MULTIVARIABLE AND VECTOR ANALYSIS

W W L CHEN
c

W W L Chen, 1997, 2008.

This chapter is available free to all individuals, on the understanding that it is not to be used for financial gain,
and may be downloaded and/or photocopied, with or without permission from the author.
However, this document may not be kept on any information storage and retrieval system without permission
from the author, unless such system is not accessible to any individuals other than its owners.

Chapter 8
VECTOR FIELDS

8.1. Introduction
In this chapter, we consider functions of the form
F : A Rn : x 7 F (x),

(1)

where the domain A Rn is a set in the n-dimensional euclidean space, and where the codomain is also
the n-dimensional euclidean space Rn . For each x A, we can write
x = (x1 , . . . , xn ),
where x1 , . . . , xn R. We can also write
F (x) = (F1 (x), . . . , Fn (x)),

(2)

where F1 (x), . . . , Fn (x) R.

Definition. A function of the type (1), where A Rn , is called a vector field in Rn . The functions
Fi : A R, defined for i = 1, . . . , n by (2), are called the component scalar fields of F .
Remark. In the special cases n = 2 and n = 3, we usually write x = (x, y) and x = (x, y, z), so that
F (x, y) = (F1 (x, y), F2 (x, y))

and

F (x, y, z) = (F1 (x, y, z), F2 (x, y, z), F3 (x, y, z))

respectively.
Example 8.1.1. Suppose that the real valued function f : Rn R is continuously differentiable. Define
the function F : Rn Rn by writing


f
f
F (x) = (f )(x) =
(x), . . . ,
(x)
x1
xn
Chapter 8 : Vector Fields

page 1 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

for every x Rn . Recall that this is the gradient of f studied in Chapter 2. This vector field F is
sometimes called a gradient vector field.
Example 8.1.2. Consider the vector field F : R2 R2 : (x, y) 7 (y, x). There is no continuously
differentiable function f : R2 R such that F = f . To see this, note that if there were, then
F (x, y) = (f )(x, y) =

f f
,
x y

so that
f
= y
x

and

f
= x.
y

2f
= 1
yx

and

2f
=1
xy

It would then follow that

which is impossible. This vector field F is an example of a non-gradient vector field.

Example 8.1.3. Newtons law of gravitation states that the force acting on a point mass m at position
x R3 due to a point mass M at the origin 0 is given by
F (x) =

M m
x,
kxk3

where  > 0 is a proportionality constant. This is an attractive force field. Note that F = f , where
f : R3 R : x 7

M m
,
kxk

so F is a gradient vector field.

Example 8.1.4. Coulombs law in electrostatics states that the force acting on a point charge q at
position x R3 due to a point charge Q at the origin 0 is given by
F (x) =

Qq
x,
kxk3

where  > 0 is a proportionality constant. This is a repulsive force field. Note that F = f , where
f : R3 R : x 7

Qq
,
kxk

so F is a gradient vector field.

Remark. Suppose that F : Rn Rn is a vector field. By a flow line of F , we mean a path (t) in Rn
such that 0 (t) = F ((t)); in other words, F yields the velocity vector of the path (t). Flow lines are
useful in understanding some of the properties of vector fields, as we shall see in the following examples.
Example 8.1.5. For the vector field F : R2 R2 : (x, y) 7 (y, x), the path (t) = (cos t, sin t) is a
flow line, for clearly
0 (t) = ( sin t, cos t)
Chapter 8 : Vector Fields

and

F ((t)) = F (cos t, sin t) = ( sin t, cos t).

page 2 of 10

Chapter 8 : Vector
Fields
83
c W

W L Chen, 1997, 2008

Multivariable and Vector Analysis

Similarly, it can be shown that for any real number c R, the path (t) = (c cos t, c sin t) is a flow line
of F . Here the flow is circular, anticlockwise about the origin.

