Notes for Classical Mechanics I, CMI, Autumn 2021

Govind S. Krishnaswami, October 19, 2021

Please let me know at [email protected] of any comments or corrections
Course website http://www.cmi.ac.in/~govind/teaching/cm1-o21

Contents
1 Primer on vectors, polar coordinates and kinematics 1
1.1 Vectors, dot and cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Position coordinates and velocity and acceleration vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Uniform circular motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Nonuniform circular motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1.5 Rotating vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1.6 Integration of kinematical equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.7 Plane polar coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.8 Spherical polar coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.9 Taylor approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.10 Some vector calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2 Newton’s laws and forces 25

2.1 Time, light, simultaneity, space & time intervals, masses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Degrees of freedom, instantaneous configurations, trajectories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3 Newton’s 1st law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

1 Primer on vectors, polar coordinates and kinematics

1.1 Vectors, dot and cross product

• Vectors provide a convenient way of writing the equations of physics in

compact form. Newton’s 2nd law Fx = max , Fy = may , Fz = maz becomes
the single equation F = ma.
• Vectors also make it easier to understand structural features (behavior
under rotations, symmetries, transformation laws) of physical quantities
and allow us to bring to bear tools from linear algebra and geometry.
• A nonzero vector v in 3d Euclidean space is a directed line segment
emanating from the origin. It has a magnitude or length denoted |v| = v
and a direction. Additionally there is one exceptional vector called the
zero vector 0 which has zero length; its direction is not defined. For some
purposes it is convenient to think of the zero vector as one that points in
all directions!
• A unit vector is one with unit length. E.g., given any nonzero vector

1
a, we have the associated unit vector â obtained by ‘normalization’, i.e.,
a
dividing it by its length: â = |a| . Conversely a = aâ. For a unit vector
|â| = 1.
• Examples of vectors include the position vector of a particle moving in
3d space, its velocity and acceleration, the force acting on the particle, the
electric field at a point in space etc.
• Vectors in 3d space form the real vector space R3 in which addition of
vectors is defined as is the multiplication of a vector by a real number.
• The multiplication of a vector v by the real number α denoted αv is
a vector in the same (or opposite) direction as v (according as α ≥ 0 or
α ≤ 0) that has the length |α||v|. For example − 12 v is a vector of half the
length that points in the direction opposite√to v . The real scalar α need
not be rational, it could be irrational like 2 or π . Moreover, 0v = 0 is
the zero vector.
• Pictorially, the sum of the vectors a and b may be obtained by con-
structing the diagonal of the parallelogram with adjacent sides a and b.

• Evidently, the order does not matter: a + b = b + a. We say that

addition of vectors is commutative or an abelian operation.
• The zero vector has the special property that 0 + v = v for any vector
v.
• The zero vector 0 is not the same as the real number 0. The former lies
at the chosen origin of 3d space while the latter is a point on the real line.
They live in different spaces.
• The space of vectors on the 2d Euclidean plane form the real vector space
denoted R2 .

2
Dot or scalar product. Geometry has to do with angles, lengths, notions
of parallel and perpendicular etc. Geometry enters through the dot product
of vectors. For two vectors in 3d space, we define their dot product as
a · b = ab cos θ where θ is the angle between the vectors.
• Notice that a · b = b · a, so the dot product is commutative.
• Turning things around, the angle between vectors can be expressed in
terms of the dot product θ = arccos( a·b
ab ).
• As a consequence of the definition, a · a = a2 . Thus the length of a
vector can also be expressed in terms of the dot product.
• The dot product is also called the scalar product since the result is a
real number (scalar) and not a vector.
• Show the law of cosines for c = a + b:
c2 = a2 + b2 + 2ab cos θ. (1)

• Two vectors are orthogonal or perpendicular if a · b = 0.

• Component of b in the direction of a Note that b cos θ = b · a/a =
b · â is the component of b in the direction of a. The component is just a
real number, it can be positive or negative or even zero. It does not depend
on the length of a.
• The vector Pâ b = b cos θâ is called the orthogonal projection of b on a.
It is a vector that points in the direction of â or −â and has a magnitude
|b cos θ| equal to the absolute value of the component of b along â.
• Similarly, a · b̂ is the component of a in the direction of b.

√
• The norm ||v|| of a vector v is defined as ||v|| = v · v . The norm is
the length of the vector, it is ≥ 0. The zero vector is the only one with
zero norm.
• Work as a scalar product. Suppose a constant force F acts on a
particle and displaces it by the vector s. Then the work done is F · s.

3
• Vector or cross product. The vector or cross product a × b is a
vector with magnitude ab sin θ where θ is the angle between the vectors.
Its direction is determined by the right hand thumb rule. If the fingers
of the right hand curl from a to b then the cross product points in the
direction of the thumb.
• In particular, a × b points in a direction perpendicular to both a and
b. Moreover, a × b = −b × a and a × a = 0 (the zero vector). Thus, the
cross product is not commutative in general.

• The torque due to a force F on a particle located at the position vector

r relative to a given origin is the cross product τ = r × F .
• How might a force point so that it imparts no torque on the particle?
• The area of a parallelogram spanned by the vectors a and b is equal to
|a × b| = |b × a|.

• Linear combination. Given vectors u and v , αu + βv where α, β are

real numbers is called a linear combination.
• Linear dependence. Two vectors u, v are linearly dependent if one
can be expressed as a multiple of the other, i.e., if they point in the same

4
or opposite directions. In other words, u = αv or v = βu for some
real numbers α and β . So when they are dependent, there is a linear
combination u − αv or v − βu that vanishes. We need to allow for both
possibilities. For instance if u = 0v and u 6= 0 then β is formally infinite
and we do not have a relation of the sort v = βu.

