Vector Calculus

I. ORDINARY INTEGRAL

In an ordinary integral we calculate the area made by the x axis and the curve. When we
calculate
Z b
f (x)dx (1)
a

We divide the length between the interval [a, b] into small segments of ∆x, where a − b = n∆x. we
calculate the area as follows.

A ≈ f (x∗1 )∆x + f (x∗2 )∆x + f (x∗3 )∆x + .... + f (x∗n )∆x (2)

When n goes to infinity we can write

Z b n
X
f (x)dx = limitn → ∞ f (x∗i )∆x (3)
a i

II. LINE INTEGRAL

A. Arc Length

Let us first understand the line integral of a curve of scalar quantity and then we move to the
Line integral of vector quantity. For this, we need to understand the arc length. We use definite
integrals to find the arc length of a curve. We can think of arc length as the distance you would

ht

Figure 1.
 2

travel if you were walking along the path of the curve. Many real-world applications involve arc
length. If a rocket is launched along a parabolic path, we might want to know how far the rocket
travels. Or, if a curve on a map represents a road, we might want to know how far we have to drive
to reach our destination.
We begin by calculating the arc length of curves defined as functions of x, then we examine the
same process for curves defined as functions of y . However, the process is identical, even if the
roles of x and y are swapped. The techniques we use to find arc length can be later extended to
find the surface area of a surface of revolution.
Let us find the arc length of the curve y = f (x). Here, the function f (x) is differentiable and
we require f ′ (x) to be continuous, i.e., smooth. Let the function be defined within the interval
[a, b] and our task is to calculate the length of the curve from the point (a, f (a)) to the point
(b, f (b)). For this, we consider n number of small segments of the curve so that each segment is
very small. Let a = x0 and b = xn and a small segment can be from the point (xi , f (xi )) to
(xi+1 , f (xi+1 )) where xi is any arbitray point in the graph and ∆x = xi+1 − xi . Similarly the
change in y direction ∆yi = f (xi ) − f (xi−1 ). From Pythagorean theorem, the length of the line
segment can be approximated as

p
dsi = (∆x)2 + (∆yi )2 (4)

which can be written as

r
∆yi 2
dsi = ∆x 1+( ) (5)
∆x
Now, we use the Mean Value Theorem, which says that there is a point x∗i within the interval
[xi−1 , xi ], such that f ′ (x∗i ) = (∆yi /∆x). Then, we can write
q
dsi = 1 + (f ′ (x∗i ))2 ∆x (6)
Pn
Then the arc length S = i=1 dsi can be written as

n q
X
S= 1 + (f ′ (x∗i ))2 ∆x (7)
i=1

Taking the limit n → ∞ leading to ∆x → 0, we have the arc length in integral form as

n q
X Z bp
S = Limit(n → ∞) ′ ∗ 2
1 + (f (xi )) ∆x = 1 + (f ′ (x))2 dx (8)
i=1 a
 3

1. Parametric equations and Curves

We can take the example of the circle

x2 + y 2 = a2 (9)

we can always write x as a function of y and vice versa. However, many such equations cannot be
written in simple forms. For this, we use the parametrization of curves. For example, for circle, we
can parametrize it using just one variable by taking x = aCosθ and y = aSinθ where θ ∈ [0, 2π].
Similarly, we can generalize the concept as x = h(t) and y = g(t) with t ∈ [ta , tb ]. This will be very
useful when we consider vector calculus.
Let us now look at the Arc length and how it can be written in a modified form using this
parametric equation. By noting that y = f (x)

s 2
Z b  Z b
dy/dt 1dx
q
S = 1+ dx = dx
(dx/dt)2 + (dy/dt)2 dt (10)
a dx/dta dt dt
Z bq Z bq
= (dx/dt)2 + (dy/dt)2 dt = (h′ (t))2 + (g ′ (t))2 dt (11)
a a

B. The Line Integral

The line integral of a curve is defined as

Z b
L= F (x, y)ds (12)
a

By using parametric equations, we can write

Z b q
L= F (h(t), g(t)) (h′ (t))2 + (g ′ (t))2 dt (13)
a

Let us now consider the vector ⃗r(t) = h(t)î + g(t)ĵ. Then,

|r′ (t)| =
p
(h′ (t))2 + (g ′ (t)))2 (14)

We can then write

Z b
L= F (h(t), g(t))|r′ (t)|dt (15)
a
 4

Figure 2.

Let us consider a 2D vector F⃗ (h(t), g(t)), then we can always write

Z b Z b
d⃗r
L= F. dt = F.d⃗r (16)
a dt a

r′ (t)
⃗
Note that d⃗
r
dt ≡ T⃗ is the tangent vector so that T̂ = |r′ (t)| . Using this, we can re-write the above as
Z b Z b Z b
d⃗r
L= F⃗ . dt F⃗ .T̂ |r′ (t)|dt = F⃗ .T̂ dr (17)
a dt a a

However, we can write

F⃗ (x, y) = M
⃗ (x, y)î + N
⃗ (x, y)ĵ (18)

⃗ = dxî + dy ĵ. In that case, we have

and dr
Z Z
L= M (x, y)dx + N (x, y)dy (19)
x y

Problem-1: Evaluate the line integral for the vector F = y 2 î + 2x(y + 1)ĵ between two points
a = (1, 1) and b = (2, 2), joining by two curves C1 and C2 .
 5

Figure 3.

Solution: Let us first take curve C1 which can be parameterized as x = t and y = t with t ∈
[1, 2]. The curve in the vector form can be written as ⃗r(t) = tî + tĵ. The vector F⃗ = t2 î + (2t2 + 2t)ĵ
d⃗
r
and dt = î + ĵ. The Line integral can then be written as
Z Z 2
d⃗r
2
F⃗ (t). dt = t2 + 2t2 + 2t dt = t3 + t2 1 = 10


L = (20)
dt t=1

Now, we go for the curve C2 .

Line-1: x = t and y = 1, i.e., ⃗r(t) = tî + ĵ and d⃗

r
dt = î. We have F⃗ = î + 4tĵ
Z 2
L= dt = 1 (21)
t=1

Line-2: x = 1 and y = t, i.e., ⃗r(t) = î + tĵ and d⃗

r
dt = ĵ. We have F⃗ = t2 î + 4(t + 1)ĵ
Z 2
2
L= 4(t + 1)dt = 4(t2 /2 + t 1 = 16 − 6 = 10 (22)
t=1

The total line integral for curve C2 is 11.

Problem-2: Evaluate the line integral for the vector F = (x − y)î + (x + y)ĵ between two points
in the curve a = (1, 1) and b = (2, 4), along the curve C represented by y = x2 .
Solution:
 6

Method-1:

Z Z
L = M (x, y)dx + N (x, y)dy (23)
x y
2 4
√
Z Z
= (x − x2 )dx + (y + y)dy (24)
x=1 y=1
2  2 4
x2 x3

y 2 3/2
= − + + y (25)
2 3 1 2 3 1
8 1 1 16 1 2 7 34
= 2− − + +8+ − − =9+ = (26)
3 2 3 3 2 3 3 3

Method-2:

d⃗
r
Let x = t and y = t2 , with t ∈ [1, 2]. Then ⃗r(t) = tî + t2 ĵ and dt = î + 2tĵ. We have
F⃗1 = (t − t2 )î + (t + t2 )ĵ.

Z 2    
L = (t − t2 )î + (t + t2 )ĵ . î + 2tĵ dt (27)
t=1
Z 2 Z 2
2 2 3
= (t − t + 2t + 2t )dt = (t + t2 + 2t3 )dt (28)
t=1 t=1
 2 2
t t3 t4 8 1 1 1
= + +2 = (2 + + 8) − ( + + ) (29)
2 3 4 1 3 2 3 2
8 1 7 34
= 10 + − 1 − = 9 + = (30)
3 3 3 3

Problem-3: Evaluate the line integral for the vector

i) F1 = y 2 î + 2xy ĵ ii) F2 = xy 2 î + 2xy ĵ between two points a = (0, 0) and b = (2, 1), along the
curve C1 and C2 .
d⃗
r
Solution-i: For C1 , x = 2t, y = t, with t ∈ [0, 1] then ⃗r(t) = 2tî + tĵ and dt = 2î + ĵ. We have
F⃗1 = t2 î + 4t2 ĵ
Z 1
L= 6t2 dt = 2 (31)
t=0

For C2 , Line-1: x = t, y = 0, with t ∈ [0, 2] then ⃗r(t) = tî and d⃗

r
dt = î. We have F⃗1 = 0 and hence
Line integral is zero.

