The Electric Displacement Field D

An auxiliary field for separating free charge from the response of matter.

In electrostatics, we usually calculate the Electric Field (\(\vec{E}\)) using Gauss’s Law. But as soon as we introduce matter, the field and the material response become coupled. Direct calculation becomes hard because the bound charge depends on the field we are trying to find.

1. The Problem: The Polarization Loop

To find \(\vec{E}\), you need to know the total charge density \(\rho_{total}\). However, in a dielectric, \(\rho_{total}\) includes both free charge (\(\rho_f\))—the stuff we control—and bound charge (\(\rho_b\))—the stuff that happens inside the atom.

                The Feedback Loop
                You bring in some external Free Charge.
This charge creates an Electric Field \(\vec{E}\).
This field polarizes the material (\(\vec{P} = \alpha \vec{E}\)).
The Polarization creates Bound Charge (\(\rho_b = -\nabla
                        \cdot \vec{P}\)).
This bound charge creates its own electric field, which modifies the total \(\vec{E}\).
                    

            

You can't find \(\vec{E}\) without \(\rho_b\), but you can't find \(\rho_b\) without \(\vec{E}\). They are locked in a circular dependency.

Visualizing the Difference: Notice how the Displacement field D depends only on the free charges (on the plates), while the Total Field E is reduced by the presence of bound charges on the dielectric surfaces.

2. The Solution: Isolate the Free Charge

We need a field that only "sees" the charges we actually control: the free charges. Let's start with the differential form of Gauss's Law:

\epsilon_0 \nabla \cdot \vec{E} = \rho_{total} = \rho_f + \rho_b

We know that \(\rho_b = -\nabla \cdot \vec{P}\). Substituting this in:

\epsilon_0 \nabla \cdot \vec{E} = \rho_f - \nabla \cdot \vec{P}

Now, let's group all the divergence terms on one side:

\nabla \cdot (\epsilon_0 \vec{E} + \vec{P}) = \rho_f

We define the quantity in the parentheses as the Electric Displacement, \(\vec{D}\).

The Displacement Vector

\vec{D} \equiv \epsilon_0 \vec{E} + \vec{P}

With this definition, Gauss's Law for \(\vec{D}\) depends only on free charge:

\nabla \cdot \vec{D} = \rho_f \quad \longleftrightarrow \quad \oint \vec{D} \cdot d\vec{a} = Q_{f,enc}

3. Physical Intuition

If \(\vec{E}\) is the "Total Field" (caused by everything), then \(\vec{D}\) is the "Administrative Field." It tracks the distribution of free charge without having to worry about how the medium is reacting internally.

In a vacuum, \(\vec{P} = 0\), so \(\vec{D} = \epsilon_0 \vec{E}\). In matter, \(\vec{D}\) "displaces" or shifts the responsibility of the bound charges into the definition of the field itself.

4. A Deceptive Parallel

Since \(\nabla \cdot \vec{D} = \rho_f\) looks like Gauss's Law, it is tempting to think \(\vec{D}\) is just "The E-field of free charges." This is wrong.

⚠️ The Trap: Non-Zero Curl

The difference between \(\vec{E}\) and \(\vec{D}\) is the curl:

\nabla \times \vec{D} = \nabla \times \vec{P}

Because matter can have "swirling" polarization, the curl of \(\vec{D}\) is not always zero. This means you cannot usually define a potential for \(\vec{D}\), and you cannot use Coulomb's simple inverse-square law for it.

Deep Dive: Wire with Insulation

5. Boundary Conditions

Because \(\vec{D}\) only cares about free charge, its behavior at boundaries is very helpful for engineering:

Normal Component

\( D^\perp_{above} - D^\perp_{below} = \sigma_f \)

Only Free Surface Charge causes a jump in the perpendicular component.

Parallel Component

\( \vec{D}^\parallel_{above} - \vec{D}^\parallel_{below} = \vec{P}^\parallel_{above} - \vec{P}^\parallel_{below} \)

Discontinuous by the jump in polarization. (Recall that \(E^\parallel\) is always continuous).