Appearance
7.9 — Nothing Is Ever Still
Cool a piece of matter towards absolute zero. Classical physics says every atom eventually stops. Chapter 3.6's third law says you can never quite get there, but the picture is still of motion dwindling towards nothing.
It does not dwindle to nothing. At absolute zero, atoms in a crystal are still moving, with an amplitude you can compute. Helium refuses to freeze at all under its own vapour pressure, no matter how cold you make it, because the residual motion is enough to shake the crystal apart. And empty space — space with nothing in it whatsoever — has fields fluctuating in it, and those fluctuations push on real objects with a force that has been measured.
This chapter is about that residual, irreducible motion: where it comes from, what it does, and where the energy inside an atom actually is.
Zero-point energy
Chapter 7.4 solved the harmonic oscillator and got:
E_n = \left(n+\tfrac{1}{2}\right)\hbar\omega
The lowest state is E_0 = \frac{1}{2}\hbar\omega, not zero.
Chapter 7.5 showed why, in two lines. Sitting still at the bottom of the well would mean \Delta x = 0 and \Delta p = 0 simultaneously, and \Delta x\Delta p \geq \hbar/2 forbids it. The oscillator settles at the compromise that minimises the total.
E \approx \frac{(\Delta p)^2}{2m}+\frac{1}{2}m\omega^2(\Delta x)^2 \geq \frac{\hbar^2}{8m(\Delta x)^2}+\frac{1}{2}m\omega^2(\Delta x)^2
Minimising over \Delta x gives (\Delta x)^2 = \hbar/2m\omega and E_{\min} = \frac{1}{2}\hbar\omega. The exact answer, from an inequality.
How big is it?
A carbon atom in a diamond lattice. Diamond's Debye temperature is 2230 K, so its characteristic vibrational frequency is:
\omega = \frac{k_B\Theta_D}{\hbar} = \frac{(1.381\times10^{-23})(2230)}{1.055\times10^{-34}} = 2.92\times10^{14}\ \text{rad/s}
E_0 = \frac{1}{2}\hbar\omega = \frac{1}{2}(1.055\times10^{-34})(2.92\times10^{14}) = 1.54\times10^{-20}\ \text{J} = 0.096\ \text{eV}
Amplitude:
\Delta x = \sqrt{\frac{\hbar}{2m\omega}} = \sqrt{\frac{1.055\times10^{-34}}{2(1.99\times10^{-26})(2.92\times10^{14})}} = \sqrt{9.08\times10^{-24}} = 3.0\times10^{-12}\ \text{m}
3 picometres, at absolute zero, against a carbon–carbon bond length of 154 pm. About 2 % of the bond length, jiggling forever, in a perfect crystal at zero kelvin.
And it is measurable. X-ray diffraction from a cooled crystal shows the Bragg peaks weakened by exactly the amount this displacement predicts — the Debye–Waller factor. Extrapolate to absolute zero and the weakening does not go to zero.
Helium never freezes
Helium is the clearest demonstration, and it is a bulk-scale quantum effect visible in a laboratory.
Two things make helium extreme. It is very light, so \omega = \sqrt{k/m} is large and E_0 = \frac{1}{2}\hbar\omega is large. And it is a closed-shell noble gas, so the attraction between atoms is the weakest of any element — a van der Waals well only about 10^{-3} eV deep.
Compare:
E_0 \approx 7\times10^{-4}\ \text{eV per atom}, \qquad E_{\text{binding}} \approx 9\times10^{-4}\ \text{eV per atom}
The zero-point motion is comparable to the binding. The atoms shake themselves out of any lattice you try to form.
So helium remains liquid down to absolute zero at atmospheric pressure. It is the only element that does. To freeze it you must apply 25 atmospheres, squeezing the atoms into a well deep enough to hold them against their own jiggling.
This is not an exotic laboratory curiosity — it is why liquid helium exists at all, and liquid helium is what cools every MRI magnet, every superconducting accelerator magnet, and most quantum computers.
