Armin Reichold

Classical Mechanics
Lecture notes HT2023, 18th January 2023
ii ARMIN REICHOLD
Contents

0.1 Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1 Newton’s law with variable mass and non-inertial reference frames (lectures 11-12) 1
1.1 Variable mass: a body acquiring mass continuously . . . . . . . . . . . . . . . . . 1
Acquiring mass without external forces . . . . . . . . . . . . . . . . . . . . . . . . 1
Acquiring mass with external forces . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Example acquiring mass: The raindrop . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Variable mass: a body ejecting mass and the rocket equation . . . . . . . . . . . 4
1st Example: The horizontal rocket launch . . . . . . . . . . . . . . . . . . . . . . 5
2nd Example: The vertical rocket launch . . . . . . . . . . . . . . . . . . . . . . . 5
Rocket Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Single-stage vs. Two-stage rockets . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Notable Rockets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Non-inertial reference frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Common Examples of non-inertial reference frames . . . . . . . . . . . . . . . . 13

2 Motion of charged particles in B and E fields (lecture 13) 16

2.1 Magnetic Force on a Charge Particle . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2 Motion in Magnetic and Electric fields . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3 Kinetic Energy in Magnetic and Electric fields . . . . . . . . . . . . . . . . . . . . 21
2.4 Applications of motion in B and E fields: The Cyclotron . . . . . . . . . . . . . . 22

3 Vector differentiation, Torque, Angular momentum and circular motion (lectures 14-
15) 24
3.1 Differentiation of vectors with respect to time (lecture 14) . . . . . . . . . . . . . 24
Differentiating a unit vector with respect to time . . . . . . . . . . . . . . . . . . 25
Differentiating the radius vector with respect to time - aka velocity . . . . . . . . 26
Differentiating the velocity with respect to time - aka acceleration . . . . . . . . 26
3.2 Angular momentum and torque (lecture 14) . . . . . . . . . . . . . . . . . . . . . 27
3.3 Angular motion variables and circular motion (lecture 14-15) . . . . . . . . . . . 28
Angular velocity in circular motion (lecture 14) . . . . . . . . . . . . . . . . . . . 28
Angular momentum in circular motion (lecture 14) . . . . . . . . . . . . . . . . . 29
Angular acceleration in circular motion (lecture 15) . . . . . . . . . . . . . . . . . 29
Work and Power in circular motion (lecture 15) . . . . . . . . . . . . . . . . . . . 30
3.4 Correspondences between linear and angular quantities . . . . . . . . . . . . . . 31
3.5 Newton’s laws for angular motion and their use . . . . . . . . . . . . . . . . . . . 31
Example of angular motion: The simple pendulum . . . . . . . . . . . . . . . . . 32
Example of angular motion: The ladder . . . . . . . . . . . . . . . . . . . . . . . . 33

4 Central Forces (lectures 15-20) 35

iv ARMIN REICHOLD

4.1 Definition of Central Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.2 Conservative Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Central Forces are conservative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.4 The equation of motion under central forces . . . . . . . . . . . . . . . . . . . . . 38

5 Appendix A 39

0.1 Foreword

This text is a set of lecture notes for the second half of the first year course
in classical mechanics. It is not a substitute for any text books. It is also
not a substitute for attending the lectures. It has been prepared by writ-
ing up the lectures given in 2022 and 2023 which draw heavily on the
materials prepared by Prof Neville Harnew and Prof Guy Wilkinson and I
acknowledge their good work here. The text books I have used to prepare
my lectures are:

• Budo, Theoretische Mechanik (written in German and not available in

English)

• Recknagel, Physik Mechanik (also only in German)

• Hand and Finch, Analytical Mechanics

Other text books for the course have been recommended by Moritz Riede
on the Canvas page for this course. I strongly recommend that you look
also at this lecture’s materials on the Canvas page which include software
for making the plots as well as recordings of the lectures and updated
lecture slides. At this time these notes are a work in progress and are
being written as the lectures are given. Please recheck my canvas pages
frequently for more up to date versions of these notes.
1 Newton’s law with variable mass and non-inertial
reference frames (lectures 11-12)

The heroes of this chapter will be Sir Isaac Newton (1643 to 1727) and
Konstantin Tsiolkovsky (1857 to 1935). Newton features here for his
ground braking fundamental work on mechanics which we will use right
away and Tsiolkovsky for his application of Newtons work to the mo-
tion of rockets which makes him one of the founding fathers of modern
rocketry.
In this chapter we will start by extending Newton’s second law of
motion to the case in which the mass of the moving body is no longer
constant as was hitherto assumed. The mass of the body could increase
or decrease and for educational reasons we will deal with both cases
separately although we will see that there is no difference between the
two cases as far as the equations of motion are concerned. We will then
use what we have learned and do some genuine “Rocket Science”. But
Figure 1.1: Sir Isaac Newton (1643
right now, let us start by deriving the equations of motion of a body
to 1727)
increasing its mass by merging with a another small body moving at a
different speed.

1.1 Variable mass: a body acquiring mass continuously

Acquiring mass without external forces

We consider a simplified one-dimensional problem as shown in figure

1.3 in which body of mass m is moving at speed v and during a short time
δt is merging inelastically with a very much smaller body of mass δm
moving in the same direction with a different speed u.
In the absence of any external forces we know that the total moment
of the system must be conserved and hence the momentum before the
merger p must equal the momentum after the merger p ′ which gives us.

δp = p ′ − p = {(m + δm)(v + δv)} − {mv − δmu} = 0

Figure 1.2: Konstantin Tsi-
If we multiply the above expression out and ignore the quadratically olkovsky (1857 to 1935)
small quantity δmδv we find

δp = mδv + wδm = 0

Where we have introduced the relative velocity of the two bodies w =

v − u. We now divide the above equation by δt to find:
δp δv δm
=m +w =0
δt δt δt Figure 1.3: A body acquiring mass
from a smaller body
2 CLASSICAL MECHANICS

Now we take the limit of δt → 0 and note that limδt →0 δv

δt =
dv
dt = a and
limδt →0 δm
δt = dm
dt which gives us our first result:

dv w dm ṁ In words equation 1.1 equation would

a= =− = −w (1.1)
dt m dt m read:
” In the absence of external forces a
This is the equivalent of Newton’s second law for a variable mass object bodies acceleration is strictly proportional
in the absence of external forces. We note that the acceleration is no to the relative rate of change of its mass
and the constant of proportionality is the
longer zero although there are no external forces acting. relative speed between the body and its
mass increments.“

Acquiring mass with external forces

The extension of the previous example to include external forces is rather

straight forward. We assume that during the small time δt a constant,
total force F acts on both bodies. If we now use Newton’s second law to
calculate the change of total momentum we find:

δp = F δt = mδv + wδm

Taking the limits of vanishing δt again as before we now get:

ò F = ma + w ṁ (1.2)
In words equation 1.3 would read:
” In the presence of a constant external
force a bodies acceleration is a linear
Alternatively we can rearrange equation 1.2 to yield an equation of function of the relative rate of change of its
motion (expression for acceleration) as done in the previous section and mass. The constant of proportionality is
the relative speed between the body and its
we would find: mass increments. The constant term is the
F ṁ
a= +w (1.3) ratio of the external force to its mass. “
m m
We notice that we only regain F = ma as
Equations 1.2 and 1.3 are the equivalent of Newton’s second law for a the old, familiar form of Newton’s second
variable mass object in the presence of a constant external force F. law if either ṁ = 0 or w = 0 but the form
F = ṗ remains true as it directly arises
from the conservation of momentum. To
Example acquiring mass: The raindrop see this we would of course need to also
consider Newton’s third law and include
We assume that the raindrop starts at height h with initial mass m o and the other body upon which the opposite
force acts and hence balances the change
initial speed v 0 = 0. It then falls under the influence of gravity. We now of overall momentum back to zero.
need to specify the rate of mass accretion because this is now not a single The vapour or the droplets do not fall
accretion event as before but a continuous accretion process. It stands to under gravity as they act like a gas that is
kept afloat by the gas surrounding it.
reason that the rate at which the drop will accrete mass is proportional
to its “size” which in term is proportional to some power of its mass. The
rate of accretion should also be proportional to the relative speed of the
drop and the vapour it accretes. A simple model for the accretion rate
may therefore be:
dm
= bmv (1.4)
dt
Here b is a constant of proportionality with dimension of an inverse
length. It encodes how dense the vapour is through which the drop falls.
One can think of it as the inverse of the mean free path of the drop inside
the vapour cloud. it may have been more realistic to con-
sider the accretion to be proportional
Next we need to specify the forces acting upon the drop. They are the
to the cross-sectional area of the drop
gravitational attraction FG = mg and some air resistance F R . We model 2
which is proportional to m 3 but that
complicates the solution significantly as
m no longer drops out of the equation
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 3

the air resistance as proportional to the mass and to the squared speed.

