The occupation-number language of the previous page is already close to field theory. It labels states by how many particles occupy each one-particle mode. The next step is to stop treating the mode label as an abstract index and package all creation and annihilation operators into position-dependent operators.
For nonrelativistic particles, this step is not forced by Lorentz invariance. It is a gentler fact: ordinary N-body quantum mechanics can be rewritten as an operator theory on Fock space. The field operator ψ(x) annihilates a particle at position x, while ψ†(x) creates one there. Once this is done, the one-body Hamiltonian, the density, and pairwise interactions all become compact operator expressions.
The payoff is large. A many-body Schrödinger equation with a different wavefunction for each particle number becomes one Hilbert-space equation
idtd∣Ψ(t)⟩=H∣Ψ(t)⟩
on Fock space. Later, relativistic QFT will use the same logic, but with a crucial upgrade: particle number need not be conserved, and antiparticles and relativistic causality will become unavoidable.
The word “field” here should not be mystical. It means that we have an annihilation operator for every point in space, just as ap is an annihilation operator for every momentum mode. The field is operator-valued because creating and annihilating particles changes the state in Hilbert space.
Using the inverse transform in H0 gives the coordinate-space Hamiltonian. Since multiplication by p2 in momentum space is the same as acting with −∇2 in position space,
H0=2m1∫d3x∇ψ†(x)⋅∇ψ(x).
Integrating by parts, assuming periodic boundary conditions or fields that vanish sufficiently fast at infinity,
H0=∫d3xψ†(x)(−2m∇2)ψ(x).
This is the first example of the second-quantization rule: a one-particle operator is placed between ψ† and ψ and integrated over space.
This equation looks like the one-particle Schrödinger equation, but its interpretation is different. In ordinary quantum mechanics, a wavefunction is a complex-valued coefficient of a state. Here ψ(x,t) is an operator-valued distribution. Matrix elements of this operator obey Schrödinger-type equations because the operator creates and destroys particles whose single-particle energy is p2/(2m).
The distinction is algebraic, not merely notational: the field has the nontrivial commutator [ψ(x),ψ†(y)]=δ(3)(x−y) and changes particle number. Indeed, because [N,ψ]=−ψ,
eiαNψ(x)e−iαN=e−iαψ(x).
This is the action of the number-symmetry generator on an operator. It should not be confused with the physically irrelevant overall phase of a state vector.
In momentum space,
idtdap(t)=2mp2ap(t),
so
ap(t)=e−ip2t/(2m)ap(0),
and
ψ(x,t)=V1p∑eip⋅x−ip2t/(2m)ap(0).
This is the operator version of the free Schrödinger wave expansion.
The subscript u reminds us that this basis is unnormalized in the same distributional sense that ∣x⟩ is unnormalized. Because the creation fields commute, the ket is symmetric under permutations of x1,…,xN.
Now commute ψ(x) through the product of creation fields:
This is the promised local meaning of ψ(x): it removes a particle at x, and the delta functions search through the unordered list of particle positions.
The annihilation field ψ(x) acts on an N-particle coordinate ket by summing over all possible particles it could remove. Each term carries δ(3)(x−xk) and leaves an (N−1)-particle ket with xk omitted.
A normalized N-boson wavefunction ΨN can be embedded into Fock space by
The inverse map recovers the ordinary wavefunction as a vacuum-to-state matrix element:
ΨN(x1,…,xN)=N!1⟨0∣ψ(xN)⋯ψ(x1)∣ΨN⟩.
The matrix element contains N! equal contractions. Together with the 1/N! in the definition of ∣ΨN⟩, they produce N!ΨN; the inverse prefactor above removes precisely that factor. Thus the operator and wavefunction descriptions contain exactly the same information within a fixed-N sector.
The factor N is the same occupation-number square root encountered earlier. It appears because there are N indistinguishable ways to remove one boson.
So ρ(x) measures a spike of density at each particle position.
The density operator ρ(x)=ψ†(x)ψ(x) acts on a coordinate basis state as ∑kδ(3)(x−xk). In an expectation value, these spikes are smeared by the many-body wavefunction.
