SEHS2340 Mathematics II

Chapter 3

Vector Calculus

Contents
3.1 Scalar and Vector fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3.2 Gradient of a scalar field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.3 Divergence and Curl of a vector field . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.4 Conservative Vector Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.5 Curves and Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.6 Line Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.7 Properties of Line Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.8 Line Integrals in vector field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.9 Path Independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.10 Green’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.11 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.12 Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.13 Surface Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.14 Surface Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.15 Surfaces Integrals of Vector Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.16 The Divergence Theorem of Gauss . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.17 Stokes’ Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

1
3.1 Scalar and Vector fields

Definition 1 A scalar field is a function which associates a unique scalar φ(P ) with each point
P in a region of the space or the plane. Therefore, if one uses rectangular coordinates, a scalar field
is nothing but a scalar valued function φ(x, y) or φ(x, y, z) of two or three independent variables.

Example 1 Temperature in a room is a scalar field T = T (x, y, z) . We may use 3D graph or 2D

graph with intensity of the field is represented by different hues of color.

1
Example 2 φ (x, y, z) =
is a scalar field defined everywhere in the xyz-space with
x2 + y 2 + z 2
the exception of the origin.

Example 3 A vector field is a function which associates a unique vector F(P ) with each point P
in a region of the space or the plane. In rectangular coordinates, a vector field may be expressed as

F(P ) = F(x, y, z) = F1 (x, y, z)i + F2 (x, y, z)j + F3 (x, y, z)k

where F1 (x, y, z), F2 (x, y, z) and F3 (x, y, z) are given scalar fields.

2
Example 4 Sketch some vectors in the vector field given by

F (x, y) = −yi + xj

We may plot vectors at several random points in the plane. However, it is more enlightening to
plot vectors of equal magnitude. This corresponds to finding level curves in scalar fields. In this
case, vectors of equal magnitude lie on circles.

F = c


x2 + y 2 = c

x2 + y 2 = c2

To begin making the sketch, choose a value for c and plot several vectors on the resulting circle.
For instance, the following vectors occur on the unit circle.

Point Vector
(1, 0) F (1, 0) = j
(0, 1) F (0, 1) = −i
(−1, 0) F(−1, 0) = −j
(0, 1) F(0, 1) = i

These and several other vectors in the vector field are shown in the following figure.

Example 5 The vector field given by

 
K K
F= r= (xi + yj + zk)
r 3
(x2 + y 2 + z 2 )3/2


where r = x2 + y 2 + z 2 = r and K is a constant, arises from an inverse-square law. Special
examples are gravitational field and electrostatic field.

3
3.2 Gradient of a scalar field

Definition 2 Given a scalar field φ(x, y, z), one may define a vector field
∂φ ∂φ ∂φ
∇φ = i+ j+ k
∂x ∂y ∂z
called the gradient field (or simply gradient) of φ. In particular, when the scalar field φ is
∂φ ∂φ
independent of z, then ∇φ = i+ j is a two dimensional vector field.
∂x ∂y
Example 6 If φ = x2 − y, find ∇φ.

Solution: Since
∂φ ∂ (x2 − y) ∂φ ∂ (x2 − y)
= = 2x and = = −1
∂x ∂x ∂y ∂y
thus
∂φ ∂φ
∇φ = i+ j = 2xi − j.
∂x ∂y
Example 7 If φ(x, y, z) = x + y 2 z − 1, find ∇φ.

Solution: Since
∂φ ∂ (x + y 2 z − 1)
= =1
∂x ∂x
∂φ ∂ (x + y 2 z − 1)
= = 2yz
∂y ∂y
∂φ ∂ (x + y 2 z − 1)
= = y2
∂z ∂z
thus
∂φ ∂φ ∂φ
∇φ = i+ j+ k = i + 2yzj + y 2 k.
∂z ∂y ∂

4
 K K
Example 8 Taking φ(x, y, z) = − = −
and after a simple calculation, we obtain
r x2 + y 2 + z 2

K
∇φ = (xi + yj + zk)
(x2 + y 2 + z 2 )3/2

which is nothing but the vector field arising from an inverse square law as considered in Example 5.
Note:
1 1
− −
∂φ ∂1 ∂ (x2 + y 2 + z 2 ) 2 ∂ (x2 + y 2 + z 2 ) 2 ∂ (x2 + y 2 + z 2 )
= r = =
∂x ∂x ∂x ∂ (x2 + y 2 + z 2 ) ∂x
1 − 1 −1  − 3
= − x2 + y 2 + z 2 2 × 2x = −x x2 + y 2 + z 2 2
2

Example 9 Consider z = f(x, y) = 4x2 + y 2 . The surface defined by this function is an elliptical
paraboloid. This is a bowl-shaped surface. The bottom of the bowl lies at the origin. The figure
below shows the level curves, defined by f (x, y) = C, of the surface. The level curves are the ellipses
4x2 + y 2 = C.

√   √ 
The gradient vector ∇f = 8xi + 2yj is plotted at the 3 points 1.25, 0 , (1, 1), 0, 5 . As the
plot shows, the gradient vector at (x, y) is normal to the level curve through (x, y). As we will see
below, the gradient vector points in the direction of greatest rate of increase of f(x, y).

3.3 Divergence and Curl of a vector field

Definition 3 Let F(x, y, z) = F1 (x, y, z)i + F2 (x, y, z)j + F3 (x, y, z)k be a vector field. We define
the divergence of F(x, y, z) by

∂F1 ∂F2 ∂F3

div F = + +
∂x ∂y ∂z

5
and the curl of F(x, y, z) by
     
∂F3 ∂F2 ∂F1 ∂F3 ∂F2 ∂F1
curl F = − i+ − j+ − k
∂y ∂z ∂z ∂x ∂x ∂y

Note that div F is a scalar field while curl F is a vector field.

Remark: ∇ is known as the del operator, and may be formally written as

∂ ∂ ∂
∇=i +j +k
∂x ∂y ∂z

With this notation, both div F and curl F may be respectively regarded as the dot and cross
∂ ∂ ∂
product of the vectors ∇ = i +j +k and F = F1 (x, y, z)i + F2 (x, y, z)j + F3 (x, y, z)k.
∂x ∂y ∂z

div F = ∇ · F(x, y, z)

and
curl F = ∇ × F(x, y, z)

For easily memorize, we write

 
 
 i j k 
 
 ∂ ∂ ∂ 
∇ × F =  ∂x ∂y ∂z 
 
 
 F1 F2 F3 

Example 10 Suppose F = x2 yi + 2y 3 zj + 3zk. Find div F and curl F.

Solution:
 
∂ ∂ ∂  
∇·F= i +j +k · x2 yi + 2y 3 zj + 3zk
∂x ∂y ∂z
2 3
∂ (x y) ∂ (2y z) ∂ (3z)
= + + = 2xy + 6y 2 z + 3
∂x ∂y ∂z

6
and
 
∂ ∂ ∂  
∇×F= i +j +k × x2 yi + 2y 3 zj + 3zk
∂x ∂y ∂z
 
 
 i j k 
 
 ∂ ∂ ∂ 
=  ∂x
 ∂y ∂z 
 2 3 
 x y 2y z 3z 
     
 ∂ ∂   ∂ ∂   ∂ ∂ 
 ∂y     
=  ∂z  i− 
  2
∂x ∂z j + 
  2
∂x ∂y k

3
 2y z 3z   x y 3z  3
 x y 2y z 
     
∂ (3z) ∂ (2y 3 z) ∂ (3z) ∂ (x2 y) ∂ (2y 3 z) ∂ (x2 y)
= − i− − j+ − k
∂y ∂z ∂x ∂z ∂x ∂y
 
= −2y 3 i− (0) j + −x2 k = − 2y 3 i−x2 k

Example 11 Suppose F1 = −yi + xj, F2 = xi + yj and F3 = −xi − yj. Find the divergence of F1 ,
F2 and F3 .

Solution:  
∂ ∂ ∂ (−y) ∂ (x)
∇ · F1 = i +j · (−yi + xj) = + =0
∂x ∂y ∂x ∂y
and  
∂ ∂ ∂ (x) ∂ (y)
∇ · F2 = i +j · (xi + yj) = + =2
∂x ∂y ∂x ∂y
and  
∂ ∂ ∂ (−x) ∂ (−y)
∇ · F3 = i +j · (−xi − yj) = + = −2
∂x ∂y ∂x ∂y

−yi + xj xi + yj −xi − yj

Interpretation: The divergence measures how much a vector field “spreads out” or “diverges”
from a given point. For example, the figure in the centre has positive divergence at any point P
except origin, since the vectors of the vector field are all spreading as they move away from P . The

7
figure on the left has zero divergence everywhere since the vectors are not spreading out at all. The
field on the right has negative divergence since the vectors are coming closer together instead of
spreading out.

Definition 4 A vector field F is said to be solenoidal or incompressible if div F = 0.

Note: If div F (P ) > 0, the net flow is outward near P and P is called a source. If div F (P ) < 0,
thee net flow is inward near P and P is called a sink.

Example 12 Suppose F1 = −yi + xj and F2 = xi + yj. Find the curl of F1 and F2 .

Solution:
 
∂ ∂ ∂
∇ × F1 = i +j +k × (−yi + xj)
∂x ∂y ∂z
 
 
 i j k 
 
 ∂ ∂ ∂ 
=  ∂x
 ∂y ∂z 
 
 −y x 0 
     
 ∂ ∂   ∂ ∂   ∂ ∂ 
 ∂y ∂z   ∂x ∂z  ∂x ∂y
=   i− 
 
j+ 
 
k

 x 0   −y 0   −y x 
     
∂ (0) ∂ (x) ∂ (0) ∂ (−y) ∂ (x) ∂ (−y)
= − i− − j+ − k
∂y ∂z ∂x ∂z ∂x ∂y
= (0) i − (0) j + (2) k = 2k

and
 
∂ ∂ ∂
∇ × F2 = i +j +k × (xi + yj)
∂x ∂y ∂z
 
 
 i j k 
 
 ∂ ∂ ∂ 
=  ∂x ∂y ∂z 
 
 
 x y 0 
     
 ∂ ∂   ∂ ∂   ∂ ∂ 
     
=  ∂y ∂z  i−  ∂x ∂z  j +  ∂x ∂y  k
 y 0   x 0   x y 
     
∂ (0) ∂ (y) ∂ (0) ∂ (x) ∂ (y) ∂ (x)
= − i− − j+ − k
∂y ∂z ∂x ∂z ∂x ∂y
= (0) i − (0) j + (0) k = 0

8
 Interpretation: The curl of a vector field measures the tendency for the vector field to swirl
around. Imagine that the vector field represents the velocity vectors of water in a lake. If the vector
field swirls around, then when we stick a paddle wheel into the water, it will tend to spin. The
amount of the spin will depend on how we orient the paddle. Thus, we should expect the curl to
be vector valued.
The vector field F1 has ∇ × F1 with positive k component. To see this, use the right hand rule.
Place your right hand at any point P except origin. Point your fingers toward the tail of one of
the vectors of F1 . Now curl your fingers around in the direction of the tip of the vector. Stick your
thumb out. It points toward the +z axis, so the curl should have positive component k.
The vector field F2 has no swirling tendency at all, so we found ∇ × F2 = 0.