Where it shows up
Isotope effects in chemistry. Chapter 7.4 mentioned this; here is the mechanism. A C–H bond and a C–D bond have the same electronic potential well, but deuterium is twice as heavy, so \omega is \sqrt{2} smaller and its zero-point energy is lower. The C–D bond therefore starts deeper in the well and needs more energy to break — typically 5 kJ/mol more, which at room temperature slows a reaction by a factor of 6 or 7. Measuring this ratio tells a chemist whether breaking that particular bond is the slow step.
It also changes physical properties. Heavy water freezes at 3.8 °C rather than 0 °C, and boils at 101.4 °C, entirely because of zero-point energy differences in the hydrogen bonding.
Neutron stars. The neutrons are so densely packed that their zero-point motion — which is what degeneracy pressure is, in a different language (Chapter 7.7) — supports the entire star against gravity.
Nuclear fusion in the Sun benefits from it: the protons' zero-point motion helps them tunnel through the Coulomb barrier (Chapter 7.4).
The quantum vacuum
Now push the same argument to fields.
Chapter 4.7 showed that light is a wave in the electromagnetic field. Quantum field theory treats each mode of the field as a harmonic oscillator — one for every wavelength and direction and polarisation.
Every oscillator has \frac{1}{2}\hbar\omega in its ground state. So the ground state of the electromagnetic field — the vacuum, with no photons in it at all — has:
E_{\text{vac}} = \sum_{\text{modes}}\frac{1}{2}\hbar\omega
Empty space contains fluctuating fields. Not because anything is there, but because "no photons" is not the same as "no field", any more than "an oscillator in its ground state" is the same as "an oscillator at rest".
This is not a metaphor. The consequences are measured.
The Casimir effect
Hendrik Casimir, working at Philips in 1948, asked what happens if you put two uncharged conducting plates very close together in a vacuum.
The argument. Between the plates, the electromagnetic field must vanish at both conducting surfaces, so only standing waves that fit are allowed — exactly the box condition of Chapter 7.4. Outside, every wavelength is available.
So there are fewer modes inside than outside. Fewer modes means less zero-point energy density inside, and the imbalance pushes the plates together.
The force per unit area:
\boxed{\frac{F}{A} = -\frac{\pi^2\hbar c}{240\,d^4}}
Note what is in it: \hbar, c, and the separation. No charge, no material property, no coupling constant. It is a pure vacuum effect.
The d^{-4} makes it negligible far apart and large very close.
Worked numbers.
At $d = 1\ \mu$m:
\frac{F}{A} = \frac{\pi^2(1.055\times10^{-34})(3\times10^{8})}{240(10^{-6})^4} = \frac{3.124\times10^{-25}}{2.4\times10^{-22}} = 1.30\times10^{-3}\ \text{Pa}
About 1.3 millipascals — a hundred-millionth of atmospheric pressure. Small.
At d = 10 nm:
\frac{F}{A} = \frac{3.124\times10^{-25}}{240(10^{-8})^4} = \frac{3.124\times10^{-25}}{2.4\times10^{-30}} = 1.3\times10^{5}\ \text{Pa}
1.3 atmospheres. At ten nanometres the vacuum presses on the plates as hard as the air does.
Measured. Steve Lamoreaux in 1997 used a torsion pendulum with a sphere and a plate, and matched theory to 5 %. Later experiments with atomic force microscopes have reached about 1 %.
And it matters industrially. Micro-electromechanical systems — the accelerometers and gyroscopes in your phone, the mirror arrays in projectors — have moving parts separated by micrometres. Casimir forces contribute to "stiction", where parts stick together and will not release, which is a real failure mode designers must account for. The vacuum is a mechanical engineering problem.
