12.05.04 · quantum / relativistic-fields

Free Klein-Gordon scalar quantum field

shipped3 tiersLean: none

Anchor (Master): Weinberg, *The Quantum Theory of Fields, Vol. 1* (Cambridge 1995), §5.2 (the scalar-field case of the Poincaré-irrep construction); Streater & Wightman, *PCT, Spin and Statistics, and All That* (Benjamin 1964; Princeton Landmarks 2000), Ch. 3 (Wightman-axiom realization); Reed & Simon, *Methods of Modern Mathematical Physics, Vol. II*, §X.7 (Fock-space functor and second quantization); Folland, *Quantum Field Theory: A Tourist Guide for Mathematicians* (AMS 2008), Ch. 5

Intuition Beginner

A quantum field is a quantum harmonic oscillator at every point of space, with all those oscillators wired together by a relativistic wave equation. The free Klein-Gordon scalar field is the simplest example: one real oscillator per spatial point, coupled by the wave equation that says a disturbance travels with a speed bounded by the speed of light and an inertia set by a mass parameter $m$ . The excitation count of each oscillator gives the count of particles of that momentum, and the rule that adding excitations adds energy gives the rule that adding particles adds energy. The field is a relativistic upgrade of the familiar single oscillator.

Why bother? Ordinary quantum mechanics tracks a fixed roster of particles, but high-energy experiments make and destroy particles routinely. A photon scatters off an electron and a second photon is born; an unstable particle decays into two lighter ones. A theory pinned to a fixed particle count cannot represent these events. Promoting the wave function to a field operator, with creation and annihilation operators built in, gives a theory in which particle number is a measured quantity rather than an input parameter.

The free Klein-Gordon field is the cleanest place to learn this idea. Interactions are switched off, so the calculation reduces to a sum over independent oscillators, one per momentum value, and every quantity of interest is computable. Every more elaborate field theory (electrons, photons, the Higgs boson) is the free Klein-Gordon construction generalised to spin, charge, or self-interaction. Learn this case carefully and the rest is a sequence of upgrades.

Visual Beginner

A schematic showing the positive mass shell in four-dimensional momentum space (an upward-opening hyperboloid) with dots marking the momenta of single-particle excitations. Each dot corresponds to one quantum of the field with three-momentum $p$ and energy $E = (p^{2} + m^{2})^{1/2}$ . The hyperboloid is the on-shell condition; the dots are how the Fock space gets populated by creation operators.

The picture captures the operational content of the field. The mass shell is the set of allowed particle momenta, and a state of the field is a list of how many particles sit at each momentum. The mode expansion of the field operator is the bookkeeping that turns this list into an operator-valued function on spacetime.

Worked example Beginner

Compute the ground-state energy and the one-particle energies of the free Klein-Gordon field on a box of side $L$ with periodic boundary conditions, and compare with the single quantum harmonic oscillator.

Step 1. Fix $m = 1$ and $L = 2 π$ in natural units. Periodic boundary conditions force the allowed momenta to be $p_{n} = n$ for $n$ an integer vector with three components. The energy of a single particle with momentum $p_{n}$ is $E_{n} = (p_{n}^{2} + 1)^{1/2}$ , where $p_{n}^{2} = n_{1}^{2} + n_{2}^{2} + n_{3}^{2}$ .

Step 2. Each momentum mode is a quantum harmonic oscillator with angular frequency $ω_{n} = E_{n}$ . The energy levels of the $n$ -th oscillator are $E_{n} (k + 1/2)$ for $k = 0, 1, 2, \dots$ .

Step 3. The state with $k = 0$ in every mode is the vacuum; its energy is the sum of zero-point energies $E_{vac} = (1/2) E_{0} + (1/2) E_{1} + \dots$ , where the $E_{0}$ contribution comes from $n = (0, 0, 0)$ and gives $E_{0} = 1$ .

Step 4. A one-particle state with momentum $p_{1} = (1, 0, 0)$ is the state with $k = 1$ in that one mode and $k = 0$ in every other mode. Its excitation energy above the vacuum is $ω_{1} = (1 + 1)^{1/2} = 2^{1/2}$ .

Step 5. A two-particle state with both particles at $p = (1, 0, 0)$ has $k = 2$ in that mode. Its excitation energy is $2 \cdot 2^{1/2}$ , the sum of the two individual energies. Particles do not feel each other because the field is free.

What this tells us: the free Klein-Gordon field is a parallel array of harmonic oscillators, one per allowed momentum. The vacuum is the joint ground state, single-particle energies are the angular frequencies of individual modes, and multi-particle energies add. The relativistic content is the formula $ω_{n} = (p_{n}^{2} + m^{2})^{1/2}$ for the angular frequencies, which forces every excitation to carry both momentum and energy on the mass shell.

Check your understanding Beginner

Formal definition Intermediate+

Fix four-dimensional Minkowski space $M = R^{1, 3}$ with the mostly-plus metric $η = diag (- 1, + 1, + 1, + 1)$ in Weinberg conventions, so that $p \cdot q = - p^{0} q^{0} + p \cdot q$ and $p^{2} = - m^{2}$ on the mass shell. The positive mass shell of mass $m > 0$ is $$ X_m = {p \in M^* : p^2 = -m^2, ; p^0 > 0}, $$ the upper sheet of a two-sheeted hyperboloid in momentum space. The unique (up to scalar) $S O^{+} (1, 3)$ -invariant Borel measure on $X_{m}$ , in the coordinates $p \in R^{3}$ with $p^{0} = ω_{p} = (p^{2} + m^{2})^{1/2}$ , is $$ d\mu_m(\mathbf{p}) = \frac{d^3\mathbf{p}}{(2\pi)^3 , 2\omega_{\mathbf{p}}}. $$ The one-particle Hilbert space of mass $m$ and spin $0$ is $H_{1} = L^{2} (X_{m}, d μ_{m}; C)$ , on which the Poincaré group $P_{+}^{+} = R^{1, 3} ⋊ S O^{+} (1, 3)$ acts by the unitary irreducible representation $(U_{1} (Λ, a) ψ) (p) = e^{ia \cdot p} ψ (Λ^{- 1} p)$ on the scalar orbit.

The bosonic Fock space built on $H_{1}$ is the orthogonal Hilbert sum $$ \mathcal{F}_s(\mathcal{H}1) = \bigoplus{n \geq 0} S^n \mathcal{H}_1, $$ with $S^{n} H_{1}$ the symmetric $n$ -fold tensor product and $S^{0} H_{1} = C \cdot Ω$ the one-dimensional vacuum subspace. The annihilation and creation operators on $F_{s} (H_{1})$ , smeared against $f \in H_{1}$ , are densely defined unbounded operators satisfying the canonical commutation relations $$ [a(f), a(g)] = 0, \qquad [a^\dagger(f), a^\dagger(g)] = 0, \qquad [a(f), a^\dagger(g)] = \langle f, g \rangle_{\mathcal{H}1}\cdot I. $$ In the improper $δ$ -function normalization customary in physics, $a (p)$ and $a^{†} (p)$ satisfy $[a(\mathbf{p}), a^\dagger(\mathbf{q})] = (2\pi)^3 , 2\omega{\mathbf{p}} , \delta^3(\mathbf{p} - \mathbf{q})$ on the dense finite-particle subspace.

