← quantum · contents

§ Phase xiv · Fields as Particles · Chapter 03

The leap to QFT

The Schrödinger equation tracks one electron through a fixed cast of characters. Pair production breaks that promise the moment a photon turns into matter. Quantum field theory rewrites the script: the field is the actor, particles are its lines, and the number of speakers can change between one line and the next.

When you learn quantum mechanics for the first time, you learn it the way Erwin Schrödinger wrote it in 1926. There is an electron, somewhere in space, described by a wavefunction ψ(x, t). The wavefunction evolves under a partial differential equation, the Schrödinger equation, that looks like a smoothed version of Newton’s second law. You solve the equation, you square the modulus of the wavefunction, and the result tells you the probability of finding the electron in any small box around any point in space. The bookkeeping is clean. You hand the universe one particle, and the universe hands you back a function on three-dimensional space.

For the hydrogen atom this works flawlessly. For helium, with two electrons, it still works if you are careful with the antisymmetry rule. For a benzene ring, with thirty-six electrons all interacting with each other and with six nuclei, the calculation is hard but the framework is unchanged: you write a wavefunction ψ(x₁, x₂, …, x₃₆, t) on a space of one hundred and eight dimensions, you apply the Schrödinger equation, and you read off probabilities. The cast of characters is fixed. You decided how many electrons were in the problem at the start, and you will have the same number at the end.

This is the architecture of single-particle quantum mechanics. It is the architecture every undergraduate textbook teaches. And in the autumn of 1928, two years after Schrödinger published his equation, Paul Dirac sat down to write a relativistic version of it. He wanted an equation that respected Einstein’s E² = (pc)² + (mc²)², which Schrödinger’s equation does not. He found one. It described the electron beautifully, predicted the existence of spin without putting it in by hand, and matched the fine structure of hydrogen to four decimal places. It also predicted something Dirac could not, at first, get rid of: a set of solutions with negative energy, infinitely many of them, lying like a sealed basement under the floor of every electron’s allowed states.

Dirac wrestled with the negative-energy states for three years. In 1931 he announced what he had decided they were: not a bug but a feature. They were, he said, occupied states, an infinite sea of negative-energy electrons filling the vacuum. If you put enough energy into one of them, you could lift it out, leaving a hole. The hole would behave like a particle with the opposite charge of the electron. He had, he wrote, “introduced a new particle, hitherto unobserved.” A year later, in 1932, Carl Anderson photographed exactly such a particle in a cloud chamber at Caltech, called it the positron, and won the Nobel Prize. Dirac’s equation had predicted antimatter from pure mathematics.

Look at what the prediction of the positron really demanded. A photon of sufficient energy, passing near a nucleus, could split into an electron and a positron: light turning into matter, two particles where there had been one. A few microseconds later the positron might find another electron, annihilate it, and release two photons: two particles vanishing, two new ones appearing. Particle number was no longer conserved. And the moment you accept that, the entire framework of single-particle quantum mechanics is in trouble, because the framework assumes that you decide, once and for all, how many particles you have, and writes a wavefunction on a space of that fixed size. You cannot describe pair production with a function ψ(x, t). The function does not know how to grow a second argument while you watch.

You can see the same crack from a different angle. Quantum mechanics says energies are quantised but particle counts are not. Relativity says energy and mass are interchangeable through E = mc². Put those two statements together and you find that at high enough energies you can always afford to make new particles, and you cannot forbid the universe from doing so. The fixed-cast bookkeeping of Schrödinger has to give way to something that lets the cast change while the show goes on. The replacement, the framework that Dirac and his successors spent twenty years building, is quantum field theory: QFT.

To explain the leap, it helps to start with a system you already know: a guitar string, or a violin string, stretched between two fixed ends. The string vibrates in a discrete set of normal modes, the fundamental and its harmonics. Each mode is independent of the others. Each mode is, mathematically, a simple harmonic oscillator with its own frequency. When you quantise the string (treat each mode as a quantum harmonic oscillator), each mode acquires an integer-labelled energy ladder: 0, 1, 2, 3, … quanta. The quanta of vibration in a crystal lattice are called phonons, and they behave exactly like particles. They have momentum, they scatter off impurities, they can be counted. They are not “really” particles in the sense of being little balls; they are excitations of a field (the lattice displacement field) that we choose to call particles because they obey particle bookkeeping.

That is the entire idea of QFT in one sentence. Replace the guitar string with a field defined at every point in space, replace its discrete modes with a continuum of modes, and call the quanta of those modes “particles.” A photon is a quantum of the electromagnetic field. An electron is a quantum of the electron field. A Higgs boson is a quantum of the Higgs field. The field is the primary object. Particles are the field’s countable excitations. You can have one, two, or seventeen of them, and you can convert between counts, because adding or removing an excitation is just acting on the field with the right operator, and the field is already defined whether or not any excitations are present.

