Physics 1.

Mechanics
Math0. Useful mathematical formulae

Here you will find a (limited) number of useful mathematical formulae which you must know in
order to understand physics.
• Let f = f (x) and g = g(x) and f 0 = df /dx, . . . . Then

(f + g)0 = f 0 + g 0 , (f g)0 = f 0 g + f g 0

• Let f = f (y), and y = y(x), then

  
df df dy
=
dx dy dx

• Following two previous rules  0

f gf 0 − f g 0
=
g g2

• Differential:  
df
df = dx
dx

• Taylor expansion
   
df 1 2 d2f
f (x) = f (x0 ) + (x − x0 ) + 2
(x − x0 ) + ...
dx |x=x0 dx2 |x=x0

or for the infinite series

∞  n 
X 1 n d f
f (x) = f (x0 ) + (x − x0 )
i=1
n! dxn |x=x
0

• Useful approximations for |x|  1

sinx ≈ x, cosx ≈ 1 − x2 /2, ex ≈ 1 + x. ln(1 + x) ≈ x, (1 + x)α ≈ 1 + αx

• Let f = f (x, y). Partial derivative ∂f /∂x is the derivative with respect to x while y is assumed
constant.

1
Physics 1. Mechanics Math0

• Total differential    
∂f ∂f
df = dx + dy
∂x ∂y

• First term of the Taylor expansion

   
∂f ∂f
f (x, y) = f (x0 , y0 ) + (x − x0 ) + (y − y0 )
∂x |x=x0 ,y=y0 ∂y |x=x0 ,y=y0

• Features of mixed derivatives for normal functions f x, y:

∂2f ∂2f
=
∂x∂y ∂y∂x

• Indefinite integral Z
f 0 dx = f + C

• Definite integral Z x2
f 0 dx = f (x2 ) − f (x1 )
x1

• Substitution. Let f = f (y) and y = y(x) then

Z Z
dy
f (y)dy = f (y(x)) dx
dx

• Integration by parts Z Z
0
f g dx = f g − gf 0 dx

2
 Physics 1. Mechanics
Lecture 0. Partial derivatives

1 Functions
Let x be some variable (like time), and for each x we measure another variable y which depends on x (like
a body temperature). The dependence of y on x is described by a function y = y(x). It is essential that x
be continuous, and that for each x from some interval x1 < x < x2 there exist a corresponding y.

2 Derivative
Let the argument x change from x1 to x2 , ∆x ≡ x2 − x1 . The function change is ∆y = y(x2 ) − y(x1 ). The
ratio ∆y/∆x shows how fast the function changes on average when the argument changes on the interval
(x1 , x2 ). Alternatively, let us choose a point x and shift from this point ∆x. The above ratio then will be
written as
∆y y(x + ∆x) − y(x)
=
∆x ∆x
as depends on x and ∆x as well. If we now require ∆x → 0 (infinitesimal change) then the ratio (if it exists
at all) will show the local rate of change of the function y in the point x,

dy ∆y
y0 ≡ = lim , (1)
dx ∆x→0 ∆x

and is called a derivative. In this notation dy/dx is not a ratio but a new function. However, it is obtained
by calculating the ratio of two infinitesimal values.
Starting with the obvious expression  
∆y
∆y = ∆x
∆x
we can extrapolate this onto infinitesimal changes and obtain a differential :
 
0 dy
dy = y dx = dx, (2)
dx

which is the change of y when x changes by infinitesimal dx → 0 (but dx 6= 0 !).

Example: Let y(x) = xn , then (dy/dx) = nxn−1 , dy = nxn−1 dx.
Example: The function y = |x| does not have a derivative at x = 0.

1
Physics 1. Mechanics Lecture 0

3 Partial derivatives
Let now y be a function of more than one variable xi , i = 1, 2, . . . , n (here n is the number of independent
variables). We define a partial derivative (∂y/∂xl ) of y with respect to the variable xl as a regular derivative,
taken as if y were function of xl only while other variables remain constant. In this case dy is produced by
the change of each xi independently, so that one has:
    X  ∂y 
∂y ∂y
dy = dx1 + dx2 + . . . = dxi . (3)
∂x1 ∂x2 ∂xi
i

Example: Let y = x21 + x1 x2 , then (∂y/∂x1 ) = 2x1 + x2 , (∂y/∂x2 ) = x1 , dy = (2x1 + x2 )dx1 + x1 dx2 .
Q: Partially differentiate and write down the full differential for the following functions:

y = x21 + 2x1 x22 + x52 , (4)

y= exp(−x21 − x22 ) · [sin x1 + 2 cos(2x1 x2 )]. (5)

Each partial derivative is itself a function.

4 Some applications
4.1 Differentiation along a curve
Let y = y(x1 , x2 , . . .), and each argument is, in turn, a function of a single argument t: xi = xi (t). Eventually,
y is a function of t too, and one may ask what is the derivative (dy/dt). One way is to substitute all xi (t)
into the expression for y and, after an explicit form of y(t) is obtained, differentiate it as usual. This is not
always convenient. Instead, let us write the full differential dy in two forms:

dy
dy = dt, (6)
dt
X ∂y 

dy = dxi (7)
∂xi
i

However,  
dxi
dxi = dt (8)
dt
and therefore
  X  ∂y 
dy
dt = dxi (9)
dt ∂xi
i
X  ∂y   dxi 
=[ ]dt ⇒ (10)
∂xi dt
i
  X  
dy ∂y dxi
= (11)
dt ∂xi dt
i

2
Physics 1. Mechanics Lecture 0

4.2 Implicit function

Sometimes it is impossible to give an explicit form of a function y = y(x1 , x2 , . . .). Instead, and implicit
function is given in the form of an equation F (y, x1 , x2 , . . .) = 0, which is difficult or impossible to solve
with respect to y. In order to find derivatives of y we use again the full differential:
  X  ∂F 
∂F
dF = 0 = dy + dxi ⇒ (12)
∂y ∂xi
i
X  ∂F  
∂F

dy = −[ dxi ]/ ⇒ (13)
∂xi ∂y
i
   
∂y ∂F ∂F
=− / (14)
∂xi ∂xi ∂y

5 Taylor expansion
We use the full differential to approximate a function variation:
   
∆y dy
∆y = ∆x ≈ ∆x (15)
∆x dx

or, in a more appropriate form,

 
dy
y(x) = y(x0 ) + |x=x0 (x − x0 ) + . . . (16)
dx

Here |x=x0 means that the derivative is evaluated in the point x = x0 , and . . . means that the expression is
not exact and (hopefully) smaller terms should be added.
For a function of many variables one has, respectively

X  ∂y 
y(xi ) = y(xi,0 ) + |xi =xi,0 (xi − xi,0 ) (17)
∂xi
i

3
 Physics 1. Mechanics
Lecture 1. Coordinates

1 Coordinates. General principles

We start with the description of the motion of a point mass (point or particle). First thing to do is
to learn how to describe the position of the point. In order to do that one needs coordinates. In the
simplest case of a motion along a line (not necessarily straight line) one needs to choose an origin O
and a method of assigning a coordinate, that is, a rule to establish a correspondence of each point of
the curve to a real number (positive of negative). In order to be useful the coordinate should change
continuously and, if possible, different points have to have different coordinates and each point has
to have only one coordinate.

Figure 1: Coordinate on a curve: one-dimensional space.

1.1 Examples
1.1.1 Straight line

Any point can be taken as the origin O. Let us choose the positive direction to the right (see Fig. 2),
and assign to each point a number corresponding to the distance from O with the sign + if the point
is to the right of O, and − if it is to the left. Let us denote this coordinate as x, then −∞ < x < ∞.
Q: Is this choice unique ? Give an example of another choice. Try to find a coordinate x′ which
would change in the range −1 ≤ x′ ≤ 1 and yet cover the whole line.

1
 O x
Figure 2: Coordinates on a straight line

ϕ
O

Figure 3: Coordinates on a circle.

1.1.2 Circle

A simple choice of the coordinate would be the angle ϕ measured counterclockwise from some radius
(see Fig. 3). In order to make this coordinate unique we have to require, for example, 0 ≤ ϕ < 2π.
However, in this case the coordinate is not continuous at ϕ = 0. On the other hand, in many
applications it is more desirable to have the coordinate continuous. To do that we may say that the
coordinates ϕ and ϕ + 2πn (where n is an arbitrary integer) describe the same point.

1.1.3 Closed curve with no intersections

Let us consider a simple closed curve which does not intersect itself. A simple coordinate choice is
the angle ϕ measured counterclockwise from some straight line intersecting the curve. In some cases
this coordinate may be ambiguous, and a better coordinate would be the curve length measured from
one chosen point in a chosen direction. In this case the coordinate is always positive.

2 Proceeding further
In the above examples we needed only one coordinate to describe a point position. In this case
it is said that the space is one-dimensional. If two-coordinates are necessary, the space is called
two-dimensional and if n coordinates are needed, the space is called n-dimensional of nD. The space
we live in is three-dimensional, that is, in general three coordinates are needed to describe a point
position. In what follows we shall denote a point as P and its coordinates as (xi ), where i = 1, . . . , n.

2
 θ
O

Figure 4: Coordinates on a closed curve with not self-intersections.

x2

x1

Figure 5: General 2D coordinates.

2.1 Examples of 2D coordinates

In the 1D case we had a single curve. In the 2D case we build two family of “parallel” curves (see
Fig. 5), one family for each coordinate. “Parallel” means that the curves from the same family do
not intersect. For each curve from one family the second coordinate is constant.

2.1.1 Cartesian coordinates

The two families are straight lines intersecting and 90◦ . The coordinates (x, y) are chosen along each
line as in the 1D case of a straight line (distance with sign). The curves of x = const are straight
lines parallel to y-axis. The curves of y = const are straight lines parallel to x-axis.

2.1.2 Polar coordinates

One family is the radii coming out of the origin, the other is the circles with the origin as a center.
One coordinate, r, is a distance from the origin (nonnegative), the second is the angle ϕ as in the

3
 P (x, y)
y

Figure 6: 2D Cartesian coordinates.

P (r, ϕ)

Figure 7: Polar coordinates.

case of a circle (see above). The curves of r = const are circles. The curves of ϕ = const are radii.

2.2 3D examples
For 3D case xi = const gives a surface.

2.2.1 3D Cartesian coordinates

Similar to the 2D case. x = x0 = const corresponds to a plane (2D space !) parallel to the y − z
plane and crossing the x axis at x = x0 .

2.2.2 Cylindrical coordinates

The coordinates (r, ϕ, z) are the polar coordinates (r, ϕ) with the addition of the “height” z. r = const
are cylindrical surfaces around z-axis. z = const are planes parallel to x − y plane.
Q: What are ϕ = const surfaces ?

4
 z

P (x, y, z)

Figure 8: 3D Cartesian coordinates.

P (r, ϕ, z)

ϕ
r

Figure 9: Cylindrical coordinates.

5
 P (R, θ, ϕ)

Figure 10: Spherical coordinates.

2.2.3 Spherical coordinates

The coordinates (R, θ, ϕ) are the distance from the origin, the angle measured downward from z-axis,
and the polar angle ϕ measured in the projection on the plane x − y counterclockwise from x-axis.
Q: What are R = const, θ = const, and ϕ = const surfaces, respectively ? What are the
dimensions of the (sub)space r = const and ϕ = const ? What is this space ?

2.3 Relation between different coordinate systems

The same point can be assigned coordinates with the help of different coordinate systems, and it
is necessary to know the relation between those coordinates. In the following we assume that the
coordinate origin is the same for all chosen systems.

2.3.1 2D Cartesian - polar

The two systems are shown in the figure. It is accepted to choose x-axis as the base of the polar
coordinates. The following relation holds:

x = r cos ϕ,
(1)
y = r sin ϕ.

2.3.2 3D

The three coordinate systems are shown in the same figure. One has

6
 P
y

ϕ
x

Figure 11: 2D Cartesian - relation.

P
R

θ y

x ϕ
r

Figure 12: 3D relation.

7
 x = r cos ϕ = R sin θ cos ϕ,
y = r sin ϕ = R sin θ sin ϕ, (2)
z = z = R cos θ.

2.4 Distance element

Let P (xi ) and P ′(xi + dxi ), i = 1, 2, 3, be two (infinitesimally) close points. We shall denote the
distance between these points as ds and call it distance element. If the coordinates are cartesian, the
distance element is given by
ds2 = dx2 + dy 2 + dz 2 , (3)

and such space is called Euclidean. The expression for the distance element will be different in
different coordinate systems. As an example we derive the distance element for spherical coordinates.
Let P (r, θ, ϕ) and P ′(r + dr, θ + dθ, ϕ + dϕ) be to close points. If we knew dx, dy, dz in the cartesian
coordinates the distance element would be obtained from (3). However, in this case (r, θ, ϕ) are
independent variables, while (x, y, z) are functions of these variables. Thus, we have to derive the
full differentials, according to the rule (??) and using the relations (2):
     
∂x ∂x ∂x
dx = dR + dθ + dϕ
∂R ∂θ ∂ϕ (4)
= sin θ cos ϕdR + R cos θ cos ϕdθ − R sin θ sin ϕdϕ,

and similarly for dy and dz. After substitution into (3) one gets

ds2 = dR2 + R2 dθ2 + R2 sin2 θdϕ2 . (5)

Q: Derive ds2 in cylindrical coordinates.

3 Curvilinear orthogonal coordinates

In all cases above the distance element had the form
X
ds2 = h2i dx2i (6)
i=1,2,3

Indeed, let us take the cylindrical coordinates with x1 = r, x2 = ϕ, and x3 = z. One has

ds2 = dr 2 + r 2 dϕ2 + dz 2 (7)

that is,
h1 ≡ hr = 1, h2 = hϕ = r, h3 ≡ hz = 1 (8)

8
For spherical coordinates x1 = R, x2 = θ, x3 = ϕ one has

ds2 = dR2 + R2 dθ2 + R2 sin θ2 dϕ2 (9)

and
h1 ≡ hR = 1, h2 ≡ hθ = R, h3 ≡ hϕ = R sin θ (10)

The coordinates where the distance element takes the form (6) are called orthogonal coordinates.
Cartesian coordinates with h1 = h2 = h3 = 1 are the special case of orthogonal coordinates. If at
least one of hi 6= 1 then the coordinates are called curvilinear coordinates.
Let us assume that we have previously introduced Cartesian coordinates (x1 , x2 , x3 ) = (x, y, z)
but it would be more convenient to use some other orthogonal coordinates (X1 , X2 , X3 ). We have to
derive the distance element with the use of these new coordinates. In order to do that we start with

ds2 = dx2 + dy 2 + dz 2 (11)

Unless we are given the relation between the new and old coordinates we cannot do anything further.
Thus, we assume that the functions

x = x(X1 , X2 , X3 ), y = y(X1 , X2 , X3 ), z = z(X1 , X2 , X3 ) (12)

are given. If so, we have

∂x ∂x ∂x
dx = dX1 + dX2 + dX3 , (13)
∂X1 ∂X2 ∂X3
∂y ∂y ∂y
dy = dX1 + dX2 + dX3 , (14)
∂X1 ∂X2 ∂X3
∂z ∂z ∂z
dx = dX1 + dX2 + dX3 , (15)
∂X1 ∂X2 ∂X3

9
and substituting into (11) one has
 
2 ∂x 2 ∂y 2 ∂z 2
ds = ( ) +( ) +( ) dX12
∂X1 ∂X1 ∂X1
 
∂x 2 ∂y 2 ∂z 2
+ ( ) +( ) +( ) dX22
∂X2 ∂X2 ∂X2
 
∂x 2 ∂y 2 ∂z 2
+ ( ) +( ) +( ) dX32
∂X3 ∂X3 ∂X3
 
∂x ∂x ∂y ∂y ∂z ∂z
+2 ( )( )+( )( )+( )( ) dX1 dX2
∂X1 ∂X2 ∂X1 ∂X2 ∂X1 ∂X2
 
∂x ∂x ∂y ∂y ∂z ∂z
+2 ( )( )+( )( )+( )( ) dX1 dX3
∂X1 ∂X3 ∂X1 ∂X3 ∂X1 ∂X3
 
∂x ∂x ∂y ∂y ∂z ∂z
+2 ( )( )+( )( )+( )( ) dX2 dX3
∂X2 ∂X3 ∂X2 ∂X3 ∂X2 ∂X3

which means
 1/2
∂x 2 ∂y 2 ∂z 2
h1 = ( ) +( ) +( ) ,
∂X1 ∂X1 ∂X1
 1/2
∂x 2 ∂y 2 ∂z 2
h2 = ( ) +( ) +( ) ,
∂X2 ∂X2 ∂X2
 1/2
∂x 2 ∂y 2 ∂z 2
h2 = ( ) +( ) +( )
∂X3 ∂X3 ∂X3

and also gives the conditions which have to be fulfilled for (X1 , X2 , X3 ) to be orthogonal coordinates:
 
∂x ∂x ∂y ∂y ∂z ∂z
( )( )+( )( )+( )( ) = 0,
∂X1 ∂X2 ∂X1 ∂X2 ∂X1 ∂X2
 
∂x ∂x ∂y ∂y ∂z ∂z
( )( )+( )( )+( )( ) = 0,
∂X1 ∂X3 ∂X1 ∂X3 ∂X1 ∂X3
 
∂x ∂x ∂y ∂y ∂z ∂z
( )( )+( )( )+( )( ) =0
∂X2 ∂X3 ∂X2 ∂X3 ∂X2 ∂X3

10
 Physics 1. Mechanics
Advanced 1. Curvilinear coordinates

1 Advanced material
Let us assume that we are interested in describing the same n-dimensional space with two different
coordinate systems, xi and x0i . For simplicity we assume that xi are cartesian, that is, ds2 = i dx2i .
P

Of course, there should be a relation between the two systems, that is, xi = xi (x0j ), i = 1, . . . , n,
j = 1, . . . , n. We are interested in the distance element which is ds2 = 2
P
i dxi in the cartesian
coordinates. Now dxi is a full differential:
X  ∂xi 
dxi = dx0j , (1)
j
∂x0j

so that
X
ds2 = dx2i
i
! !
X X  ∂xi  X  ∂xi 
= 0
dx0j · dx0k
i j
∂x j k
∂x0k
" # (2)
X X  ∂xi   ∂xi 
= 0
· 0
dx0j dx0k
j,k i
∂xj ∂xk
X
= gjk dx0j dx0k
jk

Now we can forget the initial cartesian coordinates and conclude that the most general form for the
distance element is
X
ds2 = gij dxi dxj , (3)
i,j

where gij = gji . The construction gij is called a metric tensor. The metric tensor is symmetric
gij = gji , which means that it has 6 independent components in 3-dimensional space.
Q: How many independent components has gij in n-dimensional space ?
Q: Derive metric tensor for x = x0 − y 0 , y = y 0 .
Q: Derive metric tensor for elliptical coordinates ρ, ϕ, where ρ2 = x2 /a2 + y 2 /b2 , tan ϕ = y/x.

1
Physics 1. Mechanics Advanced 1

In cartesian coordinates gij = δij , where


1 if i = j
δij = (4)
0 if i 6= j

In cylindrical coordinates (1 = r, 2 = ϕ, 3 = z) one has

g11 = 1, g22 = r2 , g33 = 1,

(5)
g12 = g13 = g23 = 0.

In spherical coordinates (1 = R, 2 = θ, 3 = ϕ)

g11 = 1, g22 = R2 , g33 = R2 sin2 θ,

(6)
g12 = g13 = g23 = 0.

Coordinates, for which gij = 0 if i 6= j are called orthogonal.

The space where it is possible to choose global (that is, which cover the whole space) cartesian
coordinates, such that gij = δij everywhere, is called Euclidean.
General remarks. The space of special relativity is pseudo-Euclidean, since it is possible to choose
coordinates (1 = x, 2 = y, 3 = z, 4 = ct - time) so that g11 = g22 = g33 = 1, g44 = −1:

ds2 = dx2 + dy 2 + dz 2 − c2 dt2 .

This space is called Minkowsky space. General relativity states that the space is locally Minkowskian.

2
 Physics 1. Mechanics
Lecture 2. Vectors

1 A simplest vector: Displacement

Let us assume that a particle starts moving from the point P1 (x1 , y1 ) and proceeds subsequently
to the point P2 (x2 , y2 ) and further to P (x3 , y3 ) (we shall work in the two-dimensional space for
simplicity of graphical representation, see Fig. 1). As a result of the first move the coordinates

P2(x2, y2)
−−→
P 1P
2
−−→
P 2P

P1(x1, y1 )
3

−
P1−P→ P3(x3, y3)
2
x

Figure 1: Displacement vector.

change by (x2 − x1 , y2 − y1 ). The same can be described by saying that the particle moved by
p
the distance (x2 − x1 )2 + y2 − y1 )2 in certain direction. The two parameters (two coordinates or
distance and direction) completely determine the particle displacement from the initial point. We
−−→
shall say that a displacement vector P1 P2 is defined either as the two components (x2 − x1 , y2 − y1 )
or the distance+direction (see Fig. 2).
The numbers x2 − x1 and y2 − y1 are called the vector x and y components, respectively. In
order to assign any physical meaning to this “displacement vector” we have to say what can be done
−−→ −−→
with it. First, let us notice that P2 P3 = (x3 − x2 , y3 − y2 ) and P1 P3 = (x3 − x1 , y3 − y1 ) are also
−−→ −−→ −−→
displacement vectors, and P1 P3 = P1 P2 + P2 P3 if we define the vector summation as the summation

1
 y

y2 P2

y1 P1

x
x1 x2

Figure 2: Displacement vector and coordinates.

of the corresponding components: x3 −x1 = (x2 −x1 )+(x3 −x2 ), and similarly for y. This means that
moving by some distance in some direction and afterwards by another distance to another direction
we as a result move by some third distance to some third direction, and the rule to find this resulting
move is known (and independent of the order of the two movements).
−−→
Let us denote the distance between P1 and P2 as |P1 P2 | and call that the magnitude of the vector.
−−→
If we want to move from the point P1 in the same direction but by a distance λ|P1 P2 |, where λ is some
real number, we shall say that the movement is in the same direction, if λ > 0, and in the opposite
−−→
direction, if λ < 0. It is easy to see that the same can said as follows: λP1 P2 = (λ(x2 −x1 ), λ(y2 −y1 )),
that is, multiplication of a vector by a number is multiplication of all its components by this number.

These two operations, vector summation and multiplication by a number, make the construction
“something with direction” meaningful. In fact, a construction may be called a vector only if these
two operations are properly defined.

2 Vectors more generally

In analogy with the displacement vector we shall call an object A a vector if a) it has several
components (3 in the three-dimensional space), b) the summation is defined, and c) multiplication by
a real number is defined. We shall write A = (Ax , Ay , Az ) or A = (A1 , A2 , A3 ) or A = (Ai , i = 1, 2, 3).
Then, for each A and B, the vector sum C = A + B means that Ci = Ai + Bi , i = 1, 2, 3. For
each A and real α, D = αA means Di = αAi (we shall omit i = 1, 2, 3 where this is obvious). The

2
 y

y3 P3
y2 P2

y1 P1

x
x1 x2 x3

Figure 3: Multiplication by a number

important features of the summation and multiplication are the following:

A + B = B + A, (1)
α(A + B) = αA + αB, (2)
(α + β)A = αA + βA. (3)

The graphical representation of vectors would be the same as for the case of the displacement
vector. Fig. 4 shows the relation between the vector components (Ax , Ay ) (REMEMBER!: it is the
same as (A1 , A2 ) !), the vector magnitude |A| and the angle Ax
c between the vector and x-axis for
the two-dimensional case:
Ax = |A| cos(Ax),
c Ay = |A| sin(Ax).
c (4)
p
It is also clear that |A| = A2x + A2y .
Geometrically vector summation is shown in Fig. 5.