Definition 5 A vector field F is said to be irrotational if curl F = 0.

Theorem 1 Let φ and ψ be a scalar fields and u and v be some vector fields. Then

(a) ∇ (φψ) = φ∇ψ + ψ∇φ

(b) ∇ · (φu) = ∇φ · u + φ∇ · u

(c) ∇ × (φu) = ∇φ × u + φ∇ × u

(d) ∇ × ∇φ = curl(∇φ) = 0

(e) ∇ · (∇ × u) = div(curl u) = 0

(f) ∇ · (u × v) = v · ∇u − u · ∇v

Proof of (b):
 
∂ ∂ ∂
∇ · (φu) = i + j + k · (φu1 i + φu2 j + φu3 k)
∂x ∂y ∂z
∂ (φu1 ) ∂ (φu2 ) ∂ (φu3 )
= + +
∂x ∂y ∂z
∂u1 ∂φ ∂u2 ∂φ ∂u3 ∂φ
=φ + u1 +φ + u2 +φ + u3
∂x ∂x ∂y ∂y ∂z ∂z
∂u1 ∂u2 ∂u3 ∂φ ∂φ ∂φ
=φ +φ +φ + u1 + u2 + u3
∂x ∂y ∂z ∂x ∂y ∂z
= φ (∇ · u) + ∇φ · u

9
 Proof of (d):
 
∂φ ∂φ ∂φ
∇ × ∇φ = ∇ × i+ j+ k
∂x ∂y ∂z
 
 
 i j k 
 
 ∂ ∂ ∂ 
=  ∂x 
 ∂y ∂z 
 ∂φ ∂φ ∂φ 
 ∂x ∂y ∂z 
     
 ∂ ∂   ∂ ∂   ∂ ∂ 
 ∂y ∂z   ∂x ∂z   ∂x ∂y 
=  i−    
 ∂φ ∂φ  j +  ∂φ ∂φ  k
∂φ 
 ∂φ
∂y ∂z
  ∂x ∂z
  ∂x ∂y

     
∂ ∂φ ∂ ∂φ ∂ ∂φ ∂ ∂φ ∂ ∂φ ∂ ∂φ
= − i− − j+ − k
∂y ∂z ∂z ∂y ∂x ∂z ∂z ∂x ∂x ∂y ∂y ∂x
=0

Proof of (e):

∇ · (∇ × u) = ∇ · (∇ × (u1 i + u2 j + u3 k))
 
 
 i j k 
 
 ∂ ∂ ∂ 
= ∇ ·  ∂x 
 ∂y ∂z 
 
 u1 u2 u3 
      
 ∂ ∂   ∂ ∂   ∂ ∂ 
 ∂y ∂z   ∂x ∂z   ∂x ∂y 
= ∇ ·  i− 
 
j + 
 
 k

 u2 u3   u1 u3   u1 u2 
        
∂ ∂ ∂ ∂u3 ∂u2 ∂u3 ∂u1 ∂u2 ∂u1
= i+ j+ k · − i− − j+ − k
∂x ∂y ∂z ∂y ∂z ∂x ∂z ∂x ∂y
     
∂ ∂u3 ∂u2 ∂ ∂u3 ∂u1 ∂ ∂u2 ∂u1
= − − − + −
∂x ∂y ∂z ∂y ∂x ∂z ∂z ∂x ∂y
2 2 2 2 2 2
∂ u3 ∂ u2 ∂ u3 ∂ u1 ∂ u2 ∂ u1
= − − + + −
∂x∂y ∂x∂z ∂y∂x ∂y∂z ∂z∂x ∂z∂y
=0

3.4 Conservative Vector Fields

Definition 6 A vector field F is said to be a conservative vector field if there is a scalar field
φ such that F = ∇φ. We say that φ is a potential function (or simply potential) for the vector
field F.

It is clear that if φ is a potential for the vector field F, then φ + K (where K is a constant) is
also a potential for F.

10
 x2 + y 2 + z 2
Example 13 Let φ = , then
2
 2 
x + y2 + z 2
∇φ = ∇
2
 2   2   2 
x + y2 + z2 x + y2 + z 2 x + y2 + z 2
∂ ∂ ∂
2 2 2
= i+ j+ k
∂x ∂y ∂z
2x 2y 2z
= i + j + k = xi + yj + zk
2 2 2
=r

Hence, the position vector r = xi + yj + zk is a conservative vector field.

Example 14 Is F = (2xy + 1) i + (x2 + 2y) j conservative?

Solution: Assuming F = (2xy + 1) i + (x2 + 2y) j = ∇φ, then we have

  ∂φ ∂φ
(2xy + 1) i + x2 + 2y j = i+ j
∂x ∂y

Hence,

∂φ
= 2xy + 1
∂x
∂φ
= x2 + 2y
∂y

After integrating, we have


φ= (2xy + 1) dx = x2 y + x + f (y)

 
φ= x2 + 2y dy = x2 y + y 2 + g (x)

where f (y) is an “arbitrary function” of y and g (x) is an “arbitrary function” of x. Thus

 
F = (2xy + 1) i + x2 + 2y j

is a conservative vector field, with potential function

φ = x2 y + x + y 2 + C

Example 15 Show that F = (yz + 2x) i + (xz − 4y) j + (xy + 1) k is conservative by constructing
a potential function for F.

11
 Solution: Observe that F = ∇φ if and only if

∂φ ∂φ ∂φ
(yz + 2x) i + (xz − 4y) j + (xy + 1) k = i+ j+ k
∂x ∂y ∂z

Hence

∂φ
= yz + 2x
∂x
∂φ
= xz − 4y
∂y
∂φ
= xy + 1
∂z

which is a system of first order partial differential equations. Integrating them, we have

φ = (yz + 2x) dx = xyz + x2 + f (y, z)

φ = (xz − 4y) dy = xyz − 2y 2 + g (x, z)

φ = (xy + 1) dz = xyz + z + h (x, y)

where f (y, z) is an arbitrary function of y and z, g (x, z) is an arbitrary function of x and z and
h (x, y) is an arbitrary function of x and y. Combining these, we conclude that

φ = xyz + x2 − 2y 2 + z + C

is a potential function of F and F is conservative.

Theorem 2 A vector field F is conservative if and only if ∇ × F = 0.

Corollary 1 Let F1 (x, y) and F2 (x, y) have continuous first partial derivatives on a region. The
vector field given by F (x, y) = F1 (x, y) i + F2 (x, y) j is conservative if and only if

∂F2 ∂F1
=
∂x ∂y

Example 16 The vector field given by F = x2 yi + xyj is not conservative because

∂F1 ∂ (x2 y) ∂F2 ∂ (xy) ∂F2 ∂F1

= = x2 and = =y⇒ = .
∂y ∂y ∂x ∂x ∂x ∂y

Example 17 The vector field given by F = (2x + x2 ) i + (y + y 3 ) j is conservative because

∂F1 ∂ (2x + x2 ) ∂F2 ∂ (y + y 3 ) ∂F2 ∂F1

= =0 and = =0⇒ = .
∂y ∂y ∂x ∂x ∂x ∂y

12
3.5 Curves and Paths

In the xy-plane R2 , we may represent a curve C by a vector function

r(t) = x (t) i + y (t) j

on I = [a, b] ⊂ R, where t is called a parameter of the parametric representation of the

curve. More examples are introduced as follows.
The simplest curve is a straight line

r(t) = a + tb,

where a and b are two fixed vectors and t varies in R.

The line is parallel to b. If a = a1 i + a2 j, b = b1 i + b2 j then

x (t) i + y (t) j = a1 i + a2 j + t (b1 i + b2 j)

x (t) i + y (t) j = (a1 + tb1 ) i + (a1 + tb1 ) j

Hence
x(t) = a1 + tb1 , y(t) = a2 + tb2 .

Thus the equation can be written as

x(t) − a1 y(t) − a2
=
b1 b2

provided, that b1 and b2 are non-zero.

Example 18 It is easy to understand that

r(t) = (cos t) i + (sin t) j

13
represents an unit circle.
When t = 0, we have
r(0) = (cos (0)) i + (sin (0)) j = i

When t = π4 , we have
π   π   π √ √
2 2
r = cos i + sin j= i+ j
4 4 4 2 2

When t = π2 , we have
π   π   π
r = cos i + sin j=j
2 2 2
3π
When t = 4
, we have
        √ √
3π 3π 3π 2 2
r = cos i + sin j=− i+ j
4 4 4 2 2
5π
When t = 4
, we have
        √ √
5π 5π 5π 2 2
r = cos i + sin j=− i− j
4 4 4 2 2

In the space R3 , a curve may be analogously represented by a vector function

r(t) = x (t) i + y (t) j + z (t) k

for t ∈ R.

Example 19
r(t) = (cos t) i + (sin t) j + 4tk, 0 ≤ t ≤ 4π

represents a helix in R3 .

14
Definition 7 A path C with parametric representation r (t) is said to be smooth on the interval
[a, b] if r (t) is continuous and r (t) = 0 in [a, b]. A path is said to be piecewise smooth if the
interval [a, b] can be partitioned into a finite number of subintervals in each of which the path is
smooth.

Example 20 Find a piecewise smooth parametrization of the graph C shown as follow

Solution: Because C consists of three line segments C1 , C2 and C3 , we can construct a smooth
parametrization for each segment and piece them together by making the last t-value in Ci corre-
spond to the first t-value in Ci+1 , as follows.

C1 : x (t) = 0, y (t) = 2t, z (t) = 0, 0≤t≤1

C2 : x (t) = t − 1, y (t) = 2, z (t) = 0, 1≤t≤2
C3 : x (t) = 1, y (t) = 2, z (t) = t − 2, 2≤t≤3
Therefore, C is given by


 2tj, 0≤t≤1


r (t) = (t − 1) i + 2j 1≤t≤2



 i + 2j + (t − 2) k 2≤t≤3

15
Because C1 , C2 and C3 are smooth, it follows that C is piecewise smooth.
A curve C may be represented, by more than one path.

Example 21 The following are three continuous paths

√
C1 : r1 (t) = xi + 1 − x2 j x ∈ I = [0, 1]
C2 : r2 (t) = (cos t) i + (sin t) j t ∈ J = [0, π/2]
C3 : r3 (t) = (sin t) i + (cos t) j t ∈ J = [0, π/2]

Suppose r(t) on I = [a, b] and R(u) on I  = [c, d] are two paths representing one and the same
curve C. Then r and R are said to be equivalent. In this case

r(t) = R(u),

for some t ∈ I and for some u ∈ I  . In other words u is a function of t, say, u(t) : I → I  . If u (t) > 0
for all t ∈ I, then from r (t) = R (u) u (t) we see that r (t) and R (u) have the same sign. We
say that r and R trace out the curve in the SAME direction and u is said to be orientation-
preserving. If u (t) < 0 for all t ∈ I, then r (t) and R (u) have different signs. We say that r
and R trace out the curve in the OPPOSITE direction and u is said to be orientation-reversing.
Example 21 shows that r1 and R1 trace out the curve in the same direction while r1 and r2 in the
opposite direction.

