Why Perpetual Motion Fails - and What "Free Energy" Actually Means in Physics

What People Mean by "Free Energy" - and the Ambiguity Behind It

When most people search for "free energy" they are not asking about chemistry. They mean a machine that runs indefinitely without fuel, or one that returns more electrical output than was put into it: a perpetual motion machine, or an "over-unity" device. The promise made for such devices is energy from nothing - no fuel, no input, no limit. No such claim has ever survived independent measurement, and the physics explains why.

The confusion is compounded because similar words get attached to things that are real. Sunlight and wind cost nothing at the point of collection, so a solar panel is loosely described as making "free" electricity. It does no such thing: it converts a small fraction of a flux arriving from the Sun. Exotic-sounding physics is invoked too - zero-point energy, dark matter, antimatter - but each of these is studied inside the same conservation laws that rule out over-unity devices. Borrowing the vocabulary of frontier physics does not suspend its accounting, and that rule applies to every organisation using it, this one included.

The Technical Term: Gibbs and Helmholtz Free Energy

In thermodynamics, free energy is a precise technical quantity: a bound on the work a process can deliver under stated constraints. The Helmholtz free energy, A = U − TS, named after Hermann von Helmholtz, bounds the total work obtainable at constant temperature and volume. The Gibbs free energy, G = H − TS, developed by Josiah Willard Gibbs in the 1870s in his work on chemical equilibrium, bounds the non-expansion work - electrical work, for instance, which is the quantity that actually matters for any energy-conversion device - at constant temperature and pressure. A negative change in Gibbs energy tells a chemist that a reaction can proceed spontaneously under those conditions.

The key word is available, not costless. Free energy is not a hidden reservoir; it is a ceiling on extractable work. The TS term accounts for heat exchanged with the surroundings at temperature T, and depending on the sign of the entropy change it can lower that ceiling or raise it. In an entropy-increasing reaction the extractable work can exceed the system's own enthalpy drop - the oxidation of carbon to carbon monoxide in a solid-oxide fuel cell is the standard example, and it is why endothermic reactions can run spontaneously at all - but only because heat flows in from the surroundings to make up the difference. The books balance exactly, every time. When a physicist says "free energy", they are describing a budget with an external ledger, not a windfall, and no process delivers more total energy than it takes in.

The First Law: Energy Is Converted, Never Created

The first law of thermodynamics states that the total energy of an isolated system is constant. Energy changes form - chemical to thermal, kinetic to electrical - but the total never changes: it can be neither created nor destroyed. Established in the 1840s through the work of Julius Robert von Mayer, James Prescott Joule and Helmholtz, it was shown by Emmy Noether in 1918 to follow from a deeper symmetry: the laws of physics do not change over time.

Physicists take conservation seriously enough to postulate new particles rather than abandon it. In 1930 Wolfgang Pauli proposed an unseen, electrically neutral, very light particle to explain why the electrons emitted in beta decay carried a continuous spread of energies rather than a single fixed value. The apparently missing energy had to be going somewhere. The name neutrino was coined by Edoardo Amaldi and popularised by Enrico Fermi, who built the particle into his 1933-34 theory of beta decay - a theory now understood as the low-energy limit of the electroweak sector of the Standard Model, the modern description of the weak nuclear force. Clyde Cowan and Frederick Reines detected the particle in 1956. Conservation held, as it has every time it has been tested.

Two later results are worth stating carefully here, because they are frequently misused in energy marketing. In 2015 the Nobel Prize in Physics went to Takaaki Kajita and Arthur B. McDonald for the discovery of neutrino oscillation, which established that neutrinos have non-zero mass - at least two of the three mass states - though the absolute scale remains unmeasured. In 2017 the COHERENT collaboration reported the first observation of coherent elastic neutrino-nucleus scattering, the faint recoil a neutrino imparts to a nucleus as a whole. Both are significant physics, and neither has any bearing on whether useful power can be drawn from neutrinos. The governing number is the interaction cross section: of order 10⁻⁴⁴ cm² at MeV energies. That is why a neutrino from one of the natural neutrino sources, such as the Sun, can cross the entire Earth with roughly a one-in-10¹¹ chance of interacting at all, and why the neutrino power deposited in a kilogram of ordinary matter is on the order of 10⁻²⁰ watts - a shortfall of some twenty orders of magnitude against anything useful, not an engineering hurdle. Kajita, McDonald and the COHERENT collaboration have no connection with the publisher of this page, and nothing in their work supports or endorses any energy-conversion programme, ours included.

