Lesson 3.4 · 3. Electromagnetism

Relativistic Formulation

Electromagnetism was relativistic before relativity was discovered. When Einstein formulated special relativity in 1905, he did not need to modify Maxwell's equations. They were already perfectly consistent with Lorentz invariance. It was Newtonian mechanics that needed fixing. In this lesson, we'll see that when electromagnetism is written in the language of spacetime tensors, Maxwell's four equations collapse into just two, the potentials form a natural 4-vector, and the entire theory fits into a Lagrangian of stunning simplicity.

This is more than an exercise in elegant notation. The relativistic formulation reveals that $\mathbf{E}$ and $\mathbf{B}$ are not separate entities. They are components of a single object, the electromagnetic field strength tensor, mixed together by Lorentz transformations. What one observer calls a purely electric field, another observer moving at a different velocity calls a mixture of electric and magnetic fields. The distinction between electricity and magnetism is frame-dependent.

The Central Theme

Electromagnetism is not merely compatible with special relativity. It demands it. The speed of light $c$ appears naturally in Maxwell's equations, Lorentz invariance is built into their structure, and the theory achieves its most natural form when written in the language of 4-vectors and tensors.

What is a tensor?

A tensor is a generalization of a vector. A scalar (like temperature) is a single number at each point: it is a rank-0 tensor. A vector (like velocity) has components that transform in a specific way when you change reference frames: it is a rank-1 tensor. A rank-2 tensor, like the electromagnetic tensor $F^{\mu\nu}$, is an array of numbers (a matrix) that transforms according to well-defined rules. It encodes more information than a simple vector. Here, it combines all six components of $\mathbf{E}$ and $\mathbf{B}$ into a single object.

What is a 4-vector?

A 4-vector is a vector with four components (one temporal and three spatial) that transforms predictably when you switch from one inertial frame to another via the Lorentz transformations. In classical physics, we use 3-component vectors $(x, y, z)$. In relativity, space and time are intertwined, so we must use 4-component vectors $(ct, x, y, z)$ to correctly describe the physics.

The 4-Potential

In the previous lesson, we introduced the scalar potential $\phi$ and the vector potential $\mathbf{A}$. These four quantities, one scalar and three vector components, combine naturally into a spacetime 4-vector:

The Electromagnetic 4-Potential

$$A^\mu = \left(\frac{\phi}{c}, \mathbf{A}\right) = \left(\frac{\phi}{c}, A_x, A_y, A_z\right)$$

This transforms as a 4-vector under Lorentz transformations, just like the spacetime position $x^\mu = (ct, \mathbf{r})$.

That $(\phi/c, \mathbf{A})$ forms a 4-vector is not a choice. It is forced by the requirement that the theory be Lorentz invariant. Under a Lorentz boost with velocity $v$ along the $x$-axis:

$$\phi' = \gamma(\phi - vA_x), \qquad A_x' = \gamma\left(A_x - \frac{v\phi}{c^2}\right)$$

$A_y$ and $A_z$ are unchanged. The scalar and vector potentials mix under boosts, just as time and space do. This is the first hint that electricity and magnetism are two aspects of the same thing.

The gauge transformation also takes a compact 4-dimensional form:

$$A^\mu \to A^\mu + \partial^\mu \chi$$

where $\partial^\mu = (\partial/\partial(ct), -\nabla)$ and $\chi$ is an arbitrary scalar function. The Lorenz gauge condition becomes:

$$\partial_\mu A^\mu = 0$$

This is manifestly Lorentz invariant, a single equation in spacetime, rather than a relation mixing spatial and temporal derivatives.

The Field Strength Tensor

The electric and magnetic fields are obtained from the potentials by differentiation. In 4-dimensional notation, they combine into a single antisymmetric tensor:

The Electromagnetic Field Tensor

$$F^{\mu\nu} = \partial^\mu A^\nu - \partial^\nu A^\mu$$

Written as a $4 \times 4$ matrix (with $\mu$ labeling rows and $\nu$ labeling columns, and $x^0 = ct$):

$$F^{\mu\nu} = \begin{pmatrix} 0 & -E_x/c & -E_y/c & -E_z/c \\ E_x/c & 0 & -B_z & B_y \\ E_y/c & B_z & 0 & -B_x \\ E_z/c & -B_y & B_x & 0 \end{pmatrix}$$

This is remarkable. The six independent components of the antisymmetric tensor $F^{\mu\nu}$ are precisely the three components of $\mathbf{E}$ and the three components of $\mathbf{B}$. They are unified into a single geometric object that lives in spacetime.

