Ampere-Maxwell Law: ∇×H = J + ∂D/∂t

Why Maxwell Modified Ampère's Law

The symmetry problem: Faraday's law versus Ampère's law

Before Maxwell, Ampère's law stated that a magnetic field circulates around any current: ∇×H = J. Faraday's law, meanwhile, said that a changing magnetic field induces a circulating electric field: ∇×E = −∂B/∂t. These two laws look asymmetric. Faraday's law has a time-derivative term driving the curl of E, but Ampère's law has no equivalent time-derivative term driving the curl of H. In the mid-1800s this asymmetry was more than aesthetic — it produced a contradiction.

Apply Ampère's law to a circuit charging a capacitor. Draw an Amperian loop around the wire feeding the capacitor plate, then try to close the surface bounded by that loop two different ways: one surface slices through the wire (where current J flows), the other bulges between the capacitor plates (where no conduction current flows at all, just a changing electric field). Ampère's original law gives two different answers for the same loop depending on which surface you pick. That's not allowed — a physical law can't depend on an arbitrary choice of surface.

Displacement current: the missing term

Maxwell resolved this in 1861 by proposing a new term: the displacement current density, ∂D/∂t, where D is the electric flux density (electric displacement field). Adding it gives the complete Ampere-Maxwell law:

∇×H = J + ∂D/∂t

In the capacitor-gap surface, J is zero but the electric field between the plates is changing rapidly as charge accumulates, so ∂D/∂t carries the current continuity across the gap. Both surface choices now give the same result. The fix isn't a patch — it's a statement that a time-varying electric field produces a magnetic field just as surely as a conduction current does.

Predicting electromagnetic wave propagation

The displacement current term did more than patch a bookkeeping error. Combined with Faraday's law, it predicted that electric and magnetic fields could sustain each other in empty space, with no wires or charges required. Maxwell computed the propagation speed of this disturbance from purely electrical and magnetic constants and got a number matching the measured speed of light. That was the moment electromagnetism and optics merged into one theory, and it is the reason every RF and wireless system you design ultimately traces back to this one added term.

Mathematical Form and Physical Interpretation

The complete equation: curl of H equals conduction current plus displacement current

∇×H = J + ∂D/∂t says that the circulating magnetic field around any point is sourced by two physically distinct mechanisms: actual moving charge (J) and a changing electric flux density (∂D/∂t). An engineer can read this as a current-continuity statement generalized to include "current" that flows through dielectrics without any charge carriers crossing the gap.

J: conduction current density (free charges)

J is the current density you already compute with Ohm's law — amperes per square meter of free charge motion in a conductor, a doped semiconductor, or a lossy dielectric. It's the only term present in DC circuit analysis, and it's where resistive heating originates.

∂D/∂t: displacement current density (time-varying electric field)

∂D/∂t has units of A/m² as well, but no charge physically crosses the region where it's evaluated. It represents the rate of change of electric flux density. In a vacuum or an ideal dielectric, this is the entire right-hand side — there is no J to speak of, yet ∇×H is nonzero whenever the field is changing.

Why both terms matter in RF circuits and free space

At DC, ∂D/∂t vanishes and you recover plain Ampère's law — which is why DC intuition (current flows only where there's a conductive path) fails at RF. At high frequency, a capacitor "passes" current precisely because displacement current closes the loop through the dielectric. In free space, J is exactly zero, and ∂D/∂t alone sustains propagating waves. Every RF design decision about coupling, isolation, and radiation hinges on correctly weighting these two terms.

Displacement Current in Dielectric Media

D and its relationship to E and material permittivity

For a linear, isotropic dielectric, D = εE, where ε = ε₀εᵣ is the material's permittivity. This constitutive relation, along with B = µH and J = σE, is what lets you convert the abstract field quantities in Maxwell's equations into measurable circuit behavior — the full set is covered in Constitutive Relations: D = εE, B = µH, J = σE. Substituting D = εE into the displacement current term gives ∂D/∂t = ε∂E/∂t: the displacement current density is proportional to how fast the electric field is changing and to the material's permittivity.

How displacement current dominates in capacitors and low-loss dielectrics

In an FR-4 or Rogers substrate with negligible conductivity, σ is small enough that J = σE contributes almost nothing compared to ε∂E/∂t at RF frequencies. This is exactly why a capacitor, or a gap in a microstrip trace, behaves as a low-impedance path at high frequency — the "current" through the dielectric is entirely displacement current.

Physical mechanism: polarization lag and field acceleration

Physically, ∂D/∂t reflects the motion of bound charge: as E oscillates, the dielectric's internal dipoles reorient (polarize) in response. That reorientation itself constitutes a tiny, bound current even though no charge crosses the dielectric volume. At high enough frequency the dipoles can't keep up, producing a phase lag that shows up macroscopically as dielectric loss (captured in the loss tangent, tan δ).

Quantifying with constitutive relations

Worked example: a parallel-plate capacitor with plate area 1 cm², separation 0.2 mm, filled with a dielectric of εᵣ = 4.3 (typical FR-4), driven at 1 GHz with a 1 V amplitude sine wave.

E = V/d = 1 V / 0.2 mm = 5000 V/m (peak) D = εᵣε₀E = 4.3 × 8.854×10⁻¹² × 5000 ≈ 1.90×10⁻⁷ C/m² ∂D/∂t (peak) = ωD = 2π×10⁹ × 1.90×10⁻⁷ ≈ 1.20×10³ A/m²

Multiply by plate area (1×10⁻⁴ m²) to get a displacement current of roughly 0.12 A — entirely due to the changing field, with zero free charge crossing the dielectric.

