Maxwell's Equations, Boundary Conditions, Potentials and Gauge Invariance
CSIR NET Maxwell's equations: displacement current, boundary conditions, Poynting theorem, scalar and vector potentials, Coulomb and Lorenz gauges.
- 20 min
On this page · 15 sections
Prerequisites
- Electrostatics: Gauss’s Law, Laplace and Poisson: boundary conditions on \( \mathbf E \) and \( \mathbf D \).
- Magnetostatics and Electromagnetic Induction: Ampère’s and Faraday’s laws, and the vector potential in the Coulomb gauge.
- Vector identities: \( \nabla\cdot(\nabla\times\mathbf A) = 0 \), \( \nabla\times(\nabla f) = 0 \), \( \nabla\times(\nabla\times\mathbf A) = \nabla(\nabla\cdot\mathbf A) - \nabla^2\mathbf A \), \( \nabla\cdot(\mathbf E\times\mathbf B) = \mathbf B\cdot(\nabla\times\mathbf E) - \mathbf E\cdot(\nabla\times\mathbf B) \).
The Hook: The Capacitor That Broke Ampère’s Law
Displacement Current
Take the divergence of \( \nabla\times\mathbf B = \mu_0\mathbf J \): the left side vanishes identically, so \( \nabla\cdot\mathbf J = 0 \). That contradicts the continuity equation \( \nabla\cdot\mathbf J = -\partial\rho/\partial t \) whenever charge piles up. Using Gauss’s law, \( \partial\rho/\partial t = \varepsilon_0\nabla\cdot(\partial\mathbf E/\partial t) \), so the combination \( \mathbf J + \varepsilon_0\partial\mathbf E/\partial t \) always has zero divergence. Maxwell’s corrected law is
$$\boxed{\nabla\times\mathbf B = \mu_0\mathbf J + \mu_0\varepsilon_0\frac{\partial\mathbf E}{\partial t}}, \qquad \mathbf J_d = \varepsilon_0\frac{\partial\mathbf E}{\partial t}$$Between the plates of a capacitor of area \( A \) carrying charge \( Q \), \( E = Q/\varepsilon_0A \) and \( J_dA = dQ/dt = I \): the displacement current between the plates equals the conduction current in the wire.
In a conductor the ratio of conduction to displacement current amplitude for fields varying as \( e^{-i\omega t} \) is \( \sigma/\varepsilon\omega \). Large means “good conductor”; small means “good dielectric”. This ratio controls everything about waves in matter on the next page.
Maxwell’s Equations
In vacuum (with sources)
| Name | Differential form | Integral form |
|---|---|---|
| Gauss (E) | \( \nabla\cdot\mathbf E = \rho/\varepsilon_0 \) | \( \oint\mathbf E\cdot d\mathbf a = Q_{\text{enc}}/\varepsilon_0 \) |
| Gauss (B), no monopoles | \( \nabla\cdot\mathbf B = 0 \) | \( \oint\mathbf B\cdot d\mathbf a = 0 \) |
| Faraday | \( \nabla\times\mathbf E = -\partial\mathbf B/\partial t \) | \( \oint\mathbf E\cdot d\mathbf l = -\dfrac{d}{dt}\int\mathbf B\cdot d\mathbf a \) |
| Ampère–Maxwell | \( \nabla\times\mathbf B = \mu_0\mathbf J + \mu_0\varepsilon_0\partial\mathbf E/\partial t \) | \( \oint\mathbf B\cdot d\mathbf l = \mu_0I_{\text{enc}} + \mu_0\varepsilon_0\dfrac{d}{dt}\int\mathbf E\cdot d\mathbf a \) |
Together with the Lorentz force law \( \mathbf F = q(\mathbf E + \mathbf v\times\mathbf B) \), these are the whole of classical electrodynamics.
In matter
Separate bound from free sources using \( \mathbf D = \varepsilon_0\mathbf E + \mathbf P \) and \( \mathbf H = \mathbf B/\mu_0 - \mathbf M \). A time-varying polarisation also produces a polarisation current \( \mathbf J_p = \partial\mathbf P/\partial t \). The result:
$$\nabla\cdot\mathbf D = \rho_f, \qquad \nabla\cdot\mathbf B = 0, \qquad \nabla\times\mathbf E = -\frac{\partial\mathbf B}{\partial t}, \qquad \nabla\times\mathbf H = \mathbf J_f + \frac{\partial\mathbf D}{\partial t}$$with constitutive relations \( \mathbf D = \varepsilon\mathbf E \), \( \mathbf B = \mu\mathbf H \) and, in conductors, Ohm’s law \( \mathbf J_f = \sigma\mathbf E \).