F R = kmv 2 (1.5)
2
We can now use equation 1.2 to find: it would again be more accurate to us m 3
in here for the same reasons as above.
dv dm
FG + F R = m +w ⇔
dt dt
dv dm
mg − kmv 2 = m +w ⇔
dt dt
dv
= g − kv 2 − bv 2
dt
The last line of the equation above is the equation of motion we were
looking for. To get to it we used the fact that the vapour is at rest and
hence w = v and equation 1.4. ¯
dv ¯
To find the terminal velocity we demand that d t ¯v = 0 and find:
t

r
g
vt = (1.6)
b +k
Next we seek to find the relation between the mass of the drop and the
distance x it has fallen. In equation 1.4 we have a differential equation
for the time evolution of the mass but it is in terms of the speed and not
the distance. We can use the chain rule to transform equation 1.4 into a
differential equation relating mass to distance.

dm dm dt
= ⇔
dx dt dx
dm dt
= bmv ⇔
dx dx
dm 1
= bmv ⇔
dx v
dm
= bm
dx
The last equation in the block above is readily separable into m ad x
dependent terms and can then be integrated which we do below. We
let the mass go from the initial mass m 0 to a general mass m and the
distance fallen changes from 0 to a general distance x.

dm
= bd x ⇒
m
Z m dm
Z x
= bd x ⇔
m0 m 0

[ln m]m x
m 0 = [bx]0 ⇔
m
µ ¶
ln = bx ⇔
m0
m = m 0 e bx

We find that the mass of the drop grows exponentially with the distance
and the scale factor in the exponent is the constant of proportionality
from the from equation 1.4 which governs the mass accrual rate. This
looks very much as one might expect.
Next we want to evaluate the speed of the drop as a function of the
distance fallen and in particular its final speed after falling from a height
4 CLASSICAL MECHANICS

h above the ground. As before we have in the equation of motion 1.3 a

differential equation for the evolution of the speed as function of time
dv dv dx dv
and not as a function of distance. Using the chain rule dt = dx dt = dx v
equation 1.3 becomes
dv
g − (k + b)v 2 =
v (1.7)
dx
Which we again separate and integrate as follows:

vd v
dx = ⇒
g − (k + b)v 2
Z h Z vh
vd v
dx = ⇔
0 0 g − (k + b)v 2
" Ã ¡ ¢ !#v h
ln g − (b + k)v 2
h= −
2(b + k)
0

You should check the final step in the above chain of equations yourself.
When you are convinced that it is right you may insert the limits and
isolate v h to find
r
g £ ¤
vh = 1 − exp (−2h(b + k)) (1.8)
b +k

This looks rather complicated and the dependence of v on h is shown in

figure 1.1.
It is good to see that it has a horizontal asymptote which indicated the
terminal velocity. We can calculate the terminal velocity as follows:
r
g
lim v h = (1.9) Figure 1.4: Speed of drop fall-
h→∞ b +k
ing through moist air in m/s vs.
Which agrees nicely with our earlier finding from equation 1.6. We now height in m
have heard enough about raindrops and will go to the more interesting
rockets.

1.2 Variable mass: a body ejecting mass and the rocket equation

When dealing with rockets, or more generally bodies ejecting mass, our
approach is identical to the one we used so far. We assume a body of Backwards velocity here means a velocity
in the direction opposite to the motion
mass m ejects a small mass δm in a short time δt with a relative back-
of the original body. We define w to be a
wards velocity w. The setup is shown in figure 1.5. positive speed.
We denote the initial momentum of the body before ejecting the small
mass with p and the final momentum of the system of the body and the
ejected mass with p ′ . The change of momentum δp = p ′ − p is then:

δp = δm(v − w) + (m − δm)(v + δv) − mv (1.10)

If we multiply the above out and neglect the doubly small term δmδv we
find
δp = −δmw + mδv (1.11)

If we divide the above by δt ′ and take the limit for δt → 0 we should Figure 1.5: Body ejecting mass in
δp
remember that limδt →0 δv dv dp the lab-frame
δt = d t and limδt →0 δt = d t as in equation 1.3
δm dm
but limt →0 δt = − d t . The negative sign is needed since both δm and
δt are defined as positive quantities but the rate of change of mass of a
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 5

rocket must clearly be negative. The last ingredient to get to Tsiolkovsky’s

rocket equation is to remember that the rate of change of change of
momentum is still equal to the force acting on the system.

Tsiolkovsky’s rocket equation:

ò F = ma + w ṁ (1.12)
F = w ddm
t
+ m dd vt

You may find figure 1.5 rather unintuitive if you are thinking about a
rocket, because the direction of the ejected mass is the same as that of
the original body.
This is merely an artefact of the reference frame we have chosen. We
chose the same frame we used for the example of the drop gaining mass Figure 1.6: Body ejecting mass
which is the lab frame in which the “vapour” was at rest and the drop (lab-frame)
moved. If you transform figure 1.5 into the CM-frame in which the rocket
is initially stationary you get figure 1.6 which looks much more “rocket-
like”.

1st Example: The horizontal rocket launch

We will now apply Tsiolkovsky’s rocket equation to the launch of a rocket

in the absence of any external forces. We depict this as a horizontal
rocket launch as shown in figure 1.7.
We first want to evaluate the increase in speed ∆v = v f − v i of the
rocket after its mass reduced from an initial mass m i to a final mass m f .
Tsiolkovsky’s rocket equation with no external force reads: 0 = w ddmt +
Figure 1.7: horizontal rocket
m dd vt which can now be multiplied with d t and then separated to give:
launch
dm
d v = −w (1.13)
m
This can be easily integrated and we find the dependence of the hori-
zontal rocket’s speed increase on the initial to final mass ratio:

The maximum horizontal speed of a single stage rocket with

ò exhaust speed w and initial and final masses m i and m f

mi
µ ¶
∆v = w ln (1.14)
mf

2nd Example: The vertical rocket launch

Let us now turn our attention to the much more realistic case of a rocket
being launched vertical in the gravitational field of the earth. In what
follows we will limit our calculations to the case of a constant gravita-
tional acceleration g ≈ 9.8 ms −1 which is a realistic assumption if we
consider rockets that reach at most the low earth orbits of a few hundred
km above ground. Satellites in this region fly from 215 km (Sputnik) to
595 km (Hubble) and only polar orbiting satellites make it up to 1700 km
above ground. This will give us distances from the centre of the earth that
6 CLASSICAL MECHANICS

vary from 6,586 km (Sputnik) to at most 6,966 km (Hubble). Across this

range of distances the the gravitational acceleration varies by only 12%.
If you want to launch an actual satellite you of course have to consider
the variation in g . We take positive velocities and accelerations to point
upwards and the setup of the rocket ejecting a small mass δm in a small
time δt with backwards velocity w is shown in figure 1.8
In this case the rocket equation now reads:
Figure 1.8: Vertical rocket launch
−mg = ma + w ṁ (1.15)

As before we assume a constant mass ejection speed w as well as con-

dm
stant mass ejection rate dt = −α which gives us a simple form for the
mass as a function of time.

m = m 0 − αt (1.16)

Where m 0 is the initial mass of the rocket with all its fuel still on board.
At this point it is convenient to introduce the concept of thrust which is
best described as the rate at which the rocket ejects momentum or more
simply as the force which the engine generates which is F t = wα. We see
that the thrust corresponds to the second term on the RHS of equation
1.15. We will now separate and integrate equation 1.15 to find the change
in speed.

dv
−mg = m − wα ⇔
dt
md v = (−mg + wα)d t ⇒
Z vf Z tf µ
wα
¶
dv = −g + dt ⇔
vi ti m 0 − αt
m 0 − αt f
v f − v i = −g (t f − t i ) − w ln ⇔
m 0 − αt i
mf
∆v = −g ∆t − w ln (1.17) Equation 1.17: The speed increment of a
mi vertical rocket as a function of time and
final to initial mass ratio
The first term on the RHS of equation 1.17 is the speed a free falling
rocket would have reached in time ∆t and the second term is the speed
from equation 1.14 which a free rocket would achieve in time ∆t .
We are now in principle ready to integrate our result once more and
find the position as a function of time. Before we do so we simplify our
notation by setting the initial time t i = 0 and the initial speed v i = 0. We
dx
further denote the general speed at time t with v = dt . With this notation
equation 1.17 reads:

dx mf
= −g t − w ln
dt mi
α
= −g t − w ln (1 − t) (1.18)
m0

Before we integrate equation 1.18 we should check that our rocket is

actually going to take off at all. We need a thrust that is at least a little
bigger than the weight of the rocket and hence F t > m 0 g ⇔ wα > m 0 g
For a given mass ejection rate α we therefore need a minimal ejection
m0 g
speed w mi n > α .
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 7

It is further convenient to introduce the fuel fraction f of a rocket

which is defined as the ratio of the initial mass of fuel to the initial total
mass of the rocket.
m f uel m f uel
f = (1.19) The fuel fraction f = m0
m0
This leads us nicely to yet another useful characteristic of a rocket
which is the burn-out time T . It is simply the duration of the rocket’s
burn and is expressed via the fuel fraction as:

m0 f m f
T= (1.20) Burnout Time T = α0
α
Having found the burnout time and knowing the speed as function of
time from equation 1.18 we immediately get the maximum speed v max of
the rocket (the speed at burnout) which is:

The maximum vertical speed of a single stage rocket with

ò mass m 0 , fuel fraction f , exhaust speed w and mass ejection
rate α
g m0 f
v max = − − w ln(1 − f ) (1.21)
α

Armed with the burnout time and the fuel fraction we now will finally
separate and integrate equation 1.18 from t = 0 to T .

xf T α
Z Z
dx = −g t − w ln (1 −
t )d t (1.22)
0 0 m 0
R
To do this integral we make use of standard integral ln zd z = z ln z − z
with the substitution z = (1 − mα0 t ) → dz
dt = − mα0 → d t = − mα0 d z This gives
us the final height of the rocket as a function of T .

