In 1949, Westinghouse engineer Joseph Slepian sketched a spacecraft powered by coils, capacitor plates, and radio-frequency electricity. His hypothetical machine appeared to offer roughly 150 newtons of force without expelling propellant, but Slepian supplied the catch himself: its alternating recoil couldn’t carry it on an interplanetary journey. Half a century later, James F. Corum and his collaborators proposed an extension of the Slepian Drive, drawing on electromagnetic field theory, nonlinear circuits, and precision measurement. They wanted to know whether carefully timed changes to the apparatus could turn an exchange of momentum between fields and hardware into a useful propulsion cycle. Their proposal brought nineteenth-century physics into contact with high-power electronics and a demanding question about what an entire machine can push against.

What Is the Slepian Drive?

Slepian knew what an electrical machine could do. As associate director of Westinghouse Research Laboratories, he worked in an industrial setting where forces on conductors and magnetic structures were everyday engineering concerns. His 1949 Electrical Engineering essay began with precisely that familiar machinery. Send current through a wire placed across a magnetic field, and the wire experiences a sideways force. The magnetic structure receives an opposing reaction. In a conventional motor, the stator is secured while the rotor turns, allowing the machine to do useful work without leaving the reaction force unexplained.

He then altered the machine one step at a time. Electromagnetic coils could replace permanent magnets. Alternating current could replace direct current because, when the current and magnetic field reversed together, the force could continue pointing in the same direction. Finally came the intriguing substitution: remove a section of the wire and replace it with two capacitor plates separated by empty space. Alternating electricity would charge and discharge the plates, creating a changing electric field in the gap. Maxwell’s displacement current described that changing field and its magnetic effects, even though electrons weren’t traveling through the gap as they would through a conductor.

Slepian gave his imaginary spaceship substantial specifications. The plates would be one yard square and one foot apart, driven at 100 megahertz. He assumed 1,000 amperes of displacement current and a magnetic field of 5,000 gauss, or 0.5 tesla. The resulting illustrative force estimate was 15 million dynes: 150 newtons, equivalent to about 34 pounds-force. He entertained steering by reversing selected machines and even regenerative braking. These were the numbers and possibilities of an engineering thought experiment, presented with an inventor’s enthusiasm. His original essay nevertheless supplied an unexpected conclusion: yes, an instantaneous force could exist; no, the illustrated machine couldn’t propel the spaceship on its journey.

In his subsequent explanation, Slepian stepped out of the inventor’s role and identified the overlooked interaction. The coils’ changing magnetic field also induced an electric field, which acted on the charges accumulated on the capacitor plates. Its average force opposed and canceled the average force on the magnetic structure. The two forces didn’t cancel at every instant, so the material apparatus could experience an alternating recoil while momentum shifted between hardware and fields. Over a complete cycle, however, the average thrust vanished. Slepian had constructed a puzzle that exposed both the reality of electromagnetic momentum and the difficulty of turning its temporary exchange into propulsion.

How Electromagnetic Fields Store and Transfer Momentum

The physical foundation reaches back to Faraday’s lines of force and Maxwell’s 1873 Treatise on Electricity and Magnetism. Maxwell described tension along field lines and pressure across them, giving the space between electrical objects a mechanical role. The Maxwell stress tensor organizes those directional stresses. Imagine an invisible surface surrounding a capacitor plate or part of a motor: the stresses across that surface help account for the momentum passing into or out of the enclosed region. Some can change the motion of matter; some can change the momentum stored in the fields. The tensor provides a systematic description of those effects.

Energy flow supplied another part of the picture. Poynting’s 1884 work described electromagnetic energy transport, while Heaviside developed related ideas about stresses, energy, and momentum. Corum’s historical account emphasizes Heaviside’s 1886 treatment of time-dependent electromagnetic forces and his explicit identification of field momentum by 1891. The implication is initially surprising: momentum needn’t belong exclusively to visibly moving objects. A combination of electric and magnetic fields can possess momentum even when the apparatus producing them appears stationary. The gap between capacitor plates can therefore contain more than an electric influence waiting to act on a charge; its fields can participate in the system’s momentum balance.

Feynman’s disk paradox makes that idea easier to picture. A freely rotating disk carries a current-producing coil and electrically charged objects around its circumference. When the magnetic field is changed or removed, an induced electric field acts on the charges and can set the disk turning. The angular momentum associated with the original field configuration is transferred to the hardware. The disk illustrates how electromagnetic fields can store rotational momentum and later exchange it with matter, even though no visible flywheel was spinning beforehand.

