V10.

1 The Divergence Theorem

1. Introduction; statement of the theorem.

The divergence theorem is about closed surfaces, so let’s start there. By a closed surface
S we will mean a surface consisting of one connected piece which doesn’t intersect itself, and
which completely encloses a single finite region D of space called its interior. The closed
surface S is then said to be the boundary of D; we include S in D. A sphere, cube, and
torus (an inflated bicycle inner tube) are all examples of closed surfaces. On the other hand,
these are not closed surfaces: a plane, a sphere with one point removed, a tin can whose
cross-section looks like a figure-8 (it intersects itself), an infinite cylinder. n
A closed surface always has two sides, and it has a natural positive
direction — the one for which n points away from the interior, i.e., points D
toward the outside. We shall always understand that the closed surface has
been oriented this way, unless otherwise specified. S

We now generalize to 3-space the normal form of Green’s theorem (Section V4).
Definition. Let F(x, y, z) = M i + N j + P k be a vector field differentiable in some
region D. By the divergence of F we mean the scalar function div F of three variables
defined in D by

∂M ∂N ∂P
(1) div F = + + .
∂x ∂y ∂z

The divergence theorem. Let S be a positively-oriented closed surface with interior

D, and let F be a vector field continuously differentiable in a domain contatining D. Then
�� ���
(2) F · dS = div F dV
S D

We write dV on the right side, rather than dx dy dz since the triple integral is
often calculated in other coordinate systems, particularly spherical coordinates.
The theorem is sometimes called Gauss’ theorem.
Physically, the divergence theorem is interpreted just like the normal form for Green’s
theorem. Think of F as a three-dimensional flow field. Look first at the left side of (2). The
surface integral represents the mass transport rate across the closed surface S, with flow out
of S considered as positive, flow into S as negative.
Look now at the right side of (2). In what follows, we will show that the value of div F
at (x, y, z) can be interpreted as the source rate at (x, y, z): the rate at which fluid is being
added to the flow at this point. (Negative rate means fluid is being removed from the flow.)
The integral on the right of (2) thus represents the source rate for D. So what the divergence
theorem says is:

(3) flux across S = source rate for D ;

i.e., the net flow outward across S is the same as the rate at which fluid is being produced
(or added to the flow) inside S.
1
2 V10.1 THE DIVERGENCE THEOREM

To complete the argument for (3) we still have to show that

(3) div F = source rate at (x, y, z) .

n
To see this, let P0 : (x0 , y0 , z0 ) be a point inside the region D where F is defined.
(To simplify, we denote by (div F)0 , (∂M/∂x)0 , etc., the value of these functions at P0 .)
n
Consider a little rectangular box, with edges Δx, Δy, Δz parallel to the coordi- Δz
Δy
nate axes, and one corner at P0 . We take n to be always pointing outwards, as P0 Δ x
usual; thus on top of the box n = k , but on the bottom face, n = − k . n

The flux across the top face in the n direction is approximately

F(x0 , y0 , z0 + Δz) · k ΔxΔy = P (x0 , y0 , z0 + Δz) ΔxΔy,

while the flux across the bottom face in the n direction is approximately

F(x0 , y0 , z0 ) · − k ΔxΔy = −P (x0 , y0 , z0 ) ΔxΔy .

So the net flux across the two faces combined is approximately

� �
ΔP
[P (x0 , y0 , z0 + Δz) − P (x0 , y0 , z0 )] ΔxΔy = ΔxΔyΔz .
Δz

Since the difference quotient is approximately equal to the partial derivative, we get the
first line below; the reasoning for the following two lines is analogous:
�
�
∂P
net flux across top and bottom ≈ ΔxΔyΔz;
∂z 0
� �
∂N
net flux across two side faces ≈ ΔxΔyΔz;
∂y 0
� �
∂M
net flux across front and back ≈ ΔxΔyΔz;
∂x 0

Adding up these three net fluxes, and using (3), we see that

source rate for box = net flux across faces of box

� �
∂M ∂N ∂P
≈ + + ΔxΔyΔz .
∂x ∂y ∂z 0

Using this, we get the interpretation for div F we are seeking:

source rate for box

source rate at P0 = lim = (div F)0 .
box→0 volume of box

Example 1. Verify the theorem when F = x i + y j + z k and S is the sphere ρ = a .

xi + yj + zk
��
Solution. For the sphere, n = ; thus F · n = a, and F · dS = 4πa3 .
a S
 V10.1 THE DIVERGENCE THEOREM 3

4
���
On the other side, div F = 3, 3 dV = 3 · πa3 ; thus the two integrals are equal. �
D 3

Example 2. Use the divergence theorem to evaluate the flux of F = x3 i + y 3 j + z 3 k

across the sphere ρ = a.
Solution. Here div F = 3(x2 + y 2 + z 2 ) = 3ρ2 . Therefore by (2),
a
12πa5
�� ��� �
F · dS = 3 ρ2 dV = 3 ρ2 · 4πρ2 dρ = ;
S D 0 5

we did the triple integration by dividing up the sphere into thin concentric spheres, having
volume dV = 4πρ2 dρ.

Example 3. Let S1 be that portion of the surface of the paraboloid z = 1 − x2 − y 2

lying above the xy-plane, and let S2 be the part of the xy-plane lying inside the unit circle,
directed so the normal n points upwards. Take F = yz i + xz j + xy k ; evaluate the flux of
F across S1 by using the divergence theorem to relate it to the flux across S2 .
Solution. We see immediately that div F = 0. Therefore, if we let S2′ be the same
surface as S2 , but oppositely oriented (so n points downwards), the surface S1 + S2′ is a
closed surface, with n pointing outwards everywhere. Hence by the divergence theorem,
�� �� ��
F · dS = 0 = F · dS − F · dS
S1 +S2′ S1 S2

Therefore, since we have n = k on S2 ,

S1 n
�� �� �� ��
F · dS = F · dS = F · k dS = xy dx dy
S1 S2 S2 S2
= 0,
S
by integrating in polar coordinates (or by symmetry). n 2
MIT OpenCourseWare
http://ocw.mit.edu

18.02SC Multivariable Calculus

Fall 2010

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.