1 Vector Calculus

1.1 Derivative of a vector

~ A
If a vector A ~ psq such that A
~  Ax psq î Ay psq ĵ Az psq k̂, where Ax , Ay and Az are scalar
functions of s and are the components of A ~ in the x, y and z directions, respectively, then the
~
derivative of A with respect to s is given by
# +
~ ~ ps δsq
A
~ ps q
dA
ds
 δslim
Ñ0
A
δs
 dA
ds
x
î
dAy
ds
ĵ
dAz
ds
k̂ (1)

Example: Let px, y, z q be the coordinates of a moving particle at time t. Then x, y and z are
functions of time. The vector displacement of the particle from the origin (called the position
vector) at time t is ~r  x î y ĵ z k̂, and therefore, the velocity of the particle is given by

~v  d~
r
dt
 dxdtptq î
dy ptq
dt
ĵ
dz ptq
dt
k̂  vx î vy ĵ vz k̂. (2)
The direction of this velocity vector is along the tangent to the path ~r at any instant. Similarly,
the acceleration of the particle is given by
2 d2 y d2 z
~a   ddtx2 î  ax î
d~v
ĵ k̂ ay ĵ az k̂. (3)
dt dt2 dt2

Remarks:
~ is a vector function, i.e., φ  φpsq and A
1. If φ is a scalar function and A ~A
~ psq, then

d   ~
φ ps qA
~ psq  dφ A
~ φ dA . (4)
ds ds ds
In the case where φ is a constant (with respect to s), i.e., independent of s, then dφ{ds  0
and the above equation reduces to
~
d
ds
p ~ q  φ dA .
φA
ds
(5)

~A
2. For two vectors A ~ psq and B
~  B~ psq,
# +
~ ps δsq  B
~ ps δsq
A ~ ps q  B
~ psq
ds
pA  B q  δslim
d ~ ~
Ñ0
A
δs
# +
~ ps δsq  B
~ ps δsq
A
~ ps q  B
~ ps δsq ~ ps q  B
~ ps δsq
A
~ ps q  B
~ ps q
 δslim
Ñ0
A
δs
A
δs
# +
 δslim rA~ ps δsq
A
~ psqs  B
~ ps δsq ~ ps q  r B
A ~ ps δsq
B
~ psqs
Ñ0 δs δs
#  + #  +
~ ps δsq
A
~ ps q ~ ps δsq
B
~ psq
 lim
δs Ñ0
A
δs
 B~ ~
A lim
δs Ñ0
B
δs
~ ~ ~ ~
 ddsA  B~ ~  dB
A
ds
 B~  ddsA ~  dB
A
ds

1
 Therefore
~ ~
ds
p
d ~ ~
A  Bq  A
~  dB
ds
~  dA
B
ds
(6)
~ ps q  B
If A ~ psq, then
~ ~ ~
p A  Bq  A
d ~ ~
ds
~  dA A
ds
~  dA  2A
ds
~  dA
ds
 dsd pA~  A~ q.
For a constant vector, A  constant, then dA{ds  0 and
~ ~
~
p
d ~ ~
ds
A  Aq  2A
~  dA
ds
0
This means that dA ~ {ds is a vector perpendicular to vector A. ~ An example, is uniform (cir-
cular) motion, where the acceleration is directed towards the center of the circular path, and
the velocity of the particle is tangential to the circular path. In that case, the kinetic energy,
K  p1{2qmv 2  constant, and dK {dt  m~v  d~v {dt  0.
~A
3. Similarly, if A ~ psq and B
~  B~ psq, then
# +
~ ps δsq  B ~ ps δsq
A
~ ps q  B
~ ps q
ds
pA  B q  lim
d ~ ~
δs
A
Ñ0 δs

 δslim
!A
~ sp δsq  B
~ ps δsq
A~ psq  B ~ ps δsq
Ñ0 δs
Apsq  B ps δsq
A
~ ~ ~ psq  B ~ ps q )

# δs +
 δslim rA~ ps δsq
Apsqs  B ps δsq A
~ ~ ~ ps q  r B~ ps δsq
B
~ psqs
Ñ0 δs δs
#  + #  +
~ ps δsq
A
~ ps q ~ ps δsq
B
~ psq
 lim
δs Ñ0
A
δs
 B~ ~
A lim
Ñ0
δs
B
δs
~ ~ ~ ~
 ddsA  B~ ~  dB
A
ds
 A~  ddsB
B~  ddsA
Therefore
~ ~
ds
p
d ~ ~
A  Bq  A~  dB dA  B
ds ds
~ (7)

As an example, consider angular momentum, L ~  ~r  p~, where p~ is the linear momentum,

given by p~  m~v . The torque, ~τ is given by
~
~τ  ddtL  dtd p~r  p~q  ~r  d~
p
dt
d~r
dt
 p~  ~r  d~
p
dt
~v  p~  ~r 
d~
p
dt
mp~v  ~v q  ~r 
d~
p
dt
,

since ~v ~v  0 for parallel vectors. Also, Newton’s second law of motion is given by F~  d~p{dt,
where F~ is the net force acting on the particle. Therefore,

~τ  ~r  d~
p
dt
 ~r  F~ (8)

Hence the rate of change of angular momentum is equal to the applied torque.

2
1.2 Space Curve and Curvature
Consider a particle moving on a space curve as shown in the figure[To be inserted]. At certain time,
the particle is at some point P . In a time interval ∆t the particle moves to another point P 1 along
the path. The arc length between P and P 1 is ∆s. Let t̂ be the unit vector in the direction of the
tangent of the curve at P . The velocity of the particle is of course in the direction of t̂,

~v  d~
r
dt
 v t̂, (9)

where v is the magnitude of the velocity, which is given by

v  ∆tlim ∆s

Ñ0 ∆t dt
ds
. (10)

The point P can also be specified by s, the distance along the curve measured from a fixed point
to P . Then by chain rule,
~v   
d~r ds d~r d~r
v . (11)
dt dt ds ds
Comparing equations (9) and (11), we have

t̂ 
d~r
. (12)
ds
Clearly, as ∆s Ñ 0, we get |∆~r{∆s|  1.
Since, t̂ is a unit vector, it means its magnitude is not changing, and therefore its derivative must
be perpendicular to itself. Suppose n̂ is a unit vector perpendicular to t̂, then we can write

dt̂
ds
 κ n̂ (13)

where κ is the magnitude of dt̂{ds and is called the curvature. The vector n̂ is known as the normal
vector and is perpendicular to t̂. The reciprocal of the curvature ρ  1{κ is known as the radius of
the curvature. Equation (13) defines both κ and n̂, and tells us how fast the unit tangent vector
changes direction as we move along the curve.
As a result, the acceleration of the particle is

~a   dtd pv t̂q  dv  v t̂
d~v dt̂
t̂ v 9 v t9 (14)
dt dt dt
where
t9   ds   vκ n̂
dt̂ dt̂ dt̂
v (15)
dt dt ds ds
and Eq. (14) becomes
v2
~a  v9 t̂ v 2 κ n̂  v9 t̂ n̂ (16)
ρ
In Eq. (16), the tangential component of ~a (in this case, v) corresponds to the change in the
magnitude of ~v , and the normal component of ~a (in this case, v 2 {ρ) corresponds to the change in
direction of ~v . Note that the normal component is the familiar centripetal acceleration.

