Gauss's law

Find a field from its symmetry.

The last lesson found the field of a few charges by adding their arrows. A charged ball holds billions of billions of charges, and adding that many arrows means a hard integral. Gauss's law is the same physics said another way, and for anything with enough symmetry it gives the field in one line. It is also the first of Maxwell's four equations.

Field lines, drawn as dots, pour out of positive charges and into negative ones: 8 lines for every nanocoulomb. The dashed amber curve is an imaginary closed surface. Drag its amber point to move it and its white point to resize it, and drag any charge. A mark shows each place a line crosses the surface: amber going out, blue coming in. This is a flat world, where lines spread only across the page. See the note further down for what that leaves out.
  • It starts with one charge of \(+2\) nC inside the surface. The tally reads 16 lines out and none in.
  • Change the surface to an oval, then to a lumpy shape. Still 16.
  • Drag the white point to make the surface bigger. The marks spread out, and the count stays at 16. Shrink it again, and the count holds for as long as the charge is still inside.
  • Drag the amber point until the charge is left outside. Lines now come in on one side and go out on the other, and the net count is 0.
  • Choose A + and a −, and put the surface back to a circle. It starts round the positive charge: \(+16\). Drag it over the negative one: \(-16\). Now stretch it until it holds both. Lines still cross it, but the net is 0.
  • Choose Three charges. Before you move anything, work out what the net count should be for a surface holding the \(+2\) and the \(-1\), and then for one holding all three. Then try them.
  • Put a charge inside the surface and drag it about in there. The marks slide round, and the count doesn't move until the charge crosses the surface.

What the count is measuring

The amount of field passing through a surface is its flux, from line and surface integrals: cut the surface into small patches, multiply the part of the field that goes squarely through each patch by the patch's area, and add up. For the electric field it is written

\[ \Phi = \oint \mathbf{E}\cdot d\mathbf{A} \]

where \(d\mathbf{A}\) is a patch's area vector, pointing outwards, and the ring means the surface is closed, like a balloon. Field coming out counts as positive and field going in as negative.

Field lines are a way of drawing flux. If each line stands for a fixed amount, then counting lines out minus lines in measures the flux. That's what the tally in the figure does.

Now work it out for the simplest case: one point charge \(q\), and a sphere of radius \(r\) centred on it.

From the last lesson, the field at distance \(r\) has strength \(\frac{kq}{r^2}\) and points straight away from the charge. So it crosses the sphere squarely everywhere, with the same strength everywhere. The flux is just strength times area.

\[ \Phi = \frac{kq}{r^2} \times 4\pi r^2 \]

The \(r^2\) cancels. A bigger sphere meets a weaker field, but has more area by exactly the same factor.

\[ \Phi = 4\pi k q \]

That \(4\pi k\) turns up so often that it gets a name of its own. Define a new constant \(\varepsilon_0\) (epsilon nought) by \(k = \frac{1}{4\pi\varepsilon_0}\). Then \(4\pi k = \frac{1}{\varepsilon_0}\).

\[ \Phi = \frac{q}{\varepsilon_0} \qquad\qquad \varepsilon_0 \approx 8.85 \times 10^{-12} \text{ C}^2/(\text{N m}^2) \]

So the flux through the sphere depends on the charge and nothing else. The figure showed that the same holds for any closed surface, and the last chapter says why. Away from the charge itself, this field has no divergence: it's the three-dimensional version of the source in divergence and curl, whose arrows spread but pass on exactly what they receive. So by the divergence theorem, no flux is gained or lost between the sphere and any other surface wrapped round the same charge. And a charge outside a surface adds nothing, because all the flux it sends in has to come out again.

Fields from several charges add, so their fluxes add too. That gives the law.

\[ \oint \mathbf{E}\cdot d\mathbf{A} = \frac{Q_{\text{inside}}}{\varepsilon_0} \]

The flux of the electric field out of any closed surface is the total charge inside it, divided by \(\varepsilon_0\). The surface can be any size and any shape. It doesn't have to be a real object: you imagine it, wherever it's useful. For the figure's 2 nC, the flux is \(\frac{2 \times 10^{-9}}{8.85 \times 10^{-12}} \approx 226\) N m²/C through every surface that holds it.

