PHYS 419: Classical Mechanics Lecture Notes

I. WORK AND ENERGY

We now want to introduce the concept of energy and derive the mechanical energy conser-
vation theorem. Note the words derive and theorem. In contrast to most elementary courses
on classical mechanics which present this subject as a new physical law, we will show that
it is the consequence of Newton’s laws. Thus, it is a theorem.

A. Force field

When we have discussed vectors, we stressed that in 3-dimensional space vectors are
objects that can be determined by specifying exactly three components (three numbers).
These numbers determine the direction, sense, and length of the vector, but not “the orig-
ination point” for a vector. Thus, all vectors with the same direction, sense, and length
are equivalent. In other words, the origin of a vector is arbitrary and we can for simplicity
assume that all vectors originate from the same point in space, for example the point [0,0,0].
Then the complete vector space is the set of all possible vectors originating from this point
(see the figure).

While this definition is sufficient for many purposes, in physics one often encounters
objects that are vectors but are connected with a given point in space. For example, the
gravitational force that a body exerts on another body depends on the position in space
where we measure this force, see the figure. The force at point r1 is different from the force
at point r2 . If instead of this body (body 1 in the figure), the field is due to another body
(body 2), the force vectors are different (broken-line arrows in the figure). By changing the
position and mass of the body, one can produce a force of arbitrary direction and magnitiude.
Thus, at each point of the “position” space there is located a separate vector space. We
differentiate between the vector spaces located at different r’s by writing the vectors as
F (r). Now we have six components in each object F (r), so we can have vectors assigned to
different origins. If we chose some value for F at each r, we say that a vector field has been

1
determined in space. Note that the vector spaces at different r are separate vector spaces,
so it does not make sense (and there is no need) to talk about the sums like F (r1 ) + F (r2 ).
In a general case, the force field can depend on time and it will then be denoted as
F (r, t). The force may also depend on the velocity of the particle interacting with the field:
F (r, v, t).

B. Work performed by a force field

From everyday experience, work is connected with the magnitude of the force acting
during the process of performing the work and the magnitude of the displacement of a body.
Thus, the following definition should appear to be “natural”. We first define infinitesimal
work by the force F (r) displacing the body from r to r + dr as

dW = F · dr. (1)

Now assume that a body is moved from point r1 to point r2 along path C, as shown on the
figure. Then the work done by the force field F on the body is
Z r2
W12 = F (r) · dr. (2)
r1 :C

In general, the work does depend on the path, so the path has to be specified. The definition
of work can be used both for the net force acting on the body (i.e., the sum of all forces)
and for each of the particular forces acting on this body.

Note on line integrals:

The line integral appearing in Eq. (2) is defined as follows. We first write the curve C in a
parametric form, i.e., the points on the curve are given by three functions

r(t) = [x(t), y(t), z(t)].

The begining and end points correspond then to the values of t equal to t1 and t2 , respectively,
i.e., ri = r(ti ). Then the definition is
Z r2 Z r2
F (r) · dr = [Fx (x, y, z)dx + Fy (x, y, z)dy + Fz (x, y, z)dz] =
r1 :C r1 :C
Z t2 £ ¤
= Fx (x(t), y(t), z(t)) x0 (t) + Fy (x(t), y(t), z(t)) y 0 (t) + Fz (x(t), y(t), z(t)) z 0 (t) dt (3)
t1

2
where x0 (t) = dx(t)/dt. The last integral is just an integral of a single variable, so we know how to
compute it. Note that an often made mistake in calculating the line integrals is to use the middle
form in Eq. (3) and treat each of the three terms appearing there as one-dimensional integrals (with
appropriate assumptions about the values of the other two variables and the limits of integration).
While in principle this procedure can lead to the correct answer, in practice it is very easy to make
errors in this approach. The only case when such approach can be recommended is when the path
of integration consists of linear segments parallel to the axes of the coordinate system.
A useful way of looking at the line integrals is also the Riemann definition. One divides the
integration path into a number of segments of length δr. Then the integral is approximated by a
sum of (scalar) contributions F (rζ ) · (ri+1 − ri ), where ri are the endpoints of the segments and
rζ is a point between ri and ri+1 . As the number of segments goes to infinity, the Riemman sum
approaches the value of the integral.

