Ch.

2--Derivatives

Chapter 2

DERIVATIVES, DIFFERENTIALS, AND

DIFFERENTIAL EQUATIONS

Note: THIS CHAPTER IS FOR REFERENCE ONLY. If you are

feeling insecure about your under-
standing of Calculus, read Chapters 2
and 3 closely. Otherwise, skim them. position
In either case, don't panic if the
material looks obscure. We'll ease
into it gently (besides, the first test
will primarily be over Chapter 1). x(t+ t) Point B

Point A
A.) Derivatives--Preliminaries: x(t)

1.) An eccentric's pet ant is con-

t t+ t time
strained to move in one dimension.
Figure 2.1 presents a graph of its
motion in time. FIGURE 2.1

a.) At time t, the ant is lo-

cated at Point A. While
there, its position coordinate
is x(t). At time t + ∆ t, the position
ant is located at Point B with
a position coordinate x(t +
∆ t). Pt. B
x(t + t)
b.) In Figure 2.2, a se- secant
cant is drawn between Points rise is x

A and B. x(t)
Pt. A
run is t

i.) Note that the

t t+ t time
slope of the secant is
equal to the secant's
"rise over run," or: FIGURE 2.2

25
 x ( t + ∆t ) − x ( t )
slope =
( t + ∆t ) − t
x ( t + ∆t ) − x ( t )
= .
∆t

2.) If we hold Point A fixed while allowing ∆ t to become very small,

Point B approaches Point A and the secant approaches the tangent to the curve
at Point A. Writing this out mathematically, we get:

slope of tangent = slope of secant

x(t + ∆t) - x(t)
= lim ∆t → 0 .
∆t

Note: Although the denominator goes to zero, the ratio of the nu-
merator and denominator converges to a finite number.

a.) This operation is called a derivative. In general, a derivative

yields a new function that defines the rate of change of the original
function with respect to one of its variables. In the above case, it defines
the rate of change of "x" with time.

b.) Books from different disciplines denote derivative operations

in different ways.

i.) Some math books, for instance, use the notation:

x(t + ∆t) - x(t)

x' (t) = lim ∆t→0 .
∆t

ii.) Most physics books, on the other hand, use the notation:

d(x) x(t + ∆t) - x(t)

= lim ∆t→0 .
dt ∆t

Note: In the latter notation, dt is NOT d multiplied by a time t any

more than ∆ t is ∆ multiplied by a time t.

c.) Mathematicians treat dx/dt as a single symbol denoting a

particular mathematical operation (a derivative). As you will see
shortly, physicists are not so formal in their treatment of the notation.

26
 Ch. 2--Derivatives

3.) Bottom line: In general, the time derivative of a time-varying

function f(t) gives you a second function df(t)/dt that defines the slope of the
tangent to f(t) at any point on f(t)'s curve. Put another way, it gives us a
general function that defines the rate at which f(t) changes with time. This
function can then be evaluated at any given time-of-interest t.

B.) Derivatives--In the Beginning . . . :

1.) In most beginning Calculus classes, the definition of the derivative:

df (t) f (t + ∆t) - f (t)

= lim ∆t→0 ,
dt ∆t

is used to generate derivative-type expressions for relatively common func-

tions. Although the approach is limited, it is effective in presenting the
theoretical underpinnings of the derivative. As such, we will do a few ex-
amples to show how it works.

a.) Consider the function f(t) = 3t +2 (see Figure 2.3). What is the
time rate of change of the function? That is, what is the new function
that defines how f(t) changes as t changes?

i.) This is an admit-

tedly easy problem. Seeing
position
that the function is linear
and noting that its slope
(i.e., its rate of change) is f(t) = 3t + 2
constant and equal to 3, we x

know the answer to the

derivative calculation before
t
we start. Nevertheless:

ii.) Following through time

with the definition of the
derivative: FIGURE 2.3

27
 df (t) f (t + ∆t) - f (t)
= lim ∆t → 0
dt ∆t
[3(t + ∆t) + 2] - [3t + 2]
= lim ∆t → 0
∆t
3t + 3∆t + 2 − 3t − 2
= lim ∆t → 0
∆t
3∆t
= lim ∆t → 0
∆t
= 3.

iii.) As the ∆ t's in the fourth line canceled, we didn't need to

invoke the limit (i.e., have ∆ t go to zero). Still, the process leads
us to a function (a constant in this case) that defines the rate at
which x = 3t + 2 changes with time.