For charges that aren't moving, Gauss's law and Coulomb's law are two ways of saying the same thing. You've just got Gauss's law from Coulomb's, and with the symmetry of a ball, the working below gives Coulomb's law back.

What the flat picture leaves out

In the figure, each charge's lines spread only across the page, so they thin out in proportion to the distance. In that flat world the field falls as \(\frac{1}{r}\), and the "surface" is a closed curve. Real field lines spread in every direction, up out of the page as well, so they thin out as \(\frac{1}{r^2}\), and a real Gauss surface is a closed skin with an inside, like a balloon. The circle stands for a sphere.

The counting rule works the same way in both worlds, because all it needs is that lines begin and end only on charges. But a flat drawing can't show real lines honestly: most of them leave the page. So treat the figure as a model of the rule, not as a picture of a real field.

The same law at a point

The divergence theorem turns the flux on the left into the divergence of \(\mathbf{E}\) added up over the volume inside. The charge on the right is the charge density \(\rho\) (rho, the charge in each cubic metre) added up over the same volume. Those two sums agree for every volume you could choose, however small, and that can only happen if the things being added agree at each point:

\[ \nabla\cdot\mathbf{E} = \frac{\rho}{\varepsilon_0} \]

Charge is the source of the electric field. Where there's positive charge, field is being made. Where there's none, whatever comes in goes out.

Letting symmetry do the work

Gauss's law is always true, but on its own it only gives you a total. To get the field itself, you need to know two things before you start: which way the field points, and a surface over which its strength is the same everywhere. Symmetry can tell you both. Then the flux is simply \(E\) times an area, and one division finishes the job.

Three charged objects, each with the surface that suits it, drawn to scale in centimetres. Drag the amber point to change the surface. The white arrows are the field on the surface. They get longer where the field is stronger, but not in proportion. The drifting specks only show which way the field points. The graph is the field's strength against distance. Each picture is a flat slice through something solid: the dashed ovals are there to suggest the sphere and the cans.
  • A ball. The surface starts 8 cm from the centre, where the field is 5,620 N/C. Drag it in to 4 cm, the edge of the ball. It holds all 4 nC the whole way, but its area shrinks to a quarter, so the field is four times stronger: 22,500 N/C.
  • Carry on inside the ball, to 2 cm. It holds only an eighth of the charge. The field is half its value at the edge (the screen shows 11,200 N/C, because the 22,500 was itself rounded up, from 22,480), and it carries on falling to nothing at the centre.
  • A long line. At 6 cm the field is 12,000 N/C. Double the distance to 12 cm and it halves (the screen shows 5,990, because the 12,000 was itself rounded up, from 11,989). It doesn't fall to a quarter, as it would for a ball.
  • A flat sheet. Drag the ends of the can in and out. The field is 22,600 N/C wherever they are.

Here is the working for each, with letters in place of the figure's numbers.

A ball of charge. A ball of radius \(a\) holds a charge \(Q\) spread evenly through it. Nothing about a ball picks out one direction over another, so the field must point straight out from the centre, and its strength can depend only on the distance \(r\) from the centre. A sphere of radius \(r\) is the surface that fits.

The field crosses the sphere squarely and has the same strength \(E\) all over it, so the flux is \(E\) times the sphere's area.

\[ \oint \mathbf{E}\cdot d\mathbf{A} = E \times 4\pi r^2 \]

Outside the ball, the sphere holds all of \(Q\). Set the flux equal to \(\frac{Q}{\varepsilon_0}\) and divide by the area.

\[ E \times 4\pi r^2 = \frac{Q}{\varepsilon_0} \qquad \Rightarrow\quad E = \frac{Q}{4\pi\varepsilon_0 r^2} = \frac{kQ}{r^2} \]

That's exactly the field of a point charge \(Q\) at the centre. From outside, you can't tell a ball from a point.