3 Scalar product
Given two vectors A = (Ai ) and B = (Bi ) (see Fig. 6) one can construct a scalar
X
A·B = Ai Bi . (5)
i

This construction is called scalar product or dot-product. Scalar is an object which does not depend on
the choice of coordinates. Let us check whether the construction (5) is indeed coordinate independent.

3
 y

Ay
A

x
Ax

Figure 4: Vector and components.

For simplicity we consider a two-dimensional case, for which A1 = |A| cos(Ax),

c A2 = |A| sin(Ax),
c
B1 = |B| cos(Bx),
d and B2 = |B| sin(Bx).
d Substituting into (5), we have

A · B = |A| · |B| · cos(Ax

c − Bx),
d (6)

or, noticing that Ax d = AB

c − Bx d is the angle between A and B, one gets

A · B = |A| · |B| · cos(AB).

d (7)

In the right hand side of the last expression the vector magnitudes and the angle between the two
vectors are completely independent of the coordinate choice (coordinates or vector components do
not even appear there) which means that the scalar product is also coordinate independent, that is,
invariant. This is true not only in the two-dimensional space but in any dimensions.
Properties of the scalar product. From the very definition it is clear that

A · B = B · A, (8)
(αA) · B = α(A · B), (9)
A · (B + C) = A · B + A · C. (10)

A2i = |A|2, so that

P P
Vector magnitude (length). From the definition A · A = i Ai Ai = i

√
|A| = A · A. (11)

This property allows to define a unit vector e such that e · e = 1: for any nonzero vector A/|A|

4
 C =A+B

B
A

A
C

C
B
Figure 5: Vector summation.

will have a unit length. Of special use are the unit vectors along the axes of the cartesian coordinate
system ei , i = 1, 2, 3 (or i = x, y, z - the same !).
With the help of these unit vectors we can represent any vector as follows (see Fig. 8):
X
A= Ai ei . (12)
i

Now we have several equivalent representations of a vector:

X
A = (Ai , i = 1, 2, 3) = Ai ei . (13)
i

It is important to understand that all these representations mean exactly the same. Use the one
which is most convenient.
Unit vectors and vector decomposition. We saw above that with the use of ex , ey , and ez , a vector
A can be decomposed as
A = Ax ex + Ay ey + Az ez .

The choice of the unit vectors ei in the cartesian coordinates is global, that is, their direction does
not depend on the point. However, we can choose three unit vectors in each point of the space
independently and decompose a vector in the same way. The three unit vectors do not even have to
be orthogonal, that is, the condition ei · ej = 0, if i 6= j, is not necessary (it is very convenient though
and we shall use it in what follows). The only necessary condition is that they are independent, that
is, if a1 e1 +a2 ej +a3 e3 = 0, then a1 = a2 = a3 = 0. Thus, in general, we can choose three unit vectors
ei which are not position dependent (that is, their directions are different in different points) and

5
 A

(AB)
d
B
Figure 6: Two vectors: notation.

decompose a vector as in (13). In practice, such unit vectors are usually related to some coordinate
frame, that is, the direction is chosen along the tangential to the curves building the frame.
Example - spherical coordinates. In spherical coordinates a curve θ = const, ϕ = const, is a
radius, which means that we choose the unit vector e1 ≡ eR parallel to the radius-vector r:
r
eR =
|r|
= (sin θ cos ϕ, sin θ sin ϕ, cos θ) (14)

= sin θ cos ϕex + sin θ sin ϕey + cos θez .

The vector e2 ≡ eθ is tangential to R = const and ϕ = const, and the vector e3 ≡ eϕ is tangential
to R = const and θ = const. They can be found from geometrical considerations but it is instructive
to derive these two from the orthogonality conditions. We shall start with eϕ which can be written as
eϕ = a1 ex + a2 ey (it is always in a plane parallel to x − y plane since θ = const. The two conditions
are

eϕ · eR = a1 sin θ cos ϕ + a2 sin θ sin ϕ = 0, (15)

eϕ · eϕ = a21 + a22 = 1. (16)

The solution is a1 = − sin ϕ, a2 = cos ϕ, or a1 = sin ϕ, a2 = − cos ϕ. In the special case eR = ex

(θ = π/2, ϕ = 0) one has eϕ = ey , which shows that we have to choose the first set:

eϕ = − sin ϕex + cos ϕey .

6
 z

êz
êx y
êy

x
Figure 7: Unit vectors êi .

The third unit vector can be expressed as eθ = b1 ex + b2 ey + b3 ez and there are three conditions:

eθ · eR = b1 sin θ cos ϕ + b2 sin θ sin ϕ + b3 cos θ = 0, (17)

eθ · eϕ = −b1 sin ϕ + b2 cos ϕ = 0, (18)
eθ · eθ = 1. (19)

The second condition gives b2 = b1 sin ϕ/ cos ϕ. Substituting into the first equation, one gets

b1 sin θ/ cos ϕ + b3 cos θ = 0,

and after substituting into the third condition one would get

b3 = − sin θ, b1 = cos θ cos ϕ, b2 = cos θ sin ϕ.

Q: Why the sign of b3 is chosen as above and not b3 = sin θ ?

Now
eθ = cos θ cos ϕex + cos θ sin ϕey − sin θez .

With the use of these unit vectors the radius-vector can be written as r = ReR .
Exercise. Find er , eϕ , and ez for cylindrical coordinates.
Scalar product - scientific notation. We can use (12) and (8)-(10) to write the scalar product as

7
 A = Ax êx + Ay êy
y

A
êy
x
êx
Figure 8: Vector, components, and unit vectors.

follows:
X X
A·B =( Ai ei ) · ( Bj ej )
i j
X
= Ai Bj (ei · ej )
i,j
X (20)
= Ai Bj δij
i,j
X
= Ai Bi .
i

Here we used the properties of the unit vectors ei : ei · ej = 1 is i = j and zero otherwise, and have
defined the famous Kronecker delta-symbol :

1, if i = j,
δij = (21)
0, if i =
6 j.

Properties of δij . The delta-symbol is fully symmetric: δij = δji . The delta-symbol is used to
filter indices:
X
δij Aj = Ai . (22)
j

8
Useful formulae:
X
δii = 3, (23)
i
X
δij δjk = δik . (24)
j

Scalar product - angle between two vectors. From the relation A · B = |A| · |B| cos(AB)
d one
immediately finds:
A·B
cos(AB)
d =√ √ . (25)
A·A B·B
Thus, if the components of the two vectors A and B are known it is straightforward to find the
angle.
Cosine theorem. Let there is a triangle built on two vectors running from the same point, A and
B. The third vector is then C = B − A. The length of C can be found using the scalar product:

C 2 = (B − A)2 = A2 + B 2 − 2A · B.

3.1 Decomposition for two vectors

Let there are two vectors, A and B, and we want to represent the vector B as a sum B = Bk + B⊥ ,
where = Bk k A, and = B⊥ ⊥ A (see figure).

B⊥

Bk A
Figure 9: Decomposition for two vectors.

9
4 Vector product
Besides the above described scalar product, another object can be constructed which can be shown
to be a vector. This object is called a vector product C = A × B and is derived according to the
following rule

C1 = A2 B3 − A3 B2 ,
C2 = A3 B1 − A1 B3 , (26)
C3 = A1 B2 − A2 B1 .

From the definition we can see that such vector product can be constructed only in three-dimensional
space. From (26) one has

e1 × e2 = −e2 × e1 = e3 ,
e2 × e3 = −e3 × e2 = e1 , (27)
e3 × e1 = −e1 × e3 e2 ,
P
and using A = i Ai ei and similar representation for B it is possible to write

e1 e2 e3

C = A × B = det A1 A2 A3 . (28)


B1 B2 B3

Q: Prove (27) and (28).

Properties of the vector product. The following properties of the vector product follow from the
definition:

A × B = −B × A,
(αA) × B = α(A × B), (29)
A × (B + C) = A × B + A × C.

Magnitude and direction. Let for simplicity A and B be in x − y plane, that is, A3 = 0, B3 = 0.
We shall also write

A1 = |A| cos(Ax),
c A2 = |A| sin(Ax),
c
(30)
B1 = |B| cos(Bx),
d B2 = |B| sin(Bx).
d

Substituting (30) into (28) one finds (see Fig. 10):

10
 A×B

90◦

Figure 10: Direction of the vector product.

C = A×B
(31)
= |A||B| sin(AB)e
d 3.

Q: Prove (31)
Q: Show that the length of A × B is equal to the area of the parallelogram built on the two
vectors.
Q: Show that C ⊥ A and C ⊥ B, and A, B, and C obey the right hand rule.
Vector product -scientific notation. One most important construction is defined usually: the
Levy-Chivita symbol 


 1, if ijk = 123, or 231, or 312

εijk = −1, if ijk = 132, or 321, or 213 (32)


0,

if at least two from ijk are equal.
With this symbol the vector product C = A × B can be written as follows:
X
Ci = εijk Aj Bk . (33)
j,k

Q: Prove (33)
Another, probably even better definition is ε123 = 1 and it is completely antisymmetric: εijk =
−εjik = −εikj . In other words, each transposition of two indices multiplies by −1.
Q: How many independent components has εijk ?

11
4.1 EXERCISES.
The below exercises with the indices and symbols are very important for better understanding.
P P
1. Calculate i δii and ij δij .
P
2. Calculate ij (Ai Bj − Aj Bi )δij .
P
3. Calculate ijk εijk .
P
4. Calculate jk εijk δjk .
P
5. Calculate ijk εijk εijk .
P
6∗ . Calculate m εmij εmkl . HINT: try
X
εmij εmkl = K(δik δjl − δil δjk )
m

and find K.
P
7∗ . Calculate mn εimn εjmn . HINT: try
X
εimn εjmn = Xδij
mn

and find X.

5 More complex operations with vectors

Let A, B, and C are three arbitrary vectors. We know how to build a scalar from two vectors:
A · B. This scalar can be multiplied by the third vector to give another vector

(A · B)C

which is parallel to C.
On the other hand, we can build a vector A × B. This vector can be multiplied by the third
vector either to give a scalar:
C · (A × B)

or to give a new vector

C × (A × B)

Q: Show that |C · (A × B)| is a volume of the parallelepiped built on the three vectors.
Q: Show that
C · (A × B) = (C × A) · B

12
 Q∗ : Derive the formula:

A × (B × C) = B(A · C) − C(A · B). (34)

ATTENTION: this expressions are very important and useful and we will need them in future.

6 Displacement vector and curvilinear coordinates

Displacement vector may be used for derivation of the scale factors hi in curvilinear coordinates.
Indeed, dr 2 = ds2 , thus calculating
 2
2 ∂r ∂r ∂r
dr = ( )dX1 + ( )dX2 + ( )dX3
∂X1 ∂X2 ∂X3
∂r 2 2 ∂r 2 2 ∂r 2 2
=( ) dX1 + ( ) dX2 + ( ) dX3
∂X1 ∂X2 ∂X3
∂r ∂r ∂r ∂r ∂r ∂r
+ 2( )·( )dX1 dX2 + 2( )·( )dX1 dX3 + 2( )·( )dX2 dX3
∂X1 ∂X2 ∂X1 ∂X3 ∂X2 ∂X3

we have

∂r 2 ∂r 2 ∂r 2
h21 = ( ) , h22 = ( ) , h23 = ( ) ,
∂X1 ∂X2 ∂X3
∂r ∂r ∂r ∂r ∂r ∂r
( )·( )=( )·( )=( )·( )=0
∂X1 ∂X2 ∂X1 ∂X3 ∂X3 ∂X3

This becomes even more obvious if we write

r = xx̂ + y ŷ + z ẑ,
∂r ∂x ∂y ∂z
= x̂ + ŷ + ẑ,
∂X1 ∂X1 ∂X1 ∂X1
∂r ∂x ∂y ∂z
= x̂ + ŷ + ẑ,
∂X2 ∂X2 ∂X2 ∂X2
∂r ∂x ∂y ∂z
= x̂ + ŷ + ẑ,
∂X3 ∂X3 ∂X3 ∂X3
∂r 2 ∂x 2 ∂y 2 ∂z 2
( ) =( ) +( ) +( ) ,
∂X1 ∂X1 ∂X1 ∂X1
...

and compare the obtained expressions with Lecture 1.

13
 Physics 1. Mechanics
Advanced 2. Coordinate transformation and vectors, tensors

1 Advanced material
Let us consider two cartesian coordinate frames (x, y) (“old”) and (x
, y
) (“new”), which are related
by a rotation to the angle θ (see Fig. 1). Let A is a vector in this plane. This vector can be
represented by its components in both frames A = (Ax , Ay ) = (A
x , A
y ). Obviously, the should be
some relations between the old and new coordinates, old and new vector components, and old and
new unit vectors (ex , ey ) and (e
x , e
y ).
From the figure one can find:

x = x
cos θ − y
sin θ, y = x
sin θ + y
cos θ, (1)

Ax = A
x cos θ − A
y sin θ, Ay = A
x sin θ + A
y cos θ, (2)
ex = e
x cos θ − e
y sin θ, ey = e
x sin θ + e
y cos θ. (3)

We see that the vector components transform similarly to the coordinate transformation.
Q: Show that i Ai ei = i A
i e
i .

This is a basic feature of vectors and another, in many cases better, definition of a vector. In
a more general way, let x
i = x
i (xj ) in an n-dimensional space. The differentials dx
i depend on the

Figure 1: Two coordinate systems - same vector.

1
Physics 1. Mechanics Advanced 2

differentials dxj as follows:

  ∂x

i
dx
i = dxj , (4)
j
∂xj

which is just the rule for the full differential.

Then a vector Ai , i = 1, . . . , n, is an object which transforms like the differentials:

  ∂x

i
A
i = Aj . (5)
j
∂xj

A vector in n-dimensional space has n components.

It is possible to define a more general construction which is called tensor. For example, a 2-rank
tensor is a two-index object Tij (n2 components) which transforms as follows:

  ∂x
  ∂x
j 
i
Tij
= Tkl . (6)
kl
∂xk ∂xl

In previous lecture we introduced the metric tensor gij . Let us check whether it is indeed a tensor.
The distance element ds2 = ij gij dxi dxj is invariant under coordinate transformations, that is, ds2

is independent of the coordinate choice. Therefore, writing the same distance element in two different
coordinate frames, one has


dx
k dx
m ,
 
gij dxi dxj = gkm
ij km

and using
  ∂xi    ∂xj 
dxi = dx
k , dxj = dx
m ,
k
∂x
k m
∂x
m

we obtain  
   ∂xi   ∂xj 
gij dx
k dx
m .
km ij
∂x
k ∂x
m


dx
k dx
m we find that


Comparing this with km gkm

  ∂xi   ∂xj 


gkm = gij , (7)
ij
∂x
k ∂x
m

which does not look exactly as (6). Nevertheless, gij is a tensor. In fact, in curvilinear coordinates
exist tensors (and vectors too) of two kinds (they are not completely independent but related through
gij ). We shall call them up and down (these are not real names which we will not use now). For
example, up-vector transforms as (5), while down-vector transforms as follows:

  ∂xj 
A
i = Aj . (8)
j
∂x
i

2
Physics 1. Mechanics Advanced 2

ATTENTION: index i in both expressions goes with the primed component A
i and coordinate x
i .
Example - two-dimensional cartesian coordinates. For convenience let us introduce new notation:

∂x
i
   
∂xj
Lij = , L
ji = . (9)
∂xj ∂x
i

Then the up- and down-vectors transform as follows:

up: A
i = down A
i = L
ji Aj .

 
Lij Aj , (10)
j j

As in the beginning we consider two cartesian frames related by rotation to the angle θ:

x1 = x
1 cos θ − x
2 sin θ, x2 = x
1 sin θ + x
2 cos θ, (11)

x
1 = x1 cos θ + x2 sin θ, x
2 = −x1 sin θ + x2 cos θ. (12)

Using these relations, we obtain

L11 = L22 = L
11 = L
22 = cos θ,

(13)
L12 = −L21 = −L
12 = L
21 = − sin θ.

The up- and down-vectors transform as (see (10))

A
1 = L11 A1 + L12 A2 , A
2 = L21 A1 + L22 A2 , up (14)

A
1 = L
11 A1 + L
21 A2 , A
2 = L
12 A1 + L
22 A2 , down (15)

and, because of (13), the two transformation laws are the same. Thus, there is no difference between
up- and down-vectors in this case. Transformations, for which Lij = L
ji , that is, (∂x
i /∂xj ) =
(∂xj /∂xi ), are called orthogonal. Rotations are orthogonal transformations.

3
 Physics 1. Mechanics
Lecture 3. Motion

1 General concepts
Any change of the position with time is called motion. This “definition” implies that we have a good
method to measure time everywhere and always. For the time being let us assume that indeed such
method exists so that we can observe xi (t), that is, coordinates of a particle are functions of time.
If the coordinates do not change we say that the particle is in the rest and consider that as a special
case of motion.
By providing xi (t) we describe the motion completely, since we know where is the particle at
each moment of the time. However, sometimes we want to know some particular things about
the motion, for example, how quickly the particle position is changing. For the description of this
“quickly” we define the velocity vi ≡ (dxi /dt). It is important to understand that, in general, the
velocity is defined using coordinates and can differ from our usual perception of the speed. Thus,
in cartesian coordinates one has the following components of the velocity: (dx/dt), (dy/dt), and
(dz/dt). In spherical coordinates the components are (dr/dt), (dθ/dt), and (dϕ/dt), that is, even
do not have the same dimensions. If necessary, we shall emphasize this by calling this velocity a
coordinate velocity, and the other one (see below) a physical velocity. In general relativity, however,
it is not always possible to define globally the physical velocity as we understand it in our everyday
life (measured in km/h, for example), and one has to use the velocity introduced above.
In a similar way one introduces acceleration ai ≡ (d2 xi /dt2 ) which shows how quickly the velocity
is changing. One might go further and introduce higher derivatives. However, this is useful only
in very few cases, because of the structure of classical mechanics, where external influence (force)
determines acceleration.
Once we know xi (t), a simple procedure (derivation) gives us subsequently the velocity and
acceleration. The inverse procedure is a little bit more complicated but still remains only a technical
one. Namely, if we know vi (t) and the initial conditions xi (t = t0 ) = xi0 , one immediately gets:

t
xi (t) = xi0 + vi (t
)dt
. (1)
t0

ATTENTION: for not to make a mistake one has to make clear the distinction between the moment
of time t in which we would like to have xi (t), and the time variable t
which is the integration

1
Physics 1. Mechanics Lecture 3

variable only. Similarly, if we know the acceleration ai (t) and the initial conditions vi (t = t0 ) = vi0 ,
we have:
t
vi (t) = vi0 + ai (t
)dt
. (2)
t0

Calculation of the distance is a little more complicated since we have to know how to write the
distance element ds. We shall delay the discussion of this until the next session where our ordinary
space is considered (in which it is always possible to choose cartesian coordinates). It has to be
mentioned, however, that both velocity and acceleration are vectors.
All points which are passed by the particle in space constitute a trajectory (or path or orbit).
Examples of trajectories are shown in Fig. 1

Figure 1: Trajectories in two-dimensional (left) and one-dimensional (right) cases.

2 Alternative approach
An alternative approach is usually used in the case where cartesian coordinates are possible (but not
necessarily chosen). In our three-dimensional space we can define a displacement vector (see previous
lecture). We can go even further and determine a particle position by the displacement vector from
the origin of the coordinates to the point where the particle is now. Following general principles
this vector is P0 P , where P = (x, y, z) (in cartesian coordinates) and P0 = (0, 0, 0) (origin !). Thus,
P0 P = (x, y, z). This vector is denoted r and is called a position vector or radius-vector. Using this
vector, we can define velocity and acceleration in a more concise way:

dr dv d2 r
v= , a= = 2. (3)
dt dt dt

2
Physics 1. Mechanics Lecture 3

Now (1)-(2) will take the following shape:

t
r(t) = r0 + v(t
)dt
, (4)
t0

t
v(t) = v0 + a(t
)dt
, (5)
t0

It is now even sufficiently easy to calculate the distance. Indeed, from ds2 = dx2i (in cartesian

i
coordinates), one has
1/2
  dxi 2

ds
vi2 )1/2 = (v · v)1/2 = |v|,

= =( (6)
dt i
dt i

so that
t
s= |v|(t
)dt
. (7)
t0

In practice, in order to calculate the distance one always has to substitute

 2  2  2 1/2
dx dy dz
|v| = + + . (8)
dt dt dt

In spherical coordinates one has to use

 2  2  2 1/2
dr dθ dϕ
|v| = + r2 + r 2 sin2 θ . (9)
dt dt dt

Q: Write |v| in cylindrical coordinates.

2.1 Usage of unit vectors

The velocity vector can be written as

v = ẋex + ẏey + żez .

Here we use the short notation Ȧ ≡ dA/dt for any A. Instead of using the cartesian unit vectors
one could use another set of those, similarly as we did in Lecture 2. We shall show how it is done
on the example of two-dimensional polar coordinates. We define er = r/r as a unit vector along the
radius-vector (tangential to the curve ϕ = const, and eϕ as a unit vector tangential to r = const
(there is no z - two-dimensional case). It is easy to find

er = cos ϕex + sin ϕey , eϕ = − sin ϕex + cos ϕey . (10)

3
Physics 1. Mechanics Lecture 3

From these expressions we find

ex = cos ϕer − sin ϕeϕ , ey = sin ϕex + cos ϕey . (11)

Now the velocity

dr
v= = ẋex + ẏey , (12)
dt
where
x = r cos ϕ, y = sin ϕ, (13)

and

ẋ = ṙ cos ϕ − r sin ϕϕ̇,

(14)
ẏ = ṙ sin ϕ + r cos ϕϕ̇.

Substituting (11) and (14) into (12) we have

v = ṙer + r ϕ̇eϕ . (15)

Q: Express v in terms of eR , eθ , and eϕ (spherical coordinates).

3 Motion - examples
3.1 Constant acceleration
Let a = const. This means that neither the magnitude nor the direction of the acceleration do not
change. From (4)-(5) one has

v = v0 + a(t − t0 ), (16)
r = r0 + v0 (t − t0 ) + 12 a(t − t0 )2 . (17)

Q: Prove these expressions.

The most famous example of such motion is the motion near the earth surface, where the free-fall
acceleration g = const. Let us choose the coordinate system as in Fig. 2. Then g = (0, −g) and one
has for the components and coordinates

vx = v0 cos α, (18)
vy = v0 sin α − g(t − t0 ), (19)
x = x0 + v0 cos α(t − t0 ), (20)
y = y0 + v0 sin α(t − t0 ) − 12 g(t − t0 )2 . (21)

4
Physics 1. Mechanics Lecture 3

v0

α x

Figure 2: Motion near the earth surface.

