Dirac field $\psi$ and the Dirac adjoint $\bar{\psi }$

Intuition Beginner

A photon is a wave; an electron is a particle that turns out to behave like a wave too. The wave equation that describes a slow electron — the Schrödinger equation — treats time and space differently, with one time derivative on one side and two space derivatives on the other. Special relativity insists that time and space sit on equal footing. Paul Dirac in 1928 looked for a wave equation that is first order in time AND first order in space. He found that the price of getting it is to replace the single complex number at each point with a column of four complex numbers. That column is the Dirac field $\psi$.

The four entries are not four particles; they are four channels that encode two physical pieces of data at once. Two of the channels describe an electron, with spin up and spin down. The other two describe its mirror partner, the positron — the antiparticle whose existence Dirac predicted from the equation and which Carl Anderson detected in cosmic rays in 1932. The four-component column is the smallest object that can hold an electron, a positron, and the two spin states of each, while transforming consistently under Lorentz boosts and rotations.

To turn $\psi$ into a number you can write inner products with, you need a partner. The naive partner ${\psi }^{†}$ (column flipped to row, with complex conjugation) does not work — the product ${\psi }^{†}\psi$ does not behave as a Lorentz scalar. The right partner is $\bar{\psi }={\psi }^{†}{\gamma }^0$, called the Dirac adjoint. With this small adjustment, $\bar{\psi }\psi$ is a Lorentz scalar, $\bar{\psi }{\gamma }^{\mu }\psi$ is a Lorentz vector (a conserved current), and the whole zoo of gamma-bilinears builds the quantities physicists measure.

Visual Beginner

A schematic of the Dirac field as a four-component column attached to every point of spacetime. At each point, the column has two "upper" entries that describe an electron in two spin states and two "lower" entries that describe a positron in two spin states. Arrows indicate how the column transforms under a Lorentz boost — it does not stay rigid, but it mixes its four entries according to a $4\times 4$ matrix $S(\Lambda )$ called the spinor representation of the Lorentz group. The mixing is the four-component analogue of how a three-vector rotates under a rotation of space.

The picture also names the partner. Below the column $\psi$ sits the row $\bar{\psi }={\psi }^{†}{\gamma }^0$. The single matrix ${\gamma }^0$ that converts the naive flipped-and-conjugated row into the Dirac adjoint is small, but it carries all the load: without that one ${\gamma }^0$, the bilinear ${\psi }^{†}\psi$ is the time component of a four-vector, not a scalar. With it, $\bar{\psi }\psi$ is a number every observer agrees on.

Worked example Beginner

Verify the anticommutator $\{{\gamma }^0,{\gamma }^1\}={\gamma }^0{\gamma }^1+{\gamma }^1{\gamma }^0=0$ in the standard Dirac representation, where ${\gamma }^0$ and ${\gamma }^1$ are explicit $4\times 4$ matrices built from $2\times 2$ Pauli blocks.

Step 1. Write the matrices. In the Dirac representation, $$ \gamma^0 = \begin{pmatrix} I & 0 \ 0 & -I \end{pmatrix}, \qquad \gamma^1 = \begin{pmatrix} 0 & \sigma^1 \ -\sigma^1 & 0 \end{pmatrix}, $$ where $I$ is the $2\times 2$ identity and ${\sigma }^1=\left(\begin{array}{cc}0& 1\\ 1& 0\end{array}\right)$ is the first Pauli matrix. Each block is $2\times 2$, so each full matrix is $4\times 4$.

Step 2. Multiply ${\gamma }^0{\gamma }^1$ using block multiplication. $$ \gamma^0 \gamma^1 = \begin{pmatrix} I & 0 \ 0 & -I \end{pmatrix} \begin{pmatrix} 0 & \sigma^1 \ -\sigma^1 & 0 \end{pmatrix} = \begin{pmatrix} 0 & \sigma^1 \ \sigma^1 & 0 \end{pmatrix}. $$ The upper-left block is $I\cdot 0=0$, the upper-right is $I\cdot {\sigma }^1={\sigma }^1$, the lower-left is $(-I)\cdot (-{\sigma }^1)={\sigma }^1$, the lower-right is $(-I)\cdot 0=0$.

Step 3. Multiply ${\gamma }^1{\gamma }^0$ the other way. $$ \gamma^1 \gamma^0 = \begin{pmatrix} 0 & \sigma^1 \ -\sigma^1 & 0 \end{pmatrix} \begin{pmatrix} I & 0 \ 0 & -I \end{pmatrix} = \begin{pmatrix} 0 & -\sigma^1 \ -\sigma^1 & 0 \end{pmatrix}. $$ The upper-right block is ${\sigma }^1\cdot (-I)=-{\sigma }^1$, the lower-left is $(-{\sigma }^1)\cdot I=-{\sigma }^1$. The other two blocks vanish.

Step 4. Add the two products. $$ \gamma^0 \gamma^1 + \gamma^1 \gamma^0 = \begin{pmatrix} 0 & \sigma^1 - \sigma^1 \ \sigma^1 - \sigma^1 & 0 \end{pmatrix} = \begin{pmatrix} 0 & 0 \ 0 & 0 \end{pmatrix} = 0. $$

Step 5. Read off the result. The anticommutator is the zero matrix. Compare to the Clifford rule $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }1$ with ${\eta }^{01}=0$ in the mostly-minus Minkowski metric. The rule predicts $\{{\gamma }^0,{\gamma }^1\}=2\cdot 0\cdot 1=0$, which matches.

What this tells us: the gamma matrices are not ordinary numbers — they anticommute when their indices differ, and they have a fixed square when their indices match. The Clifford rule packages every product into one of these two cases, and the whole algebra of Dirac spinors is built on this rule plus the four-component column $\psi$.

Check your understanding Beginner

Formal definition Intermediate+

Fix Minkowski spacetime ${\mathbb{R}}^{1,3}$ with mostly-minus metric $\eta =\mathrm{diag}(+1,-1,-1,-1)$ and coordinates $x^{\mu }$ with $\mu \in \{0,1,2,3\}$. The Clifford algebra $\mathrm{Cl}(1,3)$ is the unital associative algebra generated by symbols ${\gamma }^{\mu }$ subject to the relation $$ {\gamma^\mu, \gamma^\nu} := \gamma^\mu \gamma^\nu + \gamma^\nu \gamma^\mu = 2 \eta^{\mu\nu} \mathbb{1}, $$ where $1$ is the identity. A complex representation $\rho :\mathrm{Cl}(1,3){\otimes }_{\mathbb{R}}\mathbb{C}\to M_4(\mathbb{C})$ as $4\times 4$ matrices exists, is irreducible, and is unique up to similarity transformation; we abuse notation and write ${\gamma }^{\mu }\in M_4(\mathbb{C})$ for the images. A Dirac spinor is a map $\psi :{\mathbb{R}}^{1,3}\to {\mathbb{C}}^4$. The Dirac equation is the first-order PDE $$ (i \gamma^\mu \partial_\mu - m \mathbb{1}) \psi(x) = 0, $$ with $m\ge 0$ the mass parameter and Einstein summation over $\mu$. Applying the differential operator $(i{\gamma }^{\mu }{\partial }_{\mu }+m1)$ to both sides, and using the Clifford relation to symmetrise the indices in ${\gamma }^{\mu }{\gamma }^{\nu }{\partial }_{\mu }{\partial }_{\nu }\psi =\frac{1}{2}\{{\gamma }^{\mu },{\gamma }^{\nu }\}{\partial }_{\mu }{\partial }_{\nu }\psi ={\eta }^{\mu \nu }{\partial }_{\mu }{\partial }_{\nu }\psi =\square \psi$, one obtains $-(\square +m^2)\psi =0$: every component of a Dirac-equation solution satisfies the Klein-Gordon equation.

