Warm fusion

A device that could fit in your lab-coat pocket uses nuclear fusion, and just a little heat, to produce neutrons. The advantages in simplicity and portability over conventional neutron generators could be considerable.

On page 1115 of this issue, Naranjo, Gimzewski and Putterman1 report the successful demonstration of an intriguingly simple neutron generator that produces neutrons possessing an energy of 2.5 mega-electronvolts (MeV) from reactions involving the fusion of two nuclei of deuterium. This device, it must be stressed, will not generate net energy, and is not related to past controversies about ‘cold fusion’.

Neutrons can penetrate significant quantities of matter, and interact primarily with the nucleus rather than the electronic structure of an atom. As a result, portable neutron generators have found a wide range of applications, including well-logging for oil exploration, and the screening of baggage for airline security. Several commercial devices are available that use fusion reactions of deuterium (D) and tritium (T), whose nuclei contain one and two neutrons respectively (ordinary hydrogen nuclei have none). The reactions generate helium and a single neutron that carries away most of the reaction energy:

These neutron generators rely either on an ion beam from a miniature accelerator producing reactions in a solid target loaded with deuterium and/or tritium, or on the electrostatic confinement of a D–D or D–T plasma. In both cases high-voltage power is required, and the apparatus is fairly complex.

The device reported by Naranjo et al.1 falls into the solid-target category, only without much of the complexity. Indeed, in some ways it is remarkably low-tech — the only input is a few tens of volts, to bias an electron-suppression grid, and some gentle heat (around 2 watts). A minute or two after the heat is applied, neutron emission starts, reaching a peak of about 1,000 per second; once the heat source is removed, the device gradually switches itself off. The key to the device's simplicity lies in the replacement of the miniature ion-source and accelerator in existing generators by a system based on a combination of two well-known phenomena — the pyroelectric effect and field ionization.

The pyroelectric effect — the fact that some materials become charged when heated — was probably first recorded in 314 BC by Theophrastus2, Aristotle's student and successor, from his studies of the gemstone tourmaline. More recently, various man-made materials have been investigated, and potentials of around 100,000 volts reported for crystals such as lithium tantalate (LiTaO3), with the emission of energetic electrons under suitable conditions. This effect was used by Brownridge3,4 to produce a small pyroelectric X-ray generator, of which a commercial version, powered by a 9-volt battery, is now available5.

Field ionization of gases occurs when a potential difference of a few volts exists over atomic distances — equivalent to a field greater than 10 gigavolts (1×1010 volts) per metre. The effect is widely used as the basis of field-ion microscopy. Modest voltages applied to electrodes of very small radius can produce these extremely high fields near the electrode tips, ensuring the ionization of essentially all gas molecules entering the high-field region.

Figure 1 shows how Naranjo et al. combined these effects to generate fusion neutrons. The authors grounded one face of a 1-cm-thick pyroelectric crystal to the inside of a vacuum chamber containing deuterium gas at a pressure of 0.7 pascals (for comparison, Earth's atmospheric pressure is around 105 pascals). They then attached a tiny tungsten electrode to a plate on the positive face of the crystal. A solid target containing deuterium in the form of erbium deuteride (ErD3) was placed a few centimetres in front of this electrode.

Figure 1: Naranjo and colleagues' apparatus for neutron generation1.

The chamber is filled with deuterium gas at low pressure (0.7 pascals). As the crystal is heated, the potential builds across the crystal. Deuterium ions (deuterons) are generated at the tungsten tip, and accelerated towards the target; the electrons fall back to the crystal electrode. The ions strike the deuterium target (ErD3), and some generate 2.5-MeV neutrons. Electrons knocked from the target surface are repelled by the suppression grid and fall back on to the target rather than being accelerated back to the crystal.

Raising the temperature of the crystal at a rate of 12.4 °C per minute changed the spontaneous polarization of the crystal, and raised the potential of the positive electrode at a rate of about 50 kilovolts per minute. As the potential rose, the field near the tungsten electrode increased to a value — around 25 gigavolts per metre — sufficient to produce field ionization of the deuterium gas. The positively charged ions (deuterium nuclei, or ‘deuterons’) produced in this process were accelerated towards the target across essentially the full potential generated by the crystal; the electrons stripped from the deuterium atoms by the ionization experienced a potential drop of only a few volts as they fell back to the crystal. On hitting the target on the opposite wall of the device, the energetic deuterons interacted with the deuterium target to produce 2.5-MeV neutrons via the D+D reaction.

The maximum current obtained in this experiment was about 4 nanoamperes, leading to a maximum neutron production rate of around 1,000 neutrons each second. The accelerating potential can be maintained only while the crystal temperature is changing; thus, the duration of the pulse at this current level was limited to a few minutes by the attainable temperature rise. Although this output is too small for most applications, the authors outline plans to increase the yield to a million neutrons per second, comparable to that of some commercial portable neutron generators. Nevertheless, even at the level already attained, there are laboratory uses, such as measuring neutron detector response or for student practical demonstrations, for which a simple, inexpensive, monoenergetic neutron source would be most valuable.


  1. 1

    Naranjo, B., Gimzewski, J. K. & Putterman, S. Nature 434, 1115–1117 (2005).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Donnay, G. Acta Crystallogr. A 33, 927–932 (1977).

    ADS  Article  Google Scholar 

  3. 3

    Brownridge, J. D. Nature 358, 287–288 (1992).

    Article  Google Scholar 

  4. 4

    Brownridge, J. D. & Raboy, S. J. Appl. Phys. 86, 640–647 (1999).

    ADS  CAS  Article  Google Scholar 

  5. 5

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Saltmarsh, M. Warm fusion. Nature 434, 1077–1079 (2005).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing