DEPARTMENT OF MATHEMATICS

MATH2000
Flux integrals and Gauss’ divergence theorem (solutions)

(1) The hemisphere can be represented as

V = {(r, θ, φ) | 0 ≤ r ≤ 1, 0 ≤ θ ≤ 2π, 0 ≤ φ ≤ π/2}.

We have by direct calculation

divF = 3(x2 + y 2 + z 2 ) = 3r2

in terms of spherical coordinates.

y Z π/2 Z 2π Z 1
⇒ divF dV = 3 r2 · r2 sin φ dr dθ dφ
V 0 0 0
ÃZ ! µZ ¶ µZ ¶
π/2 2π 1
4
= 3 sin φ dφ dθ r dr
0 0 0
1 6π
= 3 × 1 × 2π × = .
5 5
{
Now to evaluate F · n dS. In this case the surface comprises of two parts: the base of
S
the hemisphere which lies in the x-y plane, denoted S1 , and the part of the sphere itself,
denoted S2 . So that
{ x x
F · n dS = F · n1 dS + F · n2 dS
S S1 S2

where n1 and n2 are outwardly pointing unit normal vectors to the surfaces S1 and S2
respectively.
x
We expect the integral F · n1 dS to be zero since the k component of F is 0 when F
S1
is restricted to the x-y plane, so there is no flux across that surface.
To verify this by direct calculation, we must first parametrise the surface S1 . Since it is
just a circular disc in the x-y plane, we have

r(r, θ) = r cos θi + r sin θj, 0 ≤ r ≤ 1, 0 ≤ θ ≤ 2π.

We then take the tangent vectors

r θ = −r sin θi + r cos θj
r r = cos θi + sin θj

1
We can calculate r r × r θ = rk. However, this is directed into the solid. We require this
vector to be directed outwards from the solid, so instead we’ll take r θ × r r = −rk.
In terms of our parametrisation,

F (r, θ) = r3 cos3 θi + r3 sin3 θj,

so that the dot product

F (r, θ) · (r θ × r r ) = 0
which tells us that the flux across S1 will be zero as originally thought.
To calculate the flux across S2 , we parametrise S2 (compare with the spherical coordinate
transformation) as

r(θ, φ) = cos θ sin φi + sin θ sin φj + cos φk, 0 ≤ θ ≤ 2π, 0 ≤ φ ≤ π/2.

The tangent vectors are

r θ = − sin θ sin φi + cos θ sin φj

r φ = cos θ cos φi + sin θ cos φj − sin φk.

To find a vector normal to the surface S2 , take

r φ × r θ = cos θ sin2 φi + sin θ sin2 φj + sin φ cos φk.

We should check the direction to make sure it is directed outwards from the surface. Take
for example the parameter values φ = π/2 and θ = 0. This gives r φ × r θ = i which is
directed out, so the direction is ok.
In terms of the parameters, we can write

F (θ, φ) = cos3 θ sin3 φi + sin3 θ sin3 φj + cos3 φk.

so that the dot product

F (θ, φ) · (r φ × r θ ) = cos4 θ sin5 φ + sin4 θ sin5 φ + sin φ cos4 φ.

The flux integral we need to evaluate is then

Z 2π Z π/2
((cos4 θ + sin4 θ) sin5 φ + sin φ cos4 φ) dφ dθ.
0 0

Using cos2 θ = 21 (1 + cos 2θ) we have

1 1
cos4 θ = (1 + cos 2θ)(1 + cos 2θ) = (1 + 2 cos 2θ + cos2 2θ)
4 4
1 1
= (1 + 2 cos 2θ + (1 + cos 4θ))
4 2
3 1 1
= + cos 2θ + cos 4θ.
8 2 8
2
It follows that

cos4 θ + sin4 θ = cos4 θ + (1 − cos2 θ)(1 − cos2 θ)

= 1 − 2 cos2 θ + 2 cos4 θ
3 1
= 1 − (1 + cos 2θ) + + cos 2θ + cos 4θ
4 4
3 1
= + cos 4θ.
4 4
Finally note that we can write sin5 φ = sin φ(1 − 2 cos2 φ + cos4 φ).
Putting this together, the flux across S2 is
Z 2π Z π/2 µ ¶
3 1
+ cos 4θ sin φ(1 − 2 cos2 φ + cos4 φ) dφ dθ
0 0 4 4
Z 2π Z π/2
+ sin φ cos4 φ dφ dθ
0 0
µZ 2π µ ¶ ¶ ÃZ π/2 !
3 1
= + cos 4θ dθ sin φ(1 − 2 cos2 φ + cos4 φ)dφ
0 4 4 0
µZ 2π ¶ ÃZ π/2 !
+ dθ sin φ cos4 φ dφ
0 0

Using the substitution u = cos φ in both φ integrals gives

÷ ¸2π Z 1 ! µ Z 1 ¶
3 1 2 4 4
= θ+ sin 4θ × (1 − 2u + u )du + 2π × u du
4 16 0 0 0
· ¸1 · ¸1
3π 2 3 1 5 1 5
= u− u + u + 2π u
2 3 5 0 5 0
µ ¶
3π 2 1 2π 6π
= 1− + + = .
2 3 5 5 5

Therefore
{ x x 6π 6π
F · n dS = F · n1 dS + F · n2 dS = 0 + = .
5 5
S S1 S2

So we have shown that for this example both sides of the equation in Gauss’ theorem are
equal.