Conduction Current and Lossy Media

Ohmic loss: J = σE in conductors and lossy dielectrics

In a conductor, J = σE dominates overwhelmingly because σ for copper (~5.8×10⁷ S/m) is enormous compared to ωε for any realistic dielectric at RF frequencies. This is why trace conductors are modeled with conduction current and insulating substrates are modeled with displacement current — the same equation, two completely different regimes.

How conductivity affects the relative magnitudes of J and ∂D/∂t

The ratio |J|/|∂D/∂t| = σ/(ωε) is the standard "loss tangent" proxy used to classify a material as a good conductor or good dielectric at a given frequency. When σ/(ωε) >> 1, the medium behaves as a conductor; when it's << 1, it behaves as a lossless dielectric. Seawater, for instance, shifts from conductor-like to dielectric-like behavior as frequency climbs from a few MHz into the GHz range — a fact that matters directly in maritime and ground-penetrating radar links.

Skin effect and attenuation in RF transmission lines

In copper traces, the dominance of J = σE combined with finite conductivity gives rise to the skin effect: current crowds into a surface layer of depth δ = √(2/(ωµσ)), increasing effective resistance with frequency. That resistance, not the displacement current in the metal, is what you model when you compute conductor loss in a transmission line.

EM Wave Generation and Propagation

How Ampere-Maxwell law couples with Faraday's law to form wave equations

Taking the curl of Faraday's law (∇×E = −∂B/∂t) and substituting the Ampere-Maxwell law into it eliminates H and produces a second-order wave equation purely in E (and a mirror-image one in H). The two laws are inseparable partners: Faraday's law says a changing B drives a circulating E; Ampere-Maxwell says a changing E (via ∂D/∂t) drives a circulating H. Each field's time variation regenerates the other.

Condition for self-sustaining waves: both ∂D/∂t and ∂B/∂t present

This is the crux: without the ∂D/∂t term, a changing E could not regenerate H, the loop would break, and no self-propagating wave could exist in source-free space — you'd be left with static or quasi-static fields that don't travel. Both time-derivative terms must be present and nonzero for a wave solution to emerge from the two curl equations. Remove either one and propagation stops.

Wave speed prediction: c = 1/√(με)

Solving the wave equation gives a propagation speed v = 1/√(µε). In vacuum this evaluates to the speed of light, c ≈ 3×10⁸ m/s. In a dielectric with εᵣ = 4.3, the speed drops to c/√4.3 ≈ 1.45×10⁸ m/s — directly explaining why microstrip traces have an effective dielectric constant and electrical length shorter than their free-space wavelength would suggest.

Practical implications for antenna design and waveguides

Antennas work because an oscillating current (J) at the feed point launches a changing E field into space, which (via Ampere-Maxwell) generates a changing H field, which (via Faraday) regenerates E farther out — a self-propagating handoff that carries energy away from the structure. Waveguide and cavity design relies on the same coupling to establish standing-wave mode patterns bounded by conducting walls.

Application to RF Circuit Design

When displacement current dominates: capacitive coupling and PCB traces

Any place you intentionally or unintentionally have two conductors separated by a dielectric — coupled lines, via stubs, decoupling capacitors, or a trace running near a ground plane — displacement current is the mechanism carrying signal across the gap. At high frequency, "isolated" traces are never truly isolated; capacitive coupling via ∂D/∂t can dominate crosstalk budgets.

When conduction current dominates: resistive losses in conductors

Inside the copper itself, J = σE governs, and conductor loss (I²R-type heating, modified by skin effect) is the term you optimize against with wider traces, thicker plating, or smoother copper roughness profiles.

Frequency dependency: displacement current grows with frequency

Because ∂D/∂t scales with ω, displacement effects that are negligible at 60 Hz power-line frequencies become dominant at GHz RF frequencies. A via stub or an unused component pad that looks like an open circuit at low frequency can become a significant capacitive coupling path or resonant stub at the frequencies your design actually operates at.

Design trade-offs in transmission line routing

Trace spacing, dielectric thickness, and copper conductivity all trade off between minimizing conduction loss and controlling unwanted displacement-current coupling. Tight trace spacing reduces board area but increases capacitive crosstalk; thicker dielectric reduces coupling but increases characteristic impedance variation and via parasitic length. These trade-offs are exactly why transmission-line geometry (not just material selection) is a first-order design variable at RF.

Connecting to EM Simulation and Measurement

How EM solvers enforce Ampere-Maxwell law in numerical form

Full-wave solvers (FDTD, FEM, MoM) discretize Maxwell's equations directly, enforcing ∇×H = J + ∂D/∂t at every mesh node or cell face on every time step or frequency point. The displacement current term is what allows these solvers to compute radiation and propagation in source-free regions — without it, the solver would only ever predict static or quasi-static field distributions tied to explicit conductors.

Role in predicting S-parameters and impedance discontinuities

S-parameter extraction depends on the solver correctly apportioning energy between conduction loss (resistive, showing up as insertion loss) and displacement-current-driven coupling or radiation (showing up as return loss, crosstalk, or far-field emission). An impedance discontinuity — a via, a connector transition, a trace width step — is, physically, a local mismatch in how J and ∂D/∂t are balanced, which the solver resolves by computing local field redistribution.

Why violation of this law leads to unphysical solutions

If a meshing error, boundary condition, or software bug breaks current continuity — i.e., makes ∇·(J + ∂D/∂t) nonzero where it shouldn't be — the solver will produce results that violate charge conservation: energy appearing or vanishing at nodes, spurious resonances, or non-convergent S-parameters that don't satisfy passivity. When your simulation results look physically implausible (gain where there should be loss, S-parameters exceeding 0 dB at a passive port), checking whether the mesh properly enforces the Ampere-Maxwell current-continuity condition is one of the first debugging steps, alongside verifying your constitutive relations and port definitions.