The Dirac adjoint is $$ \bar\psi(x) := \psi(x)^\dagger \gamma^0, $$ a row-valued spinor on ${\mathbb{R}}^{1,3}$. The defining role of the factor ${\gamma }^0$ is the identity $({\gamma }^{\mu }{)}^{†}={\gamma }^0{\gamma }^{\mu }{\gamma }^0$, valid in every representation of the Clifford algebra (it follows from ${\gamma }^0$ being Hermitian and the spatial ${\gamma }^i$ being anti-Hermitian in the standard physicists' normalisation). The Lorentz transformation law $\psi (x)\to S(\Lambda )\psi ({\Lambda }^{-1}x)$, with $S(\Lambda )\in \mathrm{GL}(4,\mathbb{C})$ the spinor representation of $\mathrm{SL}(2,\mathbb{C})$ obeying the intertwining identity $S(\Lambda {)}^{-1}{\gamma }^{\mu }S(\Lambda )={\Lambda }_{\nu }^{\mu }{\gamma }^{\nu }$, combines with the relation $S(\Lambda {)}^{†}{\gamma }^0={\gamma }^{0}S(\Lambda {)}^{-1}$ to give the dual transformation $\bar{\psi }(x)\to \bar{\psi }({\Lambda }^{-1}x)S(\Lambda {)}^{-1}$.

Five independent Lorentz-covariant bilinears in $\bar{\psi }$ and $\psi$ span the 16-dimensional space of $4\times 4$ matrix bilinears:

• Scalar: $\bar{\psi }\psi$, transforming as $\bar{\psi }\psi \to \bar{\psi }\psi$.
• Vector: $\bar{\psi }{\gamma }^{\mu }\psi$, transforming as $\bar{\psi }{\gamma }^{\mu }\psi \to {\Lambda }_{\nu }^{\mu }\bar{\psi }{\gamma }^{\nu }\psi$. This is the conserved Noether current associated with the global $\mathrm{U}(1)$ phase symmetry $\psi \to e^{i\alpha }\psi$, and ${\partial }_{\mu }(\bar{\psi }{\gamma }^{\mu }\psi )=0$ on solutions of the Dirac equation.
• Antisymmetric tensor: $\bar{\psi }{\sigma }^{\mu \nu }\psi$ with ${\sigma }^{\mu \nu }:=\frac{i}{2}[{\gamma }^{\mu },{\gamma }^{\nu }]$, transforming as a rank-2 antisymmetric tensor.
• Pseudoscalar: $\bar{\psi }{\gamma }^5\psi$ with ${\gamma }^5:=i{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3$ satisfying $\{{\gamma }^5,{\gamma }^{\mu }\}=0$ and $({\gamma }^5{)}^2=1$; under parity $P$ the bilinear changes sign.
• Axial vector: $\bar{\psi }{\gamma }^{\mu }{\gamma }^5\psi$, transforming as a vector under ${\mathrm{SO}}^{\uparrow }(1,3)$ but with opposite sign under parity.

Plane-wave solutions of the Dirac equation come in two branches. For each on-shell momentum $p^{\mu }$ with $p^2=m^2,p^0>0$, the positive-energy spinors $u_s(p)\in {\mathbb{C}}^4$ for $s\in \{1,2\}$ satisfy $({\gamma }^{\mu }{p}_{\mu }-m)u_s(p)=0$ and the negative-energy spinors $v_s(p)$ satisfy $({\gamma }^{\mu }{p}_{\mu }+m)v_s(p)=0$. They are normalised by ${\bar{u}}_r(p)u_s(p)=2m{\delta }_{rs}$ and ${\bar{v}}_r(p)v_s(p)=-2m{\delta }_{rs}$, with completeness relations ${\sum }_s{u}_s(p){\bar{u}}_s(p)={\gamma }^{\mu }{p}_{\mu }+m$ and ${\sum }_s{v}_s(p){\bar{v}}_s(p)={\gamma }^{\mu }{p}_{\mu }-m$.

Second quantisation promotes $\psi$ from a classical ${\mathbb{C}}^4$-valued function to an operator-valued tempered distribution on the fermionic Fock space 12.13.02 over the positive-energy one-particle Hilbert space: $$ \psi(x) = \int \frac{d^3 p}{(2\pi)^3 \sqrt{2 E_p}} \sum_{s=1}^{2} \Bigl[ a_{s,p}, u_s(p), e^{-i p \cdot x} + b_{s,p}^\dagger, v_s(p), e^{+i p \cdot x} \Bigr], $$ where $E_p=(|\mathbf{p}{|}^2+m^2{)}^{1/2}$, the operators $a_{s,p}$ annihilate electrons of spin $s$ and momentum $p$, and $b_{s,p}^{†}$ create positrons. The canonical anticommutation relations are $\{a_{r,p},a_{s,q}^{†}\}=\{b_{r,p},b_{s,q}^{†}\}=(2\pi {)}^3{\delta }_{rs}{\delta }^{(3)}(\mathbf{p}-\mathbf{q})$ with all other anticommutators vanishing. The fermionic choice (anticommutators rather than commutators) is forced by the spin-statistics theorem ^{[Streater-Wightman 1964]}.

Counterexamples to common slips

• The bilinear ${\psi }^{†}\psi$ is not a Lorentz scalar; it is the zero-component of a four-vector current. The ${\gamma }^0$ in the Dirac adjoint is mandatory, not cosmetic.
• The pseudoscalar $\bar{\psi }{\gamma }^5\psi$ looks like a scalar at first inspection — it is a number, after all — but under a parity transformation $\mathbf{x}\to -\mathbf{x}$, the spinor field transforms as $\psi (t,\mathbf{x})\to {\gamma }^0\psi (t,-\mathbf{x})$ and the bilinear changes sign. The "pseudo" prefix records this sign flip.
• The positive-energy spinors $u_s(p)$ and negative-energy spinors $v_s(p)$ are not orthogonal in the standard scalar sense — instead ${\bar{u}}_r(p)v_s(p)=0$ uses the Dirac-adjoint pairing. Crossing this with the usual $†$ inner product gives the wrong answer; the Dirac structure has its own bilinear.
• The mode expansion mixes $a$ with the conjugate of a different operator $b^{†}$, not with $a^{†}$. The reason is that the negative-frequency branch of the classical field is reinterpreted as creation of an antiparticle (a positron) of opposite charge: the same one-particle solution $v_s(p)$ that looks "negative energy" in classical theory becomes a positive-energy positron state in the quantised theory. This is the operational content of Dirac's hole picture, but cleaner.