• u and v are said to be linearly independent if they point in different

directions (i.e., are neither parallel not antiparallel.).
• Show that u and v are linearly independent if and only if u × v 6= 0.
• The zero vector 0 and any other vector u are always linearly dependent,
since 0 = 0u.
• Linear independence. Vectors v1 , v2 , v3 , · · · are said to be linearly
independent if the only linear combination c1 v1 + c2 v2 + c3 v3 + · · · that
vanishes is the one where c1 = c2 = c3 = · · · = 0. In other words, the only
linear combination that vanishes, is the trivial one.
• In R3 we can have at most three linearly independent vectors.
• Cartesian axes. Given a choice of mutually perpendicular Cartesian
axes x, y, z in R3 we denote the unit vectors along the axes by x̂, ŷ and ẑ
or î, ĵ and k̂ .

• Cartesian orthonormal basis. These three vectors are orthonormal

in the sense that they each have unit norm and they are mutually perpen-
dicular:
x̂ · x̂ = ŷ · ŷ = ẑ · ẑ = 1 and x̂ · ŷ = ŷ · ẑ = ẑ · x̂ = 0. (2)
The (x̂, ŷ, ẑ) frame is called right-handed if x̂ × ŷ = ẑ (rather than −ẑ , in
which case it is left-handed). What are ŷ × ẑ and ẑ × x̂? We will work
with right-handed frames.

5
• x̂, ŷ and ẑ form a basis for R3 in the sense that they are linearly inde-
pendent and any vector can be written (uniquely) as a linear combination
of them:
a = ax x̂ + ay ŷ + az ẑ. (3)
The three real numbers ax , ay and az are the components of a along the
three coordinate axes, verify that
ax = a · x̂, ay = a · ŷ and az = a · ẑ. (4)

• Notice that ax x̂ = Px̂ a is the projection of a along x̂. We say that the
vector has been resolved into its components and written as a sum of its
orthogonal projections along the orthonormal basis vectors.
• In fact, (ax , ay , az ) are the Cartesian coordinates of the location of the
tip of the vector a.
• Express the Cartesian components of the cross product a × b in terms
of those of a and b.

1.2 Position coordinates and velocity and acceleration vectors

• The instantaneous location of a particle moving in 3d Euclidean space

may be specified by its Cartesian coordinates (x(t), y(t), z(t)). As the par-
ticle moves, it traces out a curve parametrized by time, called its trajectory.
It is a directed curve, the direction being that of increasing time.
• The instantaneous location of a particle is the same no matter which
coordinate system we use to describe it. The latter is simply a conve-
nient way of specifying its ‘address’. The coordinates x, y, z depend on

6
the choice of origin and orientation of coordinate axes. If we change the
origin of our coordinates or orientation of the axes, we will get a differ-
ent set of coordinates to describe the location of the particle. Two people
following different coordinate systems will nevertheless meet each other at
the common instantaneous location of the particle. [For instance, a courier
delivers a letter to the same geographic location irrespective of whether
the address on the envelope says CMI, Old number 2, 2nd Avenue or CMI,
New number 5, 2nd Avenue].

• The vector that points from the origin of Cartesian coordinates to the
instantaneous position of the particle, has components (x(t), y(t), z(t)). It
is called the position vector and is denoted r(t). While such a designation
is convenient for some purposes, it is important to bear in mind that the
location of a particle is not really a vector: it is not physically associated to
a direction and the location of the particle does not come with any intrinsic
notion of an origin.
• The infinitesimal displacement of a particle over a time [t, t + δt] does
define a vector, albeit a vector with infinitesimal length:

δr(t) = r(t + δt) − r(t). (5)

The concept of infinitesimal displacement defines an origin, namely the

initial location of the particle (at time t). The infinitesimal displacement
vector then points from this origin to the final location of the particle (at
time t + δt).
• The concept of infinitesimal displacement does not define a coordinate
frame, it only defines an origin and a vector δr . We may resolve δr along
the axes of any frame. Here, we will parallel transport δr to the origin of
our Cartesian coordinate frame, its components with respect to this frame

7
are
δr(t) = r(t+δt)−r(t) = (x(t+δt)−x(t), y(t+δt)−y(t), z(t+δt)−z(t)) (6)
• The arbitrarily chosen origin of the Cartesian coordinate system (x, y, z)
has no physical relevance to the infinitesimal displacement vector. We have
parallel transported it to this origin for ease of some later calculations.
• The velocity of the particle is defined as the limiting value of the differ-
ence quotient
δr(t)
v(t) = lim . (7)
δt→0 δt

• The velocity of the particle is a vector. It is the time derivative of the

position along the trajectory. The velocity
dr(t)
v(t) = (8)
dt
defines a vector that points from the instantaneous location in the direction
of motion. It is a tangent vector to the trajectory. Its magnitude is called
the instantaneous speed of the particle.
• The origin from which the velocity vector points moves with the particle.
• For many purposes, it is convenient to resolve v along the Cartesian
coordinate axes by moving the origin of the Cartesian frame to the current
location of the particle. With this understanding,
v(t) = (ẋ(t), ẏ(t), ż(t)) = ẋx̂ + ẏ ŷ + ż ẑ (9)
• The time-derivative of the velocity is the acceleration, which may be
viewed as the vector
a(t) = ẍ(t)x̂ + ÿ(t)ŷ + z̈(t)ẑ. (10)
It is the limit of the difference quotient (v(t + ∆t) − v(t))/∆t as ∆t → 0.
It may be regarded as a vector emanating from the instantaneous location
of the particle.