3
 In the case of circular motion, where a particle is rotating around the z-axis with a constant
angular velocity ω, then ~r  r cos ωt î r sin ωt ĵ and v  ωr. We then get the tangential unit
vector, t̂ as
t̂   p
rω sin ωt î rω cos ωt ĵ q 
sin ωt î cos ωt ĵ
~v 1
v ωr
Since t9  dt̂{dt  v dt̂{ds, it implies that
dt̂
ds
 v1 dt
dt̂
 ωr
1
p
ω cos ωt î
ω sin ωt ĵ q 
r12 pr cos ωt î r sin ωt ĵ q 
r12 ~r 
1r ~rr
Also note that dt̂{ds  κn̂. Then in this case, we get κ  1{r, ρ  r, and n̂ 
~r{r  r̂. In other
words, the radius of curvature in a circular motion is equal to the radius of the circle.

1.2.1 Partial derivatives

Take a function f  f px, y, zq. Then

Bf  lim " f px δx, y, zq
f px, y, zq *
Bx δxÑ0 " δx
(17)

Bf  lim f px, y δy, zq
f px, y, zq *

By δyÑ0 " δy
(18)

Bf  lim f px, y, z δzq
f px, y, zq *

Bz δzÑ0 δz
(19)

If the function is continuous,

B2f  B2f
B xB y B y B x
For example, given f px, y, z q  x7 y 5 z
2 , then
Bf  7x6y5z
2; Bf  5x7y4z
2 Bf 
2x7y5z
3
Bx By and
Bz
giving
B2f  B  Bf  B p5x7y4z
2q  35x6y4z
2.
B xB y B x B y Bx
Also,
B2f  B  Bf  B p7x6y5z
2q  35x6y4z
2.
ByBx By Bx By
Therefore, f px, y, z q is a continuous function.
In the case of vector functions, say, A ~A ~ px, y, z q, we have
# +
BA~  lim A~ px δx, y, zq
A~ px, y, zq
Bx δxÑ0 δx
(20)
# +
BA~  lim A~ px, y δy, zq
A~ px, y, zq
By δyÑ0 δy
(21)
# +
BA~  lim A~ px, y, z δzq
A~ px, y, zq
Bz δzÑ0 δz
(22)

4
 For example, if  
~
E  k px 2 x
y 2 q3{2
î
px 2
y
y 2 q3{2
ĵ

where k is a constant, then

BE~   B  x B y

 pxy2
y2x
2 2
î
2
3xy
B x B x p x2 y 2 q3{2
î
B x p x2 y 2 q3{2
ĵ 2 q5{2 px y2q5{2 ĵ
and

BE~   B  x B y


px2 3xyy2q5{2 î pxx2
y2y
2 2

B y B y px 2 y 2 q3{2
î
B y px 2 y 2 q3{2
ĵ 2 q5{2
ĵ

1.3 ~
The Del operator, ∇
This is a differential vector operator called “del” or “nabla”, defined by

~  î B B B.
∇
Bx ĵ
By k̂
Bz (23)

1.3.1 Gradient of a scalar function

Let φpx, y, z q be a single-valued scalar function that is defined at every point of x, y, z of region
Q (i.e., its first derivative is continuous in a certain region of space). If ~r is the position vector
of an object located at P px, y, z q, then ~r  x î y ĵ z k̂ and d~r  dx î dy ĵ dz k̂. The total
differential of φpx, y, z q is
  
dφ 
B φ B φ B φ
dz 
B B B 




~  d~r
Bx dx
By dy
Bz î
Bx ĵ
By k̂
Bz φ dx î dy ĵ dz k̂ ∇φ (24)

~ is called the gradient of φ, denoted grad φ. Therefore,

where ∇φ

grad φ  î
Bφ Bφ Bφ  ∇φ,
~
Bx ĵ
By k̂
Bz (25)

which is a vector function whose px, y, z q components are the first partial derivatives of φpx, y, z q
with respect to x, y and z, respectively.
Example 1: (a) Given φpx, y, z q  xyz x2 y, find grad φ.
Solution:

grad φ  î
Bφ Bφ Bφ  pyz 2xy q î pxz x2 q ĵ pxyq k̂  p2xy yz q î pxz x2 q ĵ
Bx ĵ
By k̂
Bz xy k̂

(b) Find a unit vector normal to the surface φ  x2 y 2
4z at P p1, 2,
3q

Solution: If φ  x2 y 2
4z, then ∇φ
~  2x î 2y ĵ
4 k̂.

5
 At the point P p1, 2,
3q, x  1, y  2, z 
3, and


~
∇φ
atP
 2p1q î 2p2q ĵ
4 k̂  2 î 4 ĵ
4 k̂

The unit vector, n, normal to φ at P p1, 2,
3q is given by

 a 22î 4 ĵ 2
4 k̂ 2  ?2 pî
~
n̂  2 ĵ
2 k̂ q  pî 2 ĵ
2 k̂ q.
∇φ 1
|∇φ
~ | p2q p4q p
4q 36 3

Note that if φ1 and φ2 are two arbitrary scalar functions, then

~ pφ 1
(i) ∇ φ2 q  ∇φ
~ 1 ~ 2
∇φ
~ pφ1 φ2 q  φ1 ∇φ
(ii) ∇ ~ 2 ~ 1
φ2 ∇φ
~ pcφ1 q  c∇φ
(iii) ∇ ~ 1 , where c is a constant.