Structure worth knowing
- Homogeneous pair (\( \nabla\cdot\mathbf B = 0 \), Faraday) contains no sources. It is solved identically by introducing potentials.
- Inhomogeneous pair (Gauss, Ampère–Maxwell) contains the sources and becomes the equations for the potentials.
- Charge conservation is built in: taking the divergence of Ampère–Maxwell and using Gauss gives \( \partial\rho/\partial t + \nabla\cdot\mathbf J = 0 \). You cannot feed Maxwell’s equations a non-conserved current.
- Linearity: fields of separate sources add.
- Duality: in empty space the equations are unchanged under \( \mathbf E \to c\mathbf B \), \( c\mathbf B \to -\mathbf E \). With hypothetical magnetic charge \( \rho_m \) and current \( \mathbf J_m \), one would have \( \nabla\cdot\mathbf B = \mu_0\rho_m \) and \( \nabla\times\mathbf E = -\mu_0\mathbf J_m - \partial\mathbf B/\partial t \), and the symmetry would extend to sources. No monopole has been observed.
Boundary Conditions
Apply the integral forms to a thin pillbox (for normal components) and a thin rectangular loop (for tangential components) straddling the interface. With \( \hat n \) pointing from medium 2 into medium 1, free surface charge \( \sigma_f \) and free surface current \( \mathbf K_f \):
| Condition | General form | Linear media |
|---|---|---|
| Normal D | \( D_{1\perp} - D_{2\perp} = \sigma_f \) | \( \varepsilon_1E_{1\perp} - \varepsilon_2E_{2\perp} = \sigma_f \) |
| Normal B | \( B_{1\perp} = B_{2\perp} \) | same |
| Tangential E | \( \mathbf E_{1\parallel} = \mathbf E_{2\parallel} \) | same |
| Tangential H | \( \mathbf H_{1\parallel} - \mathbf H_{2\parallel} = \mathbf K_f\times\hat n \) | \( \dfrac{\mathbf B_{1\parallel}}{\mu_1} - \dfrac{\mathbf B_{2\parallel}}{\mu_2} = \mathbf K_f\times\hat n \) |
These hold for time-dependent fields too, because the pillbox and loop shrink to zero area and the \( \partial/\partial t \) terms drop out.
Refraction of field lines. With no free surface charge, \( E_\parallel \) and \( \varepsilon E_\perp \) are continuous, so \( \tan\theta_1/\tan\theta_2 = \varepsilon_1/\varepsilon_2 \) (angles from the normal). The magnetic analogue with no free surface current is \( \tan\theta_1/\tan\theta_2 = \mu_1/\mu_2 \): field lines leaving iron (\( \mu \gg \mu_0 \)) into air emerge almost along the normal.
Energy and Momentum: Poynting’s Theorem
The work done by the fields on charges in a volume \( \mathcal V \) per unit time is \( \int\mathbf E\cdot\mathbf J\,d\tau \). Eliminating \( \mathbf J \) with Ampère–Maxwell and using Faraday gives
$$\boxed{\int_{\mathcal V}\mathbf E\cdot\mathbf J\,d\tau = -\frac{d}{dt}\int_{\mathcal V}u\,d\tau - \oint_S\mathbf S\cdot d\mathbf a}$$$$u = \frac{1}{2}\left(\varepsilon_0E^2 + \frac{B^2}{\mu_0}\right), \qquad \boxed{\mathbf S = \frac{1}{\mu_0}\mathbf E\times\mathbf B}\ \ (\text{in matter } \mathbf S = \mathbf E\times\mathbf H)$$Work done on charges equals the decrease of field energy minus the energy flowing out through the surface. Differentially: \( \partial u/\partial t + \nabla\cdot\mathbf S = -\mathbf E\cdot\mathbf J \).