F (x,y)=(y,x)

Example 8.1.6. Let

Letususreturn
returnto
to Example
Example 8.1.4,
8.1.4, and
and consider
consider again
again Coulombs law of electrostatics,
where
F (x) =

Qq
"Qq
x.
kxk3
"x"

(3Qqt)1/3 u is a flow line of F , for clearly

Let u R3 be a fixed unit vector. The path (t) = (3"Qqt)
"0(t) =

Qq
"Qq
u
(3Qqt)2/3
(3"Qqt)

and

F ((t)) =

Qq
"Qq
(3Qqt)1/3 u.
(3"Qqt)
(3Qqt)
(3"Qqt)

This shows that the flow lines are radial and away from the origin.
(x, y).
y). For a path (t) =
Example 8.1.7. Consider
Consider the
the vector
vector field
field FF : : RR22
RR22 :: (x,
(x,y)
y) 7
$ (x,
"0
(1 (t), 2 (t)) to be a flow line of F , we must have (t) = F ((t)), so that
("10 (t), "20 (t)) = F ( 1(t), 2(t)) = ( 1(t), 2(t)),
whence "10 (t) = 1 (t) and "20 (t) = 2 (t); in other words, we need
d1
= 1
dt

d2
= 2 .
dt

and

These two differential equations have solutions 1 (t) = C1 et and 2 (t) = C2 et , where C1 , C2 R are
constants. It follows that the flow lines of F are of the form (t) = (C1 et , C2 et ), where C1 , C2 R are
constants. Note that the curve of the path (t) = (C1 et , C2 et ) is given by the hyperbola xy = C1 C2 .
The picture below shows F (x, y) at some points along the same flow line xy = 1.

Chapter 8 : Vector Fields

page 3 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

For the sake of convenience, we make the following definition.

Definition. The operator in Rn is given by
=

,...,
x1
xn

Remarks. (1) In the special cases n = 2 and n = 3, we have respectively

=

,
x y

and

,
,
x y z

(2) Note that for any real valued function f (x1 , . . . , xn ), the gradient of f is equal to
f =

,...,
x1
xn

f=

f
f
,...,
x1
xn

8.2. Divergence of a Vector Field

Suppose that F is the vector field of a gas or fluid. Then we may wish to discuss the rate of expansion
of volume under flow. This is a scalar valued function of a vector field.
Definition. Suppose that F = (F1 , . . . , Fn ) is a vector field in Rn . Then the divergence of F is the
scalar field



Fn
F1
div F = F =
,...,
+ ... +
.
(F1 , . . . , Fn ) =
x1
xn
x1
xn
Example 8.2.1. For the vector field F : R3 R3 : (x, y, z) 7 (yz, xz, xy), we have
div F =

(yz) +
(xz) +
(xy) = 0.
x
y
z

Example 8.2.2. For the vector field F : R2 R2 : (x, y) 7 (x, 0), we have div F = 1. Consider
next the flow lines of this vector field. Any flow line must be a path (t) = (1 (t), 2 (t)) satisfying
0 (t) = F ((t)), so that
(01 (t), 02 (t)) = F (1 (t), 2 (t)) = (1 (t), 0).
It follows that 1 (t) = C1 et and 2 (t) = C2 , where C1 , C2 R are constants. The flow is therefore in
the x-direction. If we think of F as a velocity field, then the speed is greater as we move further away
from the line x = 0. This corresponds to an expansion which is consistent with div F > 0.
Example 8.2.3. For the vector field F : R2 R2 : (x, y) 7 (x, y), we have div F = 2. Consider
next the flow lines of this vector field. Any flow line must be a path (t) = (1 (t), 2 (t)) satisfying
0 (t) = F ((t)), so that
(01 (t), 02 (t)) = F (1 (t), 2 (t)) = (1 (t), 2 (t)).
Chapter 8 : Vector Fields

page 4 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

It follows that 1 (t) = C1 et and 2 (t) = C2 et , where C1 , C2 R are constants. The flow is therefore
radial and towards the origin. This corresponds to a contraction which is consistent with div F < 0.
Example 8.2.4. For the vector field F : R2 R2 : (x, y) 7 (y, x), we have shown in Example 8.1.5
that the paths of the type (t) = (c cos t, c sin t), where c R, are flow lines of F . It can actually be
shown that these are all the flow lines of F . It follows that the flow is circular and anticlockwise around
the origin, with no expansion or contraction. Note now that div F = 0.