The free Klein-Gordon scalar quantum field is the operator-valued tempered distribution $ϕ : S (M; R) \to L (F_{s} (H_{1}))$ defined by smearing the mode expansion $$ \phi(x) = \int_{X_m} d\mu_m(\mathbf{p}) , \bigl[ a(\mathbf{p}) , e^{i p \cdot x} + a^\dagger(\mathbf{p}) , e^{-i p \cdot x} \bigr] $$ against a Schwartz test function $f$ , giving $ϕ (f) = \int d^{4} x f (x) ϕ (x)$ . The integral is well-defined as a sesquilinear form on the dense subspace of finite-particle vectors with Schwartz one-particle wave functions, and extends to a closable operator-valued tempered distribution.

The canonical momentum density is the time-derivative $π (x) = \partial_{t} ϕ (x)$ , again a tempered distribution. On the smeared field operators, the equal-time canonical commutation relations read $$ [\phi(t, \mathbf{x}), \pi(t, \mathbf{y})] = i , \delta^3(\mathbf{x} - \mathbf{y}) \cdot I, \qquad [\phi, \phi] = [\pi, \pi] = 0, $$ in the conventional $ℏ = 1$ units. The Feynman propagator is the Lorentz-invariant Green's function $$ \Delta_F(x - y) = \langle \Omega | T\phi(x) \phi(y) | \Omega \rangle = \int \frac{d^4 p}{(2\pi)^4} , \frac{i , e^{i p \cdot (x - y)}}{p^2 + m^2 - i \varepsilon}, $$ where $T$ denotes time-ordered product and the $i ε$ prescription selects the Feynman contour.

Counterexamples to common slips

The mode expansion of $ϕ (x)$ is not an operator at a sharp spacetime point but an operator-valued distribution; only the smeared operator $ϕ (f)$ for $f \in S (M)$ is densely defined on $F_{s} (H_{1})$ . Statements like " $ϕ (x) ∣Ω ⟩$ is a one-particle state" are correct only when $ϕ (x)$ is read as a distribution paired against a test function.
The Klein-Gordon negative-energy difficulty is resolved at second quantization, not at the wave equation. The first-quantized Klein-Gordon equation has negative-energy solutions and a current density that fails to be positive; the second-quantized field reinterprets the negative-frequency modes as creation operators for antiparticles, and for the real scalar the antiparticle is identified with the particle.
The Feynman $i ε$ prescription is not a regularization in the sense of cutoff theory; it is a contour choice in the $p^{0}$ integral that selects the time-ordered Green's function out of the four-parameter family of Green's functions of $□ + m^{2}$ . Retarded, advanced, Wightman, and Feynman propagators are four distinct contour selections of the same complex pole integrand.
Microcausality is the commutator vanishing for spacelike separations, not the field operators themselves vanishing. The field at spacelike-separated points is generically nonzero; the commutator is zero, expressing the impossibility of causal influence across spacelike intervals.

Key theorem with proof Intermediate+

Theorem (microcausality of the free Klein-Gordon field; Streater-Wightman Ch. 3). Let $ϕ$ be the free Klein-Gordon scalar quantum field of mass $m > 0$ on Minkowski space, smeared as an operator-valued tempered distribution on $S (R^{1, 3})$ with values in unbounded densely defined operators on the bosonic Fock space $F_{s} (L^{2} (X_{m}, d μ_{m}))$ . If $f, g \in S (R^{1, 3})$ are real-valued Schwartz functions whose supports are spacelike separated — that is, every pair $(x, y) \in supp (f) \times supp (g)$ satisfies $(x - y)^{2} = - (x^{0} - y^{0})^{2} + ∣ x - y ∣^{2} > 0$ — then the commutator vanishes: $$ [\phi(f), \phi(g)] = 0. $$

Proof. The argument has three steps. First, reduce the commutator to a $c$ -number distribution via the canonical commutation relations. Second, identify that $c$ -number distribution as a Lorentz-invariant tempered distribution. Third, show that this Lorentz-invariant distribution is supported on the closed light cone ${x : x^{2} \leq 0}$ .

Step 1: $c$ -number reduction. Substitute the mode expansion of $ϕ (x)$ and compute. With the abbreviation $\overset{a}{^}^{\pm} (p) = e^{\pm i p \cdot x}$ for the plane-wave factors evaluated at the relevant spacetime point and $d μ_{m} = d^{3} p / [(2 π)^{3} 2 ω_{p}]$ , the commutator of unsmeared field operators is $$ [\phi(x), \phi(y)] = \int d\mu_m d\mu_{m'} , \bigl[ [a(\mathbf{p}), a^\dagger(\mathbf{p}')] , \hat a^+(\mathbf{p}, x) \hat a^-(\mathbf{p}', y) + [a^\dagger(\mathbf{p}), a(\mathbf{p}')] , \hat a^-(\mathbf{p}, x) \hat a^+(\mathbf{p}', y) \bigr]. $$ The other two commutators in the expansion vanish identically. Using the canonical relation $[a (p), a^{†} (q)] = (2 π)^{3} 2 ω_{p} δ^{3} (p - q)$ and collapsing the $δ$ -functions, one obtains $$ [\phi(x), \phi(y)] = i \Delta(x - y) \cdot I, \qquad i \Delta(z) = \int \frac{d^3 \mathbf{p}}{(2\pi)^3 , 2\omega_{\mathbf{p}}} \bigl[ e^{i p \cdot z} - e^{-i p \cdot z} \bigr]{p^0 = \omega{\mathbf{p}}}. $$ The commutator is the $c$ -number distribution $i Δ$ (the Pauli-Jordan commutator function) times the identity operator on Fock space. Smearing against $f$ and $g$ gives $[ϕ (f), ϕ (g)] = i ⟨ Δ * (\tilde{f} \otimes g)⟩ \cdot I$ in tempered-distribution language; vanishing of the $c$ -number on the spacelike support is what remains to be checked.

Step 2: Lorentz invariance of $Δ$ . Both summands in the integral are Lorentz-invariant tempered distributions on $R^{1, 3}$ , since the measure $d μ_{m}$ and the inner product $p \cdot z$ are both invariant under $S O^{+} (1, 3)$ . Hence $Δ (z)$ is itself a Lorentz-invariant tempered distribution. The full Poincaré group also includes translations, but $Δ$ depends only on the difference $z = x - y$ , so translation invariance is automatic.

Step 3: spacelike vanishing. A Lorentz-invariant tempered distribution on $R^{1, 3}$ takes the same value on every $z$ with a fixed value of $z^{2} = - (z^{0})^{2} + ∣ z ∣^{2}$ and a fixed sign of $z^{0}$ . Pick a spacelike $z = (0, z) \in R^{1, 3}$ with $∣ z ∣ > 0$ . There is a continuous $S O^{+} (1, 3)$ rotation that sends $(0, z)$ to $(0, - z)$ (rotate by $π$ about an axis perpendicular to $z$ ), and this rotation is connected to the identity. By Lorentz invariance, $Δ (0, z) = Δ (0, - z)$ . But the integrand of $Δ$ at $z = (0, z)$ is $$ \Delta(0, \mathbf{z}) = -i \int \frac{d^3 \mathbf{p}}{(2\pi)^3 , 2\omega_{\mathbf{p}}} , 2i \sin(\mathbf{p} \cdot \mathbf{z}) = \int \frac{d^3 \mathbf{p}}{(2\pi)^3 , \omega_{\mathbf{p}}} , \sin(\mathbf{p} \cdot \mathbf{z}), $$ which is odd under $z \mapsto - z$ . The two conditions — odd under $z \mapsto - z$ by direct calculation, equal under $z \mapsto - z$ by Lorentz invariance — force $Δ (0, z) = 0$ for every spacelike $z$ .