−g T 2 wm 0 α α α
xf = + [(1 − t ) ln (1 − t ) − (1 − t )]T (1.23)
2 α m0 m0 m0 0

From the above we remove T using equation 1.20 to find:

The maximum vertical height of a single stage rocket with

ò mass m 0 , fuel fraction f , exhaust speed w and mass ejection
rate α

wm 0 f £ ¤ g m 02 f 2
xf = (1 − f ) ln (1 − f ) + f − (1.24)
α 2α2

It is useful to analyse equation 1.24 in more detail and extract some

meaning from the various terms. The final term on the RHS of equa-
tion 1.24 is again the height that the rocket would fallen in time T and it
subtracts from the achievable height. The right angle brackets have two
terms in them. The first if the product (1 − f ) ln (1 − f ) which is overall
negative as f < 1 and hence (1 − f ) > 0 but ln(1 − f ) < 0. So this product
term reduces the achievable height and it vanishes when f = 1. It de-
scribes the detrimental effect that the “non-fuel” part of the rocket has on
8 CLASSICAL MECHANICS

its achievable height. The only positive term in equation 1.24 is therefore
wm 0 f 2
α and it grows quadratically with f .
In the limit of f → 1 we now find:

wm 0 g m 02
lim x f = −
f →1 α 2α2

As the maximal height which a rocket made 100% of fuel could achieve.

Rocket Plots

We will now look at graphs showing the velocity and height vs. time of an
example rocket, for a range of fuel fractions at a given mass ejection rate
and ejection speed.

Figure 1.9: velocity (left) and height (right) of a rocket vs. time for various fuel fractions

We assume a start weight: m 0 = 10 kg, an ejection rate: α = 1 kg/s

and an ejection speed of w = 99 m/s which is just about the require
minimum of w mi n = 98 m/s. The marginal thrust exemplifies the initially
low acceleration of the rocket. We vary the fuel fraction from 0.4 to 0.9.
We observe that the rockets with lower fuel fractions of course burn out
earlier (the thicker the line, the lower the fuel fraction). We also see the
nearly quadratic advantage in achievable height from lower fuel fraction
as predicted by the limiting form of equation 1.24.
Next we will look at similar plots but this time varying the exhaust
speed at fixed fuel fraction. This time we fix the fuel fraction at f = 0.9
and vary the exhaust speed from a low value of 33 m/s that produces an
insufficient thrust to lift the locket. We see that in these cases speeds and
heights are initially turning negative and we have to wait until the rocket
has become light enough for the fixed thrust to exceed its weight.

Single-stage vs. Two-stage rockets

We now compare a two stage rocket to an equivalent single stage rocket.

With equivalent here we mean that both rockets have the same amount
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 9

Figure 1.10: velocity (left) and height (right) of a rocket vs. time for various exhaust speeds

of fuel and the same total mass. We will also stay in low earth orbit and
hence use a constant gravitational acceleration. We will see that the two
stage rocket has a distinct advantage because it can get rid of the non-
fuel mass in its first stage before igniting its second stage. We chose the
same propulsion parameters for both rocket stages and for the single
stage rocket and we add the same additional payload to the top of either
rocket.

• fuel ejection rate: α − 500 kg/s

• exhaust speed: w = 2500 m/s

• fuel fractions of either stage: f 1 = f 2 = 0.75

• Total mass of first stage: M 0,1 = 9100 kg

• Total mass of second stage: M 0,2 = 900 kg Figure 1.11: Two stage rocket

• Total mass of single stage rocket: M 0 = M 0,1 + M 0,2 = 10, 000 kg

• Payload Mass; m = 100 kg

• Grant total mass of either rocket including payload: M i = 10, 100 kg

Because we have chosen to add a payload to the top of the rocket and but
we have defined our fuel fractions for the stages without the payload. This is common practice for rockets
because typically the same rocket can
In the calculations done so far our fuel fraction was defined as the frac-
carry a variety of payloads so it makes
tion of the initial total mass made up by the fuel. To continue using our sense to define parameters of the rockets
equations we must now define an effective fuel fraction f e f on their own

M 0,1 + M 0,2
fe f = f = 0.726 (1.25)
Mi

It is also useful to have the single stage rocket’s final mass

M f = (1 − f e f )M i = 2, 600 kg (1.26)
10 CLASSICAL MECHANICS

We will now compute the final speed of the single stage rocket. From
equation 1.20 we find:

(M 0,1 + M 0,2 ) f 10, 000 · 0.75

T= = s = 15s (1.27)
α 500
Together with equation 1.21 we can translate this into a final speed:

Mf
v max = −g T − w ln
Mi
= 3245 m/s

We briefly compare this to the escape speed for earth

1 2 G Me m
mv esc ≥ →
2 RE
s
2G M E
v esc ≥ = 11200 m/s
RE

and we notice that we are very far away from escaping earth with the
single stage rocket.
Now we repeat the above but stage by stage for the two stage rocket.

f 1 M 0,1
• The time to 1st -stage burnout is T1 = α = 13.65 s

• Mass after 1st -stage burnout: M f ,1 = (1 − f )M 0,1 + M 0,2 + m = 3275 kg

• Speed after 1st -stage burnout:

M f ,1
v max,1 = −g T1 − w ln Mi = 2681 m/s

Now address the second stage burn:

f f M 0,2
• Time to 2nd -stage burnout: T2 = α = 1.35 s

• Initial mass of 2nd -stage (1st -stage detached) M i ,2 = M 0,2 + m = 1000 kg

• Final mass of 2nd -stage M f ,2 = (1 − f )M 0,2 + m = 325 kg

• Speed after 2nd -stage burnout:

M f ,2
• v max,2 = v max,1 + −g T2 − w ln M = 4483 m/s
i ,2

This is significantly faster than the single stage but still not nearly
fast enough to escape so we would need a third stage and a better fuel
fraction to escape.
Even if we the two stage rocket does not escape it may still reach a
large height and our approximation of constant g may turn out to be
invalid. Let us check how high the two stage rocket would get. From
equation 1.24 we can find the height at which the first stage burnt out

−g T12 w Mi M (t ) M (t ) M (t ) T1
·µ ¶ µ ¶ µ ¶¸
x f ,1 = + ln −
2 α Mi Mi Mi 0
2 M f ,1
−g T1 w Mi
· ¸
= + M f ,1 ln + −1
2 α Mi M f ,1
= 14769 m
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 11

From this we can get the height at which the second stage burns out

g T12
w Mf M f ,1
· ¸
x f ,2 = x f ,1 + v max,1 T2 − + M f ln + −1
2 α M f ,1 Mf
= 17850 m

(R E2
At this final height gravity has fallen to g 2 = g ∗ (R 2 = 9.755 m/s
E +x f ,2 )
which is only 0.56% less than our assumed initial g = 9.8

Notable Rockets

Let us end this section with some famous rockets (figures 1.12 and 1.13))
from the last 50 years and note that their performance parameters have
not changed so much. What has changed however is the cost for rocket
launches which a topic beyond this course. I hope this section has given

Figure 1.12: Left: Saturn V, Apollo moon missions, 3 stages, 111 m high, 3,000 T mass, 140 T payload (incl. 3rd stage) Right: Titan III-E, Inter-
planetary probes, e.g. Voyager, 3-4 stages, 48 m high, 630 T mass, 4 T payload (incl. last stage)

you some insights into rocket science and if it has also triggered your
curiosity for this subject I strongly recommend the following resources.

• MIT lecture notes on Rockets: https://ocw.mit.edu/courses/aeronautics-

and-astronautics/16-07-dynamics-fall-2009/lecture-notes/MIT16_07F09_Lec14.pdf

• Plotting and computation Matlab scripts for rockets on this lectures

Canvas pages

• An excellent video of the James Webb Space Telescope Launch:

https://www.youtube.com/watch?v=7nT7JGZMbtM
12 CLASSICAL MECHANICS

Figure 1.13: Left: Energia carrying Buran, Buran Spacecraft, 4 Boosters, 3 stages, 59 m high, 2400 T mass, 105 T payload to LEO (incl. Buran).
Right: Falcon Heavy, Partially re-useable, 2.5 stages, 70 m high, 1,420 T mass, 63.8 T payload to LEO
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 13

1.3 Non-inertial reference frames

A non-inertial reference frame is defined as a frame in which Newton’s

laws are not satisfied because it is itself accelerating due to external
forces. We are already familiar with the Galilean transformations which
link positions and velocities of points in two inertial-frames S and S ′ .
Here we will learn how the equivalent transformations into a non-inertial
frame S ′′ will look. For simplicity we assume that all frames are initially
(at t = 0) at the same position and in the same orientation. S ′ is assumed
to be moving relative to S with constant velocity u as measured in S. S ′′
however is moving relative to S with time dependent velocity w . A point
is described in S via its position r , its speed v and its acceleration a. The
same point can also be described in frame S ′ with position r ′ , velocity v ′
and acceleration a ′ which will be given by the Galilean transformations.
To describe the same point in S ′′ we need a new set of transformations
which we given in the table below next to the Galilean transformations

Expressions in S ′ Expressions in S ′′
Table 1.1: Galilean and non iner-
(Galilean transformations) (accelerating frame)
tial transformations
r ′ = r − ut r ′′ = r − w d t
R
Position
Velocity v′ = v −u v ′′ = v − w (t )
dv′ dv d v ′′ dv dw
Acceleration dt = dt dt = dt − dt
dv dw
Force on m F (r ′ ) = F (r ) = m d t
′ ′′ ′′
F (r ) = F (r ) − m dt

The crucial difference is that the Galilean transformations left the

acceleration and hence the forces unchanged. This is not the case when
transforming into an accelerating frame. Here the force transforms and
receives a new additive term which we refer to as a fictitious force F f i c t .
This is also sometimes called an “inertial force”. The force only appears
to be present if we are describing the motion of an object in the co-
ordinates of a non-inertial reference frame.