Key theorem with proof Intermediate+

Theorem (Klein-Gordon from Dirac; Dirac 1928, ^{[Dirac 1928]}). Let $\psi :{\mathbb{R}}^{1,3}\to {\mathbb{C}}^4$ be a smooth solution of the Dirac equation $(i{\gamma }^{\mu }{\partial }_{\mu }-m1)\psi =0$ with ${\gamma }^{\mu }$ satisfying the Clifford relation $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }1$. Then every component ${\psi }_a$ for $a\in \{1,2,3,4\}$ separately satisfies the Klein-Gordon equation $$ (\Box + m^2) \psi_a(x) = 0, \qquad \Box := \eta^{\mu\nu} \partial_\mu \partial_\nu = \partial_0^2 - \nabla^2. $$

Proof. Apply the operator $(i{\gamma }^{\nu }{\partial }_{\nu }+m1)$ to both sides of the Dirac equation. The left side becomes $$ (i \gamma^\nu \partial_\nu + m \mathbb{1})(i \gamma^\mu \partial_\mu - m \mathbb{1}) \psi = (- \gamma^\nu \gamma^\mu \partial_\nu \partial_\mu - m^2 \mathbb{1}) \psi + i m \gamma^\mu \partial_\mu \psi - i m \gamma^\nu \partial_\nu \psi. $$ The last two cross terms cancel after relabelling the dummy indices, leaving $$ (i\gamma^\nu \partial_\nu + m)(i \gamma^\mu \partial_\mu - m) \psi = - \gamma^\nu \gamma^\mu \partial_\nu \partial_\mu \psi - m^2 \psi. $$ The right side is zero (the original equation gives $(i{\gamma }^{\mu }{\partial }_{\mu }-m)\psi =0$, so multiplying by the conjugate operator preserves zero).

Symmetrise the index pair in ${\gamma }^{\nu }{\gamma }^{\mu }{\partial }_{\nu }{\partial }_{\mu }$. The partial derivatives commute, so ${\partial }_{\nu }{\partial }_{\mu }={\partial }_{\mu }{\partial }_{\nu }$ is symmetric in $\mu ,\nu$. Therefore, only the symmetric part of ${\gamma }^{\nu }{\gamma }^{\mu }$ contributes, and the symmetric part is exactly half the anticommutator: $$ \gamma^\nu \gamma^\mu \partial_\nu \partial_\mu = \tfrac{1}{2} {\gamma^\nu, \gamma^\mu} \partial_\nu \partial_\mu = \tfrac{1}{2} \cdot 2 \eta^{\nu\mu} \partial_\nu \partial_\mu \psi = \eta^{\mu\nu} \partial_\mu \partial_\nu \psi = \Box \psi. $$

Combining: $-\square \psi -m^2\psi =0$, hence $(\square +m^2)\psi =0$. The $1$ in the operator acts component-wise, so each ${\psi }_a$ separately satisfies $(\square +m^2){\psi }_a=0$. $\square$

Bridge. This computation builds toward every Dirac-field application that follows, and appears again in 08.10.11 as the operator-side substrate for QED scattering amplitudes. The foundational reason it works is exactly the Clifford relation: the anticommutator $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }1$ is the algebraic content of "taking the square root of the d'Alembertian", and the Dirac equation factors the second-order Klein-Gordon operator into two first-order pieces $(i\gamma \cdot \partial -m)(i\gamma \cdot \partial +m)=1\cdot (-\square -m^2)$. This is exactly the same factor-the-quadratic trick that appears again in 03.09.02 (Clifford algebra) at the level of pure algebra, with no derivatives. The central insight is that the Dirac field $\psi$ identifies a relativistic spin-$1/2$ wavefunction with the Klein-Gordon scalar field's "square root": each spinor component satisfies the same wave equation as a free Klein-Gordon scalar, but the four components together encode spin and the particle/antiparticle distinction that scalar fields lack. The bridge is that the gamma-matrix algebra, the Dirac equation, and the Dirac adjoint $\bar{\psi }={\psi }^{†}{\gamma }^0$ together produce the Lorentz-covariant bilinears (scalar, vector, tensor, pseudoscalar, axial vector) used by every QFT interaction Lagrangian. Putting these together, the Dirac field generalises the Klein-Gordon scalar to spin $1/2$, is dual to its own antiparticle via charge conjugation, and identifies the four-component column with the direct sum $(1/2,0)\oplus (0,1/2)$ of two Weyl-spinor representations of $\mathrm{SL}(2,\mathbb{C})$.

Exercises Intermediate+

Advanced results Master

Theorem (Lorentz covariance of the Dirac equation; Dirac 1928, ^{[Dirac 1928]}). Let $\Lambda \in {\mathrm{SO}}^{\uparrow }(1,3)$ be a proper orthochronous Lorentz transformation with $\mathrm{SL}(2,\mathbb{C})$-lift $A\in \mathrm{SL}(2,\mathbb{C})$, and let $S(\Lambda )\in \mathrm{GL}(4,\mathbb{C})$ be the spinor representation determined by $$ S(\Lambda)^{-1} \gamma^\mu S(\Lambda) = \Lambda^\mu_\nu \gamma^\nu. $$ Then $\psi (x)\to {\psi }^{\prime }(x):=S(\Lambda )\psi ({\Lambda }^{-1}x)$ maps solutions of the Dirac equation to solutions: $$ (i \gamma^\mu \partial_\mu - m) \psi(x) = 0 \quad \Longrightarrow \quad (i \gamma^\mu \partial_\mu - m) \psi'(x) = 0. $$ The double-cover map $\mathrm{SL}(2,\mathbb{C})\to {\mathrm{SO}}^{\uparrow }(1,3)$ realises $S$ as a genuine (not merely projective) representation of the universal cover.

The intertwining identity is the algebraic content of "the gamma matrices transform as a four-vector". Its existence (and uniqueness up to overall sign) is the Pauli fundamental theorem on Clifford-algebra automorphisms ^{[Pauli Handbuch 1933]}. Lorentz invariance of the Dirac equation is then a one-line consequence: ${\gamma }^{\mu }{\partial }_{\mu }{\psi }^{\prime }(x)=S(\Lambda )\cdot {\Lambda }_{\nu }^{\mu }{\gamma }^{\nu }\cdot {\Lambda }_{\mu }^{\lambda }{\partial }_{\lambda }\psi ({\Lambda }^{-1}x)=S(\Lambda ){\gamma }^{\lambda }{\partial }_{\lambda }\psi ({\Lambda }^{-1}x)$ using ${\Lambda }_{\nu }^{\mu }{\Lambda }_{\mu }^{\lambda }={\delta }_{\nu }^{\lambda }$.