The fields also carry momentum, with density \( \mathbf g = \mu_0\varepsilon_0\mathbf S = \mathbf S/c^2 \), and angular momentum density \( \mathbf r\times\mathbf g \). The momentum flux is the (negative of the) Maxwell stress tensor, \( T_{ij} = \varepsilon_0\left(E_iE_j - \tfrac{1}{2}\delta_{ij}E^2\right) + \frac{1}{\mu_0}\left(B_iB_j - \tfrac{1}{2}\delta_{ij}B^2\right) \); the force on charges in a volume is \( \oint T_{ij}da_j - \frac{d}{dt}\int g_i\,d\tau \).
The same theorem shows that, when a capacitor charges, energy flows in from the sides of the gap, and that a static charge next to a static magnet has non-zero \( \mathbf E\times\mathbf B \) and stores field angular momentum (the basis of “hidden momentum” puzzles such as Feynman’s disc paradox).
Scalar and Vector Potentials
The homogeneous equations are solved identically by potentials:
$$\nabla\cdot\mathbf B = 0 \Rightarrow \mathbf B = \nabla\times\mathbf A; \qquad \nabla\times\left(\mathbf E + \frac{\partial\mathbf A}{\partial t}\right) = 0 \Rightarrow \boxed{\mathbf E = -\nabla V - \frac{\partial\mathbf A}{\partial t}}$$Six field components are replaced by four potential components. Substituting into the inhomogeneous equations gives
$$\nabla^2V + \frac{\partial}{\partial t}(\nabla\cdot\mathbf A) = -\frac{\rho}{\varepsilon_0}$$$$\left(\nabla^2\mathbf A - \mu_0\varepsilon_0\frac{\partial^2\mathbf A}{\partial t^2}\right) - \nabla\left(\nabla\cdot\mathbf A + \mu_0\varepsilon_0\frac{\partial V}{\partial t}\right) = -\mu_0\mathbf J$$These look messy because they still contain the gauge freedom.
Gauge Transformations
For any scalar function \( \lambda(\mathbf r, t) \),
$$\boxed{\mathbf A' = \mathbf A + \nabla\lambda, \qquad V' = V - \frac{\partial\lambda}{\partial t}}$$leaves \( \mathbf E \) and \( \mathbf B \) unchanged: \( \nabla\times\nabla\lambda = 0 \), and the extra \( -\nabla(-\partial_t\lambda) - \partial_t\nabla\lambda = 0 \). This is gauge invariance: the fields, and all measurable quantities, are unchanged; the potentials are not unique. One uses the freedom to impose a convenient condition.
| Gauge | Condition | Equation for \( V \) | Equation for \( \mathbf A \) | Character |
|---|---|---|---|---|
| Coulomb (transverse) | \( \nabla\cdot\mathbf A = 0 \) | \( \nabla^2V = -\rho/\varepsilon_0 \): instantaneous Coulomb potential | wave equation with source \( -\mu_0\mathbf J + \mu_0\varepsilon_0\nabla(\partial_tV) \) | simple \( V \), messy \( \mathbf A \); natural for radiation in free space and in atomic physics |
| Lorenz | \( \nabla\cdot\mathbf A = -\mu_0\varepsilon_0\dfrac{\partial V}{\partial t} \) | \( \Box^2V = -\rho/\varepsilon_0 \) | \( \Box^2\mathbf A = -\mu_0\mathbf J \) | both wave equations with speed \( c \); Lorentz covariant |
| Temporal (Weyl) | \( V = 0 \) | none | \( \mathbf E = -\partial_t\mathbf A \) | convenient for some wave problems |
Here \( \Box^2 = \nabla^2 - \frac{1}{c^2}\frac{\partial^2}{\partial t^2} \). The Lorenz condition is named after Ludvig Lorenz; it is often misprinted “Lorentz”.
In the Coulomb gauge \( V \) responds instantly to distant charge, which looks acausal. It is harmless: \( V \) is not measurable, and the extra instantaneous piece in \( \mathbf A \) cancels it in \( \mathbf E \), which propagates at \( c \).
Retarded potentials
In the Lorenz gauge the solutions for localised sources are
$$V(\mathbf r, t) = \frac{1}{4\pi\varepsilon_0}\int\frac{\rho(\mathbf r', t_r)}{\mathscr R}d\tau', \qquad \mathbf A(\mathbf r, t) = \frac{\mu_0}{4\pi}\int\frac{\mathbf J(\mathbf r', t_r)}{\mathscr R}d\tau', \qquad t_r = t - \frac{\mathscr R}{c}$$the static formulas evaluated at the retarded time. The fields are not simply the static fields at retarded time (Jefimenko’s equations contain extra terms), which is the origin of radiation. Radiation itself belongs to the advanced electrodynamics unit.
