Miscellaneous examples

Math 120 Section 4

Stokes’ theorem
Example 1. Let F~ be any differentiable vector field defined in R3 , and let S be
the unit sphere x2 + y 2 + z 2 = 1 oriented outward. Show that
ZZ
curlF~ · ~ndS = 0.
S

Solution. Divide up the sphere S into the upper hemisphere S1 and the lower
hemisphere S2 , by the unit circle C that is the “equator”. Note that each hemi-
sphere has a boundary curve (the equator C).

Applying Stokes’ theorem to the surface S1 gives:

ZZ Z
curlF~ · ~ndS1 = F~ · d~r (1)
S1 Cccw

where Cccw indicates that the curve C is oriented “counterclockwise” as seen

from above. (Check that this is the positive orientation of the curve, since the
normal vectors to S1 point outward.)

Stokes’ theorem for the surface S1 gives:

ZZ Z
~
curlF · ~ndS2 = F~ · d~r (2)
S2 Ccw

where this time the boundary C is oriented clockwise - that is the positive ori-
entation when the normal vectors for S2 point “outward” (check with the Left
Hand Rule!).

1
But we know that reversing orientation changes the sign of the line integral:
Z Z
F~ · d~r = − F~ · d~r
Ccw Cccw

So on adding (1) and (2) we get:

ZZ ZZ ZZ Z Z
~
curlF ·~ndS = ~
curlF ·~ndS1 + ~
curlF ·~ndS2 = F~ ·d~r+ F~ ·d~r = 0.
S S1 S2 Cccw Ccw

Note. This argument in fact works for any closed surface, by dividing the surface
into two using any closed curve C. In fact, we shall see another solution to this
in class, using the Divergence Theorem.
Example 2 (Exercise 5 in Section 16.8). The surface S consists of the top and
four sides (but not the bottom) of the cube with vertices (±1, ±1, ±1), oriented
outward. Evaluate ZZ
curlF~ · ~ndS
S

where F~ (x, y, z) = hxyz, xy, x2 yzi.

Solution. Here the curl of the vector field is:

curlF~ = hx2 z, xy − 2xyz, y − xzi

Here it is probably tedious to compute the flux of this vector field over S di-
rectly, because one has to do it separately on each of the five faces of the cube.
But whenever the question asks to compute the flux of the curl of a vector field,
one should think of Stokes’ theorem, and that helps here.

Stokes theorem gives you two ways to avoid computing the flux of the curl over S:

One is to compute a line integral instead:

ZZ Z
curlF · ~ndS = F~ · d~r
~
S C

Here, C is the boundary of S which is a “square” on the z = −1 plane (check

what the orientation should be, given that S is oriented outward!).

The other way (which we use here) is to replace S with a simpler surface S2
that has the same boundary curve:
ZZ ZZ
~
curlF · ~ndS = curlF~ · ~ndS2
S S2

Here, we can take S2 to be the “missing face” of the cube, which is the part of
the plane z = −1 enclosed by the square C, with the normal vector pointing

2
“downward”.

Computing the flux of the curl over S2 is easy! Treat S2 as the graph of the
function z = −1 over the domain D which is the square −1 ≤ x ≤ 1, −1 ≤ y ≤ 1
on the xy-plane. The normal vectors are −~k everywhere (S2 is flat!).

So we have:
ZZ ZZ
~
curlF · ~ndS2 = hx2 (−1), xy − 2xy(−1), y − x(−1)i · h0, 0, −1idA
S2 D

where we have plugged in z = −1 in the expression for curlF~ since we are look-
ing along the graph of z = −1.

This now becomes:

ZZ Z1 Z1
(−x − y)dA = (−x − y)dxdy = 0
D −1 −1

and that’s the answer!

Triple integrals
Example 3 (From the Final Exam, Spring ’11). Convert the triple integral
√
3
√3 2 √ 2 2
Z2 Z4 −x 1−x −y
Z
(x2 + y 2 + z 2 )dzdydx
√
3
√ 3 2 1
− 2 − 4 −x 2

into cylindrical coordinates. Do not evaluate the integral.

Solution. Looking at the limits for the variables x and y and z, we first
p figure out
that the solid region is the one between the graphs z = 21 and z = 1 − x2 − y 2
√
over a disk D of radius 23 on the xy-plane centered at (0, 0). (See the figure!)

3
Figure 1: The plane z = 21 intersects the sphere x2 + y 2 + z 2 = 1 in a circle of
√
radius 23 . Our solid region is the one between the plane and the sphere, and
√
its “shadow” is a disk of radius 23 on the xy-plane.

We start with inside the integral, where we convert the function into cylindrical
coordinates, and replace the dzdydx by the volume element rdzdrdθ. We now
need to set up the limits.

Cylindrical coordinates essentially means one uses polar coordinates on the xy-
plane: the limits for r and θ should give you p
the disk D (the “shadow” of the
solid region). The graphs z = 1/2 and z = 1√− x2 − y 2 when converted to
cylindrical coordinates become z = 1/2 and z = 1 − r2 respectively.

So the answer is:

√
3 √
Z2π Z2 Z1−r2
(r2 + z 2 )rdzdrdθ
0 0 1
2

Example 4 (Done in class, except the calculation). Chop up the unit ball
x2 + y 2 + z 2 ≤ 1 by the plane z = √12 . What is the volume of the top portion E
?

Solution. The picture is similar to the one for the previous example, except that
the plane is at a different height.

In class, we figured out that we can set up the triple integral in spherical coor-
dinates as follows:
π
ZZZ Z2πZ4 Z1
dV = ρ2 sinφdρdφdθ
E 0 0 √ 1
2 cos θ

4
We do the calculation here:

π π π
Z2πZ4 Z1 Z2πZ4 Z2πZ4 
ρ3 1 sinφ 1 1 
ρ2 sinφdρdφdθ = sinφ 1 dρdφdθ = − √ · 3 sin φ dφdθ

3 √
2 cos θ
3 6 2 cos φ
0 0 √ 1 0 0 0 0
2 cos θ

Doing the integral of the two terms separately (the second term by the substi-
tution u = cos φ), we get:

Z2π Z2π
− cos φ  π4 1 1  π4  1 1  1 1 
=  + √ · dθ = − √ + + √ − √ dθ
3 6 2 2 cos2 φ 0 3 2 3

0 6 2 12 2
0 0

 1 1 1  2π π
= 2π − √ + + √ = − √ .
3 2 3 12 2 3 2 2

That’s the answer I promised in class!