Under a Lorentz transformation $\Lambda$, the tensor transforms as:

$$F'^{\mu\nu} = \Lambda^\mu{}_\alpha \Lambda^\nu{}_\beta F^{\alpha\beta}$$

This mixing of components under boosts is why electric and magnetic fields transform into each other. A pure electric field in one frame becomes a combination of $\mathbf{E}$ and $\mathbf{B}$ in another.

Frame S (charge at rest) + E only, B = 0 boost v Frame S' (charge moving) + E' and B' mixed
A charge at rest produces only E. In a boosted frame, the same charge (now moving) produces both E' and B'. Electric and magnetic fields are frame-dependent projections of the single tensor F.

Maxwell's Equations in Covariant Form

The greatest payoff of the tensor formulation is the compression of Maxwell's four equations into two. We also need the 4-current density:

$$J^\mu = (c\rho, \mathbf{J})$$

which combines the charge density $\rho$ and current density $\mathbf{J}$ into a single 4-vector. Charge conservation becomes the simple 4-divergence condition $\partial_\mu J^\mu = 0$.

Maxwell's Equations: Covariant Form

All four of Maxwell's equations reduce to just two tensor equations:

$$\partial_\mu F^{\mu\nu} = \mu_0 J^\nu \qquad \text{(inhomogeneous equations)}$$ $$\partial_{[\alpha} F_{\beta\gamma]} = 0 \qquad \text{(homogeneous equations)}$$

The first equation encodes Gauss's law ($\nu = 0$) and the Ampere-Maxwell law ($\nu = 1,2,3$). The second encodes the no-monopole law and Faraday's law.

Let's unpack the first equation. Setting $\nu = 0$:

$$\partial_\mu F^{\mu 0} = \mu_0 J^0 = \mu_0 c\rho$$

Writing out the sum: $\partial_1 F^{10} + \partial_2 F^{20} + \partial_3 F^{30} = \mu_0 c\rho$. Reading off the components of $F^{\mu\nu}$, this gives:

$$\frac{\partial(E_x/c)}{\partial x} + \frac{\partial(E_y/c)}{\partial y} + \frac{\partial(E_z/c)}{\partial z} = \mu_0 c\rho$$

which simplifies to $\nabla \cdot \mathbf{E} = \rho/\varepsilon_0$, Gauss's law. Setting $\nu = 1, 2, 3$ recovers the three components of the Ampere-Maxwell law.

The homogeneous equation $\partial_{[\alpha} F_{\beta\gamma]} = 0$ (the antisymmetrized derivative) is actually an identity. It follows automatically from $F_{\mu\nu} = \partial_\mu A_\nu - \partial_\nu A_\mu$ because partial derivatives commute. This is the Bianchi identity, and it encodes both $\nabla \cdot \mathbf{B} = 0$ and Faraday's law.

An equivalent and often more convenient form uses the dual tensor $\tilde{F}^{\mu\nu} = \frac{1}{2}\epsilon^{\mu\nu\alpha\beta}F_{\alpha\beta}$:

$$\partial_\mu \tilde{F}^{\mu\nu} = 0$$

Lorentz Invariants

From $F^{\mu\nu}$, we can construct scalar quantities that are the same in all reference frames: Lorentz invariants. There are exactly two independent ones:

$$F_{\mu\nu}F^{\mu\nu} = 2\left(B^2 - \frac{E^2}{c^2}\right)$$ $$\tilde{F}_{\mu\nu}F^{\mu\nu} = -\frac{4}{c}\mathbf{E} \cdot \mathbf{B}$$

Electromagnetic Invariants

Two quantities are the same in all inertial frames:

1. $B^2 - E^2/c^2$ is Lorentz invariant

2. $\mathbf{E} \cdot \mathbf{B}$ is Lorentz invariant

If $\mathbf{E} \perp \mathbf{B}$ in one frame, they are perpendicular in all frames. If $|\mathbf{E}| > c|\mathbf{B}|$ in one frame, this inequality holds in all frames. These invariants classify electromagnetic fields into types that no Lorentz transformation can change.

