Electromagnetic Theory: Boundary-Value Problems, Images, Media, Multipoles, Potentials and Gauges, Waves, Fresnel Coefficients and the Poynting Vector

Section 5 of the GATE Physics paper. The electromagnetics chapters of the engineering papers start from Maxwell’s equations and head for transmission lines and antennas; this one starts from the boundary-value problems of electrostatics and heads for the field theory. It follows the syllabus: electrostatic and magnetostatic boundary-value problems, the method of images and separation of variables; dielectrics, conductors and magnetic materials; the multipole expansion; Maxwell’s equations; scalar and vector potentials with the Coulomb and Lorenz gauges; electromagnetic waves in free space, in non-conducting media and in conductors; reflection and transmission at normal and oblique incidence; polarisation; and the Poynting vector and theorem, with the energy and momentum a wave carries. SI units throughout, with ε₀ = 8.854 × 10⁻¹² F/m, μ₀ = 4π × 10⁻⁷ H/m and c = 3.00 × 10⁸ m/s.

1. Boundary-value problems: uniqueness, images and separation of variables

In a charge-free region the potential obeys Laplace’s equation ∇²V = 0 (Poisson’s ∇²V = −ρ/ε₀ where there is charge). The uniqueness theorem — V is fixed once it, or the total charge on each conductor, is given on every boundary — licenses any trick that satisfies the equation and the boundary conditions, because whatever works is the answer.

  • Point charge q at height d above a grounded plane: replace the plane by an image −q at depth d. The induced surface charge totals −q, and the force on q is q²/[4πε₀(2d)²], attractive. For 1 μC at 0.1 m this is 8.988 × 10⁹ × 10⁻¹²/0.04 = 0.225 N.
  • Point charge q at distance a from the centre of a grounded sphere of radius R: the image is q′ = −qR/a at distance R²/a from the centre. An isolated neutral sphere needs a second image +qR/a at the centre.
  • Separation of variables: in Cartesian coordinates, V = X(x)Y(y) turns ∇²V = 0 into X″/X = −Y″/Y = k², giving sines in one direction and exponentials in the other; the boundary conditions fix k and a Fourier series fixes the coefficients. In spherical coordinates with azimuthal symmetry, V = Σ(Aₗrˡ + Bₗr−(l+1))Pₗ(cos θ).
ℹ️ The dielectric sphere in a uniform field
Only the l = 1 term survives, and the field inside is uniform: E_in = 3E₀/(ε_r + 2), weaker than E₀ because the bound surface charge sets up an opposing field. A conducting sphere is the limit ε_r → ∞: E_in = 0, and the sphere acquires a dipole moment 4πε₀R³E₀.

2. Dielectrics, conductors, magnetic materials and the multipole expansion

A dielectric’s polarisation P produces bound charges ρ_b = −∇·P and σ_b = P·n̂. The displacement D = ε₀E + P has ∇·D = ρ_free, and in a linear medium D = εE with ε = ε₀(1 + χ_e). A magnetised material has bound currents J_b = ∇×M and K_b = M×n̂; H = B/μ₀ − M has ∇×H = J_free, and B = μH with μ = μ₀(1 + χ_m). In a conductor in electrostatic equilibrium, E = 0 inside, any charge sits on the surface, and E just outside is σ/ε₀ along the normal.

Boundary conditions at an interface (no free surface charge or current)
ContinuousDiscontinuous
E_tangential, D_normalE_normal jumps by the ratio of permittivities
H_tangential, B_normalB_tangential jumps by the ratio of permeabilities

Far from a localised charge distribution, the multipole expansion V = (1/4πε₀)Σ (1/rl+1) ∫ r′ˡ Pₗ(cos α) ρ dτ′ sorts the potential by its fall-off: monopole (total charge) as 1/r, dipole as 1/r², quadrupole as 1/r³. The dipole term is V = p cos θ/(4πε₀r²), and its field falls as 1/r³. The lowest non-vanishing moment is independent of the origin; a neutral pair of charges has zero monopole, so its dipole moment is origin-independent. The magnetic expansion has no monopole term at all, and its dipole term uses m = I × area.