By Lorentz invariance, $Δ$ vanishes on every spacelike $z$ , not only the equal-time slice. Hence $[ϕ (x), ϕ (y)] = 0$ for spacelike-separated $(x, y)$ , and smearing against Schwartz functions of spacelike-separated support gives $[ϕ (f), ϕ (g)] = 0$ . $□$

Bridge. Microcausality builds toward every causality-based result in relativistic quantum field theory: the Reeh-Schlieder theorem on vacuum cyclicity, the cluster-decomposition derivation of the connected $S$ -matrix in 12.13.01 (bosonic Fock space and second quantisation), and the spin-statistics theorem of Lüders-Pauli-Zumino. The foundational reason microcausality holds is exactly the Lorentz invariance of the Pauli-Jordan distribution $i Δ$ : spacelike-separated arguments lie on a single Lorentz orbit on which the distribution must be odd by direct calculation and even by invariance, so it vanishes. This is exactly the same odd-even-conflict mechanism that appears again in 12.13.02 (fermionic Fock space and Pauli exclusion), where the analogous odd-even conflict forces the field anticommutator to vanish at spacelike separation only when half-integer spin is paired with anticommutators rather than commutators — the spin-statistics theorem. The central insight is that the canonical commutation relations on Fock space, combined with the Poincaré-invariant mode expansion over the mass shell, identifies the field commutator with a Lorentz-invariant $c$ -number distribution, and Lorentz invariance plus the odd-under-spatial-inversion structure forces vanishing on spacelike orbits. The bridge is the recognition that microcausality on a free field generalises to the locality axiom of the Wightman framework, and putting these together one obtains the operator-algebra formulation of Haag-Kastler in which local observable algebras at spacelike-separated regions commute. Microcausality is dual to Lorentz covariance in the sense that the two together cut out the Wightman free field uniquely among scalar quantum fields with a given two-point function.

Exercises Intermediate+

Exercise 3 (medium, symbolic).

Compute the Wightman two-point function $W_{2} (x, y) = ⟨ Ω∣ ϕ (x) ϕ (y) ∣Ω ⟩$ of the free Klein-Gordon field in terms of the Lorentz-invariant measure on the mass shell.

Hint

Substitute the mode expansion of $ϕ$ and use $a (p) ∣Ω ⟩ = 0$ and $⟨ Ω∣ a^{†} (p) = 0$ to keep only the term with annihilation on the right and creation on the left.

Answer

Using the mode expansion and $a (p) ∣Ω ⟩ = 0$ , only the $a (p) a^{†} (q)$ ordering contributes. By the canonical commutator $[a (p), a^{†} (q)] = (2 π)^{3} 2 ω_{p} δ^{3} (p - q)$ , $$ W_2(x, y) = \int \frac{d^3 \mathbf{p}}{(2\pi)^3 , 2\omega_{\mathbf{p}}} , e^{i p \cdot (x - y)} \big|{p^0 = \omega{\mathbf{p}}} = \Delta_+(x - y). $$ This is the positive-frequency Wightman function, a Lorentz-invariant tempered distribution depending only on $z = x - y$ . Rubric: full credit for the Fourier-integral form plus the identification as $Δ_{+}$ .

Exercise 4 (medium, short-answer).

Show that the Feynman propagator $Δ_{F} (x - y) = ⟨ Ω∣ T ϕ (x) ϕ (y) ∣Ω ⟩$ is a Green's function of the Klein-Gordon operator $□ + m^{2}$ , where $□ = η^{μν} \partial_{μ} \partial_{ν}$ .

Hint

Differentiate the momentum-space representation $Δ_{F} (z) = \int d^{4} p / (2 π)^{4} \cdot i / (p^{2} + m^{2} - i ε) \cdot e^{i p \cdot z}$ under the integral and use the Fourier representation of the four-dimensional delta function.

Answer

Applying $□_{z} + m^{2} = η^{μν} \partial_{z^{μ}} \partial_{z^{ν}} + m^{2}$ to the momentum-space integrand brings down $- p^{2} + m^{2}$ . With the mostly-plus signature $η = diag (- 1, + 1, + 1, + 1)$ , $p^{2} = η^{μν} p_{μ} p_{ν}$ , and the on-shell relation reads $p^{2} = - m^{2}$ , so $- p^{2} + m^{2} = - (p^{2} - m^{2})$ . Picking the sign convention consistent with $(□ + m^{2})$ acting through partial derivatives of $e^{i p \cdot z}$ , the operator brings down $- (p^{2} + m^{2})$ , which cancels the denominator of the integrand and leaves $$ (\Box + m^2) \Delta_F(z) = \int \frac{d^4 p}{(2\pi)^4} , (-i) , e^{i p \cdot z} = -i \cdot \delta^4(z). $$ Up to the conventional sign, $Δ_{F}$ is the Feynman Green's function of $□ + m^{2}$ . The $i ε$ prescription selects the contour that produces the time-ordered analytic structure: positive-frequency modes propagate forward in time, negative-frequency modes propagate backward. Rubric: full credit for the cancellation argument plus identification of the Green's-function equation.

Exercise 5 (medium, short-answer).

The Hamiltonian density of the free Klein-Gordon field is $H = (1/2) π^{2} + (1/2) (\nabla ϕ)^{2} + (1/2) m^{2} ϕ^{2}$ . Show that the normal-ordered Hamiltonian on Fock space reads $H = \int d μ_{m} (p) ω_{p} a^{†} (p) a (p)$ .

Hint

Substitute the mode expansion of $ϕ$ and $π = \partial_{t} ϕ$ into the Hamiltonian, integrate over spatial coordinates to collapse the plane-wave products to $δ$ -functions, and subtract the zero-point energy by normal ordering.

Answer

Substituting the mode expansion of $ϕ$ and $π$ into $H = \int d^{3} x H$ produces a sum of four integrals labelled by the four sign combinations of the two mode operators. The spatial integral $\int d^{3} x e^{i (p \pm q) \cdot x}$ gives $(2 π)^{3} δ^{3} (p \pm q)$ , collapsing each integral to a single momentum integral. The cross terms with $a (p) a (- p)$ and $a^{†} (p) a^{†} (- p)$ cancel between $π^{2}$ and $(\nabla ϕ)^{2} + m^{2} ϕ^{2}$ because of the on-shell relation $ω_{p}^{2} = p^{2} + m^{2}$ . The remaining terms produce $H = \int d μ_{m} (p) ω_{p} [a^{†} (p) a (p) + (1/2) [a (p), a^{†} (p)]]$ . Normal ordering drops the $(1/2)$ -commutator zero-point energy and leaves the stated expression. Rubric: full credit for the cancellation argument plus the normal-ordering step.

Exercise 7 (hard, short-answer).