Theorem (representations of the gamma matrices; Pauli 1933, ^{[Pauli Handbuch 1933]}). Three explicit $4\times 4$ realisations of the Clifford algebra $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }1$ are in common use:

• Dirac (standard) representation: ${\gamma }^0=\mathrm{diag}(I,-I)$, ${\gamma }^i=\left(\begin{array}{cc}0& {\sigma }^i\\ -{\sigma }^i& 0\end{array}\right)$. Convenient for the non-relativistic limit; the upper two components dominate for positive-energy slow particles.
• Weyl (chiral) representation: ${\gamma }^0=\left(\begin{array}{cc}0& I\\ I& 0\end{array}\right)$, ${\gamma }^i=\left(\begin{array}{cc}0& {\sigma }^i\\ -{\sigma }^i& 0\end{array}\right)$. Diagonalises ${\gamma }^5$, exposing the left/right Weyl decomposition.
• Majorana representation: all ${\gamma }^{\mu }$ purely imaginary; allows the reality condition ${\psi }^{\ast }=\psi$ for charge-self-conjugate (Majorana) particles.

All three are related by similarity transformations ${\gamma }_{\mathrm{B}}^{\mu }=U{\gamma }_{\mathrm{A}}^{\mu }{U}^{-1}$ for an explicit unitary $U$, since the irreducible Clifford representation in dimension $4$ is unique up to similarity.

The choice of representation is a basis choice for the four-component column, but it commits the user to specific conventions for the chirality projector, the parity operator, and the charge-conjugation operator. The Weyl representation is preferred when chirality is the central object (Standard Model fermions, neutrino physics); the Dirac representation is preferred for non-relativistic limits and atomic physics; the Majorana representation is preferred when the field is real or self-conjugate.

Theorem (Weyl-spinor decomposition; Weyl 1929, ^{[Weyl 1929]}). Define the chirality element ${\gamma }^5:=i{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3$. It satisfies $({\gamma }^5{)}^2=1$ and $\{{\gamma }^5,{\gamma }^{\mu }\}=0$ for every $\mu$. The chirality projectors $$ P_L := \tfrac{1}{2}(1 - \gamma^5), \qquad P_R := \tfrac{1}{2}(1 + \gamma^5), $$ satisfy $P_L+P_R=1$, $P_L{P}_R=0$, $P_{L,R}^2=P_{L,R}$, and split every Dirac spinor as $\psi ={\psi }_L+{\psi }_R$ with ${\psi }_L:=P_L\psi$, ${\psi }_R:=P_R\psi$. In the Weyl basis the two two-component pieces $({\psi }_L,{\psi }_R)$ transform under the $(1/2,0)\oplus (0,1/2)$ representation of $\mathrm{SL}(2,\mathbb{C})$ as a pair of independent Weyl spinors. The Dirac equation in chiral form reads $$ i \bar\sigma^\mu \partial_\mu \psi_L = m \psi_R, \qquad i \sigma^\mu \partial_\mu \psi_R = m \psi_L, $$ with ${\sigma }^{\mu }:=(1,{\sigma }^i)$ and ${\bar{\sigma }}^{\mu }:=(1,-{\sigma }^i)$. At $m=0$ the two Weyl pieces decouple, giving two independent massless Weyl equations.

The chiral decomposition is the Standard Model's organising principle for weak interactions: the $W^{\pm }$ bosons couple only to the left-handed pieces ${\psi }_L$ of all known fermions. The mass term $m\bar{\psi }\psi =m({\psi }_L^{†}{\psi }_R+{\psi }_R^{†}{\psi }_L)$ mixes the two chiralities, so a mass eigenstate is not a chirality eigenstate; massless particles can be chirality eigenstates.

Theorem (CPT theorem for the Dirac field; Lüders 1954, Pauli 1955, ^{[Lüders 1957]}). The free Dirac field $\psi (x)$, second-quantised on the fermionic Fock space 12.13.02 with the canonical mode expansion above, is invariant under the combined action of charge conjugation $C$, parity $P$, and time reversal $T$: $$ (CPT), \psi(x), (CPT)^{-1} = i \gamma^5 \psi^(-x), \qquad \langle \Omega, \phi_1(x_1) \cdots \phi_n(x_n) \Omega \rangle = \langle \Omega, \phi_n(-x_n)^ \cdots \phi_1(-x_1)^* \Omega \rangle, $$ where ${\varphi }_i\in \{\psi ,\bar{\psi }\}$. CPT is the only discrete-symmetry combination guaranteed to be a symmetry of every Lorentz-invariant local quantum field theory with positive-energy spectrum (Lüders-Pauli theorem).

CPT is the deepest discrete symmetry of relativistic QFT. Its individual factors are not preserved by the Standard Model — parity $P$ is maximally violated by the weak interaction (Wu 1957), charge conjugation $C$ alone is also violated, and the combination $CP$ is violated by the CKM matrix in the quark sector (Cronin-Fitch 1964) — but CPT is exact. Empirically the mass equality of an electron and a positron to one part in $10^8$ is one of the sharpest CPT tests.

Theorem (Spin-statistics for the Dirac field; Pauli 1940, ^{[Pauli 1940 spin-statistics]}). The Dirac field $\psi (x)$, if quantised on a Hilbert space carrying a positive-energy representation of the Poincaré group and equipped with local commutation/anticommutation at spacelike separations, must satisfy anticommutators: $$ {\psi_a(x), \psi_b^\dagger(y)} = 0, \qquad x - y \text{ spacelike}, a, b \in {1, 2, 3, 4}. $$ Quantising with commutators forces either negative norms in the state space or violation of positivity of energy. The Dirac field is therefore a fermionic field, with creation and annihilation operators satisfying canonical anticommutation rather than commutation.

Spin-statistics is the structural reason a Dirac field lives on the antisymmetric Fock space ${\mathcal{F}}_a(\mathcal{H})$ of 12.13.02, and not on the symmetric Fock space ${\mathcal{F}}_s(\mathcal{H})$ of 08.10.01. The Pauli exclusion principle — at most one electron per quantum state — is the multi-particle expression of this antisymmetric structure: $(a_{s,p}^{†}{)}^2=0$ for every $(s,p)$.

Theorem (Dirac field as Wightman field; Streater-Wightman 1964, ^{[Streater-Wightman 1964]}). The free Dirac field $\psi (x)$ of mass $m>0$, second-quantised on ${\mathcal{F}}_a({\mathcal{H}}_m^{1/2})$ over the positive-energy spin-$1/2$ one-particle Hilbert space ${\mathcal{H}}_m^{1/2}=L^2(X_m,d{\lambda }_m)\otimes {\mathbb{C}}^2$, satisfies the Wightman axioms 08.10.07 for a relativistic quantum field. It is Poincaré-covariant under the natural Fock-space representation, the vacuum is the unique Poincaré-invariant cyclic vector, and $\{{\psi }_a(x),{\bar{\psi }}_b(y)\}=0$ for $x-y$ spacelike.