The Second Law, Perpetual Motion and What Patent Offices Do

The second law adds that the entropy of an isolated system never decreases. In practice, heat flows from hot to cold on its own, and no cyclic engine converts heat entirely into work. Carnot's limit puts a number on it: the maximum efficiency of a heat engine working between two temperatures is 1 − Tc/Th, with Tc and Th measured in kelvin, that is, on the absolute temperature scale. Inserting Celsius values produces nonsense, including impossible efficiencies above one.

That yields the standard taxonomy. A perpetual motion machine of the first kind would produce work from nothing, violating the first law. One of the second kind would draw work continuously from a single heat reservoir at uniform temperature with no waste heat: it respects energy conservation perfectly and violates the second law - which is why energy conservation alone is never a sufficient defence of an energy claim. A third kind would merely eliminate all friction and dissipation so that motion continues indefinitely; persistent currents in superconductors approach this in a narrow sense, but no net work can be extracted from them, which is precisely the point.

Patent authorities treat the matter as settled. United States examination practice singles out perpetual motion as a case where an examiner may require a working model, and such applications fail on utility and enablement. The UK Intellectual Property Office refuses applications that are contrary to well-established physical laws. The European Patent Office rejects them for insufficient disclosure and lack of industrial applicability. Offices do not grant patents for machines whose operation would require breaking these laws. It is worth being equally clear about the converse: a granted patent establishes novelty and formal sufficiency of disclosure, not that a device works, is efficient, or produces the output claimed for it. That caveat applies to every patent number cited in this field, including any filed by the publisher of this page.

Historic Claims That Failed, and Why

The dream is old. Around 1150 the Indian mathematician Bhāskara II described an overbalanced wheel with mercury-filled spokes, expecting the shifting weight to keep it turning. It cannot: gravity is a conservative force, so the net work around one full revolution is exactly zero even in the ideal frictionless case - and with friction the balance is negative, so the wheel slows and stops. In 1618 the English physician Robert Fludd published a design for a closed-loop water mill in which a screw pump would lift the very water that drove it. Both were sincere errors, made before energy conservation was understood.

Not every case has been judged an honest error. From 1872 John Worrell Keely of Philadelphia raised substantial investment through the Keely Motor Company for a motor said to run on "etheric force". After his death in 1898, investigators who examined his workshop reported finding concealed compressed-air machinery. The episode is documented in contemporary press accounts and remains a study in how a compelling demonstration can conceal a mundane power source.

A different case, and one where no fraud was alleged or found: in 1989 Martin Fleischmann and Stanley Pons announced excess heat from palladium electrodes in heavy water, the episode remembered as "cold fusion". This was not an over-unity claim. It proposed a nuclear energy source, which is fully compatible with conservation of energy, and the announcement was made in good faith. It failed on a different test. A United States Department of Energy review later that year found the evidence unpersuasive, and laboratories worldwide could not reliably reproduce the result. That is the same test any new energy claim must face - including the research described further down this page, which has not yet faced it - and it is worth noting that Fleischmann and Pons at least published a measurement for others to attempt to replicate.

What Is Real: Harvesting Energy From an Open Environment

None of this makes ambient energy a fantasy. The Earth's surface is bathed in flux, and capturing part of it is ordinary engineering. Photovoltaics convert sunlight; turbines convert moving air; geothermal plants tap crustal heat; thermoelectric generators exploit temperature differences through the Seebeck effect; piezoelectric elements convert mechanical vibration; radio-frequency harvesters recover a trickle of ambient transmitter power. Every one of them is flux-driven or gradient-driven: it intercepts something arriving from outside, or sits across a measurable temperature difference. None of them draws work from a bath at uniform temperature, because nothing can.

The decisive point is that these are open systems: they intercept an external flux, they do not create one. That is why they are bounded by efficiency limits rather than exempt from them. The Shockley-Queisser detailed-balance limit caps an ideal single-junction cell at roughly 33% under standard sunlight, including for silicon's band gap; the often-quoted figure of about 29% for silicon is the tighter limit obtained once unavoidable Auger recombination and free-carrier absorption are included as well. Betz's law caps a free-stream wind turbine at 59.3% of the kinetic energy flux through its swept area.