b.) Consider the function x(t) = kt3 (see Figure 2.4 below), where k
is a proportionality
constant equal to 1
m/s3 (without k, x
position
would have the units
of seconds cubed).
What is the rate of
change of the func- x(t) = kt 3
tion? That is, what is
the new function that
defines how x changes where k = 1 m/s 3
as t changes?

i.) This is not time

a trivial situa-
tion. Setting k = FIGURE 2.4
1 (we don't need
to carry it through the calculation as it was included only to make
the units acceptable), and following through with the definition of
the derivative, we get:

28
 Ch. 2--Derivatives

dx(t) x(t + ∆t) − x(t)

= lim ∆t→0
dt ∆t

= lim ∆t→0
[ (t + ∆t) ] − [ t ]
3 3

∆t
t + 3t (∆t) + 3t(∆t)2 + (∆t)3 − t 3
3 2
= lim ∆t→0
∆t
3t (∆t) + 3t(∆t)2 + (∆t)3
2
= lim ∆t→0
∆t
[
= lim ∆t→0 3t 2 + 3t(∆t) + (∆t)2 ]
= 3t 2 2
(or 3kt if we replace the " k").

ii.) What happened to all the ∆ t's? Some were canceled; some
went away when we invoked the limit.

iii.) What does this mean? At t = 2 seconds, for instance, the

body is at x = kt3 = (1 m/s3)(2 s)3 = 8 meters. In addition, according
to our derivation, the slope of the tangent to the curve at that time
(i.e., the body's rate of change of position with time, a.k.a. its
velocity) is dx/dt = 3kt2 = 3(1 m/s3)(2 s)2 = 12 m/s.

c.) If we had done this process for k2t2, we would have found the
derivative equal to 2k2t (try it!). If we had done the process for k4t4,
the derivative would have been 4k4t3.
Generalizing, we can write:

If
x = k1t, then dx/dt = 1k1.
If
x = k2t2, then dx/dt = 2k2t.
If
x = k3t3, then dx/dt = 3k3t2.
If
x = k4t4, then dx/dt = 4k4t3.
etc.

From this we can deduce that if:

29
 x = kntn, then dx/dt = nkntn-1.

d.) This is exactly what Calculus students do for the first month
or so of the Derivatives section of their course. They use the definition
of the derivative to determine specific derivative functions, then they
generalize if possible.

e.) Other derivative rules that might come in handy:

i.) d(Ae-kt)/dt = -Ake-kt, where A and k are constants;

ii.) d[sin (kt)]/dt = k cos (kt);

iii.) d[cos (kt)]/dt = -k sin(kt);

iv.) d[ln (kt)]/dt = k (1/t), where "ln" designates a natural log.

v.) Noting the (k/t) = kt-1, we get d(kt-1)/dt = (-1)kt-2 = -k/(t2);

vi.) d[f(x)g(x)]/dx = [df(x)/dx][g(x)] + [f(x)][dg(x)/dx] (this is

called the product rule). Example: d[(4x)(x2)]/dx = (4x)(2x) + 4(x2)
= 12x2 (note that (4x)(x2) = 4x3, and that the derivative of 4x3 is,
indeed, 12x2 . . . it works).

C.) The Chasm Between Mathematicians and Physicists:

1.) As has already been mentioned, mathematicians have very def-

inite ideas about how calculus-oriented terms are to be defined and ma-
nipulated. On the other hand, physicists are (within limits) willing to play
fast and loose with the notation. Specifically:

a.) Mathematicians demand that dt be read as "an infinites-

imally small time interval" (i.e., no real duration at all). Physicists
think of dt as a very, very small time interval. The appeal of the latter
description? It implies there exists a tiny period of time during which
something can happen.

b.) A similar situation holds for the symbol dx. From a

mathematician's perspective, dx is "an infinitesimally small dis-
placement with no physical reality to it at all." Physicists treat dx as a
very, very small displacement.

30
 Ch. 2--Derivatives

c.) Although dx/dt denotes a specific operation, physicists are more

than happy to treat dx and dt as separate entities, even to the extent of
manipulating them algebraically. For instance, dx/dt = c implies that the
derivative of a function x equals a constant c. Mathematicians use for-
mal algorithms and transforms to show that if this is true, it is also true
that dx = c dt (this reads a differential change of position "dx" equals a
constant "c" times the differential time interval "dt" over which the dis-
placement takes place). Physicists know how to use the algorithms, etc.,
but aren't willing to bother. Instead, they short cut the process by simply
multiplying both sides of dx/dt = c by dt yielding dx = c dt. Mathema-
ticians hate this kind of nonchalance, but it works if you're only inter-
ested in the bottom line.

