See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/259666484

A primer on work-energy relationships for introductory physics

Article  in  The Physics Teacher · January 2005

DOI: 10.1119/1.1845983

CITATIONS READS

25 426

1 author:

Carl Mungan
United States Naval Academy
168 PUBLICATIONS   1,500 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Carl Mungan on 23 January 2014.

The user has requested enhancement of the downloaded file.

A Primer on Work-
Energy Relationships for
Introductory Physics
Carl E. Mungan, U.S. Naval Academy, Annapolis, MD

T
here has been and continues to be consider- in an inertial reference frame, where the center-of-
able discussion in the educational commu- mass work done on the system is
nity about different ways of relating the con-
cepts of work and energy in introductory physics.1 The Wc.m. ≡ ∑ ∫ Fi ⋅ dri , (2)
present article reviews a consistent and streamlined parts

treatment of the subject, drawing particular attention

to aspects seldom covered in textbooks. The paper is and the translational kinetic energy of the system is
intended to clarify the central equations for introduc-
tory courses and to put the wider literature in context. K ≡ ∑ 12 mi υi2 . (3)
It is specifically designed to tie closely in terminology parts

and order of presentation to standard texts, so that it Note that Eq. (1) is perfectly general. In particular,
complements rather than supplants them. In brief, it is applicable to deformable objects such as a verti-
the key point is that there are two major categories of cal chain falling into a pile on a surface and to open
work, center-of-mass work and particle work.2 After systems undergoing irreversible processes such as a
an overview of these two approaches, I illustrate them block sliding on a rough table.5
with a couple of instructive examples that can be used Several other comments on these equations help
in group problem-solving sessions in class. clarify them further:

Center-of-Mass Work (i) The work on individual part i is due both to

In what is usually called the work-energy theo- external forces (i.e., exerted by agents not included
rem, one is concerned with center-of-mass work and in the system) and to internal forces (e.g., the force
mechanical energy. This relation is most useful in part j exerts on part i where i  j). That is, the con-
mechanics, and it is a theorem in that it can be derived cept of internal work is well defined and useful in
starting from Newton’s laws.3 Suppose an object4 i has mechanics, unlike in thermodynamics. For example,
mass mi and that net force Fi acts on it while its center consider a system of two parts, a book and a rough
of mass (c.m.) undergoes a differential displacement table. After being given an initial push, the book
dri , so that its c.m. velocity is υi  dri /dt. If we take comes to rest because of the internal force of friction.
our system to be composed of a set of objects, which To the extent that the table is heavy enough that
will be referred to from now on as parts, it is easy to we can neglect its recoil, (negative) center-of-mass
prove that work is done only on the book and the system loses
kinetic energy.
Wc.m. = K (1)
(ii) The kinetic energy of a system depends on how