For C2 , Line-2: x = 2, y = t, with t ∈ [0, 1] then ⃗r(t) = 2î+tĵ and d⃗

r
dt = ĵ. We have F⃗1 = t2 î+4tĵ
 7

Figure 4.

and hence Line integral is

Z 1
L= 4tdt = 2 (32)
t=0

Hence, this line integral is independent of the path. Solution-ii: For C1 , x = 2t, y = t, with
t ∈ [0, 1] then ⃗r(t) = 2tî + tĵ and d⃗
r
dt = 2î + ĵ. We have F⃗2 = 2t3 î + 4t2 ĵ

1 1
4t3
Z 
3 2 4
L= (4t + 4t )dt = t + = 7/3 (33)
t=0 3 0

For C2 , Line-1: x = t, y = 0, with t ∈ [0, 2] then ⃗r(t) = tî and d⃗

r
dt = î. We have F⃗1 = 0 and hence
Line integral is zero.

d⃗
r
For C2 , Line-2: x = 2, y = t, with t ∈ [0, 1] then ⃗r(t) = 2î + tĵ and dt = ĵ. We have
F⃗1 = 2t2 î + 4tĵ and hence Line integral is
Z 1
L= dt = 2 (34)
t=0

But the line integral of the vector F⃗2 is dependent on the path.

How do we know which line integral is independent of the path? What is the physical signifi-
cance of it? We answer those questions soon.
 8

C. Scalar field and Vector field

A physical quantity can be expressed as a continuous function of the position of a region of space.
If x, y, z are the coordinates of a point P , and the physical quantity is A, then if A = A(x, y, z),
then this function is called a point function and the region in which it specifies the physical quantity
is known as field.

D. Scalar field and Gradient of a scalar field

When a scalar physical quantity is expressed from point to point in a region of space by a
point function, say, T = T (x, y, z), which gives the value of the physical quantity at every point,
the region is called a scalar field. Example. density, temperature, electric potential, gravitational
potential etc. At every point the value is unique. It may be possible that in a region of space, the
value of the physical quantity is the same, but at the same point, there cannot be two values.
Before introducing the gradient, let us recapitulate what we know about the ordinary derivative
of the function of a single variable. Let us have a function f (x). and
df
df = dx (35)
dx
df
What does this dx tell us? It tells how fast or slow the function varies. If f increases rapidly, the
df df
dx is large. Geometrically, dx gives us the slope.
Now, consider that there is a fire source in one corner of this room. The temperature can be
the function of all x, y, z, and we denote this point function as T (x, y, z) - which is our scalar field.
If I move one side, the temperature may be less than in the other direction. What happens? The
question is how fast the temperature varies. It is difficult to answer this question as the change
in temperature is the combination of three variables and not one variable. In fact, there can be
infinitely many answers possible.
However, we have a powerful theorem from partial derivative, which says that
∂T ∂T ∂T
dT = dx + dy + dz (36)
∂x ∂y ∂z
This simply gives how T changes if we alter the three variables by infinitesimal amounts dx, dy, dz.
Interestingly, we show that we do not need an infinite number of derivatives. Let us rearrange the
above equation as
  
∂T ∂T ∂T 
dT = î + ĵ + k̂ . dxî + dy ĵ + dz k̂
∂x ∂y ∂z
= ∇T.d⃗r (37)
 9

where
 
∂ ∂ ∂
∇T = î + ĵ + k̂ T (38)
∂x ∂y ∂z

is known as the gradient of T - the generalized notion of derivative we are looking for. Note that
∇ is the vector differential and a ∇T is a vector quantity having direction and magnitude.

Interpretation: Like every vector ∇T has magnitude and direction. Let us write

dT = |∇T ||d⃗r|Cosθ (39)

If we fix the |dr| and search for it in various directions, the maximum change in T comes when
θ = 0. Hence, for a fixed distance |dr|, dT is the greatest when T moves in the same direction as
∇T . Thus, the gradient ∇T points in the direction of maximum increase of the function T , and
the magnitude |∇T | gives the slope along this maximal direction.

p
Problem-4: Find the gradient of r = x2 + y 2 + z 2 and interpret. Think!

Solution:

1 2x 1 2y 1 2z
∇r = p î + p ĵ + p k̂ (40)
2 x +y +z
2 2 2 2 x +y +z
2 2 2 2 x + y2 + z2
2

xî + y ĵ + z k̂ ⃗r
= p = = r̂ (41)
x2 + y 2 + z 2 r

Let us now ask a question what happens if our vector field F⃗ (x, y, z) = ∇T . Replacing F⃗ (x, y, z)
by ∇T , we find
Z b b Z
d⃗r
L = ∇T.d⃗r = ∇T. dt (42)
a a dt
Z b 
∂T dx ∂T dy ∂T dz
= + + dt (43)
a ∂x dt ∂y dt ∂z dt
Z b ⃗
d(T (r(t)))
= dt = T (b) − T (a) (44)
a dt

This implies that there is a special kind of vector field which can be written as a gradient of a
scalar field. If so, the line integral of such vector field is independent of path and depends only on
the end points values of the scaler filed.
 10

Rb
Problem-5: Let a scalar field is defined as T = Sin(πx) + Cos(πy) − xyz. Find a ∇T.d⃗r
where a = (1, 2, 2) and b = (2, 1, −1).

Solution: How do you proceed? The answer is simple.

T (b) − T (a) = (Sin(π) + Cos(2π) − 4) − (Sin(2π) + Cos(π) + 2) (45)

= −3 + 1 = −2

What we find is that if we can write any arbitrary vector field F⃗ as ∇T , then the calculation of
line integral is very simple. Can we always do that? The answer is negative. A vector field F⃗ is
called conservative if and only if it can be written as a gradient of scalar function as ∇T . Such a
scalar function is called potential. Let us find a general way to find the scalar field T from F⃗ . We
discuss it through the problem below.

Problem-6: Consider the vector field

F⃗ (x, y, z) = (2x + y)î + (x + z 3 )ĵ + (3yz 2 + 1)k̂ (46)

Determine whether the F⃗ is conservative or not. If so, find its potential.

Solution: If the F⃗ can be written as F⃗ = ∇T then following relations holds simultaneously.

∂T
= 2x + y (47)
∂x
∂T
= x + z3 (48)
∂y
∂T
= 3yz 2 + 1 (49)
∂z
From Eq. (47), one can find T (x, y, z) by integrating 2x + y by x, regarding y to be a constant.
In single-variable calculus, an integration constant will be added after the integration. However,
we are now considering partial derivatives, and not only constants but also y or z will vanish after
differentiating by x! In other words, the “integration constant” is no longer a constant but instead
a function of y and z. Precisely, we have:

T (x, y, z) = x2 + yx + g(y, z) (50)

By differentiating both sides we get

∂T ∂g
=x+ (51)
∂y ∂y
 11

Comparing with Eq. (48) we get

∂g
= z3 (52)
∂y

Integrating again by y we get

g(y, z) = yz 3 + h(z) (53)

By differentiating both sides with respect to z, we have

∂g
= 3yz 2 + h′ (z) (54)
∂z

Note that from Eq. (50),

∂T ∂g
= (55)
∂z ∂z

By comparing with Eq. (49), we get h′ (z) = 1, that is, h(z) = z + C. This time, this constant is
independent of x, y. Putting together, we have,

T (x, y, z) = x2 + yx + g(y, z) = x2 + yx + yz 3 + h(z) = x2 + yx + yz 3 + z + C (56)

Note that only a handful of vector fields are conservative. Let’s discuss an example.

Problem-7: Let F⃗ (x, y) = y î + xĵ. Determine whether this field is conservative.

Solution: if F⃗ = ∇T , we have

∂T ∂T
= y; =x (57)
∂x ∂y

Solving T (x, y) = xy + g(y). Differentiating it with respect to y, we get

∂T
= x + g ′ (y) (58)
∂y

Comparing, we get

x + g ′ (y) = x ⇒ g ′ (y) = 0 (59)

This gives g = C and we get

T = xy (60)

Problem-8: Let F⃗ (x, y, z) = xî + y ĵ + z k̂. Determine whether this field is conservative.
 12

Solution: if F⃗ = ∇T , we have

∂T ∂T ∂T
= x; = y; =z (61)
∂x ∂y ∂z

x2
Solving T (x, y, z) = 2 + g(y, z). Differentiating it with respect to y, we get

∂T
= g ′ (y, z) (62)
∂y

Comparing, we get

y2
g ′ (y, z) = y ⇒ g(y, z) = + h(z) (63)
2

Differentiating

∂g
= h′ (z) (64)
∂z
∂g z2
Since ∂T
∂z = ∂z , we have h′ (z) = z implying h(z) = 2 +C
Putting all together, we get

x2 + y 2 + z 2
T (x, y, z) = +C (65)
2

Problem-9: Let F⃗ (x, y) = −y î + xĵ. Determine whether this field is conservative.