This is the bridge from the Lagrangian / canonical setup to the axiomatic / structural treatment of the Dirac field. Reconstruction theorems then go the other way: a Wightman field whose two-point function matches the Dirac propagator reconstructs the free Dirac field on ${\mathcal{F}}_a({\mathcal{H}}_m^{1/2})$, uniquely up to unitary equivalence.

Synthesis. The Dirac field $\psi$ and its adjoint $\bar{\psi }$ together package the relativistic spin-$1/2$ one-particle data into a single four-component spinor and its dual row, and this is exactly the foundational reason every QED computation that follows can be done. The central insight is that the gamma matrices satisfy the Clifford relation $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }1$ — the algebraic content of taking the square root of the d'Alembertian — and the Dirac equation $(i{\gamma }^{\mu }{\partial }_{\mu }-m)\psi =0$ factors the second-order Klein-Gordon operator into a pair of first-order operators acting on a spinor. Putting these together, the Dirac field generalises the Klein-Gordon scalar to spin $1/2$, identifies the four-component column with the $(1/2,0)\oplus (0,1/2)$ representation of $\mathrm{SL}(2,\mathbb{C})$ via the Weyl decomposition $\psi =({\psi }_L,{\psi }_R)$, and is dual to its own antiparticle via charge conjugation. The bridge between the classical wave equation and the second-quantised operator field is the fermionic mode expansion $\psi (x)={\sum }_{s,p}(au{e}^{-ipx}+b^{†}v{e}^{+ipx})/\sqrt{2E_p}$, with CAR on the operators $a,b$ packaging the spin-statistics constraint and the Pauli exclusion principle.

The five gamma-bilinears — scalar $\bar{\psi }\psi$, vector $\bar{\psi }{\gamma }^{\mu }\psi$, tensor $\bar{\psi }{\sigma }^{\mu \nu }\psi$, pseudoscalar $\bar{\psi }{\gamma }^5\psi$, axial vector $\bar{\psi }{\gamma }^{\mu }{\gamma }^5\psi$ — span the 16-dimensional space of $4\times 4$ bilinears via the Fierz identities, and identify the conserved Noether currents with explicit operator-side expressions: the vector bilinear $\bar{\psi }{\gamma }^{\mu }\psi$ is the $\mathrm{U}(1)$ Noether current of $\psi \to e^{i\alpha }\psi$, conserved on shell. This recipe generalises in two directions. On the algebraic side, the Clifford-algebra construction extends to higher-dimensional Minkowski space and to curved spacetime, producing the Dirac operator $D={\gamma }^{\mu }{\nabla }_{\mu }$ on spin-manifolds 03.09.08 that is dual to the Atiyah-Singer index theorem in topology. On the operator side, the trace identities $\mathrm{tr}({\gamma }^{\mu }{\gamma }^{\nu })=4{\eta }^{\mu \nu }$ and $\mathrm{tr}({\gamma }^{\mu }{\gamma }^{\nu }{\gamma }^{\rho }{\gamma }^{\sigma })=4({\eta }^{\mu \nu }{\eta }^{\rho \sigma }-{\eta }^{\mu \rho }{\eta }^{\nu \sigma }+{\eta }^{\mu \sigma }{\eta }^{\nu \rho })$ are the workhorses of perturbative QED, with every spin-summed cross-section reducing to a polynomial in these traces. The bridge is that the algebraic Clifford structure, the operator-side Fock representation, and the discrete-symmetry structure (CPT, spin-statistics) all encode the same content: the Dirac field is the unique relativistic spin-$1/2$ field consistent with locality, positivity of energy, and Lorentz covariance.

Full proof set Master

Proposition (computation of $\mathrm{tr}({\gamma }^{\mu }{\gamma }^{\nu })=4{\eta }^{\mu \nu }$).

Proof. Write ${\gamma }^{\mu }{\gamma }^{\nu }=\frac{1}{2}\{{\gamma }^{\mu },{\gamma }^{\nu }\}+\frac{1}{2}[{\gamma }^{\mu },{\gamma }^{\nu }]$. The first term equals ${\eta }^{\mu \nu }1$ by the Clifford relation. The second term, the commutator, has trace zero by cyclicity: $\mathrm{tr}[{\gamma }^{\mu },{\gamma }^{\nu }]=\mathrm{tr}({\gamma }^{\mu }{\gamma }^{\nu })-\mathrm{tr}({\gamma }^{\nu }{\gamma }^{\mu })=0$. So $\mathrm{tr}({\gamma }^{\mu }{\gamma }^{\nu })={\eta }^{\mu \nu }\mathrm{tr}(1)=4{\eta }^{\mu \nu }$, using $\mathrm{tr}(1)=4$ in dimension $4$. $\square$

Proposition (vanishing of odd traces). For any odd integer $n$, $\mathrm{tr}({\gamma }^{{\mu }_1}{\gamma }^{{\mu }_2}\cdots {\gamma }^{{\mu }_n})=0$.

Proof. Insert the identity $1=({\gamma }^5{)}^2$ at the left of the product: $$ \mathrm{tr}(\gamma^{\mu_1} \cdots \gamma^{\mu_n}) = \mathrm{tr}((\gamma^5)^2 \gamma^{\mu_1} \cdots \gamma^{\mu_n}). $$ Anticommute one ${\gamma }^5$ through to the right end. Each pass picks up a sign $(-1)$ from $\{{\gamma }^5,{\gamma }^{\mu }\}=0$, so $n$ passes contribute $(-1{)}^n=-1$ when $n$ is odd: $$ \mathrm{tr}((\gamma^5)^2 \gamma^{\mu_1} \cdots \gamma^{\mu_n}) = - \mathrm{tr}(\gamma^5 \gamma^{\mu_1} \cdots \gamma^{\mu_n} \gamma^5). $$ By cyclicity of the trace, $\mathrm{tr}({\gamma }^5{\gamma }^{{\mu }_1}\cdots {\gamma }^{{\mu }_n}{\gamma }^5)=\mathrm{tr}({\gamma }^5{\gamma }^5{\gamma }^{{\mu }_1}\cdots {\gamma }^{{\mu }_n})=\mathrm{tr}({\gamma }^{{\mu }_1}\cdots {\gamma }^{{\mu }_n})$. So the trace equals its own negative, hence equals zero. $\square$