8
1.3 Uniform circular motion

Suppose a particle moves counterclockwise on the circle x2 + y 2 = `2 of

radius ` in the x-y plane at a constant angular speed ω > 0 radians per
second. Assuming it starts from the point (`, 0) at t = 0, its instantaneous
location may be given by the Cartesian coordinates

x(t) = ` cos ωt and y(t) = ` sin ωt. (11)

Sometimes, it is convenient to regard r(t) = ` cos ωt x̂+` sin ωtŷ as a vector

that points radially outwards from the center of the circle. Thus r(t) is
the position vector of the particle relative to the origin chosen to lie at the
center of the circle.

• Let us see why this formula is justified. Notice that x2 + y 2 = `2 at

all times and that the motion is counterclockwise. The velocity vector is
given by
v(t) = −`ω sin ωt x̂ + `ω cos ωt ŷ. (12)
√
The speed of the particle v = |v(t)| = v · v = `ω is constant ensuring
uniform circular motion. The particle goes round the circle once in a time
T = 2π`/v = 2π/ω . Thus, the particle covers 2π radians in a time 2π/ω
resulting in an angular speed of ω radians per second (angular speed is
sometimes called angular frequency).
• Notice that v · r = 0. Thus, the velocity vector is tangent to the circle.
The acceleration is given by

a(t) = v̇(t) = −`ω 2 cos ωtx̂ − `ω 2 sin ωtŷ = −ω 2 r(t). (13)

We see that the acceleration has the constant magnitude |a(t)| = `ω 2 and
points radially inwards towards the center of the circle. The latter feature
justifies the name centripetal acceleration. Centripetal means ‘seeking the
center’ in Latin.

9
• The time derivative of acceleration ȧ is sometimes called jerk or jolt.
Show that ȧ · a = 0. Which way does the jerk point in uniform circular
motion?
• We observe that if there is a radially inward force F (like a person tug-
ging at a string with a stone tied at the other end and rotated) that is
responsible for this circular motion, then F · ds = 0. Here ds is the in-
finitesimal displacement of the particle, which is tangent to the circle. This
dot product is called the infinitesimal work done by the force in displacing
the particle ds. Thus we see that a radially inward force does no work
in moving a particle along a circular trajectory. This is not the case of
the motion is due to a force that is tangential - like an agent pushing the
particle along the rim of the circle.

1.4 Nonuniform circular motion

• We may model nonuniform circular motion of a particle around a circle

of radius ` via the position vector
r(t) = `(cos θ(t), sin θ(t)) = `(cos θ(t)x̂ + sin θ(t)ŷ). (14)
If we denote ω(t) = θ̇, then the angular speed of such a particle is |ω(t)| =
|θ̇|, which we suppose is not constant.
• The velocity of such a particle is
v(t) = `θ̇(− sin θ(t), cos θ(t)). (15)
Notice that v · r = 0. So v always points tangent to the circle as it must
for a particle confined to the circle. However, the speed `|θ̇| may vary with
time.
• The acceleration is given by
a = v̇ = `θ̈(− sin θ, cos θ) − `θ̇2 (cos θ, sin θ). (16)

10
The first term points tangentially, and is called the angular acceleration
while the second term points radially inwards and is called the centripetal
acceleration. Thus a · v 6= 0 in general for nonuniform circular motion.

1.5 Rotating vectors

• We observed that the velocity of a uniformly rotating particle is orthog-

onal to its radius vector, i.e., the time derivative of the position vector r
is perpendicular to r : r · (dr/dt) = r · v = 0. In fact, this is true even if
the rotation is not uniform. Let us comment on the significance of this.
• Suppose A is any vector such that its time derivative is orthogonal to
A. Then the infinitesimal change in A in a short time is perpendicular to
A and not along A. This means the magnitude of A cannot change, and
the vector can only rotate. Let us obtain a formula for |dA/dt|. Suppose
∆A = A(t + ∆t) − A(t) is the infinitesimal change in A. Since the length
of A does not change, the three vectors A(t), A(t + ∆t) and ∆A form an
isosceles triangle with ∆A as base. Let us denote the angle at the apex of
this isosceles triangle by ∆θ, which is the angle of rotation.

• Then

|∆A| = |2A sin(∆θ/2)| ≈ A|∆θ| for small ∆θ. (17)

We have used the linear Taylor approximation for the sine function (more
on this soon). Taking the limit ∆t → 0,
   
 dA 
 = A  dθ  .
 

 dt   dt  (18)

Here, |dθ/dt| is the angular speed of A.

• We may apply this to circular motion where A = r is the radius vector
of the particle and dA/dt = v is its velocity. Then the linear speed of the
particle is v = r|ω| where ω = dθ/dt is the angular speed (positive for

11
counterclockwise motion). Note that v, ω need not be constant. Uniform
circular motion is a special case where ω is a constant and
r = r(cos ωt x̂ + sin ωt ŷ) and v = rω(− sin ωt x̂ + cos ωt ŷ) (19)
Notice that v · r = 0 since v points tangentially/azimuthally while r is
radial. This ensures that the length of r does not change with time. What
is more, we showed that the acceleration a = v̇ = −ω 2 r so that v̇ is
perpendicular to v for uniform circular motion. Thus, the velocity vector
cannot change in magnitude and must also simply rotate! Verify that the
same is true of a as well, for uniform circular motion.