Prove that the four standard Green's functions of the Klein-Gordon operator $□ + m^{2}$ — retarded $Δ_{R}$ , advanced $Δ_{A}$ , Feynman $Δ_{F}$ , and Dyson $Δ_{D}$ — correspond to four distinct closed contours in the complex $p^{0}$ plane that all pick up the same poles $p^{0} = \pm ω_{p}$ but with different residue combinations.

Hint

In the $d^{4} p$ integral representation, isolate the $p^{0}$ integration. The integrand has poles at $p^{0} = \pm ω_{p}$ on the real axis. Each Green's function is obtained by displacing one or both poles by $\pm i ε$ in different combinations, equivalently by choosing one of four contours.

Answer

The $p^{0}$ integrand of the momentum-space representation of any Green's function of $□ + m^{2}$ has poles at $p^{0} = \pm (p^{2} + m^{2})^{1/2} = \pm ω_{p}$ . Four prescriptions: $Δ_{R}$ : both poles pushed to $p^{0} \to p^{0} + i ε$ , contour closed above for $z^{0} < 0$ (zero) and below for $z^{0} > 0$ (both poles contribute). Retarded support $z^{0} > 0$ . $Δ_{A}$ : both poles pushed to $p^{0} \to p^{0} - i ε$ , supports $z^{0} < 0$ by the analogous contour argument. $Δ_{F}$ (Feynman): positive pole pushed down ( $+ ω \to + ω - i ε$ ), negative pole pushed up ( $- ω \to - ω + i ε$ ); positive-frequency modes propagate forward in time, negative-frequency modes backward. Time-ordered. $Δ_{D}$ (Dyson): the opposite prescription to Feynman; anti-time-ordered. The four Green's functions satisfy $Δ_{F} - Δ_{D} = i (Δ_{+} - Δ_{-}) = i Δ$ (the Pauli-Jordan commutator function), $Δ_{R} - Δ_{A} = i Δ$ , and $Δ_{F} = Δ_{R} + i Δ_{-} = Δ_{A} + i Δ_{+}$ . Rubric: full credit for identifying the four contour prescriptions plus at least one of the linear relations.

Exercise 8 (hard, short-answer).

Show that the Lorentz transformation property $U (Λ, 0) ϕ (x) U (Λ, 0)^{- 1} = ϕ (Λ x)$ of the free Klein-Gordon field follows from the Poincaré-covariant transformation $U (Λ, 0) a (p) U (Λ, 0)^{- 1} = a (Λ^{- 1} p)$ of the annihilation operators, plus the Lorentz invariance of the mass-shell measure $d μ_{m}$ .

Hint

Apply $U (Λ, 0)$ to the mode expansion, transform each $a$ and $a^{†}$ via the given covariance, and change variables in the integration over the mass shell using the Lorentz invariance of $d μ_{m}$ .

Answer

Acting with the Poincaré representation on the mode expansion, and using $U (Λ, 0) a (p) U (Λ, 0)^{- 1} = a (Λ^{- 1} p)$ with the analogous relation for $a^{†}$ , gives $$ U(\Lambda, 0) \phi(x) U(\Lambda, 0)^{-1} = \int d\mu_m(\mathbf{p}) , [a(\Lambda^{-1} \mathbf{p}) e^{i p \cdot x} + a^\dagger(\Lambda^{-1} \mathbf{p}) e^{-i p \cdot x}]. $$ Change variables $q = Λ^{- 1} p$ . The measure is Lorentz-invariant, $d μ_{m} (p) = d μ_{m} (q)$ , and the exponent transforms as $p \cdot x = (Λ q) \cdot x = q \cdot (Λ^{- 1} x)$ because the Minkowski inner product is preserved under $Λ \in S O^{+} (1, 3)$ . Substituting, $$ U(\Lambda, 0) \phi(x) U(\Lambda, 0)^{-1} = \int d\mu_m(\mathbf{q}) , [a(\mathbf{q}) e^{i q \cdot \Lambda^{-1} x} + a^\dagger(\mathbf{q}) e^{-i q \cdot \Lambda^{-1} x}] = \phi(\Lambda^{-1} x). $$ Reading the relation with $Λ$ replaced by $Λ^{- 1}$ gives the stated form $U (Λ, 0) ϕ (x) U (Λ, 0)^{- 1} = ϕ (Λ x)$ . Adding translations is the same argument with the additional exponential factor $e^{ia \cdot p}$ tracking translations. Rubric: full credit for the measure-invariance change of variables plus the inner-product invariance.

Advanced results Master

Theorem (Reeh-Schlieder; Reeh-Schlieder 1961 Nuovo Cim. 22, 1051). Let $ϕ$ be the free Klein-Gordon field on Minkowski space. For every nonempty open region $O \subset R^{1, 3}$ , the vacuum vector $Ω \in F_{s} (H_{1})$ is cyclic for the polynomial algebra $A (O)$ generated by smeared field operators with test functions supported in $O$ . That is, the set ${P Ω : P \in A (O)}$ is dense in $F_{s} (H_{1})$ .

The proof rests on the analyticity of $⟨ Ω∣ ϕ (x_{1}) \dots ϕ (x_{n}) ∣Ω ⟩$ as a function of the spacetime arguments in a forward tube domain (the spectrum condition: energy-momentum support in the closed forward light cone), combined with the edge-of-the-wedge theorem. A vector $Ψ$ orthogonal to $A (O) Ω$ has vanishing inner products against polynomials in $ϕ$ smeared in $O$ ; analytic continuation of those vanishing inner products in the test-function spacetime arguments forces them to vanish for arbitrary spacetime supports, so $Ψ$ is orthogonal to a dense set and is zero. The theorem expresses a striking non-locality of the vacuum: a localized operator in any arbitrarily small region prepares, by acting on $Ω$ , a state arbitrarily close to any chosen Fock-space vector. Causality is preserved through microcausality of the field commutator, not through localization of states.

Theorem (Wightman reconstruction). Given a sequence of tempered distributions $(W_{n})_{n \geq 0}$ on $(R^{1, 3})^{n}$ satisfying (i) Hermiticity, (ii) Lorentz covariance, (iii) the spectrum condition, (iv) positivity, (v) cluster decomposition, and (vi) microcausality, there exists, unique up to unitary equivalence, a quintuple $(H, U, Ω, D, ϕ)$ consisting of a Hilbert space, a continuous unitary Poincaré representation, a cyclic Poincaré-invariant vacuum, a dense invariant domain, and an operator-valued tempered distribution $ϕ$ on $D$ , such that the $W_{n}$ are the vacuum expectation values $W_{n} (x_{1}, \dots, x_{n}) = ⟨ Ω∣ ϕ (x_{1}) \dots ϕ (x_{n}) ∣Ω ⟩$ .

For the free Klein-Gordon field, the Wightman functions are Gaussian: $W_{2 k + 1} = 0$ for all odd $n$ , and $W_{2 k}$ is the sum over pair contractions of products of two-point functions $Δ_{+}$ . The reconstruction theorem identifies the free Klein-Gordon field as the unique scalar Wightman field with two-point function $Δ_{+} (x - y)$ on the mass shell of mass $m$ and zero higher truncated $n$ -point functions. The Wightman functions form a Gaussian generalized random process indexed by Schwartz functions on Minkowski space.