1.3.2 Divergence of a vector function

~ px, y, z q is a vector function. Then the divergence of A,
Suppose A ~ denoted div A
~ is defined by

~∇
div A ~ A
~ î
B
Bx ĵ
B
By k̂
B 
A î
Bz x Ay ĵ Az k̂  BBAxx BBAyy BBAzz (26)

~  x3 y î
Example: Take A y 2 z 2 ĵ z 4 x k̂. Then

~ A
~ Bpx3yq Bpy2z2q Bpz4xq  3x2y 2yz 2 4xz 3
∇
Bx By Bz
~  ∇φ.
Note that the divergence of a vector field is a scalar field. Suppose A ~ Then
~ A
∇ ~∇
~  ∇φ
~  div grad φ. (27)

This equation is called the Laplacian of φ, ∇2 φ. That is,


∇2 φ  ∇
~  ∇φ
~  B B B  î Bφ Bφ Bφ  B2φ B2φ B2φ
î
Bx ĵ
By k̂
Bz Bx ĵ
By k̂
Bz B x2 B y 2 B z 2 (28)

or simply 
∇ φ2 B2 B2 B2
B x2 B y 2 B z 2 φ (29)

from which we obtain

 BBx2 BBy2 BBz2 ,
2 2 2
∇2 (30)

6
 a scalar operator, called the Laplacian operator. It is one of several mathematical physics expres-
sions (equations) that are of importance. That is,
#
∇2 φ 

ερ
Poisson’s equation


0
0 Laplace’s equation
1 B2φ
∇2 φ 

a2 B t2
wave equation
1 Bφ
∇2 φ  2

a Bt
diffusion equation (or equation of heat conductivity)
∇2 φ 
k 2 φ
Helmholtz equation

1.3.3 Curl of a vector function

~ px, y, z q is a vector function, then the curl of A
If A ~ is a vector function defined by

î ĵ k̂
~ B B B
curl A Bx By Bz (31)
Ax Ay Az
  
 î BBAyz
BBAzy ĵ
BAx
BAz
Bz Bx k̂
B Ay
B Ax
Bx By (32)

1.3.4 Vector identities involving div and curl

~ and B
If A ~ are two vector functions, then

~  pA
1. ∇ ~ ~q  ∇
B ~ A
~ ~ B
∇ ~

~  pA
2. ∇ ~ ~q  ∇
B ~ A
~ ~ B
∇ ~

~  p∇
3. ∇ ~ A ~ q  div curl A
~  0.
~ ~  0, then B
If B is a vector function such that div B ~ is the curl of some vector function A. ~
Such vector function (B)~ whose divergence is identically zero is said to be solenoidal, and the
~ ~
vector function A is known as the vector potential of the field B.
~  p∇
4. ∇ ~ φq  curl grad φ  0.
~ ~ must be the gradient of some
If B is a vector function whose curl is identically zero, then B
scalar function, say φ. A vector function (B) ~ whose curl is identically zero is said to be
irrotational. An irrotational field is also known as a conservative vector field.
~  pφ A
5. ∇ ~ q  φ p∇
~ A
~q ~ φA
∇ ~

~  pφ A
6. ∇ ~ q  φ p∇
~ A
~q ~ φA
∇ ~

~  pA
7. ∇ ~B
~q  B
~  p∇
~ A
~q
A
~  p∇
~ B
~ q.

7
 ~  pA
8. ∇ ~B
~q  A
~ p∇
~ B
~q
B
~ p∇
~ A
~q pB~  ∇~ qA~
pA~  ∇~ qB~
~ pA
9. ∇ ~B
~ q  pA
~∇
~ qB
~ pB~  ∇~ qA~ ~  p∇
A ~ B ~q B ~  p∇ ~ A ~q

10. ~  p∇
∇ ~ A
~q  ∇
~ p∇
~ A
~ q
∇2 A.~

2 Vector integration
Suppose F~ puq  d
du pG~ puqq. Then the indefinite integral of F~ puq is given by
» »
d
~
F~ puqdu  Gpuq du  G
~ puq ~
C, (33)
du
~ is a constant vector.
where C
Similarly, the definite integral of F~ puq from u  u1 to u  u2 is given by
» u2 » u2
d
~
F~ puqdu  Gpuq du  G
~ pu2 q
G
~ pu1 q. (34)
u1 u1 du

Note that if F~ puq  F1 puq î F2 puq ĵ F3 puq k̂, then

» » » »
F~ puqdu  F1 puqdu î F2 puqdu ĵ F3 puqdu k̂. (35)

There are three types of vector integration: line integrals, surface integrals, and volume integrals.

2.1 Line integrals

Consider a vector F~  F~ px, y, z q along a curve C from point A to point B. The line integral
(or tangential line integral), Iλ corresponding to this vector in the definite integral of the scalar
component of F~ in the direction of the curve at P px, y, z q. Thus,
»B
Iλ  F~  d~r (36)
A

where d~r is an element of the displacement vector at P px, y, z q, given by d~r  dx î dy ĵ dz k̂.
If F~ is a variable force vector, then the elemental work done in moving an object through an
infinitesimal displacement, d~r is given by dW  F~  d~r. So, the total work done in moving the
object from A to B is evaluated by the integral
»B
W  F~  d~r. (37)
A

Generally, a line integral is an integral evaluated along a curve. Practical applications of this
integral in aerodynamics and fluid mechanics represent the suppression of F~ about a curve C, and

8
therefore F~ represents the fluid velocity. (Note that F~ was chosen as an arbitrary vector, not
necessarily the force.)
³
Example 1: Given that F~  p3x2
6yz qî p2y 3xz qĵ p1
4xyz 2 qk̂. Find C F~  d~r, where C
is the path joining the points (0, 0, 0) and (1, 1, 1) defined by x  t, y  t2 and z  t3 .
Solution: Take F~  F1 î F2 ĵ F3 k̂; ~r  x î y ĵ z k̂, implying that d~r  dx î dy ĵ dz k̂.
Also define ~r1  p0, 0, 0q and ~r2  p1, 1, 1q. Then
» » » »
F~  d~r  F1 dx F2 dy F3 dz
C

or » ~r2 » ~r2 » ~r2 » ~r2

F~  d~r  p3x
6yzqdx
2
p2y 3xz qdy p1
4xyz2qdz
~
r1 ~
r1 ~
r1 ~
r1

For x  t, y  t2 , z  t3 we have dx  dt; dy  2tdt and dz  3t2dt. In terms of time, t, the limits
are from t  0 to t  1. Therefore,
» ~r2 » t1 » t1 » t1
F~  d~r  p3t
6t qdt
2 5
p2t 2
3t 4
qp2tdtq p1
4t9qp3t2dtq
~
r1 
t 0 
t 0 
t 0
»t 1
 »t 1

 p3t2
6t5 4t3 6t5 3t2
12t11 qdt  p6t2 4t3
12t11 qdt  2

t 0 t 0

Example 2: Show that the work done on an object by a resultant force F~R during a displacement
from A to B is equal to the change in the kinetic energy (mv 2 {2) of the object.
Solution: The work done is given by
»B
W  F~R  d~r, (38)
A

where

F~R  m~r:  m
d~v
(Newton’s second law), and
dt
d~r  dt  ~v dt
d~r
dt
Then
F~R  d~r  m  ~v dt  m ~v  dt  m ~v  d~v .
d~v d~v
dt dt
Since dt pA  Aq  ddtA~  A~ A~  ddtA~  2A~  ddtA~ , it implies that dpA~  A~ q  2A~  dA~ or
d ~ ~

