Skip to content

Problems

Problem 3.24

Solution Since we are in cylindrical coordinates, we will write Laplace's equation in cylindrical coordinates (s,ϕ,z) (s, \phi, z) :

2V=1ss(sVs)+1s22Vϕ2=0 \laplacian V = \frac{1}{s} \pdv{}{s} \left( s \pdv{V}{s} \right) + \frac{1}{s^2} \frac{\partial ^2 V}{\partial \phi ^2} = 0

We'll try the method of separation of variables on s and ϕ \phi by searching for solutions which are products of the form

V(s,ϕ)=S(s)+Φ(ϕ) V(s, \phi) = S(s) + \Phi(\phi)

1sΦdds(sdSds)+1s2Sd2Φdϕ2=0 \frac{1}{s} \Phi \dv{}{s} \left( s \dv{S}{s} \right) + \frac{1}{s^2} S \frac{d^2 \Phi}{d\phi ^2} = 0

to separate the variables, we need to divide by V and multiply by s2 s^2

sSdds(sdSds)+1Φd2Φdϕ2=0 \frac{s}{S} \dv{}{s} \left( s \dv{S}{s} \right) + \frac{1}{\Phi} \frac{d^2 \Phi}{d \phi ^2} = 0

We define

f(s)=sSdds(sdSds) f(s) = \frac{s}{S} \dv{}{s} \left( s \dv{S}{s} \right)

and

g(ϕ)=1Φd2Φdϕ2 g(\phi) = \frac{1}{\Phi} \frac{d^2 \Phi}{d \phi ^2}

Since we have separated our independent variables and the sum is equal to zero, they must both be constant

f(s)=C1g(ϕ)=C2C1+C2=0 f(s) = C_1 \qquad g(\phi) = C_2 \qquad C_1 + C_2 = 0

Cylindrical symmetry implies that

 when ϕϕ+2π:Φ(ϕ+2π)=Φ(ϕ) \text{ when } \phi \rightarrow \phi + 2 \pi : \qquad \Phi(\phi + 2 \pi) = \Phi(\phi)

So C2 C_2 must be the positive one, since we know that will give us the periodic solutions to Laplace's equation. We write our constant as k2 k^2 so

d2Φdϕ2=k2ΦΦ(ϕ)=Acos(kϕ)+Bsin(kϕ),k=0,1,2,3, \frac{d^2 \Phi}{d \phi ^2} = - k^2 \Phi \\ \rightarrow \Phi(\phi) = A \cos (k \phi) + B \sin (k \phi), \quad k = 0, 1, 2, 3, \ldots

Back to the S part, we need a solution to

sdds(sdSds)=k2S s \dv{}{s} \left( s \dv{S}{s} \right) = k^2 S

A convenient solution would be a power function, S(s)=sn S(s) = s^n if we choose the power n appropriately

sdds(sdsnds)=sdds(snsn1)=sdds(nsn)=sn2sn1=n2sn=k2S=k2snn=±k \begin{aligned} s \dv{}{s} \left( s \dv{s^n}{s} \right) & = s \dv{}{s} \left( s n s^{n-1} \right) \\ & = s \dv{}{s} (n s^n) \\ & = s n^2 s^{n-1} \\ & = n^2 s^n \\ & = k^2 S = k^2 s^n \\ & \rightarrow n = \pm k \end{aligned}

So, our general solution for S is

S(s)=Csk+Dsk S(s) = C s^k + D s^{-k}

And our general solution will be an infinite series over k. But we have to now be careful, because previously we've expressed our general solution in terms of strictly non-zero k, but here we have k=0 k = 0 , which gives us a constant solution

k=0:S(s)=Cs0+Ds0=const. k = 0: \qquad S(s) = C s^0 + D s^0 = \text{const.}

But we should get two solutions for a second-order ordinary differential equation. If we go back to the differential equation for S,

sdds(sdSds)=k2SsdSds= const. =CdSds=csdS=CdssS(s)=Clns+D s \dv{}{s} \left( s \dv{S}{s} \right) = k^2 S \\ \rightarrow s \dv{S}{s} = \text{ const. } = C \\ \rightarrow \dv{S}{s} = \frac{c}{s} \\ \rightarrow \dd S = C \frac{ds}{s} \\ S(s) = C \ln s + D

This gives us our second solution for S for k=0 k = 0 .