Trajectory Trajectory (path) is the curve along which the particle moves. The equation for the
trajectory can be obtained by excluding time from the equation for x and further substituting into
the equation for y:
g(x − x0 )2
y = y0 + (x − x0 ) tan α − 2 . (22)
2v0 cos2 α
Q: Derive (22).
Distance In order to calculate the distance along the path (trajectory) we have to integrate

t
s= |v(t
)|dt

t

0t  (23)
= v02 cos2 α + (v0 sin α − g(t
− t0 ))2 dt
.
t0

3.2 Circular motion

This motion is in a plane. It is natural to use polar coordinates. In this coordinates r = const,
while the angle ϕ = ϕ(t) is time dependent. The derivative ω = dϕ/dt is (naturally) called angular
velocity (if the radius r is changing, dr/dt is called radial velocity). The velocity magnitude is found
from |v| = (dr/dt)2 + r 2 (dϕ/dt)2 = ωr. In most cases, however, this motion is described in terms

of r and v. In order to do that let us write down the relation between the cartesian and cylindrical
coordinates of the particle:
x = r cos ϕ, y = r sin ϕ, z = z, (24)

5
Physics 1. Mechanics Lecture 3

from which (with the use of simple differentiation) one gets

vx = −ωr sin ϕ, vy = ωr cos ϕ, vz = 0. (25)

Now we introduce the angular velocity vector ω = ωez and notice that (25) can be written as

v = ω × r. (26)

The physical sense of this vector is that it shows the rate of the angle change and also the direction
of the rotation axis.
Further differentiating (26) with respect to time one has

dω
a= × r + ω × (dr/dt)
dt
dω
= ×r+ω×v (27)
dt
dω
= × r + ω × (ω × r)
dt

Let us now write r = r + r⊥ , where r = zez  ω, r⊥ = xex + yey ⊥ ω. Then application of the
formulae of the vector algebra gives

dω
a= × r − ω 2 r⊥ . (28)
dt

Q: Prove (28).
In the special case when the angular velocity does not change ω = const the centripetal acceler-
ation is a = −ω 2 r⊥ .
Let us now consider a more general case when ω = ω(t)ez . In the polar coordinates r⊥ = rer ,
and ez × er = eϕ , so that (28) will give

a = ω̇eϕ − ω 2rer . (29)

The first term in the fight hand side of (29) is called tangential acceleration, the second term is called
centripetal acceleration.

4 General
A particle motion is given by r(t) and the velocity is v = dr/dt. From the very definition the velocity
is tangent to the trajectory. What about the acceleration ? The acceleration is defined as a = dv/dt,
and, in general, it is neither parallel nor perpendicular to the velocity. However, we can decompose

6
Physics 1. Mechanics Lecture 3

it similarly to what has been done in the case of the circular motion. For this to do let us write

v = vv̂, (30)

where v = |v| is the velocity magnitude, and v̂ = v/v is the unit vector in the direction of the velocity.
Of course, so far we have not done anything new by writing the identity (30). Now, however, the
acceleration can be written as follows:

d(vv̂) dv dv̂
a= = v̂ + v . (31)
dt dt dt

We see that the first term in (31) is parallel to v̂  v. The second term is perpendicular to the
velocity. Indeed, since v̂ · v̂ = 1 one has

d dv̂
(v̂ · v̂) = 0 = 2v̂ · .
dt dt

which means that dv̂/dt ⊥ v̂. Thus, we have decomposed the acceleration into the tangential and
normal components:

dv
at = v̂, (32)
dt
dv̂
an = v . (33)
dt

The first one shows the change in the velocity magnitude, the second one gives the change of the
direction. The centripetal acceleration is the special case of the normal acceleration when a particle
moves along a circle. In the circular motion the magnitudes of the velocity and centripetal acceleration
are related through the circle radius: |a| = v 2 /R. We can define similarly the curvature radius of the
trajectory in the general case as R = v 2 /|an |, so that

1 dv̂
R= .
v dt

Let us now pay attention that the vector dv̂ is a unit vector, while the vector dv̂/dt is not, in
general. We can build a new unit vector

dv̂/dt
N̂ = , (34)
|dv̂/dt|

which is called the vector of the normal. In order to complete these two vectors we have to add a
third one, which can be done as b̂ = v̂ × N̂.

7
 Physics 1. Mechanics
Advanced 3. Motion in curvilinear coordinates

1 Advanced material
1.1 General coordinates
In general coordinates vi = dxi /dt and ds2 =
P
ij dxi dxj . Therefore,

ds X dxi dxj
=( gij )1/2
dt dt dt
ij (1)
X
=( gij vi vj )1/2
qP
Thus, the velocity magnitude is |v| = ij gij vi vj . However, is velocity a vector ? From the previous
lecture we know what should be the transformation law when we change coordinates. Indeed,

dx0i X ∂x0i dxj X ∂x0i

vi0 = = = vj ,
dt j
∂xj dt j
∂xj

provided that the time t does not change when we change the coordinates. This means that the
velocity defined as vi = dxi /dt is a vector only if t is invariant under coordinate transformation. In
nonrelativistic classical mechanics t is an absolute time, which is the same for any coordinate choice,
so that vi is a vector.
If xi = xi (t) this defines a one-dimensional curve in the space. Is the choice of time the only
method to define a curve ? Obviously, any substitution t = t(τ ) is also good, if there is one-to-one
correspondence. Since s(t) is a monotonic function, s can be used for the curve parametrization with
equal success. Let us say that s is the new “time” and see what happens if we define new velocity
as dxi /ds. The only effect of our choice would be in the magnitude of the velocity:
X dxi dxj
v·v = gij = 1.
ij
ds ds

1
Physics 1. Mechanics Advanced 3

1.2 Differentiation along the curve

In section 1.1 we encountered the derivative of the kind v(dA/dt) (where A is some vector). Such
derivative is called the derivative along the curve. Using the definition of the curve length ds/dt = v
we can rewrite such derivative in terms of derivation with respect to the path:

d d
v = v̂ . (2)
dt ds

With this definition the curvature radius takes the form

 
−1
 dv̂ 
R =   . (3)
ds

Since we live in a three-dimensional space, any curve is three-dimensional, in general, and one
curvature radius is not sufficient to describe the curve behavior. Indeed, a second curvature is
defined related to the twisting  
 db̂ 
T =  (4)
 
 ds 

(see section 1.1 for the definition of b̂ and N̂).

The first and second curvatures are defined as C1 = 1/R, C2 = T , and the total curvature is
C 2 = C12 + C22 .
Exercise: Using N̂ = b̂ × v̂ and differentiating with respect to s show that
 
 dN̂ 
C= .
 
 ds 

2
 Physics 1. Mechanics
Lecture 4. Relativity

1 Reference frames
A reference frame is an observer with a coordinate system. This implies that a) some coordinates
are chosen to describe the position of any particle, b) there is some physical body relative to which
the motion is considered (it is possible to perform physically meaningful measurements), c) there is
a method of time measurements. The last means that a clock should exist in each point of the space,
and all these clocks should be synchronized. They are synchronized by sending some signal from
some reference clock to all others, telling them what is the correct time in the reference point. Such
procedure requires that the clocks in all points of the space be identical. In Newtonian mechanics the
signal speed is assumed to be infinite, so that the synchronization is straightforward and easy. Let us
now assume that there are two frames, S and S
, so that S
moves with the velocity V relative to S.
We also assume that both observers put themselves in the coordinate origin of each of these systems,
and both choose cartesian coordinates building the axes so that x
is parallel to x, y
is parallel to
y, and z
is parallel to z. If the signal velocity is infinite it is possible to synchronize the clocks in
S and S
as well so that the two observers measure the same time: t = t
. On the other hand, it is
clear (see Fig. 1) that for each point P , r = r
+ R, where r is the radius-vector of P measured by
the observer S, r
is the radius-vector of P measured by the observer S
, and R is the radius-vector
of S
(coordinate origin) measured by the observer S. Thus, one has the Galileo transformation in
the following form:
t = t
, r = r
+ R. (1)

Differentiating (1) with respect to time t and taking into account that t = t
, we immediately find

v = v
+ V, (2)

where v = dr/dt is the velocity of P measured by S, v
= dr
/dt
is the velocity of P measured by S
,
and V = dR/dt is the velocity of S measured by S
, that is, the relative velocity of the two frames.
The two frames, S and S
, are said to be related by the Galileo transformation if V = const. In
Newtonian mechanics the following Galileo relativity principle holds: all physics laws have the same
shape in two systems related by the Galileo transformation. In particular, this means that the second
Newton law looks identical, F = ma and F
= m
a
, in both frames. Indeed, mass m is a scalar

1
Physics 1. Mechanics Lecture 4

V
y

y
r

r

R x

x

Figure 1: Galilean transformations.

and the same in both frames. The acceleration can be obtained by differentiating (2) with respect
to time,which gives a = a
(if V = const). From this we derive the law of the force transformation
F = F
.
In what follows we assume that the frame S is inertial. This, in particular, means that the force
in the second Newton law is caused by other physical bodies. In the absence of such bodies there is
no any force and a = 0. The frame S
is also inertial if V = const. What happens if the frame S
is
non-inertial, that is V is time-dependent ? Direct differentiation of (2) gives

a = a
+ A, (3)

where A is the acceleration of the frame S
measured by S. Substituting in the Newton law, we have

Fext = ma = ma
+ mA,

where the index ext emphasizes the nature of the force: it is produced by other physical bodies.
If we want to have the Newton law i in the accelerated frame S
similar to its usual form, that
is, F
= ma
, we have to define F
= Fext − mA. The second term in this expression is the inertia
force which is not caused by any physical body but is due to the the acceleration of the frame.

2
Physics 1. Mechanics Lecture 4

a F = −ma
y

y
r

r

R x

x

Figure 2: Inertial force in the accelerated system. d

This expression can be also considered as the transformation rule for the force.

1.1 Examples of inertial forces

Hanging body. Let a body hang on a rope from a ceiling (see Fig. 3) and let the ceiling move

−ma

mg

Figure 3: Hanging body in non-inertial frame.

with constant acceleration a. What would be the angle θ between the rope with the vertical ? The
simplest (and most effective) way is to consider the situation in the non-inertial accelerated frame,
where the position of the body is determined by the balance of the rope tension T (directed along
the rope), vertical gravity force mg and the inertia force −ma. This means that the vector sum
g + (−a) should be directed along the rope. It is clear now that tan θ = |a|/|g|.

Sliding down accelerated slope. What should be the acceleration of the slope to prevent sliding
of the body (see Fig. 4)? In the accelerated frame the vector sum of the gravity force mg and inertia

3
Physics 1. Mechanics Lecture 4

−ma

mg
α

Figure 4: Sliding down accelerated slope.

force −ma should be normal to the slope, which gives |a| = |g| tan α.

2 Rotating frames
A special and one of the most important case of non-inertial frames is a rotating frame. Let S be
again an inertial (“standing”) frame and S
rotates with the constant angular velocity ω around the
axis z (z
= z), so that the coordinate origins in S and S
are the same (see Fig. 5). According to

y
y

x

Figure 5: Standing x, y, z and rotating x
, y
.z
frames.

our definition of the angular velocity, ω = ωez .

Let in the moment t the angle between x and x
axes is ϕ. Let r and r
be the radii-vector of the
particle P as measured in S and S
, respectively. Attention: although both are shown by the same
“arrow” in the figure, the vectors cannot be said to be the same since the components are different,

4
Physics 1. Mechanics Lecture 4

instead we have to say that r = r
r , where the index r means “rotated from S
into S”. In the
following we shall not emphasize this difference which should, nevertheless, be kept in mind.
Our objective is to establish the relation between the velocities and accelerations measured in
both frames. Let us start with the simplest case where the particle P does not move at all in the
rotating frame, v
= 0. In this case, from the point of view of S, the particle P simple experiences
circular motion with the constant angular velocity. We know that the observer S would measure the
velocity v = ω × r, and the acceleration a = ω × v, The nature of this acceleration is simply the
rotation of the velocity vector. The radius-vector r and the velocity vector v rotate by the angle
ωdt during the infinitesimal time dt. Since the rotation of the both vectors is the same, the relation
between the velocity change dv and v is the same as between dr and r. The latter can be found
from what we already know: dr = vdt and v = ω × r, so that dr = ω × rdt. We conclude that for
the velocity vector one would have dv = ω × vdt. Moreover, for any vector A rotating with S
we
will have in the standing frame S: dA = ω × Adt. In a more general way, if a vector A is changing
in the rotating frame then
dA dA

= +ω×A (4)
dt dt
where (dA/dt) is the rate of the vector change as measured by the standing observed and (dA
/dt)
is the rate of the vector change as measured by the rotating observer and rotated into the standing
frame.
Let now the particle P move in the frame S
with the velocity v
in the moment t. The observer
S will measure in the same moment the velocity v = v
+ ω × r. During dt this velocity changes
by dv
in the frame S
. The observer S will measure the velocity change dv which is due to a) the
change dv
, b) the change due to the rotation of the velocity vector ω × vdt = ω × (v
+ ω × r),
and c) due to the change of the rotational term d(ω × r)m = ω × dr = ω × dr
= ω × v
dt =. The
last contribution simply says that the particle is on another circler in the moment t + dt, so that its
rotation velocity is not the same as before. Combining all these contributions together, dividing by
dt and taking into account that dv/dt = a, dv
/dt = a
, we finally obtain

a = a
+ 2ω × v
+ ω × (ω × r). (5)

The last term is already known to us: it is the centripetal acceleration and can be written as −ω 2 r.
The second term is called Coriolis acceleration. If now we wish to write the second Newton law in
the usual form F
= ma
, we have to write

F
= Fext − 2mω × v
+ mω 2 r
, (6)

where the first term is the force cause by other bodies, the second term is called the Coriolis force,
and the last one is the centrifugal force. The last two are inertia forces and are caused by the frame
rotation.

5
Physics 1. Mechanics Lecture 4

The below figures show once again the change of the velocity in the rotating and nonrotating
frame.

y
t y
t + dt

v
v

r
r

x
x

Figure 6: Velocity change in the rotating frame. The left figure shows the position r
(blue) and
velocity v
(red) of the particle in the moment t. The right panel shows what happens in the moment
t + dt. The particle moved into the new position r
+ v
dt and has the new velocity v
+ ∆v
.

y t y t + dt


y

y
v v

x


x
r r

x x
Figure 7: Velocity change in the nonrotating frame. In order to find the velocity in the nonrotating
frame we have to rotate the velocity in the rotating frame by the angle ϕ and add ω × r. The left
figure shows the velocity in the moment t: the velocity v
(red) is rotated by the angle ϕ, the position
r
(blue) is rotated by the same angle (and gives r) and ω × r (green) is added. The total velocity v
is shown in magenta. The right figure shows the velocity in the moment t + dt: v
+ dv
and r
+ dr

are rotated by the angle ϕ + dϕ, and ω × (r + dr) is added.

6
 Physics 1. Mechanics
Advanced 4. Rotations, general noninertial frames

1 Advanced material: Rotations

Two cartesian coordinate systems S(x, y, z) and S 0 (x0 , y 0 , z 0 ) with the common origin can be trans-
formed one into another by a rotation. Let us start with the rotation around z axis by the angle ϕz .
Then the transformation reads

x0 = x cos ϕz + y sin ϕz ,
y 0 = −x sin ϕz + y cos ϕz , (1)
z 0 = z.

It is convenient (and not only convenient !) to represent this transformation in a matrix form. Let
us define row  
r̃ = x y z ,

and column  
x
r = y  ,
 

z
and similarly for S 0 .
The transformation (1) can be now written as follows:
 0   
x cos ϕz − sin ϕz 0 x
 0 
y  =  sin ϕz cos ϕz 0 y  (2)
 

z0 0 0 1 z

or in the following symbolic form

r0 = Rz (ϕz )r, (3)

 
cos ϕz − sin ϕz 0
Rz (ϕz ) =  sin ϕz cos ϕz 0 . (4)
 

0 0 1

1
Physics 1. Mechanics Advanced 4

The construction Rz (ϕz ) is called an operator, and (4) is the matrix form of the rotation operator.
In a similar way we can find the operators (matrices) for rotations around x and y axes:
 
1 0 0
Rx (ϕx ) = 0 cos ϕx − sin ϕx  (5)
 

0 sin ϕx cos ϕx
 
cos ϕy 0 sin ϕy
Ry (ϕy ) =  0 1 0  (6)
 

− sin ϕy 0 cos ϕy

If we want to rotate, say, first around z by ϕz and after that around x by ϕx , the result will be

r00 = Rx (ϕx )r0 = Rx (ϕx )Rz (ϕz )r. (7)

In the component form (3) is written as follows:

X
x0i = Rij xj . (8)
j

For the rows one would have

r̃0 = r̃R̃, (9)

(remember matrix multiplication !) or

X
x0i = xj R̃ji . (10)
j

Thus,
R̃ji = Rij → R̃ = RT , (11)

where RT is the transposed matrix: Rij

T
= Rji .
P P 0 0
For any rotation i xi xi = k xk xk , which gives
X X X X
x0k x0k = Rki xi Rkj xj = T
Rik Rkjxi xj = δij xi xj . (12)
k ijk ijk ij

Thus,
X
T
Rik Rkj = δij , (13)
k
or
RT R = I, (14)

where I is the unit matrix (all diagonal elements equal to one, all nondiagonal equal to zero).

2
Physics 1. Mechanics Advanced 4

Definition. If for a matrix R exists a matrix R̃ such that R̃R = I, the matrix R̃ is called the
inverse matrix and denoted R−1 . Thus, for rotations we have R−1 = RT .
Since the matrix product is noncommutative, that is, for two matrices A and B, in general,
AB 6= BA, the result of two successive rotations depends on their order. However, if we consider
infinitesimal rotations ϕ → dϕ, the situation is different. Using sin(dϕ) → ϕ, cos(dϕ) → 1, we
obtain
Ri (dϕ) = I + dϕJi , (15)

where
 
0 0 0
J x = 0 0 1 (16)
 

0 −1 0
 
0 0 −1
J y = 0 0 0 (17)
 

1 0 0
 
0 1 0
J z = 0 −1 0 (18)
 

0 0 0

Q: Prove that Ri (dϕi )Rj (dϕj ) = Rj (dϕj )Ri (dϕi ).

Now we can define the infinitesimal rotation angle vector as dϕ = (dϕx , dϕy , dϕz ) and a vector
of matrices R = (Rx , Ry , Rz ) (and similarly for J ), so that the general infinitesimal rotation will be

r0 = r + (J · dϕ)r, (19)

where J · dϕ is an operator which acts on r.

Q: Prove that (19) can be written as dr = dϕ × r.

2 Advanced material: noninertial frames

The Galilean relativity (Newtonian mechanics) state that there is a class of special frames (inertial
frames) in which the motion laws have the same shape. What happens in noninertial frames and
why they are worse than inertial ones ? Actually, if we want that the physical laws be universal we
should find a way to describe these laws in any frame. We already know how to include inertia forces
in a uniformly accelerated frame and in a uniformly rotating frame. Now let us try to do the same
in an arbitrary frame. In order to do that we shall write down the second Newton law in an inertial

3
Physics 1. Mechanics Advanced 4

frame S(xi ) (where xi are cartesian coordinates) in the form

d2 r
m = F, (20)
dt2

where F is an external (“true”) force, caused by other bodies. Let the observer in a noninertial
frame S 0 also chooses cartesian coordinates x0i , and we know the relation r(r0 , t). The dependence on
t shows that this is not a simple coordinate transformation but the relation between the coordinates
in two frames with some relative motion. The only transformation between the two frames, which
conserves the scale, is the time dependent translation (this is not precise, we shall look into it more
closely later) and rotation:
r = X + Rr0 , (21)

where X(t) is the radius-vector of the coordinate origin of S 0 according to the observer S. Now one
has
dR 0
v=V+ r + Rv0 , (22)
dt
where V is the relative velocity of the two frames as measured by S. Differentiating once again with
respect to time we get
d2 R dR
a = A + 2 r0 + 2 v0 + Ra0 , (23)
dt dt
where A = dV/dt. Left multiplying by R−1 and taking into account (20) we have finally

d2 R 0 dR
ma0 = R−1 Fext − mR−1 A − mR−1 2
r − 2mR−1 v0 . (24)
dt dt

The first term in the right hand side of (24) is the true force measured in S 0 , the second term is
the generalization of the inertia force due to the relative motion of the coordinate origin of the two
frames, the third term is the generalized centrifugal force, and the last term is the generalized Coriolis
force. We see that there no other inertia forces.
Exercise: Derive the usual centrifugal and Coriolis force for A = 0 and
 
cos(ωt) − sin(ωt) 0
R =  sin(ωt) cos(ωt) 0 (25)
 

0 0 1

4
 Physics 1. Mechanics
Lecture 5. Particle dynamics I

1 Newton’s laws, momentum, kinetic energy

This section is to remind the basics of the particle dynamics. We shall proceed in an inertial frame,
so that there are no inertia forces. The second Newton law says that the equation is proportional to
the force
F = ma. (1)

Momentum p is defined as follows:

p = mv, (2)

which immediately gives the second Newton law in a more general form:

dp
F= . (3)
dt

Now the momentum change
t2

∆p = p2 − p1 = F(t)dt. (4)
t1

Kinetic energy is a scalar:

mv 2 mv · v
K= = . (5)
2 2
The rate of change of the kinetic energy

dK dv
= mv · = mv · a = (ma) · v = F · v. (6)
dt dt

During the time dt the energy change is

dK = F · vdt = F · dr = dW. (7)

The scalar quantity dW is called the infinitesimal work. If the particle moves from r1 to r2 the
kinetic energy change ∆K is equal to the work which is done along the trajectory:

r2
∆K = W = F · dr. (8)
r1

1
Physics 1. Mechanics Lecture 5

The last integral is taken along the trajectory. It, and therefore, the work depends, in general, on
the path (and not only on the initial and final points r1 and r2 ). The work produced in unit time
P = dW/dt is called power. For our case it is

P = F · v. (9)

Third Newton law states that if two particles interact then F1→2 = −F2→1 . Here F1→2 is the
force which particle 1 exerts on particle 2. This has to be completed by the superposition principle:
if there are a number of particles acting on particle i the total force which particle i experiences is
the vectorial sum of the forces from all other particles:

Fi = Fj→i . (10)
j

Two additional important concepts are the concept of a physical system which may include a number
of particles, and the concept of the resulting force acting on the whole system, which is the vectorial
sum of all the forces acting on all parts of the system. If the system consists of a number of particles,
the resulting force would be
 
F= (Fi,ext + Fj→i ), (11)
i j

where Fi,ext is the external force (that is, not by any particle from the same system) on particle i,
and Fj→i is the force exerted by particle j on particle i (both i and j belong to the physical system
under consideration). Using the third Newton law Fi→j = −Fj→i we obtain

F= Fi,ext , (12)
i

that is, the internal forces do not contribute to the resulting forced acting on the system as a whole.

2 Work as a path integral

Work is calculated differently from the integrals we have seen before. In the expression

2
W = F · dr
1

the integral is done along the curve. The force can be a function of r and t, in general, F = F(r, t)
and in this case the work can be calculated only if we know r(t) as follows

t2
W = F(r(t), t) · v(t)dt, (13)
t1

2
Physics 1. Mechanics Lecture 5

so that the integrand becomes a function of only one variable t.

In many cases, however, F = F(r) and is independent of t. If this is the case, the path (trajectory)
should be expressed in a parametric form.

2.1 Example: friction

We assume the a body is moving in a horizontal plane so that the friction force magnitude is constant
while it is always directed against the velocity F = −f v/v. Now F · vdt = −f vdt = −f ds, where
ds is the length of the path. Thus, W = −f s, where s is the length of the whole path made by the
particle.

2.2 Example: constant force

For a constant force F one has

2
2
W = F · dr = F · dr = F · (r2 − r1 )
1 1

and does not depend on the path but only on the initial and final positions. This is an example of a
conservative force for which work does not depend on the path. In other words, a conservative force
does not produce work along any closed path.

3 Momentum conservation, center of mass

The total momentum of the system is defined as the vectorial sum of the momenta of all its parts:
 
P= pi = mvi . (14)
i i

The change of the total momentum with time is given by dP/dt:

P d 
= pi
dt dt i
 dpi  (15)
= = Fi = Fext ,
i
dt i

where Fext is the sum of the forces caused by the bodies which do not belong to the system (external
forces). If this resulting external force Fext = 0 the total system momentum does not change
dP/dt = 0, that is, the momentum is conserved.