3. Maxwell’s equations, potentials and gauges

Maxwell’s equations in matter
Differential formContent
∇·D = ρ_freeGauss: charge is the source of D
∇·B = 0no magnetic monopoles
∇×E = −∂B/∂tFaraday: a changing B induces a curling E
∇×H = J_free + ∂D/∂tAmpère–Maxwell: the displacement current completes the law

Without the displacement current ∂D/∂t, the divergence of Ampère’s law would force ∇·J = 0, contradicting the continuity equation ∇·J + ∂ρ/∂t = 0 whenever charge accumulates, as on a charging capacitor. Because ∇·B = 0, one can write B = ∇×A; Faraday’s law then allows E = −∇V − ∂A/∂t. The potentials are not unique: the gauge transformation A → A + ∇λ, V → V − ∂λ/∂t leaves E and B unchanged.

  • Coulomb gauge ∇·A = 0: V obeys Poisson’s equation with the instantaneous charge density — convenient for statics and radiation in source-free regions, and not manifestly relativistic.
  • Lorenz gauge ∇·A + (1/c²)∂V/∂t = 0: both potentials obey wave equations, □²V = −ρ/ε₀ and □²A = −μ₀J, with the same operator — the Lorentz-covariant choice, and the one whose solutions are the retarded potentials.

4. Waves in free space, dielectrics and conductors

In a source-free linear medium Maxwell’s equations give ∇²E = με ∂²E/∂t², a wave of speed v = 1/√(με) = c/n with n = √(ε_rμ_r). For a plane wave E, B and k are mutually perpendicular, B = E/v, and the intrinsic impedance is η = √(μ/ε): 376.7 Ω in vacuum and 376.7/2 ≈ 188 Ω for ε_r = 4, μ_r = 1.

In a conductor of conductivity σ the wavenumber is complex, k = β + iα. For a good conductor (σ ≫ ωε) α = β = 1/δ with the skin depth δ = √(2/(ωμσ)) = 1/√(πfμσ): the amplitude falls as e−z/δ, B lags E by 45°, the phase velocity is ωδ, and almost all the field energy is magnetic. For copper (σ = 5.8 × 10⁷ S/m) at 1 MHz, δ = 1/√(π × 10⁶ × 4π × 10⁻⁷ × 5.8 × 10⁷) = 66 μm; at 50 Hz it is about 9 mm, which is why high-frequency current flows only in the surface of a wire.

⚠️ Skin depth falls with frequency and with conductivity
δ ∝ 1/√(fσ). Quadrupling the frequency halves δ; a better conductor has a thinner skin. The intuition that a better conductor lets the wave in further is exactly backwards: a perfect conductor has δ = 0 and reflects everything.

5. Reflection, transmission and polarisation

At normal incidence from medium 1 to medium 2 (non-magnetic), the amplitude coefficients are r = (n₁ − n₂)/(n₁ + n₂) and t = 2n₁/(n₁ + n₂), and the power coefficients are R = r² and T = (n₂/n₁)t², with R + T = 1. Air to glass (n = 1.5) reflects R = (0.5/2.5)² = 4%. The negative r for n₂ > n₁ is the half-wave phase change on reflection from a denser medium.

  • Oblique incidence splits into s (TE, E perpendicular to the plane of incidence) and p (TM, E in it), each with its own Fresnel coefficients, e.g. r_s = (n₁cos θᵢ − n₂cos θₜ)/(n₁cos θᵢ + n₂cos θₜ).
  • Brewster angle tan θ_B = n₂/n₁: r_p = 0, so reflected light is purely s-polarised, and the reflected and refracted rays are perpendicular. For n₂/n₁ = √3, θ_B = 60°.
  • Total internal reflection for n₁ > n₂ beyond sin θ_c = n₂/n₁: |r| = 1, and an evanescent field decays into medium 2 without carrying power on average.
  • Polarisation is the pattern traced by E: linear when the two transverse components are in phase, circular when they are equal and 90° apart, elliptical otherwise.

6. The Poynting theorem, energy and momentum of waves

The Poynting theorem, −∂u/∂t = ∇·S + J·E, is energy conservation for fields: the field energy u = ½(ε₀E² + B²/μ₀) in a volume falls by the flux of S = E × H out through its surface plus the work done on charges inside. For a plane wave in vacuum the electric and magnetic energies are equal, and the time-averaged intensity is I = ⟨S⟩ = ½cε₀E₀². A 1 mW beam spread over 1 mm² has I = 1000 W/m² and E₀ = √(2I/(cε₀)) ≈ 868 V/m.

