You are on page 1of 5

The Gauss Divergence Theorem

Theorem 1 Let F(X) be a continuously differentiable vector field in
a domain D ⊂ R
n
. Let R ⊂ D be a closed, bounded region whose
boundary is a smooth surface, Σ ⊂ D. For each point x ∈ Σ let ν(X) be
the unit outward, or “exterior”, normal to Σ with respect to the region
R. Then, with dX ≡ dx
1
dx
2
...dx
n
and with dσ indicating integration
with respect to surface area on Σ,
_
R
∇· F(X) dX =
_
Σ
F(X) · ν(X) dσ.
Proof We will give the proof for the case n = 3; much the same
proof can be given in the general n-dimensional situation. Supposing
that
F(X) = F(x, y, z) =
_
_
_
_
f(x, y, z)
g(x, y, z)
h(x, y, z)
_
_
_
_
,
we have
∇· F(X) =
_
∂f
∂x
+
∂g
∂y
+
∂h
∂z
_
(x, y, z).
Thus
_
R
∇· F(X) dX =
_
R
∂f
∂x
dX +
_
R
∂g
∂y
dX +
_
R
∂h
∂z
dX.
Initially we will concentrate our attention on the last integral appearing
here.
We will also initially suppose (see figure on page 5) that the bounding
surface Σ for R consists of an “upper” surface and a “lower” surface:
Σ
τ
: φ(x, y, z) = z − τ(x, y) = 0; Σ
ρ
: ψ(x, y, z) = z − ρ(x, y) = 0,
together with some possible surface components parallel to the z-axis
which we will denote collectively by Σ
z
.
1
If we take a point on the upper surface, namely (x, y, τ(x, y)), the unit
normal vector there is
ν
τ
(X) =
−i
∂τ
∂x
− j
∂τ
∂y
+ k
¸
_
∂τ
∂x
_
2
+
_
∂τ
∂y
_
2
+ 1
.
On the lower surface the unit outward normal vector is
ν
ρ
(X) =
i
∂ρ
∂x
+ j
∂ρ
∂y
− k
¸
_
∂ρ
∂x
_
2
+
_
∂ρ
∂y
_
2
+ 1
.
On the upper surface, z = τ(x, y), the element of surface area is
dσ =
¸
¸
¸
¸
_
_
∂τ
∂x
_
2
+
_
∂τ
∂y
_
2
+ 1 dxdy,
while on the lower surface the corresponding surface area element is
dσ =
¸
¸
¸
¸
_
_
∂ρ
∂x
_
2
+
_
∂ρ
∂y
_
2
+ 1 dxdy.
Therefore, with R
xy
denoting the projection of R onto the x, y plane,
_
R
∂h
∂z
(x, y, z) dxdy dz =
_
R
xy
_
_
z=τ(x,y)
z=ρ(x,y)
∂h
∂z
(x, y, z) dz
_
dxdy
=
_
R
xy
(h(x, y, τ(x, y)) − h(x, y, ρ(x, y))) dxdy
=
_
Σ
τ
h(x, y, τ(x, y))

¸
_
∂τ
∂x
_
2
+
_
∂τ
∂y
_
2
+ 1

_
Σ
ρ
h(x, y, ρ(x, y))

¸
_
∂ρ
∂x
_
2
+
_
∂ρ
∂y
_
2
+ 1
.
2
But we clearly have
1
¸
_
∂τ
∂x
_
2
+
_
∂τ
∂y
_
2
+ 1
= k · ν
τ
(X),
−1
¸
_
∂ρ
∂x
_
2
+
_
∂ρ
∂y
_
2
+ 1
= k · ν
ρ
(X).
Consequently
_
R
∂h
∂z
(x, y, z) dxdy dz
=
_
Σ
τ
h(x, y, z) k · ν
τ
(X) dσ +
_
Σ
ρ
h(x, y, z) k · ν
ρ
(X) dσ.
On Σ
z
we have k · ν(X) ≡ 0, so the above can be extended to
_
R
∂h
∂z
(x, y, z) dxdy dz =
_
Σ
h(x, y, z) k · ν(X) dσ.
Complicated regions R can be reduced, by means of cuts parallel to the
z-axis to regions satisfying the hypothesis above. So we may assume
the result obtained so far is not limited with respect to the geometry
of R. (See figure on p. 5.)
In the same way we can see that
_
R
∂f
∂x
(x, y, z) dxdy dz =
_
Σ
f(x, y, z) i · ν(X) dσ;
_
R
∂g
∂y
(x, y, z) dxdy dz =
_
Σ
g(x, y, z) j · ν(X) dσ.
Adding the three results separately obtained for f, g and h we have
_
R
∇·F(X) dX =
_
R
_
∂f
∂x
(x, y, z) +
∂g
∂y
(x, y, z) +
∂h
∂z
(x, y, z)
_
dxdy dz
=
_
Σ
(f(x, y, z) i + g(x, y, z) j + h(x, y, z) k) · ν(X) dσ
=
_
Σ
F(X) · ν(X) dσ
and the theorem is proved.
3
Example 1 Let F(x, y, z) = xi + y j + z k and let Σ be the sphere
of radius 1 in R
3
. We first compute =
_
Σ
F(X) · ν(X) dσ. Here we
have
ν(X) =
1

x
2
+ y
2
+ z
2
_
_
_
_
x
y
z
_
_
_
_
−→ F(X) · ν(X) =
x
2
+ y
2
+ z
2

x
2
+ y
2
+ z
2
=
_
x
2
+ y
2
+ z
2
≡ 1 on Σ.
So we have
_
Σ
F(X) · ν(X) dσ = 1 ·
_
Σ
dσ = 4π.
On the other hand, with R denoting the unit ball inside the unit sphere
Σ as described above, we have
_
R
∇· F(X) dX =
_
R
∇·
_
_
_
_
x
y
z
_
_
_
_
dxdy dz =
_
R
3 dxdy dz
= 3
_
R
dxdy dz = 3 ·
4
3
π = 4π.
The Divergence Theorem equates two different integrals; one over
the region R and one over Σ, the boundary of R. It can happen that
one or the other of these integrals is substantially easier to calculate
than the other, so it may be advantageous to switch from one integral
to the other to facilitate computation.
Example 2 Let R and Σ be as in the preceding example, but now
F(X) = F(x, y, z) =
_
_
_
_
x
3
y
3
z
3
_
_
_
_
.
Then
F(X) · ν(X) =
x
4
+ y
4
+ z
4

x
2
+ y
2
+ z
2
‘.
4
In this case the surface integral,
_
Σ
x
4
+ y
4
+ z
4

x
2
+ y
2
+ z
2

is not easy to compute. But ∇ · F(x, y, z) = 3
_
x
2
+ y
2
+ z
2
_
and
thus, using spherical coordinates,
_
R
∇· F(x, y, z) dxdy dz =
_

0
_
π
0
_
1
0
3r
2
r
2
sin φdr dφdθ
= 2π
_
π
0
sin φdφ
_
1
0
3r
4
dr = 12π
_
1
0
r
4
dr = 12π
r
5
5
¸
¸
¸
¸
¸
¸
1
0
=
12π
5
.
5