Inside the ball, the sphere holds only part of the charge. Volume goes as the radius cubed, so the share is \(\frac{r^3}{a^3}\).

\[ E \times 4\pi r^2 = \frac{Q}{\varepsilon_0}\cdot\frac{r^3}{a^3} \qquad \Rightarrow\quad E = \frac{Q\,r}{4\pi\varepsilon_0 a^3} = \frac{kQ\,r}{a^3} \]

So inside, the field grows in step with \(r\), and the charge further out than you are adds nothing at all.

A long line of charge. Let it carry \(\lambda\) (lambda) coulombs on every metre. Turn the line about its own length, or slide it along itself, and it looks the same. So the field points straight out from the line and depends only on the distance \(r\) from it. The surface that fits is a can of radius \(r\) and length \(L\) with the line along its axis.

The field skims along the can's flat ends without crossing them, so they add nothing. It crosses the curved wall squarely, with the same strength everywhere. The wall's area is its circumference times its length.

\[ \oint \mathbf{E}\cdot d\mathbf{A} = E \times 2\pi r L \]

The can holds a length \(L\) of the line, so the charge inside is \(\lambda L\).

\[ E \times 2\pi r L = \frac{\lambda L}{\varepsilon_0} \]

The can's length cancels, as it must, since you made the can up.

\[ E = \frac{\lambda}{2\pi\varepsilon_0 r} \]

A flat sheet of charge. Let it carry \(\sigma\) (sigma) coulombs on every square metre. Slide a wide flat sheet sideways and it looks the same, so the field points straight away from it on both sides, and can depend only on the distance from it. The surface that fits is a can pushed through the sheet, with a flat end of area \(A\) the same distance out on each side.

This time nothing crosses the curved wall. The field goes squarely out through both flat ends.

\[ \oint \mathbf{E}\cdot d\mathbf{A} = E \times 2A \]

The can cuts a patch of area \(A\) out of the sheet, so the charge inside is \(\sigma A\).

\[ E \times 2A = \frac{\sigma A}{\varepsilon_0} \quad\Rightarrow\quad E = \frac{\sigma}{2\varepsilon_0} \]

No distance appears anywhere. The field of a wide flat sheet is the same strength however far from it you are.

\[ E_{\text{ball, outside}} = \frac{Q}{4\pi\varepsilon_0 r^2} \qquad\qquad E_{\text{line}} = \frac{\lambda}{2\pi\varepsilon_0 r} \qquad\qquad E_{\text{sheet}} = \frac{\sigma}{2\varepsilon_0} \]

The three go together. Lines from a ball spread in all three directions, and the field falls as \(\frac{1}{r^2}\). Lines from a long line can only spread in two, and it falls as \(\frac{1}{r}\). Lines from a sheet have nowhere to spread at all, and it doesn't fall.

No real line or sheet goes on for ever. The results hold when you're much closer to the object than to its ends or edges. And without symmetry the method fails: Gauss's law is still true for a charged teapot, but there's no surface on which \(E\) is constant, so the law gives you a total and nothing more.

The ball result is why gravity could treat a planet as a point. Gravity is an inverse-square law too, so everything here carries over, and a ball-shaped planet pulls on things outside it as if all its mass were at its centre. Newton had to work hard to prove that. With Gauss's law it takes two lines.

The sheet result builds the capacitor: two flat plates with opposite charges, facing each other. Between them the two fields point the same way and add up to a steady \(\frac{\sigma}{\varepsilon_0}\). Outside they cancel. A steady field in a known space is the simplest place to study voltage, which comes next.

And unlike Coulomb's law, Gauss's law stays true when charges move, however fast. That's why it, and not Coulomb's law, is one of Maxwell's equations.

Problems

A closed surface has three charges inside it: \(+3\) nC, \(-5\) nC and \(+4\) nC. A fourth charge of \(+10\) nC sits just outside. What is the flux of the electric field through the surface? Use \(\varepsilon_0 = 8.85 \times 10^{-12}\) C²/(N m²).

