Electrical Engineering › Section Y · Time-varying fields, Maxwell's equations, and plane waves › Topic Y.5

The Maxwell-Faraday equation and displacement current

Two laws from the static world, Kirchhoff's voltage law and Ampère's law, break when fields change in time. Repairing them gives two of Maxwell's equations, and the repair to Ampère's law, the displacement current, is the term that makes electromagnetic waves possible.

Kirchhoff's voltage law as a field law

In field terms, the voltage gained going once round a closed path C is the line integral of the electric field along it, \(\oint_C \mathbf{E} \cdot d\mathbf{l}\), where \(d\mathbf{l}\) is a short step along the path. Static fields always give zero, which is Kirchhoff's voltage law:

\oint_C \mathbf{E} \cdot d\mathbf{l} = 0 \quad \text{(static fields)}

But the previous topics showed that a changing flux through a loop induces a voltage round it. So when the magnetic flux varies, the static law must be wrong, and "Kirchhoff's voltage law" holds exactly only for lumped circuits whose loops enclose no changing flux.

The Maxwell-Faraday equation

The fix is to set the voltage round the path equal to Faraday's result. For a path C that does not move, bounding a surface S whose normal follows the right-hand rule from the direction of C:

\oint_C \mathbf{E} \cdot d\mathbf{l} = -\frac{d}{dt}\int_S \mathbf{B} \cdot d\mathbf{A}

This is the integral form of the Maxwell-Faraday equation. The minus sign is Lenz's law, exactly as in the gap-voltage form \(v_T = -d\Phi/dt\). When the flux is constant the right side is zero and Kirchhoff's law returns, so the new equation contains the old one as its static case. The path need not follow any wire.

From a loop to a point: the curl

The integral form describes whole loops. To say what happens at each point, use Stokes's theorem from vector calculus: the circulation of a field round a closed path equals the flux of its curl, \(\nabla \times \mathbf{E}\), through any surface bounded by that path. The curl measures how much a field swirls at a point, as circulation per unit area. Applying Stokes's theorem on the left, moving the time derivative inside the integral on the right, and noting that the result must hold for every possible surface, the two integrands must be equal everywhere:

\nabla \times \mathbf{E} = -\frac{\partial \mathbf{B}}{\partial t}

Wherever the magnetic field changes in time, the electric field swirls round it, point by point.

Ampère's law and the capacitor puzzle

The static Ampère's law, from "Ampere's circuital law and magnetic circuits", says the circulation of \(\mathbf{H}\) round a path equals the current passing through any surface bounded by it: \(\oint_C \mathbf{H} \cdot d\mathbf{l} = I_{enc}\). Now let an ac current charge and discharge a parallel-plate capacitor, and take C as a circle round one lead. A flat surface pierced by the lead encloses the current i. But a surface ballooned out to pass between the plates, still bounded by the same circle, encloses no conduction current at all, because no charge crosses the gap. The same path gives two different answers, so the law is incomplete whenever fields change.

Displacement current

Maxwell's resolution, in the early 1860s, was a new term, the displacement current

I_d = \int_S \frac{\partial \mathbf{D}}{\partial t} \cdot d\mathbf{A}

where \(\mathbf{D} = \varepsilon\mathbf{E}\) is the electric flux density. It has units of amperes, vanishes for steady fields, and depends on the electric field. Between the capacitor's plates the field is \(E = v/d\), so \(D = \varepsilon v/d\), and

I_d = \frac{\varepsilon A}{d}\frac{dv}{dt} = C\frac{dv}{dt}

which is exactly the conduction current i in the leads. Both surfaces now give the same answer. No charge crosses the gap; the name is historical, and \(I_d\) is best read as a changing electric field that acts as a source of magnetic field.

The Ampère-Maxwell law

Adding the new term gives the general law. In integral form, with \(I_c\) the conduction current through S:

\oint_C \mathbf{H} \cdot d\mathbf{l} = I_c + \int_S \frac{\partial \mathbf{D}}{\partial t} \cdot d\mathbf{A}

Writing \(I_c = \int_S \mathbf{J} \cdot d\mathbf{A}\), with \(\mathbf{J}\) the current density in A/m², and applying Stokes's theorem as before gives the point form:

\nabla \times \mathbf{H} = \mathbf{J} + \frac{\partial \mathbf{D}}{\partial t}

The magnetic field swirls round currents and round changing electric fields alike. Paired with the Maxwell-Faraday equation, each field's change in time drives the other's swirl in space, and that mutual drive is what lets a wave carry itself.