Graham and Lahoz supplied an experimental connection in a 1980 Nature paper reporting electromagnetic angular momentum associated with independent electric and magnetic fields in the vacuum gap of a cylindrical capacitor. Their measurement made the mechanical implications of that field momentum tangible. A change in the electromagnetic configuration could be associated with a measurable torque. For Corum, this was a reason to investigate electromagnetic momentum as something an engineer might manipulate. The experiment concerned angular momentum and torque, however; it didn’t show a complete spacecraft acquiring sustained linear thrust. That further step would require a different kind of demonstration.

James Corum’s EM Stress-Tensor Space Drive

Corum’s interest in these questions had developed long before the propulsion papers. His resource letter on the Maxwell stress tensor lists a 1965 senior research paper at Lowell Technological Institute on the role of the Lorentz force, followed by a 1967 Ohio State master’s thesis examining the stress tensor in Minkowski spacetime. A 1974 doctoral dissertation explored rotating Lorentz transformations and their applications to electrodynamics. The space-drive proposal grew from this sustained interest in the physical interpretation of electromagnetic forces. Corum approached the subject as an electrical engineer asking what familiar field theory might permit engineers to build.

A 1998 technical memorandum at SARA became the basis of the six-page paper EM Stress-Tensor Space Drive, presented at the 1999 Space Technology and Applications International Forum. Corum’s collaborators were John P. Dering, Philip Pesavento, and Alexsana Donne. They argued that conventional electromagnetic theory already contained an interaction suitable for propellantless propulsion, and asked whether a physical structure could exploit it. Their proposed avenues included short electrical pulses, radio-frequency resonators, standing and traveling waves, and nonlinear materials. The paper described a research direction with several possible implementations, all organized around the controlled exchange of electromagnetic momentum.

The authors focused on the field-momentum term they called the Heaviside force. Ordinary electric and magnetic forces act on charges and currents, but changing field momentum remains meaningful in a vacuum region containing electromagnetic fields. Corum interpreted this as a possible basis for interaction with free space. Standard electrodynamics already includes that momentum; his proposal concerned its engineering use. He drew an analogy with electrical machinery, moving from conduction current in a wire to polarization in a dielectric and displacement current in a capacitor gap. Could a machine exploit those interactions while carrying its field sources along with it?

By 2001, the proposal had become a more detailed laboratory program. AIAA paper 2001-3654, The Electromagnetic Stress-Tensor as a Possible Space Drive Propulsion Concept, listed Corum and six collaborators at the Institute for Software Research in Fairmont, West Virginia. It acknowledged partial NASA support under Contract NAS3-000124 and discussed apparatus, power electronics, measurement strategies, and university research. Its opening was candid: the authors had “no startling breakthrough to disclose.” They were seeking an experimental route from an established property of electromagnetic fields to an ambitious new application.

How Corum Proposed Turning Alternating Forces into Net Thrust

Slepian’s difficulty gave Corum a specific engineering problem. In the fixed capacitor-and-coil arrangement, the electric and magnetic fields were a quarter-cycle out of step, and the momentum exchange alternated. As the configuration repeated, the useful mechanical reaction was followed by a compensating one. Corum proposed changing the apparatus during that exchange. A boundary could move, a capacitance could vary, or a material could respond differently during different portions of the cycle. The objective was rectification: produce a directional average from an alternating effect by controlling more than the applied electrical waveform.

The inspiration came partly from radio engineering. Corum cited R. V. L. Hartley’s work on nonlinear and time-varying reactive systems, including a 1936 paper, and the later Manley–Rowe work of 1956 and 1958. Such circuits can transfer energy among frequencies when an electrical property varies with time or responds nonlinearly. Corum highlighted Hartley’s discussion of a mechanical offset in an AC-excited capacitor and asked whether related behavior could help modify Slepian’s force cycle. These studies established useful ways to understand modulation and frequency conversion. Corum’s proposed application was to coordinate those effects with an electromagnetic momentum exchange.

Appendix II of the 2001 paper illustrated the idea with an idealized changing field region. One boundary remained fixed while another varied at twice the electrical drive frequency. The selected field-momentum contribution, integrated over that changing volume, acquired a nonzero average. Physically, the ambition was to give opposing portions of the cycle different conditions, creating a persistent directional bias. That was a reason to investigate a time-dependent apparatus, but it wasn’t yet a complete vehicle calculation. Evaluating a changing region requires counting momentum transported across its boundary; physically changing that region also involves the forces on the hardware that makes the change.

