MA-111 Calculus II

(D3 & D4 )

Lecture 8

B.K. Das

Department of Mathematics
Indian Institute of Technology Bombay
Powai, Mumbai - 76

February 8, 2022
Vector analysis

Curve and path

Line integrals of vector fields

Orientation of Curves
Rn
Let n ∈ N and Rn be the Euclidean space defined by

Rn := {(x1 , x2 , · · · , xn ) | xj ∈ R; ∀ j = 1, 2, · · · , n},

equipped with the norm

n
X  12
kxk := |xj |2 .
j=1

I Any real number is called a scalar.

I For n ∈ N, any element from Rn is called vector. Note this means
elements of R can be thought of both as a scalar and vector. To
avoid confusion we will talk about vectors in Rn for n > 1.
Basic structure:
For any x := (x1 , x2 , · · · , xn ) ∈ Rn and y := (y1 , y2 , · · · , yn ) ∈ Rn and
any a ∈ R:

x +y := (x1 +y1 , x2 +y2 , · · · , xn +yn ) ∈ Rn , sum of two elements in Rn

ax := (ax1 , ax2 , · · · , axn ) ∈ Rn , Scalar multiplication.

Scalar fields and Vector fields

Let D be a subset of Rn .
Definition: A scalar field on D is a map f : D → R.
Definition A vector field on D is a map F : D → Rn . We choose n ≥ 2.

I A scalar field associates a number to each point of D, whereas a

vector field associates a vector (of the same space) to each point of
D.
I The temperature at a point on the earth is a scalar field.
I The velocity field of a moving fluid, a field describing heat flow, the
gravitational field, the magnetic field etc are examples of various
vector fields.
Vector fields: Examples

−y
F1 (x, y ) = (2x, 2y ) F2 (x, y ) = ( x 2−x
+y 2 , x 2 +y 2 )
Vector fields: Examples

F3 (x, y ) = (y , x) F4 (x, y ) = (−y , x)

Vector fields: Examples

F5 (x, y ) = (sin y , cos x)

The vector fields also occur in nature. Some of this you may have seen in
MA 109 as well.
Gravitation fields

The first figure describes the gravitational field of the earth whereas the
second one describes that of a body with mass M. The red lines denote
the direction of the force exerted on the small particles around the body.
Del operator on Functions
We will assume from now on that our vector fields are smooth wherever
they are defined.
One important class of vector fields are those that are given by the
gradient of a scalar function. We will study these in some detail later.
The del operator on functions: Gradient field We define the del
operator restricting ourselves to the case n = 3:
∂ ∂ ∂
∇= i+ j+ k.
∂x ∂y ∂z

The del operator acts on functions f : R3 → R to give a gradient vector

field :
∂f ∂f ∂f
∇f = i+ j+ k.
∂x ∂y ∂z
Thus the del operator takes scalar functions to vector fields.
Definition
Let f : D ⊂ Rn → R, n = 2, 3 be a differentiable function. Then the
vector field associated to ∇f is called a gradient vector field.
Gradient fields

F1 (x, y ) = (2x, 2y ) = ∇(x 2 + y 2 )

Gradient fields

 
−x −y  p 
F2 (x, y ) = , 2 = ∇ − ln( x 2 + y 2 )
x + y x + y2
2 2
Gradient Vector fields
Gravitational force field is a gradient field: Placing the origin of a
coordinate system at the center of the earth (assumed spherical) with
mass M, the force of attraction of the earth on a mass m whose position
vector is r(x, y , z) = (x, y , z) can be given by Newton’s Law
mMG
F(x, y , z) = − r(x, y , z) = −∇V (x, y , z),
|r(x, y , z)|3
mMG
where V (x, y , z) = − |r(x,y ,z)| . Note that the gravitational force exerted
on the mass m acts towards the origin.
Gradient Vector fields contd.
Coulomb’s law says that the force acting on a charge e at a point
r(x, y , z) = (x, y , z) due to a charge Q at the origin is
F(x, y , z) = −∇V (x, y , z)
where V (x, yz) = Qe/|r(x, y , z)| is the potential. For like charges
Qe > 0 force is repulsive and for unlike charges Qe < 0 the force is
attractive.
Definition (Conservative vector field)
A vector field F is called a conservative vector field if it is a gradient of
some scalar function, i..e, there exists a differentiable scalar function f
such that F = ∇f . In this case, f is called a potential function for F.