The fictitious force observed in a non inertial reference frame

ò moving relative to an inertial frame with time dependent velo-
city w(t ).
dw Equation 1.28 gives the form for a ficti-
F f i ct = −m (1.28)
dt tious force in a frame moving with time
dependent velocity w .

Common Examples of non-inertial reference frames

In this section we will try to show how the description of forces acting
on a body changes when we change the reference frame in which we
describe them from the usual inertial frame to an accelerating frame.
In figure 1.14 we show a train carriage which is accelerating with a
constant acceleration a towards the right. Inside the carriage is a ball
hanging from a thread, tied to the ceiling. The external observer is as-
sumed to be in an inertial frame S. She uses her observations of the An observer on the surface of the earth is
of course not in an inertial frame but we
motion of the ball to infer the forces acting on the ball and she describes
assume the acceleration on the surface of
the earth to be negligibly small
14 CLASSICAL MECHANICS

Figure 1.14: A ball hanging on

a thread inside an accelerating
train carriage with an external
(inertial) observer and an in-
ternal, (co-accelerated) observer

these forces in the co-ordinates of the black, inertial co-ordinate system.

She sees that the only forces acting on the ball are the tension from the
thread and its weight. Since the thread makes an angle θ with the normal,
the tension exerts both horizontal and vertical forces F x = T sin θ and
F y = T cos θ on the ball. Since the ball is not moving vertically the force
from the tension must be balanced by the weight. She also observes the
ball to move with constant acceleration a towards the right so that F x is
not balances but causes this acceleration. The final set of force equations
she will formulate are:

F x = T sin θ = ma
X

F y = T cos θ − mg = 0
X

The internal, accelerated observer ignores the fact that he uses an accel-
erating reference frame. He observes that a force appears to be pulling
the ball away from him and that this force is balanced by the horizontal
component of the tension. In the vertical his observations are identical to
those of the inertial observer. He will therefore describe the motion of he
observes in his accelerating frame S ′ with the following force equations:
X ′
F x = T sin θ − F f i ct i t i ous
X ′
F y = T cos θ − mg = 0

In figure 1.15 we see a mass held on a string and moving at constant Figure 1.15: A mass held on a

angular speed in a circle. If we describe the forces on this mass in an string and moving with constant

inertial frame we see that the string tension is constantly accelerating angular velocity in a circle

the mass towards the centre of the circle. We see one centripetal force at
work which is given by F cent r i p = T = −mr ω2 The negative sign in the centripetal force
indicates that it points against the radial
If we describe this inside a co-rotating frame we see a "fictitious"
direction (towards decreasing radii)
force that is pulling the mass outwards and generates the tension in the
string. In this perspective the forces is a centrifugal force of the same
magnitude but opposite orientation as the above centripetal force.
F cent r i f = +mr ω2 .
Let us now consider figure 1.16 in which an observer looks at a ball
that is free floating next to her. This time the observer really cannot know
 N E W T O N ’ S L AW W I T H VA R I A B L E M A S S A N D N O N - I N E RT I A L R E F E R E N C E F R A M E S ( L E C T U R E S 11-12) 15

that she is accelerating and has no reason to assume that there are any
forces acting on either the ball or on her. She would write down F = 0 for
the total forces acting on any body in the lift.
But maybe the observer knows that she is in a lift on earth. In this case
she would have to postulate fictitious forces which act on her and on
the ball which exactly cancel the weights, causing her to write down the
following force equation for any object in the lift.

F f i ct + mg = 0

An observer in an inertial frame however knows about gravity and that

d u(t )
both the internal observer and the ball are accelerating with dt = −g
and therefore would write down the following force equation.

du
µ ¶
m − +g =0
dt

The example of an observer with a ball in an upwards accelerated lift is

left for the lectures and your homework.

Figure 1.16: an observer with a

ball in a free falling lift
2 Motion of charged particles in B and E fields (lec-
ture 13)

In this short chapter we will describe the non-relativistic motion of

charged particles in constant electric and magnetic fields. If you do
not want to take the force laws presented here merely because they are
consistent with observation, then you can look forward to your third
year course on symmetry and relativity. At that stage you will be able to
take the Lagrangian mechanics from the end of this lecture course and
apply it the Lagrangian of the electromagnetic field to obtain the equa-
tions of motion that way. Our heroes for this short topic can be seen in
figures 2.1 and 2.2 and by the end of this chapter you should be able to Figure 2.1: Hendrik Lorentz who
tell what each of these two gentlemen have contributed to the fields of first formulated the Lorentz force
electromagnetism and accelerator physics.
We start by considering only magnetic fields and then we extend
towards the case of combined magnetic and electric fields.

2.1 Magnetic Force on a Charge Particle

Figure 2.3 describes the circular trajectory of a positively charged particle

(charge = q) that starts at the origin with initial velocity u 0 in the plane
of the page (the x y plane) which is perpendicular to the magnetic field
which points out of the page (in the positive z-direction).
The magnetic force is given as F mag = q v ×B and since v is perpendic-
ular to B this force acts always perpendicular to the velocity. This means
that the magnitude of the velocity maintains a constant magnitude of Figure 2.2: Ernest Lawrence, the
|F mag | = q vB and so does the magnetic force. The magnetic force in this inventor of the cyclotron
case acts as a centripetal force and we expect the particle to perform cir-
cular motion. We can now compute the properties of this circular motion
by equating our known equation for the centripetal force in circular mo-
tion to the magnetic force. We will do so in a fairly detailed step-by-step
manner as it is the first time we do this.

F cent = F mag
mv 2
= q vB
R

The above can be solved for the radius which of circular motion s as a
function of the speed or the momentum p. We call this the Larmor or
gyro radius. Figure 2.3: Magnetic force on
a positively charged particle. B
points out of the page.
 M OT I O N O F C H A R G E D PA RT I C L E S I N B A N D E F I E L D S ( L E C T U R E 13) 17

ò R=
mv
qB
=
p
qB
(2.1)

The definition of the Larmor radius.

It is worth noting that the Larmor radius is proportional to the mo-

mentum of the particle. This can be used to determine the momentum
of the particle by measuring its radius in a magnetic field, provided you
already know what type of particle you are measuring and hence you
know its mass to charge ratio. We call this period very unimaginatively
the Larmor period.

ò τ=
2πR 2πm
v
=
qB
(2.2)

The definition of the Larmor period.

The proportionality of the Larmor radius to the particles speed led to a

speed independent period. So if you can measure the Larmor period you
can determine the particles mass to charge ratio. This ratio, if measured
sufficiently accurately, uniquely identifies each fundamental particle.
Last we state the Larmor or cyclotron frequency.

ò ω=
2π qB
τ
=
m
(2.3)

The definition of the Larmor frequency.

Armed with the above definitions we now compute the Cartesian

co-ordinates of the particles trajectory as a function from the vector
algebraic form of the magnetic force. Or in other words we will properly
solve the equations of motion. We start by defining the initial conditions
at time t = 0.

r 0 = (0, 0, 0)
v 0 = (u 0 , 0, 0)

We use Newton’s second law to relate the acceleration to the magnetic

force.

m r̈ = q v × B (2.4)

We further express the velocity v and magnetic field B in terms of co-

ordinates:

v = (ẋ, ẏ, 0)
B = (0, 0, B z )
18 CLASSICAL MECHANICS

Now we compute the cross-product using the Cartesian unit vectors i , j

and k.
¯ ¯
¯ i j k ¯¯
¯
v × B = ¯ ẋ ẏ 0 ¯
¯ ¯
¯ ¯
¯ 0 0 Bz ¯
= ẏB z i − ẋB z j (2.5)

We insert the result from equation 2.5 into equation 2.4 to find a pair of
coupled differential equations.

m ẍ = qB z ẏ
m ÿ = −qB z ẋ

If we divide the above by m and define again the cyclotron frequency

which of course is the same as the Larmor frequency we met above in
equation 2.3

qB z
ω= (2.6)
m
We get a simplified pair of differential equations:

ẍ = ω ẏ (2.7)
ÿ = −ωẋ (2.8)

We integrate ÿ with respect to time to find:

ẏ = −ωx +C (2.9)

Since ẏ(0) = 0 and x(0) = 0 we see that the integration constant C = 0 and
we can insert the newly found ẏ into equation 2.7 to find

ẍ + ω2 x = 0 (2.10)

This may look familiar to you as it is the equation for one dimensional
harmonic motion which has the solutions:

x = A 1 cos ωt + A 2 sin ωt (2.11)

Further use of the initial condition x(0) = 0 ensures that A 1 = 0 and hence

x = A 2 sin ωt (2.12)

