Plasmavolt: The Engineering Promise of Vortex Plasma Fusion

October 13, 2026

Could a carefully organized plasma vortex make nuclear reactions easier to produce and sustain? That’s the engineering question behind the Plasmavolt, an unconventional reactor concept developed by Arie de Geus and described by Michael McDonnough as a potential source of heat, electricity, and manufactured isotopes. Its appeal lies in the possibility of using plasma motion and electrode geometry to concentrate interactions in a small region, rather than relying entirely on a larger confinement apparatus. Established fusion experiments show that organized flows can improve plasma stability and confinement. Whether Plasmavolt achieves those benefits—and whether they explain its reported nuclear products and excess heat—remains the central question.

From Stellar Vortices to a Laboratory Reactor

When McDonnough discussed the Plasmavolt in his 2004 interview, he was speaking as president of Beta Voltaic Industries and describing two related but distinct technologies. One was a proposed battery that would extract electricity from stimulated radioactive decay. The other was a plasma reactor intended to generate heat through low-energy nuclear reactions. In his proposed system, conventional heat engines would convert the reactor’s thermal output into useful power. The extraordinary claim concerned the heat source, not the downstream machinery.

McDonnough credited de Geus with developing the plasma-vortex concept from his interpretation of astronomical observations. In that account, vortices associated with stars and expanding supernova material provided a model for forming elements outside stellar cores. De Geus believed that selected light elements, interacting within a hydrogen-plasma environment, could participate in analogous processes in a laboratory. That was the hypothesis motivating the apparatus; the astronomical analogy doesn’t independently demonstrate that the same reactions occurred inside it.

The electrode materials were central to McDonnough’s description. He discussed hydrogen plasma interacting with light elements, particularly beryllium and boron, and attributed the production of additional hydrogen, helium, and potassium-40 to those interactions. He wasn’t simply describing ordinary hydrogen nuclei colliding until they formed helium. His account involved a plasma interacting with solid material and allegedly transforming the constituent nuclei into different products.

That gave the project two possible routes to practical value. A sufficiently productive reaction could supply heat, while an unusual isotope product could have research or industrial applications. McDonnough also envisioned using the manufactured potassium-40 in a separate nuclear battery. These possibilities should be evaluated independently: isotope manufacturing could be useful without net electricity production, while a reproducible energy surplus would deserve investigation even if the initial explanation of its origin proved incorrect.

Inside the Plasmavolt

De Geus’s 2002 patent application, WO2002031833A1, describes a hydrogen reactor with electrodes, a tungsten ionization coil, a quartz section, and a surrounding heat-collection jacket. Its stated startup conditions are 1–1.5 Torr, 500–550 volts DC, and approximately 900 volts of superimposed AC. After startup, the described pressure rises to 2.5–3 Torr, with a current of 2–5 milliamperes. These are specifications reported by the inventor, rather than operating conditions independently verified here.

A separate patent, NL1033078C1, published in 2008, specifies 15 centimeters between electrodes, an 1,800-volt trapezoidal or square waveform, and pressure of one-tenth to one-eighth atmosphere. That corresponds to approximately 10.1–12.7 kilopascals, or 76–95 Torr. Those figures belong to the later description, not the earlier low-pressure tests. Combining them into one standardized prototype would conceal potentially important differences between the proposed operating regimes.

The later dimensions permit a useful engineering calculation. Dividing 1,800 volts by 0.15 meters gives an average potential gradient of 12,000 volts per meter, or 120 volts per centimeter. A singly charged ion crossing that entire potential difference without collisions could gain 1.8 kiloelectronvolts of energy. But that’s a hypothetical single-pass energy, not a measurement of the plasma’s temperature. The actual energy distribution would depend on collisions, local electric fields, and where ions were created.

These descriptions establish a starting point for technical analysis, but they don’t provide a complete, common operating record. Waveform frequency, duty cycle, electrode dimensions, gas flow, plasma density, and temperature distributions would all matter. The interview also sometimes calls the reactive electrode an anode, whereas the patent descriptions assign the transmutation role to the cathode. That difference needs clarification rather than silent correction. A reconstruction must identify which apparatus and which operating configuration it represents.