Solution: if F⃗ = ∇f , we have

∂T
= −y (66)
∂x
∂T
=x (67)
∂y

Solving T (x, y) = −xy + g(y). Differentiating it with respect to y, we get

∂T
= −x + g ′ (y) (68)
∂y

Comparing, we get

−x + g ′ (y) = x ⇒ g ′ (y) = 2x (69)

However, g ′ (y) is a function of y, not of x. Since it leads to inconsistency, such a T cannot exist
and so F⃗ is not conservative.
 13

1. An interesting criteria for simple determination of conservative field

Let us look at a simple check to determine if a vector field F⃗ is conservative. Let us write

F⃗ = M (x, y, z)î + N (x, y, z)ĵ + P (x, y, z)k̂ (70)

Now assume that the vector field can be written as F = ∇T . Then

F⃗ = M (x, y, z)î + N (x, y, z)ĵ + P (x, y, z)k̂ (71)

∂T ∂T ∂T
= î + ĵ + k̂ (72)
∂x ∂y ∂x

From the theorem of partial derivative, we have

   
∂ ∂T ∂ ∂T
= (73)
∂y ∂x ∂x ∂y
∂T ∂T
Identifying M = ∂x and N = ∂y we can write

∂M ∂N
= (74)
∂y ∂x

Finally, we get the conditions as

∂M ∂N ∂M ∂P ∂N ∂P
= = = (75)
∂y ∂x ∂z ∂x ∂z ∂y

If the above conditions are satisfied, we can say that the field is conservative.

Theorem 1: Given a conservative vector field F⃗ = ∇T , where T is a potential function, then

along any path C connecting from point a(x0 , y0 , z0 ) to point b(x1 , y1 , z1 ), the line integral is given
by
Z
F⃗ .d⃗r = T (b) − T (a) (76)
C

Corollary: For a conservative vector field F⃗ , if C1 and C2 are two paths with the same initial and
final positions, then
Z Z
F⃗ .d⃗r = − F⃗ .d⃗r (77)
C1 C2

Thus, if C is a closed path

I
F⃗ .d⃗r = 0 (78)
C
 14

E. Line integral: Concept of circulation (the flow) and the flux

Let us consider two vector fields.

F⃗1 = xî + y ĵ and F⃗2 = −y î + xĵ (79)

Note that, F⃗2 = −F⃗1 × k̂.

We already have
Z b Z b Z b
d⃗r
L= F⃗ . dt F⃗ .T̂ |r′ (t)|dt = F⃗ .T̂ dr (80)
a dt a a

Let us first consider the field F⃗1 and a curve represented by x2 + y 2 = a2 . For this we parameterized
the curve as x = acost and y = asint with t ∈ [0, 2π]. This gives ⃗r(t) = acostî + asintĵ and
⃗r′ (t) = −asintî + acostĵ. Then
We already have
I I
d⃗r
L = F⃗ .T̂ ds =
F⃗1 . dt (81)
dt
Z 2π    
= acostî + asintĵ . −asintî + acostĵ dt (82)
0
Z 2π
−a2 sint · cost + a2 sint · cost dt = 0


= (83)
0

Now, consider the field F⃗2 and a curve represented by x2 + y 2 = a2 . For this we parameterized
the curve as x = acost and y = asint with t ∈ [0, 2π]. This gives ⃗r(t) = acostî + asintĵ and
⃗r′ (t) = −asintî + acostĵ. Then
I I
d⃗r
L = F .T̂ ds = F⃗2 . dt
⃗ (84)
dt
Z 2π    
= −asintî + acostĵ . −asintî + acostĵ dt (85)
0
Z 2π Z 2π
a2 sint2 + a2 cost2 dt = a2 dt = 2πa2


= (86)
0 0

Now, out of curiosity, let us consider a line integral

Z b
L= F⃗ .n̂ds (87)
a

It is seen that n̂ is perpendicular to T̂ and also perpendicular to k̂, ie., n̂ = T̂ × k̂. Then we have
This gives n̂ = asintĵ + acostî Then we have
I I I
d⃗r
L = F1 .n̂ds = F1 · (T̂ × k̂)ds = F⃗1 . dt
⃗ ⃗ (88)
dt
Z 2π    
= acostî + asintĵ . asintĵ + acostî dt (89)
0
Z 2π
= a2 dt = 2πa2 (90)
0
 15

Now, consider the field F⃗2 and a curve represented by x2 + y 2 = a2 . For this we parameterized
the curve as x = acost and y = asint with t ∈ [0, 2π]. This gives ⃗r(t) = acostî + asintĵ and
⃗r′ (t) = −asintî + acostĵ. Then
I I
d⃗r
L = F · T̂ ds = F⃗2 · dt
⃗ (91)
dt
Z 2π    
= −asintî + acostĵ . asintĵ + acostî dt = 0 (92)
0

Let us again consider

I I I I
d⃗r d⃗r ds d⃗r
L= F⃗ · T̂ ds = F⃗2 . dt = F⃗2 · dt = F⃗2 ds (93)
dt ds dt ds
we can write
Z Z  
dx dy
L= F⃗ · T̂ ds = F⃗2 · î + ĵ ds (94)
ds ds
By taking F⃗ = M (x, y)î + N (x, y)ĵ we have
I I   Z b Z b
dx dy
L = F⃗ · T̂ ds = F⃗2 · î + ĵ ds = M (x, y)dx + N (x, y)dy (95)
ds ds a a

Now,

I I   
dx dy
L= F⃗ .n̂ds = F⃗2 · î + ĵ × k̂ ds (96)
ds ds
I     Z b Z b
= M (x, y)î + N (x, y)ĵ . −dxĵ + dy î = M (x, y)dy − N (x, y)dx(97)
a a

F. Green’s theorem in circulation form

We want to calculate the circulation in a square, as depicted in Fig.4.

I I
L = F⃗ · d⃗r = F⃗ .T̂ dr (98)
ZC C
Z Z Z
= F⃗ .T̂ dr + F⃗ .T̂ dr + F⃗ .T̂ dr + F⃗ .T̂ dr (99)
C1 C2 C3 C4

Now,
Z Z Z Z
L = F⃗ · T̂ dr + F⃗ · T̂ dr + F⃗ · T̂ dr + F⃗ · T̂ dr (100)
C1 C2 C3 C4
= F⃗ (x, y) · ∆xî + F⃗ (x + ∆x) · ∆y ĵ + F⃗ (x, y + ∆y) · (−∆xî) + F⃗ (x, y) · (−∆y ĵ) (101)

= M (x, y)∆x + N (x + ∆xy)∆y − M (x, y + ∆y)∆x − N (x, y)∆y (102)

= − (M (x, y + ∆y) − M (x, y)) ∆x + (N (x + ∆x, y)∆y − N (x, y)) ∆y (103)

(M (x, y + ∆y) − M (x, y)) (N (x + ∆x, y)∆y − N (x, y))
= − ∆x∆y + ∆y∆x (104)
∆y ∆x
 16

Figure 5.

Now, by taking limit of ∆y → 0 and ∆x → 0, we have,

(M (x, y + ∆y) − M (x, y)) (N (x + ∆x, y)∆y − N (x, y))

I
F⃗ .d⃗r = − ∆x∆y + ∆y∆x (105)
C ∆y ∆x
∂M ∂N
= − ∆y∆x + ∆x∆y (106)
∂y ∂x
 
∂N ∂M
= − ∆x∆y (107)
∂x ∂y

Now, I take any arbitrary region. I cut it by two (see Fig.5). Then
I I I
⃗
F .d⃗r = ⃗
F .d⃗r + F⃗ .d⃗r (108)
C C1 C2

You can see that the lines L1 and L2 are along different directions and hence they cancel out. We
can now cut it infinitely many times and all cancels out. We find
I n I n   Z Z  
⃗
X
⃗
X ∂N ∂M ∂N ∂M
F .d⃗r = F .d⃗r = − ∆x∆y = − dxdy (109)
C ∂x ∂y ∂x ∂y
i=1 Ci i=1 R
Z Z
= (∇ × F ).k̂dA (110)
R

where
 
i j k
 
∂ ∂ ∂
∇ × F =  ∂x ∂y ∂z 
 (111)

M N P

This is the Green’s theorem in Circulation-Curl form.

 17

Figure 6.

Figure 7.
 18

1. Stokes theorem

A generalization of Green’s theorem is the Stokes theorem. For a three-dimension vector field
F = M (x, y, z)î + N (x, y, z)ĵ + P (x, y, z)k̂ we can write,
I Z Z   Z Z   Z Z  
⃗ ∂N ∂M ∂P ∂N ∂M ∂P
F .d⃗r = − dxdy + − dydz + − dxdz
(112)
C R ∂x ∂y R ∂y ∂z R ∂z ∂x
Z Z
= (∇ × F ).dA (113)
R

where ∇ × F is called the curl of a vector.

F⃗ .dr = 0. This means that whenever ∇ × F = 0, we have the

H
For a conservative vector field,
line integral independent of the path.

Note that, we previously proved that if a vector field can be written as a gradient of a scale field
then, i.e., if F⃗ = ∇T then F⃗ .dr = 0.
H

Can we check ∇ × ∇T ?

 
i j k
 
∂ ∂ ∂
∇ × ∇T =  ∂x =0 (114)

 ∂y ∂z 
∂T ∂T ∂T
∂x ∂y ∂z

Hence, ∇ × F = 0 implies that F = ∇T .

Problem-10: Let F = xz î + xyz ĵ + y 2 k̂. Calculate ∇ × F .