In theoretical physics, quantum field theory (QFT) is a theoretical framework that combines field theory, special relativity and quantum mechanics. QFT is used in particle physics to construct physical models of subatomic particles and in condensed matter physics to construct models of quasiparticles. The current Standard Model of particle physics is based on QFT.

This single shift, from “particle with a wavefunction” to “field with excitations,” dissolves the puzzle of pair production. When a photon turns into an electron and a positron, what has happened is that an excitation of the electromagnetic field has been destroyed and two excitations of the electron field (one of each charge) have been created, with energy and momentum conserved across the swap. No wavefunction had to grow a new argument. The fields were already there. The interaction term in the Lagrangian (the rule that couples the electromagnetic field to the electron field) tells you the rate at which such swaps happen. You compute the probability of the process, you measure it in the lab, you compare. As of 2026, the agreement is to about ten decimal places for the electron’s anomalous magnetic moment, the most precise prediction in the history of physics.

Once you accept the field as fundamental, a tidy catalogue falls out. Every known elementary particle is a quantum of one field. The electron field has electrons (and positrons) as its quanta. The photon field is what we used to call the electromagnetic field, and its quanta are photons. There are six quark fields (up, down, charm, strange, top, bottom), each with its own antiparticle quanta. There are three charged-lepton fields (electron, muon, tau) and three neutrino fields. There are gauge fields for the strong force (eight gluons) and the weak force (three: W⁺, W⁻, Z⁰). And there is the Higgs field, a single scalar field whose quanta were finally detected at CERN in 2012. That is the full set, as of 2026. Every particle on every Feynman diagram you have ever seen is a wiggle in one of these fields.

The fields do not just sit there in isolation. They interact, and the interactions are what make physics interesting. The way interactions enter a QFT is through extra terms in something called the Lagrangian density, the same Lagrangian that classical mechanics students learn to write for a pendulum or a planet. A free field has a Lagrangian with two parts: a kinetic term (energy of change) and a mass term (the cost of existing). When you couple two fields together, you write down a small extra term, the interaction term, that depends on both of them at once. The most famous example is the interaction term of QED, quantum electrodynamics, the theory of electrons and photons. It is a single product:

L_int = − e · ψ̄ γ^μ ψ · A_μ

In English: at every point in spacetime, multiply the electron field (twice, once with a bar and once without) by a Dirac gamma matrix, multiply by the photon field A, and stick a coupling constant e (the electron charge) out front. That tiny piece of algebra is the entire content of “electrons feel photons and photons feel electrons.” Every quantum-electrodynamic process you have ever heard of (Compton scattering, the Lamb shift, the anomalous magnetic moment of the electron, the bending of starlight near the sun reinterpreted as photon scattering) is computed by expanding the consequences of that one term to higher and higher order in the small parameter e. The answers come out absurdly accurate.

The QED Lagrangian was written down by Julian Schwinger, Sin-Itiro Tomonaga, and Richard Feynman in the late 1940s, in three different but equivalent formulations whose equivalence was proved by Freeman Dyson in 1949. The same shape (kinetic + mass + interaction) extends across the Standard Model. The strong force is described by quantum chromodynamics (QCD), whose interaction term couples quark fields to eight gluon fields and was modelled on QED’s structure. The weak force was unified with electromagnetism in the electroweak theory of Glashow, Weinberg, and Salam (Nobel 1979), using a generalisation of QED’s gauge symmetry that Yang and Mills had proposed in 1954. The Higgs mechanism stitches mass onto particles that the underlying symmetry would otherwise force to be massless. Every interaction in the Standard Model is a gauge interaction, meaning it follows the same template, with different symmetry groups (U(1), SU(2), SU(3)) controlling which fields couple to which.

↘ Derive: where the QED interaction comes from

Start with the free Dirac Lagrangian for an electron field ψ of mass m:

L_free = ψ̄ (i γ^μ ∂_μ − m) ψ

Notice that the equation of motion is unchanged if you multiply ψ by a constant phase: ψ → e^{i α} ψ. This is a global U(1) symmetry, the same one that says only relative phases are observable in quantum mechanics. By Noether’s theorem, this symmetry implies a conserved current, the electron number current J^μ = ψ̄ γ^μ ψ.

Now demand more. Demand that the phase α be allowed to depend on spacetime: α = α(x). The Lagrangian is no longer invariant; the derivative ∂_μ produces an extra term proportional to ∂_μ α that does not cancel. To restore invariance, introduce a new vector field A_μ that transforms as A_μ → A_μ − (1/e) ∂_μ α, and replace the ordinary derivative ∂_μ with the gauge-covariant derivative:

D_μ = ∂_μ + i e A_μ

Substituting D_μ for ∂_μ in the Dirac Lagrangian:

L = ψ̄ (i γ^μ D_μ − m) ψ = ψ̄ (i γ^μ ∂_μ − m) ψ − e ψ̄ γ^μ A_μ ψ

The first piece is the original free Dirac Lagrangian. The second piece is the QED interaction. The field A_μ is the photon. Demanding local phase invariance has forced the existence of the photon and dictated exactly how it couples to electrons. Adding a kinetic term for A_μ (the Maxwell term −(1/4) F_μν F^μν) completes the QED Lagrangian.

