George Sassoon,


Reading the literature on cold fusion, I came across reports that electric arcs could cause transmutations and other nuclear reactions.

Being in possession of an electric arc welder, quantities of scrap steel, and Geiger-Mueller counting equipment, I decided to do some experiments.

The welder is an Italian-made Cebora type 1722, maximum welding current 140 amps. The radiation detectors are from Aware Electronics, Wilmington DE, type RM-80. These use a pancake-type tube with a mica alpha window on one side, about 30 mm. diameter. At this location in the west of Scotland near sea level, the cosmic background count rate is about 32 counts/min., corresponding to 9 microRoentgens/hour (uR/hr) according to the tube's certified calibration. This is with the tube on its side, as it is normally operated. The background is slightly higher with the tube flat, as the cosmic radiation is principally from a vertical direction. See Web site for details of the detectors.

The steel pieces used are old farm gate hinges, being strips about 50 mm. wide by 8 mm. thick. There is no detectable radiation from them before welding.

The welding rods used are from various sources - I can obtain details if required. The flux coating of these contains minerals which are slightly radioactive. I have had them for some time so there can be no short- lived radionuclides in them.


The RM-80 is clamped to a table on its side, with an unwelded portion of a gate hinge close to its alpha window. The computer logging program provided by Aware Electronics is started and run for a few hours to establish background. The raw data file is a sequence of numbers, being counts for successive minutes, with date/time and other data.

The hinge is then removed, and any rust cleaned off with an angle grinder. A large flat weld is then made, about the diameter of the tube's window, the slag knocked off, and the weld quenched in water. It is then clamped in position close to the alpha window and the program left to run for as long as possible.

The computer program allows a curve to be displayed giving successive averages over periods of an integral number of minutes. Initially, this is set to 6 mins. (0.1 hr) but after some time it is increased to one hour or more.

The raw data files can be converted to ASCII for importation into spread- sheets such as Excel.


In every case, when the weld is placed by the tube's alpha window, the count rate rises from the background 9 to typically 23 uR/hr, and then starts to fall. It reaches a minimum after about 12 hours, and then starts to rise again, tending exponentially towards a limit of about 25 uR/hr. with a half-life of about 150 hours. This latter part of the curve is like the charging of a capacitor via a resistor.

Putting various pieces of paper, aluminium sheet etc. between the weld and the detector, it appears that the radiation is partly alpha but primarily beta or low-energy gamma. There is no detectable radiation through the thickness of the hinge (about 8 mm.), behind the weld.

The thickness of the weld does not appear to matter, implying that the radiation is from the surface and not the bulk of the weld. I have tried thick welds using 10 rods (20-30 amp types) and thin ones using only two, and the results are not significantly different.

Due to the extremely low radiation levels involved, the curves obtained are noisy. One thing I have discovered is that the windows of the room should be left open, to dispel any radon or thoron. Even very low levels of these gases can affect the results. The house is in a low-radon area, but some of the modern materials used (such as plasterboard) exude significant amounts of them.


Undoubtedly some of the radioactive flux coating of the welding rods will remain in the weld, causing the initial rise in observed radiation. But why should it change thereafter? In particular, why should it drop off after the initial rise, and then after 12 hours, start to rise again? Is there some diffusion process going on? And why should broadly similar results be obtained in all the welds (about a dozen) that I have done?


Supposing that the electric arc generates neutrons, one might find some neutron-activation of the steel. Suppose that this transmutes the iron into a radioactive isotope with a half-life of about 12 hours, and another with a half-life of 150 hours, which is an alpha-emitter and decays in turn into some long-lived radioactive isotope.

We would see the initial decay of the 12-hour isotope, but not of the 150-hour one, as the alphas would not be able to escape the steel. But we would see the decay product of the 150-hour isotope, a beta/gamma emitter, and this would cause the slow rise over 500 hours which is observed. I have done a fair simulation of this curve using Excel.

The flux coating of the rods is a mixture of various minerals and a binder, and the elements in these substances could also be neutron-activated. The slight radioactivity of the flux could be due to potassium, as naturally- occurring K is slightly radioactive (K-40, 0.0012%, half-life 1.28*10^9 years according to my CRC Handbook of Chemistry and Physics).

I would be most grateful if someone could repeat these simple experiments, and if possible use a spectrum analyser on the radiation. I would also like to know what does in fact happen when Fe, K etc. are bombarded with neutrons, and the half-lives of any radionuclides generated.