1.6 Integration of kinematical equations

As we will soon learn, if the forces on a particle are known, then one
may use Newton’s second law to find its acceleration. This is called the
dynamical part of the problem of motion, since it depends on the forces and
interactions. The kinematical part of the problem of motion is to determine
the velocity of the particle and its trajectory from its acceleration.
• Suppose we are given the acceleration of a particle. Then the velocity
must satisfy dv
dt = a(t). Integrating this equation with respect to time from
t0 to t, we get Z t
v(t) = v(t0 ) + a(t0 )dt0 . (20)
t0
In addition to knowledge of the acceleration, here we needed an ‘initial
condition’ v(t0 ) (actually three ICs, the three Cartesian components of
v(t0 )) to determine the velocity. The problem of determining velocity has
been reduced to quadratures i.e., to evaluating integrals (one each for the
three Cartesian components of velocity).
• The step from velocity to position involves one more integration and
another initial condition:
Z t
ṙ(t) = v(t) ⇒ r(t) = r(t0 ) + v(t0 )dt0 . (21)
t0

Evaluating these integrals for a specific acceleration may or may not be

feasible analytically.

12
• We have solved the 2nd order ordinary differential equations r̈(t) = a(t)
in two steps. Being 2nd order, the process required two initial condition or
pieces of initial data v(t0 ) and r(t0 ) (each of which is a vector with three
components).
• A simple example is that of uniform acceleration, i.e., where a(t) is a
constant vector a. In this case,
v(t) = v(t0 ) + (t − t0 )a, (22)
and integrating once more,
1
r(t) = r(t0 ) + (t − t0 )v(t0 ) + (t2 − t20 )a − (t − t0 )t0 a. (23)
2
The formula simplifies if t0 = 0:
1
r(t) = r(0) + tv(0) + t2 a (24)
2
This formula applies to the case of constant acceleration.

1.7 Plane polar coordinates

• For many problems, especially those where there is rotational symme-

try around a central object, polar coordinates are more convenient than
Cartesian coordinates.
• For simplicity, we consider polar coordinates (r, θ) on the plane. Sup-
pose we are given an origin O and horizontal and vertical x and y axes.
Given a point P (x, y), r is the distance of P from the origin, and θ is
the counterclockwise angle the radius vector r = (x, y) makes with the
horizontal axis. Note that x and y are called the abscissa and ordinate of
the point P . In other words, cos θ = x/r or tan θ = y/x. Thus,
p
r = x2 + y 2 and θ = arctan(y/x) = arccos(x/r). (25)
Conversely,
x = r cos θ and y = r sin θ. (26)
• Notice that θ is defined modulo 2π . θ = 0 and θ = 2π both corre-
spond to the positive x-axis. One often chooses a convenient ‘fundamental
domain’ for θ such as [0, 2π) or (−π, π].

13
  is important to observe that the polar coordinate system breaks down
It
(or is singular) at the origin where x = y = r = 0. At this point, θ is not
defined. In a sense, the point at the origin could be assigned any value of θ,
depending how we approach the origin. Said differently, the map between
x, y and r, θ fails to be 1-1 at the origin.
• Notice that the constant x and constant y curves (also known as the
level curves1 of x and y ) are mutually orthogonal straight lines parallel to
the y and x axes respectively. By contrast, the constant θ curves are rays
emanating radially outwards from the origin while the constant r curves are
concentric circles centered at O. This explains why polar coordinates are
called curvilinear coordinates. Despite being curvilinear, the level curves
of r and θ are mutually orthogonal.
• A real-valued function on the plane is any function of x and y that
assigns a real number to each point (x, y). The simplest of these functions
are the ‘coordinate functions’ x and y themselves.
• Analogously, r and θ are the coordinate functions in polar coordinates.
• Unit vectors r̂ and θ̂. Recall that at a point (x, y) on the plane, x̂ and
ŷ are unit vectors in the directions of increasing x holding y fixed and vice
versa. Similarly, we define the unit vectors r̂ and θ̂ at any point (r, θ).
r̂ points radially outwards while θ̂ points counterclockwise tangentially
to the circle of radius r. The direction in which θ̂ points is called the
azimuthal direction.
• x̂, ŷ furnish one basis for vectors at any point on the plane. Similarly,
r̂, θ̂ furnish another basis at points away from r = 0. We can therefore
expand r̂ and θ̂ in the x̂, ŷ basis.
1
A level curve of a quantity is a curve on which the quantity is a constant.

14
• A figure shows that we may decompose r̂ and θ̂ as
x y
r̂ = cos θ x̂ + sin θ ŷ = x̂ + ŷ
r yr x
and θ̂ = − sin θ x̂ + cos θ ŷ = − x̂ + ŷ. (27)
r r

• Check that r̂ and θ̂ are orthonormal:

r̂ · θ̂ = 0 and r̂ · r̂ = θ̂ · θ̂ = 1. (28)

• Exercise: Express x̂ and ŷ as linear combinations of r̂ and θ̂.

• Unlike x̂ and ŷ which point in the same direction everywhere, the direc-
tions of r̂ and θ̂ change with location.