Definition 8 When C is described by a continuous path r on [a, b] with r(a) = r(b), then the curve
is said to be closed. A simple closed curve means that it is a closed curve such that r (t1 ) = r (t2 )
for all t1 , t2 ∈ (a, b) and t1 = t2 . In other words distinct values of t in (a, b) lead to distinct points
on the curve. A simple closed curve is also called the Jordan curve.

16
3.6 Line Integrals

Let f (P ) be a continuous scalar function defined on a smooth curve C joining two points A, B in
R2 and
P = {A = P0 , P1 , . . . , Pn−1 , Pn = B}
   
 
a partition of C with Pi ∈ C and P0 Pi  > P0 Pi−1 , where P 0 Pi denotes the arc from P0 to Pi ,

i = 1, 2, , . . . , n.

 

Suppose Pi = (xi , yi ) , ∆si = Pi−1 Pi  (the arc length of the arc joining Pi−1 and Pi . If the limit

n

lim f (Pi∗ ) ∆si
|∆s|→0
i=1

exists, where ∆s = max ∆si and Pi∗ an arbitrary point on C between Pi−1 and Pi (This is corre-
i
sponds to a point t∗i in [ti−1 , ti ] .), then the limit is called the line integral of f along C and

17
denoted by  
f (P ) ds or f (x, y) ds
C C
To evaluate a line integral over a plane curve given by r (t) = x (t) i + y (t) j, use the fact that

   2  2 
 dx dy
ds = (dx)2 + (dy)2 =  + (dt)2
dt dt
 2  2 
 
  dx dy 
= + dt = [x (t)]2 + [y  (t)]2 dt = r (t) dt.
dt dt

Hence   
f (x, y) ds = f (x (t) , y (t)) [x (t)]2 + [y  (t)]2 dt
C C

Example 22 Evaluate C
(2 + xy ) ds, where C is the upper half of unit circle x2 + y 2 = 1.
2

Solution: The parametric form of C is

x = cos t, y = sin t, 0≤t≤π

and
dx d cos t dy d sin t
= = − sin t, = = cos t
dt dt dt dt
Hence
 2  2
  π
    dx dy
2 + xy 2 ds = 2 + cos t sin2 t + dt
0 dt dt
 π  
= 2 + cos t sin2 t (− sin t)2 + (cos t)2 dt
0 π
 
= 2 + cos t sin2 t dt ∵ cos2 t + sin2 = 1
0 π  π
= 2dt + cos t sin2 tdt
0 0
π
1 3
= 2t + sin t
3 0

= 2π

Differential Form
If ∆si is replaced by either ∆xi or ∆yi , then the integrals
 n
f (x, y) dx = lim f (x∗i , yi∗ ) ∆xi (1)
C |∆x|→0
i=1
 n
g (x, y) dy = lim g (x∗i , yi∗ ) ∆yi (2)
C |∆y|→0
i=1

18
are called the line integrals of f along C with respect to x and y.
Frequently, the line integrals in (1) and (2) occur in combination, in which case we dispense
with one of the integral signs and write:
  
f (x, y) dx + g (x, y) dy = f (x, y) dx + g (x, y) dy
C C C

The expression f (x, y) dx + g (x, y) dy is called a differential form.

Example 23 Evaluate 
I= xydx + (y − x) dy

along the paths (a) C1 : y = x from A (0, 0) and B = (1, 1) .

(b) C2 : y = x2
(c) C3 : y 2 = x
(d) C4 : y = x3

Solution:
(a) As xy = x2 , y − x = 0, dy = dx we get
  1  1
2 x3 1
xydx + (y − x) dy = x dx = =
C1 0 3 0 3
(b) As xy = x3 , y − x = x2 − x, dy = 2xdx we get
  1  
 3  3 4 2 3 1 3 2 1
xydx + (y − x) dy = 2
3x − 2x dx = x − x = (1)4 − (1)3 =
C2 0 4 3 0 4 3 12
(c) As xy = y 3 , y − x = y − y 2 , dx = 2ydy we get
  1  
 4 2
 2 5 1 3 1 2 1
xydx + (y − x) dy = 2y − y + y dy = y − y + y
C3 0 5 3 2 0
2 1 1 17
= (1)5 − (1)3 + (1)2 =
5 3 2 30

19
 (d) As xy = x4 , y − x = x3 − x, dy = 3x2 dx we get
  1  1
 1 1 3 
xydx + (y − x) dy = 3x + x − 3x dx = x6 + x5 − x4
5 4 3
C4 0 2 5 4 0
1 1 3 1
= (1)6 + (1)5 − (1)4 = −
2 5 4 20

From the example we see that in general line integrals depend not only on the end points of the
curves but the curves themselves. If a line integral depends only on the end points and not on the
curve itself, then the integral is said to be independent of path. If the statement is true for every
pair of points in a set S, then integral is said to be independent of path in S.

Line Integrals in Space

Suppose that C is a smooth space curve given by the parametric equations

x = x (t) y = y (t) z = z (t) a≤t≤b

or by vector equation
r (t) = x (t) i + y (t) j + z (t) k

If f is a continuous function of three variables that is continuous on some region containing C, then
the line integral of f along C is defined by
 n

f (x, y, z) ds = lim f (x∗i , yi∗ , zi∗ ) ∆si ∆s = max ∆si
C |∆s|→0 i
i=1

Similar to the 2 dimensional case, the integral is evaluated by

   2  2
  b
dx 2 dy dz
f (x, y, z) ds = f (x (t) , y (t) , z (t)) + + dt
C a dt dt dt

20
 
Example 24 Evaluate C
(x + 2) ds, where C is the curve represented by

4 1
r (t) = ti + t3/2 j + t2 k, 0≤t≤2
3 2

Solution: Since
4 1
x (t) = t, y (t) = t3/2 , z (t) = t2
3 2
and
dx dt dy d 4 t3/2 dz d 1 t2
= = 1, = 3 = 2t1/2 , = 2 =t
dt dt dt dt dt dt
Hence
 2  2  2
  2
dx dy dz
(x + 2) ds = (t + 2) + + dt
C 0 dt dt dt
 2 
2
= (t + 2) (1)2 + (2t1/2 ) + (t)2 dt
0 2 √
= 1 + 4t + t2 dt
(t + 2)
0
 2
1  
2 3/2
= 1 + 4t + t
3 0
1 √
= 13 13 − 1
3

Differential Form
The differential form of a line integral in space is

f (x, y, z) dx + g (x, y, z) dy + h (x, y, z) dz
C

Example 25 Evaluate C
ydx + zdy + xdz, where C consists of the line segment C1 from (2, 0, 0)
to (3, 4, 5) followed by the vertical line segment C2 from (3, 4, 5) to (3, 4, 0) .

21
 Solution: The parametric form of C1 is

x (t) = 2 + t, y (t) = 4t, z (t) = 5t, 0≤t≤1

Then
dx = dt, dy = 4dt, dz = 5dt

Thus
  1
ydx + zdy + xdz = (4t) dt + (5t) · 4dt + (2 + t) · 5dt
C1 0
 1
= (10 + 29t) dt
0
 1
29 2
= 10t + t
2 0
49
=
2
C2 can be written in the parametric form

x (t) = 3, y (t) = 4, z (t) = 5 − 5t, 0≤t≤1

Then dx = dy = 0 and dz = −5dt, so

  1  1
ydx + zdy + xdz = 3 · (−5) dt = −15 dt = −15
C2 0 0

Hence
  
49 19
ydx + zdy + xdz = ydx + zdy + xdz + ydx + zdy + xdz = − 15 =
C C1 C2 2 2

3.7 Properties of Line Integrals

1. Constant Multiple:  
kfds = k f ds
C C

2. Sum and difference:   

[f ± g] ds = f ds ± gds
C C C

3. Additivity: If C is a piecewise-smooth curve where C is a union of a finite number of smooth

curves C1 , C2 , . . . , Cn , then the integral of f along C is equal to the sum of the integrals of f
along each of the smooth pieces of C.
   
f ds = f ds + f ds + · · · + f ds
C C1 C2 Cn

22
 4. Let −C denotes the curve consisting of the same points as C but with the opposite orientation,
then  
f ds = − f ds
−C C

3.8 Line Integrals in vector field

Let F = P i + Qj + M k be a vector field. Then

P dx + Qdy + Mdz = (P i + Qj + M k) · (dxi + dyj + dzk) = F · dr

where dr = dxi + dyj + dzk. Thus it is natural to denote the line integral above in form
 
F · dr = P dx + Qdy + M dz
C C

If C is a closed curve, then we write

 
F · dr = F · dr
C C

Assume that the curve C is described by

r = x (t) i + y (t) j + z (t) k

and from A = (x (a) , y (a) , z (a)) to B = (x (b) , y (b) , z (b)) . Then

  b  b
  
P dx + Qdy + M dz = (P x + Qy + M z ) dt = (P i + Qj + Mk) · (x i + y  j + z  k) dt
C a a

= F · r (t) dt
C

In other words,   

F · dr = F · r (t) dt = F · T ds (3)
C C C

r (t)
where T is the unit tangent to curve C and T = .
r (t)
ds
Note: ds = r (t) dt ⇒ dt = .
r (t)

Work Done
Consider an object moving along a path C in the field. To determine the work done by the
force, you need consider only that part of the force that is acting in the same direction as that in
which the object is moving (or the opposite direction). This means that at each point on C, you

23
can consider the projection F · T of the force vector F onto the unit tangent vector T. On a small
subarc of length ∆si the increment of work is

∆Wi = (force)(distance)

≈ [F (x∗i , yi∗ , zi∗ ) · T (x∗i , yi∗ , zi∗ )] ∆si

where T (x∗i , yi∗ , zi∗ ) is a point in the i-th subarc. Consequently the total work done is given by
integral in (3).

Example 26 Find the work done by the force field F (x, y) = x2 i − xyj in moving a particle along
the quarter-circle r (x (t) , y (t)) = cos ti + sin tj, 0 ≤ t ≤ π/2.

Solution: Since x = cos t and y = sin t, we have

F (x (t) , y (t)) = cos2 ti − cos t sin tj

d cos t d sin t
r (t) = i− j = − sin ti + cos tj
dt dt

24
Therefore, the work done is
  π/2
F · dr = F (x (t) , y (t)) · r (t) dt
C 0
 π/2
 2 
= cos ti − cos t sin tj · (− sin ti + cos tj) dt
0
 π/2
 
= −2 cos2 t sin t dt
0
 3 
cos t 2
=2 =−
3 3

Example 27 Find the work done by the force field

1 1 1
F (x, y, z) = − xi − yj + k
2 2 4

on a particle as it moves along the helix given by

r (t) = cos ti + sin tj + tk

from the point (1, 0, 0) to (−1, 0, 3π).