Conservative forces are important as work done along a path will be only
dependent on the end points.
Several of the examples we have seen turn out to be gradient vector
fields/ conservative vector field. The natural question to ask is which
vector field is a gradient field.
There is a neat answer to the above question, which we will see later.
Application of the ‘Fundamental theorem for line integrals’.
Not all vector fields will turn out to be gradient vector field.
Not gradient fields

F4 (x, y ) = (−y , x), this vector field is not ∇f for any f .

Not gradient fields

F5 (x, y ) = (sin y , cos x), this vector field is not ∇f for any f .
How do you check this?
Flow lines for vector field
Vector fields also arise as the tangent vectors to the fluid flow.
Or conversely, given a vector field we can talk about its flow lines.
Definition If F is a vector field defined from D ⊂ Rn to Rn , a flow line or
integral curve is a path i.e., a map c : [a, b] → D such that

c0 (t) = F(c(t)), ∀t ∈ [a, b].

In particular, F yields the velocity field of the path c.

Example: Show that c(t) = (cos t, sin t) for t ∈ [0, 2π], is a flow line for
the vector field F(x, y ) = (−y , x).
Ans c0 (t) = (− sin t, cos t) and F(c(t)) = (− sin t, cos t).
Does it have other flow lines? Can you guess by looking at the vector
field?

Ans Yes! c(t) = (a cos t, sin t), t ∈ [0, 2π] and any a > 0.
Flow line: System of ODEs

Finding the flow line for a given vector field involves solving a system of
differential equations, if c(t) = (x(t), y (t), z(t)) then

x 0 (t) = P(x(t), y (t), z(t))

0
y (t) = Q(x(t), y (t), z(t))
0
z (t) = R(x(t), y (t), z(t)),

where the vector field is given by F = (P, Q, R).

Such questions are dealt with in MA108.
Curve and path
Recall a path in Rn is a continuous map c : [a, b] → Rn .
A curve in Rn is the image of a path c in Rn .
Both the curve and path are denoted by the same symbol c.
• Let n = 3 and c(t) = (x(t), y (t), z(t)), for all t ∈ [a, b]. The path c is
continuous iff each component x, y , z is continuous. Similarly, c is a C 1
path, i.e., continuously differentiable if and only if each component is C 1 .

• A path c is called closed if c(a) = c(b).

• A path c is called simple if c(t1 ) 6= c(t2 ) for any t1 6= t2 in [a, b] other
than t1 = a and t2 = b endpoints.
• If we write c(t) = x(t)i + y (t)j + z(t)k in vector notation, the tangent
vector to c(t) is c0 (t), i.e.,

c0 (t) = x 0 (t)i + y 0 (t)j + z 0 (t)k.

• If a C 1 curve c is such that c0 (t) 6= 0 for all t ∈ [a, b], the curve is
called a regular or non-singular parametrised curve.
Examples of curves