3
(2) Note that in this case we cannot use Gauss’ divergence theorem since the vector field
1
F = i is undefined at any point in the y-z plane (ie. when x = 0), part of which lies in
x {
the region enclosed by the surface. We must evaluate F · n dS directly.
S
Since the surface is the unit sphere, the position vector r = xi + yj + zk will also be an
outwardly pointing unit normal (since x2 + y 2 + z 2 = 1 on the surface). Taking n = r,
we have that F · n = 1. Therefore the flux evaluates to
{ {
F · n dS = dS
S S
= surface area of the unit sphere
= 4π.

(3) A diagram of the solid is as follows:

1 Y
0

0X 2
0
0.5
1
1.5 4
2

The outward flux can be calculated as

{
F ·n
S

where S is the closed surface of the box, F is the vector field, and n is an outwardly
pointing unit normal vector. The surface S consists of six open surfaces: the six faces
of the box. We can evaluate the flux integral directly by calculating the outward flux
through each face:
{ x x x x x x
F ·n = F ·n dS + F ·n dS + F ·n dS + F ·n dS + F ·n dS + F ·n dS.
S S1 S2 S3 S4 S5 S6

4
We represent each open surface follows:

– S1 is the “base” of the box which lies in the plane z = 3 (and is therefore parallel
to the x-y plane). An outwardly pointing unit normal is n = −k. Restricted to S1 ,
the vector field is given by

F = xi + 12yj + 9k, for 1 ≤ x ≤ 3, 0 ≤ y ≤ 1.

Therefore over S1 ,

F · n = (xi + 12yj + 9k) · (−k) = −9.

The surface integral is then

x x x
F · n dS = (−9) dS = −9 dS.
S1 S1 S1
x
Since dS is just the area of S1 , which is a rectangle of area = 2, so
S1
x x
dS = 2 ⇒ F · n dS = −9 × 2 = −18.
S1 S1

– S2 is the “lid” of the box which lies in the plane z = 5. An outwardly pointing unit
normal is n = k. Restricted to S2 , the vector field is given by

F = xi + 12yj + 15k, for 1 ≤ x ≤ 3, 0 ≤ y ≤ 1.

Therefore over S2 ,
F · n = (xi + 12yj + 15k) · k = 15.
The surface integral is then
x x x
F · n dS = (15) dS = 25 dS.
S2 S2 S2
x
Since dS is just the area of S2 , which is a rectangle of area = 2, so
S2
x x
dS = 2 ⇒ F · n dS = 15 × 2 = 30.
S2 S2

– S3 is the “back” of the box which lies in the plane x = 1. An outwardly pointing
unit normal is n = −i. Restricted to S3 , the vector field is given by

F = 1i + 12yj + 3zk, for 0 ≤ y ≤ 1, 3 ≤ z ≤ 5.

5
 Therefore over S3 ,

F · n = (1i + 12yj + 3zk) · (−i) = −1.

The surface integral is then

x x x
F · n dS = (−1) dS = − dS.
S3 S3 S3
x
Since dS is just the area of S3 , which is a rectangle of area = 2, so
S3
x x
dS = 2 ⇒ F · n dS = −2.
S3 S3

– S4 is the “front” of the box which lies in the plane x = 3. An outwardly pointing
unit normal is n = i. Restricted to S4 , the vector field is given by

F = 3i + 12yj + 3zk, for 0 ≤ y ≤ 1, 3 ≤ z ≤ 5.

Therefore over S4 ,
F · n = (3i + 12yj + 3zk) · (i) = 3.
The surface integral is then
x x x
F · n dS = (3) dS = 3 dS.
S4 S4 S4
x
Since dS is just the area of S4 , which is a rectangle of area = 2, so
S4
x x
dS = 2 ⇒ F · n dS = 3 × 2 = 6.
S4 S4

– S5 is the ‘left side” of the box which lies in the plane y = 0 (the x-z plane). An
outwardly pointing unit normal is n = −j. Restricted to S5 , the vector field is given
by
F = xi + 3zk, for 1 ≤ x ≤ 3, 3 ≤ z ≤ 5.
Therefore over S5 ,
F · n = (xi + 3zk) · (−j) = 0.
The surface integral is then
x x
F · n dS = (0) dS = 0.
S5 S5

6
 – S6 is the “right side” of the box which lies in the plane y = 1. An outwardly pointing
unit normal is n = j. Restricted to S6 , the vector field is given by

F = xi + 12j + 3zk, for 1 ≤ x ≤ 3, 3 ≤ z ≤ 5.