Problem-11: Prove that F⃗ = (y 2 cosx+z 3 )î+(2ysinz−4)ĵ+3xz 2 k̂ is a conservative vector field.

Solution:

III. SURFACE INTEGRAL AND DIVERGENCE THEOREM

Up to now, we have only dealt with the line integral. We now look at the generalization of line
integral - the surface integral. The basic idea remains the same, but the complexity will be slightly
increased. We first consider the scalar field and calculate the surface integral and then we extend
 19

it to the case of a vector field, the same way we did it for the case of the line integral.

The surface integral is defined as

Z Z
L2 = f (x, y, x)dA (115)

We need to parameterize by two variables u and v so that we obtain a vector ⃗r(u, v). In order to
derive the change of surface area, we assume a parallelogram with u to u + du in one direction and
v to v + dv in another direction. We then have a change in u is (∂⃗r(u, v))(∂u) with magnitude du
and a change in v is (∂⃗r(u, v))(∂v) with magnitude dv.

We then have to find the area of the parallelogram as base × height, i.e., |a||b|sinθ. Here, a =
(∂⃗r(u, v))(∂u)du and b = (∂⃗r(u, v))(∂v)dv. we denote ⃗ru = (∂⃗r(u, v))(∂u) and ⃗rv = (∂⃗r(u, v))(∂v)
We can then write

dA = |⃗rv du||⃗rv dv|Sinθ = |⃗ru × ⃗rv |dudv (116)

Then, the surface integral can be written as

Z Z Z Z
L2 = f (x, y, x)dA = F (⃗r(u, v))dA (117)

Problem-11: Let f (x, y, z) = xz + y 2 and the surface is defined by x2 + y 2 = a2 and 0 ≤ z ≤ 2.

Solution: Let us now parameterise by two variables. x = acosθ and y = asinθ and z = z with
θ ∈ [0, 2π] and z ∈ [0, 2]. We can then write ⃗r(θ, z) = acosθî + asinθĵ + z k̂. We calculate
∂⃗r(θ, z)
= −aSinθî + aCosθĵ (118)
∂θ
∂⃗r(θ, z)
= k̂ (119)
∂z

∂⃗r(u, v) ∂⃗r(u, v)
dA = | × |dθdz = |asinθĵ + acosθî|dθdz = adθdz (120)
∂u ∂v
Then
Z 2 Z 2π Z 2 Z 2π
L2 = f (⃗r(u, v))dA = (acosθz + a2 (sinθ)2 )adθdz (121)
z=0 θ=0 z=0 θ=0

Now, coming to the vector field. Then, the surface integral can be written as
Z Z Z Z
L2 = f (x, y, x)dA = F⃗ (⃗r(u, v)).n̂dA (122)
 20

where
⃗ru × ⃗rv
n̂ = (123)
|⃗ru × ⃗rv |
We can then write
Z Z Z Z Z Z
L2 = f (x, y, x)dA = ⃗
F (⃗r(u, v)).n̂dA = F⃗ (⃗r(u, v)). (ru × rv ) dudv (124)
R R

1. a special case

When the suface can be represented as z = g(x, y). We can then parametrized as x, y (i.e., x, y)
plays the role of u, v so that

⃗r(x, y) = xî + y ĵ + z(x, y)k̂ (125)

In that case rx = ∂⃗r(x, y)/∂x = î + ∂z(x, y)/∂x k̂ and ry = ∂⃗r(x, y)/∂y = ĵ + ∂z(x, y)/∂y k̂ This
gives

rx × ry = −∂z(x, y)/∂xî − ∂z(x, y)/∂y ĵ + k̂ (126)

and then
q
|rx × ry | = (∂z(x, y)/∂x)2 + (∂z(x, y)/∂y)2 + 1 (127)

Problem-12: Let F⃗ = x2 î + xy ĵ + z k̂ and the surface is defined by z = x + y 2 and x, y ∈ [0, 1].

Calculate the surface integral.

Solution: Note that this surface is in the form of ⃗r(x, y) = xî + y ĵ + z(x, y)k̂. Thus ⃗r(x, y) =
xî + y ĵ + (x + y 2 )k̂.
Z Z
L2 = F⃗ (⃗r(x, y)). (rx × ry ) dxdy (128)
R

Now, the rx = î + k̂ and ry = ĵ + 2y k̂ which gives rx × ry = −î − 2y ĵ + k̂. Then F⃗ .n̂ = −x2 − xy 2 + z
The integral for the surface dxdy we have,

Z 1 Z 1 Z 1
2 2 2
−x2 y − 2xy 3 /3 + xy + y 3 /3 |1y=0 dx(129)



L2 = −x − 2xy + x + y dxdy =
x=0 y=0 x=0
Z 1
−x2 − 2x/3 + x + 1/3 dx


= (130)
x=0
 1
= −x3 /3 − 2x/ 6 + x2 /2 + x/3 = −1/3 − 2/6 + 1/2 + 1/3 = 1/6 (131)
x=0
 21

IV. SPHERICAL POLAR COORDINATES

Figure 8. Spherical polar coordinates

Here θ is the angle between ⃗r with z-direction. There are two normals, AP on z-axisa and P Q
with the x − y plane. From the Figure, we have OA = z, BQ = y and CQ = x. Since ∠AOP is θ
then ∠OP Q is also θ. By construction, ∠OQP is a right angle. Then we have OQ = rsinθ where
OP is the hypotenuse and OP = r. Also P Q||OZ = 2 and hence z = rcosθ.
Now, consider the ∆OBQ. The angle ∠OBQ is the right angle and OQ is the hypotenuse. This
gives BQ = rsinθsinϕ and OB = x = rsinθcosϕ. We thus have

x = r sinθ cosϕ (132)

y = r sinθ sinϕ

z = rcosθ

We thus have the vector

⃗r = xî + y ĵ + z k̂ (133)
 
= r sinθ cosϕ î + sinθ sinϕ ĵ + cosθk̂ ≡ rr̂
 22

Thus, the magnitude of the ⃗r is r, and the unit vector is r̂. We will have two more orthogonal unit
vectors θ̂ and ϕ̂. By taking derivative of ⃗r with respect to θ we have

θ̂ = cosθ cosϕ î + cosθ sinϕ ĵ − sinθk̂ (134)

taking derivative of ⃗r with respect to θ we have

ϕ̂ = −Sinθ sinϕ î + sinθ cosϕ ĵ (135)

It can be checked that r̂.θ̂=0, r̂.ϕ̂=0 and θ̂.ϕ̂ = 0.

Now, the surface elements of a sphere are dxdy, dydz and dxdz. The volume element is dxdydz.
We derive the surface and volume elements in the spherical polar coordinates.
The small increment of ⃗r gives

d⃗r = drr̂ (136)

Let us consider a small increment dθ along θ̂ direction. Then the revised ⃗r becomes ⃗r′ = OP ′
 
⃗r′ = r sin(θ + dθ) cosϕ î + sin(θ + dθ) sinϕ ĵ + cos(θ + dθ)k̂ ≡ rr̂ (137)

Then the vector P P ′ = OP ′ − OP = r′ − r is given by

 
PP′ = r sin(θ + dθ) cosϕ î + sin(θ + dθ) sinϕ ĵ + cos(θ + dθ)k̂ (138)
 
− r sin(θ) cosϕ î + sin(θ) sinϕ ĵ + cos(θ)k̂

Rearranging we get
 
P P ′ = r ( sin(θ + dθ) − sin(θ)) cosϕ î + ( sin(θ + dθ) − sin(θ)) sinϕ ĵ + (cos(θ + dθ) − cos(θ)) (139)
k̂

By using the property f (x + dx) − f x = f ′ (x)dx when dx is infinitesimally small, we have

 
′
P P = rdθ cosθ cosϕ î + cosθ sinϕ ĵ − sinθk̂ ≡ rdθθ̂ (140)

⃗ Similarly, we can derive, ϕ

where rdθ is the magnitude of the vector θ. ⃗ = rsinθϕ
⃗

Then the elementary surface and volume can be written as dV = r2 sinθdθdϕdr and dA =
r2 sinθdθdϕ

Problem-13: Find the flux of F⃗ = xz î + yz ĵ + z 2 k̂ outward through that part of the sphere
x2 + y 2 + z 2 = a2 lying in the first octant (x, y, z, ≥ 0).
 23

Figure 9. Di section of spherical polar coordinates

Solution: While this problem can be solved other method we will follow the method of deriving
surface integral we just learned.
Let us parameterized the surface by considering x = r sinθ cosϕ y = r sinθ sinϕ and z = rcosθ.
Thus the vector ⃗r(θ, ϕ) can be written as
 
⃗r(θ, ϕ) = r sinθ cosϕ î + sinθ sinϕ ĵ + cosθk̂ (141)

Then
∂⃗r(θ, ϕ)  
rθ = dθ = rdθ cosθ cosϕ î + cosθ sinϕ ĵ − sinθk̂ ≡ rdθθ̂ (142)
∂θ
taking derivative of ⃗r with respect to θ we have
∂⃗r(θ, ϕ)  
rϕ = dϕ = rsinθdϕ − sinϕ î + cosϕ ĵ ≡ rsinθdϕϕ̂ (143)
∂ϕ
Then we have
 
rθ × rϕ = r2 sinθ cosϕ î + sinθ sinϕ ĵ + cosθk̂ = r⃗r(θ, ϕ) (144)

We can also write the parameterized vector as

F⃗ (θ, ϕ) = r2 SinθCosθCosϕî + r2 SinθCosθSinϕĵ + r2 (Cosθ)2 k̂ (145)

 24

Figure 10. For problem 14

The surface integral can then be written as

Z
L2 = ⃗
F .(rθ × rϕ )dθdϕ (146)
R
Z π/2 Z π/2 Z π/2
4 4
= r Sinθ dθdϕ = 2πr Sinθ dθ (147)
θ=0 ϕ=0 θ=0

Problem-14: Verify Stoke’s theorem in a plane for F⃗ = (y − z + 2)î − (yz + 4)ĵ − xz k̂ where S
is the open surface of the cube formed by the planes x = 0, x = 2, y = 0, y = 2 , z = 0 and z = 2
above x − y plane.