Gauge invariance in quantum mechanics
The Hamiltonian of a charge in potentials is \( H = \frac{1}{2m}(\mathbf p - q\mathbf A)^2 + qV \). Under a gauge transformation the Schrödinger equation keeps its form provided the wavefunction also changes phase:
$$\psi' = e^{iq\lambda/\hbar}\psi$$Probabilities and expectation values of gauge-invariant operators (such as the kinetic momentum \( m\mathbf v = \mathbf p - q\mathbf A \)) are unchanged. The canonical momentum \( \mathbf p \) is not gauge-invariant. The Aharonov–Bohm phase \( q\oint\mathbf A\cdot d\mathbf l/\hbar = q\Phi/\hbar \) is gauge-invariant because it depends only on the enclosed flux. More on this formalism in Operators, Commutators, Dirac Notation and Uncertainty.
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Displacement current \( \varepsilon_0\partial\mathbf E/\partial t \): no moving charge | Polarisation current \( \partial\mathbf P/\partial t \): real bound-charge motion | |
| Lorenz gauge (Ludvig Lorenz) | Lorentz force, Lorentz transformation (Hendrik Lorentz) | |
| \( \mathbf E \), \( \mathbf B \) gauge-invariant | \( V \), \( \mathbf A \), canonical \( \mathbf p \) gauge-dependent | |
| Coulomb gauge: \( V \) instantaneous | Fields still propagate at \( c \) | |
| \( E_\parallel \) continuous at every interface | \( E_\perp \) jumps by \( \sigma/\varepsilon_0 \) (total charge) | |
| \( \mathbf S = \mathbf E\times\mathbf B/\mu_0 \): energy flux | \( \mathbf g = \mathbf S/c^2 \): momentum density |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper.
Level 1: Foundation
- Write the displacement current density.
- Which two of Maxwell’s equations contain no source terms?
- State the Lorenz gauge condition.
Answers
- \( \mathbf J_d = \varepsilon_0\partial\mathbf E/\partial t \) (in matter, \( \partial\mathbf D/\partial t \) includes the polarisation current).
- \( \nabla\cdot\mathbf B = 0 \) and \( \nabla\times\mathbf E + \partial\mathbf B/\partial t = 0 \).
- \( \nabla\cdot\mathbf A + \mu_0\varepsilon_0\,\partial V/\partial t = 0 \).
Level 2: Part B Style
Practice Q1. A capacitor with circular plates of radius 5 cm is charged by a steady current of 2 A. The magnetic field between the plates at 3 cm from the axis is:
(a) 4.8 μT (b) 8.0 μT (c) 13.3 μT (d) zero
Answer: (a) 4.8 μT
\( B = \dfrac{\mu_0Is}{2\pi R^2} = 4.8\times10^{-6} \) T (worked above). Option (c) is \( \mu_0I/2\pi s \), the field of the wire at 3 cm, which ignores that only the fraction \( s^2/R^2 \) of the displacement current is enclosed. Option (d) ignores the displacement current altogether.
Practice Q2. The plane \( z = 0 \) separates medium 1 (\( z > 0 \), \( \varepsilon_r = 2 \)) from medium 2 (\( z < 0 \), \( \varepsilon_r = 5 \)). There is no free surface charge. Just above the interface \( \mathbf E_1 = (3\hat x + 4\hat y + 10\hat z) \) V/m. Just below, \( \mathbf E_2 \) is:
(a) \( 3\hat x + 4\hat y + 25\hat z \) (b) \( 3\hat x + 4\hat y + 4\hat z \) (c) \( 1.2\hat x + 1.6\hat y + 10\hat z \) (d) \( 3\hat x + 4\hat y + 10\hat z \)
Answer: (b)
Tangential \( \mathbf E \) is continuous: \( 3\hat x + 4\hat y \). Normal \( D \) is continuous: \( 2\times10 = 5E_{2z} \Rightarrow E_{2z} = 4 \) V/m. Option (a) inverts the permittivity ratio; option (c) scales the tangential part instead.