Proposition (computation of $\mathrm{tr}({\gamma }^{\mu }{\gamma }^{\nu }{\gamma }^{\rho }{\gamma }^{\sigma })$). The four-gamma trace is $$ \mathrm{tr}(\gamma^\mu \gamma^\nu \gamma^\rho \gamma^\sigma) = 4 \bigl( \eta^{\mu\nu} \eta^{\rho\sigma} - \eta^{\mu\rho} \eta^{\nu\sigma} + \eta^{\mu\sigma} \eta^{\nu\rho} \bigr). $$

Proof. Move the leftmost gamma to the right using the Clifford relation ${\gamma }^{\mu }{\gamma }^{\nu }=2{\eta }^{\mu \nu }1-{\gamma }^{\nu }{\gamma }^{\mu }$: $$ \gamma^\mu \gamma^\nu \gamma^\rho \gamma^\sigma = 2 \eta^{\mu\nu} \gamma^\rho \gamma^\sigma - \gamma^\nu \gamma^\mu \gamma^\rho \gamma^\sigma. $$ Move again past ${\gamma }^{\rho }$: $$

• \gamma^\nu \gamma^\mu \gamma^\rho \gamma^\sigma = -2 \eta^{\mu\rho} \gamma^\nu \gamma^\sigma + \gamma^\nu \gamma^\rho \gamma^\mu \gamma^\sigma.Move past $\gamma^\sigma$: \gamma^\nu \gamma^\rho \gamma^\mu \gamma^\sigma = 2 \eta^{\mu\sigma} \gamma^\nu \gamma^\rho - \gamma^\nu \gamma^\rho \gamma^\sigma \gamma^\mu.Take the trace of the assembled equation. Use cyclicity $\mathrm{tr}(\gamma^\nu \gamma^\rho \gamma^\sigma \gamma^\mu) = \mathrm{tr}(\gamma^\mu \gamma^\nu \gamma^\rho \gamma^\sigma) =: T$, and the two-gamma trace formula: T = 2 \eta^{\mu\nu} \cdot 4 \eta^{\rho\sigma} - 2 \eta^{\mu\rho} \cdot 4 \eta^{\nu\sigma} + 2 \eta^{\mu\sigma} \cdot 4 \eta^{\nu\rho} - T. $$ Adding $T$ to both sides and dividing by $2$ gives $T=4({\eta }^{\mu \nu }{\eta }^{\rho \sigma }-{\eta }^{\mu \rho }{\eta }^{\nu \sigma }+{\eta }^{\mu \sigma }{\eta }^{\nu \rho })$. $\square$

Proposition (Klein-Gordon factorisation). Every component of a Dirac solution satisfies the Klein-Gordon equation.

Proof. Given in the Intermediate-tier Key-theorem section: apply $(i{\gamma }^{\nu }{\partial }_{\nu }+m)$ to $(i{\gamma }^{\mu }{\partial }_{\mu }-m)\psi =0$, expand the product, symmetrise the gamma indices using ${\partial }_{\mu }{\partial }_{\nu }={\partial }_{\nu }{\partial }_{\mu }$, and apply the Clifford relation to obtain $(\square +m^2)\psi =0$ component-wise. $\square$

Proposition (Lorentz scalar nature of $\bar{\psi }\psi$). The bilinear $\bar{\psi }\psi$ transforms as a Lorentz scalar under $\psi (x)\to S(\Lambda )\psi ({\Lambda }^{-1}x)$.

Proof. Given in Exercise 8: derive the dual-transformation identity $S(\Lambda {)}^{†}{\gamma }^0={\gamma }^{0}S(\Lambda {)}^{-1}$ from the intertwining identity and Hermiticity $({\gamma }^0{)}^{†}={\gamma }^0$. Then $(\bar{\psi }\psi )(x)\to (\psi ({\Lambda }^{-1}x){)}^{†}S(\Lambda {)}^{†}{\gamma }^{0}S(\Lambda )\psi ({\Lambda }^{-1}x)=(\psi ({\Lambda }^{-1}x){)}^{†}{\gamma }^{0}S(\Lambda {)}^{-1}S(\Lambda )\psi ({\Lambda }^{-1}x)=\bar{\psi }({\Lambda }^{-1}x)\psi ({\Lambda }^{-1}x)$. The bilinear evaluates the scalar at the Lorentz-transformed argument; no extra spinor factors remain. $\square$

Proposition (chirality projectors are orthogonal idempotents). Define $P_L:=(1-{\gamma }^5)/2$, $P_R:=(1+{\gamma }^5)/2$ with ${\gamma }^5:=i{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3$. Then $P_L+P_R=1$, $P_L{P}_R=P_R{P}_L=0$, $P_{L,R}^2=P_{L,R}$, and $({\gamma }^5{)}^2=1$.

Proof. Compute $({\gamma }^5{)}^2=(i{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3)(i{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3)=-{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3$. Move the second ${\gamma }^0$ left past three gammas (sign $(-1{)}^3=-1$) to pair with the first ${\gamma }^0$: $({\gamma }^0{)}^2=1$, leaving ${\gamma }^1{\gamma }^2{\gamma }^3{\gamma }^1{\gamma }^2{\gamma }^3$. Repeat for ${\gamma }^1$ (move past two gammas, sign $(-1{)}^2=+1$): leaves $({\gamma }^1{)}^2{\gamma }^2{\gamma }^3{\gamma }^2{\gamma }^3=-{\gamma }^2{\gamma }^3{\gamma }^2{\gamma }^3$. Move the second ${\gamma }^2$ left past ${\gamma }^3$ (sign $(-1{)}^1=-1$), pair $({\gamma }^2{)}^2=-1$, leaves $-\cdot (-1)\cdot (-1)\cdot ({\gamma }^3{)}^2=-(-1)=1$. Tracking signs: $-(-1)(-1)(-1)(-1)(-1)(-1)=-(-1{)}^6=-1\cdot 1=-1$... re-examining the bookkeeping more carefully with the standard identity: the cleanest derivation is $({\gamma }^5{)}^2=-({\gamma }^0{)}^2({\gamma }^1{)}^2({\gamma }^2{)}^2({\gamma }^3{)}^2\cdot (\text{sign from reordering})$. The reordering of ${\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3{\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3$ to $({\gamma }^0{)}^2({\gamma }^1{)}^2({\gamma }^2{)}^2({\gamma }^3{)}^2$ requires bringing the second ${\gamma }^0$ left through three gammas, the second ${\gamma }^1$ through two, the second ${\gamma }^2$ through one, total $3+2+1=6$ anticommutations, sign $(-1{)}^6=+1$. Squares: $({\gamma }^0{)}^2=+1$, $({\gamma }^i{)}^2=-1$ for $i\in \{1,2,3\}$, product $1\cdot (-1{)}^3=-1$. Overall $({\gamma }^5{)}^2=i^2\cdot (+1)\cdot (-1)=-1\cdot -1=1$.