Solution:

Problem-15 Verify stoke’s theorem for a vector field F⃗ = (x2 − y 2 )î + 2xy ĵ in a rectangular
region in X-Y plane, bounded by x = 0, x = a, y = 0, y = b. [Ans: 2ab2 ]
Solution
To evaluate:
ZZ
∇ × F · n̂ dS
S

Given F⃗ = (x2 − y 2 )î + 2xy ĵ

 25

Figure 11. for problem 15

î ĵ k̂
∇×F = ∂ ∂ ∂ = î(0) − ĵ(0 − 0) + k̂ [2y − (0 − 2y)] = 4y k̂
∂x ∂y ∂z

x2 − y 2 2xy 0

Since the surface is a rectangle in the xy plane, n̂ = k, dS = dxdy. ∇.F⃗ · n̂ = 4y k̂ · k̂ = 4y.

Order of integration is dxdy. x varies from x = 0 to x = a. y varies from y = 0 to y = b.

Z bZ a b b b
4ay 2
ZZ Z Z 
∇ × F · n̂ dS = 4y dx dy = 4y [x]a0 dy = 4ay dy = = 2ab2
S 0 0 0 0 2 0

Here, the line integral over the simple closed curve C is bounding the surface OABCO, consisting
of the edges OA, AB, BC, and CO. Hence; we can avoid parametrization. However, I suggest you
follow the formal procedure and use parametrization for future problems.

Curve Equation Limit

OA y=0 x = 0 to x = a
AB x=a y = 0 to y = b Therefore,
BC y=b x = a to x = 0
CO x=0 y = b to y = 0
Z Z
F · dr = F · dr
C OABCO
 26

Z Z Z Z Z
F · dr = F · dr + F · dr + F · dr + F · dr
C OA AB BC CO

F · dr = (x2 − y 2 )dx + 2xydy ......... (2)

On OA: y = 0, dy = 0, x varies from 0 to a.

(2)⇒ F · dr = x2 dx

a a
x3 a3
Z Z 
2
F · dr = x dx = =
OA 0 3 0 3

On AB : x = a, dx = 0, y varies from 0 to b.
(2) ⇒ F · dr = 2aydy

b b
2ay 2
Z Z 
F · dr = 2aydy = = ab2
AB 0 2 0

On BC : y = b, dy = 0, x varies from a to 0.
(2) ⇒ F · dr = (x2 − b2 )dx
0 0
x3 a3
Z Z 
2 2
F · dr = (x − b ) dx = − b2 x =− + ab2
BC a 3 a 3

On CO : x = 0, dx = 0, y varies from b to 0.
(2) ⇒ F · dr = 0
Z
F · dr = 0
CO

a3 a3
(2) ⇒ F · dr = + ab2 − + ab2 = 2ab2 ........... (3)
3 3

From (3) and (1),

I ZZ
F · dr = ∇ × F · n̂ dS
C S

Problem-16. Verify Stoke’s theorem in a plane for F⃗ = (2xy − x2 )î − (x2 − y 2 )ĵ, where c is
the boundary of region bounded by y = x2 and x = y 2 .
 27

Figure 12. For problem 16

RR
To evaluate: S ∇ × F · n̂ dS
Given F = (2xy − x2 )î − (x2 − y 2 )ĵ
î ĵ k̂
Curl F = ∇ × F = ∂ ∂ ∂
∂x ∂y ∂z

(2xy − x2 ) −(x2 − y 2 ) 0
 2 2
    
∂(2xy−x2 ) ∂(−(x2 −y 2 )) ∂(2xy−x2 )
= î ∂(0)
∂y − ∂(−(x∂z−y )) − ĵ ∂(0)
∂x − ∂z + k̂ ∂x − ∂y

= î(0) − ĵ(0 − 0) + k̂(−2x − 2x)

= −4xk̂
Since the surface is a rectangle in the xy -plane, n̂ = k̂, dS = dx dy.
Curl F · n̂ = −4xk̂ · k̂ = −4x.
Order of integration is dxdy.

√
Limits: x varies from y 2 to y. y varies from 0 to 1. ⇒
√ √
ZZ Z 1Z y Z 1  2 y
x
∇ × F · n̂dS = −4xdxdy = −4 dy
S 0 y2 0 2 y2
1
y2 y 5 1
Z    
4 1 1 3
= −2 (y − y ) dy = −2 − = −2 − =−
0 2 5 0 2 5 5

Here the line integral over simple closed curve C bounding the surface OAO consisting of the
 28

Figure 13. For problem 17

curve OA and AO
Z Z Z
F · dr = F · dr + F · dr ...........(2)
c OA AO

F · dr = (2xy − x2 )î − (x2 − y 2 )ĵ ........(3)

On OA : y = x2 , dy = 2x dx, x varies from 0 to 1.

(3) ⇒ F · dr = (2xx2 − x2 ) dx − (x2 − x4 )2x dx = (2x3 − x2 − 2x3 + 2x5 ) dx = (2x5 − x2 ) dx.

h i1
2x6 x3
R R1 5 2 1 1
Therefore, OA F · dr = 0 (2x − x ) dx = 6 − 3 0 = 3 − 3 =0
On AO : x = y 2 , dx = 2y dy, y varies from 1 to 0.

(3) ⇒ F · dr = (3y 4 − 2y 5 + y 2 ) dy.

h 5 i0
R0 2y 6 y3
therefore, AO F · dr = 1 (3y 4 − 2y 5 + y 2 ) dy = 3y5 − = − 35 + 1 1
= − 35
R
6 + 3 1 3 − 3

Z
3 3
(2) ⇒ F · dr = 0 − =− .......... (3)
c 5 5

Problem-17. Verify Stoke’s theorem where F⃗ = y 2 î + x2 ĵ + (x + z)k̂ and C is the boundary of

the triangle with vertices at (0,0,0),(1,0,0) and (1,1,0).
 29

Figure 14.

A. Green’s theorem in flux form

Now, let us take a look at the perpendicular direction n̂ of the tangent direction T̂ .
I
L = F⃗ .n̂dr (148)
ZC Z Z Z
= ⃗
F .n̂dr + ⃗
F .n̂dr + ⃗
F .n̂dr + F⃗ .n̂dr (149)
L1 L2 L3 L4

Since n̂ = T̂ × k̂ , we can rewrite

Z Z Z Z
L = ⃗
F .n̂dr + ⃗
F .n̂dr + ⃗
F .n̂dr + F⃗ .n̂dr
L1 L2 L3 L4
       
= F⃗ (x, y).∆x î × k̂ + F⃗ (x + ∆x).∆y ĵ × k̂ + F⃗ (x, y + ∆y). − ∆x î × k̂ + F⃗ (x, y). − ∆y ĵ (150)
× k̂

Considering F⃗ (x, y) = M (x, y)î + N (x, y)ĵ,

L = −F⃗ (x, y).∆xĵ + F⃗ (x + ∆x).∆y î + F⃗ (x, y + ∆y).∆xĵ − F⃗ (x, y).∆y î (151)

= −N (x, y)∆x + M (x + ∆x, y)∆y + N (x, y + ∆y)∆x − M (x, y)∆y

(M (x + ∆x, y) − M (x, y)) (N (x, y + ∆y) − N (x, y))
= ∆x∆y + ∆x∆y
∆x ∆y
Now, by taking limit of ∆y → 0 and ∆x → 0, we have
I
∂M ∂N
F⃗ .n̂dr = ∆y∆x + ∆x∆y (152)
C ∂x ∂y
 
∂M ∂N
= + ∆x∆y (153)
∂x ∂y
 30

Now, I take any arbitrary region. I cut it by two. Then

I I I
⃗
F .n̂dr = ⃗
F .n̂dr + F⃗ .n̂dr (154)
C C1 C2

You can see that the lines ⃗n1 and ⃗n2 are along opposite directions and hence they cancel out. We
can now cut it infinitely many times and all cancels out. We find
I n I n   Z Z  
X X ∂M ∂N ∂M ∂N
F⃗ .n̂dr = F⃗ .n̂dr = + ∆x∆y = + dxdy (155)
C Ci ∂x ∂y R ∂x ∂y
i=1 i=1

This is a flux form of Green’s theorem.

