A place for peculiar ideas.

Research article / 6 minute read

US 5,734,122: Aspden's Nernst-Effect Thermoelectric Apparatus and Its Cold-Fusion Rationale (priority 1989)

US Patent 5,734,122, "Thermoelectric energy conversion apparatus", was granted to Harold Aspden as an individual applicant on 31 March 1998. The patent describes its hardware concretely: rods, transformers, heat sinks and a drive waveform. It is also a useful case study in how much theory one patent can carry on very little reported measurement. The apparatus is conventional enough. The ideas it is meant to test, ranging from vacuum spin to cold fusion, are not.

Filing history and status

The patent family traces back to a British application, GB8908571, filed on 15 April 1989 under the title "Electrically controlled ion fusion". A second British filing, GB9310734 of 25 May 1993, was titled "Hydrogen activated heat generation apparatus". The US application (08/520,008) was filed on 23 August 1995 as a continuation-in-part of an application filed on 3 February 1994. The patent office placed it in class G21B3/00, which covers low-temperature fusion reactors and whose label explicitly calls cold fusion reactors alleged. The patent lapsed for non-payment of maintenance fees, effective 31 March 2002.

So the titles of the earlier filings already point to fusion, even though the granted patent presents itself as a piece of thermoelectric equipment.

The physics it starts from

The patent rests on the Nernst effect. This is a real, established thermoelectric phenomenon: when heat flows through a conductor that sits in a magnetic field at right angles to the flow, an electric field appears in the third perpendicular direction. The patent's own explanation is that the magnetic field deflects the electrons that carry the heat sideways, through the Lorentz force. Charge then piles up at the conductor's side surfaces. In the patent's framing, heat is converted to electrical energy while the magnetic field acts only as a catalyst, steering the charge without supplying the energy.

The patent writes the relation as E = N_c H δT/δx, where N_c is a Nernst coefficient for the metal. It asserts that E can reach several volts per centimetre in a field of about one tesla with a temperature gradient of one degree C per centimetre. It also names the practical difficulties itself. A strong gradient and a strong field are hard to set up inside a metal. And if you draw power through an external circuit, that circuit disturbs the heat path and works against the conversion.

The patent illustrates the effect with transformer laminations. In electrical sheet steel about 200 microns thick, heat from hysteresis and eddy-current losses escapes sideways across the width of the sheet, while magnetic domains roughly 100 microns across are saturated along the crystal axes. That gives the crossed heat flow and magnetic field the Nernst effect needs.

The apparatus claimed

What the device is. The preferred design (the patent's FIG. 9) is a long, closed rectangle of metal. Its two long sides are solid cylindrical rods of equal diameter, typically about one centimetre, connected at each end by a short link to complete the circuit. Transformers are used to induce current in this loop; in the preferred version there are two toroidal transformers, each with one of the end links threaded through its core, so the loop itself works as their secondary winding.

What is driven through it. Because the loop's resistance is so low, the patent says a small voltage is enough to push currents in the hundreds of amps around it. Its worked example applies 5 volts to a toroidal transformer rated at 300 volts at 60 Hz and runs it at 1 Hz. The a.c. is kept below 5 Hz and shaped asymmetrically: in one direction the voltage is lower and lasts longer, in the other it is higher and briefer. The patent gives two reasons for the slow cycle: to avoid skin effects in thick, highly permeable conductors and, more importantly, to allow what it calls a vacuum field spin state to build up before the current reverses.

Heat is the second input. Finned heat sinks are attached to the rods at two different positions, one set feeding heat in and the other drawing it out, and insulation along the rods between them keeps the heat travelling lengthwise.

What is supposed to happen. The current flowing along each rod produces a magnetic field that circles the rod's axis, while the heat flows along that axis. The patent asserts that the two together produce a radial electric field inside the rod; it reports no measurement of such a field. It also says the device cannot work without a temperature gradient: with no heat flowing, the current alone yields no radial field.

One claim adds an electrolytic variant carried over from the parent application. Here a rod sits along the axis of a cylindrical anode, immersed in electrolyte and fed with a separate d.c. supply. Because the patent puts the electrolyte's resistivity (about one ohm-cm for a typical salt solution) at 100,000 times that of the metal, it argues that almost none of the heavy a.c. leaks through the liquid; nearly all of it stays in the rod.

The theory behind it

Most of the specification argues why a radial electric field inside a metal should matter. The patent proposes that the vacuum is a structured medium with synchronised quantum spins (FIG. 5), in line with Aspden's Aether Lattice (Aspden's Preferred-Frame Vacuum). It cites his 1972 paper with D. M. Eagles in Physics Letters 41A, 423 as an early success of this theory.

From there the patent moves to cold fusion. It claims that the muon has a role in creating and decaying protons. It says neutrons do not exist as such inside nuclei, a position at odds with standard nuclear physics. It also says deuterons spend one seventh of their time with a neutral core. On that basis the patent argues that the Nernst effect could be used to set up a controlled background of negative charge inside a hydrogen-loaded metal, and that single surplus electrons in that background could let pairs of deuterons get close enough to fuse. None of this is shown to occur in the document. The patent quotes a reaction energy of 4 MeV for two deuterons forming a proton and a triton, followed by triton decay to helium 3 at 17.9 kev with a 12.2 year lifetime. It even extends the idea to biology, citing Goldfine's suggestion that Mg-ATP crystallites in mitochondria behave like tiny cyclotrons that drive transmutations such as Na to Mg. The dossier treats that theme separately under Ion Cyclotron Resonance in Biology (Aspden).

What was actually demonstrated

Very little. The patent says the underlying research "demonstrated certain energy anomalies some of which are not yet well understood". The patent text reports no temperatures, no power balance, no calorimetry and no output figures for the FIG. 9 apparatus. The one experiment it describes in any detail is a test rig with a nylon rotor, a d.c. drive motor, an electrical tachometer and a declutching mechanism, so that spin-down times could be compared with the voltage priming. By the patent's own account, the results "did not meet expectation", and any vacuum spin that may have been present showed no evident mechanical coupling to the rotor.

The patent describes its own usefulness modestly: as experimental apparatus for investigating whether anomalous heating occurs. That is a claim to have designed a test rig, not a claim to have found an effect. As Patent as Evidence explains, a grant establishes novelty of design, not that a device works. Later patents that cite this one are mostly ordinary thermoelectric products from companies such as Gentherm and Infineon, plus a few fusion filings. Being cited is routine prior-art practice and implies no endorsement.

Place in the dossier

The patent cites two earlier Aspden thermoelectric patents, US 5,065,085 (assigned to Strachan-Aspden Limited) and US 5,288,336, plus US 5,376,184. It says it complements a rotating-machine design in which a magnet inside a spinning conductor produces the radial field. It fits the pattern described in Aspden Claims vs. Evidence: The Index Dossier Without Data: a detailed theoretical structure supported by argument rather than reported measurement. It is the fusion-and-heat counterpart to the Over-Unity Magnetic Motor Claim (Aspden).

Source notes & attribution
  1. https://patents.google.com/patent/US5734122A/en