8.3. Curl of a Vector Field

While the divergence of a vector field is related to expansion or contraction, so the curl of a vector field
is related to rotation there are beaches in Sydney named after this operator! Indeed, a vector field
with zero curl will be called irrotational.
Definition. Suppose that F = (F1 , F2 , F3 ) is a vector field in R3 . Then the curl of F is the vector field
curl F = F =

,
,
x y z

(F1 , F2 , F3 ) =

F3
F2 F1
F3 F2
F1

y
z z
x x
y

Remarks. (1) We can write

curl F = F = det
x

F1

F2

F3








= F3 F2 i + F1 F3 j + F2 F1 k,

y
z
z
x
x
y

where i = (1, 0, 0), j = (0, 1, 0) and k = (0, 0, 1).

(2) Unlike gradient and divergence which are valid in euclidean space Rn for any natural number
n N, curl is only defined in R3 .
(3) Suppose that F is a vector field in R2 . While we cannot define curl F , we can nevertheless regard
F as a vector field in R3 for which the third component is zero and the two other components are
independent of z. Then


F2
F1
F = 0, 0,

.
x
y
The function
F2
F1

x
y
is sometimes called the scalar curl of F .
Example 8.3.1. For the vector field F : R3 R3 : (x, y, z) 7 (yz, xz, xy), we have
curl F =

(xy)
(xz),
(yz)
(xy),
(xz)
(yz)
y
z
z
x
x
y

= 0.

Here, note that if we consider the function f : R3 R : (x, y, z) 7 xyz, then F = f . We shall show
later in Theorem 8G that (f ) = 0 for any twice continuously differentiable function f : R3 R.
Chapter 8 : Vector Fields

page 5 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

Example 8.3.2. For the vector field F : R3 R3 : (x, y, z) 7 (x2 , (x + y)2 , (x + y + z)2 ), we have



2
2
2 2
2
2
curl F =
(x + y + z)
(x + y) ,
x
(x + y + z) ,
(x + y)
x
y
z
z
x
x
y
= (2(x + y + z), 2(x + y + z), 2(x + y)).
Hence
div(curl F ) =

(2(x + y + z)) +
(2(x + y + z)) +
(2(x + y)) = 0.
x
y
z

We shall show later in Theorem 8F that (F ) = 0 for any twice continuously differentiable function
F : R 3 R3 .
Example 8.3.3. Consider again the vector field F : R2 R2 : (x, y) 7 (y, x). We have shown in
Examples 8.1.5 and 8.2.4 that the flow is circular and anticlockwise around the origin. Note now that
the scalar curl of F is equal to
F1
F2

= 2.
x
y

8.4. Basic Identities of Vector Analysis

The first three theorems do not involve curl and are therefore valid in Rn for any natural number n N.
The first two of these theorems are easy to prove.
THEOREM 8A. For any continuously differentiable functions f : A R and g : A R, where
A Rn , and for any fixed real number c R, we have
(a) (f + g) = f + g;
(b) (cf ) = cf ;
(c) (f g) = f g + gf ; and
(d) (f /g) = (gf f g)/g 2 at any point x A for which g(x) 6= 0.
THEOREM 8B. For any continuously differentiable functions F : A Rn and G : A Rn , where
A Rn , and for any fixed real number c R, we have
(a) div(F + G) = div F + div G; and
(b) div(cF ) = c div F .
THEOREM 8C. For any continuously differentiable functions F : A Rn and f : A R, where
A Rn , we have div(f F ) = f div F + F f .
Proof. Let F = (F1 , . . . , Fn ). Then

div(f F ) = div(f F1 , . . . , f Fn ) =
(f F1 ) + . . . +
(f Fn )
x1
xn




F1
f
Fn
f
= f
+ F1
+ ... + f
+ Fn
x1
x1
xn
xn

 

F1
Fn
f
f
= f
+ ... + f
+ F1
+ . . . + Fn
x1
xn
x1
xn




F1
Fn
f
f
=f
+ ... +
+ (F1 , . . . , Fn )
,...,
x1
xn
x1
xn
= f div F + F f.
This completes the proof.
Chapter 8 : Vector Fields

page 6 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

We also have the following four theorems which involve curl and are therefore restricted to R3 .
THEOREM 8D. For any continuously differentiable functions F : A R3 and G : A R3 , where
A R3 , and for any fixed real number c R, we have
(a) curl(F + G) = curl F + curl G;
(b) curl(cF ) = c curl F ; and
(c) div(F G) = G curl F F curl G.
Proof. Parts (a) and (b) are easy to check. To prove (c), let F = (F1 , F2 , F3 ) and G = (G1 , G2 , G3 ).
Then
div(F G) = div(F2 G3 F3 G2 , F3 G1 F1 G3 , F1 G2 F2 G1 )

