Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $F$ be a vector field defined in $\mathbb{R}^{3}$ minus the origin.

$$F(r)=\frac{r}{|r|^{3}}$$ where $r=(x,y,z)$.

Let $S$ be the sphere of radius $1$ and center $(2,0,0)$.

Compute $\int_{S}{F\cdot n \ dS}$ where $n$ is the vector normal to the surface of $S$.

My initial reaction was to use the divergence theorem to get:

$$\int_{S}{F\cdot n \ dS} = \int{\operatorname{div}{F} \ dV} = -\int_{S}{\frac{2}{|r|^{4}} \ dV}$$

My multivariable calculus is a bit rusty, so I was having a bit of trouble writing out the triple integral.

The sphere is defined by the equation $(x-2)^{2}+y^{2}+z^{2}=1$ which amounts to $x^{2}+y^{2}+z^{2}=4x-3$ which means $r^2=4 r\sin(\phi)\sin(\theta)-3$.

This seems not too good... Did I make a mistake in my calculations?

share|cite|improve this question
up vote 2 down vote accepted

Yes, actually you made a calculation mistake. Indeed, careful calculation shows that

$$\nabla \cdot \mathbf{F} = 0$$

away from the origin. To be specific, let us rename the coordinate variables as $(x_1, x_2, x_3) = (x, y, z)$. Then

$$\begin{align*}\nabla \cdot \mathbf{F} &= \sum_{i=1}^{3} \frac{\partial}{\partial x_i}\frac{x_i}{r^3} = \sum_{i=1}^{3} \frac{\partial x_i}{\partial x_i}\frac{1}{r^3} + x_i \frac{\partial}{\partial x_i}\frac{1}{r^3} \\ &= \sum_{i=1}^{3} \frac{1}{r^3} + x_i \left(-\frac{3}{r^4}\right)\frac{\partial}{\partial x_i}\sqrt{x_1^2+x_2^2+x_3^2} \\ &= \frac{3}{r^3} - \frac{3}{r^4} \sum_{i=1}^{3} x_i \cdot \frac{x_i}{r} \\ &= \frac{3}{r^3} - \frac{3}{r^3}=0. \end{align*}$$

Now you may apply the divergence theorem to obtain

$$\int_{S}\mathbf{F}\cdot\mathbf{n}\;da = \int_{V}\nabla \cdot \mathbf{F}\;dv = 0,$$

where $V$ is the region inside the sphere. This calculation is valid since the origin does not lie on and inside the sphere $S$.

If you made a bold and brave resolution to attack this problem by brutal integration, then you may do as follows: First, make the substitution $(x',y',z')=(x-2,y,z)$. This is equivalent to shifting the whole situation so that the center of $S$ coincides with the origin. (This makes the domain of integration simpler in expense of the simplicity of the integrand.) Then make a variant of spherical coordinate change as follows:

$$ \begin{align*} x' &= r \cos\phi \\ y' &= r \sin\phi\cos\theta\\ z' &= r \sin\phi\sin\theta. \end{align*}$$

Of course, $0 \leq \phi \leq \pi$ and $0 \leq \theta < 2\pi$. Then on $S$, with this new coordinate system, $S$ is characterizes as the locus of $r = 1$, and on $S$ we have $da = r^2\sin\phi\,d\phi d\theta = \sin\phi\,d\phi d\theta$. Thus our integral becomes

$$\begin{align*}\int_S \mathbf{F}\cdot\mathbf{n}\;da &=\int_{0}^{2\pi}\int_{0}^{\pi} \frac{(x'+2,y',z')}{((x'+2)^2+y'^2+z'^2)^{3/2}} \cdot \frac{(x',y',z')}{(x'^2+y'^2+z'^2)^{1/2}}\;r^2 \sin\phi\,d\phi d\theta\\ &=\int_{0}^{2\pi}\int_{0}^{\pi} \frac{(x'+2,y',z') \cdot (x',y',z')}{(r^2 + 4r\cos\phi + 4)^{3/2}}\;r \sin\phi\,d\phi d\theta\\ &=\int_{0}^{2\pi}\int_{0}^{\pi} \frac{r (r^2 + 2r\cos\phi) \sin\phi}{(r^2 + 4r\cos\phi + 4)^{3/2} }\;d\phi d\theta\\ &=\int_{0}^{2\pi}\int_{0}^{\pi} \frac{(1 + 2\cos\phi) \sin\phi}{(5 + 4\cos\phi)^{3/2} }\;d\phi d\theta\\ &=2\pi \int_{0}^{\pi} \frac{(1 + 2\cos\phi) \sin\phi}{(5 + 4\cos\phi)^{3/2} }\;d\phi \end{align*}$$

Finally, let $t = 4\cos\phi$. Then $dt = -4\sin\phi \, d\phi$ and therefore

$$\begin{align*} \int_S \mathbf{F}\cdot\mathbf{n}\;da &= \frac{\pi}{4} \int_{-4}^{4} \frac{2 + t}{(5 + t)^{3/2} }\;dt \\ &= \frac{\pi}{4} \left[ -2 \cdot \frac{2 + t}{\sqrt{5 + t}}\right]_{-4}^{4} + \frac{\pi}{2} \int_{-4}^{4} \frac{1}{\sqrt{5-t}}\;dt \\ &= -2\pi + \frac{\pi}{2} \left[2\sqrt{5+t}\right]_{-4}^{4} = 0. \end{align*}$$

share|cite|improve this answer
Oops, you are right! Thanks for the correction! But suppose I did have to compute that sort of integral, then do you know how I would be able to do it? It would be nice to know anyways. – neelp Aug 26 '12 at 14:20

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.