Anticommutation $\{{\gamma }^5,{\gamma }^{\mu }\}=0$: each fixed ${\gamma }^{\mu }$ appears once in the product ${\gamma }^0{\gamma }^1{\gamma }^2{\gamma }^3$ and anticommutes with the three others, so moving ${\gamma }^{\mu }$ past all four gives sign $(-1{)}^3=-1$ from the three it does not pair with, times $+1$ from the one it equals — net $-1$, i.e. anticommutation. Idempotency: $P_L^2=(1-{\gamma }^5{)}^2/4=(1-2{\gamma }^5+({\gamma }^5{)}^2)/4=(1-2{\gamma }^5+1)/4=(2-2{\gamma }^5)/4=P_L$. Similarly $P_R^2=P_R$ and $P_L{P}_R=(1-{\gamma }^5)(1+{\gamma }^5)/4=(1-({\gamma }^5{)}^2)/4=0$. Completeness $P_L+P_R=1$ is immediate. $\square$

Proposition (mode expansion satisfies the Dirac equation). The second-quantised operator-valued distribution $$ \psi(x) = \int \frac{d^3 p}{(2\pi)^3 \sqrt{2 E_p}} \sum_{s=1}^{2} \Bigl[ a_{s,p} u_s(p) e^{-i p \cdot x} + b_{s,p}^\dagger v_s(p) e^{+i p \cdot x} \Bigr] $$ satisfies $(i{\gamma }^{\mu }{\partial }_{\mu }-m)\psi (x)=0$ in the operator-valued tempered-distribution sense, given the spinor identities $({\gamma }^{\mu }{p}_{\mu }-m)u_s(p)=0$ and $({\gamma }^{\mu }{p}_{\mu }+m)v_s(p)=0$.

Proof. Apply $(i{\gamma }^{\mu }{\partial }_{\mu }-m)$ to each plane-wave term. For $u_s(p)e^{-ip\cdot x}$, the derivative pulls down ${\partial }_{\mu }\to -i{p}_{\mu }$, so $i{\gamma }^{\mu }{\partial }_{\mu }(u_s{e}^{-ipx})=i{\gamma }^{\mu }(-i{p}_{\mu })u_s{e}^{-ipx}={\gamma }^{\mu }{p}_{\mu }{u}_s{e}^{-ipx}$. Combined with the $-m$ term: $({\gamma }^{\mu }{p}_{\mu }-m)u_s(p)e^{-ipx}=0$ by the on-shell spinor identity.

For $v_s(p)e^{+ip\cdot x}$, the derivative pulls down ${\partial }_{\mu }\to +i{p}_{\mu }$, so $i{\gamma }^{\mu }{\partial }_{\mu }(v_s{e}^{+ipx})=-{\gamma }^{\mu }{p}_{\mu }{v}_s{e}^{+ipx}$. Combined: $(-{\gamma }^{\mu }{p}_{\mu }-m)v_s(p)e^{+ipx}=-({\gamma }^{\mu }{p}_{\mu }+m)v_s(p)e^{+ipx}=0$ by the on-shell identity for $v_s$. Each plane-wave term vanishes, and the integral over momenta of an identically zero integrand is zero. The result holds as an operator-valued distribution after smearing against Schwartz test functions on ${\mathbb{R}}^{1,3}$. $\square$

Connections Master

• Clifford algebra 03.09.02. The gamma matrices ${\gamma }^{\mu }$ satisfying $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }1$ are the generators of the Clifford algebra $\mathrm{Cl}(1,3)$ of Minkowski space with mostly-minus signature. The Dirac spinor space ${\mathbb{C}}^4$ is the (unique up to similarity) irreducible representation of the complexified Clifford algebra in dimension $4$. The whole gamma-matrix technology — anticommutators, traces, chirality element ${\gamma }^5$, bilinears — is the operator-side incarnation of pure Clifford-algebraic identities.

• Spin group 03.09.03. The spinor representation $S:\mathrm{SL}(2,\mathbb{C})\to \mathrm{GL}(4,\mathbb{C})$ that conjugates gamma matrices into Lorentz-transformed gammas is the double-cover lift of the vector representation ${\mathrm{SO}}^{\uparrow }(1,3)\to \mathrm{GL}(4,\mathbb{R})$. The double cover ${\mathrm{Spin}}^{\uparrow }(1,3)=\mathrm{SL}(2,\mathbb{C})$ is exactly the input that turns a projective representation of the Poincaré group on a four-component column into an honest unitary representation — the algebraic mechanism behind half-integer spin.

• Dirac equation and relativistic spin 12.11.01. This unit's QM-side prequel develops the Dirac equation as a relativistic wavefunction equation, with $\psi$ interpreted as a one-particle wavefunction in the Hilbert-space sense. The present unit takes that same $\psi$ and promotes it to an operator-valued tempered distribution on the fermionic Fock space, replacing the hole-theoretic reinterpretation of negative-energy solutions with the cleaner antiparticle-creation language of second quantisation.

• Fermionic Fock space and Pauli exclusion 12.13.02. The mode expansion of the Dirac field uses operators $a_{s,p},b_{s,p}^{†}$ satisfying the canonical anticommutation relations on the antisymmetric Fock space ${\mathcal{F}}_a({\mathcal{H}}_{+})$ over the positive-energy spin-$1/2$ one-particle Hilbert space. The spin-statistics theorem forces this choice: Lorentz invariance plus positive-energy spectrum plus locality is consistent with anticommutators but not commutators for a half-integer-spin field.

• Bosonic Fock space and second quantisation 08.10.01. The bosonic Fock-space construction for spin-$0$ fields (the Klein-Gordon scalar) is the parallel construction for the Dirac field replaced by its fermionic analogue. The two together cover the Wigner classification of free relativistic fields by mass and spin, with bosonic Fock space for integer spin and fermionic Fock space for half-integer spin.

• Wightman axioms (W1-W7) 08.10.07. The free Dirac field on ${\mathcal{F}}_a({\mathcal{H}}_m^{1/2})$ satisfies the Wightman axioms: Poincaré covariance, spectrum condition, vacuum cyclicity, operator-valued tempered-distribution structure, and local anticommutativity at spacelike separations. The Wightman framework is the structural setting in which the Lüders-Pauli CPT theorem and the Pauli spin-statistics theorem are stated as theorems rather than axioms.

Historical & philosophical context Master

Paul Dirac discovered the equation that bears his name in early 1928, publishing in Proceedings of the Royal Society A in two parts (vol. 117, 610; vol. 118, 351) ^{[Dirac 1928]}. The motivation was a long-standing dissatisfaction with the Klein-Gordon equation $(\square +m^2)\varphi =0$ as a single-particle wave equation: the second-order time derivative gave a probability current that could become negative, and the negative-energy solutions had no clear interpretation. Dirac sought a first-order equation. The discovery that the algebraic price of first-orderness was a four-component wavefunction with $4\times 4$ matrix coefficients satisfying the anticommutation relation $\{{\gamma }^{\mu },{\gamma }^{\nu }\}=2{\eta }^{\mu \nu }$ — what the mathematicians' Clifford had constructed abstractly in 1878 — was unexpected and immediate. Spin emerged as a structural consequence of the equation, not a postulate; the gyromagnetic ratio $g=2$ for the electron came out automatically in the non-relativistic limit (Foldy-Wouthuysen 1950). The negative-energy solutions were initially troubling, then reinterpreted via Dirac's hole theory (1930, Proc. Roy. Soc. A 126), and finally cleanly absorbed into second quantisation via the antiparticle. Carl Anderson detected the positron in cloud-chamber tracks in 1932 (Science 76, 238; Phys. Rev. 43, 491), one of the cleanest theory-precedes-experiment confirmations in twentieth-century physics.