Compression Without Shrinking the Chamber

A reaction chamber doesn’t have to shrink for local compression to occur. With a fixed inventory of particles, the average density remains fixed, but the internal distribution can change substantially. As an illustrative example, gathering 10% of the particles into 1% of the chamber’s volume would create a region with ten times the chamber-average density. The surrounding region would become correspondingly less dense. This is the physically reasonable core of the proposed concentration mechanism.

However, narrowing a plasma stream doesn’t necessarily increase its density. For steady flow without particle creation or loss, particle conservation requires the product of density, flow speed, and cross-sectional area to remain constant. If the area decreases tenfold while speed increases tenfold, the density needn’t change. If the flow converges without an equivalent acceleration, density can rise. A luminous filament becoming narrower therefore tells only part of the story; its velocity and particle distribution determine whether genuine compression accompanies the change.

Density matters because, for ordinary two-body fusion reactions at unchanged particle-energy distributions, the local reaction rate depends on the densities of both reacting populations. Increasing both densities tenfold can increase reactions per unit volume by a factor of one hundred. That doesn’t automatically multiply the entire reactor’s output by one hundred, because the reacting volume may shrink. Nor does fast circulation automatically create high collision energies: neighboring ions moving together can have high laboratory-frame speeds but little additional velocity relative to one another.

De Geus’s own explanation complicates a straightforward compression story. The later patent explicitly describes density falling inside the vortex, extends that account to an underlying “aether,” and invokes unconventional hydrogen states. A lower-density core could coexist with denser peripheral or electrode regions, but the required spatial measurements aren’t supplied. Local compression is therefore an engineering hypothesis worth testing—not something that can simply replace his stated mechanism. The crucial question is where particles accumulate, how energetic they become, and how long they remain there.

Why Organized Plasma Flow Deserves Attention

The broader idea of using organized motion to improve confinement has established experimental counterparts. In a 2010 study of the Gas Dynamic Trap, researchers described vortex confinement produced by electrically driven shear flows. Biased end plates and limiters helped create a flow structure surrounding the plasma core, and experiments and simulations indicated improved confinement. The important effect wasn’t that every particle spiraled toward a single point. It was that the driven flow helped control the instabilities and transport processes that otherwise degraded the plasma.

A different example comes from the University of Washington and Lawrence Livermore National Laboratory’s Fusion Z-pinch Experiment, or FuZE. Its 2019 results included approximately 200-kiloampere pinch currents, plasma temperatures of 1–2 keV, densities near 10²³ particles per cubic meter, and a pinch radius of approximately 3 millimeters. Neutron emission persisted for about 5 microseconds, with an estimated 125,000 ± 45,000 neutrons per pulse. The plasma’s current supplied magnetic compression, while a sheared axial flow helped stabilize the column. That flow wasn’t simply a rotating vortex.

Rotation itself plays a more direct role in the University of Maryland’s Centrifugal Mirror Fusion Experiment. A 2025 paper reported a maximum average neutron-emission rate of approximately 8.4 million ± 700,000 neutrons per second at a set voltage of 65 kilovolts. Interpretive modeling inferred an average ion temperature of about 720 electronvolts and electron density near 2.8 × 10¹⁸ per cubic meter for the highest-performing discharge. The distinction matters: the neutron measurements were experimental observations, while those temperature and density values were model-derived.

These experiments support a specific engineering proposition: appropriately designed flows can help maintain useful plasma conditions. They don’t establish that Plasmavolt reaches comparable conditions or produces its claimed nuclear transformations. They also show why “working with nature” needn’t mean eliminating external power. Engineers supply energy and choose fields, boundaries, and flow profiles so that the plasma’s resulting behavior assists confinement. The possible advantage is better performance for the effort expended, not the disappearance of the energy requirements.

A Tabletop Fusor—or a Different Kind of Reactor?

The comparison with a tabletop fusor is useful because conventional inertial electrostatic confinement, or IEC, also concentrates particle trajectories inside a fixed-volume chamber. In a gridded fusor, an electrostatic potential accelerates positive ions inward through openings in an inner electrode. Wisconsin’s IEC research describes this as electrostatic focusing toward a dense central region, with ion energies in the tens-of-keV range. Compact apparatus can therefore produce genuine nuclear reactions without resembling a large tokamak, although producing reactions and producing net power are different achievements.