A wave also carries momentum, with density S/c², so it exerts a radiation pressure: I/c on a perfect absorber and 2I/c on a perfect reflector at normal incidence. Sunlight at 1360 W/m² presses on an absorbing surface with 1360/(3 × 10⁸) = 4.53 μPa. A circularly polarised wave also carries angular momentum, ±ħ per photon.

Key takeaways

  • Uniqueness licenses images: −q at depth d for a grounded plane, −qR/a at R²/a for a grounded sphere; a dielectric sphere in E₀ has E_in = 3E₀/(ε_r + 2).
  • Bound charges −∇·P and P·n̂, bound currents ∇×M and M×n̂; multipoles fall as 1/r, 1/r², 1/r³, and there is no magnetic monopole term.
  • E = −∇V − ∂A/∂t, B = ∇×A; Coulomb gauge ∇·A = 0, Lorenz gauge ∇·A + (1/c²)∂V/∂t = 0 gives wave equations for both potentials.
  • Good conductor: δ = √(2/(ωμσ)), falling with f and σ; B lags E by 45°.
  • R = ((n₁ − n₂)/(n₁ + n₂))² at normal incidence, tan θ_B = n₂/n₁; I = ½cε₀E₀², pressure I/c absorbed and 2I/c reflected.

Practice questions (14)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

  1. A point charge of 1 μC is held 0.1 m above an infinite grounded conducting plane. Taking 1/(4πε₀) = 8.988 × 10⁹ N m² C⁻², the magnitude of the force on the charge, in N, to three decimal places, is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 0.225

    The image −q sits 0.1 m below the plane, 0.2 m from q, so F = 8.988 × 10⁹ × (10⁻⁶)²/(0.2)² = 0.225 N, attractive. Using the height 0.1 m as the separation gives 0.899 N, four times too large.
  2. A point charge q is at distance a from the centre of a grounded conducting sphere of radius R (a > R). The image charge and its distance from the centre are:

    1. −qR/a at R²/a
    2. −q at R²/a
    3. −qR/a at a − R
    4. −qa/R at R²/a
    Show answer

    Answer: A — −qR/a at R²/a

    Requiring V = 0 on the whole sphere forces q′ = −qR/a placed at the inverse point R²/a. The full −q is the plane’s image (the R → ∞ limit), and −qa/R would exceed q in magnitude.
  3. A linear dielectric sphere of relative permittivity ε_r is placed in a uniform field E₀. The field inside the sphere is:

    1. Uniform, 3E₀/(ε_r + 2)
    2. Uniform, E₀/ε_r
    3. Zero
    4. Non-uniform, largest at the poles
    Show answer

    Answer: A — Uniform, 3E₀/(ε_r + 2)

    Separation of variables with only the l = 1 Legendre term gives a uniform interior field 3E₀/(ε_r + 2). E₀/ε_r is the result for a slab perpendicular to the field, not a sphere, and zero is the conducting limit ε_r → ∞.
  4. Far from a neutral charge distribution whose dipole moment vanishes but whose quadrupole moment does not, the potential falls off as:

    1. 1/r³
    2. 1/r²
    3. 1/r
    4. 1/r⁴
    Show answer

    Answer: A — 1/r³

    The l-th multipole contributes to V as 1/rl+1: monopole 1/r, dipole 1/r², quadrupole (l = 2) 1/r³. 1/r⁴ is the quadrupole’s field, one power faster than its potential.
  5. Which of the following hold for time-dependent fields in vacuum?

    1. ∇·B = 0
    2. ∇×E = −∂B/∂t
    3. ∇·E = ρ/ε₀
    4. ∇×B = μ₀J
    Show answer

    Answer: A — ∇·B = 0; B — ∇×E = −∂B/∂t; C — ∇·E = ρ/ε₀

    The first three are Maxwell’s equations as they stand. ∇×B = μ₀J is Ampère’s law for steady currents only; with time-dependent fields it lacks the displacement term μ₀ε₀∂E/∂t and would contradict charge conservation.
  6. The Lorenz gauge condition is:

    1. ∇·A + (1/c²)∂V/∂t = 0
    2. ∇·A = 0
    3. ∇×A = 0
    4. ∇·A − ∂V/∂t = 0
    Show answer