Answer

Only the charge inside counts: \(3 - 5 + 4 = 2\) nC. The \(+10\) nC outside changes the field at every point of the surface, but everything it sends in comes out again.

\(\Phi = \frac{2 \times 10^{-9}}{8.85 \times 10^{-12}} \approx 226\) N m²/C.

A point charge of 6 nC sits at the exact centre of a cube. What is the flux through the whole cube? What is the flux through one of its faces?

Answer

The whole cube: \(\frac{6 \times 10^{-9}}{8.85 \times 10^{-12}} \approx 678\) N m²/C. The shape and size of the surface make no difference.

The charge is at the centre, so the six faces are all placed alike, and each must get the same share: \(\frac{678}{6} = 113\) N m²/C. Finding that by adding up the field over a flat square would be a hard integral.

Use Gauss's law to find the field 0.50 m from a point charge of 2.0 nC. Then check it with \(E = \frac{kq}{r^2}\) and \(k = 8.99 \times 10^9\) N m²/C².

Answer

Take a sphere of radius 0.50 m round the charge. Its area is \(4\pi \times 0.50^2 \approx 3.14\) m². The flux through it is \(\frac{2.0 \times 10^{-9}}{8.85 \times 10^{-12}} \approx 226\) N m²/C, spread evenly, so \(E = \frac{226}{3.14} \approx 72.0\) N/C.

Coulomb's way: \(\frac{8.99 \times 10^9 \times 2.0 \times 10^{-9}}{0.50^2} \approx 71.9\) N/C. The two agree but for the rounding along the way.

The ball in the figure holds 4.0 nC spread evenly through a radius of 4.0 cm. Find the field 8.0 cm from its centre, at its surface, and 2.0 cm from its centre. Use \(k = 8.99 \times 10^9\) N m²/C², and check all three on the figure.

Answer

At 8.0 cm you're outside, so the ball acts as a point charge: \(E = \frac{8.99 \times 10^9 \times 4.0 \times 10^{-9}}{0.080^2} = \frac{35.96}{0.0064} \approx 5{,}620\) N/C.

At the surface, \(r = 0.040\) m: \(\frac{35.96}{0.0016} \approx 22{,}500\) N/C.

At 2.0 cm you're inside. A sphere of that radius holds \(\left(\frac{2.0}{4.0}\right)^3 = \frac{1}{8}\) of the charge, which is 0.50 nC, and it acts as a point charge at the centre: \(\frac{8.99 \times 10^9 \times 0.50 \times 10^{-9}}{0.020^2} = \frac{4.495}{0.0004} \approx 11{,}200\) N/C.

A long straight wire carries 40 nC on every metre. Find the field 6.0 cm from it. How far from the wire has the field fallen to half of that?

Answer

\(E = \frac{\lambda}{2\pi\varepsilon_0 r} = \frac{40 \times 10^{-9}}{2\pi \times 8.85 \times 10^{-12} \times 0.060} = \frac{40 \times 10^{-9}}{3.34 \times 10^{-12}} \approx 12{,}000\) N/C.

The field goes as \(\frac{1}{r}\), so it halves when the distance doubles: at 12 cm.

Two wide flat sheets face each other a short distance apart. One carries \(+400\) nC on every square metre and the other \(-400\) nC. Find the field in the gap between them, and the field outside the pair.

Answer

Each sheet on its own makes a field of \(\frac{\sigma}{2\varepsilon_0} = \frac{400 \times 10^{-9}}{2 \times 8.85 \times 10^{-12}} \approx 22{,}600\) N/C on both of its sides, the same at any distance. The positive sheet's field points away from it, and the negative sheet's points towards it.

In the gap, both of those are the same direction: from the positive sheet to the negative one. They add up to about \(45{,}200\) N/C, which is \(\frac{\sigma}{\varepsilon_0}\).

Outside the pair, on either side, the two fields point opposite ways and are equally strong, so they cancel: zero. All the field is shut in the gap. This is a capacitor.

Learn more

An unhandled error has occurred. Reload 🗙