Theorem (CPT for the free scalar; Lüders 1954, Pauli 1955 — Lüders-Pauli construction in Streater-Wightman §4). The free Klein-Gordon scalar quantum field admits an anti-unitary operator $Θ$ on $F_{s} (H_{1})$ implementing the simultaneous reflection of spacetime $x \mapsto - x$ , charge conjugation (identity action for the real scalar), and complex conjugation: $Θ ϕ (x) Θ^{- 1} = ϕ (- x)$ , with $ΘΩ = Ω$ . For the complex (charged) Klein-Gordon field, $Θ$ also exchanges $a (p) \leftrightarrow b^{†} (p)$ between particle and antiparticle creation operators.

The CPT theorem in the Wightman framework expresses Lorentz covariance plus locality plus the spectrum condition as the existence of this anti-unitary symmetry. For the free scalar the theorem is direct from the Lorentz-invariance and Hermiticity of the mode expansion; the interacting case (Streater-Wightman §4) is the same statement promoted to the abstract Wightman theory. Lüders 1954 and Pauli 1955 are the originator-prose anchors.

Theorem (Haag's theorem; Haag 1955 — see Streater-Wightman §4.5). Two free Klein-Gordon fields of distinct mass $m_{1} \neq = m_{2}$ have unitarily inequivalent Fock-space representations of the canonical commutation relations.

The argument: the vacuum two-point function $Δ_{+}^{(m)}$ depends on $m$ , so two Fock representations associated with masses $m_{1} \neq = m_{2}$ have vacuum vectors with distinct expectation values for the same observables, an obstruction to the existence of a unitary intertwiner. This is the field-theoretic failure of the Stone-von-Neumann theorem (which, in finite-dimensional canonical-commutation-relation algebras, asserts uniqueness of the Fock representation). Haag's theorem motivates Wightman's distributional formulation and signals the impossibility of constructing interacting quantum field theories by perturbing free theories within a single Fock space — a key foundational issue.

Theorem (free-field two-point function as a Schwartz tempered distribution). The Wightman two-point function $Δ_{+} (z) = \int d^{3} p / [(2 π)^{3} 2 ω_{p}] e^{i p \cdot z}$ is the boundary value of the holomorphic function $Δ_{+} (z + i η)$ on the forward tube $R^{1, 3} + i V^{+}$ (with $V^{+}$ the open forward light cone in $R^{1, 3}$ ), and is a tempered distribution on $R^{1, 3}$ .

The forward-tube holomorphy follows from the spectrum condition: the support of the Fourier transform of $Δ_{+}$ in $p$ -space lies in the closed forward light cone, so the Fourier-Laplace transform converges on the forward tube. This is the analytic substrate that powers Reeh-Schlieder, the BMP theorem, and the edge-of-the-wedge arguments throughout axiomatic QFT.

Theorem (Bogoliubov axioms equivalence). The Bogoliubov-Shirkov $S$ -matrix axioms (translation invariance, Lorentz invariance, unitarity, microcausality of the $S$ -matrix, and the spectrum condition) plus the free-field asymptotic input determine the same theory as the Wightman field axioms applied to interpolating Heisenberg-picture fields.

In the free-Klein-Gordon case this equivalence is direct: the asymptotic free fields are the interpolating fields, the $S$ -matrix is the identity, and both axiom systems describe the same Fock-space theory. For interacting theories the equivalence is the LSZ-Bogoliubov-Shirkov programme.

Theorem (Garding-Wightman dense-domain property). The dense subspace $D \subset F_{s} (H_{1})$ of finite-particle Schwartz vectors (states $Ψ = (Ψ_{n})$ with $Ψ_{n} \in S (R^{3 n})$ for all $n$ and $Ψ_{n} = 0$ for all sufficiently large $n$ ) is invariant under every smeared field operator $ϕ (f)$ and is a core for each $ϕ (f)$ as an unbounded densely defined operator.

This is the technical underpinning that lets one treat the field as an unbounded operator-valued distribution rather than as an abstract sesquilinear form. The Garding-Wightman dense domain is the analogue of the Schwartz space in single-particle quantum mechanics.

Synthesis. The free Klein-Gordon scalar quantum field is the foundational reason every relativistic quantum theory begins from a free-field substrate: the construction identifies the unique Lorentz-covariant local realization of a mass- $m$ spin- $0$ Poincaré representation as a Fock-space-valued distribution, and the central insight is that microcausality, Lorentz covariance, the spectrum condition, and the cluster property of the vacuum together cut the free Klein-Gordon field out as the unique Wightman theory with the prescribed two-point function. Putting these together, the operator construction (Weinberg-style assembly from one-particle Poincaré representations), the canonical-quantization construction (Peskin-Schroeder via equal-time commutators on the classical Lagrangian), and the Wightman-axiom construction (existence-and-uniqueness via the reconstruction theorem) all produce identical Fock representations and identical Wightman functions; the three approaches are dual to one another in the sense that each one identifies the same operator-valued distribution from a distinct starting point. This is exactly the bridge that appears again in 12.13.02 (fermionic Fock space and Pauli exclusion), where the analogous three-way equivalence picks out the unique mass- $m$ spin- $1/2$ Wightman theory and forces the spin-statistics theorem as a derived consequence.

The bridge between the operator-theoretic and the path-integral framings is Wick rotation: the Wightman two-point function $Δ_{+}$ analytically continues to the Schwinger function $S_{2}$ on Euclidean spacetime, which is the Gaussian two-point function of the free Euclidean scalar field of mass $m$ , and the Osterwalder-Schrader reconstruction theorem identifies the Euclidean Gaussian process with the Lorentzian Wightman theory. The central insight here is that the free field is at once a Fock-space operator construction, a canonical-commutation realization, a Wightman tempered-distribution-valued field, and a Gaussian random process on Euclidean spacetime; the recurring pattern across these four perspectives is the role of the on-shell Lorentz-invariant measure and the resulting two-point function as the carrier of every kinematical fact about the theory.

The free Klein-Gordon field is dual to itself in a further precise sense: the Poincaré-covariant assembly of creation operators against the mass shell and the Poincaré-covariant assembly of annihilation operators against the dual mass shell are conjugate constructions, related by Hermitian conjugation of mode operators and complex conjugation of plane waves. This conjugation generalises to spin- $1/2$ as charge conjugation (electron-positron exchange) and to spin- $1$ as the same complex-conjugation pairing of polarization vectors. The pattern recurs: every higher-spin / charged-field construction is the free-scalar story re-run with the appropriate little-group representation replacing the one-dimensional $S O (3)$ representation of spin- $0$ .