=
(F2 G3 F3 G2 ) +
(F3 G1 F1 G3 ) +
(F1 G2 F2 G1 ).
x
y
z
Using the sum and product rules to the terms on the right and rearranging, we obtain





F2
F3
F1
F1
F2
F3

+ G2

+ G3

div(F G) = G1
y
z
z
x
x
y






G3
G2
G1
G3
G2
G1
F1

F2

F3

y
z
z
x
x
y


F2 F1
F3 F2
F1
F3

= (G1 , G2 , G3 )
y
z z
x x
y


G3
G2 G1
G3 G2
G1
(F1 , F2 , F3 )

y
z z
x x
y
= G curl F F curl G.


This completes the proof.

THEOREM 8E. For any continuously differentiable functions F : A R3 and f : A R, where
A R3 , we have curl(f F ) = f curl F F f .
Proof. Let F = (F1 , F2 , F3 ). Then
curl(f F ) = curl(f F1 , f F2 , f F3 ) =

(f F3 )
(f F2 ),
(f F1 )
(f F3 ),
(f F2 )
(f F1 ) .
y
z
z
x
x
y

Using the sum and product rules to the terms on the right and rearranging, we obtain

F3
F2 F1
F3 F2
F1
curl(f F ) = f

y
z z
x x
y


f
f
f
f
f
f
+ F3
F2 , F1
F3 , F2
F1
y
z
z
x
x
y

 

F3
F2 F1
F3 F2
F1
f f f
=f

+
,
,
(F1 , F2 , F3 )
y
z z
x x
y
x y z
= f curl F F f.


This completes the proof.

Chapter 8 : Vector Fields

page 7 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

THEOREM 8F. For any twice continuously differentiable function F : A R3 , where A R3 , we

have div(curl F ) = 0.
Proof. Let F = (F1 , F2 , F3 ). Then


F3
F2 F1
F3 F2
F1
div(curl F ) = div

y
z z
x x
y






F3
F2
F1
F3
F2
F1
=

x y
z
y z
x
z x
y
2 F2
2 F1
2 F3
2 F2
2 F1
2 F3

=
xy xz
yz
yx zx zy
 2



 2

2 F1
2 F2
2 F2
F3
2 F3
F1
=

= 0,
yz
zy
zx xz
xy yx
in view of Theorem 4A.
THEOREM 8G. For any twice continuously differentiable function f : A R, where A R3 , we
have curl(f ) = 0.
Proof. We have


f f f
,
,
curl(f ) = curl
x y z

 
 
 
 
 
 
f
f
f
f
f
f
=

y z
z y
z x
x z
x y
y x
 2

2
2
2
2
2
f
f f
f f
f
=

= 0,
yz
zy zx xz xy yx
in view of Theorem 4A.
Example 8.4.1. Consider the vector field F : R3 R3 : (x, y, z) 7 (x, y, z). It is easily checked that
div F = 3. It follows that there is no function G : R3 R3 such that F = curl G, for otherwise div F = 0
by Theorem 8F.
Example 8.4.2. Consider the vector field F : R3 R3 : (x, y, z) 7 (y, x, 0). It is easily checked that
curl F = (0, 0, 2). It follows that there is no function f : R3 R such that F = f , for otherwise
curl F = 0 by Theorem 8G.
For any twice continuously differentiable function f : A R, where A Rn , we have

 


f
f
2f
2f
div(f ) = (f ) =
,...,

,...,
=
+ ... +
.
2
x1
xn
x1
xn
x1
x2n
Definition. The Laplace operator 2 in Rn is defined as the divergence of the gradient, so that
2 =

2
2
+ ... +
.
2
x1
x2n

Example 8.4.3. A function f : A R, where A Rn , is said to satisfy Laplaces equation if 2 f = 0.