Hermann Weyl's Elektron und Gravitation I (Z. Phys. 56, 330, 1929) ^{[Weyl 1929]} developed the two-component spinor decomposition and introduced what we now call the chirality element ${\gamma }^5$; the parallel two-component / four-component pictures coexisted thereafter. Ettore Majorana's 1937 Nuovo Cim. paper ^{[Majorana 1937]} introduced the Majorana representation of the gamma matrices and the real (charge-self-conjugate) spinor, opening the question — still open empirically in 2026 — whether the neutrino is a Majorana or a Dirac fermion. Wolfgang Pauli's 1933 Handbuch der Physik article ^{[Pauli Handbuch 1933]} is the canonical pre-second-quantisation exposition, in which the Dirac adjoint $\bar{\psi }={\psi }^{†}{\gamma }^0$ is established, the conserved current $\bar{\psi }{\gamma }^{\mu }\psi$ is derived, and the gamma-matrix trace identities are tabulated. The second-quantised Dirac field on the fermionic Fock space — with creation operators $a^{†}$ for electrons and $b^{†}$ for positrons satisfying canonical anticommutation — was developed in the 1930s by Heisenberg, Pauli, Fock, and Furry-Oppenheimer (Phys. Rev. 45, 245, 1934), and reached textbook form in Bjorken-Drell Relativistic Quantum Mechanics (1964) ^{[Bjorken-Drell 1964]} and Peskin-Schroeder Ch. 3 (1995) ^{[Peskin-Schroeder 1995]}.

Two structural theorems closed the foundations. Gerhart Lüders (Dan. Mat. Fys. Medd. 28, no. 5, 1954; Ann. Phys. 2, 1, 1957) ^{[Lüders 1957]} and Wolfgang Pauli (in the Bohr Festschrift, 1955) proved the CPT theorem: the combined operation of charge conjugation, parity, and time reversal is a symmetry of every Lorentz-invariant local quantum field theory with positive-energy spectrum — the Dirac field included. Pauli's 1940 spin-statistics paper (Phys. Rev. 58, 716) ^{[Pauli 1940 spin-statistics]}, building on Fierz 1939, established that half-integer-spin fields must be quantised with anticommutators. The Streater-Wightman 1964 monograph PCT, Spin and Statistics, and All That ^{[Streater-Wightman 1964]} packaged both results into the Wightman-axiomatic framework that anchors the modern mathematical treatment of relativistic QFT. The Dirac field, half a century after its discovery, became the canonical example verifying the axioms.

Bibliography Master

@article{Dirac1928QuantumElectron,
  author  = {Dirac, P. A. M.},
  title   = {The Quantum Theory of the Electron},
  journal = {Proc. Roy. Soc. A},
  volume  = {117},
  year    = {1928},
  pages   = {610--624}
}

@article{Dirac1928QuantumElectronII,
  author  = {Dirac, P. A. M.},
  title   = {The Quantum Theory of the Electron, Part II},
  journal = {Proc. Roy. Soc. A},
  volume  = {118},
  year    = {1928},
  pages   = {351--361}
}

@article{Weyl1929ElektronGravitation,
  author  = {Weyl, Hermann},
  title   = {Elektron und Gravitation I},
  journal = {Z. Phys.},
  volume  = {56},
  year    = {1929},
  pages   = {330--352}
}

@article{Majorana1937,
  author  = {Majorana, Ettore},
  title   = {Teoria simmetrica dell'elettrone e del positrone},
  journal = {Nuovo Cimento},
  volume  = {14},
  year    = {1937},
  pages   = {171--184}
}

@incollection{Pauli1933Handbuch,
  author    = {Pauli, Wolfgang},
  title     = {Die allgemeinen Prinzipien der Wellenmechanik},
  booktitle = {Handbuch der Physik, Bd. 24, Teil 1},
  publisher = {Springer},
  year      = {1933}
}

@article{Pauli1940SpinStatistics,
  author  = {Pauli, Wolfgang},
  title   = {The Connection Between Spin and Statistics},
  journal = {Phys. Rev.},
  volume  = {58},
  year    = {1940},
  pages   = {716--722}
}

@article{Anderson1933Positron,
  author  = {Anderson, Carl D.},
  title   = {The Positive Electron},
  journal = {Phys. Rev.},
  volume  = {43},
  year    = {1933},
  pages   = {491--494}
}

@article{Lueders1957CPT,
  author  = {L{\"u}ders, Gerhart},
  title   = {Proof of the {TCP} Theorem},
  journal = {Ann. Phys. (NY)},
  volume  = {2},
  year    = {1957},
  pages   = {1--15}
}

@article{Lueders1954CPT,
  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},
  number  = {5},
  year    = {1954}
}

@article{Fierz1939,
  author  = {Fierz, Markus},
  title   = {{\"U}ber die relativistische Theorie kr{\"a}ftefreier Teilchen mit beliebigem Spin},
  journal = {Helv. Phys. Acta},
  volume  = {12},
  year    = {1939},
  pages   = {3--37}
}

@book{BjorkenDrell1964,
  author    = {Bjorken, James D. and Drell, Sidney D.},
  title     = {Relativistic Quantum Mechanics},
  publisher = {McGraw-Hill},
  year      = {1964}
}

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

@book{ItzyksonZuber1980,
  author    = {Itzykson, Claude and Zuber, Jean-Bernard},
  title     = {Quantum Field Theory},
  publisher = {McGraw-Hill},
  year      = {1980}
}

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

@book{Weinberg1995QFTI,
  author    = {Weinberg, Steven},
  title     = {The Quantum Theory of Fields, Vol. I: Foundations},
  publisher = {Cambridge University Press},
  year      = {1995}
}

@article{FoldyWouthuysen1950,
  author  = {Foldy, Leslie L. and Wouthuysen, Siegfried A.},
  title   = {On the {D}irac Theory of Spin 1/2 Particles and its Non-Relativistic Limit},
  journal = {Phys. Rev.},
  volume  = {78},
  year    = {1950},
  pages   = {29--36}
}

@article{Wigner1939UnitaryReps,
  author  = {Wigner, Eugene P.},
  title   = {On Unitary Representations of the Inhomogeneous {L}orentz Group},
  journal = {Ann. of Math.},
  volume  = {40},
  year    = {1939},
  pages   = {149--204}
}