• Let c(t) = (cos 2πt, sin 2πt) where 0 ≤ t ≤ 1. This is a simple closed
C 1 (actually smooth) curve.
• Let c(t) = (t, t 2 ) where −1 ≤ t ≤ 5 is a simple curve but not closed.
• Let c(t) = (sin(2t), sin t) where −π ≤ t ≤ π. It traces out a figure 8.
It is not a simple but a closed C 1 curve.
• Let c(t) = (t 3 , t) where −1 ≤ t ≤ 1 for some real numbers a, b is a
part of the graph of the function y = x 1/3 . This is simple but not a
1
closed curve. Though the function y = x 3 is not a smooth function at
origin, but this parametrization is regular.
Work done along a curve
• Recall from Physics, that work done by a particle on which force F is
applied is given by the F.ds where ds is the displacement.
• If this is in one variable it is just the product and given by dot-product
when it is in 2D or 3D space. This idea works when the displacement is
straight line.
• If the particle is moving along a curve c then locally the curve can be
approximated by a straight line.
• For a path c : [0, 1] → Rn for n = 2 or 3 if ∆t = t2 − t1 is very very
small then
∆s = c(t2 ) − c(t1 ) = c0 (tb)(t2 − t1 )
for some tb ∈ [t1 , t2 ] by mean value theorem.
• Then work done will have to be computed over these small intervals
[ti , ti+1 ] for i = 1, . . . , n.
Pn
• Total
Pn work done = i=1 F(tbi ) · c(ti+1 ) − c(ti )
= i=1 (F(tbi ) · c0 (tbi ))(ti+1 − ti ).
Does this remind you of something?
Riemann sum: The limit of these Riemann sums as the length of the
subintervals tends to zero, if it exists, is defined to be the line integral of
the vector field F over the curve c and is denoted by
Z
F.ds.
c
Line integrals of vector fields
Assume that the vector field F : D ⊂ Rn → Rn , for n = 1, 2, is
continuous and the curve c : [a, b] → D is C 1 .
Then we define the line integral of F over c as:
Z Z b
F · ds := F(c(t)) · c0 (t)dt.
c a

If F = (F1 , F2 , F3 ) and c(t) = (x(t), y (t), z(t)), we see that

Z b
F(c(t)) · c0 (t)dt
aZ
b
dx(t) dy (t) dz(t) 
= F1 (c(t)) + F2 (c(t)) + F3 (c(t)) dt.
a dt dt dt

Because of the form of the right hand side the line integral is sometimes
written as Z Z b
F · ds = F1 dx + F2 dy + F3 dz.
c a
The expression on the right hand side is just alternate notation for the
line integral. It does not have any independent meaning.
An example

Example 1: Evaluate Z
x 2 dx + xydy + dz,
c

where c : [0, 1] → R3 is given by c(t) = (t, t 2 , 1).

Solution: Let c(t) = (t, t 2 , 1).
Let F(x, y , z) = (F1 (x, y , z), F2 (x, y , z), F3 (x, y , z)) = (x 2 , xy , 1).
Thus F1 (t, t 2 , 1) = t 2 , F2 (t, t 2 , 1) = t 3 and F3 (t, t 2 , 1) = 1.

We have c0 (t) = (1, 2t, 0), hence

(F1 (t, t 2 , 1), F2 (t, t 2 , 1), F3 (t, t 2 , 1)) · c0 (t) = t 2 + 2t 4 + 0.

Z Z 1
x 2 dx + xydy + dz = (t 2 + 2t 4 )dt = 11/15.
c 0
Example 2 (Marsden, Tromba, Weinstein): Find the work done by the
force field F = (x 2 + y 2 )(i + j) around the loop c(t) = (cos t, sin t),
0 ≤ t ≤ 2π.
Solution: The work done is given by
Z
W = F · ds
c
Z 2π
= F(c(t)).c0 (t) dt
0
Z 2π
= (cos t + sin t)dt
0
= (sin t − cos t)|2π
0 =0
Integrating along successive paths

It is easy to see that if c1 is a path joining two points P0 and P1 and c2

is a path joining P1 and P2 and c is the union of these paths (that is, it
is a path from P0 to P2 passing through P1 ), which is C 1 then
Z Z Z
F · ds = F · ds + F · ds.
c c1 c2

This property follows directly from the corresponding property for

Riemann integrals:
Z b Z c Z b
f (t)dt = f (t)dt + f (t)dt,
a a c

where c is a point between a and b.