An example of such a function is given in the case n = 3 by
f (x) = f (x, y, z) =
Chapter 8 : Vector Fields

1
1
=
.
kxk
(x2 + y 2 + z 2 )1/2
page 8 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

THEOREM 8H. For any twice continuously differentiable functions f : A R and g : A R, where
A Rn , we have
(a) 2 (f g) = f 2 g + g2 f + 2(f g); and
(b) div(f g gf ) = f 2 g g2 f .
Proof. Note that
2 (f g)
2 (f g)
+
.
.
.
+
x2
x2n

 2

 12
f g
g
f g
g
2f
2f
= f 2 +2
+ g 2 + ... + f 2 + 2
+g 2
x1
x1 x1
x1
xn
xn xn
xn
 2

 2



g
2g
f
2f
f g
f g
=f
+
.
.
.
+
+
g
+
.
.
.
+
+
2
+
.
.
.
+
x2
x2n
x21
x2n
x1 x1
xn xn




 21
 

2g
2f
2f
f
g
f
g
g
+
.
.
.
+
+
g
+
.
.
.
+
+
2
=f
,
.
.
.
,

,
.
.
.
,
x21
x2n
x21
x2n
x1
xn
x1
xn

2 (f g) =

= f 2 g + g2 f + 2(f g).

This gives (a). On the other hand,




 
g
f
f
g
,...,
g
,...,
div(f g gf ) = div f
x1
xn
x1
xn


g
f
g
f
= div f
g
,...,f
g
x1
x1
xn
xn




f
f

g
g
g
=
f
+ ... +
f
x1
x1
x1
xn
xn
xn
 2

 2

g
f g
g f
2f
g
f g
g f
2f
= f 2+

g 2 + ... + f 2 +

g 2
x
x1 x1
x1 x1
x1
xn
xn xn
xn xn
xn
 21

 2

2
2
g
g
f
f
=f
g
+ ... +
+ ... +
x21
x2n
x21
x2n
= f 2 g g2 f.

This gives (b).

Finally, we have the following result. The proof is left as an exercise to test the readers grasp of the
techniques.
THEOREM 8J. For any twice continuously differentiable functions f : A R and g : A R, where
A R3 , we have div(f g) = 0.

Chapter 8 : Vector Fields

page 9 of 10

Multivariable and Vector Analysis

W W L Chen, 1997, 2008

Problems for Chapter 8

1. Consider the vector field F : R2 R2 : (x, y) 7 (4x, 2y).
a) Is this a gradient vector field? Justify your assertion.
b) Suppose that (t) = (1 (t), 2 (t)) is a flow line of the vector field F . Show that satisfies the
two differential equations
d1
d2
= 41
and
= 22 .
dt
dt
c) By solving the differential equations in part (b), show that the flow lines are parabolas.
d) Draw a picture to support your observation in part (c).
2. Consider the functions F : R3 R3 : (x, y, z) 7 (2xz, xyz, y 3 xz) and f : R3 R : (x, y, z) 7 x2 yz.
Compute each of the following:
a) f
b) div F
c) curl F
d) F f
e) F f
f) (F f )
g) div(curl F )
3. Discuss the question of whether curl F has to be perpendicular to F .
4. Consider the vector field F (x, y) =
a)
b)
c)
d)

x
p

x2 + y 2

,p

y
x2 + y 2

!
.

Let u R2 be a unit vector. Show that (t) = tu is a flow line of F .

Show that div F > 0 whenever (x, y) 6= (0, 0).
Show that the scalar curl of F is zero.
Draw a picture of the vector field F and briefly explain the observations in parts (b) and (c).

5. Consider the vector field F (x, y, z) = (2x, z, z 2 ).

a) Let (t) = (1 (t), 2 (t), 3 (t)) be a flow line of F . Show that 1 (t), 2 (t), 3 (t) satisfy three
ordinary differential equations.
b) Solve two of the differential equations in part (a) for 1 (t) and 3 (t). Then solve the remaining
differential equation for 2 (t).
c) Check that (t) = (e2t , log t, 1/t), where t > 0, is one of the solutions.
d) Find div F and curl F .
6. Consider the vector field F : R3 R3 : (x, y, z) 7 (3x2 y, x3 + y 3 , 0).
a) Verify that curl F = 0.
b) Find a function f : R3 R such that f = F .

Chapter 8 : Vector Fields

page 10 of 10