10 DOI: 10.1119/1.1845983 THE PHYSICS TEACHER ◆ Vol. 43, January 2005

you partition it and thus one must clearly specify is the integral of the net torque on the part over its
not only the system but also its parts. For example, macroscopic angular displacement, and
consider a Frisbee® of mass m and moment of iner-
K rot ≡ 12 I ω 2 (6)
tia I at temperature T thrown in the conventional
way so that it spins with rotational speed ω as it is the rotational kinetic energy of the part. The
sails through the air with translational speed υ. If torque , differential angular displacement dθ,
we take our system to be a single part, the Frisbee angular speed ω, and the moment of inertia I must
as a whole, then K = 12 mυ 2 ; one might call this all be evaluated about the same axis, which must
the macroscopic view. On the other hand, if we either pass through the c.m. or be the instantaneous
take this system to be composed of a large number axis of rotation, to avoid noninertial corrections.
of bits of plastic that are small compared to the Furthermore, Eq. (4) applies only to an object whose
size of the Frisbee, but large compared to molecu- moment of inertia is constant. As a counterexample,
lar dimensions, then K = 12 mυ 2 + 12 I ω 2 , which in the familiar demo where a point mass is swung in
one might call the mesoscopic view. Finally, if one a circle at the end of a string of decreasing length,
resolves the Frisbee into its N individual atoms, then the rotational kinetic energy of the mass (about the
K = 12 mυ 2 + 12 I ω 2 + 32 NkBT (at high temperatures, center of the circle) increases even though the torque
typically valid at room temperature6 ), where kB is on it is always zero. Equation (4) does not apply in
the Boltzmann constant; this is the microscopic view. this case because I decreases as the string is short-
While this is often confusing initially to students, ened.
an appreciation of the fact that kinetic energy can Equation (1) can be recast into another common
be “hidden” inside an object in this way is crucial to form by introducing potential energy. Split the net
the development of the concept of internal energy. center-of-mass work into the sum of the work done
To emphasize this, Eq. (2) has sometimes been by all conservative forces, Wc , and the work done by
called pseudowork in the literature,7 although I rec- all nonconservative forces, Wnc. Assuming that the
ommend against use of this name because it is not system is chosen to be encompassing enough that all
used in standard textbooks, and this form of work conservative forces are internal, Wc can be moved to
is no more a “false” construct than is particle work, the right-hand side to obtain
described in the next section.
Wnc = Emech. (7)
(iii) The proof of Eq. (1) from a spatial integration
of Newton’s second law exactly parallels the deri- Here Emech is the sum of the kinetic energies K of
vation of the impulse-momentum theorem start- every part and the potential energies U associated
ing from a temporal integration. However, linear with every conservative force acting between all pairs
momentum of a system is always equal to the vector of parts in the system.
sum of the momenta of the particles composing it.
Consequently one does not need to qualify impulse Particle Work
with adjectives such as “center-of-mass.” At the particle8 level, the energy of a system chang-
es only if work is done by external forces,9
The rotational analog of Eq. (1) for any individual
part rotating about a principal axis is Wparticle = E, (8)

Wrot = Krot, (4) as measured by an inertial observer. Here Wparticle

(sometimes called the external1 or real7 work) is the
where sum of the work done on every particle in the system
by all external forces,
Wrot   dθ • (5)
Wparticle particles Fexti • dri ,
 (9)

THE PHYSICS TEACHER ◆ Vol. 43, January 2005 11

where Fexti is the net external force (i.e., exerted by processes, as illustrated by myriad examples involv-
agents external to the system) on particle i during its ing ideal gases in introductory texts. But there exist
displacement dri. (I assume that all relevant objects differences of opinion among educators about the
that do work on each other via “action at a distance” magnitude of the heat transfer Q in many irrevers-
forces are included in the system to avoid the issue of ible processes.13 For example, during the sliding of
the work done on and by fields. Also, for notational a block on a rough table, the particles on the con-
consistency I continue to assume that all objects that tacting surfaces of the block and table are neither in
do work via conservative forces are included in the thermal nor mechanical equilibrium.14 It is probably
system. Both of these assumptions can be relaxed in best in such situations to discuss the energy trans-
subsequent, more advanced treatments.) The sum fer between them in terms of particle work using
of the mechanical Emech and internal Eint energies is a model such as that of Sherwood and Bernard,15
the total energy E of the system. Internal energy10 without attempting to distinguish thermodynamic
of a system is a sum over that of its parts. In turn, work from heat. Since the process is irreversible, Q
internal energy of a part is an inertial-frame-invari- is not directly related to the entropy change of the
ant state property and includes all stored energy11 system in any case.
except bulk translational kinetic energy of the part’s
c.m. (which depends on the frame of reference of the Problem 1: A pulled spool that rolls
external observer and thus is not a property of a part without slipping
alone) and bulk potential energy between the parts A free-body diagram for the horizontal forces on a
(which depends on the interactions between them cylindrically symmetric spool of mass m, outer radius
and hence is not “owned” by either part). R, and moment of inertia I is sketched in Fig. 1.
Equivalently one can think of particle work as a
sum of the line integral of each external force over the
displacement of the point of application of that force,

Wparticle = forces Fexti • dri .

 (10)

For the special case where the system is isolated, so

that no external work is done on it, then

Emech = –Eint , (11)

which is a general statement of conservation of

energy. By way of examples, the mechanical energy
lost by a block sliding on a rough table reappears
mainly as vibrational energy of the molecules on the
contacting surfaces of the table and block, and the Fig. 1. Free-body diagram of a spool (with c.m. at B) being
pulled by a constant tension T applied to its string unwrap-
mechanical energy gained by an accelerating figure
ping from the top point C of the inner cylinder of radius r. The
skater comes from the chemical energy of previously spool makes contact with a rough, horizontal table at point A
eaten food. on its outer cylinder of radius R. Since the spool rolls with-
In thermodynamics, the particle work is often out slipping, the friction f is static and is initially assumed to
categorized according to whether the energy transfer point opposite to the direction of pulling.

is adiabatic or thermal (i.e., driven by a temperature

difference),
Wparticle = Wthermo + Q. (12) The spool starts from rest and the string (wrapped
around inner radius r) is pulled horizontally with a
This distinction12 is unambiguous for reversible constant tension T, causing the spool to roll without

12 THE PHYSICS TEACHER ◆ Vol. 43, January 2005

slipping a distance L along a level, rough table. Find contrast to kinetic friction on a sliding object. This is
its final translational speed υ. not a consequence of the fact that static friction does
First consider the solution using the concept of par- zero particle work! For example, the tension also
ticle work. Taking the system to be the spool (viewed does not change the thermal energy, nor does the
macroscopically), its change in mechanical energy is static friction on, say, a box that is not slipping in
1 2
2 mυ , while the only form of internal energy that the flat bed of an accelerating truck.
1
changes is the spool’s rotational kinetic energy 2 Iω2,
where its angular speed is ω. There are two external (iv) Nevertheless, static friction does in general alter
forces, the tension T and the static friction f. The the nonthermal internal energy of the spool, namely
displacements of their points of application are L(1 its rotational kinetic energy, and this in turn changes
+ r /R) and zero, respectively.16 Therefore, Eq. (8) the energy available for translation!
becomes
We gain additional insights into the physical situ-
 r
TL 1 +  = 12 mυ 2 + 12 I ω 2 . (13) ation by using the concept of center-of-mass work.
 R
Let the system again consist of a single part, the spool.
Introducing the mass distribution factor, Equation (1) becomes
  I/(mR2), and the no-slip condition, ω = υ/R,
now leads to the solution, (T – f )L = 1
2 mυ2, (15)

2TL(1 + r / R ) while Eq. (4) evaluated about the c.m. of the spool is
υ= . (14)
m(1 + γ ) L 1 2
(Tr + fR ) = 2 Iω . (16)
This solution becomes that of an object that rolls R
without slipping down an incline whose vertical drop Adding these two equations together reproduces
is h, when we replace the external force T by mg and Eq. (13) and hence the solution Eq. (14). But we
its distance of application L(1 + r /R) by h. (Another now also learn some things about friction. If the
interesting variation with the latter solution is to sus- frictional force is directed backward (as in Fig. 1),
pend the spool vertically by its string and allow it to the negative sign on the left-hand side of Eq. (15)
fall a distance h as it freely unwinds like a yo-yo.) implies that friction slows the spool down trans-
Instructors need to be cognizant of a number of lationally (compared to what would occur if the
potential stumbling blocks for students attempting to coefficient of friction were zero), while the positive
internalize this approach: sign on the left-hand side of Eq. (16) indicates that
it simultaneously speeds the object up rotationally.
(i) The displacements of the points of application Furthermore, by multiplying Eq. (15) by  and
of the external forces are not intuitively obvious. (In equating it to Eq. (16), the frictional force is found
contrast, calculation of the center-of-mass work only to be
involves the overall translational and angular dis- γ −r /R
placements, which are familiar to students from their f = T. (17)
study of kinematics and dynamics.) γ +1
(Dividing the absolute value of this result by mg
(ii) I have implicitly assumed that the spool and gives the minimum value of the coefficient of static
table are rigid, so that rolling friction can be friction if the spool is not to slip for a given pulling
neglected. This issue may arise after the next point is force.) Note that the mass distribution factor can be
brought to the attention of students. altered within the range 0 <  < 1. In particular, if
the center of the spool is made heavy enough, the
(iii) Static friction does not change the thermal por- frictional force becomes negative, indicating that it
tion of the internal energy of the spool,17 in striking is in the direction of motion.18 In that case, friction

THE PHYSICS TEACHER ◆ Vol. 43, January 2005 13

speeds the object up translationally (at the expense since by symmetry the two blocks have equal and
of its rotations) so that this spool would outpace an opposite velocities υrel relative to the c.m. This can
identical one being pulled with the same tension on be solved for υrel and thus the velocities of each
an air table! block can be calculated from υ1 = υc.m. + υrel and
υ2 = υc.m. – υrel, which is (c). As a check, one can
Problem 2: Pushing on a deformable apply Eq. (7) or (8) with the system treated meso-
system scopically as having two parts, namely the two
Suppose two rigid blocks of mass m are at rest on blocks, to obtain
a level, frictionless surface and are connected by a
massless spring of stiffness constant k that is initially Fx1 = 1
2
mυ12 + 1
2
mυ22 + 1
2
k(x2 – x1)2. (20)
relaxed with length L. A constant inward force F is
suddenly applied to block 1, displacing it a distance x1 Substitution of the above forms for υ1 and υ2 repro-
in the direction of the second block. During this time, duces Eq. (19). Finally, with respect to (d), note that
the spring coupling causes block 2 to move a distance the c.m. moves with constant velocity υc.m. after the
x2 in the same direction, as sketched in Fig. 2. force F is removed. So jump into the center-of-mass
(a) Find the resulting velocity of the center of mass, frame. Here one sees that the length of the spring
υc.m., of the pair of blocks. (b) Find the total vibra- oscillates sinusoidally about L with an angular fre-
tional energy (kinetic plus potential) of the system, quency of 2 k/m
 and an amplitude of 2E vib / k .
Evib. (c) Find the instantaneous velocity of each This setup is analogous to two massive pistons that
block, υ1 and υ2. (d) If the force F is now removed, can slide frictionlessly along a horizontal pipe with
describe the subsequent motion of the system. an ideal gas between them. If the far piston is instead
clamped in place and the near piston is pushed to-
x1 x2
ward it, this becomes a traditional thermodynamics
k problem. The compression can be either reversible or
F m m irreversible depending on whether the pushing force
per unit area of the movable piston is infinitesimally
or arbitrarily larger than the gas pressure (and hence
Fig. 2. Free-body diagram of a pair of equal masses m con- than the clamping force per unit area of the fixed pis-
nected by an ideal spring k and resting on a level, frictionless ton), respectively. This in turn will determine via Eq.
table. A constant force F is applied to the left-hand mass, (1) whether the gas will acquire bulk kinetic energy in
causing it to translate a distance x1 rightward while the right- addition to the change in its internal energy.19
hand mass moves x2 rightward.
Conclusions
The entire system viewed as one part has mass 2m. Work is always defined as a force integrated over a
The net force acting on it is F while the c.m. moves a displacement. However, students must be brought to
distance of (x1 + x2)/2, beginning from rest and end- consciously consider which forces and displacements
ing with speed υc.m.. Hence Eq. (1) implies are involved. Depending on the context, they might
be asked whether they are including (i) internal or ex-
ternal, (ii) conservative or nonconservative, (iii) field
F 12 ( x1 + x 2 ) = 12 (2m )υc.m.
2
, (18) or contact, and (iv) random (thermal) or organized
forces in their calculations. In the case of the displace-
which immediately gives the solution to (a). On the ments, the relevant options are (i) center of mass ver-
other hand, Eq. (8) becomes sus point of application, and (ii) translational versus
angular.
Fx1 = 12 (2m )υc.m.
2
+ E vib , (19) It helps to explicitly point out that for a particle,
thereby solving (b). The nonthermal internal energy Wc.m. = Wparticle. It is only for objects that can rotate,
of the system is Evib = 2( 12 mυ2rel) + 12 k(x2 – x1)2 deform, or undergo irreversible changes that center-