2.) To what does this all come down? As aesthetically irritating as it

might be to mathematicians, the Calculus is both easier to visualize and
easier to use in the context of real world problems when dx's and dt's are
afforded a physical significance.

3.) An example: Understanding the Chain Rule:

a.) Consider the function f(x) = x2.

i.) Note that it is easy to see how the function changes with
changes of position as its derivative with respect to x is:

df (x)
= 2x.
dx

ii.) IMPORTANT: There is nothing wrong with manipulating

this into the form df(x) = (2x)dx, then replacing the 2x by df(x)/dx.
This probably seems a bit circular, but in doing so we get an
interesting differential equation. Specifically:

df (x) = (2x) dx
 df (x) 
= dx.
 dx 

This equation states that the differential change of the function

(i.e., df) as one moves some differential distance dx along the x axis
equals the rate at which the function changes with "x" (i.e., df/dx =
2x) times the SIZE of the change (i.e., the displacement dx).

31
 b.) Assume now that x changes with time (i.e., x = x(t)).
Specifically, for example, assume that x(t) = kt3. How does the
function f(x) change with time (i.e., what is df(x)/dt)?

c.) The approach that allows us to determine df/dt is called the

Chain Rule. Stated mathematically, it is:

df (x)  df (x)   dx(t) 

= .
dt  dx   dt 

The question? Where did this come from and what does it mean?

d.) In Part 3a-ii, we decided that the change of the function df as

we move some differential distance dx along the x axis equals the rate
at which the function changes as we proceed along the x axis (i.e., df/dx)
times the SIZE of the differential displacement (dx). Mathematically,
this was written:

 df (x) 
df (x) =  dx (Equation A).
 dx 

e.) We can follow a similar path in expressing the differential

displacement dx in terms of dt. That is, the change of the position
function dx as we move through some differential time interval dt
equals the rate at which x changes in time (i.e., dx/dt) times the SIZE of
the differential time interval dt during which the change occurs. This is
written:

 dx(t) 
dx =  dt (Equation B).
 dt 

f.) As Equation B is an expression for the differential change of x

(i.e., dx) in terms of time, we can use it in Equation A to write:

 df (x) 
df (x) =  dx
 dx 
 df (x)   dx (t) 
= dt .
 dx   dt 

32
 Ch. 2--Derivatives

g.) Dividing both sides by dt (oops, there goes Fermat, rolling over
in his grave again), we get the Chain Rule, or:
df (x)  df (x)   dx(t) 
= .
dt  dx   dt 

h.) Example: Consider f(x) = x2 (derivative df(x)/dx = 2x). If x(t) =

kt3 (derivative dx(t)/dt = 3kt2), then df/dt is:

df (x)  df (x)   dx (t) 

=
dt  dx   dt 
= [2 x ] [3kt ] 2

= [2(kt )] [3kt ]
3 2

= 6k 2 t 5 .

i.) Bottom Line: It is not correct to treat the symbolic represen-

tation of a derivative as though it were made up of algebraic symbols.
Nevertheless, doing so has allowed us to derive a Chain Rule that
looks and acts just like the one derived by mathematicians.

D.) Partial Derivatives and the Del Operator:

Note: The del operator is a mathematical entity that most students

first encounter at the university level. It is,
nevertheless, a tool we will use later in this
course. Of the material presented below, the
only operations you will be expected to know f(x,y)

will be the basic execution of a partial

derivative and the basic execution of a f(x1 ,y1)

gradient. Both will be discussed in class;

neither will show up until much later in the
course. The rest of the material has been
plane tangent to curve
included for the sake of completeness and for
your own personal amusement. y
1
y
1.) Consider a function f(x,y) similar x1
Curve A
to the one graphed in Figure 2.5 to the right.
Note that a plane parallel to the y-z plane x
has been placed at x1, and that it cuts the

FIGURE 2.5

33
surface of f(x,y) along a curve as shown (the sketch labels this curve "Curve
A").

a.) The question? How do we derive a function that tells us the rate of
change of f(x,y) while holding the x parameter constant? Put another way, how
do we mathematically define the slope of the tangent to Curve A in the sketch?

b.) In Figure 2.5, a tangent has been drawn on the curve at

an arbitrary point f(x1, y1). The slope of that tangent is:

f (x 1 , y 1 + ∆y) − f (x 1 , y 1 )
lim ∆y→0 .
∆y

c.) This somewhat unusual operation (i.e., determining the slope of

f(x,y) holding the x variable constant) is called a partial derivative. The
symbol and formal definition for a partial derivative is shown below:

∂f (x, y) f (x 1 , y 1 + ∆y) − f (x 1 , y 1 )
= lim ∆y→0 .
∂y ∆y

d.) The partial derivative quoted above reads: the change of f(x,y)
with respect to y holding x constant. Put another way, it is the
derivative of f(x,y) with respect to y treating all other variables as
constants.

e.) Example: ∂(x 2 y 4 ) / ∂y = (x 2 )(4y 3 ).

