Why Introduce the Dirac Equation?

The Dirac equation combines local first-order evolution with relativistic dispersion by introducing a multicomponent amplitude. Its positive density and spin structure make it useful for relativistic fermion wave mechanics. They do not make it a complete theory of interacting particles. This page explains the physical choice behind the construction; the algebraic derivation belongs to Gamma Matrices.

Required background. The Energy–Momentum Relation fixes the dispersion; the Klein–Gordon Equation provides the scalar comparison. Helpful background. The Square-Root Hamiltonian shows the cost of retaining only the positive scalar branch.

Local first-order evolution is an additional requirement

The mass shell $E^2=c^2\mathbf p^2+m^2c^4$ does not uniquely select a wave equation or a spin. The scalar Klein–Gordon equation realizes it with a second time derivative. A positive square-root Hamiltonian realizes one branch with unitary first-order time evolution but a nonlocal spatial operator.

Dirac’s construction asks for a finite-component equation that is first order in both time and space, with a Hamiltonian linear in momentum,

$$i\hbar\partial_t\psi=H_D\psi,\qquad H_D=c\boldsymbol\alpha\cdot\mathbf P+\beta mc^2.$$

The coefficients are constant matrices acting on internal components, not on spatial coordinates. To reproduce the mass shell for every momentum, their mixed products must cancel. The Clifford relations derived in the canonical owner achieve $H_D^2=(c^2\mathbf P^2+m^2c^4)I$. In three spatial dimensions the massive irreducible construction needs four complex components.

These requirements explain a useful route to the equation, not a proof that every relativistic particle must be a Dirac particle. Scalar fields remain legitimate, higher-spin fields require different structures, and a second-order equation can also be rewritten as a first-order system by enlarging its Cauchy data. Such a rewriting need not have the Dirac equation’s positive Hilbert metric or spatial first-order form.

A two-component example in one spatial dimension

The essential algebra is visible in a smaller model:

$$H_{1{\rm D}}=c\sigma_x P+mc^2\sigma_z.$$

The two Pauli matrices are Hermitian, square to the identity, and anticommute. Their cross term therefore vanishes on squaring:

$$H_{1{\rm D}}^2=(c^2P^2+m^2c^4)I_2.$$

At a fixed momentum, the eigenvalues are still $\pm E_p$. At rest the two components diagonalize the two mass-energy signs; they are not two spatial directions. This reduced model does not by itself carry the full three-dimensional electron-spin representation. A similar matrix Hamiltonian in a material can use components for sublattice or orbital labels, so matrix size alone does not identify physical spin.

The example also shows why the construction does not discard negative energy: replacing a scalar square root by a local matrix operator retains both eigenvalue branches.

What the spinor equation gains

With the standard self-adjoint free realization, initial data in $D(H_D)$ determine a strong time derivative; arbitrary $L^2$ data still have unitary evolution without necessarily being strongly differentiable. The conserved density is $\psi^\dagger\psi\geq0$ and its spatial current is $c\psi^\dagger\boldsymbol\alpha\psi$. The Covariant Dirac Equation derives conservation and explains how spinors transform under Lorentz transformations. Their four components are not a Lorentz four-vector.

After electromagnetic coupling and a controlled low-energy expansion, the spinor structure gives the Pauli spin interaction and tree-level magnetic moment. Those consequences are derived in Dirac to Pauli. They should not be assumed merely because a model has two or four components.

What remains to be supplied

A positive one-particle density does not make its Hamiltonian bounded below: the free Dirac operator has both energy continua. Projection onto the positive continuum is nonlocal, and backgrounds can couple chosen sectors. Quantization reinterprets the negative-frequency modes through antiparticle creation operators and introduces many-particle states. A wave equation alone does not prove fermionic statistics or describe the quantum electromagnetic field.

Dirac’s 1928 paper developed the relativistic electron equation in connection with atomic spectroscopy and the transformation theory of quantum mechanics. The current derivation organizes that motivation in modern operator language; it should not be mistaken for the entire historical sequence of the antiparticle interpretation.

Exercises

1. Could real scalar constants $a,b$ replace $\sigma_x,\sigma_z$ in $H=caP+bmc^2$ while retaining the massive dispersion for all momenta?

Solution

The coefficients of $P^2$, the constant term, and the linear cross term would require $a^2=b^2=1$ and $ab=0$. These conditions are incompatible. Noncommuting matrices permit anticommutation without either factor vanishing.

2. Does $H_{1{\rm D}}^2=E_p^2I$ prove that every state has positive energy? Check at zero momentum.

Solution

No. At $p=0$, $H_{1{\rm D}}=mc^2\sigma_z$ has eigenvalues $+mc^2$ and $-mc^2$. Squaring loses the branch sign. Positivity of the ordinary spinor norm is independent of this spectral sign.

References

• J. D. Bjorken and S. D. Drell, Relativistic Quantum Mechanics, McGraw–Hill, 1964 — first-order relativistic wave mechanics.
• P. A. M. Dirac, “The Quantum Theory of the Electron,” Proceedings of the Royal Society A 117, 610–624, 1928, doi:10.1098/rspa.1928.0023 — the original electron equation and its motivation.
• B. Thaller, The Dirac Equation, Springer, 1992 — operator and spectral interpretation.