Therefore over S6 ,
F · n = (xi + 12j + 3zk) · (j) = 12.
The surface integral is then
x x x
F · n dS = (12) dS = 12 dS.
S6 S6 S6
x
Since dS is just the area of S6 , which is a rectangle of area = 4, so
S6
x x
dS = 4 ⇒ F · n dS = 12 × 4 = 48.
S6 S6

Putting all of this information together gives

{ x x x x x x
F ·n = F ·n dS + F ·n dS + F ·n dS + F ·n dS + F ·n dS + F ·n dS
S S1 S2 S3 S4 S5 S6

= −18 + 30 + (−2) + 6 + 0 + 48 = 64.

Using the divergence theorem, we can also calculate the outward flux as
y
divF dV,
V

where V is the region enclosed by S (ie. the box). We can calculate

∂ ∂ ∂
divF = (x) + (12y) + (3z) = 1 + 12 + 3 = 16.
∂x ∂y ∂z
The outward flux is then
y y
divF dV = 16 dV
V V
= 16 × (vol. of box)
= 16 × (2 × 2 × 1) = 64.

We have therefore verified the divergence theorem. In this case, it is a lot less work to
calculate the volume integral compared to the flux integral.

7
(4) Use the divergence theorem. The region (in this case a sphere of radius 5) can be repre-
sented as
V = {(r, θ, φ) | 0 ≤ r ≤ 5, 0 ≤ φ ≤ π, 0 ≤ θ ≤ 2π}
in term of spherical polar coordinates. We also have
∂ ∂ ∂
divF = (3x) + (4y) + (5z) = 3 + 4 + 5 = 12.
∂x ∂y ∂z
Hence by the divergence theorem the flux out of the surface is
y Z 2π Z π Z 5
divF dV = 12 r2 sin φ dr dφ dθ
V 0 0 0
µZ 2π ¶ µZ π ¶ µZ 5 ¶
2
= 12 dθ sin φ dφ r dr
0 0 0
125
= 12 × 2π × 2 ×
3
= 2000π.

Alternatively, we could make the observation that

y y
divF dV = 12 dV
V V
= 12 × ( volume of sphere of radius 5)
µ ¶
4 3
= 12 × π5 = 2000π.
3
{ y
(5) We need to find F · n dS. By Gauss’ divergence theorem this is equal to divF dV .
S V

∂ ∂ ∂
divF = (x) + (3y) + (6z)
∂x ∂y ∂z
= 1 + 3 + 6 = 10.

In cylindrical polar coordinates, the cone is z 2 = (r cos θ)2 + (r sin θ)2 = r2 ⇒ z = r in

this case since 0 ≤ z ≤ 2. The region in R3 is

V = {(r, θ, z) | 0 ≤ z ≤ 2, 0 ≤ θ ≤ 2π, 0 ≤ r ≤ z}.

8
 So flux
Z 2 Z 2π Z z
= 10 r dr dθ dz
0 0 0
Z 2
2π Z
1 2
= 10 z dθ dz
0 0 2
Z 2 Z 2π
2
= 5 z dz dθ
0 0
· ¸2
1 3
= 5 z · 2π
3 0
80π
= .
3

(6) F = (x3 + xy 2 + xz 2 )i + (x2 y + y 3 + yz 2 )j + (x2 z + y 2 z + z 3 )k so

∂ 3 ∂ 2
divF = (x + xy 2 + xz 2 ) + (x y + y 3 + yz 2 )
∂x ∂y
∂
+ (x2 z + y 2 z + z 3 )
∂z
= (3x2 + y 2 + z 2 ) + (x2 + 3y 2 + z 2 )
+(x2 + y 2 + 3z 2 )
= 5r2 .

The sphere is described by

V = {(r, θ, φ) | 0 ≤ r ≤ a, 0 ≤ θ ≤ 2π, 0 ≤ φ ≤ π}.

So by Gauss’ divergence theorem, the flux across the surface of the sphere
Z a Z 2π Z π
= (5r2 ) · r2 sin φ dφ dθ dr
0
Z a0 0
Z 2π Z π
4
= 5 r dr dθ sin φ dφ
·0 ¸a 0 0
1
= 5 r5 · [θ]2π 0 · [− cos φ]0
π
5 0
1
= 5 · a5 · 2π · 2
5
= 4πa5 .