Now what about for Φ \Phi ? Looking at the k = 0 case for the Φ \Phi ODE,

d2Φdϕ2=k2Φ=0 for k=0dΦdϕ= const. =BΦ(ϕ)=Bϕ+A \frac{d^2 \Phi}{d \phi ^2} = - k^2 \Phi = 0 \quad \text{ for } k = 0 \\ \frac{d \Phi}{d \phi} = \text{ const. } = B \\ \rightarrow \Phi(\phi) = B \phi + A

But this doesn't meet our periodicity requirement! This isn't a physically acceptable solution. For k = 0, Φ=B \Phi = B is the only 'physically acceptable' solution (we discard Bϕ+A B \phi + A out of hand.)

Finally, our general solution looks like

V(s,ϕ)=a0+b0lns+k=1[sk(akcoskϕ+bksinkϕ)+sk(akcoskϕ+bksinkϕ)] V(s, \phi) = a_0 + b_0 \ln s + \sum_{k=1} ^\infty \left[ s^k (a_k \cos k \phi + b_k \sin k \phi) + s^{-k} (a_k \cos k \phi + b_k \sin k \phi) \right]

We've only been asked for the general solution in cylindrical coordinates (from which we can tell that our solution is independent of a), and we must be given boundary conditions in order to solve for the constants ak,bk a_k, b_k .

Problem 3.27

A sphere of radius R, centered at the origin, carries charge density ρ(r,θ)=kRr2(R2r)sinθ \rho(r, \theta) = k \frac{R}{r^2} (R - 2r) \sin \theta where k is a constant, and r,θ r, \theta are the usual spherical coordinates. Find the approximate potential for points on the z axis, far from the sphere.

We are asked for the approximate potential for points on the z-axis far from the charge distribution, so we'll calculate the terms of our potential from Eq 3.95, and stop when we find the first non-zero term, replacing θ \theta for α \alpha and z z for r r as we go.

V(r)=14πϵ0n=01r(n+1)(r)Pn(cosα)ρ(r)dτ V(\vec{r}) = \frac{1}{4 \pi \epsilon_0} \sum_{n=0} ^\infty \frac{1}{r^{(n+1)}} \int (r') P_n(\cos \alpha) \rho(\vec{r'}) \dd{\tau'}

Let's start with the monopole term. The integral we have to calculate is simply the charge density integrated over the charge distribution

ρ(r)dτ=kR0R0π02π1r2(R2r)sinθ(r2sinθ)drdθdϕ0R(R2r)dr=(Rrr2)0R=0 \int \rho(r) \dd \tau = k R \int_0 ^R \int _0 ^{\pi} \int_{0} ^{2 \pi} \frac{1}{r^2} (R - 2r) \sin \theta (r^2 \sin \theta ) \dd r \dd \theta \dd \phi \\ \int _{0} ^R (R - 2r) \dd r = \left.(R r - r^2)\right|_{0} ^R = 0

So the monopole term comes out to zero. Next, we try calculating the dipole term:

rcosθρ(r)dτ=kRrcosθ1r2(R2r)sinθ(r2sinθ)drdθdϕ \int r \cos \theta \rho(r) \dd \tau = k R \iiint r \cos \theta \frac{1}{r^2} (R - 2r) \sin \theta (r^2 \sin \theta) \dd r \dd \theta \dd \phi

The θ \theta integral will come out to

0πsin2cosθdθ=0πsin2θd(sinθ)=13sin3θ0π=0 \int_0 ^{\pi} \sin ^2 \cos \theta \dd \theta = \int _0 ^\pi \sin ^2 \theta \dd (\sin \theta) = \left. \frac{1}{3} \sin ^3 \theta \right|_0 ^\pi = 0