3
Physics 1. Mechanics Lecture 5

Let us now define the center-of-mass (CM) as the point with the radius-vector
 
i mi ri mi ri
Rcm =  = i , (16)
i mi M

where M is the total mass of the system. There does not have to be any particle in this point - it
is completely imaginary. Differentiating (16) with respect to time we find that CM is moving with
the velocity 
mi vi P
Vcm = = . (17)
M M
Thus, the physical sense of the center-of-mass is that this is a point which may be assigned the total
momentum and the total mass of the system. From (15) we find the acceleration of the center-of-mass
as follows
acm = Fext /M. (18)

In the absence of external forces the center-of-mass moves with constant velocity. If it was in the
rest in the beginning it would remain in the rest if only internal forces are present.
While the total momentum can be attributed to the motion of the center-of-mass, the kinetic
energy cannot. Indeed, the total kinetic energy
 mi
vi2

K= Ki =
i i
2
 mi
= [(vi − Vcm ) + Vcm ]2 (19)
i
2
 mi  mi
(vi − Vcm )2 + 2

= mi (vi − Vcm ) · Vcm + Vcm .
i
2 i i
2

The second term in the last expression of (19) vanishes (SHOW THAT !), and the first term can be
written in terms of the velocity relative to the center-of-mass ui = vi − Vcm , so that we get finally:

M 2  mi
K= Vcm + u2i . (20)
2 i
2

Thus, the total kinetic energy of the system is the sum of the kinetic energy of the center-of-mass
and the kinetic energy of the relative motion of the system parts. The last contribution is called
internal energy.

4 Center-of-mass: continuous systems

Most of the systems we deal with do not consist of point-like particles, but are continuous bodies
of finite size. The generalization onto such systems is straightforward, although not always simple
for calculations. We divide the whole body onto small (infinitesimal) parts with the mass dm,

4
Physics 1. Mechanics Lecture 5

radius-vector r , and velocity v , so that the sums are substituted by integrals:

M= dm, (21)


P= v dm, (22)


Rcm = r dm/M. (23)

The integrals should cover all material, they are written in a symbolical form. In practice, one has
always to express the integrals in terms of coordinates.

4.1 Center-of-mass: examples

A rod 1. A zero thickness rod 0 < x < l of the mass m (see Fig. 1).

O dx
x

Figure 1: Center of mass for a rod.

In the integral (23) dm = (m/l)dx (λ = dm/dx = m/l is the linear density), and we have

l
1 xmdx l
Xcm = = , (24)
m 0 l 2

as could be expected.
2. Same but the linear density is not constant λ = ax. Then
l
xλdx 2l
Xcm = 0 l = . (25)
λdx 3
0

5
Physics 1. Mechanics Lecture 5

A triangle A uniform density triangle has the vertices (0, 0), (l, 0), and (l, h), and the mass m.
The surface density σ ≡ dm/dA = m/A, where A = lh/2 is the triangle area. We divide the triangle
onto infinitesimal dxdy (see Fig. 2) so that for each of which dm = σdxdy. We have (for any general

dy
x
O dx

Figure 2: Center of mass for triangle.

surface density σ(x, y)


xσdxdy
Xcm =   , (26)
σdxdy

yσdxdy
Ycm =   , (27)
σdxdy

where the integral limits should be chosen according to the conditions. In our case



l 
hx/l

(. . .)dxdy → dy(. . .) dx, (28)
0 0

so that we have
 l  hx/l
( xσdy)dx
Xcm = 0 l 0 hx/l = 2l/3 (29)
(
0 0
σdy)dx
 l  hx/l
( yσdy)dx
Ycm = 0 l 0 hx/l = h/3. (30)
(
0 0
σdy)dx

A half-disk Given a uniform thin half-disk with the radius a (see Fig. 3). It is obvious that

6
Physics 1. Mechanics Lecture 5

Ycm = 0. For Xcm we use (26):

 a  √a2 −x2
( √ 2 2 xσdy)dx 4a
Xcm = 0 −√ a2 −x2 = . (31)
a a
√ −x 3π
0
( − a2 −x2
σdy)dx

5 Reactive motion
From the general form of the second Newton law F = dp/dt and p = mv it follows that the velocity
can change if the mass changes. A rocket is based on this principle. Let us consider one-dimensional
motion of a rocket which ejects backwards gas with the velocity u relative to the rocket. Let v be
the velocity of the rocket in the moment t and v + dv is its velocity in the moment t + dt. Let also m
be its mass in the moment t and m + dm is its mass in the moment t + dt (dm < 0 since the rocket
ejects gas). The mass of the ejected gas is −dm (so that the total mass (m + dm) + (−dm) = m
does not change), while the velocity of the gas is v − u (here v > 0 and u > 0). The momentum
conservation (no external force) gives

mv = (m + dm)(v + dv) + (−dm)(v − u) → dv = −udm/m. (32)

Differentiating the last equation with respect to time one gets the rocket acceleration

a = µu/m, (33)

where µ = −dm/dt is the rate of the gas ejection.

On the other hand, integrating (32) one gets

mi
vf − vi = u ln , (34)
mf

where i and f mean initial and final, respectively.

7
Physics 1. Mechanics Lecture 5

y
√
y = a2 − x2

dx
x
O

√
y = − a2 − x2

Figure 3: Center of mass for halfdisk.

8
 Physics 1. Mechanics
Math1. Complex numbers

In this lecture we bring (without proof) the basics of the usage of complex numbers, which is
necessary for a physicist.
A complex number is a construction (it is widely accepted use z for general complex numbers)
z = x + iy, where x and y are real numbers, and i is a special number such that i2 = −1. For any
two complex numbers z1 = x1 + iy1 and z2 = x2 + iy2 we can find z1 + z2 and z1 · z2 as follows:

z1 + z + 1 = (x1 + x2 ) + i(y1 + y2 ), (1)

z1 z2 = (x1 x2 − y1 y2 ) + i(x1 y2 + x2 y1 ). (2)

The absolute value of the complex number z = x + iy is

p
|z| = x2 + y 2 .

The number x is called the real part and the number y is called the imaginary part:

x = Re z, y = Im z.
p
Thus, z = Re z +i Im z, and |z| = (Re z)2 + (Im z)2 . A complex conjugate (cc) number z ∗ is defined
as z ∗ = Re z − Im z. It is easy to see that |z|2 = zz ∗ . Now the division of the complex numbers if
defined as follows:
z1 z1 z2∗ z1 z2∗
= =
z2 z2 z2∗ |z2 |2
Exercise: Express z1 /z2 in terms of x1 , y1 , x2 , y2 , i.
We can also write
Re z = (z + z ∗ )/2, Im z = (z − z ∗ )/2i.

Any complex number can be represented as a point in the complex plane, where x axis is for Re z
and y axis is for Im z. The same point in the x − y plane can be represented by its polar coordinates
r and ϕ, so that x = r cos ϕ, y = r sin ϕ. With this one can write

z = x + iy = |z|(cos ϕ + i sin ϕ). (3)

1
Physics 1. Mechanics Math1

This is completed with the (most important) Euler formula

cos ϕ + i sin ϕ = exp(iϕ) = eiϕ . (4)

With this one can write z = |z| exp(iϕ), z ∗ = |z| exp(−iϕ), 1/z = (1/|z|) exp(−iϕ), z1 z2 =
|z1 ||z2 | exp[i(ϕ1 + ϕ2 )], z n = |z|2 exp(inϕ), and so on.

2
 Physics 1. Mechanics
Lecture 6. Particle dynamics II

In this lecture we shall study several important cases of particle dynamics in electric and magnetic
fields.

1 Field
We say that an electric field E exists in the space if we can put an electric charge q (charged particle)
in any point and it will experience the force F = qE. The concept of the field is one of the central
concepts in physics. Since we shall deal with fields a lot in future it is worth to devote some tome to
this concept now.
We start with the example of the Earth and satellites around it. It is known that each two masses
are attracted to each other. For the time being the precise shape of the force is not important. Let
us say that the two bodies have masses M and m, M > m, and are in the positions R and r,
respectively. Then the force acting on, say m, is proportional to m and M and depends on R and r.
The same force but with the opposite direction acts on M. Although the magnitude of the force is
the same, the magnitudes of the accelerations are inversely proportional to masses, so that the larger
mass is less affected by this interaction. Let us assume for simplicity that M
m, so that the effect
of the interaction on M can be neglected. Then we can regard the large mass M as a “source” of the
force acting on the small mass m. Thus, we can take this small mass (which is called test particle
or test mass), put it in an arbitrary point of the space and measure the force acting on it. Since we
fixed the position and the mass of the large body, the only meaningful dependence of the measured
force would be on the mass of the test particle and its position. If we put different test particles in
the same point, each would feel the force proportional to its mass, that is, the force can be written
as F→m = mE, where E does not depend at all on the test particle mass m. If we take the same
test particle in two different points, say 1 and 2, it will experience forces F1 = mE1 , and F2 = mE2 ,
which are, in general, different. That means that the vector E depends, in general, on the position,
E = E(r). If the large body moves somehow, this vector can depend on time too, E = E(r, t). Let
us now forget about the body which causes E and look at everything from the point of view of the
test particle. The test particle does know anything about what causes the forces, the only think it
knows is that in each point of space r and every moment t it would feel the force F = mE(r, t). The
test panicle (and we too) concludes that the vector E(r, t) exists in space even without presence of

1
Physics 1. Mechanics Lecture 6

this test particle. Indeed, any test particle could be placed in any position at any time, and the
relation F = mE(r, t) does not change because of our bringing the test particle into consideration.
Thus, E(r, t) is real, depends only on other bodies (source) and is independent of test measurements.
We say that there is a field E(r, t). This field exists in space in time and produced by other bodies
which are not of interest to us and the motion of which is usually assumed to be known. The mass m
which, being multiplied by the field, gives the force acting on the test particle, is called, in general, a
charge. When certain field is considered, the characteristic word should be added to make clear what
is studied. Thus, the field produced by the gravitational forces is called gravitational field and the
mass is the gravitational charge. In this lecture we will study the motion in the electric and magnetic
fields acting on the electric charge.

Fields in general Fields are not necessarily related to forces. In general, we say that there is
a field in space if some quantity can be measured at different positions and times, but the very
existence of which is not determined by such measurement. In the above consideration what was
measure was the force vector, thus the field was a vector field. A well-known example of another field
is temperature, which can be measured at any position and any time. Temperature is a scalar field
which exists even if we do not measure it. A “test particle” would be in this case a thermometer.

2 Motion in a constant electric field

When we consider a motion in a constant electric field we assume that the field vector E does not
depend on r and t, at least in that part of the space where our test particles moves. The electric
force is F = qE, where q is the electric charge of the particle (we omit the word “test” for brevity).
This case is not especially interesting since the second Newton law states

F = ma → ma = qE → a = qE/m = const,

that is, the particle moves with constant acceleration, the case we studied earlier.

3 Motion in a uniform oscillating field

A field is called uniform is it does not depend on r. It can still depend on time. In the section we
shall try the field of the form
E = E0 sin(ωt), (1)

where E0 = const is a constant vector (both the magnitude and direction does not change), and ω
is the frequency (or angular frequency). It is convenient to define a unit vector ê in the direction of

2
Physics 1. Mechanics Lecture 6

E0 , ê = E0 /|E0 |, so that E0 = E0 ê. The equation of motion reads

d2 r


dv qE0
a= = 2 = ê sin(ωt). (2)
dt dt m

It is convenient (this procedure will be widely used in future) to split the position vector r = xê + r⊥ ,
where r⊥ ⊥ ê. Then the equation of motion (2) splits into

d2 r⊥
= 0, (3)
dt2

d2 x

qE0
= sin(ωt). (4)
dt2 m

Eq. (3) describes the motion with zero acceleration (constant velocity), that is, we immediately know
the answer:
r⊥ = r⊥,0 + v⊥,0 (t − t0 ). (5)

Here v⊥,0 ⊥ ê. Thus, the particle moves with the constant velocity v⊥,0 in the direction perpendicular
to the electric field direction (!!! it is only a part of the motion !!!).

Frame transformation. Here we introduce a very important method of studying particle dy-
namics. We should remember that it is always possible to switch to another inertial frame, and
the second Newton law will not change the form. If we now switch to the frame moving with the
velocity v⊥,0 with respect to our original frame, we will see the particle not moving in the direction
perpendicular to the electric field. In fact, any motion with constant velocity can be excluded from
the consideration by switching to the corresponding moving inertial frame. (WARNING: we have
to know, in general, how the fields transform with the frame transformation. For the time being,
we shall leave this issue aside, assuming that our electric field remains the same.) Thus, in our case
we may say that the particle does not move in the perpendicular direction in the properly chosen
reference frame.
With all above, the only equation of motion to be solved is the equation (4) for single coordinate
in the direction of the electric field. This equation is easily integrated to give

qE0
vx = vx0 + (1 − cos(ωt)) , (6)

mω 
qE0 qE0
x = x0 + vx0 t + t− sin(ωt), (7)
mω mω 2

where we added the initial conditions x = x0 , vx = vx0 at t = 0. We see that the motion in x direction
is the motion with the constant velocity vd = vx0 + qE0 /mω and periodic oscillations. Again we can
switch to the reference frame moving in x direction with the velocity vd . In this frame the particle
only oscillates with the frequency ω.

3
Physics 1. Mechanics Lecture 6

4 Constant magnetic field

The electric and magnetic field are the two parts of the electromagnetic field, and act on electric
charges. However, while the electric field acts on any charges, the magnetic field acts only on moving
charges. The magnetic force looks as follows (SI units)

F = qv × B, (8)

where B is the magnetic field and v is the particle velocity.

[In the Gaussian (CGS) system of units magnetic and electric fields are measured in the same
units and the force is written as F = qv × B/c, where c is the speed of light. In our course we shall
use SI units throughout.]
The force (8) is perpendicular both to the velocity and magnetic field direction. As an obvious
consequence, the magnetic force cannot change the kinetic energy of the particle, that is, the magnetic
force does not produce work:

P = dW/dt = F · v = q(v × B) · v = 0.

Since the kinetic energy K = mv 2 /2 does not change, the magnitude of the particle velocity v = |v|
remains constant.
Let is now follow the principles of the previous section and split r = r⊥ + r (and similarly for the
velocity) where the indices ⊥ and  mean perpendicular and parallel to the magnetic field (remember:
the direction of the magnetic field is constant).
Exercise: Show that the equations of motion will take the form

dv
= 0, (9)
dt

dv⊥ qB
= − × v⊥ . (10)
dt m

Eq. (9) shows that the velocity in the direction of magnetic field does not change at all. We
already know that it is possible to move to another inertial reference frame in which this motion is
absent. Let us concentrate on the motion in the plane perpendicular to the magnetic field, which is
described by (10). Because of the importance of the case we shall do that in two ways.

4.1 Analogy
This is simple. We have only to remember the equation for circular motion with constant angular
velocity which reads (dv/dt) = ω × v. Comparing with (10) we see that the motion is the circular
motion, gyration, with the angular velocity ω = −qB/m, called gyrofrequency. Respectively, the
radius of gyration, gyroradius, is r⊥ = mv⊥ /|qB|.

4
Physics 1. Mechanics Lecture 6

4.2 Coordinate representation

Let B = Bêz , than (10) will be written as (we use the widely accepted notation Ẋ ≡ dX/dt for any
variable X):

v˙x = ωvy , (11)

v˙y = −ωvx , (12)
ω = qB/m. (13)

In order to proceed we shall use complex numbers. See mathematical

appendix Math1 as a separate lecture. Proceed only after having learnt
the material.
It is convenient to define vs = vx + isvy , where s = ±1. Multiplying the equation for vy by is and
adding to the equation for vx we get
v˙s = −isωvs , (14)

which has the solution in the form

vs = v⊥ exp(isωt + isφ), (15)

where v⊥ = const and φ = const, so that vx = v⊥ cos φ and vy = v⊥ sin φ at t = 0 . In order to get
vx and vy we have to use the Euler formula:

vs = v⊥ cos(ωt + φ) + isv⊥ sin(ωt + φ), (16)

or
vx = v⊥ cos(ωt + φ), vy = v⊥ sin(ωt + φ). (17)

5 Electric and magnetic field together

Here we consider the case where there are both constant electric E and magnetic B fields. The total
force (Lorentz force) is then F = qE + qv ×B. As before we split all vectors onto components parallel
(index ) and perpendicular (index ⊥) to the magnetic field. Let b = B/B be the unit vector in the
direction of the magnetic field, then we shall write r = zb + r⊥ , v = v b + v⊥ , and E = E b + E⊥ .
The equation or motion reads
q
v̇ = (E + v × B) →
m (18)
q q
v̇ b + v̇⊥ = E b + [E⊥ + v⊥ × B].
m m

5
Physics 1. Mechanics Lecture 6

Separating parallel and perpendicular components we obtain

q
v̇ = E , (19)
m
q
v̇⊥ = (E⊥ + v⊥ × B. (20)
m

We see that (19) contains only parallel components, while (20) contains only perpendicular ones.
This allows to study separately the motion along () and across (⊥) the magnetic field.
The motion along the magnetic field is the simple constant acceleration motion (see (19)) and
the analysis remains as an exercise for the reader.
In order to study the motion across the magnetic field, eq. (20) it is convenient and instructive to
switch to another frame which is moving with some velocity Vd ⊥ B. We shall choose this frame so
that the equations of motion reduce to what we already know. Let us denote the particle velocity in
the new frame as u. According to the rules of Galilean transformation, v⊥ = Vd + u. Substituting
into (20) and taking into account that Vd is constant, we obtain

q
u̇ = (E⊥ + Vd × B) + ω × u, (21)
m

where ω = −qB/m. The first term in parentheses is constant. If it were absent we would obtain the
equation identical to the equation of motion for a particle in the magnetic field, without any electric
field. Now we can choose the new frame velocity to eliminate this term. We require

E ⊥ + Vd × B = 0 (22)

from which it immediately follows

E⊥ × B
Vd = . (23)
B2
This can be done only if E⊥ /B < c where c is the speed of light !. The obtained velocity is called the
drift velocity. In the frame moving with this velocity, the particle is gyrating around the magnetic
field, as if the electric field was absent (in the course Physics 2 we shall see that the electric and
magnetic fields change when switching frames, and the electric field indeed vanished for the frame
transformation that we did). Thus, we can conclude that the particle gyrates around the magnetic
field and in addition drifts with the constant velocity Vd . This drift is perpendicular to the electric
and magnetic field as well.

6
Physics 1. Mechanics Lecture 6

Figure 1: Gyration, drift, and motion along the magnetic field.

7
 Physics 1. Mechanics
Advanced 6. Perturbation theory in particle dynamics

Not all problems in physics can be solved exactly. In fact, only very few problems have been solved
exactly and not many can be solved exactly. Thus, approximate methods play a very important role
in physics. Here we shall study two simple examples of such methods.

1 General principles of perturbation theory

Let us assume that the equation of motion can be written in the form

mẍ = F0 (x, t) + F1 (x, t), (1)

where  is a symbolic small parameter, that is, shows that the addition F1 is small relative to the
main force F0 . Let us also assume that we know the solution x = x0 (t) of the unperturbed equation

mẍ = F0 . (2)

We proceed assuming that the perturbation F1 perturbs the solution only weakly, that is, that the
solution of the complete equation (1) can be presented as x = x0 (t)+x1 (t), where  again symbolizes
the smallness of the change. It is important that it is the same  as in the force perturbation.
Substituting into (1) we get

m(ẍ0 + ẍ1 ) = F0 (x0 + x1 , t) + F1 (x0 + x1 , t)

dF0 (3)
≈ F0 (x0 (t), t) + x1 + F1 (x0 (t), t)
dx |x=x0 (t)

and therefore
dF0
mẍ1 = x1 + F1 (x0 (t), t). (4)
dx |x=x0 (t)
The new equation is usually easier to solve than the complete one. Although the above is written
for a one-dimensional motion for simplicity, it can be two- and three-dimensional as well.

1
Physics 1. Mechanics Advanced 6

2 Example 1: Magnetic field

Let us start with the particle gyration in the constant magnetic field

mv̇x = qvy Bz , (5)

mv̇y = −qvx Bz , (6)

with the solution

v0,x = v⊥ cos(ωt), (7)

v0,y = −v⊥ sin(ωt), (8)

where ω = qBz /m (Bz is assumed constant). Now let us add perturbation in the form of the electric
force F1,x = 0, F1,y = qEy . Following the method outlined above we write vx = v0,x + v1,x and
vy = v0,y + v1,y . Substituting this into the equations of motion we obtain

mv̇1,x = qv1,y Bz , (9)

mv̇1,y = −qv1,x Bz + qEy , (10)

which has the special solution of the form v1,x = Ey /Bz , v1,y = 0. We rediscovered the E × B drift.
This example is simple but not very informative since we know how to solve the equations exactly
in this case.

3 Example 2: Wave field

Let us now solve the one-dimensional equation of motion for the force of the kind F = F1 cos(kx),
where k is a constant parameters. (Actually, in this case the problem can be solved exactly but the
situation changes if there are two forces of the kind Fi cos(ωi t − ki x), i = 1, 2, where ω2 /k2 6= ω1 /k1 .
The method of solution in this case is the same as the method we are going to learn now.) The solution
of the unperturbed equation ẍ = 0 is x0 (t) = v0 t (for simplicity we ignore the initial coordinate).
Substituting this into the force we have for the perturbation (x = x0 + x1 )

mx¨1 = F1 cos(kv0 t) (11)

and the solution can be found by direct integration:

F1
v1 = sin(kv0 t), (12)
mkv0
F1
x1 = − 2 2 cos(kv0 t). (13)
mk v0

2
Physics 1. Mechanics Advanced 6

This approximation is valid if the perturbation is weak, that is, max|v1 |  v0 , which gives

F1
1
mkv02

Let us now notice that for the potential force F1 cos(kx) the potential U = −(F1 /k) sin(kx) therefore
(F1 /k) has the meaning of the maximum potential energy. More physically speaking, F1 /k is of
the order of the typical potential energy F1 /k ∼ U1 (this is not a precise statement but an estimate
which is one of the most powerful instruments of a physicist). In the same sense mv02 ∼ K, thus
the applicability criterion reads U1 /K  1, where U1 is the potential energy of the particle in the
perturbed state and K is the kinetic energy of the particle in the unperturbed state. The small
parameter  becomes now a quite real parameter  ∼ U1 /K.

4 Averaging
Let us consider a particle moving in the x − y plane perpendicular to the magnetic field, only in this
case the magnetic field will be allowed to be weakly inhomogeneous, Bz = Bz (x). Later we shall see
what does it mean: “weakly”. If the magnetic field was constant the motion would be a gyration,
circular motion around the magnetic field,

vx = v⊥ cos(ωt), (14)
vy = −v⊥ sin(ωt), (15)
x = x0 + r⊥ sin(ωt), (16)
y = y0 + r⊥ cos(ωt), (17)

where the gyrofrequency ω = qBz /m, and gyroradius r⊥ = v⊥ /ω. Because of the magnetic field
inhomogeneity the particle “feels” different magnetic fields in different points of its circular trajectory
which eventually would distort the circle. If the magnetic field does not change substantially from
one point of the circle to another, that is, |Bz (x0 + r⊥ ) − Bz (x0 − r⊥ )|  |Bz (x0 )| we may Taylor
expand to obtain
dBz
Bz (x) = Bz (x0 ) + (x − x0 ) . (18)
dx |x=x0
Now we see that the condition for weak inhomogeneity is

dBz
r⊥  Bz , (19)
dx

that is, the magnetic field does not change significantly on the gyroradius. If we estimate dBz /dx ∼
Bz /L, where L is the typical length of the magnetic field inhomogeneity then the condition will be
L  r⊥ .