The Plasmavolt descriptions place more emphasis on interaction with electrode material. A tapered cathode appears in de Geus’s 2002 drawings, while McDonnough describes the plasma interacting directly with a reactive electrode. Concentrating ions onto a solid surface isn’t equivalent to allowing them to pass through an open grid into a central reaction region. For the proposed apparatus, an experiment would need to determine whether the significant interactions occur in the plasma, at the surface, within the electrode, or across several of those regions.

Scale introduces another distinction. As a purely geometrical example, shrinking every dimension of a device by a factor of ten reduces its volume by 1,000, but its surface area by only 100. Surface area per unit volume therefore rises tenfold. That could help remove heat from an electrode-driven process, while making wall interactions more consequential for a plasma intended to retain its energy. A smaller version wouldn’t necessarily preserve the balance of heating, confinement, and material loss found in a larger apparatus.

McDonnough’s own descriptions reflect that tension. In 2004, he said the space and material requirements for useful output made larger applications, such as boats and stationary installations, more practical than small cars. In the later APEC discussion, he recalled an approximately 85-gram anode and proposed dividing the concept into much smaller vortex devices. That figure referred to an electrode, not the complete reactor’s mass. A small experimental cell and a complete power system—with pumps, cooling, controls, and conversion machinery—remain different engineering objects.

What the Reported Energy Gain Establishes

McDonnough presented excess heat as an important part of the Plasmavolt’s appeal, but the supplied interviews don’t provide a complete numerical energy balance. When asked about experimental performance in 2004, he declined to disclose details because of commercial confidentiality. The later discussion refers to significant heat gain without giving the input waveforms, calibrated thermal output, run duration, or measurement uncertainty. His testimony identifies the claim to investigate, but it doesn’t supply the records needed to independently verify it.

The 2002 patent states a reactor input of 1–3 watts and photoelectric output of 20–120 milliwatts. It cites Applied Technical Services report M14343, dated March 1, 2001, in connection with testing and describes a thermal-output/input ratio of 2.19 for one region. It separately claims approximately 5:1 thermal output in April 2001. Those are the inventor’s accounts of results. The underlying ATS report would be needed to determine precisely what the laboratory measured and concluded.

Electrical accounting is especially important for an apparatus using superimposed AC and DC. Input energy must be calculated from voltage and current measured together over time, rather than by assuming that a nominal voltage multiplied by an average current captures everything. Mathematically, the average of the instantaneous product isn’t generally equal to the product of two separate averages. Any auxiliary ionization supply, bias circuit, or stored electrical energy also needs to be included within a clearly defined measurement boundary.

The thermal side needs an equally explicit accounting of heat stored in the apparatus, heat leaving through the cooling system, radiation, and material changes. A useful test would compare the reactor with a calibrated electrical heater over the same output range and operating duration. It should also determine whether electrode consumption or another conventional energy source could explain the result. A verified energy surplus would be an important finding, but attributing that surplus to nuclear reactions would require a second body of evidence.

The Potassium-40 Claim

The most striking material claim is McDonnough’s statement that an analyzed sample contained as much as 97% potassium-40. For comparison, NIST lists potassium-40 at approximately 0.0117% of natural potassium. If the 97% figure represents an isotope fraction among potassium atoms, it would be roughly 8,300 times the natural atom fraction. That would be an extraordinary result. However, the transcript doesn’t clearly establish whether the reported percentage referred to elemental composition, isotope composition, or some other basis.

During the APEC Q&A, McDonnough attributes the figure to analysis involving a scanning electron microscope and X-ray spectroscopy. Ordinary SEM energy-dispersive X-ray spectroscopy identifies elements through their characteristic X-rays. It can establish that a region contains potassium, but it doesn’t establish that the potassium is predominantly potassium-40. Consequently, the measurement method he describes doesn’t substantiate the isotope-specific conclusion. An additional assay might resolve that gap, but no such report is included in the supplied record.