For electromagnetic waves, both invariants are zero: $E = cB$ and $\mathbf{E} \perp \mathbf{B}$. These properties are Lorentz invariant. Electromagnetic waves look like electromagnetic waves in every inertial frame.

The Lorentz Force as 4-Force

The Lorentz force on a charged particle, $\mathbf{F} = q(\mathbf{E} + \mathbf{v} \times \mathbf{B})$, takes a compact covariant form. The equation of motion for a particle with charge $q$ and 4-velocity $u^\mu$ is:

$$\frac{dp^\mu}{d\tau} = qF^{\mu\nu}u_\nu$$

where $p^\mu = mu^\mu$ is the 4-momentum and $\tau$ is the proper time. The spatial components give the ordinary Lorentz force, and the time component gives the rate at which the field does work on the particle:

$$\frac{dE}{dt} = q\mathbf{E} \cdot \mathbf{v}$$

Note that the magnetic force does no work (it is always perpendicular to $\mathbf{v}$). Only the electric field can change a particle's energy.

The Electromagnetic Lagrangian

What is a Lagrangian?

A Lagrangian is a single function from which all the equations of motion of a system can be derived via the principle of least action. Instead of stating each law separately, you write one Lagrangian and Maxwell's equations, the Lorentz force, and energy conservation all follow automatically. Think of it as a master recipe from which an entire meal can be prepared.

The entire theory of classical electromagnetism, Maxwell's equations, the Lorentz force, energy-momentum conservation, follows from a single Lagrangian density.

The EM Lagrangian Density

$$\mathcal{L} = -\frac{1}{4\mu_0}F_{\mu\nu}F^{\mu\nu} - J_\mu A^\mu$$

The first term describes the free electromagnetic field. The second describes its coupling to charged matter. All of electromagnetism flows from this one expression via the principle of least action.

Let's see how. The action is $S = \int \mathcal{L} \, d^4x$. Varying with respect to $A_\nu$ and setting $\delta S = 0$ gives the Euler-Lagrange equation:

$$\partial_\mu \frac{\partial \mathcal{L}}{\partial(\partial_\mu A_\nu)} = \frac{\partial \mathcal{L}}{\partial A_\nu}$$

The left side gives $-\frac{1}{\mu_0}\partial_\mu F^{\mu\nu}$ and the right side gives $-J^\nu$, yielding:

$$\partial_\mu F^{\mu\nu} = \mu_0 J^\nu$$

This is precisely the inhomogeneous Maxwell equation! The homogeneous equation is automatic from the definition $F^{\mu\nu} = \partial^\mu A^\nu - \partial^\nu A^\mu$. The entire content of Maxwell's equations emerges from a single variational principle.

L = -1/(4μ₀) F F - J A The entire theory in one line Maxwell's equations Lorentz force energy & momentum gauge invariance Lorentz symmetry U(1) gauge symmetry least action
The electromagnetic Lagrangian density, determined by Lorentz symmetry and U(1) gauge invariance, contains all of electromagnetism. Maxwell's equations, the Lorentz force, energy-momentum conservation, and gauge structure all follow from this single expression.

The Lagrangian $\mathcal{L} = -\frac{1}{4\mu_0}F_{\mu\nu}F^{\mu\nu}$ is essentially the simplest possible Lorentz-invariant, gauge-invariant scalar that can be built from the field strength. Higher-order terms like $(F_{\mu\nu}F^{\mu\nu})^2$ are allowed by symmetry but are suppressed at low energies. They correspond to quantum corrections (the Euler-Heisenberg effective Lagrangian).