14 THE PHYSICS TEACHER ◆ Vol. 43, January 2005

of-mass and particle work provide distinct and com-
plementary information about the behavior of a sys-
tem.20 Broadly speaking, center-of-mass work relates
to the bulk kinetic and potential energies of a system.
This is primarily of interest in mechanics problems
(particularly when direct solution of Newton’s laws
would prove difficult). For example, center-of-mass
work tells us that a net external force (usually static
friction) is needed if a car is to accelerate along a level
road. On the other hand, particle work is useful when
we are seeking to account for the sources and sinks of
energy. Returning to the same example, it is the inter-
nal energy of the gasoline that powers the car along
the road. Together then, Wc.m. and Wparticle give us a
balanced view of the mechanical universe, and both
should be presented in an introductory course.

Acknowledgments
My thinking has been sharpened by discussions with
David Bowman, John Denker, Chuck Edmondson,
John Fontanella, Jim Green, Ludwik Kowalski,
John Mallinckrodt, Gene Mosca, Joel Rauber,
Bob Sciamanda, and Bruce Sherwood. I thank the
Research Corporation for its support.

References
1. For example, see A.J. Mallinckrodt and H.S. Leff, “All
about work,” Am. J. Phys. 60, 356–365 (April 1992).
2. In partial agreement with R.C. Hilborn, “Let’s ban
work from physics!” Phys. Teach. 38, 447 (Oct. 2000), I
encourage a careful verbal distinction between different
kinds of work and an avoidance of the unsubscripted
symbol W.
3. See the discussion of the “point-particle system” in
R.W. Chabay and B.A. Sherwood, Matter & Interac-
tions I: Modern Mechanics (Wiley, New York, 2002),
Chap. 7.
4. In this context, objects are assumed to be classical and
to have mass. Thus, gravitational and electromagnetic
fields, and massless particles, springs, strings, and rods
are excluded; their role is limited to mediating the in-
teractions between objects.
5. A.B. Arons, “Development of energy concepts in intro-
ductory physics courses,” Am. J. Phys. 67, 1063–1067
(Dec. 1999).
6. This corresponds to a high-temperature specific heat
per atom of 3kB = 4.14  10-23 J/K (known as the
Dulong-Petit rule), evident, for example, in Fig. 3 of
R.W. Chabay and B.A. Sherwood, “Bringing atoms