The proposed second experimental phase made the practical demands clearer. The researchers wanted to begin with their existing 30-kilohertz electrical system, which would require synchronized boundary modulation at 60 kilohertz under the twice-frequency scheme. Later operation at electrical frequencies between 10 and 30 megahertz was contemplated, with new amplifiers and at least several tens of kilowatts anticipated for a detectable effect. The challenge involved field strength, relative phase, and a boundary or parameter response that could follow the required timing. Higher frequency alone wouldn’t establish the proposed rectification. The apparatus had to change in the right way at the right part of each cycle.

Inside the Experiments: Capacitors, Coils, and Power Electronics

The laboratory starting point was an Ampère–Pixii rail motor. In a conducting version, current passes through a movable bridge or axle, and electromagnetic forces move that component along rails. Corum’s group reported operating an ordinary conducting arrangement with both AC and DC excitation. Their proposed modification was to replace the metallic axle between rolling wheels with a capacitor. Comparing that configuration with a conducting baseline could help separate forces on the metallic portions from any additional contribution associated with the dielectric or gap. About 50 amperes had been needed to make the metallic axle roll consistently, helping establish the capacitor experiment’s target current.

Obtaining suitable capacitors constrained the design. The largest available two-plate component they acquired had a capacitance of 20,000 picofarads, or 20 nanofarads, and a dielectric with relative permittivity around 5,000. Other samples covered values from approximately 1,000 down to 10. The team selected 30 kilohertz because operation at 60 hertz couldn’t provide the desired capacitor current within its practical limits. Inductors were built to tune the circuit to resonance. That distinction matters: a large current can circulate reactive energy through a resonant circuit, so the capacitor’s current doesn’t directly specify the power consumed by the apparatus.

The magnetic system also required development. At 50 amperes, the self-generated field was only a few gauss, so the researchers constructed modified spiral Helmholtz coils intended to provide approximately 50–100 gauss. They could vary the phase between the capacitor and coil currents, and they reported checking the magnetic field’s uniformity. Two Class D power generators used full-bridge switching electronics. The DC supply was limited to 300 volts and 150 amperes, while the insulated-gate bipolar transistors were rated for 1,200 volts and 150 amperes. Those figures describe supply limits and component ratings; they don’t establish the voltage across the resonant capacitor or a measured continuous input-power figure.

The intended vacuum limit proved harder to reach directly. The group reported that it couldn’t obtain a suitable parallel-plate vacuum capacitor above approximately 100 picofarads with a gap exceeding a few millimeters. Instead, it planned to compare dielectric samples and extrapolate toward vacuum’s relative permittivity of one. The lowest reported purchased value was still about 10, so this would be an indirect investigation rather than a gradual progression all the way to an empty-gap measurement. A piezoelectric transducer was proposed to measure an oscillating force. The first capacitor apparatus was therefore designed to investigate the interaction even if it failed to produce a steady push.

Measuring Electromagnetic Momentum with a Torsion Balance

A second part of the program pursued a much smaller and more carefully isolated effect. In an appendix to the AIAA paper, West Virginia University physicist Alan Barnes described a torsion-pendulum experiment intended to reexamine electromagnetic angular momentum. The estimated field momentum was sufficient for an equivalent free rotation of only one revolution in approximately 160 years. The experiment wouldn’t wait for that rotation to unfold. Instead, it would use repeated excitation and sensitive optical measurements to detect tiny oscillations. The comparison conveyed how small the underlying effect was and why the instrument’s design mattered as much as the theoretical prediction.

Barnes’s apparatus used a coaxial cylindrical capacitor in a magnetic field. Charging the capacitor changed the angular momentum associated with the overlapping fields. To reduce problems from external power leads and alignment effects, the design placed permanent magnets, battery power, and relevant circuitry onboard the pendulum, with optical control from outside. A thin suspension allowed small torques to produce measurable angular movement. This was a different experimental strategy from the rail apparatus: instead of pursuing a relatively large linear response, it concentrated on separating a tiny angular response from electrical and mechanical influences outside the instrument.

The quantitative targets were demanding. Barnes estimated an angular-momentum content of about five times ten to the minus thirteen kilogram square meters per second, corresponding to roughly 1.25 billionths of a radian per second of equivalent free rotation. The optical readout was described as sensitive to angles much smaller than 0.1 microradian. Driving the pendulum at resonance and measuring its response synchronously could accumulate the effect of repeated excitation at a known frequency. The goal was a one-percent statistical error limited by thermal noise. That was a measurement objective, supported by preliminary checks of the laboratory’s vibration environment, rather than an accuracy already demonstrated in the paper.

