AP Physics C: E&M · Unit 8 lecture · Part 6 of 11 · 11 min

8.4 Electric fields of charge distributionsRods and rings: chop, cancel, integrate

Mr. Thompkins
Real charge doesn't sit on dots. It's smeared along wires and over surfaces. This is the section that makes Physics C a calculus course. The idea is simple: chop the charge into tiny pieces, treat each piece as a point charge, and let an integral add them up.

First, describe how charge is spread out with a density. For charge along a line, the linear charge density is $\lambda$ (coulombs per meter), and a small length $dx$ holds $dq = \lambda\,dx$. For charge on a surface, $\sigma$ is charge per area and $dq = \sigma\,dA$. For charge through a volume, $\rho$ is charge per volume and $dq = \rho\,dV$. If the charge is spread uniformly, the density is just total charge divided by total length, area or volume.

The recipe for the field of a distribution:

  1. Draw the object and the point P. Pick one small piece $dq$ at a general position, not at a special spot.
  2. Write the field of that piece: $dE = \frac{k\,dq}{r^2}$, where $r$ is the distance from the piece to P, written in terms of your variable.
  3. Draw $d\vec E$ at P. Use symmetry: does a partner piece cancel some component? Keep only the components that survive.
  4. Write $dq$ in terms of the variable ($\lambda\,dx$, $\lambda R\,d\theta$), set the limits, and integrate.
  5. Check: units, and a limit. Far away, every distribution must look like a point charge, $kQ/r^2$.
?Wait, how do I integrate $1/x^2$?
$\int x^{-2}\,dx = -x^{-1} + C$. So $\int_a^b \frac{dx}{x^2} = \frac{1}{a} - \frac{1}{b}$. You'll use this one constantly, along with $\int \frac{dx}{x} = \ln x$.

A rod, at a point on its axis

Put the point P at the origin and a rod of length L and uniform density $\lambda$ along the x-axis from $x = d$ to $x = d + L$. A piece at position x is a distance x from P. Every piece pushes P in the same direction (for a positive rod, in $-x$), so there's nothing to cancel, and the sizes simply add.

P$x = 0$$x = d$$x = d + L$$dq = \lambda\,dx$distance $x$$d\vec E$every piece pushes P the same way: no cancelling
Putting P at the origin makes the distance to each piece just x. All the dE vectors point the same way.

$$E = \int_d^{d+L} \frac{k\lambda\,dx}{x^2} = k\lambda\left[\frac{1}{d} - \frac{1}{d+L}\right] = \frac{k\lambda L}{d(d+L)} = \frac{kQ}{d(d+L)}$$

Check the limit. If P is far away ($d \gg L$), then $d(d+L) \approx d^2$ and $E \approx kQ/d^2$: a point charge. The physics passes the test.

00.10.20.30.40.50.6050010001500200025003000$d$ (m)$E$ (N/C)rodpoint charge at center1200 N/C
Field beyond the end of a 0.20 m rod holding 4.0 nC, compared with a point charge at the rod's center. Close up, the rod is stronger: its near end is very close. Far away, the two curves merge.
Pause and think
For the rod above, Q = 4.0 nC, L = 0.20 m and d = 0.10 m. A student models the rod as a point charge at its center, 0.20 m from P, and gets 900 N/C. Is the true field bigger or smaller?
I thought about it. Show me
Bigger: 1200 N/C. Because of the $1/r^2$, the near half of the rod contributes much more than the far half loses. Averaging the distance underestimates the field. That's why you need the integral.

A ring, at a point on its axis

Now a ring of radius $a$ with total charge Q, and P on its axis a distance z from the center. Every piece is the same distance $r = \sqrt{z^2 + a^2}$ from P. Each $d\vec E$ points away from its piece, tilted. The piece on the opposite side of the ring has a $d\vec E$ whose sideways part points the opposite way. So all sideways parts cancel, and only the parts along the axis survive.

P$d\vec E$$d\vec E$$\vec E$$z$$a$$r = \sqrt{z^2 + a^2}$ring, edge onsideways parts cancel
Pieces on opposite sides of the ring cancel each other's sideways fields. The axial parts all add.
?Wait, how do I get the axial component?
The component of $d\vec E$ along the axis is $dE\cos\theta$, where θ is the angle between $d\vec E$ and the axis. From the right triangle with legs z and a and hypotenuse r, $\cos\theta = z/r$.

$$E = \int \frac{k\,dq}{r^2}\cdot\frac{z}{r} = \frac{kz}{(z^2+a^2)^{3/2}}\int dq = \frac{kQz}{(z^2 + a^2)^{3/2}}$$

Here z and a are the same for every piece, so they come out of the integral, and $\int dq$ is just Q. Check the limits: at the center (z = 0), E = 0, because every piece is canceled by the one across from it. Far away ($z \gg a$), $E \to kQ/z^2$, a point charge again. In between the field rises to a maximum (at $z = a/\sqrt2$, found by setting $dE/dz = 0$) and falls off.

00.10.20.30.40500100015002000$z$ (m)$E$ (N/C)max at $z = a/\sqrt2$1591 N/Czero at the center
On-axis field of a ring with Q = 5.0 nC and a = 0.10 m. Zero at the center, a peak of about 1730 N/C, then a point-charge falloff.
Ring on its axis
A ring of radius 0.10 m holds 5.0 nC spread uniformly. Find the field on the axis 0.10 m from the center, and compare it with a point charge of 5.0 nC at the same distance.
  1. Start from the derived result: $E = \frac{kQz}{(z^2 + a^2)^{3/2}}$.
    Why?
    On an FRQ that asks you to derive it, start one step earlier, from dE = k dq / r², and show the cos θ = z/r and the cancellation argument.
  2. $z^2 + a^2 = 0.010 + 0.010 = 0.020\ \text{m}^2$, and $(0.020)^{3/2} = 2.83 \times 10^{-3}\ \text{m}^3$.
  3. $E = \frac{(9.0 \times 10^9)(5.0 \times 10^{-9})(0.10)}{2.83 \times 10^{-3}} = 1.59 \times 10^3$ N/C, along the axis, away from the ring.
  4. Point charge: $kQ/z^2 = \frac{(9.0 \times 10^9)(5.0 \times 10^{-9})}{0.010} = 4500$ N/C. The ring's field is much smaller.
    Why?
    The ring's charge is farther away (0.141 m, not 0.10 m) and tilted, so much of each piece's field cancels.

Answer: About $1.6 \times 10^3$ N/C along the axis, about one third of the point-charge value.

Stuck? Get a hint
  1. Every piece of the ring is at the same distance r = √(z² + a²).
  2. Only the axial component survives: multiply by cos θ = z/r.
  3. z and a are constants, so ∫dq = Q.
Check yourself
A ring of charge Q and radius a. Where on its axis is the field zero?
  1. AAt z = a
  2. BAt the center, z = 0
  3. CAt $z = a/\sqrt{2}$
  4. DNowhere
Show the answer

B. At the center, each piece's field is canceled by the piece directly across. $z = a/\sqrt2$ is where the field is largest, a popular trap.

  • Densities: $dq = \lambda\,dx$ (line), $\sigma\,dA$ (surface), $\rho\,dV$ (volume).
  • Recipe: pick dq, write $dE = k\,dq/r^2$, cancel by symmetry, integrate, check a limit.
  • Rod on its axis: $E = kQ/[d(d+L)]$. Nothing cancels.
  • Ring on its axis: $E = kQz/(z^2+a^2)^{3/2}$. Sideways parts cancel; zero at the center.
  • Far away, every finite distribution acts like a point charge.