~v  d~v  12 dp~v  ~v q  21 dpv 2 q, thus F~R  d~r  m ~v  d~v  m

2 dpv q, and the work done becomes
2

»B »B
m 2 vB
W  F~R  d~r 
m
2
dpv 2 q 
2
v 
v vA
 12 mvB2
12 mvA2  change in kinetic energy. (39)
A A

This is the work energy theorem of mechanics.

9
Remark on conservative forces

If the force F~ is the gradient of a single-valued scalar potential φpx, y, z q, i.e., F~  ∇φ,
~ and the
path of integration is closed, we obtain
¾ ¾ ¾
Work  F~  d~r  ~  d~r 
∇φ dφ  0. (40)
¶
(The symbol means integration around a closed path.) In this case, the work done is zero. For
the work-energy theorem that would imply that the final kinetic energy of the object is equal to the
initial kinetic energy. Hence energy is conserved. Such a force, F~  ∇φ
~ is said to br a conservative
force.
To see why an irrotational field is also called a conservative vector field, consider F~ px, y, z q as
the force in Newton’s equation of motion

F~ px, y, z q  m ~a  m .
d~v
(41)
dt
Since F~ is irrotational, we have

F~ px, y, z qq 
∇φ
~ px, y, z q. (42)

Therefore,
d~v
dt

∇φ
~ px, y, z q.
m (43)

Now, take the dot product of both sides of (43) with d~r and integrate. The left-hand side becomes
» » »
m
d~v
dt
 d~r  
d~v d~r
m
dt dt
dt  m
d~v
dt
 ~v dt
» »  » 
 m ~v  dt dt  dt 2 m ~v  ~v dt  d 12 mv2
d~v d 1

 12 mv2 constant (44)

Also, the right-hand side becomes

» »
p
∇φ
~ q  d~r 
dφ 
φ constant. (45)

Equating the right hand sides of (44) and (45), we get

φ  constant.
1
mv 2 (46)
2
Note that in classical mechanics, the expression mv 2 {2 is defined as the kinetic energy, φpx, y, z q
as the potential energy, and the sum of the two is the total energy. Thus, equation (46) says that
no matter where and when the total energy is evaluated, it must be equal to the same constant.
This is the principle of conservation of energy.
Although we have used classical mechanics to introduce the idea of conservative field, the idea
can be generalized. Any vector field V~ px, y, z q which can be expressed as the gradient of a scalar

10
field φpx, y, z q is called a conservative field and the scalar function φ is called the scalar potential.
~ ∇
Since ∇φ ~ pφ constantq, the scalar potential is defined up to an additive constant.
We have seen that if ∇ ~  F~  0 in a simply connected region, we say F~ is an irrotational vector
³B
field. Also, if ∇  F  0, the line integral A F~  d~r is independent of path. This means that F~
~ ~
can be expressed as the gradient of a scalar function φ, known as the scalar potential. Because of
Stokes’s theorem (as we will see later),
¾ ¼
F~  d~r  p∇~  F~ q  n̂da,
C S

an irrotational field F~ is characterized by any of the following equivalent conditions:

~  F~
(i) ∇  0,
¶
(ii) F~  d~r  0 for any closed loop,
³B
A F  d~
(iii) ~ r is independent of path,

(iv) F~ 
∇φ. ~

Additional Examples (From Chapter 6 of Mary Boas 3rd Ed.):

8.15 Verify that the given force field, F~  2x cos2 y î
px2 1q sin 2y ĵ, is conservative. Then find
a scalar potential φ such that
∇φ ~  F~ .

Solution:
Given that F~  2x cos2 y, î
px2 1q sin 2y ĵ, then the curl of F~ is given by

î ĵ k̂ î ĵ k̂
~  F~  B B B  B B B
∇ Bx By Bx Bx By Bx
Fx Fy
"
Fz 2x cos2 y
px2 1q sin*2y 0

 0 î 0 ĵ
BBx rpx2 1q sin 2ys
BBy r2x cos2 ys k̂

 p
2x sin 2y 4x cos y sin y q k̂
 p
2x sin 2y 2x sin 2y q k̂
0
Therefore, F~ is conservative, and we can find a scalar potential, φpx, y q such that
∇φ
~  F~ .
That is,
Bφ 
F 
2x cos2 y;
Bx x (47)
Bφ 
F  px2 1q sin 2y.
By y (48)

Integrate Eq. (47) with respect to x to get

φpx, y q 
x2 cos2 y cpy q, (49)

11
 where cpy q is a constant of integration (independent of x) to be determined. To get the
constant, cpy q, differentiate Eq. (49) with respect to y, and compare the result to Eq. (48):
Bφ  x2 sin 2y d
cpy q 
Fy  px2 1q sin 2y  x2 sin 2y
By dy
sin 2y.

That is,
cpy q  x2 sin 2y cpy q  sin 2y,
d d
x2 sin 2y sin 2y, or (50)
dy dy
giving cpy q 
p1{2q cos 2y constant, and therefore

φpx, y q 
x2 cos2 y
constant 
px2 1q cos 2y

1
cos 2y constant
2

8.20 For motion near the surface of the earth, we usually assume that the gravitational force on a
mass m is F~ 
mg k̂, but for motion involving an appreciable variation in distance r from
the center of the earth, we must use F~ 
pC {r2 q êr 
pC {r2 qp~r{|~r|q 
pC {r3 q ~r, where
C is a constant. Show that both these F~ ’s are conservative, and find the potential for each
case.