To determine the constant A 2 we differentiate the above with respect to

time and find
ẋ = A 2 ω cos ωt (2.13)
u0
This allows us to use initial condition ẋ(0) = 0 to determine A 2 = ω and
thus:
u0
x= sin ωt (2.14)
ω
We can now turn our attention to finding the y-coordinate. We do this by
inserting equations 2.14 into equation 2.8 to find:

ẏ = −u 0 sin ωt
 M OT I O N O F C H A R G E D PA RT I C L E S I N B A N D E F I E L D S ( L E C T U R E 13) 19

This integrates once more to give

u0
y= cos ωt +C ′
ω

Using initial condition y(0) = 0 then defines the constant C ′ = − uω0 and
complete our solution of the y-coordinate.
u0
y= (cos ωt − 1) (2.15)
ω
We now remember the suggestively named cyclotron radius which is the
same as the Larmor Radius from from equation 2.1.
u0 Equation 2.16 defines the cyclotron radius
R= (2.16)
ω which defines the circular trajectory of a
particle with speed u 0 perpendicular to a
With this equations 2.14 and 2.15 begin to show a familiar form: pure magnetic field B .

x = R sin ωt
y + R = R cos ωt

If we square the above two equations and add them we find

x 2 + (y + R)2 = R 2

which is the very familiar equation of a circle of radius R with a centre at

(x, y) = (0, −R).
The extension of the above result to the case in which the initial ve-
locity and magnetic field are not perpendicular is straightforward. The
component of initial velocity parallel to the magnetic field does not gen-
erate any force and hence stays constant. This means the motion is a
Figure 2.4: Helical trajectory of a
constant motion in the direction along the magnetic field and a circle in
particle in a magnetic field where
the plane perpendicular to the field. This situation is shown in figure 2.4.
u 0 is not perpendicular to B

2.2 Motion in Magnetic and Electric fields

Let us now extend the problem and consider the simultaneous presence
of electric E and magnetic B fields. The force now is sum of the electric
and magnetic force and is called the Lorentz-Force.

The Lorentz-Force on acting on a particle of charge q and

ò velocity v in mangentic and electric fields B and E

F = q v × B + qE (2.17)

For simplicity of the following vector algebra we consider the con-

figuration shown in figure 2.5. We note that this is again a rather special
case because the magnetic field B will again turn out to be perpendicular
to the speed (which is initially zero) and B is also perpendicular to the
electric field E
The particle is initially at rest at the origin, E = E x i points along i
and B = B z k points along k. We first evaluate the cross product from
Figure 2.5: Charged particle in
perpendicular E and B fields
20 CLASSICAL MECHANICS

equation 2.17.
¯ ¯
¯ i j k ¯¯
¯
v × B = ¯ ẋ ẏ ż ¯ = ẏB z i − ẋB z j
¯ ¯
¯ ¯
¯ 0 0 Bz ¯
We find that we still do not have any force in the z-direction and the two
differential equations of motion are:

m ẍ = qB z ẏ + qE x (2.18)
m ÿ = −qB z ẋ (2.19)

We proceed in steps very similar to the ones we used in section 2.1 but
qB z
we start with the ÿ equation. We define the frequency ω = m , divide the
both equations by m, integrate ÿ using the initial condition ẏ(0) = 0 to
find ẏ = −ωx. We then insert this result into the equation for ẍ and this
time we find:
qE x
ẍ + ω2 x = (2.20)
m
This is an inhomogeneous version of equation 2.10. We use the harmonic
complementary function
This particular solution is just a constant
x 1 = A 1 cos ωt + A 2 sin ωt (2.21) and you can easily see that it solves
equation 2.20

And one particular solution of the inhomogeneous equation

qE x
x2 = (2.22)
mω2
The general solution is then the sum of the particular solution and the
complementary function.

x = x 1 + x 2 = A 1 cos ωt + A 2 sin ωt + x 2 (2.23)

Armed with this solution and the boundary conditions x(0) = 0 and
ẋ(0) = 0 we find
qE x
x= (1 − cos ωt ) (2.24)
mω2
Now we insert x into ẏ = −ωx and integrate under the boundary condi-
tion y(0) = 0 to find
qE x
y= (sin ωt − ωt ) (2.25)
mω2
If we re-write both equations in terms of x 2 we get

Figure 2.6: Trajectory of an ini-

x = x 2 (1 − cos ωt )
tially stationary charged particle
y = x 2 (sin ωt − ωt )
in perpendicular E and B fields.
which further simplifies to

(x − x 2 ) = −x 2 cos ωt
(y + x 2 ωt ) = x 2 sin ωt

If we square both of the above equations and add them we get:

(x − x 2 )2 + (y + x 2 ωt )2 = x 22 (2.26)

The interpretation of this equation is shown in figure 2.6.

Figure 2.7: Trajectories of an

initially stationary proton in per-
pendicular E and B = 0.1T fields.
 M OT I O N O F C H A R G E D PA RT I C L E S I N B A N D E F I E L D S ( L E C T U R E 13) 21

It is the equation of a circle with a moving centre. Its radius is x 2 and

hence it circumference S = 2πx 2 . Its centre moves parallel to the y-axis
2πx 2 S
on the line x = 2. The centre moves with speed u y = x 2 ω = τ = τ which
is exactly one circumference per period τ. We can therefore say that the
qE x
circle is "rolling" down the y-axis. We note further that since x 2 = mω2
qB z Ex
and ω = m the speed of the circle’s movement is u y = By

2.3 Kinetic Energy in Magnetic and Electric fields

As in the previous section we will initially consider a pure magnetic

field perpendicular to the initial particle velocity using the example of
from section 2.1 shown in figure 2.3. Since the motion is confined to the
x y-plane it is sufficient to compute velocity components v x and v y to
determine how the kinetic energy evolves in the magnetic field. We recall
here the solution for the x and y co-ordinate from equation 2.14 and
2.15.
u0
x= sin ωt
ω
u0
y= (cos ωt − 1)
ω
Differentiating the above with respect to time produced the desired
velocity components.

ẋ = u 0 cos ωt
ẏ = −u 0 sin ωt

From these we can calculate the kinetic energy components T x and T y

and the total kinetic energy T = T x + T y

1
T x = m [u 0 cos ωt ]2
2
1
T y = m [−u 0 sin ωt ]2
2
1 ¢ 1
T = mu 02 cos2 ωt + sin2 ωt = mu 02
¡
2 2
We notice that T =constant and we have therefore found an important
conclusion.

The kinetic energy of a charge particle in a pure magnetic

ò field and in the absence of other external forces is constant.

Now we can use the approach from above and apply it to the case of
combined electric and magnetic fields as described in section 2.2 and
equations 2.24 and 2.25 which we repeat below.

x = x 2 (1 − cos ωt )
y = x 2 (sin ωt − ωt )

We differentiate these with respect to time to find:

ẋ = x 2 ω sin ωt
ẏ = x 2 ω(cos ωt − 1)
22 CLASSICAL MECHANICS

From this we form the components of the kinetic energy.

1
T x = m [x 2 ω sin ωt )]2
2
1
T y = m [x 2 ω(cos ωt − 1)]2
2
We add them together and in the penultimate line we use the trigono-
metric relation sin2 θ2 = 1−cos θ
2 .

1
T = m(x 2 ω)2 (sin2 ωt + [cos ωt − 1]2 )
2
1
= m(x 2 ω)2 (sin2 ωt + cos2 ωt − 2 cos ωt + 1)
2
= m(x 2 ω)2 (1 − cos ωt )
ωt
= 2m(x 2 ω)2 sin2
2
Ex
Finally we use x 2 ω = Bz from the end of section 2.2 to give us the
kinetic energy in the initial parameters of the problem.
¶2
Ex qB z
µ
T = 2m sin2 t (2.27)
Bz 2m

We see that the kinetic energy in this case is oscillatory. The electric field
does work on the particle which is equal to the time derivative of the
kinetic energy W = and then the particle returns this energy back into
the field. The work done by the field therefore oscillates with have the
cyclotron frequency

2.4 Applications of motion in B and E fields: The Cyclotron

At the end of the short section on forces in magnetic and electric fields
we turn out attention to one of the many great application of this physics
- the cyclotron - shown schematically in figure 2.8 and next to Ernest
Orlando Lawrence who invented it in 1934 in figure 2.9.

Figure 2.8: Schematic drawing

of a cyclotron. The left shows a
top-down view. The right shows
the side view.

