Boundary Conditions General

At the interface between two different electromagnetic media, the field vectors E, D, H, and B do not remain continuous in general. The boundary conditions specify exactly which components of these fields are continuous across an interface and which suffer a discontinuity. These conditions are derived directly from Maxwell's integral equations by applying them to infinitesimally thin surfaces and loops straddling the interface. Without boundary conditions, field problems involving multiple media regions cannot be solved, making this topic fundamental in waveguide theory, antenna design, and microwave engineering.

Core Concept Explanation

The boundary conditions are derived by applying Maxwell's integral equations to limiting cases at the interface. For the tangential electric field, Faraday's law is applied to a rectangular loop of infinitesimal height straddling the boundary. As the height shrinks to zero, the flux integral over the enclosed area also shrinks to zero, leaving the equality of tangential E on both sides. This gives Et1 = Et2, which holds unconditionally at any interface.

For the tangential magnetic field, the Ampere-Maxwell law is applied to a similar loop. As the area shrinks, the displacement current integral vanishes but the line integral of any surface current Js on the boundary remains finite. This gives an × (H1 − H2) = Js. If no surface current exists (as at a dielectric-dielectric interface), the tangential H is continuous.

For the normal electric flux density D, Gauss's law is applied to a Gaussian pillbox (thin cylinder) straddling the interface. As the height of the pillbox shrinks, only the top and bottom faces contribute to the flux integral. The net outward flux equals the enclosed surface charge density ρs, giving Dn1 − Dn2 = ρs. At a charge-free interface (typical dielectric-dielectric boundary), Dn1 = Dn2.

For the normal magnetic flux density B, Gauss's magnetic law applied to the same pillbox gives Bn1 = Bn2 unconditionally, since there is never any enclosed magnetic charge. This is a direct consequence of the nonexistence of magnetic monopoles.

Mathematical Expression

The four general boundary conditions, with an₁₂ denoting the unit normal pointing from medium 2 to medium 1:

Et1 = Et2 (tangential E always continuous)

an × (H1 − H2) = Js (Js = 0 at dielectric interface → Ht1 = Ht2)

Dn1 − Dn2 = ρs (ρs = 0 at dielectric interface → Dn1 = Dn2)

Bn1 = Bn2 (normal B always continuous)

Since D = εE and B = μH, the normal continuity conditions can be rewritten in terms of E and H. The normal D condition ε₁En1 = ε₂En2 (for ρs = 0) shows that if ε₁ ≠ ε₂, the normal component of E is discontinuous in proportion to the permittivity ratio. Similarly, the normal B condition μ₁Hn1 = μ₂Hn2 (always) shows that normal H is discontinuous at an interface between media of different permeability.

Practical Understanding

For a perfect electric conductor (PEC) surface, the boundary conditions simplify significantly. Inside a PEC, E = 0 and H = 0 (fields cannot penetrate). Therefore, just outside a PEC surface: the tangential E is zero (Et = 0), the normal D equals the surface charge density (Dn = ρs), the normal B is zero (Bn = 0), and the tangential H equals the surface current density (Ht = Js). These PEC boundary conditions are used extensively in waveguide analysis.

For a perfect magnetic conductor (PMC), the dual conditions apply: normal B = 0, tangential H = 0 outside. PMC is a theoretical concept used in antenna design as a dual to PEC, but it does not exist in nature. However, certain metamaterials and high-impedance surfaces approximate PMC behavior at specific frequencies.

The boundary conditions also determine Snell's law of refraction for electromagnetic waves. When a wave crosses a dielectric boundary, the tangential component of the wave vector must be continuous, which is the field-theoretic derivation of the angle relationship. The ratio of refracted to incident field amplitudes (Fresnel coefficients) also follows directly from applying all four boundary conditions simultaneously.

Given:
Interface between Medium 1 (ε₁ = 4ε₀, no surface charge) and Medium 2 (ε₂ = ε₀).
In Medium 1, the electric field is E1 = 3ax + 4ay + 5az V/m.
Interface is the z = 0 plane with an = az (normal pointing from 2 to 1).
Find E2 in Medium 2.

Why this formula applies:
Tangential E is continuous: Et1 = Et2 (x and y components)
Normal D is continuous (no surface charge): Dn1 = Dn2 → ε₁En1 = ε₂En2

Formula:
Et2 = Et1 (tangential components unchanged)
ε₁ Ez1 = ε₂ Ez2  → Ez2 = (ε₁/ε₂) × Ez1

Substitution:
Tangential: Ex2 = 3 V/m, Ey2 = 4 V/m (unchanged)
Normal (z): Ez2 = (4ε₀ / ε₀) × 5 = 4 × 5 = 20 V/m

Calculation:
E2 = 3ax + 4ay + 20az V/m

Final Answer with units:
E2 = 3ax + 4ay + 20az V/m
(Normal component amplified by ratio ε₁/ε₂ = 4 because D must be continuous)

Exam Tip: Always identify the tangential (parallel to interface) and normal (perpendicular to interface) components separately before applying boundary conditions. A common GATE mistake is applying the wrong condition to the wrong component. Remember: tangential E and normal B are always continuous regardless of interface type. Only Ht and Dn depend on surface current and surface charge.

• Tangential E is always continuous across any interface: Et1 = Et2 (from Faraday's law applied to a shrinking loop).
• Normal B is always continuous across any interface: Bn1 = Bn2 (from Gauss magnetic law, no monopoles).
• Tangential H has a discontinuity equal to surface current: an × (H1 − H2) = Js. For zero surface current, Ht1 = Ht2.
• Normal D has a discontinuity equal to surface charge: Dn1 − Dn2 = ρs. For charge-free interface, ε₁En1 = ε₂En2.
• At a PEC surface: Et = 0, Bn = 0, Dn = ρs, Ht = Js — only normal D and tangential H survive at the PEC boundary.

Quick Revision

• Always continuous: tangential E (Et1 = Et2) and normal B (Bn1 = Bn2).
• Conditionally continuous: tangential H (= if Js = 0) and normal D (= if ρs = 0).
• Normal D relation at charge-free interface: ε₁En1 = ε₂En2 — normal E is NOT continuous if ε₁ ≠ ε₂.
• Normal B relation: μ₁Hn1 = μ₂Hn2 — normal H is NOT continuous if μ₁ ≠ μ₂.
• PEC boundary (outside): Et = 0; Dn = ρs; Bn = 0; an × H = Js.
• Derivation tools: Faraday/Ampere loop for tangential; Gaussian pillbox for normal.
• Exam trap: Normal component of E is discontinuous at a dielectric interface even when ρs = 0, because ε₁En1 = ε₂En2 implies En1 ≠ En2 when ε₁ ≠ ε₂.