Solution:
Let F~1 
mg k̂ and F~2 
rC êr 
rC
2 2
~
r
|~r| 
rC ~r.
Note that F~2 is expressed in spherical
3

coordinates, (r, θ, φ), although it is a function of r only. To show that F~1 or F~2 is a conser-
vative vector field, we obtain the curl of F~1 and of F~2 . If the curl of the vector is zero, then
the vector field is a conservative vector field, otherwise it is a nonconservative vector field.
For F~1 
mg k̂, use rectangular (or Cartesian) coordinates, and get the curl of F~1 :
 
~  F~1
∇


 ∇~  p
mg k̂q 
mg ∇~  k̂ B B

mg Bx î By ĵ Bz k̂  k̂ B
 
Bp 1q Bp 1q Bp

mg Bx pî  k̂q By pĵ  k̂q Bz pk̂  k̂q1q

0
since k̂  k̂  0, ĵ  k̂  î, î  k̂ 
ĵ, and BpBx1q  BpBy1q  BpBz1q  0. Hence F~1 
mg k̂
is a conservative vector field. It can therefore be expressed in the form: F~1 
∇Φ, ~ where
Φ  Φpx, y, z q is a scalar function. That is,
B Φpx, y, zq î B Φpx, y, zq ĵ B Φpx, y, zq k̂  mg k̂
Bx By Bz
or
BΦ  0
Bx (51)
BΦ  0
By (52)
BΦ  mg
Bz (53)

12
Integrate equation (3) with respect to get

Φpx, y, z q  mgz C px, y q (54)

where C px, y q is a constant of integration that does not depend on z. At this point, if we
know C px, y q, we shall have found our solution. So, to proceed, we differentiate Eq. (4) with
respect to y and compare the result with Eq. (2). That is,
BΦ  B C px, yq  0
By By
which implies that C px, y q is independent of y, i.e., C px, y q  C pxq, a constant that may be
depending only on x. Substitute this into Eq. (4) to get

Φpx, y, z q  mgz C px q (55)

Again, differentiate Eq. (5) with respect to x, and use Eq. (1) to get
B Φ  B C px q  0
Bx By
The last equality implies that C pxq i a constant with respect to x, call it C. Our Eq. (5) then
becomes Φpx, y, z q  mgz C, as the required scalar function.
In the case of F~ 
C2 êr , use spherical coordinates, and get the curl of F~ using:
r

êr r êθ r sin θ êφ

~  F~  1 B B B
∇ 2
r sin θ Br Bθ Bφ
Fr rFθ r sin θFφ

where Fr 
C {r2, Fθ  Fφ  0. This gives

êr r êθ r sin θ êφ
~  F~  r2 sin
1 B B B
∇ Br Bθ Bφ
θ

C { r 2 0 0

with components along êθ and êφ only. That is,

"    *
~  F~
∇  r2 sin
1 B B

r êθ 0 Bφ pC {r q r sin θ êφ 0 Bφ pC {r q
2 2
θ
 r2 sin θ
1
t
r êθ p0q r sin θ êφp0qu
 0.

Hence F~ 
pC {r2 q êr is also a conservative vector field. It can be expressed in the form:
F~ 
∇Φ,
~ where Φ  Φpx, y, z q is a scalar function, and in spherical coordinates, the grad
Φpr, θ, φq is given by
~  B Φ êr 1 B Φ êθ 1 BΦ
∇Φ
Br r Bθ r sin θ B φ
êφ

13
 With Fr 
C {r2, Fθ  Fφ  0, it follows that F~ 
∇Φ
~ gives or

BΦ  C
Br r2 (56)
BΦ  0
Bθ (57)
BΦ  0.
Bφ (58)

We then get
Φpr, θ, φq 

C
constant. (59)
r
Note that you can as well use Cartesian coordinates by first simply transforming the expres-
from spherical coordinates to Cartesian coordinates, using ~r  x î y ĵ z k̂ and
sion of F~ a
r  |~r|  x2 y 2 z 2 . That is,

F~ 
rC3 ~r 
Cpxp2x î y2y ĵ z2qz3k̂{2q


px 2 Cx
y 2 z 2 q3{2
î
2
px
Cy
y 2 z 2 q3{2
ĵ
2
px
Cz
y 2 z 2 q3{2
k̂

 Fx î Fy ĵ Fz k̂

and use

î ĵ k̂
~  F~  B B B
∇ Bx By Bx
Fx Fy Fz

~  B Φ î BΦ ĵ BΦ k̂.
and
∇Φ
Bx By Bz
In this case, you get
Φpx, y, xq 
a
C
constant. (60)
x2 y2 z2

2.2 Surface integrals

Let S be a two-sided surface and let one side be arbitrarily positive (the outer surface). A unit
vector n̂ normal to any point of the positive side of S is termed positive or outward-drawn normal.
If dσ is an element of area of the boundary surface, then a vector d~σ is such that d~σ  n̂ dσ, where
n̂ is a unit vector normal to the surface (or to the element of area). The integral
¼ ¼
~  d~σ
A  ~  n̂ dσ
A (61)
S S

14
 ~ over a closed boundary S. That
is known as the surface integral (or normal surface integral) of A
~
is, the surface integral of the scalar component of A in the direction of a normal to the surface.
Other surface integrals are:
¼ ¼ ¼
~
φ dS, φ n̂ dS, ~  S,
A ~
S S S

where φ is a scalar function.

´
Example 3: Given A ~  px2
y 2 qî 2xy ĵ, find the curl of A.
~ Hence, evaluate ∇~ A
~  d~σ over
a rectangle in the (x, y) plane bounded by the lines x  0, x  a, y  0 and y  b.
Solution: The curl of A~ is

î ĵ k̂ î ĵ k̂
~ A
~ B B B  B B B  p0q î
p0q ĵ  4y k̂
∇ Bx By Bz Bx By Bz 4y k̂
Ax Ay Az p
q
x2 y2 2xy 0
´
To evaluate ∇ ~ A ~  d~σ over a rectangle in the (x, y) plane bounded by the lines x  0, x 
a, y  0 and y  b, notice that ∇ ~ A ~  4y k̂. Then, the area element is d~σ  n̂ dσ  dxdy k̂, and
therefore ∇  A  d~σ is equal to p4y k̂ q  pdxdy k̂ q  p4ydxdy qpk̂  k̂ q  4ydxdy. The integral becomes
~ ~
¼ »a »b
~ A
∇ ~  d~σ 4 dx ydy  4paqpb2{2q  2ab2.
0 0
´
Example 4: Evaluate ~r  n̂ dσ over the whole surface of the cylinder bounded by x2 y 2  1,
z  0 and z  3. Here, ~r  x î y ĵ z k̂.
´
Solution: To evaluate ~r  n̂ dσ over the whole surface of the cylinder bounded by x2 y 2  1,
z  0 and z  3, where ~r  x î y ĵ z k̂, it is much easier to use cylindrical coordinates
(ρ, φ, z) instead of rectangular coordinates (x, y, z). Now, in cylindrical coordinates, x  ρ cos φ,
y  ρ sin φ, x2 y 2  ρ2 , ~r  ρ êρ z êz , and the elemental area, d~σ is given by
$
'
&ρdφdz êρ
d~σ  n̂ dσ  'dρdz êφ
%
ρdρdφ êz

In this case, x2 y 2  1 implies ρ2  1 or |ρ|  1.