2.) Partial derivatives are interesting in and of themselves, but they

become really useful when used in the context of what is called the del operator.
a.) The del operator is defined as:

 ∂() ∂() ∂() 

∇= i+ j+ k .
 ∂x ∂y ∂z 

Note 1: Bold-face letters denote vectors. The bold-face letters i, j,

and k denote unit vectors in the x, y, and z directions respectively. As such,
the del operator is a vector operator.

Note 2: An operator has no significance by itself. That is, it must

operate on a function, be it scalar or vector, to derive any meaning.

34
 Ch. 2--Derivatives

Note 3: The "()'s" found in the del operator expression normally house
the function being operated upon. Examples are coming.

3.) The del operator and the GRADIENT of a scalar:

a.) A scalar field is a function that defines a magnitude at every

point in a given volume. Temperature is a scalar field as there exists a
value for it everywhere.

b.) Consider the temperature scalar field T = 3ky2 + 20 (this could model
a room in which the temperature increases as one moves upward away from
the floor). Assume that k = 1 o/m2 (it is included for the sake of units).

c.) Note that if we travel in the +x direction, the temperature stays the
same. Only when we move upward or downward does it change. In fact, the
MAXIMUM POSITIVE CHANGE of T is along the "+j" direction.

d.) Using the del operator in conjunction with T, we get:

 ∂(3ky 2 + 20) ∂(3ky 2 + 20) ∂(3ky 2 + 20) 

∇T =  i+ j+ k
 ∂x ∂y ∂z 
= (6ky) j.

e.) Two things to observe:

i.) The DIRECTION of ∇ T is +j. This is the same direction

as the MAXIMUM POSITIVE CHANGE OF T at any point.

ii.) Though it may not be immediately evident, the RATE OF

CHANGE OF THE FUNCTION IN THE DIRECTION OF THE
MAXIMUM POSITIVE CHANGE at any point is equal to the
MAGNITUDE of ∇ T evaluated at that point.

f.) Bottom Line: When a del operator acts on a scalar function, it

yields a vector function. The direction of the vector is THE
DIRECTION IN WHICH THE SCALAR FUNCTION CHANGES THE
FASTEST (in a positive sense), and the magnitude of the vector
equals the RATE AT WHICH THE SCALAR FUNCTION CHANGES
IN THAT FASTEST-CHANGING DIRECTION.

g.) This operation is called the GRADIENT.

35
NOTE: THE REST OF THIS CHAPTER IS FOR YOUR OWN
EDIFICATION. READ IT BUT DO NOT STRESS OVER IT!

4.) The del operator and the DIVERGENCE of a vector:

a.) A vector field is a vector function that defines a magnitude and a

direction at every point in a given volume. As an example, the earth's
gravitational field is a vector field in the sense that at every point around
the earth: i.) there is a number that defines the magnitude of the
acceleration an object will experience if released at that point, and ii.)
there is a direction defined at every point for that acceleration.

b.) Consider a cylindrical cavity that is closed at both ends and

that holds air at high
pressure (see Figure 2.6a).
capped cylinder
Note: This example was
inspired by a problem in Hugh
cavity filled with high pressure air
Skilling's book Fundamentals of
Electric Waves.
cut-away view from side
i.) When the cavity
is opened, the com- FIGURE 2.6a
pressed air will expand uncapped
out and rush from the v = kx 2 i

cavity in the process.

ii.) Just after the

cavity is opened, the
air's velocity near the x=1 x=2 x=3 x

closed end will be velocity magnitude diverges from point to point

relatively small while
the air's velocity near FIGURE 2.6b
the open end will be
large. For the sake of
argument, assume the spatially varying velocity is defined as v =
kx2i, where k is a constant with the appropriate units and x is
defined from the closed end (see Figure 2.6b).

iii.) The air's velocity is clearly diverging from point to point.

36
 Ch. 2--Derivatives

iv.) If we dot the del operator into the velocity vector, we get:

 ∂() ∂() ∂() 

∇•v =  i+ j+ k  • (kx 2 i + 0 j + 0k)
 ∂x ∂y ∂z 
 ∂(kx 2 ) 
= 
 ∂x 
= 2kx.

v.) This scalar quantity tells us the rate at which the velocity
changes as we move along the line of the velocity vector. Put another
way, it gives us the divergence of the velocity--the rate of change of
velocity with position. This is the DIVERGENCE of the velocity vector.

vi.) In general, the DIVERGENCE yields the rate of change of a

vector function (in the direction of the vector) with position.