• Position coordinate and velocity vector. The position vector can

now be expressed in polar coordinates.
r = xx̂ + y ŷ = r cos θx̂ + r sin θŷ = rr̂. (29)
We wish to find the velocity and acceleration vectors in plane polar co-
ordinates. These are the polar coordinate analogues of v = ẋx̂ + ẏ ŷ and
a = ẍx̂ + ÿ ŷ .
• Now, suppose r(t) = r(t)r̂(t) is the position of a particle at time t. As
it moves along a trajectory, the radial coordinate r can change, but so can

15
the unit vector r̂(t). Thus, its velocity is given by
dr̂
v(t) = ṙ(t) = ṙ r̂ + r . (30)
dt
ṙr̂ is what we might naively guess as the radial velocity. The other term
comes from the change in direction of the basis vector r̂.
• Let us take a moment to find the rates of change of the basis unit vectors
r̂ and θ̂. Being unit vectors, their change can come only from a change in
their direction. For instance,
dr̂
r̂ = cos θx̂ + sin θŷ ⇒ = − sin θ θ̇ x̂ + cos θ θ̇ ŷ = θ̇θ̂, (31)
dt
where we recalled that θ̂ = − sin θx̂ + cos θŷ .
• Thus, the change in r̂ is always in the azimuthal θ̂ direction. Neither
the magnitude nor the direction of r̂ changes in the radial direction. Un-
derstand this through the figure

• Similarly,
dθ̂
θ̂ = − sin θx̂ + cos θŷ ⇒ = − cos θ θ̇ x̂ − sin θ θ̇ ŷ = −θ̇ r̂. (32)
dt
The rate of change of θ̂ always points radially.

• Putting these together, we get the decomposition of the velocity of the

particle in the polar coordinate basis:
d(rr̂) dr̂
v = ṙ(t) = = ṙ r̂ + r = ṙ r̂ + r θ̇ θ̂. (33)
dt dt

16
• We say that v · r̂ = ṙ is the radial velocity while v · θ̂ = rθ̇ is the angular
or azimuthal component of velocity.
• For circular motion at a constant radius `, ṙ = 0 and v = `θ̇θ̂ is purely
azimuthal. The speed of the particle is v = `|θ̇| while its angular speed is
v/` = |θ̇|.
• Acceleration in polar coordinates. We may differentiate the velocity
to obtain the acceleration in polar coordinates.
d dr̂ dθ̂
a = v̇ = (ṙ r̂ + r θ̇ θ̂) = r̈ r̂ + ṙ + ṙθ̇θ̂ + rθ̈θ̂ + rθ̇ (34)
dt dt dt
˙
Now we use r̂˙ = θ̇θ̂ and θ̂ = −θ̇r̂ to combine the first and last terms ∝ r̂
and the remaining terms ∝ θ̂:
a = (r̈ − rθ̇2 )r̂ + (rθ̈ + 2ṙθ̇)θ̂. (35)

• The various terms in a have physical interpretations.

• r̈r̂ is the radial acceleration.
• −rθ̇2 r̂ is called the centripetal acceleration, which we met in the context
of uniform circular motion. It always points radially inwards.
• rθ̈θ̂ is the angular acceleration that comes from the varying angular
speed θ̇ of the particle.
• The last term 2ṙθ̇θ̂ is called the Coriolis acceleration, it is present when-
ever both r and θ change with time.

1.8 Spherical polar coordinates

• The analogue of plane polar coordinates in 3d (R3 ) are called spherical

polar coordinates (r, θ, φ). They are called the radial, polar and azimuthal
coordinate respectively. They are particularly useful in dealing with sys-
tems where there is spherical symmetry about a central object such a the
Sun in the solar system of the nucleus in an atom.
• Given a point P with Cartesian coordinates (x,
p y, z), the radial coordi-
nate r is the distance of P from the origin r = x2 + y 2 + z 2 . Evidently,
0 ≤ r < ∞.

17
• If P has position vector r relative to the origin, then the polar angle θ
is the angle r makes with respect to the upward vertical z axis. Thus θ =
arccos(z/r). Notice that 0 ≤ θ ≤ π with θ = 0 and θ = π corresponding
to the positive and negative z axis.

• Finally, suppose we orthogonally project the position vector onto the

x-y plane. This projected vector has length r sin θ = x2 + y 2 .
• The azimuthal angle φ is defined as the angle that this projection
p makes
with the x axis, measured counterclockwise. Thus, φ = arccos(x/ x2 + y 2 )
or φ = arctan(y/x). Notice that 0 ≤ φ < 2π .
• φ is the azimuthal angle for plane polar coordinates on the x-y plane
(it was called θ in that discussion!)
• Unfortunately, the spherical polar coordinate system breaks down along
the z axis, where φ is not uniquely defined (it can be assigned any value
0 ≤ φ < 2π ). [When a coordinate system does not cover the whole of R3 ,
we could introduce another set of coordinates that work in the excluded
region.]
• Alternatively, we may write

z = r cos θ, x = r sin θ cos φ and y = r sin θ sin φ. (36)

p
Check that tan θ = x2 + y 2 /z .
• The position vector of a particle located at (x, y, z) is then given by
r = xx̂ + y ŷ + z ẑ = rr̂.
• The other formulae we derived for plane polar coordinates may be gen-
eralized to spherical polar coordinates.
• For instance, if r̂, θ̂, φ̂ are the unit vectors in the directions of increasing

18
r, θ, φ, then the figure helps us express

r̂ = cos θẑ + sin θ(cos φx̂ + sin φŷ),

θ̂ = − sin θẑ + cos θ(cos φx̂ + sin φŷ) and
φ̂ = − sin φx̂ + cos φŷ. (37)

• Verify that (r̂, θ̂, φ̂) is a right-handed orthonormal system.