A caution about interpretation. The Casimir force can also be derived without ever mentioning vacuum energy, as a retarded van der Waals force between the atoms in the two plates. Both derivations give the same answer. So the experiment confirms the calculation, and whether it proves the vacuum "really contains" energy is a matter of some debate. What is not in doubt is that the prediction works.
The Lamb shift
Chapter 7.6 noted that the 2s and 2p levels of hydrogen, exactly degenerate in both Schrödinger's and Dirac's theories, are actually split.
Willis Lamb and Robert Retherford measured it in 1947 using microwave techniques developed for wartime radar. The splitting is 1057 MHz — about 4\times10^{-6} eV, a part in 10^{7} of the level energy.
The cause is the vacuum. The electron is continuously buffeted by vacuum fluctuations of the electromagnetic field, which smear its position by about 10^{-12} m. That smearing averages the Coulomb potential over a small region, and since the 2s state has non-zero density at the nucleus while the 2p does not, the two are affected differently.
The Lamb shift is why quantum electrodynamics exists. The measurement was announced at the Shelter Island conference in June 1947, and within weeks Bethe had a rough calculation, and within two years Feynman, Schwinger and Tomonaga had built the renormalised theory that handles it. A part-per-ten-million measurement launched the most successful theory in physics.
The same vacuum fluctuations produce the electron's anomalous magnetic moment from Chapter 7.7, now agreeing with experiment to twelve figures.
Virtual particles, honestly
The vacuum is usually described as full of particle–antiparticle pairs popping in and out of existence. This picture is useful and it is a metaphor for a calculation, and it is worth being precise about which parts are real.
What is real: the vacuum has measurable effects — Casimir, Lamb shift, anomalous moment, vacuum polarisation. Its fields fluctuate. These are experimental facts.
What is a calculational device: the picture of definite particles appearing for a definite time. "Virtual particles" are lines in Feynman diagrams, which are a bookkeeping scheme for terms in a perturbation series. They are not observable, they do not obey E^2 = p^2c^2+m^2c^4, and they cannot be detected.
The energy–time uncertainty argument — that energy \Delta E can be borrowed for time \hbar/\Delta E — is a heuristic. It gives the right scaling and it is not a literal loan.
What definitely happens: the vacuum polarises. Put a charge in it and the surrounding vacuum responds, screening it. So the measured charge of an electron depends on how closely you probe it.
This has been measured. The fine structure constant, \alpha \approx 1/137 at low energy, rises to about 1/128 at the energy of the Z boson, 91 GeV. The electron's charge is genuinely larger when probed harder, because you have penetrated the screening cloud. This is called the running of the coupling, and Chapter 8.3 shows the same thing happens for the strong force, with the opposite sign, producing quark confinement.
Hawking radiation
Chapter 12.3 does this properly. The vacuum connection is worth stating here.
Near a black hole's event horizon, the vacuum fluctuations are affected by the extreme spacetime curvature. What one observer calls the vacuum, another accelerating observer does not — the notion of "no particles" is not observer-independent in curved spacetime.
The result is that a black hole radiates thermally at temperature:
T_H = \frac{\hbar c^3}{8\pi GMk_B}
For a solar-mass hole this is 6\times10^{-8} K — far colder than the cosmic microwave background at 2.7 K, so real black holes currently absorb more than they emit.
The related Unruh effect says an observer accelerating through flat empty space sees a thermal bath at:
T = \frac{\hbar a}{2\pi ck_B}
Even in perfectly empty flat space. At a = g this is 4\times10^{-20} K, unmeasurably small, which is why nobody has seen it. But the principle is the same: what counts as "empty" depends on how you are moving.
Where the energy in an atom comes from
The question in this chapter's title deserves a direct answer, because it is one people ask and rarely get.
An atom is not full of energy in the sense of a wound spring. It is in its ground state, which is the lowest energy available. There is nothing to extract.