Solution: Since x (t) = cos t, y (t) = sin t and z (t) = t. Thus, the force field can be written as

1 1 1
F (x, y, z) = − cos ti − sin tj + k
2 2 4

To find the work done by the force field in moving a particle along the curve C, use the fact that

d cos t d sin t dt
r (t) = i+ j + k = − sin ti + cos tj + k
dt dt dt

25
and write the following
 
W =F · dr = F (x (t) , y (t) , z (t)) · r (t) dt
C3π  C

1 1 1
= − cos ti − sin tj + k · (− sin ti + cos tj + k) dt
0 2 2 4
 3π  
1 1 1
= sin t cos t − sin t cos t + dt
0 2 2 4
 3π
1
= dt
0 4
 3π
1 3π
= t =
4 0 4

Example 28 Let F(x, y) = yi + x2 j and evaluate the line integral C F · dr for each of the following
parabolic curves.

(a) C1 : r1 (t) = (4 − t) i + (4t − t2 ) j, 0 ≤ t ≤ 3

(b) C2 : r2 (t) = ti + (4t − t2 ) j, 1 ≤ t ≤ 4

d(4−t) d(4t−t2 )
Solution: (a) Because r1 (t) = dt
i + dt
j = −i + (4 − 2t) j and
 
F (x (t) , y (t)) = 4t − t2 i + (4 − t)2 j

the line integral is

 
W = F · dr = F (x (t) , y (t)) · r (t) dt
C3
C
  
= 4t − t2 i + (4 − t)2 j · (−i + (4 − 2t) j) dt
0 3
   
= − 4t − t2 + (4 − t)2 (4 − 2t) dt
0 3
 
= −68t + 21t2 + 64 − 2t3 dt
0
 3
2 3 1 4 69
= −34t + 7t + 64t − t =
2 0 2

26
 dt d(4t−t2 )
(b) Because r2 (t) = dt
i + dt
j = i + (4 − 2t) j and
 
F (x (t) , y (t)) = 4t − t2 i + t2 j

the line integral is

 
W = F · dr = F (x (t) , y (t)) · r (t) dt
C4 C
  
= 4t − t2 i + t2 j · (i + (4 − 2t) j) dt
1
 4   
= 4t − t2 + t2 (4 − 2t) dt
1 4
 
= 4t + 3t2 − 2t3 dt
1
 4
2 3 1 4 69
= 2t + t − t =−
2 1 2

3.9 Path Independence

 
Example 29 Let F(x, y) = yi + xj. Evaluate the line integral C
F · dr = C
ydx + xdy over the
following curves:

(a) the line segment y = x from (0,0) to (1, 1)

(b) the parabola y = x2 from (0,0) to (1. 1)

(c) the cubic y = x3 from (0, 0) to (1,1)

Solution: (a) With x = t as the parameter, the path of integration is given by r(t) = ti + tj
(0 ≤ t ≤ 1). Since F(x, y) = yi + xj, it follows that

F(x(t), y(t)) = ti + tj

27
and
dt dt
r (t) = i+ j=i+j
dt dt
Thus,
 
F · dr = F (x (t) , y (t)) · r (t) dt
C
C1
= (ti + tj) · (i + j) dt
0
 1  1
= 2tdt = t2 0
=1
0

(b) With x = t as the parameter, the path of integration is given by r(t) = ti + t2 j (0 ≤ t ≤ 1).
Since F(x, y) = yi + xj, it follows that

F(x(t), y(t)) = t2 i + tj

and
dt dt2
r (t) = i+ j = i + 2tj
dt dt
Thus,
 
F · dr = F (x (t) , y (t)) · r (t) dt
C
C1
 
= t2 i + tj · (i + 2tj) dt
0
 1  1
= 3t2 dt = t3 0
=1
0

(c) With x = t as the parameter, the path of integration is given by r(t) = ti + t3 j (0 ≤ t ≤ 1).
Since F(x, y) = yi + xj, it follows that

F(x(t), y(t)) = t3 i + tj

and
dt dt3
r (t) = i+ j = i + 3t2 j
dt dt
Thus,
 
F · dr = F (x (t) , y (t)) · r (t) dt
C
C1
3   
= t i + tj · i + 3t2 j dt
0 1
 1
= 4t3 dt = t4 0
=1
0

28
Definition 9 A line integral from a point P0 to a point P1 is said to be independent of path in
a region R containing P0 and P1 if its value is the same for any piecewise smooth, simple curve in
R joining P0 to P1 .

It is clear that if  P1
P dx + Qdy
P0

is independent of path for all points P0 and P1 in R, then


P dx + Qdy = 0
C

for every simple, closed piecewise smooth curve C in R. This follows from the fact that an arbitrary
simple, piecewise smooth, closed curve C may be considered as composed of two simple piecewise
smooth curves C1 , from P0 to P1 , and C2 , from P1 to P0 , for any two points P0 and P1 on C. Then
  
P dx + Qdy = P dx + Qdy + P dx + Qdy
C C1 C2

Now  
P dx + Qdy = − P dx + Qdy
C2 −C2

From path independence  

P dx + Qdy = P dx + Qdy
−C2 C1

Thus the result follows. Let us note that precisely the same definition may be given for path

independence in three dimensions with the corresponding result that C P dx + Qdy + N dz = 0,
where C is any simple, piecewise smooth, closed curve in R. We also note that the converse of this
is true. That is, if the line integral is zero for every simple, piecewise smooth, closed curve, then it
is path independent.

29
Theorem 3 Let P (x, y) and Q(x, y) be defined and continuous in a domain D of the xy-plane.
Consider the line integral 
L= P dx + Qdy
C
where C is a simple, piecewise smooth curve in D.

(a) L is independent of path in D if and only if there is a continuously differentiable function

φ (x, y) such that
P i + Qj = ∇φ

Then, if P0 and P1 are the initial and final points of any path in D,

P dx + Qdy = φ (P1 ) − φ (P0 )
C

(b) If P and Q are continuously differentiable, and L is independent of path in D, then

Py = Qx

Remark:

1. Theorem 3(a) is called the Fundamental Theorem of Line Integrals.

2. The result in Theorem 3(a) is also true in three dimensions, that is, if ∇f = F = F1 i + F2 j +
F3 k, then 
F · dr = f (P1 ) − f (P0 )
C

3. The three-dimensional analog of the result in Theorem 3(b): If C
F · dr is independent of
path, then ∇ × F = 0.

Example 30 Evaluate C F · dr, where C is a piecewise smooth curve from (−1, 4) to (1, 2) and
 
F (x, y) = 2xyi + x2 − y j

30
 Solution: Since
   
∂ x2 y − 12 y 2 + K ∂ x2 y − 12 y 2 + K  
∇f (x, y) = i+ j = 2xyi + x2 − y j
∂x ∂y
thus F (x, y) is the gradient of f (x, y) = x2 y − 12 y 2 + K. By the Fundamental Theorem of Line
Integrals, it follows that

F · dr = f (1, 2) − f (−1, 4)
C
   
2 1 2 2 1 2
= (1) (2) − (2) + K − (−1) (4) − (4) + K
2 2
=4

Example 31 Evaluate C
F · dr, where C is a piecewise smooth curve from (1, 1, 0) to (0, 2, 3) and
 
F (x, y, z) = 2xyi + x2 + z 2 j + 2zyj

Solution: Since
∂ (x2 y + z 2 y + K) ∂ (x2 y + z 2 y + K) ∂ (x2 y + z 2 y + K)
∇f (x, y) = i+ j+ k
∂x ∂y ∂z
 
= 2xyi + x2 + z 2 j + 2zyj

thus F (x, y, z) is the gradient of f (x, y, z) = x2 y + z 2 y + K. By the Fundamental Theorem of Line

Integrals, it follows that

F · dr = f (0, 2, 3) − f (1, 1, 0)
C
   
= (0)2 (2) + (3)2 (2) + K − (1)2 (1) + (0)2 (1) + K

= 17

31
Definition 10 A region is connected if any two points in the region can be joined by a piecewise
smooth curve lying entirely within the region.

Theorem 4 Let F (x, y, z) = F1 i + F2 j + F3 k have continuous first partial derivatives in an open

connected region R, and let C be a piecewise smooth curve in R. The following conditions are
equivalent.

1. F is conservative. That is, F = ∇φ for some function φ.


2. C
F · dr is independent of path.

3. F · dr = 0 for every closed curve C in R.
C

Example 32 Evaluate C1 F · dr, where
   
F (x, y) = y 3 + 1 i + 3xy 2 + 1 j

and C1 is the semicircular path from (0, 0) to (2, 0).

32
 Solution: We have three methods.
Method 1: We can use the parametrization r (t) = (1 − cos t)i + sin tj, where 0 ≤ t ≤ π. For
this parametrization, it follows that dr = (sin ti + cos tj) dt and
  π
 3     
F · dr = sin t + 1 i + 3(1 − cos t) sin2 t + 1 j · (sin ti + cos tj) dt
C1
0 π
 
= sin t + sin4 t + cos t + 3 sin2 t cos t − 3 sin2 t cos2 t dt
0

=2

Method 2: We can try to find a potential function and evaluate the line integral by the Funda-
mental Theorem of Line Integrals. We can find the potential function to be f (x, y) = xy 3 +x+y+K
and, by the Fundamental Theorem,

   
F · dr = (2) (0)3 + (2) + (0) + K − (0) (0)3 + (0) + (0) + K = 2
C1

Method 3: Knowing that F is conservative, we have a third option. Because the value of the line
integral is independent of path, we can replace the semicircular path with a simpler path. Suppose
we choose the straight-line path C2 from (0, 0) to (2, 0) . Then, r (t) = ti, where 0 ≤ t ≤ 2. Thus,
dr = idt and F = (y 3 + 1) i + (3xy 2 + 1)j = i + j, so that
   2
F · dr = F · dr = 1dt = [t]20 = 2
C1 C2 0

3.10 Green’s Theorem

Green’s Theorem gives a relationship between the value of a double integral over a simple con-
nected plane region D and the value of a line integral around the boundary of D.

Definition 11 A simply-connected region in the plane is a connected region D such that every
simple closed curve in D encloses only points that are in D. Notice that a simply-connected region
contains no hole and can’t consist of two separate pieces.

33
Theorem 5 (Green’s Theorem) Let D be a simply connected region with a piecewise smooth
boundary C, oriented counterclockwise. If P and Q have continuous partial derivatives in an open
region containing D, then
   
∂Q ∂P
P dx + Qdy = − dA
C D ∂x ∂y
Interpretation: Green’s theorem is simply a relationship between the macroscopic circulation
around the curve C and the sum of all the microscopic circulation that is inside C. If C is a simple
closed curve in the plane , then it surrounds some region D in the plane. D is the “interior” of the
curve C.

Green’s theorem says that if you add up all the microscopic circulation inside C (i.e., the
microscopic circulation in D), then that total is exactly the same as the macroscopic circulation
around C.