At publication, the conceptual and mechanical design was complete and the first pendulum bob was under construction. A second design, supplied through a coaxial feed along the suspension, was being pursued in parallel. The work shows the effort required to investigate even an accepted field-momentum effect in a particular geometry. A successful measurement could strengthen confidence in the electromagnetic interaction and the methods used to isolate it. Establishing a linear space drive would still require another result: a net impulse on the complete apparatus through an operating cycle, with all of its sources, supports, and internal reactions included.

The Slepian Drive and Conservation of Momentum

A propulsion cycle has to do more than produce a force during a selected interval. It must establish the fields, operate whatever switches or moving elements are required, and prepare the apparatus to repeat. Under ordinary electrodynamics, an isolated system whose internal state repeats can’t accumulate momentum if none is carried away or exchanged with its surroundings. A strong, brief push can be canceled by a weaker force acting longer. This makes impulse over the complete cycle the relevant quantity. Corum’s own resource letter explicitly recognizes that momentum conservation applies to the combined mechanical and electromagnetic system.

The moving-boundary example makes this accounting particularly subtle. A local field-momentum contribution integrated over a changing region can have a nonzero average. That selected quantity isn’t automatically the total rate of change of momentum in the region, because the moving boundary contributes to the balance as well. A physical machine introduces additional participants: the moving plate, its actuator, the magnetic structure, the wiring, and the power source. The calculation must establish what happens to each. Corum’s model showed a directional bias in one contribution; a propulsion model had to demonstrate that the complete apparatus retained a useful net effect after the other contributions were included.

The proposed pulsed backup experiment illustrates why the full sequence matters. It used disc capacitors arranged as a Marx bank and a permanent magnetic field of approximately 1,000 gauss. The capacitors would charge in parallel while the apparatus was held stationary, then discharge in series after its stops were released. The researchers anticipated newton-scale forces lasting microseconds, compared with millinewton-scale forces lasting seconds in the rail arrangement. The pulsed apparatus could weigh several hundred kilograms, making a brief displacement harder to observe. But a released kick would also have to be considered alongside the charging stage: the stationary support could receive momentum while the apparatus was restrained. These were proposed operating regimes, not reported thruster measurements.

Materials add another layer. The Abraham–Minkowski debate concerns how electromagnetic momentum is assigned between fields and matter, and Corum treated the differing formulations as experimentally significant. Hidden momentum, discussed in his resource letter through work including Shockley and James’s 1967 research, involves internal energy flow and stresses that don’t appear as obvious bulk motion. Barnett’s 2010 analysis later identified kinetic and canonical roles for the Abraham and Minkowski momentum forms. These issues show why a simple picture of a capacitor pushing against its field can be incomplete. Momentum can reside in material contributions or field-source interactions that aren’t represented by the movement of the one component being measured.

Electromagnetic Stress, Spacetime, and the Quantum Vacuum

Corum placed these engineering questions within a broader view of electromagnetic stress. In relativity, the spatial Maxwell stress tensor belongs to a larger description of electromagnetic energy, momentum, and stress. In general relativity, those physical quantities contribute to gravity. This gave the authors a reason to discuss their proposal in terms of the fabric of spacetime: electromagnetic fields have physical consequences even where ordinary material is absent. The laboratory program, however, concentrated on electromagnetic forces and momentum exchange. The papers didn’t demonstrate that the fields produced a useful gravitational propulsion effect or calculate a working spacecraft trajectory from engineered spacetime curvature.

The language of a stressed elastic medium has a longer history. Maxwell used mechanical pictures to describe tension and pressure in fields, and Corum drew on that tradition to emphasize their physical reality. It helps explain how force can be transmitted through the space between material components. It doesn’t establish that vacuum acts like a stationary rail anchored to the universe. Indeed, the 2001 paper explicitly separated its approach from the old luminiferous-aether theories. The field’s ability to carry stress and momentum was the relevant physical feature; identifying how a complete vehicle would exchange momentum with something beyond its own repeating internal state remained essential.

The 1999 paper also suggested a connection to Casimir stresses and proposals involving zero-point fields. Electromagnetic boundary conditions matter in quantum-vacuum physics, where they can affect measurable forces between objects. Corum regarded the classical stress-tensor approach as a possible bridge toward that wider field of research. The common elements were boundaries, electromagnetic stresses, and their mechanical consequences. This was a suggested extension of the program rather than a second detailed engine design. The capacitor-and-coil apparatus depended on supplied electrical power, and the papers didn’t present a demonstrated Casimir propulsion cycle or a measured source of energy extracted from the vacuum.