THE PHYSICS TEACHER ◆ Vol. 43, January 2005 15

 into first-year physics,” Am. J. Phys. 67, 1045–1050 lent stirring, and the like.
(Dec. 1999). 13. M.W. Zemansky, “The use and misuse of the word
7. B.A. Sherwood, “Pseudowork and real work,” Am. J. ‘heat’ in physics teaching,” Phys. Teach. 8, 295–300
Phys. 51, 597–602 (July 1983). (Sept. 1970). Also see R.H. Romer, “Heat is not a
8. A particle is any object having no accessible internal de- noun,” Am. J. Phys. 69, 107–109 (Feb. 2001); R.P.
grees of freedom. It need not be microscopic. The rigid, Bauman, “Physics that textbook writers usually get
isothermal, smooth, nonrotating blocks of introductory wrong: II. Heat and energy,” Phys. Teach. 30, 353–356
physics are a macroscopic example. (Sept. 1992); G.M. Barrow, “Thermodynamics should
be built on energy—Not on heat and work,” J. Chem.
9. There are no dissipative forces at the particle level. If
Educ. 65, 122–125 (Feb. 1988); and M. Tribus, “Gen-
there were, it would be possible for internal forces to
eralizing the meaning of ‘heat’,” Int. J. Heat Mass Trans-
change the energy of an isolated system! (Internal forces
fer 11, 9–14 (Jan. 1968).
can change the net mechanical energy of two interact-
ing macroscopic parts, but only with a compensating 14. H. Erlichson, “Are microscopic pictures part of mac-
change in the internal energy of these same parts.) For roscopic thermodynamics?” Am. J. Phys. 54, 665 (July
example, the prototypical nonconservative force, fric- 1986) and S.G. Canagaratna, “Critique of the treat-
tion, actually results from a sum over (conservative) ment of work,” Am. J. Phys. 46, 1241–1244 (Dec.
electrostatic forces between the electron shells of the at- 1978).
oms in the contacting surfaces. In consequence, what is 15. B.A. Sherwood and W.H. Bernard, “Work and heat
called internal work in W.H. Bernard, “Internal work: transfer in the presence of sliding friction,” Am. J. Phys.
A misinterpretation,” Am. J. Phys. 52, 253–254 (March 52, 1001–1007 (Nov. 1984). Also see the broken tail-
1984) is simply the negative of the change in the total hook spring model in R.P. Bauman, “Physics that text-
(bulk plus internal) potential energy of the system. book writers usually get wrong: I. Work,” Phys. Teach.
10. Internal energy as used here is not identical with what 30, 264–269 (May 1992).
thermodynamics texts usually refer to as internal ener- 16. While the c.m. of the spool translates by L, the spool
gy. In the present context, it includes the energy of bulk rotates through an angle of θ = L/R. During this time,
rotations and other nonthermalized internal modes. See a length l = rθ of string is unwound. The total displace-
M. Alonso and E.J. Finn, “On the notion of internal ment of point C is thus L + l = L(1 + r /R). Note that
energy,” Phys. Educ. 32, 256–264 (July 1997). this is consistent with the familiar results that point A
11. Stored energy is traditionally divided into categories (equivalent to r = –R) has zero net velocity, point B (r =
called “forms of energy” such as electric field energy, 0) has the c.m. velocity υ, and point D (r = R) has twice
gravitational potential energy, chemical energy, and the c.m. velocity.
so on. But as argued in G. Falk, F. Herrmann, and 17. As noted in C. Carnero, J. Aguiar, and J. Hierrezuelo,
G.B. Schmid, “Energy forms or energy carriers?” Am. “The work of the frictional force in rolling motion,”
J. Phys. 51, 1074–1077 (Dec. 1983), why should one Phys. Educ. 28, 225–227 (July 1993), nonconservative
distinguish forms of energy any more than, say, forms forces need not be dissipative.
of charge? It is only the carriers of the energy or charge, 18. C.E. Mungan, “Acceleration of a pulled spool,” Phys.
and not the energy or charge per se, that change. Teach. 39, 481–485 (Nov. 2001).
12. For example, H. Erlichson, “Internal energy in the first 19. C.E. Mungan, “Irreversible adiabatic compression of an
law of thermodynamics,” Am. J. Phys. 52, 623–625 ideal gas,” Phys. Teach. 41, 450–453 (Nov. 2003).
(July 1984) distinguishes work from heat on the basis
20. H.S. Leff and A.J. Mallinckrodt, “Stopping objects
of whether the force is macroscopic or microscopic.
with zero external work: Mechanics meets thermody-
With the growing prevalence of microelectromechani-
namics,” Am. J. Phys. 61, 121–127 (Feb. 1993).
cal systems (MEMS), I think most physicists would be
uncomfortable with this. A somewhat more common PACS codes: 46.04A, 46.05B, 01.40Gb
approach is that of B.A. Waite, “A gas kinetic explana-
tion of simple thermodynamic processes,” J. Chem. Carl E. Mungan grew up in Calgary and received a bach-
elor's degree in theoretical physics at Queen's University
Educ. 62, 224–227 (March 1985), who distinguishes
in Canada. He subsequently obtained a Ph.D. in experi-
work and heat on the basis of whether the interaction mental condensed matter physics from Cornell University
is organized or random. While the logic behind this is (NY). Recently he has been thinking about the thermody-
evident for gases being acted upon by pistons or bunsen namics of laser radiation and cooling.
burners, it is far less helpful when thinking about dis- Physics Department, U.S. Naval Academy
sipation via electrical resistance, kinetic friction, turbu- Annapolis, MD 21402-5040; mungan@usna.edu

16 THE PHYSICS TEACHER ◆ Vol. 43, January 2005

View publication stats