Established nuclear data provide another way to examine the claim. Potassium-40 has a half-life of approximately 1.25 billion years, and its decay includes a characteristic gamma emission at approximately 1,460.8 keV, occurring in about 10.3% of decays. It isn’t a gamma-free isotope, despite that characterization in the interview. Using its half-life and atomic mass, one milligram of pure potassium-40 would have an activity of approximately 264 decays per second. That calculation supplies a quantitative expectation for a retained sample.

A convincing investigation would combine isotope-sensitive analysis with calibrated radiometric measurements and a documented starting inventory. The question isn’t merely whether potassium appears after a discharge, but whether additional potassium-40 has been produced beyond everything initially present. Electrode material, deposits, chamber components, and sample preparation would all need to be accounted for. Ideally, the material assay and calorimetry would come from the same experimental run, allowing the claimed reaction products and energy release to be evaluated together.

From Reactor Heat to Useful Electricity

McDonnough’s proposed power-conversion architecture was comparatively conventional: collect heat and operate a Rankine or organic Rankine cycle. In an organic Rankine system, heat vaporizes a working fluid, the vapor expands through a turbine or another expander, and a generator converts the mechanical output into electricity. The fluid then condenses and returns through the cycle. This is a separate engineering subsystem; proving an unusual heat source wouldn’t remove the need to design and operate it efficiently.

Temperature would be just as important as the amount of heat produced. As an illustrative calculation—not a reported Plasmavolt operating point—a heat engine working between a 200°C hot reservoir and a 30°C cold reservoir has a maximum Carnot efficiency of approximately 35.9%. A real system would achieve less. The example shows why the temperature of a glowing discharge can’t substitute for a measurement of the useful heat delivered to an external working fluid.

An energy-surplus claim also doesn’t automatically imply a self-powered generator. Suppose a hypothetical reactor consumes 100 watts of electricity and delivers 220 watts of usable heat. A conversion system operating at 20% efficiency would generate only 44 watts, leaving a 56-watt shortfall before additional auxiliary loads. At that efficiency, a thermal-output/input ratio greater than five would be required to exceed the reactor’s own electrical demand. Direct electrical output or useful gas production could alter the calculation, but each contribution would have to be measured separately.

The proposed nuclear battery introduces another independent requirement. A long-lived isotope can contain substantial eventual decay energy while releasing it too slowly for a demanding electrical load. McDonnough’s later presentation acknowledges that he hadn’t found an effective way to stimulate beta decay while obtaining more energy than the stimulation consumed. That materially qualifies the battery proposal. It doesn’t settle the plasma-reactor question, but it prevents the two technologies from being treated as though success in isotope production would automatically deliver a practical high-power battery.

The Experiment That Would Move the Story Forward

The most productive next step would begin with the existing record: the original laboratory reports, retained samples, apparatus drawings, and operating logs. Researchers would need to establish which configuration produced each claimed result and whether the heat and isotope measurements were made during the same runs. Resolving those questions could prevent a new experiment from combining incompatible details from different versions. It could also reveal whether the historical observations were stronger—or more limited—than the interviews convey.

The plasma experiment should then ask what changes when the organized flow is established. High-speed imaging could document the discharge’s structure, while independent density and velocity measurements would test whether material genuinely converges. Suitable spectroscopic and other plasma diagnostics would be needed to distinguish ion energy, electron energy, and bulk motion. The decisive comparison would hold total input power and relevant operating conditions as closely matched as possible, then measure whether the proposed vortex improves density, stability, residence time, or losses.

The nuclear and thermal investigations should remain independently interpretable. Calorimetry would test for an energy surplus; material analysis would test for newly produced isotopes; radiation measurements would examine expected emissions and unexpected products. Control runs and independently analyzed samples would help distinguish a reproducible effect from contamination or measurement bias. The apparatus would need appropriate laboratory controls for its electrical, gas-handling, material, and radiation characteristics, rather than relying on an assumption that an unconventional reaction must be harmless.

Plasmavolt’s strongest engineering promise is therefore narrower—and more testable—than the full vision of a revolutionary generator. An organized discharge might produce a more useful plasma configuration, a reproducible material transformation, or an unexplained energy surplus. Each would represent a different finding, with different consequences. The established fusion examples justify investigating what carefully controlled flow can accomplish. The future of the Plasmavolt claim depends on showing, quantitatively, what its particular vortex does that the same apparatus without that behavior cannot.