids, so that free diffusion may be level and middle latitudes. and Pons is being taken as proof that their expected. These figures vary according to altitude conclusion is not valid, and that nuclear The fact remains that muon-induced (about twice as great at 1,500 m elevation), reactions between deuterons do not occur fusion has not yet been reported in metal­ and from time to time, mostly because of under the conditions they describe. lic compounds of hydrogen isotopes, and variations in atmospheric density and But when the kinetic energy is as small indeed has been considered unlikely solar and geomagnetic field intensity. The as in their experiment, the neutron and because of the preferential capture of e-folding thickness in the atmosphere is proton components of the deuteron do not muons by the heavier metal nuclei. On the 2 about 150 g cm- , so that, for example, behave in the same way, because the hypothesis that this loss mechanism is barometric pressure variations of nucleus of the target atom repels the suppressed by a resonance or band struc­ ±13 mm of mercury cause about ±10% proton but not the neutron. Thus, the ture effect in deuterium-loaded palladium, flux variations. Sometimes, when dealing neutron can be captured by the target it is possible to estimate the turnover with such a general source of neutrons, nucleus while the proton, which remains number required to explain the effects material intended as shielding, and even outside the Coulomb barrier, will fly off. which have been reported. detector material itself, can act as a source, This process, first recognized by Jones et a/! observe a neutron count­ so that some care in this respect is called Oppenheimer and Phillips 1 in 1935 leads rate of 4x10-' s- 1 with a neutron detection for in the measurements. to a pure (d, p) reaction and has a rela­ efficiency of about 1% in a volume of 160 As it happens, the counting rates due to tively high probability of occurrence, mi. If the neutrons observed are products cosmic-ray-induced neutrons are of the certainly much greater than that of the of the reaction, in a muon-deuterium same order of magnitude as the counting (d, n) reaction'. molecule, of two deuterons to yield 'He rates observed in the neutron and secon­ It follows that if the experiments des­ and a 2.45-MeV neutron, one should also dary radiation detectors in many of the cribed really brought the deuterium nuclei allow for the equally probable reaction measurements being made. And in detec­ close enough together to interact, one yielding 'Hand a proton, which would not tors that disperse the spectrum, the evap­ should expect no neutron emission and a have been detected by the neutron moni­ oration peak in the energy distribution reaction rate much higher than that tor. This implies five fusion events per due to cosmic-ray-induced neutrons is at evaluated on the basis of the high-energy second per litre. The required turnover nearly the same energy as that expected number is then of the order of 100, com­ from deuteron-deuteron fusion, 2.45 1. Oppenheimer, J. R. & Phillips, M. Phys. Rev. 48, 500 parable with that already known for the MeV. These observations, coupled with (1935). case of a mixture of liquid isotopes, but 2. Morrison, P. Experimental Physics Vol. 2 (ed. Segre, E.) the (admittedly weak. ± 10%) (Wiley, New York. 1953). significantly greater than that in pure NATURE · VOL 338 · 27 APRIL 1989 711 © 1989 Nature Publishing Group