There is an established way for an electrical device to exchange momentum with space: emit radiation. An ideally directed kilowatt of outgoing electromagnetic power produces about 3.3 micronewtons of recoil, a small force whose reaction is carried by the radiation. Corum’s proposal sought to exploit stored, reactive fields and their controlled variation rather than rely on that ordinary radiation-recoil mechanism. Any successful interpretation would still need to identify the momentum channel. Radiation, an external field, or surrounding matter can carry the reaction; fields that simply return to their original internal configuration don’t supply an unlimited momentum reservoir.

Did Corum’s Space Drive Demonstrate Net Thrust?

The 2001 paper didn’t demonstrate net propulsive thrust from Corum’s complete space-drive system. It records a program with theoretical analysis, simulations, power electronics, magnetic hardware, and several proposed measurement strategies. It reports ordinary AC and DC rail-motor operation and checks of its coil design, while describing the capacitor-force, rectification, and WVU investigations as work in progress. The researchers had moved beyond a purely verbal suggestion to a set of apparatus concepts and engineering choices, but a successful demonstration of propulsion remained to be established.

A dated follow-up appeared in Marc Millis’s 2004 NASA report, Prospects for Breakthrough Propulsion From Physics. Its Slepian-Drive section described the ISR work and reported that an independent Air Force Academy assessment expected no net propulsive force from the approach. Millis identified two central issues: converting oscillatory effects into a net force and obtaining an external force from internal momentum exchanges. The Air Force conclusion is known here through that NASA report. The assessment focused on the same step that separated Corum’s motivating field physics from the proposed vehicle: whether the complete machine could retain a directional effect.

Millis returned to the subject in his 2005 report, Assessing Potential Propulsion Breakthroughs, naming Corum among researchers exploring electromagnetic momentum for propulsion. He described the conversion into a net propulsive force as unproven and recognized that experiments could still be useful for investigating electromagnetic momentum. By that assessment, the program’s central propulsion question remained unresolved. Its proposed measurement techniques could clarify how fields and materials exchange momentum even if they didn’t deliver a new engine. That was a scientifically meaningful outcome, while the prospect of propulsion required an additional, more demanding result.

A convincing drive demonstration would measure a repeatable net impulse of the complete onboard assembly over full operating cycles. Its field sources and power system would be included, radiation and interaction with the laboratory would be accounted for, and controlled changes in timing would produce the predicted response. Independent apparatus would need to reproduce the effect. Slepian had shown how an appealing motor argument could overlook a reaction; Corum proposed varying the apparatus and investigating whether a useful effect could survive. The decisive question is concrete: after the fields, moving boundaries, and circuits have completed their cycle, what carries the momentum opposite to the spacecraft’s?

References

1. Electromagnetic Space-Ship and Answer to Previous Essay — Joseph Slepian, Electrical Engineering, 1949, pp. 145–146 and 245.
2. EM Stress-Tensor Space Drive — James F. Corum, John P. Dering, Philip Pesavento, and Alexsana Donne, STAIF 1999.
3. The Electromagnetic Stress-Tensor as a Possible Space Drive Propulsion Concept — J. F. Corum and colleagues, AIAA 2001-3654, 2001.
4. Resource Letter: The Classical Maxwell Stress Tensor — Part I — J. F. Corum, T. D. Keech, and S. Kapin.
5. Resource Letter: The Classical Maxwell Stress Tensor — Part II — Annotated historical references.
6. Observation of Static Electromagnetic Angular Momentum in Vacua — G. M. Graham and D. G. Lahoz, Nature 285, 154–155, 1980.
7. The Feynman Lectures on Physics: Field Energy and Field Momentum — Volume II, Chapter 27.
8. Hidden Momentum, Field Momentum, and Electromagnetic Impulse — David Babson, Stephen P. Reynolds, Robin Bjorkquist, and David J. Griffiths, American Journal of Physics, 2009.
9. Resolution of the Abraham-Minkowski Dilemma — Stephen M. Barnett, Physical Review Letters 104, 070401, 2010.
10. Prospects for Breakthrough Propulsion From Physics — Marc G. Millis, NASA/TM-2004-213082, 2004.
11. Assessing Potential Propulsion Breakthroughs — Marc G. Millis, NASA/TM-2005-213998, 2005.