The cyclotron is constructed from two opposing D-shaped metal

boxes (blue and red in figure 2.8)) that are open at their centre. The boxes
are immersed in a homogeneous magnetic field that is perpendicular Figure 2.9: Inventor of the cyclo-
to the D-shapes. The boxes are also connected to an alternating voltage tron, Ernest Orlando Lawrence
(UC Berkeley) next to one of his
early cyclotrons
 M OT I O N O F C H A R G E D PA RT I C L E S I N B A N D E F I E L D S ( L E C T U R E 13) 23

source which changes the polarity every half period of the cyclotron
qB z
frequency ωc = m which we met earlier in equation 2.6. The entire
assembly is immersed in a vacuum chamber to allow the charged particle
to travel without collisions with gas molecules.
The charged particle is injected at the centre of the apparatus and
immediately accelerated by the alternating electric field towards the
other D-shape. As it travels towards and through it the B -field keeps it
on a circular track but every time particle passes the gap between the
D-shapes it increases its kinetic energy and the cyclotron radius from
equation 2.16 increases accordingly. When the the cyclotron radius
reaches the radius of the D-shaped electrodes the particle is ejected from
the cyclotron and will have reached a non-relativistic maximal energy
E max which is limited principally by the radius of the electrodes and
shown in equation 2.28.
q 2B 2R 2 Equation 2.28 defines the maximal energy
E max = (2.28) that a particle of charge q and mass m
2m
can gain in a cyclotron of radius R and
magnetic field B z .
3 Vector differentiation, Torque, Angular momentum
and circular motion (lectures 14-15)

In this chapter we will extend the familiar rules of calculus to the dif-
ferentiation of vectors and unit vectors in Cartesian and plane polar
co-ordinates. The two heroes for this chapter are shown in figures 3 and
3. Maria Agnesi was the worlds first female physics professors. She was
appointed at the University of Bologna in 1750. She wrote text books on
differential and integral calculus for students such as yourself. I want
to not here that I have been a contemporary of Dr Carol Clark, the first
female fellow of a traditionally male Oxford college, appointed at Balliol
College in 1973 in modern languages. So for academic appointments
Bologna beats Oxford Colleges by a stunning 223 years.
Rene Descartes has contributed to more aspects of science and philo-
sophy than I dare mention but he appears as our hero for this chapter
due to his immensely useful Cartesian coordinates which we will be
using extensively.
In the content of this chapter we restrict ourselves to the two dimen-
sional case of plane polar co-ordinate in order to maintain some didactic
clarity. We will also not need any more than the plane polar co-ordinates
for the rest of this course. The extension to the three dimensional case
can be found in the excellent Lecture Notes by M. Jones from Purdue and
in particular lecture on coordinate systems and unit vectors. After this we
introduce vector-valued physical observables which arise out of the dif-
ferentiation of other vectors such as the angular momentum and torque
or the angular velocity.

3.1 Differentiation of vectors with respect to time (lecture 14)

Vectors, just like scalars follow the same rules of differentiation but they
can change in magnitude and direction. The latter is new to but it is
straight forward to implement as we show in the block of equations
below. The one difference is the appearance of the outer- or vector-
 V E C T O R D I F F E R E N T I AT I O N , T O R QU E , A N G U L A R M O M E N T U M A N D C I R C U L A R M OT I O N ( L E C T U R E S 14-15)
25

product in which the order of operation is important.

d d ax d ay d az
a= i+ j+ k = ȧ x i + ȧ y j + ȧ z k (3.1)
dt dt dt dt
d da db
(a + b) = + = ȧ + ḃ (3.2)
dt dt dt
d dc da
(c a) = a +c = ċ a + c ȧ (3.3)
dt dt dt
d da db
(a · b) = ·b +a · = ȧ · b + a · ḃ (3.4)
dt dt dt
d da db
(a × b) = ×b +a × = ȧ × b + a × ḃ (3.5)
dt dt dt

Differentiating a unit vector with respect to time

We now observe a special property of unit vectors which are vectors

whose length cannot change and is equal to one. If we take the scalar
product of a unit vector with itself, here on the example of the radial unit
vector r̂ from plane polar co-ordinates, we note from r̂ · r̂ = 1 that it is a
constant. This is true of any unit vector, not just r̂ The time derivative of
a scalar product of a unit vector with itself must therefore be zero which
we use in the equation below.

d ¡ ¢ r̂
r̂ · r̂ = 2r̂ · =0
dt dt
We also know from vector algebra that two vectors which form a zero
scalar product must be orthogonal and we conclude that Equation 3.6 states that the time derivat-
ive of a unit vector is a vector perpendicu-
d r̂ d r̂ lar to it.
⊥ r̂ → ∝ θ̂ (3.6)
dt dt

where θ̂ denotes the unit vector associated with the angle θ in plane
polar coordinates as shown in figure 3.1.
We can understand this phenomenon by realising that any unit vector
such as r̂ has a fixed length but not a fixed direction. Any infinitesimal Figure 3.1: The radial r̂ and an-
change in time of r̂ must therefore be an infinitesimal vector that per- gular θ̂ unit vectors from plane
pendicular to r̂ . Adding an infinitesimal perpendicular vector to r̂ is polar coordinates
a necessary but insufficient condition for maintaining its length. We
d r̂
have found that dt ∝ θ̂ but we have not yet found what the constant of
proportionality is. To do this we use fixed Cartesian coordinates to differ-
entiate both r = r (i cos θ + j sin θ) and r̂ = (i cos θ + j sin θ) to emphasise
the difference.
Fixed here indicates that the unit vectors indicating the directions of
the coordinates do not vary with time at all. When differentiating r we
have to vary both the direction (t het a) and the magnitude (r ). When
differentiating r̂ we have to keep the magnitude r constant and equal to
one. We will now differentiate r̂ using the chain rule while keeping i and
j fixed.
d r̂ d θ
d r̂
= θ̇(−i sin θ + j cos θ)
= (3.7)
dt dθ dt
We note that the expression in brackets in the above is another unit
vector which is perpendicular to r̂ and we denote it with θ̂ = (−i sin θ +
j cos θ).
26 CLASSICAL MECHANICS

We note that θ̇ is the rate of change of the angle of r̂ in plane polar

coordinates as shown in figure 3.2.
This means we have found the desired constant of proportionality in
the derivative of r̂
r̂˙ = θ̇ θ̂ (3.8)

Differentiating θ̂ in the same way gives us:

˙
θ̂ = −θ̇ r̂ (3.9)

Differentiating the radius vector with respect to time - aka velocity

We now continue with the differentiation of r = r r̂ via the product rule to Figure 3.2: The radial r̂ and an-
find: gular θ̂ unit vectors in relation to
v = ṙ = r˙r̂ + r r̂˙ the radius vector r
Using our result from equation 3.8 we find the expression for the velocity
in plane polar coordinates

v = r˙r̂ + r θ̇ θ̂ (3.10)

The components of the velocity in plane polar motion are shown in figure
3.3
If r describes the circular motion of a particle then the radius of this
motion is fixed so that r˙ = 0 and we denote θ̇ = ω. In this special case of
circular motion the velocity is given by:
Figure 3.3: The components of
v = r ωθ̂ (3.11)
velocity in general plane polar
The single component of the velocity in circular motion is shown in co-ordinates
figure 3.4

Differentiating the velocity with respect to time - aka acceleration

We start from equation 3.3 and differentiate its two terms separately
using equations 3.8 and 3.9 to find:

d ¡ ¢
r˙r̂ = r¨r̂ + r˙θ̇ θ̂
dt Figure 3.4: The components of
d ³ ´
r θ̇ θ̂ = −r θ̇ 2 r̂ + r θ̈ θ̂ + r˙θ̇ θ̂ velocity in circular plane polar
dt
co-ordinates
Adding these two components we get the expression for the acceleration

a = r̈ = r¨ − r θ̇ 2 r̂ + 2r˙θ̇ + r θ̈ θ̂
¡ ¢ ¡ ¢
(3.12)

As before we also consider the special case of plane and constant circular
motion in which the radius r is constant and the azimuthal acceleration
θ̈ is zero so that r¨ = r˙ = θ̈ = 0. We again denote θ̇ = const = ω and realise
that v = ωr . Inserting the above into equation 3.12 provides us with the
familiar centripetal acceleration for constant circular motion

a = −r θ̇ 2 r̂
= −r ω2 r̂
v2
=− r̂ (3.13)
r
 V E C T O R D I F F E R E N T I AT I O N , T O R QU E , A N G U L A R M O M E N T U M A N D C I R C U L A R M OT I O N ( L E C T U R E S 14-15)
27

3.2 Angular momentum and torque (lecture 14)

Let us start by defining the angular momentum in terms of observables

we already know and show it in figure 3.5.

The angular momentum J of a single point particle around

ò and origin O is defined as:

J =r ×p (3.14)

where r is the displacement of the particle from the origin and

p is the linear momentum of the particle. We note that the an-
gular momentum is dependent on the choice of origin around
which we define it. Figure 3.5: Definition of angular
momentum of a point particle
By its definition through the vector product the direction of J is per-
pendicular to the plane of motion.
Let us differentiate J with respect to time

dJ dp dr
=r × + ×p (3.15)
dt dt dt

dp dr
and remember that dt = F and dt = v and p = mv to find:

dJ
= r ×F +v ×p
dt
= r ×F + v × vm
= r ×F (3.16)

dp
In analogy to the linear force F = dt which is the time derivative of the
linear momentum we call the time derivative of the angular momentum
the "angular force" or torque τ.

The torque τ generated by a force F acting on a body at potion

ò r from an origin O is defined as: Figure 3.6: Torque around z-axis
dJ
τ= = r ×F (3.17)
dt
we note that the torque, just like the angular momentum is
dependent on the choice of axis around which we define it.

To emphasise the dependence of the torque on the arbitrarily chosen

axis around we define it we now consider the torque created by a force
acting on a given point when we change the reference axis as is shown in
figures 3.6 and 3.7.
We denote the torque which force F acting at position r generates
around the z-axis (figure 3.6) with τO and the torque from the same force Figure 3.7: Torque around a shif-
but around an axis parallel to the z-axis and shifted by displacement R ted axis
28 CLASSICAL MECHANICS

with τ A . The two torques are related by equation ??