Worked example: a capacitor's displacement current

An air capacitor has circular plates of radius 5.0 cm, 1.0 mm apart, and \(v = 10\sin(2\pi \times 10^{6}t)\,\text{V}\). Find the peak displacement current and the peak H on a circle of radius 5.0 cm between the plates. 1. Capacitance: \(C = \varepsilon_0 A/d = 69.5\,\text{pF}\), with \(A = 7.85 \times 10^{-3}\,\text{m}^2\). 2. Peak current:

\begin{aligned} I_d &= \omega C V_m \\ &= 6.28 \times 10^{6} \times 69.5 \times 10^{-12} \times 10 \\ &= 4.37\,\text{mA} \end{aligned}

3. Ignoring fringing, the circle encloses all of it, so by symmetry \(H = I_d/(2\pi r) = 13.9\,\text{mA/m}\). 4. Check another way: \(E_m = 10/0.001 = 10^{4}\,\text{V/m}\), so the peak displacement current density is \(\varepsilon_0\omega E_m = 0.556\,\text{A/m}^2\), and times the area that is 4.37 mA again.

Worked example: when is displacement current negligible?

For a sinusoidal field of peak \(E_m\) in a material of conductivity \(\sigma\), the peak conduction current density is \(\sigma E_m\) and the peak displacement current density is \(\omega\varepsilon E_m\). Compare them in copper, \(\sigma = 5.8 \times 10^{7}\,\text{S/m}\), with \(\varepsilon \approx \varepsilon_0\). 1. The ratio is \(\sigma/(\omega\varepsilon)\), independent of \(E_m\). 2. At 60 Hz it is about \(1.7 \times 10^{16}\). 3. At 1 GHz:

\frac{5.8 \times 10^{7}}{2\pi \times 10^{9} \times 8.854 \times 10^{-12}} = 1.0 \times 10^{9}

4. Conclusion: inside metals displacement current is utterly negligible, which is why circuit theory never needed it. In a vacuum or a perfect insulator, where \(\sigma = 0\), it is the only current there is.

Common mistakes to avoid

First, treating displacement current as moving charge; nothing crosses a capacitor's gap. Second, dropping the minus sign in the Maxwell-Faraday equation, or orienting the surface without the right-hand rule. Third, applying \(\oint \mathbf{E} \cdot d\mathbf{l} = 0\) to a loop that encloses changing flux. Fourth, assuming displacement current matters inside good conductors; it matters in insulators and empty space. Fifth, forgetting that the integral laws must give the same answer for every surface on the same path.

Key takeaways

Kirchhoff's voltage law, \(\oint \mathbf{E} \cdot d\mathbf{l} = 0\), is the static case of the Maxwell-Faraday equation, which sets that circulation equal to minus the rate of change of the enclosed flux; its point form is \(\nabla \times \mathbf{E} = -\partial\mathbf{B}/\partial t\). Ampère's law fails for a charging capacitor until Maxwell's displacement current, \(\int_S \partial\mathbf{D}/\partial t \cdot d\mathbf{A}\), is added; in a capacitor it equals \(C\,dv/dt\). The Ampère-Maxwell law is \(\nabla \times \mathbf{H} = \mathbf{J} + \partial\mathbf{D}/\partial t\). Changing fields drive each other, which makes waves possible.

Flashcards

Flashcard

∮ E · dl

Voltage gained once round a closed path

Zero for static fields: KVL.

Test your understanding

Question 1 of 5

A 100 nF capacitor's voltage is rising at 10,000 V/s. What is the displacement current between its plates?

Question 2 of 5

Which equation is the point form of the Maxwell-Faraday equation?

Question 3 of 5

Why did the static form of Ampère's law fail for a charging capacitor?

Question 4 of 5

A 1 GHz field of peak 100 V/m exists in a vacuum. What is the peak displacement current density?

Question 5 of 5

Inside a copper wire at 60 Hz, how does displacement current compare with conduction current?

← Previous topicNext topic →