B. Surface integral, flux and Divergence theorem

Before delving into the details of the divergence theorem, let us understand the flux or divergence
of a vector field and its connection with the surface integral.
The most important type of surface integral is the one that calculates the flux of a vector field
across a surface A. If A is a closed surface, like a sphere or cube - that is, a surface with no
boundaries, so that it completely encloses a portion of 3-space - then by convention, it is oriented
so that the outer side is the positive one, i.e., so that the normal vector n̂ always points towards
the outside of A.

Brief notion of flux: Flux is a measure of the number of field lines passing through a surface
in a given amount time. Let we have a velocity field v corresponding to a fluid passing through a
cylinder having surface area dA. Then, total volume of fluid crossing the surface area dA in time ∆t
is equal to the volume contained in cylinder of base dA and height v∆t, i.e., (v∆t)dA. We assumed
that v is along the same direction of the surface n̂dA. For an arbitrary direction of the surface n̂,
making an angle θ with the velocity filed, we have the volume (v∆t)dAcosθ = (v.n̂dA)∆t. Then
the flux (volume per unit second crossing the surface ares dA) is ∆ϕ = v.n̂dA. Then the flux for
the whole closed surface we have
I
ϕ= ⃗v .n̂dA (156)

1. An intuitive proof of divergence theorem using Green’s theorem in flow-flux form

Let us consider an intuitive proof of the divergence theorem through Green’s theorem in flow-flux
form. Then the line integral for L2 , F (x + ∆x, y)∆î and for L4 , F (x, y)∆(−î). Then we have the
 31

line integrals for L2 and L4 is given by

M (x + ∆x, y − M (x, y) ∂M
∆x∆y = ∆x∆y (157)
∆x ∂x

Now we consider the whole y-z plane so that î is the direction and for any small volume ∆z the
direction remains unchanged. The total line integral for the line ∆z is given by

∂M
(F (x + ∆x, y)i∆y + F (x, y)(−i)∆y)∆z = ∆x∆y∆z (158)
∂x

Note that, we have take a small volume with ∆x∆y∆z = dvi . By considering arbitrary less numbers
of small volume element we get
Z Z Z Z Z Z Z
∂M
F îdA + F −̂idA = dx dy dz (159)
y z y z ∂x

Similarly, we consider the lines L1 and L3 we have

Z Z Z Z Z
∂N
F⃗ .ĵdxdz − F⃗ .ĵdxdz = ∆x∆ydz (160)
xz xz z ∂y

Further by considering F⃗ = M î + N ĵ + P k̂, by considering the curve C in x − z plane From Eq.

(152), we can write
Z Z Z Z Z
∂P
F⃗ .k̂dxdy − F⃗ .ĵdxdy = ∆y∆zdx (161)
xy xy x ∂z

Then by considering the infinitely many similar cube we have the closed integral in left and integral
in right as
I Z Z Z   Z Z Z
∂M ∂N ∂P
F⃗ .n̂dA = + + dxdydz = ∇.F dV (162)
∂x ∂y ∂z

which is the divergence theorem.

Statement: Let A be a piecewise, smooth closed surface that encloses volume V in space.
Assume that A is oriented outward, and let F⃗ be a vector field with continuous partial derivatives.
Then
Z Z Z I
∇.F⃗ dV = F.n̂dA (163)

C. Divergence theorem: Alternate proof

Note that you can write any of the proofs. Both of them will carry equal marks.
 32

Figure 15. Divergence of a vector field

Let B be a small box with sides parallel to the three planes, see, Fig.14. Let the center of B
have coordinates (x,y,z) and suppose the edge lengths are ∆x∆y∆z. The normal vector out of
the top of the box is k̂ and the normal vector out of the bottom of the box is −k̂.The vector filed
F⃗ = M î + N ĵ + P k̂. The area along k̂ and −k̂ are same as ∆x∆y. The flux out of the top of the
box can be approximated by

∆z
P (x, y, z + )∆x∆y (164)
2

the flux out of the bottom of the box is

∆z
P (x, y, z − )∆x∆y (165)
2

If we denote the difference between these values as ∆P , then the net flux in the vertical direction
can be approximated by ∆P ∆x∆y. However,

∆P ∂P ∂P
∆P ∆x∆y = ∆z∆x∆y ≈ ∆x∆y∆z = ∆V (166)
∆z ∂z ∂z

Similarly, for other two sides we have

∂M
∆M ∆z∆y = ∆V (167)
∂x
∂N
∆N ∆x∆z = ∆V (168)
∂y

Then the total flux is given by

 
∂M ∂N ∂P
+ + ∆V ≡ ∇.F ∆V (169)
∂x ∂y ∂z

If we take summation of infinitely many such boxes, we have the total flux going outward
 33

X  ∂M ∂N ∂P
 Z Z Z
+ + ∆V = ∇.F dV (170)
∂x ∂y ∂z
Vi

Finally, we have
I Z Z Z
F.n̂dA = ∇.F⃗ dV (171)

⃗ = 4xz⃗i − y 2⃗j + yz⃗k over the cube bounded by

Problem 18: Verify the G.D.T for F

x = 0, x = 1, y = 0, y = 1, z = 0, z = 1.
Solution:
⃗ · ⃗n ds = ⃗ dv
⃗ ·F
RR RRR
Gauss Divergence Theorem is, S F V ∇
⃗ = 4xz⃗i − y 2⃗
Given: F 2 j + yz⃗k

⃗ = 4z − 2y + y = 4z − y
⃗ ·F
∇

Now, the right-hand side (R.H.S) is:

 34

ZZZ
(4z − y) dxdydz
V
Z 1Z 1Z 1
= (4z − y) dxdydz
0 0 0
Z 1Z 1 1
= (4xz − yz) dydz
0 0 0
Z 1Z 1
= (4z − y) dydz
0 0
y2 1
Z 1 
= 4zy − dz
0 2 0
Z 1 
1
= 4z − dz
0 2
h z i1
= 2z 2 −
 2 0

1 3
= 2− −0=
2 2

Now, the left-hand side (L.H.S) is:

ZZ ZZ ZZ ZZ ZZ ZZ ZZ
⃗ ds =
F·n ⃗ ds+
F·n ⃗ ds+
F·n ⃗ ds+
F·n ⃗ ds+
F·n ⃗ ds+
F·n ⃗ ds
F·n
S S1 S2 S3 S4 S5 S6

R h i1
⃗ · n ds + ⃗ · n ds = 1 1 0 dxdy + 1 1 y dxdy = 0 + 1 y2 dx = 1 1 dx = x 1 = 1
RR RR R R R R R  
(i) S1 F S2 F 0 0 0 0 0 2 0 2 2 0 2
0

⃗ · n ds + ⃗ · n ds = 1 1 0 dxdz + 1 1 (−1) dxdz

RR RR R R R R
(ii) S3 F S4 F 0 0 0 0

R1 R1
= 0 + 0 [−x]10 dz = − 0 dz = −[z]10 = −[1] = −1
R1R1 R1R1 R1 1 R1
⃗ · n ds + ⃗
RR RR
(iii) S5 F S6 F · n ds = 0 0 0 dydz + 0 0 4z dydz = 0 + 0 [4zy]0 dz = 0 4z dz
 35

h 2 i1
z 1


=4 2 =4 2 −0 =2
0

Therefore,
ZZ ZZ ZZ ZZ ZZ ZZ ZZ
⃗ · n ds = 1 3
F + + + + + = −1+2=
S S1 S2 S3 S4 S5 S6 2 2

Hence, by the Gauss Divergence Theorem, we have verified that

ZZ ZZZ
⃗ · n ds =
F ∇·F ⃗ dv
S V

⃗ = 4x⃗i − 2y 2⃗j + z 2⃗k taken

Problem 19: Verify the divergence theorem for the function F
over the surface bounded by the cylinder x2 + y 2 = 4 and z = 0, z = 3.
Solution :
Gauss’s divergence theorem is
ZZ ZZZ
⃗ · ⃗n ds =
F ⃗ dv
⃗ ·F
∇
S V

⃗ = 4x⃗i − 2y 2⃗j + z 2⃗k

Given: F

⃗ = 4 − 4y + 2z
⃗ ·F
∇

Limits:

z = 0 to 3

√
x2 + y 2 = 4 ⇒ y 2 = 4 − x2 ⇒ y = ± 4 − x 2
 36

√ √
Therefore, y = − 4 − x2 to 4 − x2 .