3
Physics 1. Mechanics Advanced 6

Coming back to the equations of motion we shall now add the velocity perturbation v → v + v1
and write the magnetic field as Bz = B0 + (x − x0 )(dBz /dx) where  ∼ r⊥ /L is the small parameter
of the problem. The equations of motion read

dBz
mv̇1,x = qv1,y B0 + qvy (x − x0 ) (20)
dx
dBz
= qv1,y B0 − qv⊥ sin(ωt)r⊥ sin(ωt) , (21)
dx
dBz
mv̇1,y = −qv1,x B0 − qvx (x − x0 ) (22)
dx
dBz
= −qv1,x B0 − qv⊥ cos(ωt)r⊥ sin(ωt) . (23)
dx

Instead of trying to solve these equations we shall try to find out if there is some average velocity
changes, that it, whether there is a velocity addition which is not periodic with the period T = 2π/ω.
In order to do that we perform averaging according to the recipe
Z t+T
1
v¯x = vx (t0 )dt0 (24)
T t

¯ = 0)
which immediately gives (obviously, dv/dt

2
v⊥ 1 dBz
v̄1,y = , v̄1,x = 0. (25)
2ω Bz dx

This is the so called gradient drift which is perpendicular to the direction of the magnetic field (z)
and to the direction of the magnetic field change (x). We may regard the particle motion as a rapid
gyration with local gyrofrequency ω = qBz (x) and local gyroradius, while the center of the circle
slowly drifts along y direction with the velocity given by (25).

4
 Physics 1. Mechanics
Lecture 7. Particle dynamics III

1 Potential energy

2
Let F(r) be a conservative force. This means that for each two points r1 and r2 the work W = 1
F·dr
does not depend on the path but only on the initial and final points themselves. The minimum work
that we have to do against this force to move a particle from r1 to r2 is
 r2
Wext = −W = − F · dr.
r1

Let us now fix some point r0 as a reference point and consider the minimum work we have to do to
move a particle from this reference point to any other point r. This work depends only on r0 and r
so that we can write  r
U(r) − U(r0 ) = − F · dr. (1)
r0

The function U(r) is called a potential energy, and the difference U(r) − U(r0 ) what work the force
does when moving a particle from r to r0 or what minimum work an external force must do to move
particle from r0 to r. From this analysis it is clear that there is no physical meaning to the value
of the potential energy in the reference point, since only the difference is meaningful. The reference
point is usually chosen from the conveniency arguments. For example, is a particle is very far from
bodies the interaction is negligible, and it is natural to choose U0 = 0.
Let us now consider two close points, r and r + dr. According to (1) we have

dU = U(r + dr) − U(r) = −F · dr = −(Fx dx + Fy dy + Fz dz). (2)

Thus, if we know the potential energy as a function of position r we can find the force at r as follows:

∂U
Fx = − ,
∂x
∂U
Fy = − , (3)
∂y
∂U
Fz = − .
∂z

1
Physics 1. Mechanics Lecture 7

For brevity this is usually written in a symbolical form as follows:

∂U
F=− , (4)
∂r

or

F = −∇U, (5)

where the special symbol

∂ ∂ ∂
∇ = ex + ey + ez (6)
∂x ∂y ∂z
denotes the differential operator, that is, the partial derivatives act on everything which stands to
the right or them. The construction ∇U is called a gradient of U.

1.1 Examples
Here we provide several simple examples showing how force is obtained from potential energy.

1. Potential energy related to a constant vector. U = f · r → F = −∂U/∂r = −f, that is,

the force is constant.

2. General dependence on distance. U = U(R), where R = |r − r0 |). Following the same

procedure, we find
∂U ∂R
F=− .
∂R ∂r
The last can be easily found it we take into account that

R2 = (r − r0 ) · (r − r0 ).

Differentiating the last expression with respect to r we have

∂R
R = r − r0 ,
∂r

or
∂R r − r0 r − r0
= = = (r

− r0 ),
∂r R |r − r0 |

where (r

− r0 ) is the unit vector in the direction of r − r0 . Thus,

∂U

F=− (r − r0 ).
∂R

2
Physics 1. Mechanics Lecture 7

3. Coulomb potential. The function U = A/R is called Coulomb potential. Following the
previous derivation we find
A(r − r0 )
F=− .
|r − r0 |3

4. Harmonic oscillator. The function U = k1 x2 /2 + k2 y 2 /2 + k3 z 2 /2 describes three-dimensional

harmonic oscillations. The corresponding force is given by (3):

F = −k1 xex − k2 y ey − k3 z ez .

2 Conservative (potential) forces

We have seen that if a force is conservative (work does not depend on path) a potential energy can be
defined. Thus, we can call such forces potential forces too. Let F(r) be some force (more precisely,
force field since it depends on r). In order to find out whether this force is potential or not, it is
not necessary to calculate integrals to show that they are independent of trajectory. Instead it is
possible to use the relations (3) and the identities
   
∂ ∂U ∂ ∂U
=
∂xi ∂xj ∂xj ∂xi

for any i and j. For example, from

   
∂ ∂U ∂ ∂U
=
∂x ∂y ∂y ∂x

we find

∂ ∂
Fy = Fx (7)
∂x ∂y

and similarly

∂ ∂
Fz = Fy (8)
∂y ∂z
∂ ∂
Fx = Fz (9)
∂z ∂x

3
 Physics 1. Mechanics
Advanced7. Differential operators

0.1 Advanced material

In a more compact form we can write
∂U
Fi = − , (1)
∂xi
and the condition of potentiality looks as follows:
X ∂Fk X ∂U
εijk =− εijk = 0. (2)
jk
∂xj jk
∂xj ∂xk

Using the vector operator ∇ the right hand side of (2) can be written as ∇ × F and the whole
relation takes the form
∇ × F = ∇ × (−∇u) = −(∇ × ∇)U = 0. (3)

The last relation becomes obvious since for any vector A the vector product A × A = 0.
This shows that ∇ can be treated as an ordinary vector. One has to remember, though, that it
is an operator acting on everything on the right, so that the order is important. Let us consider, for
example, the construction
∇ × (B × C)

If ∇ was an ordinary vector, it could written as follows (INCORRECT !):

∇ × (B × C) = B(∇ · C) − C(∇ · B).

However, in the left hand side ∇ always acts on B and C, while in the right hand side it is not so.
Therefore, the correct form will be

∇ × (B × C) = (∇ · C)B − (∇ · B)C,

where ∇ further acts as a derivative. Taking into account (xy)0 = x0 y + xy 0 we find eventually

∇ × (B × C) = B(∇ · C) + (C · ∇)B − C(∇ · B) − (B · ∇)C,

P
where B · ∇ = i Bi (∂/∂xi ).

1
Physics 1. Mechanics Advanced7

1 Differential operators: More systematic approach

We have defined above ∇ as a differential operator. Depending to what it is applied and how this
operator has different names. When applied to a scalar field U (r) it gives a vector field called gradient

∂U ∂U ∂U
A = grad U = ∇U = x̂ + ŷ + ẑ (4)
∂x ∂y ∂z

When it is applied to a vector field it can give a scalar field called divergence:

∂Ax ∂Ay ∂Az

V = div A = + + (5)
∂x ∂y ∂z

When applied to a vector field it can give a vector field called rotor or curl :
 
∂Az ∂Ay
B = rot A = − x̂
∂y ∂z
 
∂Ax ∂Az
+ − ŷ (6)
∂z ∂x
 
∂Ay ∂Ax
+ − ẑ
∂x ∂y

Differential operators can be applied more than once. Indeed, if we apply grad to a scalar field U
we get a vector field to which div or rot can be applied, which we can write as div grad U . Or rot can
be applied giving rot grad U . If we apply div to a vector field A we get a scalar field to which grad
can be applied giving grad div A. Finally, if rot is applied to a vector field A we get again a vector
field an div can be applied to give div rot A.
Exercise: Prove

∂2U ∂2U ∂2U

div grad U = + + ≡ ∆U, (7)
∂x2 ∂y 2 ∂y 2
div rot A = 0, (8)
rot grad U = 0. (9)

2
 Physics 1. Mechanics
Lecture 8. Conservation laws

1 Energy conservation
1.1 Simple derivation
We already know that the change of the kinetic energy K = mv 2 /2 is caused by work, that is
Z 2
K2 − K1 = W = F · dr. (1)
1

If the force is conservative, the right hand side can be written as U1 − U2 , and we have

E = K1 + U1 = K2 + U2 = const, (2)

which is the simplest form of the energy conservation. The sum E = K + U is called mechanical
energy.

1.2 More complicated (but more informative) derivation

Let us assume that the force is conservative (potential ) and can be written as F = −∂U/∂r, where
the potential energy U (r, t) can, in principle, explicitly depend on time t and must depend on r. The
rate of change of the kinetic energy K = mv 2 /2 along the particle trajectory is

dK dv dr ∂U dr
= mv · =F· =− . (3)
dt dt dt ∂r dt

Let us now notice that the total time derivative of U is

dU ∂U dr ∂U
= + , (4)
dt ∂r dt ∂t

and substitute into (3):

d ∂U
(K + U ) = (5)
dt ∂t
We see that the mechanical energy E = K + U is conserved, dE/dt = 0, if the potential energy does
not depend on time, ∂U/∂t = 0.

1
Physics 1. Mechanics Lecture 8

1.3 Nonconservative forces

If the total force can be decomposed into the conservative and nonconservative components,

∂U
F=− + Fn
∂r

the energy equation takes the following form

Z 2
E2 − E1 = Fn · dr, (6)
1

where E = K + U , or, in the local form

dE
= Fn · v (7)
dt

2 Momentum conservation
A particle momentum is not conserver if there is a potential energy U (r), since the conservation
requires F = −(∂U/∂r) = 0.

3 Angular momentum conservation

Angular momentum is defined as follows:

J=r×p (8)

Let us assume that the acting forces are potential, then the rate of change of angular momentum is

dJ
=v×p+r×F=r×F (9)
dt

since the first term vanishes identically. The construction N = r × F is called a momentum of force
or torque. Thus, in order that the angular momentum be conserved, dJ/dt = 0, the force should be
always parallel to r. In the previous lecture we have seen that ∂U/∂r k r is the potential energy
depends only on r = |r|. It is possible to show that this is the sufficient and necessary condition for
angular momentum conservation.

4 Interaction
In all above we considerer a particle moving under influence of a external force, that is, all other
bodies were excluded from our consideration. It is time to consider the interaction between bodies,

2
Physics 1. Mechanics Lecture 8

so that all they are included in the consideration. It appears that it is sufficient to study a two-
particle system, since the superposition principle states that interactions in a many body system can
be reduced to the interactions between all possible couples, that is, a many body system “consists”
of two-particle systems.
Thus, let us consider a system consisting of two particles, 1 and 2, with no other forces acting
on these particles. If particle 1 exerts a force F1→2 on particle 2, then particle 2 also exerts a
force F2→1 = −F1→2 on particle 1. We say that the two particles interact. If one of the forces is
conservative, the other one is conservative too. From the point of view of particle 1 it has a potential
energy U (r1 ), where r1 is its position vector. If the distance between the particles is infinite they
should not feel each other, and the potential energy should vanish. Thus, the potential energy of the
particle 1 can be calculated by fixing the particle 2 in the point r2 and bringing the particle 1 from
infinity into the point r1 .
On the other hand, from the point of view of particle 2 it has a potential energy U (r2 ), and this
potential energy can be calculated by fixing the particle 1 in the point r1 and bringing the particle
2 from infinity into the point r2 . In both cases we come to the same configuration (particle 1 in
r1 and particle 2 in r2 ) so that the result should be the same. With all this and noticing that this
potential energy does not exist when one of the particles is absent we come to the conclusion that
the potential energy is the energy of interaction and belongs to the system of the two particles, so
that the total mechanical energy is E = K1 + K2 + U12 (r1 , r2 ). This, in particular, means that

∂U ∂U
F2 ≡ F1→2 = − , F1 ≡ F2→1 = −
∂r2 ∂r1

Let us check the infinitesimal change of the mechanical energy E during time dt:

dE = dK1 + dK2 + dU12

∂U ∂U
= mv1 · dv1 + mv2 · dv2 + · dr1 + · dr2
∂r1 ∂r2 (10)
= mv1 · a1 dt + mv2 · a2 dt − F1 · dr1 − F2 · dr2
= F1 · (v1 dt) + F2 · (v2 dt) − F1 · dr1 − F2 · dr2 = 0,

that is, the total mechanical energy is conserved. It is easy to generalize this derivation onto systems
P
with arbitrary number of particles. We conclude that total mechanical energy E = i Ki + U is
conserved. Here U is the potential energy of the system and depends on all ri . In this derivation we
assumed that the system is closed, that is, there are no other bodies which interact with the parts
of out system and which do not belong to the system.

3
Physics 1. Mechanics Lecture 8

5 Momentum conservation
It is possible to show that the third Newton law means that the potential energy of the interaction
of two particles depends only on the difference r2 − r1 . This is the manifestation of the homogeneity
of the space. In other words, it means that we can arbitrarily shift the coordinate origin, so that
r01 = r1 + R, and r02 = r2 + R, where R is some constant vector, and the potential energy will
not change (since r02 − r01 = r2 − r1 ). The inverse statement is also correct: if the choice of the
coordinate origin does not affect the potential energy, it must depend only on the difference. Thus,
U = U (mbr2 − r1 ), so that F1 = −(∂U/∂r1 ) = (∂U/∂r2 ) = −F2 . The change of the total momentum
is
dP = dp1 + dp2 = F1 dt + F2 dt = 0,

that is, the total momentum is conserved. In this derivation the system is also assumed to be closed.

6 Angular momentum
The vector product J = r × p is called angular momentum. The importance of this quantity is
that it is also conserved, if the potential is isotropic, that is, the interaction does not depend on the
direction. For two particles this means that the potential energy U = U (X), where X = |r2 − r1 |,
that is, depends only on the distance between the two particles, and does not depend on the direction
of the vector r2 − r1 . Indeed,

dJ = dJ1 + dJ2
= dr1 × p1 + r1 × dp1 + dr2 × p2 + r2 × dp2
[dr = vdt||p → dr × p = 0] (11)
= r1 × F1 dt + r2 × F2 dt
 
∂U ∂U
= −r1 − r2 dt.
∂r1 ∂r2

Now we use

∂U ∂U r1 − r2
= ,
∂r1 ∂X X
(12)
∂U ∂U r2 − r1
= .
∂r2 ∂X X

4
Physics 1. Mechanics Lecture 8

Substituting (12) into (11) we obtain

 
r1 − r2 r2 − r1 ∂U
dJ = −r1 × − r2 × dt
X X ∂X
(13)
(r1 − r2 ) × (r1 − r2 ) ∂U
=− dt = 0.
X ∂X

Thus, angular momentum of a closed system is conserved.

7 Open systems
Open are systems which are under influence of some external bodies. In such systems energy, mo-
mentum, and angular momentum are not conserved and their changes should be determined by the
external forces.

7.1 Single particle

Single particle under the influence of external force is an open system. We already know that

dp
= Fext
dt

and
dK
= Fext · v.
dt
If the external force Fext is conservative, that is, Fext = − ∂U
∂r
(U does not depend on time) then the
second relation can be written as dE/dt = 0 where E = K + U is the particle energy.
For the angular momentum we get

dJ dr dp
= ×p+r× =v×p+r×F
dt dt dt

Since v k p only the last term survives so that

dJ
= r × Fext ≡ N
dt

where the right hand side is called momentum of force or torque.

5
Physics 1. Mechanics Lecture 8

7.2 Many particle system

We have shown that the total momentum of the system can be expressed in terms of the total mass
and the velocity of the center-of-mass:
X
P= pi = M Vcm
i

We have already shown that the total energy if the kinetic energy of the center-of-mass plus the
internal energy
2
X M Vcm X mi
K= Ki = + (vi − Vcm )2 .
i
2 i
2

In a similar way for the angular momentum we get

X X
J= Ji = ri × pi
i i
X
= [(r − Rcm ) + Rcm ] × pi (14)
i
X
= (r − Rcm ) × pi + Rcm × P.
i

Thus, the angular momentum of the system is the angular momentum of the center of mass Rcm × P
plus the internal angular momentum, that is, the angular momentum of the parts of the system with
respect to the center-of-mass. This part of the angular momentum is often called spin:
X
Jint = (r − Rcm ) × pi
i

Now for a system of particles one has

dP X
= Fext = Fi,ext .
dt i

For energy we would obtain

2
 
d M Vcm
= Fext · Vcm
dt 2
and
dEint X
= Fi,ext · (vi − Vcm )
dt i

where Eint is the internal energy, which is now the sum of the kinetic energy of the motion of the
system parts relative to the center-of-mass and the potential energy of the interaction between the
particles of the system.

6
Physics 1. Mechanics Lecture 8

For angular momentum we obtain

d
(Rcm × P) = Rcm × Fext
dt

and
d X X
Jint = (vi − Vcm ) × pi + (ri − Rcm ) × Fi .
dt i i

Q: Show that the first term in the right hand side is identically zero, so that

d X
Jint = (ri − Rcm ) × Fi ≡ Ncm .
dt i

Here Ncm is the torque relative to the center of mass.

8 Collisions
A collision is a kind of interaction where the time of interaction is negligible, that is, what is of
interest is the initial and final states. During a collision the momentum and angular momentum are
conserved (since only internal forces are working), while the kinetic energy may be converted into
the internal energy and vice versa. The collision where the kinetic energy is conserved is called an
elastic collision. If the kinetic energy is not conserved the collision is called inelastic. During inelastic
collisions the kinetic energy often dissipates to plastic deformations. An example of the collision with
the conversion of the internal energy into kinetic energy is explosion. Here we shall consider a special
case of the elastic collision of two particles. We shall call the lab frame the reference frame where we
see one particle, m1 , v1 , colliding with another particle m2 at rest, v2 = 0. In this case the collision
is often called scattering of the particle 1 on the particle 2.
The momentum and energy conservation give (primed quantities are after the collision)

p1 + p2 = p01 + p02 , (15)

p21 p2 p0 21 p0 22
+ 2 = + . (16)
2m1 2m2 2m1 2m2

Our objective is to find as much information about p0 1 and p0 2 if we know p1 and p2 .

Let me remind you that in the lab frame p2 = 0. It is extremely convenient to work with two
frames, the lab frame and CM frame (center-of-mass frame). The latter moves with the velocity
Vcm = (p1 + p2 )/(m1 + m2 ) = p1 /(m1 + m2 ) relative to the lab frame. In what follows we denote the
particle velocities in the lab frame as v1 , v2 , v0 1 , v0 2 , and in CM frame as u1 , u2 , u0 1 , u0 2 . According
to the velocity transformation rule we have v = u + Vcm and similarly for all other velocities in both
frames. Therefore, if we know the initial and final velocities in CM we can easily find them in the
lab frame.

7
Physics 1. Mechanics Lecture 8
Physics 1. Mechanics Lecture 8

Figure 1: Comparison of what is measured in the laboratory frame (left) and center-of-mass frame
Figure 1: Comparison of what is measured in the laboratory frame (left) and center-of-mass frame
(right).
(right). Attention: p2 6= 0 in this figure for better visualization, despite our definition of the
laboratory system.
In CM frame the total momentum should vanish (according to the definition of CM frame),
so that m1u1 + m2u2 = m1u! 1 + m2 u!2 . Energy conservation requires that m1u2 /2 + m2u2 /2 = 1
In CM frame the total momentum should vanish (according to the definition of CM2 frame),
m1 u!21 /2 + m2u! 22 /2. It is easy to see that the only way to satisfy this two conditions is to rotate
so that m1 u1 + m2 u2 = m1 u0 1 + m2 u0 2 . Energy conservation requires that m1 u21 /2 + m2 u22 /2 =
both vectors u1 and u2 by the same angle, while their magnitudes do not change. Moreover, all four
m1 u0 21 /2 + m2 u0 22 /2. It is easy to see that the only way to satisfy this two conditions is to rotate
vectors u, the vector Vcm , and the vectors v remain in the same plain. For convenience we choose
bothcartesian
vectors u 1 and u2 by
coordinates so the
thatsame
v1 = angle, while their
v1 eˆx (remember magnitudes
that do not
v2 = 0). Then Vcmchange.
= (m1v1Moreover,
/(m1 + m2)) alleˆfour
x,
vectors u, the vector
and u1 = v1 − Vcm =cm V , and the vectors v remain in the same plain. For convenience
(m2v1/(m1 + m2))eˆx . Respectively, u2 = −m1 u1/m2 = −(m1v1/(m1 + m2))eˆx . we choose
cartesian
If aftercoordinates
the collisionsoboth
thatvelocity eˆx (remember
v1 = v1vectors that rotate
in CM frame v2 = 0). Then
by the Vcm
angle θcm=, (m 1 vwe
then 1 /(m 1 + m2 ))eˆx ,
have
and u1 = v1 − Vcm = (m2 v1 /(m1 + m2 ))eˆx . Respectively, u2 = −m1 u1 /m2 = −(m1 v1 /(m1 + m2 ))eˆx .
m2 v1
If after the collision both velocity vectors u!1x in
= CM frame cosrotate
θcm , by the angle θcm , then we have(17)
m +m
1 2
m2 v1
u!
0 1y
= m2 v1 sin θcm , (18)
u1x = m1 + m2 cos θcm , (17)
m1 +mm 1 v21
u!2x = −m cos θcm , (19)
0 m21v1+ m2
u1y = m1v1 sin θcm , (18)
u!1y =m−1 + m2 sin θcm . (20)
0 mm v1m2
1 1+
u2x = − cos θcm , (19)
m1 + m2
Using v! = u + Vcm we find m1 v1
u01y =− sin θcm . (20)
m1 + m2
! m1 + m2 cos θcm
v1x = v1 , (21)
Using v0 = u + Vcm we find m1 + m2
! m2 sin θcm
v1y = v1 , (22)
0 mm +m
1 1+ m2 2cos θcm
v1x! = m2 − m1 cos θcm v1 , (21)
v2x = m1 + m2 v1 , (23)
m m1θ + m2
sin
0 2 cm
v1y!= v1 ,
m1 sin θcm (22)
v1y =m− 1 + m2 v 1 . (24)
m1 + m2
0 m2 − m1 cos θcm
v2x = v1 , (23)
m1 + m2
0 m1 sin θcm
v1y =− v1 . (24)
m1 + m2
8

8
Physics 1. Mechanics Lecture 8

Now the angle between the initial velocity v1 and final velocity v2 is given by
0
v1y m2 sin θcm
tan θ = 0
= . (25)
v1x m1 + m2 cos θcm

The obtained expressions fully describe the collision (scattering) with the use of only one free pa-
rameter θcm . It is impossible to find this parameter from the conservation laws, and this angle is
arbitrary, unless some additional conditions are imposed.