1.9 Taylor approximation

• Taylor series for one variable. Given a function of one real variable
f (x) that is continuous and hopefully differentiable a few times, we are
interested in approximately evaluating it in the neighborhood of a point
x0 .
• By continuity, f (x) ≈ f (x0 ) is of course our zeroth order approximation
to the value of the function for x near x0 .
• The next possibility is to approximate f by a linear function near x0 .
It is natural to take the slope of this linear function to be the derivative of
f at x0 (assuming f is differentiable at x0 ), so that we approximate the
graph of f by the tangent through the point (x0 , f (x0 )). This leads to the
first order or linear Taylor approximation

f (x) ≈ f (x0 ) + f 0 (x0 )(x − x0 ). (38)

19
• To indicate that x − x0 is small, we will denote it by ∆x = x − x0 , and
denote f (x) − f (x0 ) = ∆f . Then we have ∆f ≈ f 0 (x0 )∆x. This is only
an approximation
• It is also convenient to introduce the differential of f , which at x is
defined as df (x) = f 0 (x)dx. dx is called the differential of x. The deriva-
tive denoted df /dx is the limit of ∆f /∆x as ∆x → 0. For example,
d sin x = cos x dx. The differential of a function is also called a 1-form.
• More generally, if f is n times differentiable at x0 , we have the nth order
Taylor polynomial approximation for small x − x0 :
1 1
f (x) ≈ f (x0 ) + f 0 (x0 )(x − x0 ) + f 00 (x0 )(x − x0 )2 + f 000 (x0 )(x − x0 )3
2 3!
1 (n) n
+ · · · + f (x0 )(x − x0 ) , (39)
n!
where f (n) (x0 ) is the nth derivative of f at x0 .
• For many of the functions we encounter, the Taylor series, obtained by
letting n → ∞, converges to the function f (x) for x in a neighborhood of
x0 . Such functions are called real analytic.
• A real-valued function that is continuous in some domain is said to be of
type C 0 in that domain. A function that is differentiable with continuous
first derivative is said to be of class C 1 in that domain. Similarly we
have the notion of C k functions for k = 1, 2, 3, . . .: k times continuously
differentiable functions in some domain. A function that is C k for all
k = 1, 2, 3, 4, . . . is said to be smooth or C ∞ . A function whose Taylor
series converges to the function in some domain is said to be real analytic
or of type C ω .
• For example, show that the Taylor series for 1/(1 − x) around x = 0 is
given by a geometric series:
1
= 1 + x + x2 + · · · . (40)
1−x
This series converges to 1/(1 − x) for |x| < 1. Also verify that (1 − x)(1 +
x + x2 + · · · ) = 1 by multiplying things out and canceling. Note that
1/(1 − x) does not admit a Taylor expansion around x = 1.

20
• Show that the Taylor series for ex , is given by

x x2 xn
e =1+x+ + ··· + + ··· . (41)
2! n!
This series has an infinite radius of convergence. Find the Taylor series for
sin x and cos x.
• The binomial series is a very useful Taylor series around x = 0:
ν(ν − 1) 2 ν(ν − 1)(ν − 2) 3
(1 + x)ν = 1 + νx + x + x + ··· (42)
1·2 1·2·3
which converges for |x| < 1 and any (real or complex) number ν . For a
positive integer ν = n, this series terminates and we recover the binomial
n
expansion
Pn n r
with coefficients given by combinatorial factors: (1 + x) =
r=0 r x .
• In particular, show that
1 x x2
√ =1− + + ··· . (43)
1+x 2 8

• Taylor series for more variables. For a real function f (x, y) of two
variables, we have the Taylor expansion of f around a point (x0 , y0 ):
∂f ∂f
f (x, y) = f (x0 , y0 ) + |(x0 ,y0 ) (x − x0 ) + |(x ,y ) (y − y0 )
 2 ∂x ∂y 0 0
∂2f ∂2f ∂2f

1 ∂ f 2 2
+ (x − x 0 ) + (y − y 0 ) + (x − x 0 )(y − y 0 ) + (y − y0 )(x − x 0 ) ···
+ (44)
2 ∂x2 ∂y 2 ∂x∂y ∂y∂x

where all the partial derivatives are evaluated at (x0 , y0 ). The mixed sec-
∂2f ∂2f
ond partials ∂x∂y and ∂y∂x are equal (Clairaut’s or Schwarz’s Theorem,
assuming the second partials are continuous).
• To calculate a partial derivative with respect to y we simply differentiate
the function with respect to y treating x as fixed.
• One way to obtain this series is to treat y as fixed and first write down
a Taylor series in x around x0 with coefficients being functions of y . Then
we expand these coefficients in a Taylor series in y .
• Calculate the mixed second partials of f (x, y) = cos xy and show that
they are both equal to − sin xy − xy cos xy

21
1.10 Some vector calculus

• Scalar fields. At a given instant of time, the pressure p(r) in the

atmosphere is a real number that depends on height, and more generally on
location r . The density ρ(r) of sea water at a given instant of time depends
on depth. Similarly, the salt concentration of sea water c(r) depends on
location. The potential energy V (r) of a massive particle in Earth’s gravity
depends on height above the Earth’s surface as well as the latitude and
longitude. All these are examples of real valued functions in 3d space. We
will also refer to real-valued functions as scalar fields. A scalar field assigns
a real number to each location r . Typically, the real number would vary
smoothly (or at least continuously differentiably) as the location changes.
• Note that the notion of a field introduced here is different from the
algebraic notion of a field (e.g., field of real or complex numbers). Here,
field refers to something depends on location.
• Vector fields. Similarly, we have the concept of a vector field: a
smoothly varying vector v(r) at each location r . The gravitational force
felt by a point mass m at various locations and heights above the Earth’s
surface defines a vector field.