Well dangit, we still don't have the first non-zero term! On to the quadrupole term:

r2(32cos2θ12)ρdτ=r2(32cos2θ12)kRr2(R2r)sinθr2sinθdrdθdϕ=12kRr2(3cos2θ1)(R2r)sin2θdrdθdϕ \begin{aligned} & \int r^2 \left( \frac{3}{2} \cos ^2 \theta - \frac{1}{2} \right) \rho \dd \tau \\ = & \iiint r^2 \left( \frac{3}{2} \cos ^2 \theta - \frac{1}{2} \right) \frac{kR}{r^2} (R - 2r) \sin \theta r^2 \sin \theta \dd r \dd \theta \dd \phi \\ = & \frac{1}{2} kR \iiint r^2 (3 \cos ^2 \theta - 1)(R - 2r) \sin ^2 \theta \dd r \dd \theta \dd \phi \end{aligned}

Thankfully we don't have any cross-terms, so we can do the integrals separately. The integral in r is

0Rr2(R2r)dr=R46 \int_0 ^R r^2 (R - 2r) \dd r = - \frac{R^4}{6}

The integral in θ \theta is

0π(3cos2θ1)sin2θdθ=0π[3(1sin2θ)1]sin2θdθ=π8 \int_0 ^\pi (3 \cos ^2 \theta - 1) \sin ^2 \theta \dd \theta = \int _0 ^\pi \left[ 3 (1 - \sin ^2 \theta) - 1 \right] \sin ^2 \theta \dd \theta = - \frac{\pi}{8}

And we just get a 2π 2 \pi from the ϕ \phi integral, so converting our r to z in our coordinate system, the whole quadrupole potential is

V(r)14πϵ01z312kR(R46)(π8)(2π)=14πϵ0kπ2R548z3(Quadrupole) V(\vec{r}) \approx \frac{1}{4 \pi \epsilon_0} \frac{1}{z^3} \frac{1}{2} k R \left( - \frac{R^4}{6} \right) \left( - \frac{\pi}{8} \right) (2 \pi) \\ = \frac{1}{4 \pi \epsilon_0} \frac{k \pi ^2 R ^5}{48 z^3} \quad \text{(Quadrupole)}

Problem 3.31

In Ex. 3.9, we derived the exact potential for a spherical shell of radius R, which carries a surface charge σ=kcosθ \sigma = k \cos \theta . a) Calculate the dipole moment of this charge distribution. b) Find the approximate potential, at points far from the sphere, and compare the exact answer (Eq 3.87). What can you conclude about the higher multipoles?

By the symmetry of the problem, p is going to be in the z-direction: p=pz^;p=zρdτzσda \vec{p} = p \vu{z}; \, p = \int z \rho \dd \tau \rightarrow \int z \sigma \dd a .

p=(Rcosθ)(kcosθ)R3sinθdθdϕ=2πR3k0πcos2θsinθdθ=2πR3k(cos3θ3)0π=23πR3k[1(1)]=4πR3k3 \begin{aligned} p & = \int (R \cos \theta)(k \cos \theta) R^3 \sin \theta \dd \theta \dd \phi \\ & = 2 \pi R^3 k \int _0 ^\pi \cos ^2 \theta \sin \theta \dd \theta \\ & = 2 \pi R^3 k \left. \left( - \frac{\cos ^3 \theta}{3} \right) \right|_0 ^\pi \\ & = \frac{2}{3} \pi R^3 k [ 1 - (-1) ] \\ & = \frac{4 \pi R^3 k}{3} \end{aligned}

p=4πR3k3z^(a) \tag{a} \vec{p} = \frac{4 \pi R^3 k}{3} \vu{z}

The associated dipole potential is just

Vdip14πϵ0r^pr2=kR33ϵ0cosθr2 V_{dip} \approx \frac{1}{4 \pi \epsilon_0} \frac{\vu{r} \cdot \vec{p}}{r^2} = \frac{k R^3}{3 \epsilon_0} \frac{\cos \theta}{r^2}

Problem 3.33

Show that the electric field of a 'pure' dipole can be written in the coordinate-free form math \vec{E_{dip}} = \frac{1}{4 \pi \epsilon_0} \frac{1}{r^3} [3 (\vec{p} \cdot \vu{r}) \vu{r} - \vec{p} ]math