τ A = r A × F = (r − R) × F
= r ×F −R ×F
= τO − R × F (3.18)

If multiple forces F i act on a body at multiple positions r i the torque

dJ
becomes τ = d t = ni=1 r i × F i
P

3.3 Angular motion variables and circular motion (lecture 14-

15)

In the following section we will define angular equivalents of velocity and

acceleration and show the particularly simple forms of angular velocity,
angular momentum and angular acceleration in circular motion. We
will also show what form work and power take in this special case. It is
important to realise that circular motion comes in two forms.

Constant circular motion More accurately we should refer to this type

of motion as "constant speed" circular motion. In this form of motion
the velocity vector v only changes direction but the magnitude of
velocity v (the speed) remains constant. The acceleration vector
vec t a continuously changes direction to always point in the radial
direction but it remains fixed in magnitude. Note that in this case we
have a constant speed although we do have a non-zero acceleration.
Examples of constant speed circular motion include the circular orbits
under central forces.

Accelerated circular motion Do not be mislead by this name. All circular

motion is accelerated! So more rigorously we would call this motion
"non-constant speed" circular motion. In this type of motion we do
not only change the direction of velocity but also its magnitude. An
example could be a body that is driven to rotate around a fixed axis
with variable ω.

Angular velocity in circular motion (lecture 14)

We start by implicitly defining the angular velocity ω in circular mo-

tion by equation 3.19 which relates it to the linear velocity v as the time
derivative of the position vector r .

v = ω×r (3.19)

The above should me sense to you because in circular motion the velo-
Figure 3.8: Orientation of velocity
city is indeed always perpendicular to the radius vector and for fixed ω it
and radius in circular motion
grows proportional to r as seen in figure 3.8.
Next we remember equation 3.11 (v = r ωθ̂) which already found the
linear velocity for the case of circular motion in terms of r and ω. We seek
to match the implicit definition from equation 3.19 with that of equation
3.11.
To do so we remember that θ̂ and r̂ are perpendicular to each other
as shown in figure 3.8. This allows us construct θ̂ from r̂ and a new unit
 V E C T O R D I F F E R E N T I AT I O N , T O R QU E , A N G U L A R M O M E N T U M A N D C I R C U L A R M OT I O N ( L E C T U R E S 14-15)
29

vector n̂ which defines the plane of motion by being perpendicular to

that plane. This construction gives us:

θ̂ = n̂ × r̂ (3.20)

Using this construction in equation 3.11 we find:

v = r ωn̂ × r̂
= ωn̂ × r̂ r
= ωn̂ × r (3.21)

Comparing equations 3.21 and 3.19 we that they match each other if we
explicitly define the angular velocity in circular motion as:

ω = ωn̂ (3.22)

Angular momentum in circular motion (lecture 14)

We now look at the angular momentum in the special case of circular

motion. From its definition J = r × p and equation 3.19 we conclude that
in circular motion
J = r × (m ṙ ) = mr × (ω × r ) (3.23)

Using the vector identity for the triple vector product

a × (b × c) = (a · c)b − (a · b)c (3.24)

We can re-express J as follows:

J = mr × (ω × r ) = mr 2 ω − m(r · ω)r (3.25)

If we now recall the r and ω are perpendicular and thus r · ω = 0 the

expression for the angular momentum in circular motion becomes:

J = mr 2 ω = I ω (3.26)

where we have defined I as the moment of inertia of single point mass.

We will soon generalise this (lecture 21) to the case of multiple masses
and solid bodies.

Angular acceleration in circular motion (lecture 15)

We start by defining angular acceleration α in general and analogously to

the definition of linear acceleration a = v̇ .

α = ω̇ (3.27)

Since in circular motion the plane of motion and hence also its defining
unit vector n̂ are constant we can use equation 3.22 to evaluate the above
general definition in the case of circular motion to find:

α = ω̇n̂ (3.28)

So in contrast to the linear acceleration vector a which constantly

changes direction in circular motion its counterpart, the angular accel-
eration vector α has a constant direction. If we also consider the special
case of constant circular motion we find that α is entirely constant.
30 CLASSICAL MECHANICS

If we consider constantly accelerated circular motion then α (the mag-

nitude of α) can change at a constant rate and hence we can integrate the
equation of motion to obtain expressions for ω and θ.

dω
α= = (const .) →
dt
ω = ω0 + αt (3.29)

We can integrate this further to find

dθ
ω= →
dt
1
θ = θ0 + ω0 t + αt 2 (3.30)
2
Equations 3.29 and 3.30 are exactly analogous to two of your well known
"SUVAT" equations.
Finally we briefly mention the angular equivalent of the well known
relation between Force and linear acceleration F = ma (for constant
masses). The torque τ as the angular equivalent of Force has already
dJ
been defined via equation 3.17 (τ = dt = r × F ). But how does τ relate to
α?
We merely mention this relation here and leave its derivation as the
subject of chapter ?? (chapter not written yet!).

τ = Iα (3.31)

Here I is called the moment of inertia and this form is analogous to the
linear force equation F = ma.

Work and Power in circular motion (lecture 15)

We begin by defining an infinitesimal amount of work done by a force

F acting on a body along an infinitesimal path length d s. Remember
that here and throughout this chapter we are still in the special two
dimensional case of planar motion.

dW = F d s (3.32)

The total work done arises from the above by integration so that
I
W = Fds (3.33)

It is worth noting that the above is not a general approach. The work
done when a body is moved from one position to another is in general
dependent on the path taken if we allow non-conservative forces such
as friction to be involved. Under such circumstances there is no exact
differential for the work. So in what follows we limit ourselves to the case
of conservative forces.
Now we seek to re-express work in terms of the circular equivalents of
vec t F and s which are vec t τ and θ. We can do so using the general form
indicated in figure 3.9.

d s = d (θ × r ) (3.34)

Figure 3.9: Arc segment d s in

terms of angle change d θ
 V E C T O R D I F F E R E N T I AT I O N , T O R QU E , A N G U L A R M O M E N T U M A N D C I R C U L A R M OT I O N ( L E C T U R E S 14-15)
31

Using this in equation 3.33 gives the following

dW = F · d (θ × r )
= F · (d θ × r + θ × d r )

To progress we use the vector identity for the triple product

a · (b × c) = (c × a) · b

which applied once to the first gives us:

dW = (r × F ) · d θ + (d r × F ) · θ
= (r × F ) · d θ + θ · (d r × F )
= τ · d θ + (F × θ) · d r

In the last line we used the definition of torque from equation 3.17.
Z Z
W= τ · dθ + (F × θ) · d r (3.35)

So we see that in the general case (but on only with conservative forces)
the work gets contributions from an integral along the azimuthal d θ
direction and from one along the radial d r direction.
In circular motion however the radius is constant and the second
integral vanishes so the work in circular motion is
Z
W= τ · d θ in circular motion (3.36)

Last we recall that for constant circular motion d θ = ωd t which allows us

to write Z
W= τ · ωd t in constant circular motion (3.37)

dW
We can use the above together with the definition of Power P = dt to
find the expression for power in circular motion.

P = τ · ω in constant circular motion (3.38)

3.4 Correspondences between linear and angular quantities

To summarise the work in the last chapter we provide a table that lists the
relations between the linear and the angular quantities.

3.5 Newton’s laws for angular motion and their use

Let us state initially Newton’s three laws in their form for angular motion.

Newton-1: In the absence of a net applied torque τ, the angu-

ò lar velocity ω remains constant.

τ = 0 ⇒ ω = const. (3.39)
32 CLASSICAL MECHANICS

Figure 3.10: Table comparing linear and equivalent rotational properties

Newton-2: Torque τ is the rate of change of angular mo-

ò mentum J .
dJ
τ= (3.40)
dt
or for circular motion. The torque is equal to product of the
moment of inertia and the angular acceleration α

τ = Iα (3.41)

Newton-3: For every applied torque τi n , there is an equal and

ò opposite reaction torque τout .

τi n = −τi n (3.42)

Example of angular motion: The simple pendulum

We will now use our equations for angular motion to obtain the equa-
tions of motion for problems that we have solved before using the equa-
tions for linear motion. We start with the example of a simple pendulum
as shown in figure 3.11.
To use equation ?? we need to find expressions for torque and angular
momentum. We start with the definition of torque from equation 3.17.
With the geometry from figure 3.11 this gives us: Figure 3.11: A point mass on
a massless rigid rod forming a
τ = r × F = −mg r sin θ ẑ
simple pendulum
Now we use the definition of angular momentum from equation 3.14 and
 V E C T O R D I F F E R E N T I AT I O N , T O R QU E , A N G U L A R M O M E N T U M A N D C I R C U L A R M OT I O N ( L E C T U R E S 14-15)
33

this gives us:

J = r × p = r mv ẑ

We now take the time derivative of the angular momentum, remem-

bering that the radius and mass of the pendulum are constant and we
re-express the speed and its time derivative via v = r θ̇ → v̇ = r θ̈ to get:

˙
J˙ = r m(r θ̈ ẑ + r θ̇ ẑ)

If we now insert τ and J̇ into ?? and obtain:

−mg r sin θ ẑ + 0ẑ˙ = r 2 m(θ̈ ẑ + θ̇ ẑ)

˙ (3.43)

where I have added the zero valued term 0ẑ˙ into the left hand side to
make it clearer that equation 3.43 represents two equations, one in the
ẑ direction (in the plane of motion) and one in the ẑ˙ (out of the plane
of motion). These two directions are perpendicular to each other as we
found in equation 3.6 and hence the equations are independent. For
out-of-plane of motion we have:

r 2 m θ̇ ẑ˙ = 0

which means that the absence of out-of-plane torque keeps the motion
in a fixed plane. This is essentially Newton-1 for the out-of-plane direc-
tion. In the plane of motion we get
g
θ̈ + sin θ = 0
r
Which, after approximating for small angles (sin θ ≈ θ) becomes the
standard equation for harmonic motion.