Example 33 Use Green’s Theorem to evaluate the line integral


 
y 3 dx + x3 + 3xy 2 dy
C

where C is the path from (0, 0) to (1, 1) along the graph of y = x3 and from (1, 1) to (0, 0) along the
graph of y = x.

34
 Solution: Let P = y 3 and Q = x3 + 3xy 2 , it follows that

∂Q ∂  3 
= x + 3xy 2 = 3x2 + 3y 2
∂x ∂x
∂P ∂  3
= y = 3y 2
∂y ∂y

applying Green’s Theorem, we have

   
3
 3 2
 ∂Q ∂P
y dx + x + 3xy dy = − dA
C D ∂x ∂y
 1 x
 2 
= 3x + 3y 2 − 3y 2 dydx
3
0 1 x x
= 3x2 dydx
3
0 1 x
 2 y=x
= 3x y y=x3 dx
0
 1
 3 
= 3x − 3x5 dx
0
 
3 4 1 6 x=1
= x − x
4 2 x=0
1
=
4

Example 34 While subject to the force

 
F (x, y) = y 3 i + x3 + 3xy 2 j

a particle travels one around the circle of radius 3. Use Green’s Theorem to find the work done by
F.

35
 Solution: Let P = y 3 and Q = x3 + 3xy 2 , then

∂Q ∂ (x3 + 3xy 2 ) ∂P ∂ (y 3 )
= = 3x2 + 3y 2 and = = 3y 2
∂x ∂x ∂y ∂y

By Green’s Theorem, we have

   
3
 3 2
 ∂Q ∂P
y dx + x + 3xy dy = − dA
C D ∂x ∂y

= 3x2 dydx
D

Use polar coordinates x = r cos θ and y = r sin θ,

 2π  3
W = 3 (r cos θ)2 rdrdθ
0
 2π 0 3
=3 r3 cos2 θdrdθ
0 0
 2π  4 r=3
r 2
=3 cos θ dθ
0 4 r=0
 2π
81
=3 cos2 θdθ
0 4
  
243 2π 1 + cos 2θ
= dθ
4 0 2
 
243 sin 2θ θ=2π
= θ+
8 2 θ=0
243π
=
4

36
Example 35 Evaluate the line integral

 3   
y + ex dx + 3xy 2 + y 4 dy
C

where C is the path as shown.

Solution: Let P = y 3 + ex and Q = 3xy 2 + y 4 , then

∂Q ∂ (3xy 2 + y 4 ) ∂P ∂ (y 3 + ex )
= = 3y 2 = = 3y 2
∂x ∂x ∂y ∂y

By Green’s Theorem, we have

   
 3 x
  2 4
 ∂Q ∂P
y + e dx + 3xy + y dy = − dA
C D ∂x ∂y

 2 
= 3y − 3y 2 dA
 D
= (0) dA = 0
D

Example 36 Evaluate 
 −1   
tan x + y 2 dx + ex − x2 dy
C

where C is the path enclosing the annular region shown as follow

Solution: Let P = tan−1 x + y 2 and Q = ex − x2 , then

∂Q ∂ (ex − x2 ) ∂P ∂ (tan−1 x + y 2 )
= = −2x and = = 2y
∂x ∂x ∂y ∂y

37
By Green’s Theorem, we have
 
 −1 2
  x 2

tan x + y dx + e − x dy = (−2x − 2y) dA
C R

because
∂ (tan−1 x + y 2 ) ∂ (ex − x2 )
= 2y and = −2x
∂y ∂x
Use polar coordinates and the fact that
  π  3
(−2x − 2y) dA = −2r (cos θ + sin θ) rdrdθ
R 0 1
 π  r=3
r3
= −2 (cos θ + sin θ) dθ
0 3 r=1

52 π
=− (cos θ + sin θ) dθ
3 0
52 θ=π
= − [sin θ − cos θ]θ=0
3
104
=−
3

Line Integral for Area

∂Q ∂P
If − = 1, then
∂x ∂y
   
∂Q ∂P
P dx + Qdy = − dA
C R ∂x ∂y

= dA
R

= Area of region R

Corollary 2 If R is a plane region bounded by a piecewise smooth simple closed curve C, oriented
counterclockwise, then the area of R is given by

1
A= xdy − ydx.
2 C

x2 y 2
Example 37 Use a line integral to find the area of the ellipse + 2 = 1.
a2 b

38
 Solution: Use parametric form,

x = a cos t and y = b sin t, 0 ≤ t ≤ 2π.

Therefore, the area is


1
A= xdy − ydx
2 C

1 2π
= [(a cos t) (b cos t) dt − (a sin t) (b sin t) dt]
2 0

1 2π  2 
= ab cos t + sin2 t dt
2 0
= πab

Example 38 Let R be the region inside the ellipse (x2 /9) + (y 2 /4) = 1 and outside the circle
x2 + y 2 = 1. Evaluate the line integral

 
2xydx + x2 + 2x dy
C

where C = C1 + C2 is the boundary of R.

39
 Solution: We introduce the line segments C3 and C4 . Note because the curves C3 and C4
have opposite orientations, the line integrals over them cancel. Furthermore, we can apply Green’s
Theorem to the region R using the boundary C1 + C2 + C3 + C4 to obtain
   
 2  ∂Q ∂P
2xydx + x + 2x dy = − dA
C R ∂x ∂y

= (2x + 2 − 2x) dA
R

=2 dA
R

= 2 (area of R)

= 2 π (3) (2) − π (1)2

= 10π

3.11 Applications

An important application of the line integral of a vector field around a simple closed curve arises in
fluid dynamics. The motion of a fluid (a stream of water, for example) is described by its velocity
field F, where F (r) is the velocity of the fluid at the point with position vector r. Imagine a small
cork moving with the fluid. Then, besides moving downstream, the cork may also rotate about a
vertical axis. It occurs less dramatically in a stream as illustrated in the following figure, where the
velocity of the fluid is rather greater near one bank than near the other. It does not occur at all in
an evenly flowing stream.

Let F be a continuous vector field modelling the velocity on the surface of a fluid. Let C be an

oriented simple closed curve in S. The line integral C F · dr can be interpreted as the integral of

40
the tangential component of F around C. If C is the simple closed curve where the cork breaks the
surface of the liquid we would expect the cork to rotate in the direction of the orientation of C if

C
F · dr is positive, and in the opposite direction (or not at all) if this integral is negative (or zero).

The integral C F · dr is called the circulation of F around C.
Suppose that the curve C is a circle, centre O, radius c. We obtain an absolute measure of the
cork’s rate of rotation (about a vertical axis through O) by calculating the mean angular velocity
about O of the fluid around the circle C. Using the relation v = cω between the tangential component
of velocity v at a point P on C and the angular velocity ω of P about O, we obtain
  
1 F
mean angular velocity around C = · dr
length of C C c

1
= F · dr
2πc2 C

Notice that πc2 is the area enclosed by the circle C.

Generalizing, let C be an arbitrary oriented simple closed curve in S. The integral

1
F · dr
area enclosed by C C

is called the rotation of F around C. It provides a measure in radians per unit time of twice the
net rate of turning of the fluid around C.

Example 39 Let F be defined by

F (x, y) = −yi

and let C let be the counterclockwise oriented unit circle x2 + y 2 = 1, as pictured below.

Use the parametric form of x2 + y 2 = 1, that is x = cos t,y = sin t, dx = − sin t, dy = cos t. The

41
rotation of F around C is
 
1 1
F · dr= (−yi) · (dxi+dyj)
π C π C

1
= −ydx
π C

1 2π
=− (sin t) (− sin t) dt
π 0

1 2π 2
= sin tdt
π 0

1 2π 1 − cos 2t
= dt = 1
π 0 2

Example 40 Let F be defined by

F (x, y) = xi + yj

and let C let be the counterclockwise oriented unit circle x2 + y 2 = 1, as pictured below.

Use the parametric form of x2 + y 2 = 1, that is x = cos t,y = sin t, dx = − sin t, dy = cos t. The
rotation of F around C is
 
1 1
F · dr= (xi+yj) · (dxi+dyj)
π C π C

1
= xdx + ydy
π C

1 2π
= (cos t) (− sin t) dt + (sin t) (cos t) dt
π 0

1 2π
= (− cos t sin tdt + sin t cos t) dt
π 0
=0

Example 41 A coil spring lies along the helix

r (t) = (cos 4t) i + (sin 4t) j + tk

42
The spring’s density is ρ (x, y, z) = (π − z)2 . Find the spring’s mass and center of mass, and its
moment of inertia about the z-axis.

Solution: x = cos 4t, y = sin 4t, z = t ⇒ dx

dt
= −4 sin 4t, dy
dt
= 4 cos 4t, dz
dt
=1
 2  2  2
  2π
2 dx dy dz
M= ρ (x, y, z) ds = (π − z) + + dt
C 0 dt dt dt
 2π 
2
= (π − t) (−4 sin 4t)2 + (4 cos 4t)2 + (1)2 dt
0
 2π √
= (π − t)2 17dt
0
2 √
= π 3 17
3

 2  2  2
  2π
2 dx dy dz
Mxy = zρ (x, y, z) ds = z (π − z) + + dt
C 0 dt dt dt
 2π √
= t (π − t)2 17dt
0
2 √
= π 4 17
3

By symmetry, x̄ = ȳ = 0 and
Mxy
z̄ = =π
M

 2 
Iz = x + y 2 ρ (x, y, z) ds
C2π √
 
= (cos 4t)2 + (sin 4t)2 (π − t)2 17dt
0
2 √
= π 3 17
3

43
 
f (x, y, z) C
f (x, y, z) ds
f (x, y, z) = 1 Length of the curve C
f (x, y, z) = ρ (x, y, z) = density Mass of the wire C
f (x, y, z) = xρ (x, y, z) Myz = First moment about the yz-plane
f (x, y, z) = yρ (x, y, z) Mxz = First moment about the xz-plane
f (x, y, z) = zρ (x, y, z) Mxy = First moment about the xy-plane
xρ (x, y, z)
f (x, y, z) =  x̄ = center of gravity(x-coordinate) of wire C
C
ρ (x, y, z) ds
yρ (x, y, z)
f (x, y, z) =  ȳ = center of gravity(y-coordinate) of wire C
C
ρ (x, y, z) ds
zρ (x, y, z)
f (x, y, z) =  z̄ = center of gravity(z-coordinate) of wire C
C
ρ (x, y, z) ds
x
f (x, y, z) =  x̄ = centroid(x-coordinate) of wire C
C
ds
y
f (x, y, z) =  ȳ = centroid(y-coordinate) of wire C
C
ds
z
f (x, y, z) =  z̄ = centroid(z-coordinate) of wire C
C
ds
f (x, y, z) = (y 2 + z 2 ) ρ (x, y, z) Ix = moment of inertia about the x-axis
f (x, y, z) = (x2 + z 2 ) ρ (x, y, z) Iy = moment of inertia about the y-axis
f (x, y, z) = (x2 + y 2 ) ρ (x, y, z) Iz = moment of inertia about the z-axis
f (x, y, z) = (r (x, y, z))2 ρ (x, y, z) IL = moment of inertia about a line L
where r (x, y, z) = distance from (x, y, z) to L

3.12 Surfaces

We have seen that the graph of a function of two variables is a surface in space. Such surfaces
are known as simple surfaces. More precisely, given any function g (x, y), the equation z = g (x, y)
defines a surface S in the xyz-space, having the property that every line parallel to the z-axis may
intersect the surface at no more than one point. Such a surface S is called a z-simple surface.