Magnitudes matter more than lists. Sunlight delivers on the order of 1,000 watts per square metre at noon; ambient radio-frequency fields are typically in the microwatts per square metre; cosmic rays deposit real energy in the atmosphere, but the secondary flux reaching the ground is on the order of a millionth of a watt per square metre. Real is not the same as useful, and a source six orders of magnitude below sunlight does not become practical by being described enthusiastically.

One clarification prevents a common misreading. No device delivers more total energy than it receives. Some ratios called "efficiency" can exceed 100% - a heat pump's coefficient of performance of 4, or the wall-plug efficiency of an LED operating with electroluminescent cooling - but only because they measure output against one chosen input while heat quietly enters from elsewhere. Total output never exceeds total input. Over-unity means exactly that impossible thing, and genuine energy conversion always has a traceable source.

A Checklist for Spotting Over-Unity and "Free Energy" Pseudoscience

Anyone marketing a "free energy" device deserves determined scepticism. The six tests below are the ones working scientists apply to any extraordinary energy claim. Nobody should pay for an energy device that has not passed them.

These tests apply to the publisher of this page too, so here is the honest scorecard before the next section rather than after it. Neutrinovoltaic has not been independently replicated by a laboratory with no financial stake. No closed-loop, self-powering demonstration has been published. No peer-reviewed measurement of ambient-flux conversion in our material has been published for scrutiny. Until those boxes are ticked, scepticism is the correct position on neutrinovoltaic, and the section below says so in plain terms.

• Independent replication: has a laboratory with no financial stake rebuilt the device from a published description and measured the same result?
• A complete energy audit: are all inputs accounted for - mains supply, batteries, pre-charged capacitors, compressed gas, temperature gradients, ambient light?
• A closed-loop demonstration: does it power itself from its own output for days on end, with every external supply physically removed, under independent observation?
• Honest measurement: pulsed or non-sinusoidal waveforms require true-RMS instruments; naive meter readings are the single most common source of apparent "excess" power.
• Peer review and disclosure: are the methods, uncertainties and raw data published for scrutiny, or are there only videos and press releases?
• Red flags: appeals to suppression by industry, refusal to permit inspection, demonstrations that cannot be independently examined, and the vocabulary of physics used with no equations or data behind it.

Disclosure: Where Our Own Neutrinovoltaic Research Stands

This page is published by Neutrino Energy Group, a Berlin-based research company, and neutrinovoltaic is our own research programme. We have a commercial interest in this field, which is why the checklist above is written to apply to us. Unlike the harvesting technologies described earlier, neutrinovoltaic has not been independently validated and does not belong on that list.

Whether any useful electric current can be drawn from ambient environmental flux is an open question. Our programme is investigating it in a thin-film multilayer material combining graphene and silicon. No peer-reviewed measurement of such conversion has been published, and the hypothesis remains untested by independent laboratories. The programme studies several ambient sources together - thermal gradients relative to the surroundings (not ambient heat at equilibrium, which the second law rules out), cosmic and other radiation, and ambient radio-frequency waves alongside neutrinos. Two formulations used in earlier versions of our own material are corrected here explicitly: "thermal radiation of the environment" and "electromagnetic fields" do not work as source claims. A body in thermal equilibrium with its surroundings emits and absorbs equally on average, and static fields transport no energy. Any current observed would therefore not be attributable to neutrinos alone, and we do not claim otherwise: given the cross sections quoted above, the neutrino contribution to any measured power would be negligible. Ambient heat at a single uniform temperature contributes nothing either. Extracting work from a thermal bath in equilibrium is precisely the perpetual motion machine of the second kind described earlier, so any thermal contribution would require a measurable temperature difference across the device.

Neutrinovoltaic is an in-development research programme - not a proven technology, not a commercially available product, and not a promise of household power. No free-energy, over-unity or perpetual-motion claim is made, and none could be sustained: any output would be bounded by both laws of thermodynamics, meaning energy taken from the surrounding environment and never from nothing (first law), and nothing extractable from ambient heat at a single uniform temperature (second law). The laws of thermodynamics are not on trial here; they bound this research exactly as they bound everything else. What measurement has to settle is the far narrower question of whether any usable power density can be reached within those bounds.

Neutrino Energy Group publishes no performance figures, efficiency ratings or availability dates for neutrinovoltaic, because none have been independently verified. Claims in this field are settled by independent measurement, which neutrinovoltaic has not yet undergone; see how neutrinos are detected and the instruments used by neutrino observatories for what measurement at this scale actually involves. For the terminology used across these pages, start with the neutrino glossary and our overview of new energy technology.