C. Kinetic energy

As some forces act on a body, and the net force is nonzero, the body accelerates and the
magnitiude of the velocity is changing. We know from experience that depending on this
magnitude and the mass of the body, the effects of a collision of the body with another object
can be quite different. One quantity which depends on velocity and mass is the momentum,
a vector quantity. Here we will define a scalar quantity which will reflect this changing state
of a body.
Consider again a body of mass m which is moving from point r1 to point r2 along path C.
The body is accelerated. Let us try to relate the work W12 to the initial and final velocities
of the body, v1 and v2 . Let F (r) be the net force acting on the body. We therefore have:
F = ma and we can write Eq. (2) as
Z r2 Z r2 Z t2 µ ¶ µ ¶
dv dr
W12 = F · dr = m a · dr = m · dt
r1 :C r1 :C t1 dt dt
Z t2 Z Z
m t2 d(v · v) m t2 dv 2
= m (v̇ · v) dt = dt = dt
t1 2 t1 dt 2 t1 dt
m 2 ¯¯t2 mv22 mv12
= v ¯ = − = T2 − T1 (4)
2 t1 2 2
where we have defined
mv 2
T = . (5)
2
This quantity is called the kinetic energy. We have used in the derivation the identity
d
(v · v) = v̇ · v + v · v̇ = 2(v̇ · v).
dt

3
We have also used the definition of the integral
Z
df (x)
dx = f (x) + C.
dx
Thus, the work performed on a particle by the net force is equal to the change of the kinetic
energy of this particle:
W12 = T2 − T1 (6)

This equation is sometimes called “work–kinetic energy” theorem.

D. Conservative force field

We define the conservative force field as a field having the property that the integral (2)
is independent of the path C. In other words, the field is conservative if and only if
Z r2 Z r2
F (r) · dr = F (r) · dr (7)
r1 :C1 r1 :C2

for any path Ci between the two points (see also the figure below). The reason for using the
name “conservative” will be given later.

We can now prove a theorem

H
Theorem 1 The force F (r) is conservative ⇔ F (r) · dr = 0.

Proof (⇒):
I Z r2 Z r1
F (r) · dr = F (r) · dr + F (r) · dr
Zr1r:C
2
1
Zr2r:C
2
2

= F (r) · dr − F (r) · dr = 0 (8)

r1 :C1 r1 :C2

since the field was assumed to be conservative. The paths involved are illustrated on the
figure on the right above.
Proof (⇐):
Consider the figure below. We have two closed path integrals, one over the path C1 + C2
and another one over C1 + C20 . Each integral is zero by assumption. Since the integrals have
a common part over C1 , the integral over C2 must be equal to the integral over C20 .

4
 We need one more theorem which introduces the important concept of the potential
energy.
H
Theorem 2 F (r) · dr = 0 for any closed path ⇔ there exists a function φ(r) such that
F (r) = ∇φ(r).

Proof (⇒):
Consider the integral from r0 to r over path C. Due to Theorem 1, this integral is indepen-
dent of the choice of the path. Thus, for some fixed r0 , the integral
Z r
F (r 0 ) · dr 0
r0

(there is no C in the symbol of integral to stress the independence of path) gives always the
same number for a given r. Thus, as we change r, the integral defines a unique function of
r: Z r
F (r 0 ) · dr 0 = Hr0 (r). (9)
r0

Notice that if we choose another starting point, for example r00 marked on the figure below,
the new function Hr00 (r) differs from Hr0 (r) only by a constant, equal to the value of the
integral from r00 to r0 .

Now consider the point in space which is shifted with respect to r by x00 along the x̂ axis
(see the figure above on the right). The value of Hr0 at this point is
Z r+x00 x̂ Z r+x00 x̂
00 0 0
Hr0 (x + x , y, z) = F (r ) · dr = Hr0 (x, y, z) + F (r 0 ) · dr 0 . (10)
r0 r

The last integral in the equation above is just a one-dimensional integral since on this path
dr 0 = dx0 x̂. Therefore, we can write
Z r+x00 x̂ Z r+x00 x̂
0 0
F (r ) · dr = Fx (x0 , y, z)dx0 (11)
r r

5
where Fx denotes the x component of F . In this integral, the values of y 0 and z 0 are fixed
and equal to y and z, respectively, as utilized in writing the integrand. Thus, the integral
R
written above is equivalent to the standard integral of a single variable g(x0 )dx0 . Denoting
by G(x) the antiderivative of g(x), i.e.,
dG(x)
g(x) = ,
dx
we can use the definition of the integral:
Z x+x00 Z x+x00
0 0 dG(x0 ) 0
g(x )dx = 0
dx = G(x + x00 ) − G(x).
x x dx
Now, comparing this to Eqs. (10) and (11), we can write
∂Hr0
Fx = (12)
∂x
Similar relations can be obtained for Fy and Fz by considering appropriate displacements
from r. We therefore can write:
F (r) = ∇Hr0 (r) (13)
and the function φ(r) in the theorem is equal to Hr0 (r) plus an arbitrary constant

φ(r) = Hr0 (r) + const.