Practice Q3. The potentials \( V = 0 \), \( \mathbf A = -\dfrac{1}{4\pi\varepsilon_0}\dfrac{qt}{r^2}\hat r \) describe:
(a) a uniform magnetic field (b) a point charge at rest at the origin (c) a charge that grows linearly in time (d) no fields at all
Answer: (b)
\( \mathbf B = \nabla\times\mathbf A = 0 \) and \( \mathbf E = -\partial_t\mathbf A = \dfrac{q}{4\pi\varepsilon_0r^2}\hat r \), time-independent. The gauge function \( \lambda = -qt/(4\pi\varepsilon_0r) \) converts it to the ordinary Coulomb potential. Option (c) mistakes the explicit \( t \) in the potential for time dependence of the source.
Practice Q4. A long straight wire of radius \( a \) and resistance \( R \) carries a steady current \( I \). The Poynting vector at its surface is:
(a) along the wire, magnitude \( I^2R/\pi a^2 \) (b) radially outward, magnitude \( I^2R/2\pi al \) (c) radially inward, magnitude \( I^2R/2\pi al \) (d) zero, since the fields are static
Answer: (c)
\( E = IR/l \) along the wire, \( B = \mu_0I/2\pi a \) azimuthal; \( \mathbf E\times\mathbf B \) points inward. \( S = \dfrac{IR}{l}\dfrac{I}{2\pi a} \). Multiplying by the curved area \( 2\pi al \) gives \( I^2R \), the Joule heat. Static fields can carry steady energy flow, so (d) is wrong.
Level 3: Part C Style
Practice Q5. In vacuum the potentials are \( V = 0 \), \( \mathbf A = A_0\sin(kx - \omega t)\,\hat y \). Consider:
- \( \mathbf E = A_0\omega\cos(kx - \omega t)\,\hat y \).
- \( \mathbf B = A_0k\cos(kx - \omega t)\,\hat z \).
- The potentials satisfy the Coulomb gauge but not the Lorenz gauge.
- With no sources present, Maxwell’s equations require \( \omega = ck \).
Which are correct?
(a) 1, 2 and 4 only (b) 1 and 2 only (c) 1, 2 and 3 only (d) 2 and 4 only
Answer: (a) 1, 2 and 4 only
Statement 1: \( \mathbf E = -\partial_t\mathbf A = A_0\omega\cos(kx - \omega t)\hat y \). True.
Statement 2: \( (\nabla\times\mathbf A)_z = \partial_xA_y = A_0k\cos(kx - \omega t) \). True.
Statement 3: \( \nabla\cdot\mathbf A = \partial_yA_y = 0 \), so the Coulomb condition holds; \( \partial_tV = 0 \) too, so the Lorenz condition \( \nabla\cdot\mathbf A + \mu_0\varepsilon_0\partial_tV = 0 \) also holds. “Not the Lorenz gauge” is false.
Statement 4: in either gauge with no sources \( \Box^2\mathbf A = 0 \), giving \( -k^2 + \omega^2/c^2 = 0 \). Equivalently, \( \nabla\times\mathbf B = \mu_0\varepsilon_0\partial_t\mathbf E \) requires \( A_0k^2 = A_0\omega^2/c^2 \). True. Note also \( E_0/B_0 = \omega/k = c \).
Practice Q6. For a medium with \( \sigma = 4 \) S/m and \( \varepsilon_r = 81 \) (take \( \mu = \mu_0 \)), a student compares conduction and displacement current amplitudes for harmonic fields. Which statements are correct?
- At 1 MHz the conduction current exceeds the displacement current by a factor of about 900.
- The two are equal at a frequency of about 0.89 GHz.
- Above that frequency the medium behaves more like a dielectric than a conductor.
- The ratio is independent of frequency.
(a) 1 and 4 only (b) 1, 2 and 3 only (c) 2 and 3 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
Ratio \( = \sigma/\varepsilon\omega \), with \( \varepsilon = 81\times8.854\times10^{-12} = 7.17\times10^{-10} \) F/m.
At \( f = 10^6 \) Hz: \( \dfrac{4}{2\pi\times10^6\times7.17\times10^{-10}} \approx 888 \), about 900. Statement 1 true.
Equality: \( f = \dfrac{\sigma}{2\pi\varepsilon} = \dfrac{4}{2\pi\times7.17\times10^{-10}} = 8.9\times10^8 \) Hz. Statement 2 true.