Full proof set Master

Theorem (microcausality), proof. Given in the Intermediate-tier section. The argument has three steps: (i) reduce the operator commutator to a $c$ -number distribution by the canonical commutation relations on Fock space, (ii) identify the resulting distribution as the Pauli-Jordan tempered distribution $i Δ$ , which is Lorentz-invariant by direct computation, and (iii) show by an odd-even conflict (oddness under spatial inversion from the integrand, equality under spatial inversion from Lorentz invariance) that $Δ$ vanishes on every spacelike orbit. Smearing against Schwartz test functions of spacelike support gives $[ϕ (f), ϕ (g)] = 0$ . $□$

Theorem (Reeh-Schlieder), proof. Let $Ψ \in F_{s} (H_{1})$ be a vector orthogonal to $A (O) Ω$ for some nonempty open region $O$ , where $A (O)$ is the polynomial algebra of field operators smeared in $O$ . Then for every $n$ and every Schwartz test functions $f_{1}, \dots, f_{n}$ supported in $O$ , $⟨ Ψ∣ ϕ (f_{1}) \dots ϕ (f_{n}) ∣Ω ⟩ = 0$ . View this vanishing as the statement that the $n$ -point function $F_{n} (x_{1}, \dots, x_{n}) = ⟨ Ψ∣ ϕ (x_{1}) \dots ϕ (x_{n}) ∣Ω ⟩$ has vanishing pairings against products of Schwartz test functions with supports in $O$ . The spectrum condition (energy-momentum support in the forward light cone of the vacuum) implies that $F_{n}$ is the boundary value of a function holomorphic on the forward tube ${(x_{1}, \dots, x_{n}) : x_{j + 1} - x_{j} \in R^{1, 3} + i V^{+}}$ . The vanishing of $F_{n}$ on an open set in real spacetime, combined with the analyticity of its forward-tube extension, forces $F_{n}$ to vanish identically by the edge-of-the-wedge theorem and the analytic continuation principle. Hence $⟨ Ψ∣ ϕ (g_{1}) \dots ϕ (g_{n}) ∣Ω ⟩ = 0$ for arbitrary Schwartz test functions $g_{i}$ with arbitrary spacetime supports. By cyclicity of the vacuum for the global field algebra, the linear span of such vectors is dense in $F_{s} (H_{1})$ , so $Ψ = 0$ . $□$

Theorem (Wightman reconstruction), proof sketch — see Streater-Wightman §3.4 ^{[source pending]}. Given the sequence $(W_{n})$ satisfying the axioms, build the Hilbert space as the quotient of the free vector space on finite sequences of Schwartz test functions $⨁_{n} S ((R^{1, 3})^{n})$ by the radical of the positivity-induced sesquilinear form $⟨ f, g ⟩ = \sum_{m, n} \overline{f_{m} (x_{1}, \dots, x_{m})} g_{n} (y_{1}, \dots, y_{n}) W_{m + n} (\overset{x}{ˉ}_{m}, \dots, \overset{x}{ˉ}_{1}, y_{1}, \dots, y_{n})$ . The cyclic vacuum is the class of the unit vector in degree zero. The field operator $ϕ (f)$ acts on the equivalence class of $(f_{1}, \dots, f_{n})$ by prepending $f$ . Lorentz covariance of the $W_{n}$ produces the unitary Poincaré representation on the constructed Hilbert space. Cluster decomposition gives uniqueness of the vacuum and irreducibility of the field algebra. Microcausality gives locality. The constructed quintuple is unique up to unitary equivalence by the Gelfand-Naimark-Segal construction. $□$

Theorem (Gaussianity of the free Klein-Gordon Wightman functions), proof. Compute $W_{n} (x_{1}, \dots, x_{n}) = ⟨ Ω∣ ϕ (x_{1}) \dots ϕ (x_{n}) ∣Ω ⟩$ by repeated substitution of the mode expansion of each $ϕ (x_{j})$ and use $a (p) ∣Ω ⟩ = 0$ , $⟨ Ω∣ a^{†} (p) = 0$ to keep only the diagrams in which every $a$ is paired with an earlier $a^{†}$ via the canonical commutator. For odd $n$ every Wick contraction leaves at least one unpaired creation or annihilation operator, so $W_{2 k + 1} = 0$ . For even $n = 2 k$ the result is the sum over pair partitions of ${1, \dots, n}$ of products of two-point functions: $W_{2 k} (x_{1}, \dots, x_{2 k}) = \sum_{π} \prod_{(i, j) \in π} Δ_{+} (x_{i} - x_{j})$ , where the sum is over the $(2 k - 1)!! = (2 k)! / [2^{k} k!]$ pair partitions with $i < j$ in each pair (Wick's theorem for the free field). $□$

Theorem (Lorentz invariance of $d μ_{m}$ ), proof. The measure $d μ_{m} (p) = d^{3} p / [(2 π)^{3} 2 ω_{p}]$ is the pushforward to the mass-shell parametrization $p \mapsto (ω_{p}, p)$ of the Lorentz-invariant four-dimensional measure $θ (p^{0}) δ (p^{2} + m^{2}) d^{4} p$ , since on the upper mass shell $δ (p^{2} + m^{2}) = δ ((p^{0})^{2} - ω_{p}^{2}) = δ (p^{0} - ω_{p}) / (2 ω_{p})$ . The four-dimensional measure $θ (p^{0}) δ (p^{2} + m^{2}) d^{4} p$ is manifestly invariant under $S O^{+} (1, 3)$ : the step function $θ (p^{0})$ is preserved by the proper orthochronous component, the on-shell delta is preserved by the metric-preserving action, and the four-dimensional Lebesgue measure $d^{4} p$ is volume-preserving. Hence $d μ_{m}$ is Lorentz-invariant. $□$

Theorem (Feynman propagator as time-ordered two-point function), proof. By definition the time-ordered product is $T ϕ (x) ϕ (y) = θ (x^{0} - y^{0}) ϕ (x) ϕ (y) + θ (y^{0} - x^{0}) ϕ (y) ϕ (x)$ . Taking the vacuum expectation value and using $W_{2} (x, y) = Δ_{+} (x - y)$ and $W_{2} (y, x) = Δ_{+} (y - x) = Δ_{-} (x - y)$ , $$ \Delta_F(x - y) = \theta(x^0 - y^0) \Delta_+(x - y) + \theta(y^0 - x^0) \Delta_-(x - y). $$ Substitute the integral representations of $Δ_{\pm}$ on the mass shell, perform the $p^{0}$ integral by closing the contour in the upper half plane for $y^{0} > x^{0}$ (picking up the negative-frequency pole) and the lower half plane for $x^{0} > y^{0}$ (picking up the positive-frequency pole). The two contour-integral results combine into the Feynman-prescription momentum-space integral $$ \Delta_F(z) = \int \frac{d^4 p}{(2\pi)^4} , \frac{i , e^{i p \cdot z}}{p^2 + m^2 - i \varepsilon}, $$ where the $i ε$ prescription is precisely the contour selection. $□$

Theorem (Haag's theorem for distinct masses), proof sketch — see Streater-Wightman §4.5 ^{[source pending]}. Suppose two free Klein-Gordon fields of masses $m_{1} \neq = m_{2}$ have unitarily equivalent Fock representations: there exists a unitary $V : F_{s} (H_{1}^{(m_{1})}) \to F_{s} (H_{1}^{(m_{2})})$ such that $V ϕ_{1} (f) V^{- 1} = ϕ_{2} (f)$ for every Schwartz test function $f$ . Then the Wightman functions agree: $W_{2}^{(m_{1})} (f \otimes g) = ⟨ V Ω_{1} ∣ ϕ_{2} (f) ϕ_{2} (g) ∣ V Ω_{1} ⟩$ . Choose $V Ω_{1} = c \cdot Ω_{2}$ for some complex unit $c$ (which is forced by cyclic-vacuum uniqueness and Poincaré covariance up to a phase). Then $W_{2}^{(m_{1})} = W_{2}^{(m_{2})}$ . But the explicit form $W_{2}^{(m)} (z) = \int d^{3} p / [(2 π)^{3} 2 ω_{p}^{(m)}] e^{i p \cdot z}_{p^{0} = ω_{p}^{(m)}}$ depends on $m$ through the dispersion relation, contradicting $W_{2}^{(m_{1})} = W_{2}^{(m_{2})}$ for $m_{1} \neq = m_{2}$ . $□$