9
 Physics 1. Mechanics
Lecture 9. Motion in a potential, I. 1D and central forces

1 Potential energy, symmetries, and conservation

In what follows we shall study the motion of a particle under the potential force, that is, the case
when the force can be written as

∂U
F = −∇U = − grad U = −
∂r

(all forms are the same and we shall use the one which is more convenient). The mechanical energy
of the particle, E = K + U = mv 2 /2 + U . In general, the potential energy is a function of all
three coordinates R = (x, y, z) and time t and the energy E, momentum p = mv and the angular
momentum J = r × p are not conserved. However, if the potential does not depend on one or several
coordinates and/or time, some of the mentioned quantities are conserved.
Let us assume that U does not depend on t. Then

dE ∂E dv ∂E dr ∂E
= · + · +
dt ∂v dt ∂r dt ∂t (1)
∂U
= mv · a − F · v + = 0,
∂t

that is, the energy is conserved. In what follows we restrict ourselves with the cases where the
potential energy is time-independent, that is, energy is conserved, E = const.
Let us assume now, that the potential energy (or, for brevity, potential ) does not depend on one
coordinate, say x, that is, ∂U/∂x = 0. Then

dpx ∂U
= Fx = − =0
dt ∂x

which means that momentum component in x direction is conserved, px = const. This, in turn,
means that the velocity in this direction vx = px /m = const and the motion along x-axis is trivial
so that we can ignore this coordinate at all (such coordinate is called ignorable).
Let us assume now that the potential energy depends only on the length r = |r|, U = U (r). We

1
Physics 1. Mechanics Lecture 9

have seen earlier, that the force can be written as follows

∂U dU ∂r dU
F=− =− = − r̂
∂r dr ∂r dr

where r̂ = r/r is the unit vector in the direction of r and we substituted ∂U/∂r → dU/dr since U
now depends only on one variable and there is no sense in partial derivatives. Then for the angular
momentum we have
dJ
=r×F=0
dt
so that the angular momentum J = const.
Presence of conserved quantities (so called integrals of motion) greatly simplifies analysis of
particle motion.

2 One-dimensional potential
In this section we consider the potentials of the form U = U (x). According to what is found above,
y and z are ignorable coordinates (py and pz do not change and we can ignore them at all) and the
energy E = mvx2 /2 + U (x) = const. The straightforward approach is to write down the equation of
motion
d2 x dU
m 2 =−
dt dx
(since U depends only on one variable) and to try to solve it. However, it is much more simple and
physically meaningful to write down the energy conservation as follows
 2
m dx
+ U (x) = E, (2)
2 dt

and solve it with respect to dx/dt taking into account that E is some constant parameter:
 1/2
dx 2
=± (E − U (x)) (3)
dt m

and from here

x t
dx0
Z Z
± p = = t − t0 (4)
x0 (2/m)(E − U (x0 )) t0

where x = x0 at t = t0 . Even if we cannot perform the integration practically, the solution is known
in principle (only technical problems remain) and the motion is said to be integrable. The above
expressions require that E ≥ U , which means that all points x where U (x) > E cannot be reached
by the particle. In the point x∗ where U (x∗) = E the particle velocity vx = 0 and its sign changes,

2
Physics 1. Mechanics Lecture 9

that is, this point is the turning point. Three possible types of particle trajectories exist: a) if U ≤ E
for all points −∞ < x < ∞ the trajectory is infinite and unbound, the particle velocity does not
change sign and the particle crosses the whole space; b) if U (x) ≤ E only for x0 ≤ x < ∞ (or
−∞ < x ≤ x0 ) then the trajectory is infinite but bound from one side: the particle comes from ∞
(−∞ in the second case) turns at x = x0 and returns to ∞ (−∞); c) if U (x) ≤ E for x1 ≤ x ≤ x2
than the particle trajectory is finite and bound, the particle moves back and forth between the two
turning points. In the last case the motion becomes periodic and the period is
x2
dx0
Z
T =2 p .
x1 (2/m)(E − U (x0 ))

Attention: The last statement is correct only if in both turning points dU/dx 6= 0. If U (x∗) = E
and (dU/dx)x=x∗ = 0 simultaneously, the time which is necessary for the particle to reach the point
x∗ becomes infinite.

0.6

0.5

0.4

0.3

0.2

0.1

x
-7.5 -5 -2.5 2.5 5 7.5

Figure 1: Potential energy U (x) is shown by a thick black line. Upper thin black line - infinite
unbound motion. Green line - infinite bound motion (two cases), vertical green lines show the
coordinates of turning points. Blue line, right, - bound infinite. Blue, left, - bound, infinite, requires
infinite time to reach the turning point. Blue, middle, - bound, infinite period. Red, left, right, -
bound infinite. Red, middle, - bound finite, periodic.

3 Central forces
Central forces correspond to the potential energy of the form U = U (r) where r = |r|. As we have
found already in this case the angular momentum J = r×p is conserved. This means that the vectors

3
Physics 1. Mechanics Lecture 9

r and p are always in the the plane perpendicular to the constant vector J so that the motion is
two-dimensional. Let the particle move in the x − y plane. It is convenient to use polar coordinates
x = r cos ϕ, y = r sin ϕ. We shall also introduce the unit vectors êr = r/r = cos ϕêx + sin ϕêy
and êϕ = − sin ϕêx + cos ϕêy . It should be understood that x, y, r, ϕ are coordinates of the moving
particle and therefore all depend on time, as well as the unit vectors êr and êϕ . Using this notation
we can write

v = ṙ = ẋêx + ẏêy
(ṙ cos ϕ − r sin ϕϕ̇)êx + (ṙ sin ϕ + r cos ϕϕ̇)êy (5)
= ṙêr + rϕ̇êϕ

The angular momentum has only one component J = (0, 0, J),

J = xpy − ypx = m(xẏ − y ẋ) = m[r cos ϕ(ṙ sin ϕ + r cos ϕϕ̇) − r sin ϕ(ṙ cos ϕ − r sin ϕϕ̇)] = mr2 ϕ̇

and this immediately gives us

J
ϕ̇ =. (6)
mr2
Now we shall use the energy conservation in the form

m 2
v + U (r) = E = const
2

Substituting (5) we have

m 2
(ṙ + r2 ϕ̇2 ) + U (r) = E
2
or finally
m dr 2
( ) = E − Uef f (r), (7)
2 dt
where
J2
Uef f = U (r) + (8)
2mr2
is the effective potential. With the use of the effective potential the problem is reduced to one-
dimensional analyzed earlier with the only difference that 0 ≤ r < ∞ (and not −∞ < x < ∞).

4
Physics 1. Mechanics Lecture 9

u
40

20

r
1 2 3 4 5

-20

-40

-60

Figure 2: Potential U = −a2 /r (blue) and effective potential (red).

It is of interest to analyze possible trajectories for the potentials of the type U (r) = A/rα , where
α > 0. If A > 0 the potential is repelling: a particle cannot come close to r = 0 even in the case of
J = 0. The centrifugal part J 2 /2mr2 is always repelling and prevents the particle from coming close
to the center r = 0. Thus, for A > 0 the only possible trajectories are infinite trajectories bound
from lower r, and the energy is always positive E > 0.
If A < 0 the potential is attractive. If J = 0 the particle would fall onto the center for all
initial radial velocities (that is, all energies). Presence of the centrifugal part may prevent such
fall. If α < 2 the effective potential has a minimum at r∗ such that (dUef f /dr)r=r∗ = 0, that is,
at r∗ = (−J 2 /αmA)1/(2−α) (A < 0). The effective potential Uef f → ∞ when r → 0 and Uef f → 0
(while remaining negative) when r → ∞. As a result, all trajectories with E > 0 are infinite and
bound only from the lower r side. All trajectories with E < 0 are finite and bound, that is, the
motion occurs between the two radii r1 < r < r2 , such that Uef f (r1 ) = Uef f (r2 ) = E < 0.
Exercise: Show that if α > 2 a particle can fall onto the center even if J 6= 0. What happens
when α = 2 ?
The solution of (7) is Z r
dr0
t − t0 = ± p . (9)
r0 (2/m)(E − Uef f )
p
Using (dϕ/dt) = J/mr2 and dr/dt = ± (2/m)(E − Uef f ) we find

dϕ J
=± p (10)
dr 2
r 2m(E − Uef f )

5
Physics 1. Mechanics Lecture 9

and further r
Jdr0
Z
ϕ − ϕ0 = ± p . (11)
r0 r0 2 2m(E − Uef f )

3.1 Example: Coulomb potential

The potential energy of the form U = −k/r is called Coulomb potential. It describes a charged
particle in the field of a point charge or a point mass in the gravitational field of a spherical body.
In what follows we consider the gravitational field which is attractive, so that k > 0.
Exercise: Using substitution u = 1/r and the integral

2cu − b
Z
du 1
√ = √ arccos( √ )
a + bu − cu 2 c b2 + 4ac

find the solution of (11) in the form r = p/[1 +  cos(ϕ − ϕ0 )] where p = J 2 /mk, and  =
p
1 + 2EJ 2 /mk 2 .
The trajectory found in the last exercise is an ellipse with the major semiaxis a = (rmax +rmin )/2 =
p/(1 − 2 ) = −k/2E, if  < 1. It is obvious that the orbit is circular if  = 0, that is, E = −mk 2 /2J 2
(ATTENTION: finite orbits correspond to negative values of energy E !). When  = 1 the orbit is
an infinite parabola, and when  > 1 the orbit is a hyperbola (also infinite). In both cases the energy
is positive.

6
Physics 1. Mechanics Lecture 9

y
4

x
2 4 6 8

-2

-4

Figure 3: Trajectories in Coulomb potential. Black - circular trajectory,  = 0, red - ellipse, 0 <  < 1,
green - parabola,  = 1, blue - hyperbola,  > 1. Angular momentum J is the same for all orbits,
energy E changes.

7
 Physics 1. Mechanics
Advanced 9. Orbits in a central potential

1 Epicyclic approximation
As we have seen earlier a particle orbit in a central potential U (r) is determined by two conservation
laws. The angular momentum conservation gives

J
ϕ̇ = (1)
mr2

where J = const. The energy conservation gives

2
ṙ2 = (E − Uef f ), (2)
m
J 2
where E = const and Uef f = U + 2mr 2 . (The conserved quantities J and E are also called integrals of

motion.) Although (2) can be formally solved in the form of an integral (is fully integrable), in most
cases the integral cannot be calculated in terms of elementary functions which makes it difficult to
understand the motion. In many cases approximate methods provide more clear information.
The simplest orbit is, of course, a circular orbit. In this case ṙ = 0 which means that the orbit
radius should be find as a minimum of Uef f and E and J are not independent at the orbit. Let r0
be the minimum, that is, dUef f /dr = 0 at r = r0 . In other words, dU/dr − J 2 /mr03 = 0. The energy
and angular momentum are relate by E = Uef f (r0 ) = U (r0 ) + J 2 /2mr02 . The angular velocity on this
J
trajectory is ϕ̇ = mr 2
0
In order to move away from the circular trajectory we have to break the relation between E and
J. It is sufficient to keep J constant and change E. We are interested in the trajectories which are
close to the found circular orbit, so that we shall assume that E = E0 + E1 = Uef f (r0 ) + E1 and
E1 is sufficiently small. As usual, the condition of the smallness will be found after we find the new
trajectory. It is worth noting that E1 > 0 since we were already in the minimum. Now r = r0 + r1 ,
where r1 is the perturbation. Taylor expanding the potential near the minimum we have

d2 Uef f
Uef f (r) = Uef f (r0 ) + 12 r12 (3)
dr2 |r=r0

To make the notation more compact let us denote Uef f (r0 ) = Ū0 , U (r0 ) = U0 , (d2 U/dr2 )|r=r0 = U 00 ,

1
Physics 1. Mechanics Advanced 9

(d2 Uef f /dr2 )|r=r0 = (U 00 + 3J 2 /mr04 ) = mκ2 . Then the energy conservation will give

2E1
ṙ12 + κ2 r12 = (4)
m

The solution of this equation is r

1 2E1
r1 = cos(κt − φ). (5)
κ m
For ϕ we have

J 2J
ϕ̇ = − r1
mr0 mr03
2
r (6)
J 2J 2E1
= − cos(κt − φ)
mr02 κmr03 m

so that r
J 2J 2E1
ϕ= 2
t− 2 3 sin(κt − φ). (7)
mr0 κ mr0 m
This is the so called epicyclic approximation. It represents the motion as the “drift” on the circle on
the radius r0 and “rotation” on the epicycles (perturbations).

2 Keplerian potential
The potential U = −A/r is called Keplerian (or Coulomb). We know that all closed orbits in this
potential are ellipses. Let us apply the above method to the analysis of ellipses with small eccentricity
(deviations from circular shape). We have

A J2
Uef f = − +
r 2mr2

and
dUef f A J2
= 2− =0
dr r mr3
gives the radius of the circular orbit as r0 = J 2 /mA. Let A = mG (as for the gravitation). Then it
is appropriate to use the normalized variable l = J/m and e = E/m, so that r0 = l2 /G. The angular
velocity of the circular motion is
Ω = ϕ̇ = l/r02 = G2 /l3 ,

while the total normalized energy is e0 = −G/2r0 = −G2 /2l2 . Now

2G 3l2
 
2
κ = − 3 + 4 = Ω2
r0 r0

2
Physics 1. Mechanics Advanced 9

Thus, the Keplerian ellipse is represented with the epicyclic motion with the equal orbital Ω and
epicyclic κ angular velocities.

3
 Physics 1. Mechanics
Lecture 10. Motion in a potential, II, Harmonic oscillations

1 Importance of the oscillator potential

Let us consider a particle moving in an one-dimensional potential field, that is, having potential energy
U(x) depending only on one coordinate. As we already know we can ignore two other coordinates
(y and z) and concentrate on x only. The equation of motion is

d2 x dU
m =− , (1)
dt 2 dx

where partial derivative ∂U/∂x → dU/dx since only one coordinate is changing. The energy conser-
vation is now
1
2
mẋ2 + U(x) = E = const. (2)

Let us assume that the particle is in an equilibrium at the point x = x0 , that is, ẍ = 0, ẋ = 0.
The first condition (zero acceleration) immediately means that the force (dU/dx)x=x0 = 0 . If we
perturb a little the particle it will no longer be in the equilibrium but will move so that (hopefully) its
coordinate is close to x = x0 . If the difference |x − x0 | is sufficiently small (which should be checked
separately in each case) then we can write approximately

dU d2 U
U(x) = U(x0 ) + (x − x0 ) + 21 2 (x − x0 )2 , (3)
dx |x=x0 dx |x=x0

which is nothing but the Taylor expansion of the potential energy near the equilibrium point x =
x0 . In (3) the derivatives are taken in the equilibrium point, and, therefore, are simply constants.
Moreover, in the equilibrium point (dU/dx)x=x0 = 0. Let us denote U0 ≡ U(x0 ), k ≡ (d2 U/dx2 )|x=x0 ,
and for simplicity put x0 = 0 and U0 = 0 (we can always do that), so that

kx2
U(x) = U0 + . (4)
2

1
Physics 1. Mechanics Lecture 10

U
8

x
-2 -1 1 2

Figure 1: Approximation of the potential energy near stable equilibrium by (4).

Now the equation of motion and energy conservation take the form:

mẍ = −kx, (5)

1
2
mẋ2 + 21 kx2 = E = const. (6)

Equations (5) and (6) describe the so called harmonic oscillations. Note that the derivation of
these equations is quite general, which means that every stable system harmonically oscillates when
perturbed from equilibrium.

2 Harmonic oscillator: Direct solution

The solution of the equations for motion for a harmonic oscillator can be obtained directly from the
energy conservation (as we did in the general one-dimensional case) by solving for ẋ:
p
ẋ = ± (2E − kx2 )/m

and integrating
x
dx
Z
t − t0 = ± p .
x0 (2E − kx2 )/m
Q: Show that the solution is of the form x = A cos(ωt − φ) and find A, ω, and φ.

2
Physics 1. Mechanics Lecture 10

3 Harmonic oscillator: General method

The above method is not good if there are more than one degrees of freedom (motion along more than
one coordinate) or there are additional nonconservative forces which break the energy conservation
(we shall add such in future). There is another, a more general and very powerful method which is
based on solving linear differential equations with constant coefficients. We start with (5) and assume
that a solution can be written in the form

x = C exp(pt), (7)

where p and C are to be found upon substitution of (7) into (5). In fact, this is not an assumption:
any solution of (5) must be of the form x = C exp(pt).
Substituting, we obtain
p2 = −(k/m), (8)

while C is arbitrary (in fact, it has to be determined from initial conditions, we’ll do that later).
p
The solution of (8) is p = ±iω, ω = k/m. Thus, there are two possibilities, x = C1 exp(iωt) and
x = C2 exp(−ωt). Since we have two initial conditions, x(t = 0) = x0 and vx (t = 0) = v0 we have to
combine the two into the general solution

x = C1 exp(iωt) + C2 exp(−iωt). (9)

ATTENTION: C1 and C2 are complex constants, in general, while x must be real.

From (9) we find the velocity in the form

vx = ẋ = iω[C1 exp(iωt) − C2 exp(−iωt)], (10)

Now, taking into account the initial conditions, we have

C1 + C2 = x0 , iω(C1 − C2 ) = v0

3
Physics 1. Mechanics Lecture 10

and eventually

v0 v0
x = 12 (x0 + ) exp(iωt) + 21 (x0 − ) exp(−iωt) (11)
 iω  iω  
exp(iωt) + exp(−iωt) v0 exp(iωt) − exp(−iωt)
= x0 + (12)
2 ω 2i
v0
= x0 cos(ωt) + sin(ωt) (13)
ω
= A cos(ωt − φ), (14)
q
A = x20 + (v0 /ω)2 , (15)
x0 = A cos φ, v0 /ω = A sin φ (16)

A is called the amplitude of the oscillations and φ is called the phase. The amplitude shows the
maximum displacement from the equilibrium A = |x|max . The phase gives the relation between the
initial velocity and initial displacement.

x
1

0.5

t
2 4 6 8 10 12 14

-0.5

-1
Figure 2: Coordinate (red solid line) and velocity (blue dashed line ) as functions of time.

It is easy to show that the energy of the particle is

m 2 k 2 kA2
E= v + x = = const (17)
2 x 2 2

4 Damping oscillations
Here we are going to add a friction force of special kind, F = −γvx . When friction is present the
energy is no longer conserved and we cannot use the energy conservation but have to solve the

4
Physics 1. Mechanics Lecture 10

equation of motion which is now

mẍ = −kx − γvx (18)

or, in a more widely used form,

ẍ + Γẋ + ω 2 x = 0, (19)

where Γ = γ/m, and ω 2 = k/m. Equation (19) is also a linear differential equation with constant
coefficients and is also solved using substitution x = C exp(pt), which gives

p2 + Γp + ω 2 = 0 (20)

so that
Γ p 2
p1,2 = − ± Γ /4 − ω 2 . (21)
2
Since there are two different solutions p1 6= p2 the general solution will again have the form x =
C1 exp(p1 t) + C2 exp(p2 t), where C1 and C2 are determined from the initial conditions.
Let us first consider the case of strong friction, Γ/2 > ω. In this case both pi < 0, so that exp(pi t)
are monotonically decreasing functions of time. The total energy

m k
E = ẋ2 + x2
2 2
m 2
= [ẋ + ω 2x2 ]
2 (22)
m
= [(p1 C1 exp(p1 t) + p2 C2 exp(p2 t))2 + ω 2(C1 exp(p1 t) + C2 exp(p2 t))2 ]
2
m 2
= [(p1 + ω 2 )C12 exp(2p1 t) + (p22 + ω 2)C22 exp(2p2 t) + 2(p1 p2 + ω 2 )C1 C2 exp((p1 + p2 )t)]
2

Each term in the last expression is positive and decreases with time, which means that the total
energy monotonically decreases. The motion is strongly damped.
Let us now consider the case of weak friction, Γ/2 < ω. In this case

Γ
p1,2 = − ± iω̃ (23)
2

where
p
ω̃ = ω 2 − (Γ/2)2 , (24)

and the general solution can be written as follows:

x = [C1 exp(iω̃t) + C2 exp(−iω̃t)] exp(−Γt/2). (25)

We have seen earlier that the expression in brackets can be written as follows:

C1 exp(iω̃t) + C2 exp(−iω̃t) = A cos(ω̃t − φ)

5
Physics 1. Mechanics Lecture 10

and describes harmonic oscillations. Now (25) can be rewritten as

x = A(t) cos(ω̃t − φ) (26)

where now the “amplitude” A(t) = A0 exp(−Γt/2) monotonically decreases with time (A0 is the
amplitude at t = 0). The cos term is periodic in time and describes oscillations, while exp(−Γt/2)
describes damping. Thus, altogether (26) describes damped oscillations. During these oscillations
the displacement crosses x = 0 each T = 2π/ω̃, that is, infinite number of times, while the maximum
displacement during each period T monotonically decreases.

x
0.8

0.6

0.4

0.2

t
2 4 6 8 10 12 14
-0.2

-0.4

Figure 3: Coordinate (red solid line) and velocity (blued dashed line) in the case of weak damping.

The power of the friction force P = F · v = −γvx2 ≤ 0 so that the total mechanical energy
E = K + U should decrease monotonically. Using the expression (26) we find the velocity in the
form
Γ
vx = ẋ = − A(t) cos(ω̃t − φ) − ω̃A(t) sin(ω̃t − φ). (27)
2
Substituting (26) and (27) into E = mvx2 /2 + kx2 /2 = (m/2)(vx2 + ω 2x2 ) we find

m Γ2

E(t) = cos2 (ω̃t − φ) + ω̃ 2 sin2 (ω̃t − φ)
2 4
+Γω̃ cos(ω̃t − φ) sin(ω̃t − φ) + ω 2 cos2 (ω̃t − φ) A20 exp(−Γt)

(28)
m 2 Γ2
 
Γω̃
= ω + cos(2ω̃t − 2φ) + sin(2ω̃t − 2φ) A20 exp(−Γt)
2 4 2

It can be shown that the energy E(t) decreases but we get more physical information when speaking

6
Physics 1. Mechanics Lecture 10

about the average energy which is defined as

t+T
1
Z
Ē(t) = E(t)dt (29)
T t

Doing such averaging in our case we find

k k
Ē(t) = [A(t)]2 = A20 exp(−Γt) (30)
2 2

We see that
Ē(t + 1/Γ) 1
= ,
Ē(t) e
thus the time τ = 1/Γ describes the rate of the change of the average oscillation energy. It is called
damping or decay time.

e
1

0.8

0.6

0.4

0.2

t
2 4 6 8 10 12 14

Figure 4: Energy (red solid line) and period average energy (blue dashed line) for the weak damping
case.

5 Forced oscillations
Oscillations which occur in the system when slightly perturbed from the equilibrium and no other
force is present are called natural oscillations or a normal mode and occur at the natural frequency
p
ω = k/m. The friction force of the kind Fx = −γvx changes the oscillation frequency to ω̄ =
p
ω 2 − (Γ/2)2 and brings about damping. In this case we shall call the oscillations damping natural
oscillations or a damping normal mode. Now we are going to consider the case where such system is
under the influence of an external force of the kind Fext = F0 cos(ω̃t). That means that we have to

7
Physics 1. Mechanics Lecture 10

solve the equation of the form

ẍ + Γẋ + ω 2x = (F0 /m) cos(ω̃t) (31)

We shall try the solution of the form x = A cos(ω̃t − φ0 ). Then

x = A[cos(ω̃t) cos φ0 − sin(ω̃t) sin φ0 ], (32)

ẋ = ω̃A[− sin(ω̃t) cos φ0 − cos(ω̃t) sin φ0 ], (33)
ẍ = −ω̃ 2 A[cos(ω̃t) cos φ0 − sin(ω̃t) sin φ0 ] (34)

Substituting this into (31) and collecting all terms with cos(ω̃t) and with sin(ω̃t) separately, we arrive
at two equations

[(ω 2 − ω̃ 2 ) cos φ0 − Γω̃ sin φ0 ]A = (F0 /m), (35)

(ω 2 − ω̃ 2 ) sin φ0 + Γω̃ cos φ0 = 0 (36)

From the second equation we have

ω 2 − ω̃ 2
cos φ0 = , (37)
[(ω 2 − ω̃ 2)2 + Γ2 ω̃ 2 ]1/2
Γω̃
sin φ0 = − 2 , (38)
[(ω − ω̃ 2 )2 + Γ2 ω̃ 2 ]1/2

and substituting into (35) we find eventually

F0 1
A= . (39)
m [(ω − ω̃ ) + Γ2 ω̃ 2 ]1/2
2 2 2

The dependence of the amplitude A on ω̃/ω for several ratios Γ/ω is shown in Fig 5. When
Γ/ω  1 the amplitude reaches its maximum at ω̃ ≈ ω (not exactly but very close) and the
maximum is very sharp. This phenomenon is called a resonance.