Example of angular motion: The ladder

Note that in static problems torque is often called “Moment of force” or

“Moment”. Let us now consider a ladder leaning against a wall as shown
in figure ??
What we want to know is under which conditions the ladder will
remain stationary. This problem is more complicated than it may look
initially. We need to ensure that the ladder neither translates is centre
of gravity nor does it rotate. To not rotate we must ensure that all the
torques on the ladder sum to zero about a chosen point so we demand in
general
n
!
τt ot =
X
ri ×Fi = 0
i =1
Figure 3.12: A ladder leaning
We chose the point O in the figure to evaluate the torque balance equa-
against a wall.
tion. There are two torques acting on the ladder. The first one, τwei g ht
come from the projection of the ladder’s weight perpendicular to the
ladder which is cos θmg and it acts in the ladder’s centre of gravity.
It generates torque around O with a lever arm of L/2 and magnitude
τwei g ht = cos θmg L2 . The second torque τw al l arises from the projection
of the reaction force N2 from the vertical wall. It points in the opposite
direction to τwei g ht and has magnitude τw al l = −N2 L sin θ But careful.
Demanding that these torques sum to zero is not correct now because a
34 CLASSICAL MECHANICS

rotation in the counter clockwise direction is geometrically impossible if

the wall and point O are fixed. We must only ensure that the ladder does
not rotate clockwise. So this gives a torque balance inequality.

L
mg cos θ − N2 L sin θ ≥ 0 (3.44)
2
To not translate all the forces on the ladder must sum to zero. We
consider this in the vertical and horizontal directions separately. In the
vertical we demand that the weight is compensated by the reaction force
N1 from the ground. In the horizontal the reaction force from the wall
N2 must be compensated by the for of friction F s = µN1 where µ is the
coefficient of static friction (stiction). So the two force equations are:

N1 = mg
N2 = µN1

which nicely combine into one equation µmg = N2 which we can insert
into 3.44 to find.
L
mg cos θ − µmg L sin θ ≥ 0 ⇔
2
cos θ − 2µ sin θ ≥ 0

If we assume a typical coefficient of stiction µr ub−concr et e = 0.5 between

the rubber feet of the ladder and a wet concrete floor we get cos θ = sin θ
which is satisfied for θ ≥ 45o .
4 Central Forces (lectures 15-20)

In this long chapter we will define central forces and find out that our
definition of central forces restricts them to also being conservative
forces. We will then find the general equations of motion under central
forces and find out various properties of the resulting motion which will
turn out to be planar. We will investigate a special class of central forces
that follow an inverse square law such as gravity and electro-static forces.
For the inverse square law of gravity we will define an effective potential
and using this new concept will derive interesting orbit properties such
as the elliptical, parabolic or hyperbolic orbit geometries or the points
of closest and furthest approach. We will learn how to steer a satellite
between orbits and after learning how to reduce the "two body problem"
to an effective "one body problem" we will apply what we have learned to
scattering processes.

4.1 Definition of Central Forces

Let us begin this section by defining what we mean by a central force.

A force acting on a body is considered central if

ò • it acts parallel or anti-parallel to the radius vector r from
the origin of an inertial reference frame to the body and This definition is more restrictive than
those found in many text books where a
central force is merely acting along the
• its magnitude only depends on the magnitude of |r | = r
radius vector with magnitude allowed to
depend on the direction of r or even on ṙ
F cent r al = f (r )r̂ (4.1) or time t .

Where f (r ) is any function of r .

The two most well known examples of central forces are

m1 m2
Gravitational force : F g r av = −G r̂ (4.2)
r2
1 q1 q2
Electro − Static force : F el ect r o−st at i c = r̂ (4.3)
4πϵ0 r 2
The vector r in equation 4.1 is strictly the radius vector to the origin of an
inertial frame while in the two example forces it is the vector between the
locations of two masses or charges, none of which is generally at rest in
any inertial frame. For the time being we will assume that one of the bod-
ies is at rest in an inertial frame. This can be made approximately true if
that mass is much heavier than the other one. If such an approximation
the problem of finding and solving the equations of motion of the less
36 CLASSICAL MECHANICS

massive body is called a "Single body Problem". In lecture 19 we will

see how to drop the above approximations and reduce the proper "Two
body problem" through the consideration of the centre of mass and the
reduced mass into an equivalent "Single body Problem".

4.2 Conservative Forces

We will soon show that any central force as defined in equation 4.1 will
also be conservative. To give this statement meaning we begin by de-
fining what a conservative force is. Below we provide three statements
that define a conservative force. These statements are equivalent to each
other and as we will soon see, each one reveals particularly useful proper-
ties of conservative forces.

All conservative forces meet all of the following three equival-

ò ent conditionsF .

1. The curl of the force is zero:

∂F z ∂F y
   
∂y − ∂z  0
 ∂F ∂F z   
x
∇ × F cons = 
 ∂z − ∂x 
= 0 (4.4)
∂F y ∂F x
− 0
∂x ∂y

2. The work done by the force over any closed path is zero: Equivalently to condition 2. we could
I demand that the work done between two
W = F cons d r = 0 (4.5) points P and Q is independent of the path
taken between these points.

3. The force is the gradient of a scalar potential U (r ) :

 ∂U 
∂x
F cons = −∇U (r ) =  ∂U (4.6)
 
∂y 
∂U
∂z

We start showing the equivalences of equations 4.4,4.5 and 4.6 by

proving that statements 1. and 2. are equivalent. To do this we use Stokes
theorem from vector calculus which states.

The integral of the curl of any vector field F in R3 over a sur-

ò face A is equal to the line integral along the closed path S
surrounding that surface.
Ï I
∇×F d A = F ·ds (4.7)
A s

Figure 4.1: Surface A surrounded

The right hand side of equation 4.7 is equal to the work W done by the by line S on which two points P
force field F along the closed path S. If F meets condition 1. (F satisfies and Q are located.
equation 4.4) then the left hand side of equation 4.7 is zero so must be its
right hand side which is the work done. This means that any F satisfying
condition 1. must also satisfy condition 2.
Let us now show that conditions 1. and 3. are in fact mathematically
 CENTRAL FORCES (LECTURES 15-20) 37

equivalent. From vector calculus we will have learned that the gradient of
any scalar field has to have a vanishing curl or ∇ × ∇U (r ) = 0. We can see
this easily when we look at it in co-ordinate form.

∂ ∂U ∂ ∂U
   
− ∂z 0
 ∂y ∂z ∂y 
∂ ∂U ∂ ∂U   
(4.8)
∂z  0
 ∂z ∂x − ∂x =
∂ ∂U ∂ ∂U
∂x ∂y − ∂y ∂x 0

Equation 4.8 holds for any doubly differentiable scalar field U as the
order of taking the derivatives does not matter.
Since we have now shown that conditions 1. and 2. are equivalent and
conditions 1. and 3. are equivalent it follows that conditions 2. and 3. are
also equivalent.

4.3 Central Forces are conservative

We will now show that any force meeting a restricted form of condition 3.
(equation 4.6) will also be central. With a restricted form of condition 3.
we mean a form that identifies a subset of all conservative forces.

If a force F satisfies F cons = −∇U (|r |) it is conservative and

ò central and it can be written in the form F (r ) = f (r )r̂

The restricted condition allows the potential U only to depend on

the magnitude of r and not on its direction. Let us show the above by
∂ ∂ ∂
using Cartesian co-ordinates. We find: −∇ = î ∂x + ĵ ∂y + k̂ ∂z and r =
p
2 2 2
x + y + z from which we derive:

∂r −x x
=p =−
∂x 2 2
x +y +z 2 r
∂r −y y
=p =−
∂y 2 2
x +y +z 2 r
∂r −z z
=p =−
∂z x2 + y 2 + z2 r

We are now equipped to compute ∇U (|r |) via the chain rule as follows.

∂U (r ) ∂r ∂U (r ) ∂r ∂U (r ) ∂r
−∇U (|r |) = −î − ĵ − k̂
∂r ∂x ∂r ∂y ∂r ∂z
à !
∂U (r ) î x ĵ y k̂ z
= + +
∂r r r r
à !
∂U (r ) î x + ĵ y + k̂ z
=
∂r r
∂U (r ) r ∂U (r )
= = r̂ = f (r )r̂
∂r r ∂r
Q.E.D.

We have now shown that −∇·U (r ) = f (r )r̂ which we reformulate in words

as follows.
38 CLASSICAL MECHANICS

The gradient of a scalar potential which depends only on the

ò magnitude of the radius vector r always points along the ra-
dial direction.A force that can be written as such a gradient is
therefore central by definition and conservative via subset of
condition 3.

4.4 The equation of motion under central forces

5 Appendix A

We start from equation 3.12