Connections Master

Bosonic Fock space and second quantisation 12.13.01. The free Klein-Gordon scalar quantum field is the Fock-space-valued operator construction one obtains by applying the second-quantization functor of Cook 1953 to the one-particle Hilbert space $H_{1} = L^{2} (X_{m}, d μ_{m})$ . Every commutation relation, vacuum expectation value, and Wick-contraction identity in this unit is an instance of the general bosonic Fock-space machinery specialised to the mass-shell one-particle space. Without the Fock-space machinery there is no field operator; with it, the field operator is forced to take the mode-expansion form by Poincaré covariance.
Hilbert space 02.11.08. The state space of the free Klein-Gordon quantum field is a separable complex Hilbert space, the bosonic Fock space $F_{s} (H_{1})$ . Every inner product, projection, and self-adjoint extension result invoked in this unit comes from the general Hilbert-space framework. The technical points where unbounded operators (the smeared field, the canonical momentum density, the Hamiltonian) act on the dense Garding-Wightman domain are direct applications of the Hilbert-space adjoint theory.
Unbounded self-adjoint operators 02.11.03. The smeared field operator $ϕ (f)$ , the canonical momentum density $π (f)$ , the Hamiltonian, and the number operator are all unbounded densely defined self-adjoint operators on the Fock space. The Garding-Wightman dense domain serves as a core for each one. The technical machinery of self-adjoint extensions, deficiency indices, and Stone's theorem on one-parameter unitary groups is the analytic substrate that makes the canonical commutation relations and the Heisenberg-picture evolution mathematically rigorous.
Noether theorem 05.00.04. The conserved currents of the free Klein-Gordon field (energy-momentum tensor, angular momentum, charge for the complex case) are derived from the classical Noether construction applied to the spacetime symmetries (translations, Lorentz transformations) and internal symmetries (the global $U (1)$ for the complex field) of the Klein-Gordon Lagrangian. Quantization promotes the conserved charges to self-adjoint operators on Fock space, identified with the generators of the unitary Poincaré representation. The Poincaré covariance of the field operator $U (Λ, a) ϕ (x) U (Λ, a)^{- 1} = ϕ (Λ x + a)$ is the second-quantized version of the classical Noether-current identity.

Historical & philosophical context Master

The Klein-Gordon equation as a relativistic wave equation predates quantum field theory by a decade. Oskar Klein in March 1926 (Z. Phys. 37, 895) and Walter Gordon in September 1926 (Z. Phys. 40, 117) independently derived the equation $(□ + m^{2}) ϕ = 0$ as the relativistic completion of the Schrödinger equation ^{[source pending]}. Vladimir Fock in the same year (Z. Phys. 39, 226) ^{[source pending]} obtained the minimally coupled version with electromagnetic potentials. Schrödinger himself, in unpublished notes from late 1925, had derived the same equation but discarded it on noticing that its hydrogenic spectrum disagreed with the relativistic correction observed in Sommerfeld's fine-structure formula. The wave equation was understood as a first-quantized theory, with $ϕ$ a complex-valued single-particle wave function on Minkowski space. The serious difficulty noticed almost immediately was that the natural conserved current $j^{μ} = i (ϕ^{*} \partial^{μ} ϕ - ϕ \partial^{μ} ϕ^{*})$ is not positive-definite, so it cannot be interpreted as a probability current — a fatal flaw in any single-particle quantum-mechanical reading.

The resolution arrived only with second quantization, in the work of Pauli and Weisskopf 1934 (Helv. Phys. Acta 7, 709) ^{[source pending]}. Their paper reinterpreted the negative-frequency solutions of the Klein-Gordon equation as creation operators for antiparticles in a Fock-space framework, converting the negative-probability difficulty into a positive-energy spectrum and a conserved electric charge that can take either sign. The Pauli-Weisskopf construction was contemporaneous with Fock's 1932 (Z. Phys. 75, 622) systematic introduction of what is now called Fock space, and the operator-valued reinterpretation of the wave equation entered the mainstream of theoretical physics through Heisenberg, Dirac, and Jordan's earlier (1929-30) covariant quantum field theory programme. By 1939, Wigner's Ann. Math. 40 classification of the unitary irreducible representations of the Poincaré group ^{[source pending]} had organised the relativistic free-field landscape around mass and spin, with the free Klein-Gordon field identified as the mass- $m$ spin- $0$ case.

The modern axiomatic formulation, in which the free Klein-Gordon field is the unique scalar Wightman theory with prescribed two-point function, is due to Wightman 1956 (Phys. Rev. 101, 860) ^{[source pending]} and the consolidated Streater-Wightman 1964 PCT, Spin and Statistics, and All That ^{[source pending]}. Reeh and Schlieder 1961 (Nuovo Cim. 22, 1051) ^{[source pending]} established the cyclicity theorem that bears their names. The mathematical-physics framework — operator-valued tempered distributions on Schwartz space, the Garding-Wightman dense domain, the GNS-style reconstruction theorem — was completed by Wightman and Gårding 1964 (Ark. Fys. 28, 129) and absorbed into the constructive QFT programme of Glimm and Jaffe (1968-87) and the algebraic QFT programme of Haag and Kastler 1964 (J. Math. Phys. 5, 848). Weinberg's 1995 Quantum Theory of Fields Vol. 1 ^{[source pending]} presents the operator construction by direct assembly from Poincaré irreps as the foundational route to the free Klein-Gordon field, with canonical quantization treated as a derived computational tool — the inverse of the standard physicist's order, but the order favoured by mathematical-physics axiomatic developments since Wightman.

Bibliography Master

@article{Klein1926,
  author  = {Klein, Oskar},
  title   = {Quantentheorie und f{\"u}nfdimensionale Relativit{\"a}tstheorie},
  journal = {Z. Phys.},
  volume  = {37},
  year    = {1926},
  pages   = {895--906}
}

@article{Gordon1926,
  author  = {Gordon, Walter},
  title   = {Der Comptoneffekt nach der Schr{\"o}dingerschen Theorie},
  journal = {Z. Phys.},
  volume  = {40},
  year    = {1926},
  pages   = {117--133}
}

@article{Fock1926,
  author  = {Fock, Vladimir},
  title   = {{\"U}ber die invariante Form der Wellen- und der Bewegungsgleichungen f{\"u}r einen geladenen Massenpunkt},
  journal = {Z. Phys.},
  volume  = {39},
  year    = {1926},
  pages   = {226--232}
}

@article{PauliWeisskopf1934,
  author  = {Pauli, Wolfgang and Weisskopf, Victor},
  title   = {{\"U}ber die Quantisierung der skalaren relativistischen Wellengleichung},
  journal = {Helv. Phys. Acta},
  volume  = {7},
  year    = {1934},
  pages   = {709--731}
}