Proof (⇐):
If F (r) = ∇φ(r) then
I I I
F (r) · dr = ∇φ(r) · dr = dφ = φ(r0 ) − φ(r0 ) = 0 (14)

where we have used

∂φ ∂φ ∂φ
dφ = dx + dy + dz = ∇φ(r) · dr.
∂x ∂y ∂z
This completes the proof. The intermediate step in (14) can be understood better by writing
explicitly the definition of the line integral
I Z t2 Z t2 · ¸ Z t2
∂φ dx ∂φ dy ∂φ dz dφ
∇φ(r) · dr = ∇φ(r) · ṙdt = + + dt = dt.
t1 t1 ∂x dt ∂y dt ∂z dt t1 dt

One may remark here that the equation we have proved above
Z r2
∇φ(r) · dr = φ(r2 ) − φ(r1 )
r1

is the three-dimensional equivalent of the familiar formula

Z x2
dG(x)
dx = G(x2 ) − G(x1 ).
x1 dx

6
Furthermore, in future work we will often see Eq. (9) written as:
Z r
φ(r) = φ(r0 ) + F (r 0 ) · dr 0 .
r0

In physics, the function φ is traditionally taken with sign reversed, denoted by the letter
U , and called the potential energy, so that we have

F (r) = −∇U (r). (15)

We can therefore write the relation between work and the potential in one more form
Z r2
U (r1 ) − U (r2 ) = F (r) · dr. (16)
r1

We can now formulate the final form of Theorem 2

Theorem 2a: For any conservative force field F (r) there exists a scalar function U (r)
such that F (r) = −∇U (r). The function U (r) is called the potential energy.

As it is obvious from the theorems discussed earlier, the inverse theorem is also true, i.e.,

Theorem 2b: If a force field can be written as F (r) = −∇U (r), then this force field is
conservative.

E. Curl of F

There exists another useful relation for checking if a force field is conservative. It utilizes
the familiar Stokes theorem
Z I
(∇ × F ) · dσ = F (r) · dr (17)
S C

where the first integral is over the surface S and the second integral is over the curve C
bounding S. The differential dσ is a vector perpendicular to the surface S at a given point
and having the magnitude of the surface area element, see the figure. Recall that the curl

7
of a vector is defined as: ¯ ¯
¯ x̂ ŷ ẑ ¯
¯ ¯
¯ ∂ ∂ ∂ ¯
∇ × F = ¯ ∂x ∂y ∂z ¯ .
¯ ¯
¯ Fx Fy Fz ¯

We can now formulate the following theorem

Theorem 3:
I
∇×F =0 ⇔ F (r) · dr = 0 (⇔ F is conservative)
C

where C is an arbitrary closed contour in some region of space. The proof of this theorem
results immediately from the Stokes theorem. In ⇒ direction, ∇ × F = 0 implies that the
integral on the left-hand side of Eq. (17) is zero, so the integral on the right-hand side has
to be zero as well. In the ⇐ direction, we can choose as C any contour on the surface S.
Thus, the surface integral similar to that on the left-hand side of Eq. (17), but taken over
an arbitrary part of S, must be zero. This is possible only if the integrand is zero. If this
is not yet convincing, realize that if the integral on the left-hand side of Eq. (17) (over the
complete S) is zero as implied by the assumption, and the integrand is not zero, this means
that there are positive and negative contribution to this integral which cancel each other.
Then consider the subarea which gives a positive contribution (such that the integrand is
positive everywhere). This subarea is bounded by a contour and the line integral over this
contour is zero by assumption. Therefore, the surface integral has also to be zero and cannot,
in fact, be positive. This implies again that ∇ × F = 0.

F. Mechanical energy conservation theorem

If a force field is conservative, we can combine Eqs. (2), (15), (16), and (6) together
Z r2
W12 = F (r) · dr = U (r1 ) − U (r2 ) = U1 − U2 = T2 − T1
r1

which implies
T1 + U1 = T2 + U2 .

Since the points r1 and r2 are arbitrary, it follows that

T + U = E = const. (18)

where T and U are the kinetic and potential energies, respectively, at some arbitrary point
and we have denoted their sum by E. This sum will be called the (total) energy of the
system. Equation (18) is the mechanical energy conservation theorem.