Notes on lecture 19

Jiten Chandra Kalita and Shyamashree Upadhyay

1 Stokes theorem continued ...

Remark 1.0.1. Whenever you are given a problem in which you have to use
Stokes theorem, you should look at the following: From the inputs provided to
you, calculating the line integral is going to be simpler or the surface integral.
Depending upon whichever is simpler, you should proceed. Sometimes, it may so
happen that even without using Stokes theorem, you can evaluate the integral
in a much straight forward way.

Example 1.0.2. Evaluate S (O × F~ ) · dS ~ where F~ = xz î + yz ĵ + xy k̂,

R R
2 2 2
and S is the part of the sphere x + y + z = 4 that lies inside the cylinder
x2 + y 2 = 1 and above the xy-plane. Figure 1.0.1 shows the surface S in blue
color.

Figure 1.0.1: Figure for Example 1.0.2

By inspection, we can see that evaluating the surface integral is going to be

1
cumbersome. It is perhaps better to convert the integral into a line integral.
By Stokes theorem, we have
Z Z I
~ =
(O × F~ ) · dS ~
F~ · dr,
S C

where C is a positively oriented curve that bounds the surface S. We have to

find out such a curve C here. Consider the curve of intersection of the sphere
with the cylinder. This curve certainly bounds the surface S. If we substitute
x2 + y 2 = 1 in the equation x2 + y 2 + √z 2 = 4 of the sphere, then we will get
z 2 = 3, which in turn implies that z = 3 [We have taken the positive square
root of 3 because the surface is above the xy-plane.]. If we choose this curve
of intersection to be C, and take the anti-clockwise orientation on it, then
we get a curve which is positively oriented and bounds the surface S.
The equations for the curve C are given by:
√
x2 + y 2 = 1, z = 3.
A parametrization of C is given by:
√
~r(t) =< cos t, sin t, 3 >, 0 ≤ t ≤ 2π.
So by Stokes theorem, we have
Z Z I Z 2π
~ =
(O × F~ ) · dS ~ =
F~ · dr F~ (~r(t)) · r~0 (t)dt
S C 0
Z 2π √ √
= < 3 cos t, 3 sin t, cos t sin t > · < − sin t, cos t, 0 > dt
0
Z 2π √ √
= (− 3 sin t cos t + 3 sin t cos t)dt = 0.
0

Remark 1.0.3. Please refer to the last remark of the last lecture. It has some
significance. We are going to consider some problems related to it in the tutorial
sheet. Whenever we consider a surface, in order to find a curve which bounds
the surface, what are we going to take/choose? This choice plays a crucial role
in some problems.
In the above example, even without doing all those calculations, we could have
reached the same conclusion (that the value of the integral is 0), by choosing
the curve C bounding the surface S in a specific way. You will find a couple of
problems in the tutorial sheet related to this matter.

2
2 The divergence theorem
In this section, we are going to study the Divergence theorem, which is due
to Gauss, and is also sometimes called the Gauss’s divergence theorem.
At this point, recall the vector form of Green’s theorem. Green’s theorem
was considered on a planar region (2-dimensional) and we had found that
I Z Z
F~ · n̂ds = (O · F~ )dA.
C D

This was the result in the vector form of Green’s theorem. All this was in a
2-dimensional plane.
It makes sense to extend this result (Green’s theorem) H to the 3-dimensional
space. Then, intuitively speaking, the line integral C F~ · n̂ds should have
become a surface integral of the form S (F~ · n̂)dS (for some surface S), and
R R

the double integral D (O · F~ )dA should have become a triple integral of the
R R

(O · F~ )dV for some solid region E in R3 . It so happens that our

RR R
form E
intuition becomes true. The equation
Z Z Z Z Z
(F~ · n̂)dS = (O · F~ )dV
S E

is precisely the result stated in the Divergence theorem.

Theorem 2.0.1. Divergence theorem: Let E be a simple solid region and

let S be the boundary surface of E, given with positive (outward) orientation.
Let F~ be a vector field whose component functions have continuous partial
derivatives on an open region that contains E. Then
Z Z Z Z Z
~ ~
F · dS = div F~ dV.
S E

Remark 2.0.2. In the statement of Divergence theorem, E is a simple solid

region. That is, E can be simultaneously considered as a type 1 or 2 or 3 region
in R3 . Even if E is not a simple solid region, we can definitely express E as the
union of some simple solid regions. Then by breaking up the integral, we can
prove the divergence theorem.

Proof: (of Divergence theorem for a simple solid region:) Assume

that F~ =< P, Q, R > where P, Q, R are functions of x, y, z having continuous

3
partial derivatives. Then the left hand side (LHS) integral in the statement
of divergence theorem equals
Z Z Z Z Z Z Z Z
(F~ · n̂)dS = (P î) · n̂dS + (Qĵ) · n̂dS + (Rk̂) · n̂dS
S S S S

and the right hand side (RHS) integral in the statement of divergence theorem
equals Z Z Z  
∂P ∂Q ∂R
+ + dV.
E ∂x ∂y ∂z
If we can prove that
Z Z Z Z Z
∂P
(P î) · n̂dS = dV, (1)
S E ∂x
Z Z Z Z Z
∂Q
(Qĵ) · n̂dS = dV, (2)
S E ∂y
and Z Z Z Z Z
∂R
(Rk̂) · n̂dS = dV, (3)
S E ∂z
then we are done.
We will prove (3). The proofs of (1) and (2) are similar. To prove (3), we
are going to consider E as a type 1 region. That is,

E = {(x, y, z) ∈ R3 |(x, y) ∈ D ⊆ R2 , u1 (x, y) ≤ z ≤ u2 (x, y)}.