On the sides of the cylinder parallel to the z-axis and of constant ρ  1, we have n̂  êρ, and
the surface element directed outward is ρdφdz . That is,
~r  n̂ dσ  pρ êρ z êz q  pρdφdz êρ q  ρ2 dφdz |ρ1  dφdz.
Then, ¼ ¼ » 2π »3
I1  ~r  n̂ dσ  dφdz  dφ dz  p2πqp3q  6π.
0 0
On the top and bottom surfaces, the surface element directed outward is n̂ dσ  ρdρdφêz at
z  3 and
ρdρdφêz at z  0. That is,
~r  n̂ dσ  pρ êρ z êz q  pρdρdφêz q  zρdρdφ.

15
 ´
Then, at z  0, this gives zero, and therefore ~r  n̂ dσ  0 on the bottom surface.
For z  3, the integral becomes
¼ ¼ »1 » 2π
I2  ~r  n̂ dσ  3ρdρdφ  3 ρdρ dφ  3p1{2qp2π q  3π.
0 0

Therefore, ¼
~r  n̂ dσ  6π 0 3π  9π

2.3 Volume Integrals

Consider a closed surface in space enclosing a volume V . Then
½
~ dτ
A (62)
V

is an example of a volume integral, with dτ  dxdydz being the volume element.

µ
Example 5: If F~  2xz î
x ĵ y 3 k̂, find ~
V F dτ , where V is the region bounded by the
surfaces x  0, x  2; y  0, y  6; z  x , z  4.
2

Solution:
½ ½ ½ ½
F~ dτ  î Fx dτ ĵ Fy dτ k̂ Fz dτ
V V V V
»2 »6 »4 »2 »6 »4 »2 »6 »4
 î p2xzq dxdydz
ĵ x dxdydz k̂ y 2 dxdydz

x 0 y 0 z x2   
x 0 y 0 z x2   
x 0 y 0 z x2  
"» 6 » 2 » 4 * "» 6 » 2 » 4 *
 î dy 2x z dz dx
ĵ dy x dz dx
y 0  
x 0 
z x2 
y 0 x 0  z x2
"» 6 » 2 » 4 *
2
k̂ y dy dz dx
y 0 x 0  
z x2
" »2 * " »2 * " »2 *
 
 î p6q x 16
x 4
dx
ĵ p6q xp4
x q dx 2
k̂ p324q 4
x 2
dx
»2
x 0 »2
x 0  »2

x 0
 
 6î x 16
x 4
dx
6ĵ xp4
x q dx 2
72k̂ 4
x2 dx
x 0 x 0 
x 0
 6p64{3q î
6p4q ĵ 72p16{3q k̂  128 î
24ĵ 384 k̂

2.4 Vector theorems

2.4.1 Divergence (or Gauss’s) theorem

Consider a volume region V bounded by a solid surface S. The divergence theorem states that the
normal surface integral of a function F~ over the boundary of a closed surface of arbitrary shape is

16
equal to the volume integral of the divergence of F~ taken throughout the enclosed volume. Note
that F~ and its first derivative must be uniform and continuous.
The equation form of Gauss’s theorem is
¼ ¼ ½
F~  dS
~  F~  n̂ dS  p∇~  F~ q dτ (63)
S S V

´
Consider Example 4 above. We evaluated ~r ´n̂ dσ over the whole surface of the cylinder bounded
by x2 y 2  1, z  0 and z  3, and obtained ~r  n̂ dσ  9π.
Using the divergence theorem, we need the divergence of ~r and the volume element in cylindrical
coordinates:
~  ~r  1 B pρrρ q 1 B rφ Brz ,
ρ Bρ ρ Bφ Bz
∇ (64)

where ~r  rρ êρ rφ êφ ρ êρ z êz . The divergence of ~r then becomes

rz êz

~  ~r  1 B ρ2
 1 Bp0q Bz  2 0 1  3.
ρ Bρ ρ Bφ Bz
∇

Applying the divergence theorem (with dτ  ρdρdφdz), we get

¼ ½ ½ »1 » 2π »3
~r  n̂ dσ  p∇~  ~rq dτ  3ρdρdφdz  ρdρ dφ dz  p3{2qp2πqp3q  9π
0 0 0
V

as obtained earlier.
Example 5: As an application to electrostatics, use the divergence theorem to find the electric
field of a point charge that is a distance r from the charge Q, assuming an electric force, F~  ε0 E,
~
~
where E is the electric field intensity and ε0 is the permittivity of free space.
Solution: We are required to show that E  Q{p4πε0 r2 q using the divergence theorem.
Use Maxwell’s first equation, ∇ ~ E~  ρ{ε0 , called Gauss’s law, where ρ is the volume charge
density. In this case, the divergence theorem takes the form
¼ ½ ½ ½
~  n̂ dσ
E  p∇~  E~ q dτ  ρ
dτ  1
ρ dτ (65)
ε0 ε0
S V V V

Equation (65) is referred to as Gauss’s law in electrostatics. For convenience, we let the arbitrarily
shaped hypothetical surface S or Gaussian surface to be a spherical surface, enclosing the total
charge Q. Then the electric field at a distance r from a charge Q is given by Eq. (65), that is,
¼ ½
~  n̂ dσ
E  1
ρ dτ  εQ , (66)
ε0 0
S V
µ
~ k d~σ (i.e., E
where V ρ dτ gives the total charge, Q. With E ~  n̂ dσ
~ k n̂), we have E  Edσ, and
so the surface integral in Eq. (66) becomes
¼ ¼ ¼
~  n̂ dσ
E  Edσ E dσ  E p4πr2q (67)
S S S

17
 ´
where S dσ gives the total surface area of the spherical surface, and is equal to 4πr2 .
Combining eqtuations (66) and (67), we have

E p4πr2 q 
Q
ε0
from which we get
E  4πεQ r2 or ~
E  4πεQ r2 r̂ (68)
0 0

Example 6: Use the divergence theorem to get the heat equation

BT  ∇~  pκ∇T
~ q
cρ
Bt
from Fourier’s law, F~ 
κ∇T~ , where κ is the thermal conductivity, ρ is the density, and c is the
specific heat capacity of the substance.
Solution: This is a problem of the transfer of heat through a body by means of conduction. We
assume that