Above it, displacement current dominates: dielectric behaviour. Statement 3 true. Statement 4 contradicts the \( 1/\omega \) dependence. False.
Practice Q7. A gauge function \( \lambda = \alpha\left(x^2 + c^2t^2\right) \) (\( \alpha \) a constant) is applied to potentials that satisfy the Lorenz condition. Consider:
- The fields \( \mathbf E \) and \( \mathbf B \) change.
- The new potentials still satisfy the Lorenz condition.
- \( \mathbf A \) changes by \( 2\alpha x\,\hat x \) and \( V \) changes by \( -2\alpha c^2t \).
- The new potentials satisfy the Coulomb condition if the old ones did.
Which are correct?
(a) 2 and 3 only (b) 3 only (c) 1 and 2 only (d) 2, 3 and 4 only
Answer: (a) 2 and 3 only
Statement 1: gauge transformations never change the fields. False.
Statement 2: the Lorenz condition is preserved if and only if \( \Box^2\lambda = 0 \). Here \( \nabla^2\lambda = 2\alpha \) and \( \frac{1}{c^2}\partial_t^2\lambda = \frac{1}{c^2}(2\alpha c^2) = 2\alpha \), so \( \Box^2\lambda = 2\alpha - 2\alpha = 0 \). True.
Statement 3: \( \nabla\lambda = 2\alpha x\,\hat x \); \( -\partial_t\lambda = -2\alpha c^2t \). True.
Statement 4: \( \nabla\cdot\mathbf A' = \nabla\cdot\mathbf A + \nabla^2\lambda = \nabla\cdot\mathbf A + 2\alpha \neq 0 \). False.
The trap: with \( \lambda = \alpha(x^2 - c^2t^2) \) instead, \( \Box^2\lambda = 2\alpha + 2\alpha = 4\alpha \neq 0 \) and the Lorenz condition would be lost. Always compute \( \Box^2\lambda \); do not judge by the look of the combination.
Quick Revision Table
| Result | Formula |
|---|---|
| Displacement current | \( \mathbf J_d = \varepsilon_0\partial_t\mathbf E \) |
| Ampère–Maxwell | \( \nabla\times\mathbf B = \mu_0\mathbf J + \mu_0\varepsilon_0\partial_t\mathbf E \) |
| Matter | \( \nabla\cdot\mathbf D = \rho_f \), \( \nabla\times\mathbf H = \mathbf J_f + \partial_t\mathbf D \) |
| Continuity | \( \partial_t\rho + \nabla\cdot\mathbf J = 0 \) |
| Boundaries | \( \Delta D_\perp = \sigma_f \), \( \Delta B_\perp = 0 \), \( \Delta\mathbf E_\parallel = 0 \), \( \Delta\mathbf H_\parallel = \mathbf K_f\times\hat n \) |
| Poynting | \( \mathbf S = \mathbf E\times\mathbf B/\mu_0 \), \( u = \frac{1}{2}(\varepsilon_0E^2 + B^2/\mu_0) \) |
| Momentum density | \( \mathbf g = \mathbf S/c^2 \) |
| Potentials | \( \mathbf B = \nabla\times\mathbf A \), \( \mathbf E = -\nabla V - \partial_t\mathbf A \) |
| Gauge | \( \mathbf A + \nabla\lambda \), \( V - \partial_t\lambda \) |
| Coulomb | \( \nabla\cdot\mathbf A = 0 \); \( \nabla^2V = -\rho/\varepsilon_0 \) |
| Lorenz | \( \nabla\cdot\mathbf A = -\mu_0\varepsilon_0\partial_tV \); \( \Box^2V = -\rho/\varepsilon_0 \), \( \Box^2\mathbf A = -\mu_0\mathbf J \) |
| QM | \( \psi \to e^{iq\lambda/\hbar}\psi \) |
Key Takeaways
What’s Next
- EM Waves, Reflection, Refraction and Polarization: source-free Maxwell equations give waves, and the boundary conditions on this page give the Fresnel equations.
- Charged Particles in Static Uniform Fields: the Lorentz force and the Hamiltonian \( (\mathbf p - q\mathbf A)^2/2m \) in action.
- Electromagnetic Theory: Quick Revision.
- The second-order PDE theory behind the wave and Poisson equations is in Classification of Second-Order PDEs.