Figure 2.0.1 shows E as a type 1 region.

Now the RHS integral in equation (3) equals

Z Z Z Z Z "Z u2 (x,y) #
∂R ∂R
dV = dz dA
E ∂z D z=u1 (x,y) ∂z

Z Z
= [R(x, y, u2 (x, y)) − R(x, y, u1 (x, y))] dA. (4)
D
R R
And the LHS integral in equation (3) equals S (Rk̂) · n̂dS.
The region E is bounded by the surfaces S1 , S2 and S3 (as shown in Figure

4
 Figure 2.0.1: E as a type 1 region

2.0.1). The surface S1 is the lower one, S2 is the upper one, and S3 is the
side surface. The surface S is the union of S1 , S2 and S3 . So we have
Z Z 3 Z Z
X
(Rk̂) · n̂dS = (Rk̂) · n̂dS.
S i=1 Si

R R
Let us focus on S3 (Rk̂) · n̂dS first. On S3 , observe that any outward
normal n̂ is always going to be horizontal. And the vector k̂ is in the vertical
direction. The dot product of k̂ and n̂ is going to be 0. Therefore we have
Z Z
(Rk̂) · n̂dS = 0.
S3
R R
Let us now focus on S2 (Rk̂) · n̂dS. Since S2 is the upper surface, it fol-
lows that the normal n̂ to it will have its z-component positive (that is,
the outward normal will point upwards). In fact, the normal vector will be
< − ∂u
∂x
2
, − ∂u
∂y
2
, 1 >. We can then apply Remark 2.3.5 of the previous lecture
notes, and conclude that
Z Z Z Z
(Rk̂) · n̂dS = R(x, y, u2 (x, y))dA.
S2 D

Similarly, for the lower surface S1 , the normal n̂ will have its z-component
negative (that is, the outward normal will point downwards). In fact, the

5
normal vector will be < ∂u
∂x
1 ∂u1
, ∂y , −1 >. So we can do a calculation similar to
that shown in Remark 2.3.5 of the previous lecture notes, and conclude that
Z Z Z Z
(Rk̂) · n̂dS = − R(x, y, u1 (x, y))dA.
S1 D

Therefore,
Z Z Z Z
(Rk̂) · n̂dS = [R(x, y, u2 (x, y)) − R(x, y, u1 (x, y))]dA,
S D

which is nothing but the expression on the right hand side of equation (4)
above.
We can work out similarly for proving (1) and (2). For those proofs, we will
have to consider E as a type 3 or type 2 region and proceed. This completes
the proof. 

Example 2.0.3. Find the flux of the vector field F~ (x, y, z) = z î + y ĵ + xk̂
over the unit sphere x2 + yR2 + 2
R z = 1.
~ where S is the unit sphere. By diver-
Basically, we have to find S F~ · dS,
gence theorem, Z Z Z Z Z
~ ~
F · dS = O · F~ dV,
S E
where E is the solid unit sphere. Here,
∂ ∂ ∂
O · F~ = (z) + (y) + (x) = 1.
∂x ∂y ∂z
This implies that
Z Z Z Z Z Z
~ 4 4π
O · F dV = 1dV = V olume of E = π(1)2 = .
E E 3 3
So we could compute a complicated surface integral in a much simple way
using divergence theorem. 
~ where
Example 2.0.4. Evaluate S F~ · dS,
R R

2
F~ (x, y, z) = xy î + (y 2 + exz )ĵ + sin(xy)k̂

6
 Figure 2.0.2: Figure for Example 2.0.4

and S is the surface of the region E bounded by the parabolic cylinder

z = 1 − x2 and the planes z = 0, y = 0, and y + z = 2. Figure 2.0.2 shows
the region E in blue color.

Observe from the figure that if we take the projection/shadow of E on the xz-
plane, then the computation of the integral will become easy. That means,
we have to consider E as a type 3 region as follows:
E = {(x, y, z) ∈ R3 | − 1 ≤ x ≤ 1, 0 ≤ z ≤ 1 − x2 , 0 ≤ y ≤ 2 − z}.
Observe also that
∂ ∂ 2 2 ∂
O · F~ = (xy) + (y + exz ) + (sin(xy))
∂x ∂y ∂z
= y + 2y + 0 = 3y.
Then by divergence theorem,
Z Z Z Z Z
~ ~
F · dS = O · F~ dV
S E
Z 1 Z 1−x2 Z 2−z
= 3ydydzdx.
x=−1 z=0 y=0
The rest of the calculation is left to the reader as an exercise. 
—————-END——————