(i) the temperature, within this body is a finite and continuous point function T px, y, z, tq, i.e.,
T depends on space and time;

(ii) throughout any point, a surface on which the temperature is everywhere the same (isothermal
surface) may be drawn;

(iii) the temperature gradient, ∇T~ , and the direction of heat flow are normal to the isothermal
surface at the point in question;
~ is a finite and continuous function, and is in the direction of increasing temperature T ,
(iv) ∇T
and

(v) the rate of heat flow per unit area across the isothermal surface, F~ , is given by F~ 
κ∇T
~
(Fourier’s law), where κ is the thermal conductivity of the substance under investigation. The
negative sign indicates that heat flows from points at higher temperature to points at lower
~ points in teh direction of increasing temperature.
temperatures, i.e., the vector ∇T

Now, if σ is an arbitrary closed surface drawn entirely within a certain region of the substance
under investigation, then the total amount of heat flowing out of σ in time ∆t is given by
¼ ¼ ½
Q  ∆t F~  n̂ dσ 
∆t ~  n̂ dσ
κ∇T 
∆t ~  pκ∇T
∇ ~ q dτ (69)
σ σ τ

(the last equality obtained after using the divergence theorem). Also, Q can be obtained from
½
Q 
∆t
BT dτ
τ
cρ
Bt (70)

18
which represents the amount of heat flowing out of σ in time ∆t, provided c is the specific heat
capacity of the substance, ρ is the density, and B T {B t is the rate of increase in temperature within
the surface σ.
Equating Eqs. (69) and (70), we have
½  
B
Bt
∇  pκ∇T q dτ  0
T ~ ~
cρ
τ

which holds only if

BT  ∇~  pκ∇T
~ q,
cρ
Bt (71)
called the heat equation in thermodynamics.

2.4.2 Stokes’s theorem

It states that if F~ is a vector function which is uniform, finite and continuous along with its first
derivative in any direction, then the tangential line integral of F~ over any closed surface S bounded
by a curve C is equal to the normal surface integral of curl F~ over S. That is
¾ ¼ ¼
F~  d~r  p∇~  F~ q  d~σ  p∇~  F~ q  n̂ dσ (72)
C S S

where n̂ is the unit vector at any point of S drawn in the required sense of description of C.
´ 3, where, given A  px
y qî 2xy ĵ, we obtained the curl of A, and
Example 7: In Example ~ 2 2 ~
went ahead to evaluate p∇  Aq  d~σ over a rectangle in the (x, y) plane bounded by the lines
~ ~
x  0, x  a, y  0 and y  b. We obtained.
¼
p∇~  A~ q  d~σ  2ab2.
¶
For this problem, evaluate I  A ~  d~r around the boundary of the rectangle defined in the (x, y)
plane, and bounded by the lines x  0, x  a, y  0 and y  b, and comment if Stokes’ theorem is
verified by vector A.~
We use A ~  d~r  px2
y 2 qdx 2xydy and integrate along a closed path: y  0, x  a, y  b and
x  0.
Along y  0, we have dy  0 and x : 0 Ñ a. Also, A ~  d~r  x2 dx, so
» »a
a3
I1  ~  d~r 
A x2 dx 
3
C1 0

Along x  a, we have dx  0, y : 0 Ñ b, and A

~  d~r  2aydy, so
» »b
I2  ~  d~r 
A 2aydy  ab2
C2 0

Along y  b, we have dy  0, x : a Ñ 0, and A~  d~r  px2
b2qdx, so

» »0 3
I3  ~  d~r 
A px2
b2qdx  ab2
a3 .
C3 a

19
 Along x  0, we have dx  0, y : b Ñ 0, and A
~  d~r  0, so
»
I4  ~  d~r  0
A
C4

Combining these, we get

¾ 3 a3
I  ~  d~r  I1
A I2 I3 I4  a3 ab2 ab2

3
0  2ab2

´
This gives the same result as ∇ ~ A~  d~σ , and therefore, Stokes’s theorem is verified. Note that,
as given by Eq. (72), Stokes’s theorem states that
¾ ¼
~  d~r 
A p∇~  A~ q  d~σ
C σ

where C denotes the curve bounding the surface, σ.

´
Example 8: Evaluate S p∇ ~  F~ q  n̂ dS, for F~  py 2
z qî pyz 4qĵ
xz k̂, where S is the
surface of the cube x  y  z  0, x  y  z  2 above the px, y q plane.
By Stokes’s theorem,
¼ ¼ ¾
p∇~  F~ q  dS~  p∇~  F~ q  n̂ dS  F~  d~r
S S C

In this case, the boundary C of the surface S is a square bounded by the lines x  0, x  2;
y  0, y  2 in the px, y q plane. Thus
¾ ¾ ¾ ¾ ¾
F~  d~r  F~  d~r F~  d~r F~  d~r F~  d~r
C OP PQ PR RO

where OP is the path along y  0; P Q is the path along x  2; QR is the path along y  2, and
RO is the path along x  0, and F~  d~r  py 2
z qdx pyz 4qdy
xzdz.
Along OP , y  0, dy  0, z  0, dz  0 and x varies from 0 to 2. With F~  d~r  2dx, we get,
¾ »2
F~  d~r  2 dx  4.
0
OP

Along P Q, x  2, dx  0, z  0, dz  0 and y varies from 0 to 2. With F~  d~r  4dy, we get,

¾ »2
F~  d~r  4 dy  8.
0
OP

Also, along QR, y  2, dy  0, z  0, dz  0 and x varies from 2 to 0. With F~  d~r  4dx, we get,
¾ »0
F~  d~r  4 dx 
8,
2
OP

20
and along RO, x  0, dx  0, z  0, dz  0 and y varies from 2 to 0. With F~  d~r  4dy, we get,
¾ »0
F~  d~r  4 dy  8.
2
OP

Hence ¼ ¾
p∇~  F~ q  dS~  F~  d~r  4 8
8
8 
4
S C

2.4.3 Green’s theorem

Consider Stokes’s theorem

¾ ¼ ¼
F~  d~r  p∇~  F~ q  d~σ  p∇~  F~ q  n̂ dσ (73)
C S S

where C is a simple closed path containing a region σ. If we define F~ as a 2-dimensional vector

field in x and y, given by
F~  P px, y q î Qpx, y q ĵ, (74)
where P px, y q and Qpx, y q are smooth functions, i.e., the first derivatives of P and Q are contin-
uous, then the left hand side of Eq. (73) becomes
¾ ¾ ¾
F~  d~r  pP î Q ĵ q  pdx î dy ĵ q  pP dx Qdy q (75)
C C C

and the line integral direction is taken to be anticlockwise or counterclockwise.