We still assume the dipole is pointing in the z-direction and start with spherical coordinates, and then move to a coordinate-free system

p=pz^ \vec{p} = p \vu{z}

p=prr^+pθθ^+pϕϕ^ \vec{p} = p_r \vu{r} + p_\theta \vu{\theta} + p_{\phi} \vu{\phi}

Since p is in the z-direction, we can safely say pϕ=0 p_\phi = 0

pr=pr^=pcosθpθ=pθ^=psinθp=pcosθr^psinθθ^ p_r = \vec{p} \cdot \vu{r} = p \cos \theta \\ p_\theta = \vec{p} \cdot \vu{\theta} = - p \sin \theta \\ \vec{p} = p \cos \theta \vu{r} - p \sin \theta \vu{\theta}

So we can directly check this expression against the expression we got as Eqn 3.103 (Edip(r,θ)=p4πϵ0r3(2cosθr^+sinθθ^) (\vec{E_{dip}}(r, \theta) = \frac{p}{4 \pi \epsilon_0 r^3} (2 \cos \theta \vu{r} + \sin \theta \vu{\theta} ) :

3(pr^)r^p=3pcosθr^pcosθr^+psinθθ^=2pcosθr^+psinθθ^Edip=14πϵ01r3[3(pr^)r^p] \begin{aligned} 3 ( \vec{p} \cdot \vu{r}) \vu{r} - \vec{p} = & 3 p \cos \theta \vu{r} - p \cos \theta \vu{r} + p \sin \theta \vu{\theta} \\ & = 2 p \cos \theta \vu{r} + p \sin \theta \vu{\theta} \\ \rightarrow \vec{E_{dip}} & = \frac{1}{4 \pi \epsilon_0} \frac{1}{r^3} [3 (\vec{p} \cdot \vu{r}) \vu{r} - \vec{p} ] \end{aligned}

So it all checks out.

Problem 3.34

Three point charges are located as shown in Fig 3.38, each a distance a from the origin. Find the approximate electric field at points far from the origin. Express your answer in spherical coordinates, and include the two lowest orders in the multipole expansion.

Figure 3.38

We'll get to the electric field by writing down the multipole expansion of the potential, and then using the approximate potential to get the electric field. The total charge is -q, so the monopole term will be

Vmon=14πϵ01r(q) V_{mon} = \frac{1}{4 \pi \epsilon_0} \frac{1}{r} (-q)

The dipole moment is given by

p=i=13qiri=(q)ay^+(q)a(y^)+qaz^=qaz^ \begin{aligned} \vec{p} & = \sum_{i=1} ^3 q_i \vec{r_i} \\ & = (-q) a \vu{y} + (-q) a (-\vu{y}) + q a \vu{z} \\ & = qa \vu{z} \end{aligned}

The dipole term in the multipole expansion of V is then

Vdip=14πϵ0pr^r2=14πϵ0qaz^r^r2=14πϵ0qacosθr2 \begin{aligned} V_{dip} & = \frac{1}{4 \pi \epsilon_0} \frac{\vec{p} \cdot \vu{r}}{r^2} \\ & = \frac{1}{4 \pi \epsilon_0} \frac{q a \vu{z} \cdot \vu{r}}{r^2} \\ & = \frac{1}{4 \pi \epsilon_0} \frac{q a \cos \theta}{r^2} \end{aligned}

V(r,θ)q4πϵ0(1r+acosθr2) V(r, \theta) \approx \frac{q}{4 \pi \epsilon_0 } \left( - \frac{1}{r} + \frac{a \cos \theta}{r^2} \right)

The field is given by

E=Vq4πϵ0(1r2r^+2acosθr^r3r^+ar3sinθθ^) \vec{E} = - \grad V \approx \frac{q}{4 \pi \epsilon_0} \left( - \frac{1}{r^2} \vu{r} + \frac{2 a \cos \theta \vu{r}}{r^3} \vu{r} + \frac{a}{r^3} \sin \theta \vu{\theta} \right)