This allows us to define integration over peicewise differentiable curves
for example the perimeter of a square.
Let the curve c be a union of curves c1 , . . . , cn . We often write this as
c = c1 + c2 + . . . cn , where end point of ci is the starting point of ci+1
for all i = 1, . . . , n − 1.
Then we can define
Z Z Z
F · ds := F · ds + . . . + F · ds.
c c1 cn

Divide the curve c at Ra pointRp into two Rcurves c1 and c2 . Then there it
is easy to verify that c F· = c1 F · ds + c2 F · ds.
Let c be a curve on [a, b] and e c(t) = c(a + b − t), that is the curve e
c
traversed
R in
R the reverse direction and is denoted by −c. What is
c
F · ds + −c F · ds?
R R
c
F · ds = − −c F · ds (use change of variables formula).
Different parametrizations of the same path

Example 1: Let c1 (t) = (cos t, sin t) for 0 ≤ t ≤ 2π. Then

c2 (t) = (cos 2t, sin 2t) for 0 ≤ t ≤ π, the paths are different as a
function but the curves traversed are the same.
Example: 2: Here is an example of a simple path with three different
parametrisations with the same domain.
Take the straight line segment between (0, 0, 0) and (1, 0, 0).
Here are three different ways of parametrising it:

{t, 0, 0)}, {(t 2 , 0, 0)} and {(t 3 , 0, 0)},

where 0 ≤ t ≤ 1.
Reparametrisation
Let c : [a, b] → Rn be a path which is non-singular, that is, c0 (t) 6= 0 for
all t ∈ [a, b].
I Suppose we now make change of variables t = h(u), where h is C 1
diffeomorphism (this means that h is bijective, C 1 and so is its
inverse) from [α, β] to [a, b]. We let γ(u) = c(h(u)).
I We will assume that h(α) = a and h(β) = b.

I Then γ is called a reparametrisation of c.

I Because h is a C 1 diffeomorphism, γ is also a C 1 curve.

The line integral of a vector field F along γ is given by

Z Z β Z β
F · ds = F(γ(u)) · γ 0 (u)du = F(c(h(u)) · c0 (h(u))h0 (u)du,
γ α α

where the last equality follows from the chain rule. Using the fact that
h0 (u)du = dt, we can change variables from u to t to get
Z Z b Z
F · ds = F(c(t)) · c0 (t)dt = F · ds.
γ a c
Orientation of Curves

For given two points P and Q on Rn for n = 2, 3, and a path connecting

them, we can ask whether the path is traversed from P to Q or from Q
to P?
Since a path from P to Q is a mapping c : [a, b] → Rn with c(a) = P and
c(b) = Q, (or vice-versa), it allows us to determine the direction in which
the path is traversed. This direction of the path is called its Orientation.

If the reparametrization γ(·) = c(h(·)) preserves the orientation of c, then

Z Z
F.ds = F.ds.
γ c

If the reparamtrization reverses the orientation, then

Z Z
F.ds = − F.ds.
γ c
Curves on plane
Let us consider the paths lying in R2 , namely, Planar curves.
For a simple closed planar curve, we get a choice of direction- clockwise
or anti-clockwise.
Ex. γ(θ) = (cos(θ), sin(θ)), θ ∈ [0, 2π]. This is a circle with direction
anti-clockwise.
Set γ1 (θ) = (cos(θ), − sin(θ)), θ ∈ [0, 2π]. It is circle with clockwise
direction.
The argument just made shows that the line integral is independent of
the choice of parametrisation - it depends only on the image of the
non-singular parametrised path.

I A geometric curve C is a set of points in the plane or in the space

that can be traversed by a parametrized path in the given direction.
Often the line integral
R of a vector Rfield F along a ‘geometric curve’
C is represented by C F.ds or by C F1 dx + F2 dy + F3 dz.
R
I To evaluate C F.ds, choose a convenient parametrization c of C
traversing C in the given direction and then
Z Z
F.ds := F.ds.
C c
I
I ‘ ’ means the line integral over a closed curve C .
C