The Energy-Momentum Tensor

The energy and momentum of the electromagnetic field are encoded in the stress-energy tensor:

$$T^{\mu\nu} = \frac{1}{\mu_0}\left(F^{\mu\alpha}F^\nu{}_\alpha - \frac{1}{4}\eta^{\mu\nu}F_{\alpha\beta}F^{\alpha\beta}\right)$$

The components have physical meanings:

  • $T^{00} = \frac{1}{2}\left(\varepsilon_0 E^2 + \frac{B^2}{\mu_0}\right)$ is the energy density
  • $T^{0i}/c = \varepsilon_0(\mathbf{E} \times \mathbf{B})_i$ is the momentum density (the Poynting vector divided by $c^2$)
  • $T^{ij}$ is the Maxwell stress tensor (the flow of momentum)

Conservation of energy and momentum is expressed compactly as $\partial_\mu T^{\mu\nu} = -F^{\nu\alpha}J_\alpha$, where the right side is the rate at which the field transfers energy-momentum to matter.

Why Electromagnetism Naturally Lives in Spacetime

It is no accident that electromagnetism is naturally relativistic. Historically, the inconsistency between Maxwell's equations and Galilean relativity was the puzzle that led Einstein to special relativity. Maxwell's equations predict a universal speed of light $c$. They don't say "the speed of light relative to any particular observer" but just "the speed of light." This demanded a new understanding of space and time.

In hindsight, the relativistic structure was always there. The fact that $\mu_0\varepsilon_0 = 1/c^2$ is not a coincidence. It's the signature of a theory built on Lorentz symmetry. The fact that four equations collapse into two tensor equations is not just notation. It reflects the underlying unity of spacetime. And the Lagrangian $-\frac{1}{4\mu_0}F_{\mu\nu}F^{\mu\nu}$ is the unique simplest theory consistent with Lorentz and gauge invariance.

The Structure of Modern Physics

The relativistic formulation of electromagnetism is the template for all fundamental physics. Every force in the Standard Model is described by a Lagrangian of the form $-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}$, with the field strength tensor $F$ generalized to non-Abelian gauge groups. The path from Maxwell to the Standard Model is remarkably direct: write the simplest Lagrangian consistent with Lorentz invariance and gauge symmetry, and the laws of physics follow.

Key Insights

  • The potentials $(\phi/c, \mathbf{A})$ form a Lorentz 4-vector $A^\mu$
  • The fields $\mathbf{E}$ and $\mathbf{B}$ are components of the antisymmetric tensor $F^{\mu\nu} = \partial^\mu A^\nu - \partial^\nu A^\mu$
  • Maxwell's four equations compress to two: $\partial_\mu F^{\mu\nu} = \mu_0 J^\nu$ and $\partial_{[\alpha}F_{\beta\gamma]} = 0$
  • The Lorentz invariants $B^2 - E^2/c^2$ and $\mathbf{E} \cdot \mathbf{B}$ classify electromagnetic fields
  • The Lorentz force law becomes $dp^\mu/d\tau = qF^{\mu\nu}u_\nu$
  • The entire theory follows from the Lagrangian $\mathcal{L} = -\frac{1}{4\mu_0}F_{\mu\nu}F^{\mu\nu} - J_\mu A^\mu$
  • Electromagnetism was the first, and remains the clearest, example of a relativistic gauge field theory

Looking Ahead

With electromagnetism now in its natural relativistic language, we're ready to dive deep into special relativity itself. In Phase 4, we'll develop the full framework of Lorentz transformations, spacetime geometry, 4-vectors, and relativistic mechanics. We'll see how the concepts we've just used (4-vectors, tensors, invariants) apply to all of physics, and how the simple postulate that the laws of physics are the same in all inertial frames reshapes our understanding of space, time, energy, and mass.

Key Takeaways
  • The electric and magnetic fields are not separate entities but components of a single antisymmetric tensor $F^{\mu\nu}$, mixed together by Lorentz transformations.
  • Maxwell's four equations compress into just two covariant tensor equations, revealing the natural relativistic structure of electromagnetism.
  • The entire theory of classical electromagnetism follows from the Lagrangian $\mathcal{L} = -\frac{1}{4\mu_0}F_{\mu\nu}F^{\mu\nu} - J_\mu A^\mu$ via the principle of least action.
  • The Lorentz invariants $B^2 - E^2/c^2$ and $\mathbf{E} \cdot \mathbf{B}$ classify electromagnetic fields into types that no change of reference frame can alter.