44
 Similarly, equations such as y = h(x, z) or x = l(y, z), where h and l are functions of x, z and y, z
respectively, define y-simple and x-simple surfaces. It is clear that every straight line parallel to
the y-axis (x-axis) may intersect a y-simple (x-simple) surface at no more than one point.

We also learned that given a scalar function F (x, y, z), the equation

F (x, y, z) = C (constant)

defines a surface in space. Such a surface is called a level surface of the given function F (x, y, z).
It is clear that every simple surface may be regarded as a level surface.

45
 For instance, a z-simple surface of the form z = g(x, y) may be re-written as g(x, y) − z = 0,
which is a level surface of the function

F (x, y, z) = g(x, y) − z.

Example 42 A sphere with centre at the origin and radius a > 0 satisfies the equation x2 +y 2 +z 2 =
a2 . It may be decomposed into two disjoint z-simple surfaces, the upper hemisphere


z= a2 − x2 − y 2

and the lower hemisphere

z = − a2 − x2 − y 2

Normal Vector to a Surface

If P0 = (x0 , y0 , z0 ) is a point on a simple surface S with equation given by z = g(x, y) (or
F (x, y, z) = g (x, y) − z = 0), then a normal vector at P0 to S may be obtained by taking gradient,
i.e.
n = ∇ [g (x, y) − z]P0 = gx (x0 , y0 ) i + gy (x0 , y0 ) j − k

or
n = ∇ [F (x, y, z)]P0 = Fx (x0 , y0 , z0 ) i + Fy (x0 , y0 , z0 ) j + Fz (x0 , y0 , z0 ) k

Sometimes, we need to use unit normal vectors and they are given by
gx i + gy j − k
n=  (4)
(gx )2 + (gy )2 + 1
or
Fx i + Fy j + Fz k
n=  (5)
(Fx )2 + (Fy )2 + (Fz )2

46
Example 43 For the surface z = x2 + y 2 , we may write z = g (x, y) = x2 + y 2 or F (x, y, z) =
x2 + y 2 − z = 0. The normal vector is calculated by either

∂ (x2 + y 2 ) ∂ (x2 + y 2 )
n = gx (x0 , y0 ) i + gy (x0 , y0 ) j − k = i+ j−k
∂x ∂y
= 2xi + 2yj − k

or

n = Fx (x0 , y0 , z0 ) i + Fy (x0 , y0 , z0 ) j + Fz (x0 , y0 , z0 ) k

∂ (x2 + y 2 − z) ∂ (x2 + y 2 − z) ∂ (x2 + y 2 − z)
= i+ j− k
∂x ∂y ∂z
= 2xi + 2yj − k

The unit vector is

2xi + 2yj − k 2xi + 2yj − k
n=  =

(2x)2 + (2y)2 + (−1)2 4x2 + 4y 2 + 1

Parametrized Surfaces
It is sometime convenient to represent a surface by parametric equations. Consider a surface S
as the locus of a point P = (x, y, z), where

x = X(u, v), y = Y (u, v), z = Z(u, v) (6)

Here, X(u, v), Y (u, v) and Z(u, v) are continuously differentiable functions defined for (u, v) be-
longing to some region Ω in the uv-plane such that

[Xu (u, v)]2 + [Yu (u, v)]2 + [Zu (u, v)]2 = 0

and
[Xv (u, v)]2 + [Yv (u, v)]2 + [Zv (u, v)]2 = 0

Eq. (6) is known as a parametric representation of the surface S.

−→
Let r be the position vector of P , i.e. r = OP = xi + yj + zk. Eq. (6) may then be re-written
in a more condensed form as

r = R (u, v) = X (u, v) i + Y (u, v) j + Z (u, v) k

47
Definition 12 If ru × rv is not 0, then the surface S is called smooth (no corners).

Definition 13 A smooth surface S is said to be orientable if, to each point on S, we can assign a
unit normal vector in such a way that the resulting vector field of normals is continuous everywhere
on S.

Example 44 Common surfaces such as spheres and ellipsoids are orientable.

Example 45 An example of a non-orientable surface is the Mobius band.

A Mobius band, gotten by taking a strip of paper, twisting it around once, and then joining the ends,

48
is not orientable. Note that if we start at P0 , proceed around the strip in the direction of the arrow
while varying n continuously, and come back to P0 , the direction of n will be reversed.

Definition 14 If S can be subdivided into finitely many smooth portions, it is called a piecewise
smooth surface. e.g. the faces of a cube.

Definition 15 A piecewise smooth surface S, consisting of a finite number of smooth surfaces

joined by piecewise smooth, simple curves, is said to be orientable if:

(a) Each smooth piece is orientable.

(b) An orientation may be defined on each part such that the positive directions induced on the
common boundary curves are opposite.

Example 46 A typical piecewise smooth orientable surface is depicted in the following figure, where
an orientation is shown for a closed cylindrical shell.

Example 47 A sphere with centre at (0, 0, 0) and radius equal to a > 0 has parametric equations
given by
x = a sin u sin v, y = a sin u sin v, z = a cos v

where 0 ≤ u ≤ 2π and 0 ≤ v ≤ π.

49
Example 48 Find a parametrization of the paraboloid z = x2 + y 2 .

Solution: We may take x = u, y = v and z = u2 + v2 OR x = v cos u, y = v sin u and z = v 2 .

Remark: Let S be a surface with parametric equation

r = R (u, v) = X(u, v)i + Y (u, v)j + Z(u, v)k,

and let P0 = (x0 , y0 , z0 ) be a point on S, with position vector r0 = R (u0 , v0 ).

It is clear that curves with equations given by

r = R (u, v0 ) and r = R (u0 , v)

lie on S, and they intersect each other at P0 . Since Ru (u0 , v0 ) and Rv (u0 , v0 ) are tangent vectors
to these curves at P0 , it follows that the vector

n = Ru (u0 , v0 ) × Rv (u0 , v0 )

is normal to the surface S at P0 .

Example 49 Consider a sphere with centre at the origin and radius equal to a. It has parametric
equations given by
r = (a cos u sin v) i + (a sin u sin v) j + (a cos v)k.

A simple calculation shows that

dr
= (−a sin u sin v) i + (a cos u sin v) j
du
dr
= (a cos u cos v) i + (a sin u cos v) j + (−a sin v)k
dv

50
Thus
dr dr
n= ×
du dv 
 
 i j k 
 
 
=  (−a sin u sin v) (a cos u sin v) 0 
 
 
 (a cos u cos v) (a sin u cos v) (−a sin v) 
 
     i j


     
 (a cos u sin v) 0   (−a sin u sin v) 0   
=  i− 
 
 j+  (−a sin u sin v) (a cos u sin v)
  k
 (a sin u cos v) (−a sin v)   (a cos u cos v) (−a sin v)   

 (a cos u cos v) (a sin u cos v) 
     
= a2 sin2 v cos u i + a2 sin2 v sin u j + a2 cos v sin v k = (a sin v) r.

In other words, n is a scalar multiple of r.

3.13 Surface Area

Let f be a differentiable function of two variables defined on

 a domain D. We wish to find the
surface area of the graph of f over D. It is simply equal to dS. Therefore we need to express
D
the differential of the surface area dS in terms of the differential dA of the domain. To do so,
take any point P  (x, y) in D and let P be the corresponding point on the graph of f. Consider
an increment dx along the x-direction and an increment dy along the y-direction at the point P  .
Thus dA = |dxdy|. These increments sweep out an increment of surface area on the surface at P .
The differential dS of this area at P is given by the corresponding area on the tangent plane to the
surface at P .

−→ −→
Let P Q be the vector on the tangent plane at P with x-component dx, and P R the vector with
−→ −→
y-component dy. Thus, P Q = dxi + fx (x, y)dxk and P R = dyj + fy (x, y)dyk. The area of the

51
 −→ −→ −→ −→
parallelogram spanned by P Q and P R is the magnitude of the cross product P Q × P R.
 
 
 i j k 
−→ −→   

P Q × P R =  dx 0 fxdx  = (−fx i − fy j + k) dxdy
 
 
 0 dy fy dy 


Therefore, dS = (−fx i − fy j + k) dxdy = fx2 + fy2 + 1dA. Consequently,
 
Surface area = fx2 + fy2 + 1dA.
D

Example 50 Find the surface area of the portion of the plane 2x + y + z = 10 contained in the
first octant of the xyz-space.

Solution: Equation of the plane is given by z = 10 − 2x − y = f (x, y) and the projection onto
the xy-plane is the triangle with vertices (0, 0), (5, 0) and (0, 10). Therefore, the required area is
equal to
   
2 2
(fx ) + (fy ) + 1dxdy = (−2)2 + (−1)2 + 1dxdy
Rxy Rxy
√  5  10−2x
= 6 dydx
0 0
√  5
= 6 (10 − 2x) dx
0
√
= 25 6

Remark: If S is given by F (x, y, z) = 0, then

∇F  ∇F  ∇F 
dS = dxdy = dydz = dzdx
|Fz | |Fx | |Fy |

52
provided that n = ∇F is nowhere parallel to the respective coordinate planes.
Consequently, the surface integral is equal to the double integral
!
 ∂F 2  ∂F 2  ∂F 2
  ∂x
+ ∂y + ∂z
dS =  ∂F  dxdy, etc. (7)
S R
 
∂z

Example 51 Find the surface area of the portion of the plane 2x + y + z = 10 contained in the
first octant of the xyz-space.

Solution: Let F (x, y, z) = 2x + y + z − 10 = 0

∂F ∂F ∂F
= 2, = 1, =1
∂x ∂y ∂z

!  2
 ∂F 2  ∂F 2
  ∂x
+ ∂F
∂y
+ ∂z
dS =  ∂F  dxdy
S R
 
∂z

 (2)2 + (1)2 + (1)2
= dxdy
R 1
 10  10−y
√
2
= 6dxdy
0 10 "√
0
# 10−y
2
= 6x dy
0 0
 10 √  
10 − y
= 6 dy
0 2
  10
√ 1 2 √
= 6 5y − y = 25 6
4 0

Area Formula for parametrized surfaces

53
 Suppose now that S is a surface with parametric representation

r = r (u, v) = X (u, v) i + Y (u, v) j + Z (u, v) k

where (u, v) ∈ Ω. Consider a small piece ∆S of S. If ∆S can be projected onto a region ∆A of the
uv-plane, then
∆S = ru ∆u × rv ∆v = ru × rv  ∆u∆v (8)

By taking limit ∆S → 0, we have

dS = ru × rv  dudv. (9)

We obtain the following formula for computing surface area of a parametrized surface:

Surface area of S = ru × rv  dudv (10)

Remark: Using the vector identity a × b2 = a2 b2 − (a · b)2 , we have

ru × rv 2 = ru 2 rv 2 − (Ru · rv )2 = EG − F 2 ,

where

E = ru 2 = (Xu )2 + (Yu )2 + (Zu )2

G = rv 2 = (Xv )2 + (Yv )2 + (Zv )2

F = ru · rv = Xu Xv + Yu Yv + Zu Zv

With this notation, the surface area formula Eq. (10) may be re-rewritten as
 √
Surface area of S = EG − F 2 dudv (11)

54
Example 52 For a sphere with radius a and centre at the origin, we have X = a sin v cos u, Y =
a sin v sin u and Z = a cos v.