    Answer: A — ∇·A + (1/c²)∂V/∂t = 0

    The Lorenz condition couples A and V so that both obey the same wave equation and the condition is Lorentz invariant. ∇·A = 0 is the Coulomb gauge; ∇×A = 0 would make B vanish and is not a gauge choice at all.
  7. Copper has σ = 5.8 × 10⁷ S/m and μ = μ₀ = 4π × 10⁻⁷ H/m. The skin depth at 1 MHz, in μm, to the nearest integer, is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 66

    δ = 1/√(πfμσ) = 1/√(3.1416 × 10⁶ × 1.2566 × 10⁻⁶ × 5.8 × 10⁷) = 1/√(2.29 × 10⁸) = 6.61 × 10⁻⁵ m = 66 μm. Dropping the π, 1/√(fμσ), gives 117 μm; using ω in place of πf in 1/√(πfμσ) gives 47 μm.
  8. For a plane wave in a good conductor (σ ≫ ωε), which statements are correct?

    1. B lags E by 45°
    2. The skin depth is proportional to 1/√f
    3. The skin depth increases with conductivity
    4. The magnetic energy density far exceeds the electric
    Show answer

    Answer: A — B lags E by 45°; B — The skin depth is proportional to 1/√f; D — The magnetic energy density far exceeds the electric

    k = (1 + i)/δ has phase 45°, and B = kE/ω lags E by that angle. δ = √(2/(ωμσ)) falls as 1/√f and as 1/√σ, so a better conductor has a thinner skin. |B|/|E| = |k|/ω is large, so B²/μ ≫ εE²: the energy is mostly magnetic.
  9. The intrinsic impedance of free space is 376.7 Ω. For a non-magnetic lossless dielectric with ε_r = 4, the intrinsic impedance in Ω, to the nearest integer, is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 188

    η = √(μ/ε) = η₀/√ε_r = 376.7/2 = 188 Ω. Dividing by ε_r instead of its square root gives 94 Ω.
  10. Light falls normally from air onto glass of refractive index 1.5. The fraction of the incident power reflected is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 0.04

    R = ((n₁ − n₂)/(n₁ + n₂))² = (−0.5/2.5)² = 0.04. The amplitude coefficient is −0.2; reporting it, or its magnitude, instead of its square is the usual error.
  11. Light in air is incident on a medium of refractive index 1.732. The Brewster angle, in degrees, to the nearest integer, is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 60

    tan θ_B = n₂/n₁ = 1.732 = √3, so θ_B = 60°. At this angle r_p = 0 and the reflected light is s-polarised. Using sin θ = 1/n (a critical-angle formula) gives 35°.
  12. A wave has E = E₀(x̂ cos(kz − ωt) + ŷ sin(kz − ωt)). It is:

    1. Circularly polarised
    2. Linearly polarised at 45°
    3. Elliptically polarised with unequal axes
    4. Unpolarised
    Show answer

    Answer: A — Circularly polarised

    The two components have equal amplitude and are 90° out of phase, so the tip of E traces a circle of radius E₀ at fixed z. Linear polarisation at 45° needs the components in phase; unequal amplitudes would make it elliptical.
  13. A 1 mW laser beam of uniform intensity has a cross-section of 1 mm². With c = 3.00 × 10⁸ m/s and ε₀ = 8.854 × 10⁻¹² F/m, the amplitude of its electric field, in V/m, to the nearest integer, is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 868

    I = 10⁻³ W/10⁻⁶ m² = 1000 W/m², and I = ½cε₀E₀² gives E₀ = √(2 × 1000/(3 × 10⁸ × 8.854 × 10⁻¹²)) = √(7.53 × 10⁵) = 868 V/m. Leaving out the ½ gives the rms-like value 614 V/m.
  14. Sunlight of intensity 1360 W/m² falls normally on a perfectly absorbing surface. With c = 3.00 × 10⁸ m/s, the radiation pressure in μPa, to two decimal places, is ____.

    Numerical answer — type the value.

    Show answer

    Answer: 4.53

    P = I/c = 1360/(3 × 10⁸) = 4.53 × 10⁻⁶ Pa = 4.53 μPa. A perfect reflector would feel twice this, 9.07 μPa, because the momentum is reversed rather than absorbed.