This is the deepest lesson of QFT. Interactions are not added by hand; they are demanded by symmetry. Yang and Mills generalised the same argument to non-Abelian symmetry groups in 1954, and out fell the structure of the strong and weak forces. Every gauge field in the Standard Model is a Yang-Mills field with a different symmetry group.

The interaction terms make QFT work. They also nearly broke it. When you actually try to compute the consequences of an interaction beyond the leading order, you have to sum over all the ways the fields can fluctuate in the intermediate steps: virtual photons, virtual electron-positron pairs, virtual everything. Each of those intermediate fluctuations gets integrated over its momentum, and the momentum integral runs from zero to infinity. At high momentum, the integrals diverge. The first-order correction to the electron’s self-energy is infinite. The first-order correction to the photon’s mass is infinite. By 1947 it looked, briefly, as though QED was internally inconsistent. The bare parameters in the Lagrangian (the mass m, the charge e) seemed to be infinite. Calculations gave infinite answers. The theory predicted, formally, that the electron has infinite mass and infinite charge.

The rescue, worked out between 1947 and 1949, is called renormalization, and it is one of the strangest stories in twentieth-century physics. The argument goes like this. The bare parameters in the Lagrangian (the m and e written down at the start) are not what you measure. What you measure is the dressed electron, an electron surrounded by a halo of its own virtual photons and virtual pairs. The halo contributes infinities of its own. But the bare parameters can be chosen to be infinite in just the right way that the infinities cancel between the bare parameters and the halo, leaving a finite, measured mass and a finite, measured charge. In short: you write the Lagrangian with infinite parameters, you compute observables, and the infinities cancel in pairs to leave finite predictions. You never observe the bare parameters. You only observe the dressed ones, and the dressed ones come out right.

Most physicists, when they first encounter renormalization, do not like it. Dirac himself never approved of it. He called it “a process that takes one infinity from another and gets a finite answer, and that is not mathematics.” Feynman, who had helped invent the technique, agreed that it was “a dippy process,” but pointed out that it worked, that it had matched the Lamb shift to four decimal places and the anomalous moment of the electron to seven, and that whatever its philosophical status it was at least an extraordinarily successful piece of bookkeeping. By the 1970s, the work of Kenneth Wilson on the renormalization group recast the whole business in a deeper light: the bare parameters are functions of the energy scale at which you choose to define them, and the physical observables are scale-invariant quantities that do not care which scale you picked. The “infinities” were artefacts of trying to define a theory all the way down to arbitrarily small distances. Cut the theory off at any finite distance and it is perfectly finite.

The reason for going through all this in a chapter that promised to be intuitive is that QFT is what physics actually uses now. The single-particle Schrödinger equation is the right tool for chemistry, where electron numbers do not change, but the moment you walk into a particle collider, a stellar interior, the cosmic microwave background, or a superconductor near its critical point, you need fields. The electron field, the photon field, the quark fields, the Higgs field: each one is a thing in its own right, defined at every point in spacetime, with its own equation of motion and its own ladder of excitations. Particles are how the field appears when you ask it to commit to a count. The fact that a vacuum is not nothing, that empty space is a roiling sea of zero-point fluctuations of every field at once, is a direct consequence of treating the field as primary. It is also why the Casimir force pushes two plates together, why the vacuum has an energy density that nobody can compute correctly, and why every textbook from 1960 forward starts not with ψ(x, t) but with the field operator ψ̂(x).

The bridge between this chapter and the next, “Stellar fusion,” is one of the cleanest in the book. Quantum field theory tells us that the vacuum is not empty and that particles are excitations of fields that live everywhere at once. The very next place to test that picture is not in a collider but in a star. The sun shines because protons tunnel through the Coulomb barrier that classical mechanics would have insisted is impassable, and the rate at which they tunnel is set by the same wave-mechanical amplitudes we have been computing all along. Before the book closes on the open questions of dark matter and quantum gravity, it takes a tour through the most extreme quantum experiments nature runs for us, in the cores of stars and the final moments before a black hole forms.

§

Quantum field theory built the Standard Model and the Standard Model built our picture of the vacuum. The next chapter steps outside the laboratory and into the core of the sun, where tunneling is not a curiosity but a power source, and where the same equations we have been writing for two-electron atoms run a star.

next chapter → Stellar fusion