The total particle-number operator is
N=∫d3xρ(x)=∫d3xψ†(x)ψ(x)=p∑ap†ap.
It satisfies
[N,ψ†(x)]=ψ†(x),[N,ψ(x)]=−ψ(x).
Thus ψ† raises particle number by one, while ψ lowers it by one.
For a normalized N-particle wavefunction,
⟨ΨN∣ρ(x)∣ΨN⟩=N∫d3x2⋯d3xN∣ΨN(x,x2,…,xN)∣2.
This integrates to N, not to 1:
∫d3x⟨ρ(x)⟩=N.
The density is a particle-counting density. To get the probability density for one randomly selected particle, divide by N.
The colons denote normal ordering: all creation operators are moved to the left of all annihilation operators. This is not just cosmetic. The un-normal-ordered product ρ(x)ρ(y) includes a self-contraction proportional to δ(3)(x−y)ρ(x). Normal ordering removes the spurious interaction of a particle with itself. For a contact interaction this distinction is especially important, because the self-term would be proportional to the ill-defined quantity δ(3)(0).
The interaction Hint annihilates two particles at x and y, weights the pair by V(x−y), and recreates the two particles. The factor 1/2 prevents double-counting the unordered pair (x,y).
Let us verify the action of this term. Acting on an N-particle coordinate ket, the two annihilation operators remove an ordered pair of particles. The result is
This recovers exactly the first-quantized pair potential. The second-quantized formula is not an approximation. It is the same Hamiltonian written in a language that can act on all particle-number sectors at once.
This equivalence is the main lesson of the page. A field theory can be a rewriting of familiar many-body quantum mechanics. Once that rewriting is in place, it becomes natural to consider Hamiltonians that change particle number, and later relativistic fields whose quanta are particles and antiparticles.
The nonrelativistic field operator is the bridge from many-body quantum mechanics to QFT. It is built by Fourier-transforming the creation and annihilation operators of momentum modes. Its equal-time algebra is local:
[ψ(x),ψ†(y)]=δ(3)(x−y).
The operator ψ†(x) creates a particle at x, while ψ(x) removes one. The density
ρ(x)=ψ†(x)ψ(x)
counts particles locally, and N=∫d3xρ(x) counts particles globally.
The notation is heavier at first, but the conceptual payoff is enormous: particles become excitations created and destroyed by fields, and interactions become operator expressions built from local densities. Nothing mystical has happened yet—this is still ordinary many-body quantum mechanics—but the language is now ready for relativistic QFT, where particle number and even the distinction between particles and antiparticles become dynamical.
A field operator is not a many-body wavefunction. The free field equation
i∂tψ=−2m∇2ψ
is an operator equation. Wavefunctions are obtained as coefficients of Fock-space states or as matrix elements.
The density ρ(x) is not normalized to one. In an N-particle state,
∫d3x⟨ρ(x)⟩=N.
For a probability density of one randomly selected particle, divide by N.
The expression 21∫ρVρ must be normal ordered, or otherwise corrected, if it is meant to represent an interaction between distinct particles. Without this care, it includes a self-interaction term.
Continuum coordinate kets are distributions. Delta functions in formulas involving ∣x1,…,xN⟩u are not pathologies; they are the many-particle version of ⟨x∣y⟩=δ(3)(x−y).
Mark Srednicki, Quantum Field Theory, Section 1, for a compact discussion of rewriting fixed-particle-number nonrelativistic quantum mechanics as a field theory.
Sidney Coleman, Lectures of Sidney Coleman on Quantum Field Theory, Chapter 2, for a systematic construction of the many-particle Hilbert space and occupation-number language.
Alexander L. Fetter and John D. Walecka, Quantum Theory of Many-Particle Systems, Chapters 1–2, for a many-body perspective on field operators, density operators, and second quantization.
A. Zee, Quantum Field Theory in a Nutshell, especially the discussion of field theory without relativity, for physical motivation and condensed-matter connections.