Xu = −a sin v sin u, Yu = a sin v cos u, Zu = 0

Xv = a cos v cos u, Yv = a cos v sin u, Zv = −a sin v

Therefore, one has E = a2 , G = a2 sin2 v and F = 0. By Eq. (11), the surface area of the sphere is
given by
 2π  π √  2π  π
 2π  π
2 2
EG − F 2 dudv = a4 sin vdvdu = a sin vdvdu
0 0 0 0 0 0
 2π
=a 2
[− cos v]v=π
v=0 du = 4πa
2
0

3.14 Surface Integrals

Let f (x, y, z) be a continuous function and S be a surface. We shall define the surface integral of
f (x, y, z) over S as follows:

1. Sub-divide S into small pieces ∆S1 , ∆S2 , ∆S3 , . . . , ∆Sn .

2. For each j = 1, 2, 3, . . . , N . choose a point (xj , yj , zj ) on ∆Sj and consider the sum

f (xj , yj , zj ) ∆Sj

3. Define the integral of f (x, y, z) over S by

 
f (x, y, z) dS = lim f (xj , yj , zj ) ∆Sj
S ∆S→0

where ∆S = maxj ∆Sj . It is important to note that the limit should be independent of the
way S is being subdivided.

Definition 16 S
f (x, y, z) dS is known as the surface integral of f (x, y, z) over S.

Remark: Let S a thin shell made by metal with non-uniform density δ, i.e., S is a function of

x, y and z. It follows that S δdS represents the total mass of the shell.
Properties of surface integrals

1.   
[f (x, y, z) ± g (x, y, z)] dS = f (x, y, z) dS ± g (x, y, z) dS
S S S

55
 2.  
kf (x, y, z) dS = k f (x, y, z) dS
S S

3.   
f (x, y, z) dS = f (x, y, z) dS + f (x, y, z) dS
S1 ∪S2 S1 S1

if the interiors of S1 and S2 are non-overlapping.

Evaluation of surface integrals

Method 1: If S is a surface with parametric representation

r = r(u, v) = X(u, v)i + Y (u, v)j + Z(u, v)k.

√
where (u, v) ∈ Ω, then dS = EG − F 2 dudv and
  √
f (x, y, z) dS = f (X (u, v) , Y (u, v) , Z (u, v)) EG − F 2 dudv (12)
S Ω

Method 2: If S is a simple surface given by an equation z = g (x, y) , then

  
f (x, y, z) dS = f (x, y, g (x, y)) (gx )2 + (gy )2 + 1dxdy (13)
S Ω

Method 3: If S is given by F (x, y, z) = 0, then

∇F  ∇F  ∇F 
dS = dxdy = dydz = dzdx
|Fz | |Fx | |Fy |
provided that n = ∇F is nowhere parallel to the respective coordinate planes.
Consequently, the surface integral is equal to the double integral
!
 ∂F 2  ∂F 2  ∂F 2
  + ∂y + ∂z
∂x
f (x, y, z) dS = f (x, y, z)  ∂F  dxdy, etc. (14)
S R
 
∂z

56
 
Example 53 Evaluate S
x2 y 2 zdS where S is the upper hemisphere with radius equal to a.

Solution: S has parametric equations given by X = a sin v cos u, Y = a sin v sin u and Z =
π
a cos v, where 0 ≤ v ≤ 2
and 0 ≤ u ≤ 2π. By Method 1, we have

Xu = −a sin v sin u, Yu = a sin v cos u, Zu = 0

Xv = a cos v cos u, Yv = a cos v sin u, Zv = −a sin v

√
With EG − F 2 = a2 sin v, one has
  2π  π/2
2 2
x y zdS = a7 sin5 v cos v cos2 u sin2 udvdu
S 0 0
 2π  π/2
  
= a7 sin5 v cos v cos2 u sin2 u dvdu
0 0
 π/2  2π
7
 5   2 
=a sin v cos v dv × cos u sin2 u du
0 0
  2π
1 1 1
= a7 u− sin 2u
6 8 16 0
πa7
=
24
π
Note: u = sin v ⇒ du = cos vdv, when v = 0 ⇒ u = sin 0 = 0; when v = 2
⇒ u = sin π2 = 1.
 π/2  1  u=1
 5  5 1 6 1
sin v cos v dv = u du = u =
0 0 6 u=0 6
and
    
 2 2
 1 1 2 1 − cos 2u 1 1 1
cos u sin u du =
sin 2udu = du = u − sin 2u + C
4 4 2 4 2 4

Example 54 Evaluate the surface integral S (y 2 + 2yz) dS where S is the first-octant portion of
the plane 2x + y + 2z = 6.

57
 Solution: Begin by writing S as

F (x, y, z) = 2x + y + 2z − 6 = 0

Using the partial derivatives Fx (x, y, z) = 2, Fy (x, y, z) = 1 and Fz (x, y, z) = 2, you can write
!
 ∂F 2  ∂F 2  ∂F 2 
+ + (2)2 + (1)2 + (2)2
∂x ∂y ∂z 3
 ∂F  = = .
  2 2
∂z
Hence,
!  2
 ∂F 2  ∂F 2
  + ∂F
+
  ∂x ∂y ∂z
y 2 + 2yz dS =
f (x, y, z)  ∂F  dxdy
S R
 
∂z
   
2 1 3
= y + 2y (6 − 2x − y) dxdy
R 2 2
 3  2(3−x)
=3 y (3 − x) dydx
0 0
 3
=6 (3 − x)3 dx
0
 3
3 4 243
= − (3 − x) = .
2 0 2


Example 55 Evaluate S zdS where S is the portion of the cone z = x2 + y 2 below the plane
z = 1.

Solution: The projection of S onto the xy-plane is the region within the unit circle x2 + y 2 = 1.
1/2 1/2

∂ (x2 + y 2 ) x ∂ (x2 + y 2 )
2 2
Let z = g (x, y) = x + y , then gx = =
and gy = =
∂x x2 + y 2 ∂y
y

. We conclude from Method 2 that
x + y2
2
  
zdS = f (x, y, g (x, y)) (gx )2 + (gy )2 + 1dxdy
S R
 2  2


 x y
= x2 + y 2 
+
+ 1dxdy
Rxy x2 + y 2 x2 + y 2


x2 y2
= x2 + y 2 + + 1dxdy
Rxy x2 + y 2 x2 + y 2


x2 + y 2
= x2 + y 2 + 1dxdy
Rxy x2 + y 2

√
= x2 + y 2 2dxdy
Rxy
√
√  2π  1 2 2 2π
= 2 r drdθ =
0 0 3

58
3.15 Surfaces Integrals of Vector Fields

Definition 17 Let F (x, y, z) = F1 (x, y, z) i + F2 (x, y, z) j + F3 (x, y, z) k be a vector field, and

suppose that dS is a small piece of surface with n being a unit normal vector on dS. Then the scalar
(F · n) dS is known as the flux of the vector field F across dS. More generally, if S is a surface
and n is a unit normal vector defined on S, then the surface integral

(F · n) dS
S

is known as the flux of F across S.

Example 56 Find the flux of the vector field F = xi + yj + 3k out of the region T bounded by the
paraboloid z = x2 + y 2 and by the plane z = 4.

Solution: Let S1 denote the circular top, which has outer unit normal vector n1 = k. Let S2
be the parabolic part of this surface, with outer unit normal vector n2 . The flux across S1 is
  
F · n1 dS = (xi + yj + 3k) ·kdS = 3dS = 12π
S1 S1 S1

59
because S1 is a circle of radius 2.
∂g ∂g
Let z = g (x, y) = 2x + y − 6, then gx = = 2x and gy = = 2y. By Eq. (4),
∂x ∂y
gx i + gy j − k 2xi + 2yj − k
n2 =
2 =

gx + gy2 + 1 4x2 + 4y 2 + 1
n2 is the unit outer normal vector. With parameters (x, y) in the circular disk x2 + y 2 ≤ 4 in the



xy-plane, the surface area element is dS = gx2 + gy2 + 1dxdy = 4x2 + 4y 2 + 1dxdy. Therefore,
the outward flux across S2 is
 
2xi+2yj − k
2
F · n2 dS = (xi + yj + 3k) ·
4x + 4y 2 + 1dxdy
S2 R 4x2 + 4y 2 + 1

 2 
= 2x + 2y 2 − 3 dxdy
R

We change to polar coordinates and find that

  2π  2  2
 2  1 4 3 2
F · n2 dS = 2r − 3 rdrdθ = 2π r − r = 4π
S2 0 0 2 2 0
Hence the total flux of F out of T is 16π.
Remark: The above definition is motivated by following examples in physics.

1. Let S be a surface immersed in a fluid with non-uniform density ρ(x, y, z) and velocity vector

v (x, y, z). If n is a unit normal vector on S, then S (ρv · n) dS is the quantity of fluid flowing
across the surface S per unit time.

Example 57 Let S be that portion of the paraboloid z = 4 − x2 − y 2 lying above the xy-plane,
oriented by an upward unit normal vector. A fluid of constant density ρ is flowing through the
surface S according to the velocity field v (x, y, z) = xi + yj + zk. Find the rate of mass flow through
S.

60
 Solution: Projecting S onto the xy-plane, we have the following.

S : z = 4 − x2 − y 2 = g (x, y)

R : x2 + y 2 ≤ 4
∂ (4−x2 −y2 ) ∂ (4−x2 −y2 )
Then gx = ∂x
= −2x and gy = ∂y
= −2y
gx i + gy j − k 2xi + 2yj + k
Unit normal: n = −  =
. This is the unit outer normal vector.
2 4x 2 + 4y 2 + 1
2
(gx ) + (gy ) + 1


And dS = (gx2 ) + (gy )2 + 1dA = 4x2 + 4y 2 + 1dA
 
2xi + 2yj + k 2x2 + 2y 2 + z
v · n = (xi + yj + zk) ·
=

4x2 + 4y 2 + 1 4x2 + 4y 2 + 1
Therefore, the rate of mass flow through S is
   


2x2 + 2y 2 + z
ρv · ndS = ρ
4x2 + 4y 2 + 1dA
S R 4x2 + 4y 2 + 1

 2  
=ρ 2x + 2y 2 + 4 − x2 − y 2 dA
 R
 
=ρ 4 + x2 + y 2 dA
R
 2π  2
 
=ρ 4 + r2 rdrdθ
0 2π 0
=ρ 12dθ = 24πρ.
0

2. Let E (x, y, z) denote the electrostatic field at any point (x, y, z) induced by a point charge
situated at the origin. Then
r xi + yj + zk
E = KQ = KQ (15)
r 3
(x2 + y 2 + z 2 )3/2
where K is a physical constant depending on the units chosen. If S is a surface which does

not contain the origin and if n is a unit normal vector on S, then S (E · n) dS is known as
the electric flux across S.