• In figure, we have displayed three vector fields on the plane. Since such
a vector field has two components in Cartesian coordinates, a vector field
on the plane may be regarded as a map from R2 → R2 . If x and y are
the horizontal and vertical directions, then the first vector field is plausibly
v ∝ x̂. The second vector field points radially outwards with a magnitude
increasing with radial distance and is circularly symmetric. It is plausible
that the 2nd vector field v ∝ xx̂ + y ŷ = r . The 3rd vector field could be
the velocity vector field of a steadily flowing fluid.
• Gradient of a scalar field. Given a scalar field φ(r), its gradient
is a kind of derivative that produces a vector field denoted ∇φ(r). In

22
Cartesian coordinates r = (x, y, z),
∂φ ∂φ ∂φ
grad φ = ∇φ(r) = x̂ + ŷ + ẑ. (45)
∂x ∂y ∂z
• Example 1: If φ(x, y) = x then ∇φ = x̂ is a constant vector field
pointing in the x direction at all points of R2 .
• Example 2: If φ = 21 (x2 + y 2 + z 2 ), then
∇φ = xx̂ + y ŷ + z ẑ = rr̂ (46)
is a radially outward pointing vector field on R3 , with magnitude equal to
the distance from the origin.
• At any location r , ∇φ is a vector that points in the direction of most
rapid increase of φ. For a vector field v = ∇φ on the plane, ∇φ must
point orthogonally to the level curves (or equipotentials if φ is regarded as
a potential function) of φ, since φ is constant along its level curves. For
φ(x, y) = x, the level curves are lines parallel to the y axis, and ∇φ = x̂
points perpendicular to these lines. For a vector field on R3 , v = ∇φ must
point perpendicular to the level surfaces of φ. For φ = 12 (x2 + y 2 + z 2 ), the
level surfaces are concentric spheres centered at the origin and ∇φ = r is
perpendicular to these surfaces.
• Line integral. Given a vector field v(r) = (vx , vy , vz )(r) in 3d space
and a parametrized curve γ(t) = (x(t), y(t), z(t)) for 0 ≤ t ≤ 1, we may
define the ‘line integral’ of v along γ as the real number
Z Z 1 Z 1 
dγ dx dy dz
v · dγ = v· dt = vx + vy + vz dt. (47)
γ 0 dt 0 dt dt dt

• Here, γ̇ = dγ
dt = ẋx̂ + ẏ ŷ + ż ẑ is a vector field along the curve γ (it is
not defined elsewhere in R3 ).
• For example, if γ is the helix (cos t, sin t, t), then dγ = (− sin t, cos t, 1)dt.
We may consider dγ as the differential of the map γ : [0, 1] → R3 .

23
• The work done by a force field F (r) in moving a particle
R along a curve
γ is an important example of a line integral: WF (γ) = γ F · dγ .
• In general, the line integral depends on the values of v all along the
curve γ . However, if v is the gradient of a scalar, v = ∇φ, then the line
integral can be evaluated in terms of the values of φ at the endpoints:
Z Z 1 
∂φ dx ∂φ dy ∂φ dz
∇φ · dγ = + + dt
γ 0 ∂x dt ∂y dt ∂z dt
Z 1
dφ(r(t))
= dt = φ(r(1)) − φ(r(0)). (48)
0 dt
Here, we viewed φ(x(t), y(t), z(t)) as a function of t and used the chain
rule to differentiate it wite respect to t.
• In particular, if γ is a closed curve, then r(0) = r(1) and the line integral
of a gradient vanishes I
∇φ · dγ = 0. (49)
γ
H
Here denotes a line integral around a closed contour.

• A vector field that is the gradient of a scalar field is called a gradient

vector field. In mechanics, if a force field F (r) is the gradient of a scalar
field (or ‘potential’ φ(r)), then it is called a conservative force field. The
work done by a conservative force field ∇φ depends only on the initial and
final locations of the particle, and not on the rest of the details of the path
taken. A conservative force field does no work in moving a particle around
a closed curve.

24
2 Newton’s laws and forces

2.1 Time, light, simultaneity, space & time intervals, masses.

• To describe the dynamics (motion or more precisely the evolution in

time) of mechanical systems, observers find it helpful to have a notion of
time (measured with a clock) to index a sequence of events.
• In Newtonian mechanics, one assumes that if there is a flash of light
somewhere, then all observers (irrespective of their locations) receive the
flash instantaneously. In effect, light is assumed to travel infinitely fast.
• Using such flashes of light, all observers can synchronize their clocks and
assign the same time for a given event.
• Another consequence is that two events (possibly at different locations)
that occur at the same time for one observer occur simultaneously for any
other observer.
• These assumptions about light, time and simultaneity were in line with
common human experience in Newton’s time (as well as today!).
• In Newtonian mechanics, one also assumes that masses of particles, scales
of length and time are the same for all observers. In other words, distances
(like the length of a meter stick) and time intervals (like that between two
ticks of a clock) are the same for all observers.
• Causality. Our experience with physical systems indicates that they
respect the principle of causality: cause precedes effect. For example, a
stone that is stationary is seen to move when it is pushed and not before
that.
• The principle of causality postulates that it is not possible to send a
signal from an event to its past. Given our Newtonian concept of time, all
observers have a common notion of the past and future of an event. The
future of an event that occurs at time t0 consists of all events that occur
at t > t0 and the past consists of events that occurred at t < t0 .
• These notions of time, simultaneity and universality of masses, space
and time intervals had to be discarded and replaced with more accurate
concepts in the special theory of relativity, where speeds of bodies or ob-

25
servers could be comparable to that of light, which is a large but finite
constant in vacuum (c ≈ 3 × 108 m/s).
• Newtonian or nonrelativistic mechanics is a limiting case of special rela-
tivistic mechanics where the speed of light is infinite (very large compared
to other speeds). The principle of causality continues to apply in special
relativity, though the notions of past and future need to be revised.