When y = 0 ⇒ x2 = 4

⇒ x = ±2
Therefore, y = −2 to 2
ZZZ
R.H.S = ∇ ⃗ dv
⃗ ·F
V

√
Z 2 Z 4−x2 Z 3
= √
(4 − 4y + 2z) dzdydx
−2 − 4−x2 0

√ 3
2 4−x2
2z 2
Z Z 
= √
4z − 4yz + dydx
−2 − 4−x2 2 0

√
Z 2 Z 4−x2
= √
(12 − 12y + 9)dydx
−2 − 4−x2

√
Z 2 Z 4−x2
= √
(21 − 12y)dydx
−2 − 4−x2

√
Z 2 Z 4−x2
=2 21 dy dx (As 12y is an odd function)
−2 0
 37

Z a Z a Z a
Because f (x) dx = 2 f (x) dx for even functions, f (x) dx = 0 for odd functions
−a 0 −a

Z 2 √
2
= 42 [y]0 4−x dx
−2

Z 2 √
= 84 4 − x2 dx (because Even function)
0

hx√
−1 x
i2
= 84 4− x2 + 2 Sin
2 2 0

= 84 0 + 2 Sin−1 (1)
 

π
= 84 · 2 ·
2

= 84 π

ZZ ZZ ZZ ZZ
L.H.S = ⃗ · n ds =
F + +
S S1 S2 S3

Along S1 (bottom):

xy-plane ⇒ z = 0, dz = 0

And ds = dxdy, n = −⃗k

⃗ · n̂ = (4x⃗i − 2y 2⃗j + z 2⃗k) · (−⃗k) = −z 2 = 0

T heref ore F

ZZ ZZ
theref ore ⃗ · n ds =
F 0 ds = 0
S1 S1

Along S2 (top):

xy-plane ⇒ z = 3, dz = 0
 38

And ds = dx dy, n̂ = ⃗k

⃗ · n̂ = (4x⃗i − 2y 2⃗j + z 2⃗k) · (−⃗k)

T heref ore F

= z2 = 9

Therefore
⃗ ds =
RR RR RR
S2
F·n S2
9 dxdy = R
9 dxdy = 9·(Area of the circle) = 9·(πr2 ) = 36π [as, r = 2]

Along S3 (curved surface):

Given x2 + y 2 = 4

Let ϕ = x2 + y 2 − 4
⃗ = ⃗i ∂ϕ + ⃗j ∂ϕ + ⃗k ∂ϕ = 2x⃗i + 2y⃗j
∇ϕ ∂x ∂y ∂z
p √
|∇ϕ| = 4x2 + 4y 2 = 2 4 = 4
∇ϕ 2(x⃗i+y⃗j) x⃗i+y⃗j
ñ = |∇ϕ|
= 4
= 2

The cylindrical coordinates are x = 2 cos θ, y = 2 sin θ, ds = 2dzdθ Where z varies from
0 to 3 and θ varies from 0 to 2π
 
⃗ · n̂ = (4x⃗i − 2y 2⃗j + z 2⃗k) · x⃗i+y⃗j
Now F 2

= 2x2 − y 3 = 2(2 cos θ)2 − (2 sin θ)3

= 8 cos2 θ − 8 sin3 θ = 8 1+cos 2θ 3 sin θ−sin 3θ

2
− 4

Therefore
R 2π R 3
⃗ · n ds = 8 1 cos 2θ
− 3 sin θ sin 3θ
RR

S3
F 0 0 2
+ 2 4
+ 4
2 dz dθ
θ sin 2θ 3 cos θ cos 3θ

2π
= 48 2
+ 4
− 4
− 12
0
   
2π sin(2·2π) 3 cos(2π) cos(3·2π) 0 sin(2·0) 3 cos(0) cos(3·0)
= 48 2
+ 4
− 4
− 12
− 2
+ 4
− 4
− 12

= 48 π + 0 − 43 − 1
− 0 + 43 + 1


12 12
= 48π

⃗ · n ds = 0 + 36π + 48π = 84π

RR
L.H.S = F S

⃗ · n ds = ⃗ dv
⃗ ·F
RR RRR
L.H.S = R.H.S (i.e.) S F V
∇
 39

Problem 20: Using the Divergence Theorem of Gauss, evaluate

⃗ = x3⃗i + y 3⃗j + z 3⃗k and S is the sphere
⃗ · n ds, where F
RR
S
F
x 2 + y 2 + z 2 = a2 .
⃗ · n ds = ⃗ dv
⃗ ·F
R RRR
Solution : S F V
∇
⃗ = x3⃗i + y 3⃗j + z 3⃗k
F
⃗ ·F
∇ ⃗ = 3x2 + 3y 2 + 3z 2 = 3(x2 + y 2 + z 2 )

⃗ · n ds = 3
R RRR
S
F V
(x2 + y 2 + z 2 ) dxdydz

In spherical polar coordinates:

x = r sin θ cos ϕ
y = r sin θ sin ϕ

z = r cos θ
x 2 + y 2 + z 2 = a2
dx dy dz = r2 sin θ drdθdϕ

RRR
=3 V
(x2 + y 2 + z 2 ) dxdydz
RRR
=3 r4 sin θ drdθdϕ
V
R 2π R π R a
= 3 0 0 0 r4 sin θ drdθdϕ
3a5
= 5
[− cos θ]π0 · 2π
12πa5
= 5

V. ADDITIONAL PROBLEMS

1: If F⃗ = (3x2 + 6y)⃗i + (14yz)⃗j + (20xz 2 )⃗k, evaluate F⃗ · d⃗r from (0, 0, 0) to (1, 1, 1)
R
C

along the curve x = t, y = t2 , z = t3 .

Solution:
 40

Given: F⃗ = (3x2 + 6y)⃗i + 14yz⃗j + 20xz 2⃗k

⃗ = dx⃗i + dy⃗j + dz⃗k
dr

⃗ = (3x2 + 6y)dx + 14yzdy + 20xz 2 dz

F⃗ · dr

Given: x = t, y = t2 , z = t3 dx = dt, dy = 2tdt, dz = 3t2 dt

The points (0, 0, 0) to (1, 1, 1) on the curve correspond to t = 0 and t = 1.

⃗
F⃗ · dr
R
c
R1 2
= 0
(3t + 6t2 )dt + 14t5 (2tdt) + 20t7 (3t2 dt)

R1
= 0
(9t2 + 28t6 + 60t9 )dt

= 13 units

⃗ · ⃗n ds if F
⃗ = (x + y 2 )î − 2xĵ + 2yz k̂ and S is the surface of the
RR
2: Evaluate S
F
plane 2x + y + 2z = 6 in the first octant.
Solution: Given F⃗ = (x + y 2 )î − 2xĵ + 2yz k̂

Let ϕ = 2x + y + 2z − 6

Then ∇ϕ = î ∂ϕ
∂x
+ ĵ ∂ϕ
∂y
+ k̂ ∂ϕ
∂z
= 2î + ĵ + 2k̂.

√ √
|∇ϕ| = 4 + 1 + 4 = 9 = 3.

∇ϕ 2î+ĵ+2k̂
n̂ = |∇ϕ|
= 3
.

h i  
F⃗ · n̂ = (x + y 2 )î − 2xĵ + 2yz k̂ · 2î+ĵ+2
3
k̂

= 13 [2(x + y 2 ) − 2x + 4yz]
 41

= 32 (y 2 + 2yz)

= 23 y (y + 2z)

= 32 y (y + 6 − 2x − y) (since 2z = 6 − 2x − y)

= 32 y (6 − 2x)

= 34 y (3 − x).

Let R be the projection of S on the xy-plane.

dx dy
Therefore, ds = |n̂·k̂|

 
2î+ĵ+2k̂
n̂ · k̂ = 3
· k̂ = 23 .

So,
F⃗ · n̂ ds = F⃗ · n̂ dx dy
RR RR
S R |n̂·k̂|

4
− x) dx2dy
RR
= R 3
y(3
3

RR
=2 R
y(3 − x) dx dy.

In R1 : 2x + y = 6, x varies from 0 to 6−y

2
,
y varies from 0 to 6.

RR
=2 R
y(3 − x) dx dy

R 6 R 6−y
=2 0 0
2
y(3 − x) dx dy

R6 h i 6−y
x2 2
=2 0
y 3x − 2
dy
0
 42

R 6 h 6−y (6−y)2
i
=2 0
y 3 2 − 23
dy

R 6  18y−3y2 2

=2 0 2
− y (6−y)
8
dy

R 6  18y−3y2 2

=2 0 2
− y (36+y8 −12y dy

R 6  18y−3y2 (36y+y 3 −12y 2


=2 0 2
− 8
dy

R6 2 −36y−y 3 +12y 2
=2 0
( 72y−12y 8
) dy

R6 3
=2 0
( 36y−y
8
) dy

h i6
9y 2 y4
=2 4
− 32
0

= 81

R
3: Verify Green’s theorem in the plane for C
(3x2 − 8y 2 ) dx + (4y − 6xy) dy where C is
the boundary of the region defined by x = 0, y = 0, and x + y = 1

Solution: We have to prove that

Z ZZ  
∂N ∂M
M dx + N dy = − dx dy
C R ∂x ∂y

Here, M = 3x2 − 8y 2 and N = 4y − 6xy.