Since both P and Q are independent of z, the curl of F~ is given by

Curl F~  ∇~  F~  BBQx
BBPy k̂.

For an elemental area dσ in the (x, y) plane, the outward unit vector n̂ points in the k̂ direction,
giving  
∇  F q  n̂ dσ 
~ ~ B Q BP

k̂  k̂ dσ 
BQ
BP dσ
Bx By Bx By
and therefore, the right hand side of Eq. (73) becomes
¼ ¼ 
p∇~  F~ q  n̂ dσ  BQ
BP
Bx By dσ. (76)
S S

Substituting equations (75) and (76) in (73), we get

¾ ¼ 
pP dx Qdy q 
BQ
BP
Bx By dσ (77)
C S

21
 called Green’s theorem (in a plane). It states that if C is a simple closed path containing a region
σ, with P px, y q and Qpx, y q being smooth functions, then Eq. (77) holds, provided the direction
of the line integral is taken to be anticlockwise, and dσ is the area element.
³Example 9: Show that if C is a simple closed curve, the geometrical area it encloses is equal to
2 C pxdy
ydxq.
1

Solution: Put P  y and Q  x, then is, Green’s theorem gives

¾ ¼  ¼ ¼
1
rp
yqdx xdy s 
1 Bx
Bp
yq  12 p2q dσ  A
2 2 Bx By dσ dσ
C S S S

which is the geometrical area enclosed.

¶
Example 10: Use Green’s theorem to evaluate the integral C p2xdy
3ydxq around a square
placed at the vertices p0, 2q, p2, 0q, p
2, 0q and p0,
2q.
Green’s theorem states that
¾ ¼ 
pP dx Qdy q 
BQ
BP
Bx By dxdy
C A

For P 
3y, Q  2x, we have B P {B y 
3 and BQ{Bx  2, giving

¾ ¼ ¼
? ?
p2xdy
3ydxq  p2
p
3qq dxdy  5 dxdy  5A  5p 8qp 8q  40,
C A A

where A is the total surface area, and ? the square placed at the vertices p0, 2q, p2, 0q, p
2, 0q and
p0,
2q will have each side of length 8, and therefore, area of 8 square units.
Note that, treating this as a line integral takes a lot of time so solve (see below).
Alternatively, around a square with vertices p0, 2q, p2, 0q, p
2, 0q and p0,
2q, the integral can
be evaluated:
¾ ¾ » » » »
p2xdy
3ydxq  F~  d~r  F~  d~r
C1 C2 C3 C4
C C
» » p0, 2q » p
2, 0q » p0,
2q
p2, 0q
 F~  d~r
p0,
2q p2, 0q p0, 2q p
2, 0q

where F~  d~r  2xdy
3ydx.

Along C1 : y  x
2, dy  dx and x varies from 0 to 2. In this case,

F~  d~r  2xdy
3ydx  2xdx
3px
2qdx  p6
xqdx

and » »2
F~  d~r  p6
xqdx  10.
C1 0
Along C2 :, y  2
x, dy 
dx and x varies from 2 to 0. In this case,
F~  d~r  2xdy
3ydx  2xp
dxq
3p2
xqdx  p6
xqdx

22
and » »0
F~  d~r  px
6qdx  10.
C2 2
Along C3 : y x 2, dy  dx and x varies from 0 to -2. In this case,
F~  d~r  2xdy
3ydx  2xdx
3px 2qdx 
px 6qdx

and » »
2
F~  d~r 
px 6qdx  10.
C3 0

Along C4 :, y 
px 2q, dy 
dx and x varies from -2 to 0. In this case,

F~  d~r  2xdy
3ydx  2xp
dxq 3px 2qdx  px 6qdx

and » »0
F~  d~r  px 6qdx  10.
C1
2
Therefore, ¾
p2xdy
3ydxq  40,
C
as obtained by the Green’s theorem.

2.5 Other useful integral relations

1. ½ ¼
~ dτ
∇φ  φ n̂ dσ (78)
τ σ

Proof: Let F~  φ C,
~ where C
~ is a constant vector. On substituting this in the divergence
theorem, Eq. (63), we get
½ ¼ ¼
~  pφ C
∇ ~ q dτ  ~  n̂ dσ
φC  rC~  φ n̂s dσ
τ σ σ


~  pφ C
But, ∇ ~q  ∇φ
~ C
~ ~ C
φ ∇ ~  ∇φ
~ C
~ C
~  ∇φ, ~ C
~ since ∇ ~  0 as C
~ is a constant
looomooon
0
vector. It then follows that
½ ½
¼
~  pφ C
∇ ~ q dτ  ~  ∇φ
C ~ dτ  ~  pφ n̂q dσ
C
τ τ σ

~ is a constant vector, it is not affected by integration. So the above equation can be

Since C
written as:    
½ ¼
~ 
C ~ dτ   C
∇φ ~  φ n̂ dσ 
τ σ

23
 or  
½ ¼
~
C ~ dτ
∇φ
φ n̂ dσ   0
τ σ

Hence ½ ¼
~ dτ
∇φ  φ n̂ dσ (79)
τ σ

~ is a constant vector.
since C

2. ½ ¼
~ A
∇ ~ dτ  n̂  A
~ dσ (80)
τ σ

Proof: Let F~  A~  C,
~ where C
~ is an arbitrary constant vector. Applying the divergence
theorem on F~ , we have
½ ¼
~  pA
∇ ~C
~ q dτ  pA~  C~ q  n̂ dσ
τ σ

~ pA
Use, ∇ ~ C ~q  C
~ p∇~ A ~ q
A~ p∇
~ C ~q  C
~ p∇
~ A
~ q, since ∇
~ C
~  0 as C~ is a constant
vector. Also, pA  C q  n̂  C  pn̂  Aq. We then get
~ ~ ~ ~
½ ¼
~  p∇
C ~ A
~ q dτ  ~  pn̂  A
C ~ q dσ
τ σ

or  
½ ¼
~
C  ~
∇  A~ dτ
n̂  A
~ dσ   0
τ σ

which yields ½ ¼
~ A
∇ ~ dτ  n̂  A
~ dσ (81)
τ σ

~ is a constant vector.
since C

24