The kinetic energy of a ground-state electron is real — 13.6 eV of it in hydrogen — and it is there because of the uncertainty principle (Chapter 7.5). Confining the electron to a region of size a_0 forces a momentum spread, and momentum spread is kinetic energy. This is not extractable. It is exactly balanced by the potential energy well, and the total is -13.6 eV, which is why the atom is bound.
So what is "the energy inside an atom" that people speak of?
Three different things, and they should not be confused.
Chemical energy — a few eV per atom, released by rearranging electrons into more tightly bound configurations. This is burning, and it is what powers a car and a body.
Nuclear binding energy — a few MeV per nucleon, a million times larger, released by rearranging protons and neutrons. This is fission and fusion (Chapter 9.2).
Rest mass energy — mc^2, another factor of a hundred larger. And Chapter 6.4 established where over 98 % of it comes from: the gluon field energy and quark confinement energy inside protons and neutrons.
So the largest store of energy in ordinary matter is field energy, and it is field energy in exactly the sense this chapter has been describing — the energy of quantum fields, in a bound configuration. It is not zero-point energy, and it is not accessible without antimatter.
Can zero-point energy be extracted?
No, and the reason is definitional rather than technological.
Zero-point energy is the ground state. It is the lowest energy state of the system. Extracting energy means moving to a lower state, and there is none.
The Casimir effect does not do it. The plates release energy as they come together, once. To pull them apart you must put the same energy back. It is a conservative force, exactly like gravity — and nobody claims a rock falling proves gravity is an energy source.
A useful analogy: sea level. There is an enormous amount of water in the ocean, and you cannot extract energy from water at sea level, because there is nowhere lower for it to fall.
Claims to the contrary appear regularly and none has ever survived scrutiny. They fail on the first law, the second law, or both. It is worth saying plainly, because "zero-point energy" is a phrase that attracts a great deal of nonsense: the physics in this chapter is real, measured, and gives no energy source whatsoever.
The cosmological constant problem, again
Chapter 6.8 raised it and the vacuum is where it comes from.
Sum the zero-point energies of all field modes up to the Planck scale, and the vacuum energy density is about 10^{96} kg/m³. The measured dark energy density is 6\times10^{-27} kg/m³.
\text{Discrepancy: } 10^{122}
Something cancels the vacuum energy to 122 decimal places and then stops. No one knows what.
Some structure of the problem is worth noting. Supersymmetry, if it exists, would cancel bosonic and fermionic contributions exactly — but only if unbroken, and it is manifestly broken, which leaves a residual still around 10^{60} too large. The anthropic argument says the value must be small or galaxies could not form, and works only if there are many universes with different values (Chapter 12.9). Neither is satisfying, and the problem is arguably the sharpest indication that something fundamental is missing.
Where this shows up in your life
Liquid helium exists because of zero-point motion, and every MRI scanner needs it.
Heavy water behaves differently from ordinary water in ways used in nuclear reactors and in chemical research.
Casimir forces are a design constraint in the micro-electromechanical devices in your phone.
The Lamb shift and the anomalous magnetic moment are how quantum electrodynamics was verified, and QED underpins the design of every laser, semiconductor and atomic clock.
Vacuum polarisation means the constants of nature depend on the energy at which you measure them, which every particle physics calculation must account for.
And the vacuum is currently accelerating the expansion of the universe, which Chapter 12.7 takes up.
What the next chapter fixes
Quantum mechanics as built so far has a fixed number of particles. The Schrödinger equation of Chapter 7.3 describes one electron, or two, and the number never changes. But particles are created and destroyed constantly — an atom emits a photon that did not exist before, a neutron decays into three particles, and colliding two protons at the LHC produces hundreds. And the equation is not relativistic, which Chapter 6.4 says is not optional at high energy.
Fixing both at once forces something new. Chapter 7.10 shows that combining quantum mechanics with relativity requires fields rather than particles, predicts antimatter as an unavoidable consequence before anyone had seen it, and produces the framework that Part 8 uses to describe everything that exists.