∂M ∂N
= −16y, = −6y
∂y ∂x

Therefore,
Z Z
2 2
(3x − 8y ) dx + (4y − 6xy) dy = M dx + N dy
C C

By Green’s theorem in the plane,

 43

R RR  ∂N ∂M

C
M dx + N dy = R ∂x
− ∂y
dx dy
R 1 R 1−x 
= 10 0 0 y dy dx
R1 2
= 10 0 (1−x) 2
dx
R1
= 5 0 (1 − x)2 dx
h i1
(1−x)3
=5 3
0
5
= 3
.....(1)

R R R R
Consider M dx + N dy = OA
+ AB
+ BO

Along OA, y = 0, =⇒ dy = 0, x varies from 0 to 1.

Therefore
R R1 2 1
OA
M dx + N dy = 0
3x dx = [x3 ]0 = 1

Along AB, y = 1 − x =⇒ dy = −dx, x varies from 1 to 0.

Therefore
R R0
AB
M dx + N dy = 1 [3x2 − 8(1 − x)2 − 4(1 − x) + 6x(1 − x)] dx
 44

h i0
3x3 8(1−x)3 4(1−x)2
= 3
− −3
− −2
+ 3x2 − 2x3 )
1
8 8
= 3
+2−1−3+2= 3

Along BO, x = 0, =⇒ dx = 0, y varies from 1 to 0.

Therefore,
R R0 0
BO
M dx + N dy = 1 4y dy = [2y 2 ]1 = −2

Therefore, in total
Z
8 5
M dx + N dy = 1 + −2= ...(2)
C 3 3

From (1) and (2),

I ZZ  
∂N ∂M
M dx + N dy = − dx dy
C R ∂x ∂y

Hence Green’s theorem is verified.

4: Verify Green’s theorem in the XY-plane for

I
(xy + y 2 ) dx + x2 dy
C

where C is the closed curve of the region bounded by y = x and y = x2 .

R RR  ∂N ∂M 
Solution: We have to prove that C
M dx + N dy = R ∂x − ∂y dx dy
Here, M = xy + y 2 and N = x2
∂M
⇒ = x + 2y and ∂N
∂y ∂x
= 2x
 
R.H.S = R ∂N − ∂M
RR
∂x ∂y
dx dy
Limits:
√
x varies from y to y

y varies from 0 to 1
 45

RR  ∂N ∂M
 R 1 R √y
R ∂x
− ∂y
dx dy = 0 y
(2x − (x + 2y)) dx dy
R 1 h x2 i√ y
= 0 2
− 2xy dy
y
R1y √ y 2
  2
= (
0 2
− 2y y) − 2
− 2y dy
R1 √ 2

= 0 y2 − 2y y − y2 + 2y 2 dy
h 2 i1
y 4 5/2 y3 2 2
= 4 − 5y − 6 + 3y
0
1
= − 20
R
L.H.S = C
M dx + N dy
R R R
Consider C
M dx + N dy = OA + AO

Along OA, y = x2 ⇒ dy = 2x dx, x varies from 0 to 1

R R1
⇒ OA
M dx + N dy = 0
[(x · x2 + x4 ) dx + x2 · 2x dx]
 46

R1
= (3x3 + x4 ) dx
0
h 4 5
i1
= 3x4 + x5
0

= 34 + 15

19
= 20

Along AO, y = x ⇒ dy = dx and x varies from 1 to 0

R R0
⇒ AO M dx + N dy = 1 (x2 + x2 ) dx + x2 dx
R1
= 0 3x2 dx
1
= [x3 ]0

= −1
19 1
R
L.H.S = C
M dx + N dy = 20
− 1 = − 20

Therefore, L.H.S = R.H.S

Hence, Green’s theorem is verified.

⃗ = (2x − y)⃗i − yz 2⃗j − y 2⃗⃗k, where

5: Verify Stoke’s theorem for the vector field F
S is the upper half of the sphere x2 + y 2 + z 2 = 1 and C is the circular boundary
on the z = 0 plane.
Solution: By Stokes theorem,
Z ZZ
F ⃗ =
⃗ · dr ⃗ · n̂ dS
Curl F
C S

⃗ = (2x − y)⃗i − yz 2⃗j − y 2⃗k,

Given F

⃗i ⃗j ⃗k
⃗
⃗ =
Curl F ∂
∂x
∂
∂y
∂
∂z
= ⃗i[−2yz + 2yz] − ⃗j[0 − 0] + ⃗k[0 + 1] = ⃗k
2x − y −yz 2 −y 2 z

Here n̂ = ⃗k since C is the Circular boundary on z = 0 plane.

Therefore,
ZZ ZZ
Curl F · n̂ dS = dx dy = area of the circle
S S
 47

And C is a triangle with vertices (0, 0, 0), (1, 0, 0), and (1, 1, 0).
Since the z-coordinate of each vertex is zero, the triangle lies in the xy-plane with corners
(0, 0), (1, 0), and (1, 1).
To evaluate:
ZZ
curl F · n̂ ds
S

In the xy-plane, n̂ = ⃗k, and ds = dx dy.

⃗i ⃗j ⃗k
curl F = ∂
∂x
∂
∂y
∂
∂z
= ⃗i(0) − ⃗j(−1) + ⃗k(2x − 2y) = ⃗j + 2(x − y)⃗k
y 2 x2 −(x + z)

curl F · n̂ = (⃗j + 2(x − y)⃗k) · ⃗k = 2(x − y)

Limits x varies from y to 1.

y varies from 0 to 1.
Therefore
RR R1R1
S
curl F · n̂ ds = 0 y
2(x − y) dx dy
R1 1
= 2 0 12 x2 − xy y dy
R1
= 2 0 21 − 21 y − y + y 2 dy

1
− 21 − 16 + 1

=2 2 3

1
= 3

From this, we conclude that:

I
1
F · dr =
C 3

H
6: Evaluate the integral C
(x + y) dx + (2x − z) dy + (y + z) dz, where C is the
boundary of the triangle with vertices (2, 0, 0), (0, 3, 0), and (0, 0, 6) using Stoke’s
theorem.
 48

Solution:
Z ZZ
F ⃗ =
⃗ · dr ⃗ · n̂ ds
⃗ ×F
∇
C S

Given:

⃗ = (x + y)dx + (2x − z)dy + (y + z)dz

⃗ · dr
F
⃗ = (x + y)⃗i + (2x − z)⃗j + (y + z)⃗k
F
⃗i ⃗j ⃗k
⃗ =
∇×F ∂ ∂ ∂
∂x ∂y ∂z

(x + y) (2x − z) (y + z)

= ⃗i(1 − 1) − ⃗j(0) + ⃗k(2 − 1)

= 2⃗i + ⃗k

Given that C is the triangle with vertices (2, 0, 0), (0, 3, 0), and (0, 0, 6)
The equation of the plane is

x y z
+ + = 1 =⇒ 3x + 2y + z = 6
a b c

Let ϕ = 3x + 2y + z − 6
 49

∂ϕ⃗ ∂ϕ⃗ ∂ϕ ⃗
∇ϕ = i+ j+ k
∂x ∂y ∂z
= 3⃗i + 2⃗j + ⃗k
√ √
|∇ϕ| = 9 + 4 + 1 = 14
⃗
∇ϕ 3⃗i + 2⃗j + ⃗k
n̂ = = √
⃗
|∇ϕ| 14
Let R be the projection on the XY-plane,
dx dy dx dy
so ds = |n̂·⃗k|
= √1
.
14

(3⃗i + 2⃗j + ⃗k) · ⃗k 1

n̂ · ⃗k = √ =√
14 14
⃗ · n̂:
Now, for (∇ × F)

⃗ · n̂ = (2⃗i + ⃗k) ·
⃗ × F) 3⃗i + 2⃗j + ⃗k
(∇ √
14
1
= √ (2⃗i + ⃗k) · (3⃗i + 2⃗j + ⃗k)
14
1
= √ (6 + 1)
14
7
=√
14
Hence
⃗ · n̂ ds
⃗ × F)
RR
S
(∇

= R √714 dxdy
RR

= 7 R √114 dxdy
RR

RR
= 7 R dxdy

= 7 [Area of the triangle]

= 7 · 21 · (2) · (3)
= 21

7: Find the work done in moving a particle in the force field F⃗ = 3x2⃗i + (2xz −
y)⃗j − z⃗k from t = 0 to t = 1 along the curve ⃗r(t) = 2t2⃗i + t⃗j + 4t3⃗k.
 50

Solution: Given F⃗ = 3x2⃗i + (2xz − y)⃗j − z⃗k and dr

⃗ = dx⃗i + dy⃗j + dz⃗k

⃗ = 3x2 dx + (2xz − y)dy − zdz

F⃗ · dr

Given x = 2t2 , y = t, z = 4t3 , and dx = 4tdt, dy = dt, dz = 12t2 dt

⃗ = (48t5 dt + (16t5 − t)dt − 48t5 dt)

F⃗ · dr
R R

R1
= 0
(16t5 − t) dt

 6 1 2 1
= 16
6
t − 2t 0

16 1 13


= 6
− 2
= 6