8
Physics 1. Mechanics Lecture 10

A - amplitude
20

15

10

ω̃/ω
0.5 1 1.5 2

Figure 5: Amplitude of forced oscillations as a function of ω̃/ω for various values of γ/ω. For lower
γ/ω the maximum is higher and more narrow.

6 Oscillations in physical systems

Whenever motion of a physical system is described by equation (5) or its energy can be written
similarly to (6) it experiences harmonic oscillations.

6.1 A mass on a spring

In this case the energy is exactly as (6) so that the derived expressions can be directly applied to the
mass on a spring.

6.2 A pendulum
When a pendulum deviates by the angle θ its energy is

m 2 2
E= l θ̇ + mgl(1 − cos θ)
2

For small θ  1 we use the Taylor expansion cos θ = 1 − θ2 /2, so that

ml2 2 mgl 2
E= θ̇ + θ
2 2

9
Physics 1. Mechanics Lecture 10

Comparing with (6) we see that m → ml2 , k → mgl, and x → θ. We, therefore, immediately
conclude that the motion can be described as follows:

θ = A cos(ωt − φ),
p
where ω = g/l.

10
 Physics 1. Mechanics
Advanced 10. Oscillations, several degrees of freedom

1 Two degrees of freedom

We shall learn now how to deal with the oscillator-like problems when there is more than one degree
of freedom. In this section we restrict ourselves with two degrees of freedom, that is, the potential
U = U (x1 , x2 ). Respectively, we shall write the equations of motion in the form

x˙1 = v1 , (1)
x˙2 = v2 , (2)
∂U
mv˙1 = − , (3)
∂x1
∂U
mv˙2 = − . (4)
∂x2

The equations are deliberately written as a system of the first order differential equations.
We assume that there is an equilibrium x1 = X1 , x2 = X2 ,where ẋi = 0, i = 1, 2. This is
possible only if (∂U/∂xi )|xi =Xi = 0. Now let us, as usual, perturb the system from the equilibrium:
xi (t) = Xi + x̃i (t), vi (t) = ṽi (t). If we remain (at least in the beginning) near the equilibrium point,
the potential U (x1 , x2 ) can be Taylor expanded (we need to expand up to the second order since the
first derivatives are zero):

∂U ∂U
U (x1 , x2 ) = U (X1 , X2 ) + x̃1 + x̃2
∂x1 |xi =Xi ∂x2 |xi =Xi
∂2U ∂2U ∂2U
+ 12 x̃21 2 + x̃1 x̃2 + 12 x̃22 2
∂x1 |xi =Xi ∂x1 ∂x2 |xi =Xi ∂x2 |xi =Xi
X ∂2U (5)
= U (X1 , X2 ) + 21 x̃i x̃j
i,j
∂xi ∂xj
X
= U0 + 12 kij x̃i x̃j .
ij

1
Physics 1. Mechanics Advanced 10

The equations of motion will be rewritten as follows

˜x˙i = ṽi , (6)

X
˜v˙i = − κij x̃j , (7)
j

where κij = kij /m. A more formal approach would be to defined a column vector
 
x̃1
 
x̃2 
r=
 ṽ 
 (8)
 1
ṽ2

and a matrix  
0 0 1 0
 
 0 0 0 1
T=
−κ
 (9)
 11 −κ12 0 0

−κ21 −κ22 0 0
so that the equations of motion will be written as

ṙ = Tr, (10)

Now we start looking for a solution of the kind x̃i = Ai exp(pt), ṽi = Bi exp(pt), or in terms of column
vectors  
A1
 
 A2 
r(t) = 
B  exp(pt) = r0 exp(pt)
 (11)
 1
B2
Substituting this into (10) we have

pr = Tr → (T − pI)r = 0 (12)

The last equation has nonzero solutions only if

det ||T − pI|| = 0.. (13)

Eq. (13) is a fourth order equation with respect to p. In our particular case it looks as follows:

(p2 + κ11 )(p2 + κ22 ) − κ212 = 0 (14)

2
Physics 1. Mechanics Advanced 10

(since κ12 = κ21 ), and has the solutions

s 2
κ11 + κ22 κ11 − κ22
p21 =− − + κ212 , (15)
2 2
s 2
κ11 + κ22 κ11 − κ22
p22 =− + + κ212 . (16)
2 2

One solution, p21 < 0, which means that we have two possibilities: r ∝ exp(iω1 t) and r ∝ exp(−iω1 t)
where ω12 = −p21 > 0. These are oscillating solutions. The second option, p22 , is negative is κ11 κ22 >
κ212 and is positive otherwise.

1.0.1 p22 negative

In this case the other options are ∝ exp(±iω2 t), where ω22 = −p22 > 0. The general solution would
be
r = r1 exp(iω1 t) + r2 exp(−iω1 t) + r3 exp(iω2 t) + r4 exp(−iω2 t) (17)

where ri , i = 1, . . . , 4 are constant column vectors. In a more general way, we say that there are four
frequencies - normal modes, and the general solution is a superposition of these normal modes.

1.0.2 p22 positive

In this case we have one of the solutions of the type ∝ exp(p2 t), where p2 > 0 (the other one with
p2 < 0 tends to zero as t → ∞). This solution is called unstable.

2 General approach
Systems can consist of a larger number of particles, like, for example, several masses connected with
springs, so that there may be a rather large number of degrees of freedom. In general, such systems
can be represented near the equilibrium by the equation of motion similar to (10) only the size of
the vector r and matrix T corresponds to twice the number of degrees of freedom. The procedure
which leads to this system is called linearization since the obtained equations are linear. We always
search for solutions ∝ exp(pt), and possible p are found from the equation (13). Each imaginary
p corresponds to a normal mode of the system. If Re(p) < 0 the mode is said to be damping. If
Re(p) > 0 the mode is unstable.

3
 Physics 1. Mechanics
Lecture 11. Rigid body rotation I, Theory

1 Rotation of a rigid body

A rigid body is a body in which the distance between each two points |r(1) − r(2) | does not change
during its motion. Let us fix a point in the body, say P0 (r(0) ), and consider the motion of all other
points relative to this reference point. Since for each point r(i) the distance |r(i) − r(0) | does not
change, that means that at every moment the point i moves on a circle with the radius |r(i) − r(0) |
around the reference point.

Figure 1: Rotation of a rigid body.

As we already know such motion is described by an angular velocity vector ω (i) (which, in general,
depends on time) and its velocity (relative to the reference point) is v(i) = ω (i) × (r(i) − r(0) ). Indeed,

d|r(i) − r(0) |2 dr(i)

= 2(r(i) − r(0) ) · =0
dt dt

This is right for all points. Moreover, the distance |r(i) − r(j) | should not change for any i and j,

1
Physics 1. Mechanics Lecture 11

which gives, as above

d|r(i) − r(j) |2 dr(i) dr(j)

= 2(r(i) − r(j) ) · ( −( )
dt dt dt
= 2(r(i) − r(j) ) · (v(i) − v(j) )
= 2(r(i) − r(j) ) · [ω (i) × (r(i) − r(0) ) − ω (j) (r(j) − r(0) )] = 0

It is easy to see that in order that the last equality be satisfied ω (i) = ω (j) , that is, all point of the
body rotate with the same angular velocity ω.
Thus, at any time t each point of the body has the velocity (relative to r(0) ) v(i) = ω × r0 (i) , where
we used the notation r0 (i) = r(i) − r(0) . Let us calculate the angular momentum and kinetic energy
of this motion, assuming that the point r(0) is at rest.
X X
J= r0 (i) × p(i) = r0 (i) × m(i) (ω × r0 (i) ) (1)
(i) (i)
2
X
= m(i) [r 0 (i) ω − r0 (i) (r0 (i) · ω)] (2)
(i)
X
K= 1
2
m(i) [ω × r0 (i) ]2 (3)
(i)
2
X
= 1
2
m(i) [r 0 (i) ω 2 − (r0 (i) · ω)2 ] (4)
(i)

One important thing we can see immediately: J is not, in general, parallel to ω. It is more instructive
to write down the angular momentum and kinetic energy in components. In what follows the indices
a = 1, 2, 3 and b = 1, 2, 3 denote coordinates: r0 = (x01 , x02 , x03 ).
Q: Show that the above expressions can be written as follows:
XX
Ja = m(i) (x0 (i),b x0 (i),b ωa − x0 (i),a x0 (i),b ωb ) (5)
(i) b
2
XX
= [ m(i) (r 0 (i) δab − x0 (i),a x0 (i),b )]ωb (6)
b (i)
X
= Iab ωb , (7)
b
2
X
Iab = m(i) (r 0 (i) δab − x0 (i),a x0 (i),b ) (8)
(i)
X
1
K= 2
Iab ωa ωb (9)
a,b

The construction Iab which stands in the expressions for the angular momentum and kinetic
energy is called tensor of inertia.

2
Physics 1. Mechanics Lecture 11

2 Rotation around a fixed axis

If the axis of rotation is fixed the above expressions can be simplified. Let the body rotate around z
axis, that is, ω = ωẑ = (0, 0, ω). Then (7) and (9) give

Jz = Izz ω, (10)
2 2 2
X X
Izz = m(i) (x0 (i) + y 0 (i) ) = m(i) r0 ⊥ , (11)
(i) (i)

K= 1
I ω2
2 zz
(12)

The (single) quantity Izz is called the moment of inertia with respect to z-axis. We shall write down
immediately also the equations for the change of the angular momentum and kinetic energy. The
first one is dJ/dt = N = (i) r0 (i) × F(i) which gives for the z component
P

dJz dω X 0 X
= Izz = r i,⊥ × F(i) = (x0(i) F(i),y − y(i)
0
F(i),x = Nz (13)
dt dt
(i) (i)

For the kinetic energy one has

dK X
= F(i) · v(i)
dt
(i)
X
= F(i) · (ω × r0 (i) )
(i) (14)
X
= ω · r0 i,⊥ × F(i)
(i)

= Nz ω

We see the complete analogy with the one-dimensional particle motion:

3
Physics 1. Mechanics Lecture 11

Particle Rigid body

Mass m Moment of inertia Izz

Coordinate x Angle of rotation θ

Velocity vx = ẋ Angular velocity ω = θ̇

Acceleration ax = v˙x Angular acceleration α = ω̇

Force Fx Torque Nz

Momentum px = mvx Angular momentum Jz = Izz ω

Kinetic energy K = mvx2 /2 Kinetic energy K = Izz ω 2 /2

Thus, the motion of the rigid body rotating around a fixed axis is described by a single coordinate
(or a single equation of motion). It is said to have one degree of freedom, as well as a particle moving
in a one-dimensional space.

3 Center-of-mass in action
In the previous sections we derived the equations of motion for a rigid body rotating around an
arbitrary axis. This axis passes through some point r(0) of the body, which means that we can choose
each point as a momentary reference point. However, in many cases it is physically meaningful to
use the special point, the center-of-mass, as the reference point. Indeed, we know that

M V˙cm = Fext (15)

where Fext is the vector sum of all external forces acting on the body, M is the total mass, and Vcm
the velocity of the center-of-mass. We also know that

d
(Mr × Vcm ) = Rcm × Fext (16)
dt

and
X
J̇int = N = (r(i) − Rcm ) × F(i) (17)
(i)

Thus, a rigid body motion can be always represented as a motion of the center-of-mass and the
rotation around the center of mass. The importance of this representation is that the tensor of
inertia Icm,ab (or moment of inertia) relative to the center-of-mass is well-defined and is a quantity

4
Physics 1. Mechanics Lecture 11

which depends only on the mass distribution in the body and can be calculated once and forever for
each given body. The kinetic energy of such body
2
MVcm X
K= + 12 m(i) (v(i) − Vcm )2 (18)
2
(i)

P
where the second term (internal kinetic energy) was shown above to be equal (1/2) a,b Icm,ab ωa ωb ,
so that finally
2
MVcm X
K= + 12 Icm,ab ωa ωb (19)
2 a,b

A question can be asked: what is the relation between the tensor of inertia Iab calculated relative
to some arbitrary point r0 and the tensor of inertia Icm,ab . In order to establish the relation we just
write down the definition
X
Iab = m(i) [(r(i) − r0 ) · (r(i) − r0 )δab − (x(i),a − x0,a )(x(i),b − x0,b )]
(i)
X
= m(i) [((r(i) − rcm ) + (rcm − r0 )) · ((r(i) − rcm ) + (rcm − r0 ))δab
(i)

− ((x(i),a − xcm,a ) + (xcm,a − x0,a ))((x(i),a − xcm,a ) + (xcm,a − x0,a ))]

X
= m(i) [(r(i) − rcm ) · (r(i) − rcm )δab − (x(i),a − xcm,a )(x(i),b − xcm,b )]
(i)
X
+ m(i) [(rcm − r0 ) · (rcm − r0 )δab − (xcm,a − x0,a )(xcm,b − x0,b )]
(i)
X
+ m(i) [2(r(i) − rcm ) · (r0 − rcm ) − (x(i),a − xcm,a )(x0,b − xcm,b) − (x(i),b − xcm,b )(x0,a − xcm,a )]
(i)

= Icm,ab + M[(rcm − r0 ) · (rcm − r0 )δab − (xcm,a − x0,a )(xcm,b − x0,b )]

(20)
P
since all terms like (i) m(i) (x(i),a − xcm,a ) = 0 because of the definition of the center-of-mass. Thus,
we arrive at the generalized Steiner theorem:

Iab = Icm,ab + M[(rcm − r0 ) · (rcm − r0 )δab − (xcm,a − x0,a )(xcm,b − x0,b )] (21)

For the case when the direction of the angular velocity (rotation axis) does not change (let it be in
z direction) the above reduces to the more familiar Steiner theorem

Izz = Icm,zz + Ml2 , (22)

p
where l = (x0 − xcm )2 + (y0 − ycm )2 is the distance between the rotation axis and a parallel axis
passing through the center-of-mass.

5
Physics 1. Mechanics Lecture 11

4 Meaning of the tensor of inertia

P
Since Ja = b Iab ωb it means that, in general, the direction of the angular momentum J does not
coincide with the direction of the angular velocity ω.

Figure 2: Directions of angular velocity (red) and angular momentum (blue) do not coincide, in
general.

Let, for example, ω = ωẑ. Then

Jx = Ixz ω, Jy = Iyz ω, Jz = Izz ω (23)

The components Jx and Jy vanish only if Ixz = Iyz = 0. However, even if they vanish at some
moment they can become nonzero in other moment. Indeed, look at the definition of the tensor of
inertia. It includes summation over the distribution of masses. When a body rotates this distribution
changes with time in the inertial frame, which means that the tensor of inertia remains constant and
is good for use only in the frame rotating with the body. Let us assume that the torque N = 0 in the
inertial frame. Then the angular momentum J should be constant, that is dJ/dt = 0. Writing this
in components we have
J˙a = [I˙ab ωb + Iab ω̇b ] = 0
X
(24)
b

and since I˙ab 6= 0, in general, for a rotating body in the inertial frame, the angular velocity ω also

6
Physics 1. Mechanics Lecture 11

changes with time, dωa /dt 6= 0. There is important exclusion from this general rule. It can be shown
that for any (even extremely irregular) body it is possible to choose cartesian coordinates so that
Ixy = Iyx = Ixz = Izx = Iyz = Izy = 0 (in the reference frame related to the body). Let us denote
I1 = Ixx , I2 = Iyy m I3 = Izz in these coordinates. If the initial angular velocity is along one of
these coordinate axes (say, z), then the angular momentum is also along this axis, J = I3 ωẑ and
remains in the same direction as long as the angular momentum is conserved. In this case I˙3 = 0
and ω = const.

Figure 3: Directions of angular velocity do not coincide, of angular velocity is not directed along one
of the principal axes (left). If angular velocity is directed along on of the principal axes, the angular
momentum points in the same direction.

5 Degrees of freedom
So far we did not mention degrees of freedom. Number of degrees of freedom is simply a number of
coordinates which are necessary to describe the motion, or number of equations of motion which have
to be solved, or three components of (generalized) momenta. Single particle in a three-dimensional
space has three degrees of freedom. A system of N independently moving particles has 3N degrees
of freedom. A rigid body has six degrees of freedom. Three of these are related to the motion of
the center-of-mass (like a particle) and three other are related to three independent rotations around
three axes. Otherwise it can be said that the six degrees of freedom correspond to three components
of the momentum and three component of the angular momentum.

7
 Physics 1. Mechanics
Lecture 12. Rigid body rotation II, Applications

1 Calculation of moment of inertia

In the following examples we usually restrict ourselves with the calculation of the moment of inertia
relative to the axis passing through the center-of-mass. Moment of inertia relative to any other
parallel axis can be found from the Steiner theorem I = Icm + Ml2 where l is the distance between
the axes.

Figure 1: Steiner theorem

Calculation of moment of inertia for a system of particles is straightforward and will not be

1
Physics 1. Mechanics Lecture 12

considered here. The general expression reads

X
2
Iab = m(i) (r(i) δab − x(i),a x(i),b ) (1)
(i)

for the tensor of inertia, or

X
Izz == m(i) (x2(i) + y(i)
2
) (2)
(i)

for the moment of inertia relative to the rotation axis parallel to z-axis. In both cases the coordinate
origin is on the rotation axis. For a rigid body summation is substituted by integral as follows
Z
Iab = (r 2 δab − xa xb )dm, (3)
Z
Izz = (x2 + y 2 )dm, (4)

where the integral should be taken over all masses. (Compare with the calculation of the center-of-
mass position.)

1.1 Thin circle

A homogeneous circle (radius R, mass m) is in x − y axis. Let us find Ixx , Iyy , and Izz . Equation
(4) gives Z Z Z
2 2 2 2
Izz = (x + y )dm = R dm = R dm = MR2 . (5)

Similarly,
Z Z
2 2
Ixx = (y + z )dm = /z = 0/ = y 2dm (6)
Z Z
Iyy = (x2 + z 2 )dm = /z = 0/ = x2 dm (7)

x2 dm = y 2 dm =
R R
Straightforward calculation would be more lengthy, but we can notice that
(1/2)Izz , so that Ixx = Iyy = MR2 /2.

2
Physics 1. Mechanics Lecture 12

Figure 2: Thin circle: moments of inertia.

This is a special case of the perpendicular axis theorem for bodies with no thickness:
Z Z Z
2 2
Ixx + Iyy = y dm + x dm = (x2 + y 2 )dm = Izz

1.2 Disk
Let a homogeneous disk (mass M and radius R) lie in the x − y plane. We shall calculate Izz and
Ixx = Iyy (the rotation axis always goes through the center of the disk). In order to do so we have to
divide the disk onto small (infinitesimal) parts whose position is given by suitably chosen coordinates,
and the mass dm is expressed in terms of density and coordinate differentials.

3
Physics 1. Mechanics Lecture 12

Figure 3: Disk

It is convenient to use polar coordinates x = r cos ϕ, y = r sin ϕ. With the use of the surface
density σ = M/πR2 the mass element would be dm = σrdrdϕ, so that

R 2π
πR4 σ MR2
Z Z Z
2 2 2
Izz = (x + y )dm = ( σr rdϕ)dr = = , (8)
0 0 2 2
Z R Z 2π
MR2
Z
2 2
Ixx = (y + z )dm = ( σr 2 sin2 ϕrdϕ)dr = (9)
0 0 4

1.3 Sphere
Here we assume that all mass M is homogeneously distributed on the surface of the sphere of the
radius R, so that the surface density is σ = M/4πR2 . Because of the symmetry Ixx = Iyy = Izz so that
we need to calculate only one. We shall calculate Izz using spherical coordinates x = R sin θ cos ϕ,
y = R sin θ sin ϕ (we do not need z !). The infinitesimal surface element on the sphere surface has
the area dA = R2 dθ sin θdϕ and dm = σdA.

4
Physics 1. Mechanics Lecture 12

Figure 4: Sphere.

Now
π 2π
8 2
Z Z Z
2 2
Izz = (x + y )dm = ( R2 sin2 θσR2 sin θdϕ)dθ = πσR4 = MR2 (10)
0 0 3 3

1.4 Filled sphere

This time the mass M is homogeneously distributed in the volume of the sphere of the radius R, so
that the density is ρ = M/(4πR3 /3).

5
Physics 1. Mechanics Lecture 12

Figure 5: Filled sphere.

The only difference with the previous case is that we need the volume element dV = r 2 sin θdrdθdϕ,
so that dm = ρdV and the integration goes in the following way:
Z Z R Z π Z 2π
2 2
Izz = (x + y )dm = ( ( r 2 sin2 θρr 2 sin θdϕ)dθ)dr
0 0 0
R
(11)
8 8 2
Z
= πρr 4 dr = πρR5 = MR2
0 3 15 5

2 Rolling
Let a body with a spherical or cylindrical symmetry moves on a surface with friction so that the
point(s) which are in contact with surface at always at rest. Then this body rotates along the axis
which is parallel to the surface but perpendicular to the velocity.
Such motion is called rolling. This motion can be viewed from the two points of view. First,
we can say that the center of mass moves with some velocity, and the body rotates around the axis
passing thought the center of mass. If the distance between the point of contact and the rotation
axis is R then the center-of-mass velocity v and the angular velocity of rotation ω are related by
v = ωR in order that the contact point not slide relative to the surface. In this case the kinetic
energy is K = Mv 2 /2 + Icm ω 2 /2 = (M + Icm /R2 )v 2 /2.

6
Physics 1. Mechanics Lecture 12

Figure 6: Rolling

On the other hand, we can regards the body at each moment as rotating around the instantaneous
rotation axis, passing thought the contact point(s). Since the velocity of the center-of-mass is v and
it rotates on the radius R the corresponding angular velocity is ω = v/R, that is, the same as in
the previous method of consideration. The kinetic energy is K = Iω 2/2 = (Icm + MR2 )ω 2 /2 =
(Icm + MR2 )v 2 /2R2 . Since we obtained the same as in the previous method, both are consistent.
Let us consider, for example, a cylinder on a slope, and assume that friction is sufficiently strong to
ensure rolling (and prevent sliding). If we consider the cylinder as rotating around the instantaneous
axis of rotation passing through the contact point, the following equation of motion holds:

I ω̇ = N = MgR sin α. (12)

˙ = v̇/R = a/R so that we immediately have

On the other hand, ω̇ = v/R

MR2 1
a= g sin α = g sin α. (13)
I 1 + Icm /MR2

2.1 Physical pendulum

Let a body can rotate around the point O as shown in Fig. 8. If the center-of-mass is at the distance
l from the axis of rotation, the energy can be written as follows:

E = mgl(1 − cos θ) + 21 I θ̇2

7
Physics 1. Mechanics Lecture 12

Figure 7: Cylinder rolling down a slope.

and for small θ  1 (cos θ ≈ 1 − θ2 /2) we have

E = 12 mglθ2 + 12 I θ̇2 .

The last expression has the oscillatory form, and the oscillation frequency can immediately written
p
as ω = mgl/I.

8
Physics 1. Mechanics Lecture 12

Figure 8: Physical pendulum.