@article{Fock1932,
  author  = {Fock, Vladimir},
  title   = {Konfigurationsraum und zweite Quantelung},
  journal = {Z. Phys.},
  volume  = {75},
  year    = {1932},
  pages   = {622--647}
}

@article{Wigner1939,
  author  = {Wigner, Eugene P.},
  title   = {On unitary representations of the inhomogeneous Lorentz group},
  journal = {Ann. Math.},
  volume  = {40},
  year    = {1939},
  pages   = {149--204}
}

@article{Wightman1956,
  author  = {Wightman, Arthur S.},
  title   = {Quantum field theory in terms of vacuum expectation values},
  journal = {Phys. Rev.},
  volume  = {101},
  year    = {1956},
  pages   = {860--866}
}

@article{ReehSchlieder1961,
  author  = {Reeh, Helmut and Schlieder, Siegfried},
  title   = {Bemerkungen zur Unit{\"a}r{\"a}quivalenz von Lorentzinvarianten Feldern},
  journal = {Nuovo Cimento},
  volume  = {22},
  year    = {1961},
  pages   = {1051--1068}
}

@book{StreaterWightman1964,
  author    = {Streater, Raymond F. and Wightman, Arthur S.},
  title     = {PCT, Spin and Statistics, and All That},
  publisher = {W. A. Benjamin},
  year      = {1964},
  note      = {Princeton Landmarks in Physics reprint, 2000}
}

@book{Weinberg1995,
  author    = {Weinberg, Steven},
  title     = {The Quantum Theory of Fields, Vol. 1: Foundations},
  publisher = {Cambridge University Press},
  year      = {1995}
}

@book{ReedSimonII1975,
  author    = {Reed, Michael and Simon, Barry},
  title     = {Methods of Modern Mathematical Physics, Vol. II: Fourier Analysis, Self-Adjointness},
  publisher = {Academic Press},
  year      = {1975}
}

@book{PeskinSchroeder1995,
  author    = {Peskin, Michael E. and Schroeder, Daniel V.},
  title     = {An Introduction to Quantum Field Theory},
  publisher = {Westview Press},
  year      = {1995}
}

@book{Folland2008,
  author    = {Folland, Gerald B.},
  title     = {Quantum Field Theory: A Tourist Guide for Mathematicians},
  publisher = {American Mathematical Society},
  series    = {Math. Surveys},
  volume    = {149},
  year      = {2008}
}

@book{GlimmJaffe1987,
  author    = {Glimm, James and Jaffe, Arthur},
  title     = {Quantum Physics: A Functional Integral Point of View},
  publisher = {Springer},
  edition   = {2},
  year      = {1987}
}

@article{HaagKastler1964,
  author  = {Haag, Rudolf and Kastler, Daniel},
  title   = {An algebraic approach to quantum field theory},
  journal = {J. Math. Phys.},
  volume  = {5},
  year    = {1964},
  pages   = {848--861}
}

@article{Lueders1954,
  author  = {L{\"u}ders, Gerhart},
  title   = {On the equivalence of invariance under time reversal and under particle-antiparticle conjugation for relativistic field theories},
  journal = {Dan. Mat. Fys. Medd.},
  volume  = {28},
  year    = {1954},
  pages   = {1--17}
}

Prerequisites

12.13.01
02.11.08
02.11.03
05.00.04

Tier anchors

beginner: Zee, *Quantum Field Theory in a Nutshell*, 2e (Princeton 2010), Ch. I.3-I.7; Lancaster & Blundell, *Quantum Field Theory for the Gifted Amateur* (Oxford 2014), Ch. 4-7
intermediate: Peskin & Schroeder, *An Introduction to Quantum Field Theory* (Westview 1995), Ch. 2 (canonical quantization of the real scalar field); Tong, *Quantum Field Theory* lecture notes, §2; Srednicki, *Quantum Field Theory* (Cambridge 2007), Ch. 3
master: Weinberg, *The Quantum Theory of Fields, Vol. 1* (Cambridge 1995), §5.2 (the scalar-field case of the Poincaré-irrep construction); Streater & Wightman, *PCT, Spin and Statistics, and All That* (Benjamin 1964; Princeton Landmarks 2000), Ch. 3 (Wightman-axiom realization); Reed & Simon, *Methods of Modern Mathematical Physics, Vol. II*, §X.7 (Fock-space functor and second quantization); Folland, *Quantum Field Theory: A Tourist Guide for Mathematicians* (AMS 2008), Ch. 5

References

Weinberg, S. — The Quantum Theory of Fields, Vol. 1 · Cambridge University Press (1995), §5.2 (scalar field construction from Poincaré reps), §4.1 (cluster decomposition), §3.1 (S-matrix definition)
Klein, O. — Quantentheorie und fünfdimensionale Relativitätstheorie · Z. Phys. 37, 895 (1926); the wave equation for a relativistic scalar
Gordon, W. — Der Comptoneffekt nach der Schrödingerschen Theorie · Z. Phys. 40, 117 (1926); independent derivation of the wave equation
Fock, V. — Über die invariante Form der Wellen- und der Bewegungsgleichungen für einen geladenen Massenpunkt · Z. Phys. 39, 226 (1926); minimally-coupled relativistic wave equation
Pauli, W. & Weisskopf, V. — Über die Quantisierung der skalaren relativistischen Wellengleichung · Helv. Phys. Acta 7, 709 (1934); resolution of negative-energy/probability via second quantization, introduction of antiparticles for the charged scalar
Wightman, A. S. — Quantum field theory in terms of vacuum expectation values · Phys. Rev. 101, 860 (1956); the axiomatic distributional foundation later expanded in Streater-Wightman 1964
Streater, R. F. & Wightman, A. S. — PCT, Spin and Statistics, and All That · Benjamin (1964); Princeton Landmarks reprint (2000); Ch. 3 (Wightman axioms and the reconstruction theorem); Ch. 4 (PCT theorem); §4.1 (free field as the unique two-point Wightman theory)
Reeh, H. & Schlieder, S. — Bemerkungen zur Unitäräquivalenz von Lorentzinvarianten Feldern · Nuovo Cimento 22, 1051 (1961); cyclicity of the vacuum for local algebras
Reed, M. & Simon, B. — Methods of Modern Mathematical Physics, Vol. II · Academic Press (1975), §X.7 (Fock spaces, second quantization functor $\Gamma$, the free scalar field as $d\Gamma$ of multiplication by $\omega_p$)
Glimm, J. & Jaffe, A. — Quantum Physics: A Functional Integral Point of View · Springer (2nd ed. 1987), Ch. 6-7 (Gaussian-process / Fock-space correspondence; the free Euclidean scalar field)
Peskin, M. E. & Schroeder, D. V. — An Introduction to Quantum Field Theory · Westview Press (1995), Ch. 2 (canonical quantization of the real Klein-Gordon field), §2.4 (Feynman propagator and contour prescription)
Bogoliubov, N. N. & Shirkov, D. V. — Introduction to the Theory of Quantized Fields · Interscience (1959; 3rd ed. 1980); Ch. 1-2 (operator formulation, distributional products, the canonical free-scalar construction)

Estimated time

beginner: 22m
intermediate: 50m
master: 90m