2.2 Degrees of freedom, instantaneous configurations, trajectories

• A point particle moving in three-dimensional space has three degrees of

freedom: we need three coordinates (say (x, y, z) or (r, θ, φ)) to specify the
location of the particle at the initial instant of time. x, y and z can be
chosen arbitrarily at the initial instant of time.
• For a system of particles, the number of degrees of freedom is the number
of real parameters (coordinates) needed to specify the locations of all the
particles in the system at the initial instant of time.
• The number of degrees of freedom does not depend on the nature of
forces. So a free particle and a particle subject to Earth’s gravity moving
in three-dimensional space both have three degrees of freedom.
• On the other hand, a point particle that is constrained to move along a
fixed wire has only one degree of freedom. We need one coordinate, say the
distance (arc length) from one end of the wire, measured along the wire to
specify the location of the particle at a given instant of time.

• A particle constrained to move on a spherical surface (such as a bob at

the end of a rigid rod whose other end is attached to a pivot) has only two
degrees of freedom.

26
• Two point masses moving in three dimensional space have six degrees
of freedom. We need six coordinates (x1 , y2 , z1 ) and (x2 , y2 , z2 ) to specify
the locations of the two particles. E.g., the Sun and the Earth regarded
as point masses is a system with 6 degrees of freedom. Here, we do not
restrict to a particular orbit of the Earth around the Sun but ask how many
coordinates are needed to specify all possible locations of the Sun and the
Earth at any fixed instant of time, without reference to the nature of the
force between the two.
• A general rigid body like a stone has six degrees of freedom. For con-
venience, we may enumerate them as follows: 3 translational degrees of
freedom to fix the location of a marked point in the body and 3 rotational
degrees of freedom to orient the body holding the marked point fixed.
• A fluid consisting of N ∼ 1024 molecules in a bucket has a very large
number of degrees of freedom, which can be taken to be the 3N Carte-
sian coordinates needed to specify the instantaneous locations of the N
molecules, treated as point masses.
• An instantaneous configuration of a system of two point particles is any
possible location of the two particles.
• Zeroth law of classical mechanics. The path followed by a particle in
time is called its trajectory. It is a curve parametrized by time and directed
towards increasing time. The zeroth law of mechanics can be regarded
as saying that the trajectory r(t) of a particle is a (twice) differentiable
function of time.
• This not an assumption but rather an assertion about natural phenom-
ena, deduced by observing the motion of terrestrial and celestial bodies.
This assertion applies to the motion of planets, pendulum bobs, cricket
balls etc. But it fails for Brownian motion (movement of pollen grains in
water, which are observed to follow very jagged paths). It also fails for
electrons in an atom, which require a quantum mechanical treatment.
• Isaac Newton formulated three laws of classical mechanics in his Principia
(1687).

27
2.3 Newton’s 1st law

• Newton’s 1st law, or the law of inertia, says that “Every body continues in
its state of rest, or of uniform motion in a straight line, unless it is compelled
to change that state by a force impressed upon it”. In other words, the
momentum p = mv = mṙ of a particle that is free (isolated or far from
physical interactions) does not change with time. Note that if ṙ does not
change with time, i.e., r̈ = 0, then the trajectory r(t) = r(0) + ṙ(0)t is a
straight line that is uniformly traversed.
• In general, it is found that macroscopic interactions decrease with dis-
tance, so it is possible to isolate a particle by taking it far from other
bodies.
• We have been a bit imprecise in our statement of Newton’s first law.
Newton’s first law holds only in reference frames. To specify the compo-
nents of the position and velocity vectors, we need a frame of reference,
i.e., an origin and coordinate axes.
• It is found that a particle that is not subject to any forces (i.e., an
isolated body) could fail to follow a constant velocity trajectory in certain
reference frames.
• A frame in which Newton’s first law holds is said to be an inertial frame.
To a reasonable approximation (if one ignores some effects of the rotation
of the Earth), a frame that is fixed in a tennis court is an inertial frame
for the motion of tennis balls, racquets etc. In particular, if the effects of
gravity and friction are ignored (or somehow cancelled), then tennis balls
in this frame would always move uniformly in straight lines.
• However, it is found that Newton’s first law for tennis balls fails to hold
in a frame that is attached to a swinging pendulum or a rotating merry-
go-round beside the tennis court. Such frames are called accelerated or
noninertial.
• For instance, a frame that is attached to a bee as it flies irregularly in a
faraway spaceship is not inertial, since a free particle at rest in the same
spaceship would appear to move in a nonuniform manner.
• Note: here, we use the metaphor of the bee for the limited purpose of

28
defining a frame that moves nonuniformly relatively to the spaceship. A
flying bee is not a free particle - it does not move uniformly, it makes use
of its internal energy and friction with the air to change direction, speed
up or maintain its speed etc.
• Similarly, a frame that is attached to a top (spinning on the floor of the
spaceship) and participates in its rotational motion is noninertial.
• To summarize, Newton’s first law is the assertion that there is a frame
of reference (called an inertial frame) in which all isolated bodies (far from
physical interactions) move at constant velocity.
• As we will see shortly, from a principle enunciated by Galileo, an inertial
frame is not unique.
• Henceforth, unless otherwise stated, all quantities will be specified with
respect to an inertial frame of reference.

29