9
 Physics 1. Mechanics
Advanced 12. Rigid body rotation

1 General
Let us look at the body rotation from the point of view of the nonrotating S and rotating S 0 frames.
The rotating frame rotates with the instantaneous angular velocity of the body ω. The vectors
measure by the two observers are related by a rotation A = RA0 or A0 = R−1 A. Let us write down
the equation of motion

J̇ = N (1)

ot expressing with the use of the tensor of inertia I

d
(Iω) = N. (2)
dt

Because of the rotation of the body the tensor of inertia is time dependent in the nonrotating frame
dI/dt 6= 0 but it should be constant in the frame related to the body itself, dI0 /dt = 0 We have to
establish the rule of transformation for I. Kinetic energy is invariant under rotations. Remembering
matrix notations it can be written as follows’

K = 21 ω T Iω (3)

where the superscript T means transposed matrix, in particular, it makes a row from a column. Let
me remind also the feature of rotations RT = R−1 . Now, using ω T = ω 0 T RT we have

T
K = 21 ω 0 (RT IR)ω 0
T
(4)
= K 0 = 12 ω 0 I0 ω 0

from which we find

I0 = RT IR = R−1 IR (5)

or
I = RI0 R−1 . (6)

1
Physics 1. Mechanics Advanced 12

Now, substituting all this into (2) we have

d d d
(Iω) = (RI0 R−1 Rω 0 ) = (RI0 ω 0 ) (7)
dt dt   dt
d d
= RI0 ω 0 + R I0 ω 0 = RN0 (8)
dt dt

Multiplying by R−1 from the left we finally obtain

I0 ω̇ 0 + R−1 ṘI0 ω 0 = N0 (9)

which is the equation of motion in the rotating frame where I0 is constant. On the other hand, we
have seen earlier that
Ȧ = RȦ0 + ω × A (10)

Since a vector product ω × A transforms as a vector, we can write also

Ȧ = R(Ȧ0 + ω 0 × A0 ) (11)

We immediately find that

R−1 ṘA0 = ω 0 × A0 (12)

and further
I0 ω̇ 0 + ω 0 × (I0 ω 0 ) = N0 . (13)

In order to proceed we shall use the property of the tensor of inertia I0 which says that it is possible
to choose cartesian coordinates (x01 , x02 , x03 ) (rotating with the body) so that I11
0
= I1 , I220
= I2 , and
I33 = I3 and all other components are zero. In other words Iij = Ii δij . With this Ji = Ii ωi0 and
0 0 0

substitution into (13) gives

X
Ii ω̇ 0 i + εijk ωj0 Ik ωk0 = Ni0 (14)
jk

or in the extended form

I1 ω̇ 0 1 + (I3 − I2 )ω20 ω30 = N10 , (15)

I2 ω̇ 0 2 + (I1 − I3 )ω30 ω10 = N20 , (16)
I3 ω̇ 0 3 + (I2 − I1 )ω10 ω20 = N30 , (17)

It is of interest to consider first the case where there is external torque, that is, N0 = 0. In this case
ω 0 does not remain constant but changes according to

I1 ω̇ 0 1 = (I2 − I3 )ω20 ω30 , (18)

I2 ω̇ 0 2 = (I3 − I1 )ω30 ω10 , (19)

2
Physics 1. Mechanics Advanced 12

I3 ω̇ 0 3 = (I1 − I2 )ω10 ω20 , (20)

It is easily seen that it ω 0 is directed along one of the principal axes (x01 , x02 , x03 ) (that is, only one
component is nonzero in the beginning) it remains directed along this axis and is constant. This
happens in the rotating frame. The observed in the nonrotating frame would see the angular velocity
vector ω rotating as ω ×ω = 0. Thus, in the nonrotating frame the angular velocity remains constant
also.
Let as assume now that ω 0 = ω10 eˆ1 + ω̃ 0 , where ω̃ 0 is a small perturbation. We are already familiar
with the linearization procedure to derive the following equations for the perturbations:

˙
I1 ω̃10 = 0, (21)
˙
I2 ω̃20 = (I3 − I1 )ω˜30 ω10 , (22)
˙
I3 ω̃30 = (I1 − I2 )ω10 ω˜20 . (23)

Following the standard procedure, we substitute ω̃ 0 ∝ exp(pt) and get

2
p2 = −(I1 − I2 )(I1 − I3 )ω 0 1 /I2 I3 , (24)

which is negative (stable equilibrium) if (I1 − I2 )(I1 − I3 ) > 0 (I1 is either the largest or the smallest),
and positive (unstable) if I1 is between I2 and I3 . Thus, the rotation around the largest and smallest
moments of inertia (principal axes) is stable, the rotation around the intermediate is unstable.
It is also easy to see that if N0 = 0 the following quantity is conserved

1 2 2
1
(I ω 0
2 1 1
+ I2 ω 0 2 + I3 ω 0 3 ) = const (25)

which is nothing but the kinetic energy.

2 Precession etc.
Let us assume that a body is symmetrical, so that I2 = I3 and it rotates around I1 with angular
velocity ω10 . Let us also assume that N2 = const 6= 0 while N1 = N3 = 0. The equations of motion
read

I1 ω̇ 0 1 = 0, (26)
I2 ω̇ 0 2 = (I3 − I1 )ω10 ω30 + N2 , (27)
I3 ω̇ 0 3 = (I1 − I2 )ω10 ω20 . (28)

3
Physics 1. Mechanics Advanced 12

We see that ω10 = const. There is a solution with ω20 = 0, ω30 = N2 /(I1 − I3 )ω10 = const, which
corresponds to the rotation of the angular velocity vector (in the non-rotating frame). This motion
is called precession.

4
 Physics 1. Mechanics
Lecture 13. Gravitation

1 Basic law
Every two point masses, m1 and m2 , positioned in r1 and r2 interact (attract each other) with the
force
Gm1 m2 (r2 − r1 )
F1→2 = − . (1)
|r2 − r1 |3
Comparing with what is already known to us we find that this force is conservative and can be
derived from the potential energy

Gm1 m2
U = U(r) = − , r ≡ |r2 − r1 |. (2)
r

2 Two-body problem
Let there are two gravitationally interacting particles without any additional external forces. The
equations of motion can be written as follows:

Gm1 m2 (r1 − r2 )
m1 r̈1 = − , (3)
|r2 − r1 |3
Gm1 m2 (r2 − r1 )
m2 r̈2 = − . (4)
|r2 − r1 |3

Since the forces are internal, the center-of-mass Rcm = (m1 r1 + m2 r2 )/(m1 + m2 ) is in the rest all the
time (this can be checked once again by adding equations (3) and (4). On the other hand, dividing
(3) by m1 , (4) by m2 and subtracting the first from the second, we have

G(m1 + m2 )r
r̈ = − , (5)
r3

where r = r2 − r1 . In order to bring the force to the same expression as above it is widely accepted to
introduce the reduced mass µ = m1 m2 /(m1 + m2 ). With the help of the reduced mass the equation
of motion is written as follows:
Gm1 m2 r
µr̈ = − (6)
r3

1
Physics 1. Mechanics Lecture 13

Thus the two-body problem is reduced to the motion of a single particle with the reduced mass µ in
the same potential energy U = −Gm1 m2 /r. This problem has been solved already. Once r is found
the position vectors of the the two bodies are immediately obtained from r1 = −m2 r/(m1 +m2 )+Rcm
and r2 = m1 r/(m1 + m2 ) + Rcm .

3 Gravitational attraction by spherically symmetric distri-

butions
Let us assume that a mass M is homogeneously distributed on the surface of a thin (zero thickness)
spherical shell of the radius R. Let us put a test mass m at the distance r from the center of the
shell and calculate the potential energy of the interaction of the test mass with the whole shell. If
we divide the sphere onto small parts dM the potential energy of the interaction of the test mass
with this part is dU = −GmdM/|r − r |, where r is the position vector of m and r is the position
vector of dM. The total potential energy is then


GmdM
U =− (7)
|r − r |

In order to calculate this integral let us choose the coordinate system so that the origin is in the center
of the sphere and the test mass is on z axis, that is, r = (0, 0, r). It is convenient to use spherical co-
ordinates where r = (R sin θ cos ϕ, R sin θ sin ϕ, R cos θ). Then |r − r | = R2 sin2 θ + (r − R cos θ)2 ,


and dM = σR2 sin θdθdϕ. Here σ = M/4πR2 . With all this

π
2π
2 sin θ
U = −GmσR ( dϕ)dθ (8)
R2 sin2 θ + (r − R cos θ)2

0 0

and substituting z = cos θ, k = r/R

GmM 1


dz
U =− (9)
(1 + k 2 ) − 2kz

2R −1
GmM
=− , if r > R (10)
r
GmM
=− , if r < R. (11)
R

A particle outside the shell has the potential energy which is the same as it would be if the whole
mass of the shell be in the center. The potential energy of a particle inside the shell is constant.
Since F = −(∂U/∂r), the particle outside the shell experiences the force as if the whole mass of the
shell is in the center. There is not force on the particle inside the shell. This can be easily generalized
onto any spherical distribution of mass, since it can be divided onto many thin shells. Thus, let us

2
Physics 1. Mechanics Lecture 13

assume that the mass distributed so that the density depends on the radius only ρ = ρ(r). Let us
out a test particle m at some radius R, then the above analysis shows that only the mass inside the
sphere r < R affects the particle as if the whole mass

R
M(R) = 4πρ(r)r 2dr
0

is in the center, so that

GM(R)m
|F| = .
R2
The material which is outside the sphere, r > R, does not affect the test particle at all.

3
 Physics 1. Mechanics
Lecture 14. Basics of special relativity

1 Observer and event

Special relativity heavily uses two concepts: observer and event. An observer is a physical body
which can server as a basis for a reference frame. Since the velocity of each body cannot exceed the
speed of light, no observer can exist which moves with a superluminal speed. An event is something
which happens in the world. Unless the time duration of the event is of special importance, it is
assumed to be zero. That is, all time intervals should be measured between two events, as well as
spatial separation.

2 Basic principles
The following principles are the cornerstones of special relativity.

1. Inertial frames remain the specially chosen frames where the laws of physics are formulated.

2. No physical body can move at a velocity higher than the light speed in the vacuum. As a result
the synchronizing signals propagate with the speed of light and not with infinite velocity.

3. Speed of light is the same in all frames.

4. As a result, each observer measures his own coordinates and time. Time becomes a fourth
coordinate. Our space is a four-dimensional space-time.

5. Let S and S
be two inertial frames (observers) and let the two observers measure two events,
A1 and A2 . According to S the event A1 occurred in r1 in the moment t1 , the event A2 occurred
in r2 in the moment t2 , the corresponding measurements by S
are denoted by
. Then

(r2 − r1 )2 − c2 (t2 − t1 )2 = (r
2 − r
1 )2 − c2 (t
2 − t
1 )2 (1)

1
Physics 1. Mechanics Lecture 14

3 Lorentz transformations
Let S
move relative to S in x direction with the velocity v0 , both coordinate frames are cartesian
and the axes are parallel, and the origins coincide initially. If the event coordinates are (x, y, z, t)
and (x
, y
, z
, t
), then they are related as follows

y
= y, z
= z, (2)
x
= γ0 (x − v0 t), (3)
t
= γ0 (t − v0 x/c2 ), (4)


γ0 = 1/ 1 − v02 /c2 . (5)

For any velocity v the quantity γ = 1/ 1 − v 2 /c2 is called a Lorentz-factor. The inverse transfor-


mation reads

y = y
, z = z
, (6)
x = γ0 (x
+ v0 t
), (7)
t = γ0 (t
+ v0 x
/c2 ), (8)


γ0 = 1/ 1 − v02 /c2 . (9)

3.1 Simultaneous events

Let two events be observed by the moving observer S
simultaneously, t
1 = t
2 but at different positions
x
1
= x
2 . The observer in rest S will see this events at t1
= t2 , and t2 − t1 = γ0 v0 (x
2 − x
1 )/c2 . This
means that two events may be simultaneous in one frame and not simultaneous in another. The
spatial separation between the two events in S is x2 − x1 = γ0 (x
2 − x
1 ).
Let two events be observed by S
in the same position, x
1 = x
2 but at different times, t
1
= t
2 .
They are observed by S at different positions, x2 − x1 = γ0 v0 (t
2 − t
1 ). The time interval between the
two events in S is t2 − t1 = γ0 (t
2 − t
1 ).

3.2 World lines

In what follows we restrict ourselves with x and t only and forget about y and z for simplicity. A
world line of a particle is the line in the x − t plane, that is, x(t) (or x
(t
) in the moving frame S
).
Let us consider now two particles which are at rest in the frame S
, that is, x
1 (t
) = const
and x
2 (t
) = const. The distance between the particles which the moving observer S
measures is
l
= |x
2 − x
1 |. This distance is measured so that the two coordinates are measured simultaneously by
S
. If we want to know what distance S measures we have to require that the coordinate measurements
in S are simultaneous. Thus, writing x
1 = γ0 (x1 − v0 t1 ), x
2 = γ0 (x2 − v0 t2 ) and t1 = t2 we get

2
Physics 1. Mechanics Lecture 14

l = |x2 − x1 | = |x
2 − x
1 |/γ0 = l
/γ0 , that is the observer at rest measure small length then the
observer moving with the particles.

Example 3.1. L et a rigid ruler move with the velocity v0 = (v0 , 0, 0) and let its length, as

measured in its rest frame S
(moving with the same velocity) be l
. What is it length measured in
the non-moving frame S if it lies along x axis ? Let us denote the coordinates of the two ends of the
ruler in its rest frame S
as x
1 and x
2 . These coordinates remain constant and the ruler length in
this frame is simply l
= x
2 − x
1 . In order to measure the ruler length in S we have to measure the
coordinates of the both ends simultaneously, that is, t1 = t2 . Then we have

x
2 = γ0 (x2 − v0 t1 ), (10)

x
1 = γ0 (x1 − v0 t1 ), l
= x
2 − x
1 = γ0 (x2 − x1 ) = γ0 l (11)
y

v0


S

l

x

l
x

Example 3.2. L et the same rigid ruler lie at an angle θ
in S
. What would be the angle

in S ? According to the above we have lx = lx
/γ0 , and ly
= ly . Since tan θ = ly /lx we have
tan θ = ly
/(lx
/γ0 ) = γ0 tan θ
.

3
Physics 1. Mechanics Lecture 14

y

v0


S

x

Let is now assume that certain event occurs in some point x
of S
so that is starts at t
1 and ends
at t
2 . The observer S
measures the duration of the event T
= |t
2 − t
1 |. The observer S will the
event starting at the moment t1 = γ0 (t
1 + v0 x
/c2 ) and ending at t2 = γ0 (t
2 + v0 x
/c2 ), so that the
event duration, according to S, is T = |t − 2 − t1 | = γ0 |t
2 − t
1 | = γ0 T
.

3.3 Velocity transformation

According to the velocity definition,

dx dy dz
vx = , vy = , vz = , (12)
dt dt dt



dx dy dz

vx
=
, vy
=
, vz
=
. (13)
dt dt dt

From the Lorentz transformations we have

dy
= dy, dz
= dz, (14)

dx
= γ0 (dx − v0 dt), (15)
dt
= γ0 (dt − v0 dx/c2 ), (16)
vy
vy
= , (17)
γ0 (1 − v0 vx /c2 )
vz
vz
= , (18)
γ0 (1 − v0 vx /c2 )
vx − v0
vx
= . (19)
1 − v0 vx /c2

Example 3.3. L et one body move with the velocity v1 = (v, 0, 0) and another with v2 = (−v, 0, 0).

4
Physics 1. Mechanics Lecture 14

What is the relative velocity ? Using the above expression we find vrel = (v−(−v))/(1−v·(−v)/c2 ) =
2v/(1 + v 2 /c2 ).
If we notice that vx v0 = v · v0 , v = vx eˆx  v0 , and v⊥ = vy hatey + vz hatez ⊥ v0 , the above
expressions can be written as follows:

v − v0
v
= , (20)
1 − v · v0 /c2
v⊥
v
⊥ = . (21)
γ(1 − v · v0 /c2 )

These expressions are useful for finding relative velocity.

Example 3.4. L et v1 = v1 (cos 30◦ , sin 30◦ , 0), v2 = v2 (cos 30◦ , − sin 30◦ , 0). We shall rewrite the

expression substituting v0 → v1 , v → v2 , and v
→ vrel . Now v = v2 · v1 /v1 = v2 cos 60◦ ,

above

v⊥ = v22 − v2 = v2 sin 60◦ , and we get

v2 cos 60◦ − v1
vrel, = , (22)
1 − v1 v2 cos 60◦ /c2

v2 sin 60◦
vrel,⊥ = (23)
γ(1 − v1 v2 cos 60◦ /c2 )


where γ = (1 − v12 /c2 )−1/2 . Now, using vrel = 2
vrel, 2
+ vrel,⊥ we eventually find

v12 + v22 − 2v1 v2 cos 60◦ − v12 v22 sin2 60◦ /c2

vrel = (24)
1 − v1 v2 cos 60◦ /c2

3.4 Aberration
Let a particle move with the velocity v = v(cos θ, sin θ, 0) in S (v
= v
(cos θ
, sin θ
, 0) in S
, respec-
tively). From the velocity transformation

vy
v sin θ
tan θ =
= , (25)
vx γ0 (v cos θ − v0 )

that is, θ

= θ, in general. Of particular interest is v = c, that is, when the particle is a photon,
emitted in S at the angle θ to x axis. The above expression gives the angle θ
at which the photon
propagates in S
:
sin θ
tan θ
= . (26)
γ0 (cos θ − v0 /c)
This effect is called aberration.

5
Physics 1. Mechanics Lecture 14

3.5 Doppler effect

Let an electromagnetic emitter (at (x, y) at present) move with the velocity v = (v, 0, 0) and emits a
signal (electromagnetic wave pulse or photon) each T
according to its proper time (as measured by
the clock moving with the emitter). What is the time T between the two successive pulses received
by an observer at the coordinate origin of S ? Let us denote by A the event of the first photon
emission in (x, y) at the time tA = t
A = 0, by A
the event of the first photon reception in (0, 0) at
tA
, by B the event of the second photon emission in (x + ∆x, y) (the emitter is moving !) at tB , and
by B
the event of the second photon reception in (0, 0) at tB
. The time we are looking for is the
time between two successive receptions T = tB
− tA
.
If the first photon is emitted in the point (x, y) at t = 0 it arrives at the coordinate origin at
= lA /c, where lA = x2 + y 2 is the distance between the emitter and receiver in the emission

tA

time t = 0. If the second photon is emitted at the time t
B = T
(by the clock of the emitter - this
is the proper time), the time of the second emission, by the clock of the observer S is tB = γT
.
By that time the moving emitter will be in the point (x + ∆x, y), ∆x = vγT
, so that at the time
of the second emission the distance between the emitter and the receiver is lB = (x + ∆x)2 + y 2.


The second photon will come to the receiver at the time tB
= tB + lB /c, and the difference we are
looking for is T = tB
− tA
= tB + (lB − lA )/c. For small T
the difference lB − lA can be found as
a full differential: lB − lA = (x/ x2 + y 2 )∆x = cos θvγT
, where θ is the angle between the line of


sight and the emitter velocity, that is, between r and v. Using all this we find

T
= γ(1 + v cos θ/c).
T

For a periodic process such as a monochromatic wave T
and T are periods in the emitter rest
frame and the receiver rest frame, respectively. This period is related to the frequency ω as follows
ω = 2π/T , and to the wavelength λ as follows: λ = cT , so that we have

λ ω

= = γ(1 + v cos θ/c). (27)
λ
ω

This is called Doppler effect.

The case θ = 0 corresponds to the emitter moving away from the receiver along the line of sight.
The relation (27) gives λ/λ
= (1 + v/c)/(1 − v/c) > 1 (redshift). The case θ = π corresponds to


the emitter approaching the receiver along the line of sight, and λ/λ
= (1 − v/c)/(1 + v/c) < 1


(blueshift). The case θ = π/2 corresponds to the emitter moving perpendicularly to the line of sight,
so that the distance between the emitter and observer does not change. The relation (27) gives
λ/λ
= γ > 1 (redshift again). This redshift occurs only because of the time dilation ∆t = γ∆t
.

Example 3.5. A star is moving on a circular orbit of the radius R around another star. The

6
Physics 1. Mechanics Lecture 14

angular velocity is ω. The moving star emits electromagnetic radiation with the wavelength λ
in its
rest frame. What wavelength would measure a distant observer which is in the plane of the rotation
? Since the observer is very far from the start system we can ignore the change of the line of sight due
to the finite orbit radius, and choose x axis as the line of sight. Then star velocity is v0 = ωR, while
the angle between the velocity and x axis changes as θ = ωt. Substituting this in the expression for
the Doppler effect we have

λ 1
= [1 + ωR cos(ωt)/c] (28)
λ
1 − ω 2 R2 /c2


4 Momentum and energy

The proper expressions for the momentum and kinetic energy of a particle moving with the velocity
v are
E = mc2 γ, γ = 1/ 1 − v 2 /c2 .

p = mvγ, (29)

Momentum and energy transformation:

p
y = py , p
z = pz , (30)

p
x = γ0 (px − v0 E/c2 ), (31)
E
= γ0 (E − v0 px ). (32)

It is easy to see that the combination E 2 − p2 c2 = m2 c2 is frame independent.

If the energy is much higher than the rest energy, E  mc2 , it is not convenient to use the
velocity which is almost equal to c. Instead, it is widely accepted to use γ.

Example 4.1. W hat is the particle velocity is γ = 10 ? Since γ = (1 − v 2 /c2 )−1/2 we have

v = c 1 − 1/γ 2 ≈ 0.995c.


Some particles (photon, neutrino) have a zero rest mass, which means that for these particles
E = pc.

Example 4.2. T wo photons with p1 = −p2 collide and disappear producing an electron-positron

pair. This reaction can be written as γ + γ → e− + e+ (here γs stand for photons !). The rest masses
of electron and positron, m, are identical. What is the minimal photon energy for this to occur ?
Since the initial momentum (of the two photons) is zero the final momentum (of the electron and
positron) is also zero. Let p be the electron momentum. Then the total energy after the collision
will be 2 p2 c2 + m2 c4 . Since the energy is conserved (there is no internal structure so that it is


impossible to get of loose energy), we have Eγ = p2 c2 + m2 c4 , and the minimum is achieved when


7
Physics 1. Mechanics Lecture 14

p = 0: Emin = mc2 .

4.1 Aberration and Doppler-effect (another approach)

A photon (electromagnetic wave) is a massless particle, so that E = pc. A photon energy is related
to its frequency as E = hν, so that p = hν/c. Let a photon with the frequency ν
is emitted by S

at an angle θ
to x
axis, so that p
x = hν
cos θ
/c, p
y = hν
sin θ
/c. From the momentum-energy
transformation we have

px = hν cos θ/c = γ0 (hν
cos θ
/c + v0 hν
/c2 ) = γ0 (cos θ
+ v0 /c)hν
/c, (33)

py = hν sin θ/c = hν
sin θ
/c, (34)
E = hν = γ0 (hν
+ v0 hν
cos θ
/c), (35)

so that

ν
= γ0 (1 + v0 cos θ
/c), (36)
ν

cos θ
+ v0 /c
cos θ = . (37)
1 + v0 cos θ
/c

The relation (36) describes the change of the frequency depending on the relative velocity of the
emitter (S) and receiver (S
). The frequency decreases when the emitter moves away from the
receiver, and increases when the emitter approaches the receiver. This is called Doppler-effect. The
relation (37) describes the phenomenon which is called aberration: the angle at which a photon is
emitted is different from he angle at which it is received by a moving receiver. We can invert these
relations to have

ν

= γ0 (1 − v0 cos θ/c), (38)
ν
cos θ − v0 /c
cos θ
= . (39)
1 − v0 cos θ/c

so that by measuring the frequency and angle in the receiver frame we can know what are the
frequency and the angle at the emitter.