Example 58 If S is a sphere with centre at the origin and radius a > 0 and if E is the electrostatic
field defined in Eq. (15), then the electric flux across S is given by
  
r r KQ
(E · n) dS = KQ 3 · dS = 4
(xi + yj + zk) · (xi + yj + zk) dS
S r a S a
S
KQ  2 2 2

= 4
x + y + z dS
S a
 
KQa2 KQ KQ  2

= dS = dS = 4πa = 4πKQ.
S a4 a2 S a2

61
It is interesting to note that the electric flux is independent of the radius of the sphere.

3.16 The Divergence Theorem of Gauss

In this section, we shall describe an important theorem (due to Gauss) which states that the net
outward flux of a vector field across a closed surface is equal to the triple integral of the divergence
of the field over the region enclosed by the surface. Recall that the divergence of a vector field

F (x, y, z) = F1 (x, y, z) i + F2 (x, y, z) j + F3 (x, y, z) k

is a scalar field defined by

∂F1 ∂F2 ∂F3
∇ · F = div F = + +
∂x ∂y ∂z

Theorem 6 (Divergence Theorem of Gauss) Let S be a closed surface which encloses a solid
region R in space. Let n be the outward unit normal vector defined on S, and let F be a vector field
with continuous partial derivatives over R. Then
 
(F · n) dS = (∇ · F) dV
S R

Interpretation: If a vector field F represents the flow of a fluid, then the divergence of F
represents the expansion or compression of the fluid. The divergence theorem says that the total
expansion or compression of the fluid inside some three-dimensional region R equals the total flux
of the fluid out of the boundary of R.

Example 59 Verify the Divergence Theorem for the vector field F = xi + yzk over the sphere
x2 + y 2 + z 2 = a2 .

62
 Solution: On the sphere S, let F (x, y, z) = x2 + y 2 + z 2 − a2 = 0. Then
∂ (x2 + y 2 + z 2 − a2 )
Fx = = 2x
∂x
∂ (x2 + y 2 + z 2 − a2 )
Fy = = 2y
∂y
∂ (x2 + y 2 + z 2 − a2 )
Fz = = 2z
∂z
Fx i+Fy j+Fz k
By Eq. (5), we have unit normal n = √ 2 2 2
= √xi+yj+zk = xi+yj+zk
√
a2
= xi+yj+zk
a
.
2 Fx +Fy +Fz
2 2 x +y +z
√
dS = EG − F 2 = (a sin v) dvdu

Therefore,
   
xi + yj + zk
(F · n) dS = (xi + yzk) · dS
S S a
  2 
x + yz 2
= dS
S a
 2π  π  
(a cos u sin v)2 + (a sin u sin v) (a cos v)2
= (a sin v) dvdu
0 0 a
 2π  π
 2 
= a cos2 u sin3 v + a3 sin u sin2 v cos2 v dvdu
0 0
 2π  π  2π  π
 2 2 3
  3 
= a cos u sin v dvdu + a sin u sin2 v cos2 v dvdu
0
 2π0  π 0
 2π0  π
2 2
 3
 3
 2 
=a cos udu sin v dv + a sin udu sin v cos2 v dv
0 0 0 0
 π
4  
= a2 × π × + a3 (0) × sin2 v cos2 v dv
3 0
4π 2
= a.
3
Note:  2π
sin udu = [− cos u]2π
0 = − cos 2π − (− cos 0) = 0
0

and
 2π  2π  2π
2 1 + cos 2u 1 sin 2u
cos udu = du = u +
0 0 2 2 4 0
 π  π  π
 
sin3 vdv = − sin2 vd (cos v) = − 1 − cos2 u d (cos v)
0
 0 π 0
cos3 u
= − cos u +
3 0
3
 
cos π cos3 0 4
= − cos π + − − cos 0 + =
3 3 3

63
 On the other hand, since ∇ · F = 1 + y, we have
 
(∇ · F) dV = (1 + y) dxdydz
D D
  2π  π  a
= dxdydz + ρ sin θ (sin φ) ρ2 sin φdρdθdφ
D
 2π  π 0 a 0 0
4π 3  
= a + ρ3 (sin θ) sin3 φ dρdθdφ
3
0 a 0 0  π   2π 
4π 3 3 3
= a + ρ dρ sin θdθ sin φdφ
3 0 0 0
4π 3 4π 3
= a +0= a
3 3
 2π  
because 0 sin3 φ dφ = 0.
Applications:

1. Taking F = r in the Divergence Theorem, we obtain

  
(r · n) dS = (∇ · F) dV = 3dxdydz = 3 × volume of D.
S D D

In other words, 
1
volume of D = × (r · n) dS (16)
3 S

This is known as the volume formula for a solid region.

Example 60 Let S be the sphere defined by x2 + y 2 + z 2 = 1. Eq. (16) now implies that

(r · n) dS = 3 × volume of the sphere = 4πa3 .
S

Note that direct evaluation of the surface integral is more difficult.

2. The Continuity Equation of Hydrodynamics

Let B be any sphere with boundary surface S. Suppose B is immersed in a fluid with time-
dependent density ρ (x, y, z, t) and velocity vector v (x, y, z).

If n is the outward unit normal vector on S, then the surface integral S
(ρv · n) dS is the
rate at which mass of fluid leaves B across S. Since the mass of fluid contained in B is equal

to B
ρdxdydz, using the Divergence Theorem, we obtain
  
d
− ρdxdydz = (ρv · n) dS = ∇ · (ρv) dxdydz,
dt B S B

64
 where the minus sign indicates loss of fluid from B. As the above identity holds for any sphere
in the fluid, we may now conclude that

∂ρ ∂ρ
− = ∇ · (ρv) , or ∇ · (ρv) + =0 (17)
∂t ∂t

everywhere inside the fluid. Eq. (17) is known as the continuity equation.

3.17 Stokes’ Theorem

Definition 18 A surface S is said to be a oriented surface if it is possible to define a field n of

unit vectors that varies continuously with position. Once such a vector field n is chosen, the side on
which n stands is known as the positive side of S. It is clear that −n would be another continuous
unit normal vector field of S. The side of S where −n stands is called the negative side of S.

When S is a closed surface, the usual convention is to choose n to be the normal vector that
points outward (called the outward unit normal). On the other hand, when S is a oriented open
surface, then S has a boundary curve C which separates the positive side from the negative side. C
is said to be positively oriented if the chosen unit normal vector n always lies on one’s left hand
side as one travels along C while standing on the positive side of S.

Example 61 Let S be the upper hemisphere defined by x2 + y 2 + z 2 = 1 and z > 0. If n is chosen

to be the unit normal vector with positive k component, then the boundary curve C is the unit
circle on the xy-plane, which may be parametrized by the parametric equations x = cos t, y = sin t
(0 ≤ t ≤ 2π) when oriented positively.

The following theorem, due to Stokes, may be regarded as a generalization of Green’s Theorem
in 3 dimensions.

65
Theorem 7 (Stokes’ Theorem) Let S be an open oriented surface with a chosen unit normal
vector n and a positively oriented boundary curve C. Let F (x, y, z) be a smooth vector field defined
on S, then  
F · dr = (∇ × F) · ndS
C S

Stokes’ theorem is a generalization of Green’s theorem from circulation in a planar region to

circulation along a surface. The relationship between the macroscopic circulation of a vector field
F around a curve and the microscopic circulation of F along a surface in three dimensions must
hold for any surface whose boundary is the curve. The total microscopic circulation of F along the
surface must equal the circulation of F around the curve.

Example 62 Let C be the oriented triangle lying in the plane 2x + 2y + z = 6, as shown in the
$
following figure. Evaluate C F · dr where F (x, y, z) = −y 2 i + zj + xk.

66
 Solution: Using Stokes’ Theorem, begin by finding the ∇ × F.
 
 
 i j k 
 
 ∂ ∂ ∂ 
∇ × F =  ∂x ∂y ∂z  = −i − j + 2yk
 
 2 
 −y z x 

Considering z = 6 − 2x − 2y = g (x, y), by Eq. (4), for an upward normal vector to obtain

(gx (x, y) i + gy (x, y) j − k) −2i − 2j − k 2i + 2j + k

n=−
2 =− =
2
gx + gy + 1 3 3

and
  √
dS = gx2 + gy2 + 1dxdy = (−2)2 + (−2)2 + 1dxdy = 9dxdy = 3dxdy

 
F · dr = (∇ × F) · ndS
C S
  
2i + 2j + k
= (−i − j + 2yk) · dS
S 3
  
2i + 2j + k
= (−i − j + 2yk) · × 3dxdy
R 3

= (2y − 4) dxdy
R
 3  3−y
= (2y − 4) dxdy
0 0
 3
 
= 10y − 12 − 2y 2 dy
0

= −9
$
Example 63 Use Stokes’ Theorem to evaluate C F · dr where C is the intersection of the plane


z = 2 and the cone z = x2 + y 2 , with anticlockwise orientation as viewed from a point high above
the positive z-axis, and F = (x2 − y) i + 4zj + x2 k.

67


Solution: We take S to be the finite portion of the plane z = 2 as cut by the cone z = x2 + y 2 .
It is clear that the projection of S onto the xy-plane is just the disc a enclosed by the circle
x2 + y 2 = 4. Since n = k and
 
 
 i j k 
 
 ∂ ∂ ∂ 
∇ × F =  = −4i − 2xj + k
 ∂x ∂y ∂z 
 2 
 (x − y) 4z x2 

Stokes’ Theorem now gives

 
 2  2
x − y i + 4zj + x k · dr = (−4i − 2xj + k) · kdS
C S 
= 1dS = 1dxdy = 4π
S Rxy

Remark: Suppose F = P (x, y)i+Q(x, y)j is a vector field defined on a region R of the xy-plane.
Let C be the boundary curve of R, positively oriented with respective to the unit normal vector k.
Since  
 
 i j k   
  ∂Q ∂P
 ∂ ∂ 
∂  =
∇×F= − k.
 ∂x ∂y ∂z  ∂x ∂y
 
 P (x, y) Q (x, y) 0 
Stokes’ Theorem implies
 
F · dr = P (x, y) dx + Q (x, y) dy
C C

= (∇ × F) · kdS
S
  
∂Q ∂P
= − dS
R ∂x ∂y
which is nothing but the Green’s Theorem in the plane.

68