Note: Although the above example is that of a diverging velocity, the

divergence is most often associated with the differential form of Gauss's Law as it
pertains to electric fields generated by symmetric, extended charge configurations.
Most Calculus based physics texts (Halliday and Resnick's Fundamentals of
Physics, for instance) do not present this form of Gauss's Law (they only give
Gauss's Law in INTEGRAL form), but the DIFFERENTIAL form presented below
is useful as well as being the form of choice in most advanced electricity and
magnetism books. It has been included here because it is a practical example of
the use of the divergence operation, and because it allows the student a peek at a
relatively sophisticated mathematical tool physicists have at their disposal.

c.) In the world of static electricity, the evaluation of the diver-

gence of an electric field E at a point is found to be proportional to the
charge density r at the point. Mathematically, this is written:

ρ
∇•E = ,
εo
where the symbol ε o is a constant. Called the differential form of
Gauss's Law, this is one of Maxwell's equations.

Note: Again, when we get to Gauss's Law later in the year, we will
deal with it in its integral form. Even so, the differential version definitely
has its use.

37
 5.) The del operator and the CURL of
a vector:
y
Note: This operation is useful when
dealing with the differential form of
Maxwell's equations as they pertain to
magnetic fields. As before, YOU WILL NOT
BE TESTED ON THIS CONCEPT.
x

a.) Consider the force field F = (-

2)j
bx shown in Figure 2.7 to the right.

b.) Assume b = 1 newton/meter

(i.e., it is there solely for units a force field F = (-bx 2) j
conformity). Both the del operator and in the x-y plane
F are vectors, so let's do a cross product
between the two and see what we get: FIGURE 2.7

i j k i j k

∂() ∂() ∂() ∂() ∂() ∂()

∇xF = =
∂x ∂y ∂z ∂x ∂y ∂z

Fx Fy Fz 0 − bx 2 0

 ∂(0) ∂(− bx 2 )   ∂(0) ∂(0)   ∂(− bx 2 ) ∂(0) 

= i − + j − + k  ∂x − ∂y 
 ∂y ∂z   ∂z ∂x   
= (−2bx)k.

c.) Physical interpretation: Assume our field models the force of wa-
ter flowing in a stream. That is, assume there is something in the
middle of a stream which makes the flow negligible down the middle
while allowing the water to flow faster as one proceeds out away from the
central axis (we are obviously ignoring the fact that the stream's flow will
slow down at the stream's shoreline). With this in mind, consider:

38
 Ch. 2--Derivatives

i.) Three paddle-

wheels are placed in the looking down on the stream from above
field (see Figure 2.8)--
one at Point A, one at
Point B, and one at
Point C. Each is po-
sitioned so that its axis
Point B
is perpendicular to the
Point A Point C
stream's surface (i.e., in
the z direction).

ii.) Due to the

symmetry of the situa-
paddle-wheels at
tion, there is no net Points A, B, and C
force-of-rotation (i.e., no
torque) on the paddle-
wheel at Point B. That FIGURE 2.8
is, there is as much
force pushing on the
paddle to the left of the origin as there is on the paddle to the
right of the origin.
Note also that at this point, the CURL of the force field
∇ xF = (-2bx)k is zero as x = 0.

iii.) Due to the asymmetry of the situation, there is a net

torque acting on the paddle-wheel at Point A that makes it turn
counterclockwise.

Note: The sign convention presented in the last chapter for the di-
rection of a torque is as follows: A torque's UNIT VECTOR defines the direc-
tion of the axis about which the rotation occurs: the SIGN of the torque's unit
vector defines the sense of the rotation (i.e., a negative sign implies clockwise
rotation whereas a positive sign implies counterclockwise rotation).

iv.) Using our sign convention for torque, the stream-produced

torque at Point A will, by inspection, be in the +k direction (the
paddle-wheel will rotate counterclockwise).

39
 v.) The curl of F (i.e., ∇ xF = (-2bx)k) has a direction of +k for
all negative values of x (including that at Point A). In other words,
the curl of F at Point A has the same sign and axis direction as the
torque on the paddle-wheel at Point A.

vi.) A similar analysis for positive x (i.e., for the paddle-wheel

at Point C) yields a curl direction equal to -k. This is the same di-
rection as that of the paddle-wheel torque in that situation.

vii.) Lastly, although it is probably not as obvious, the

MAGNITUDE of the curl yields the rate at which a paddle wheel
will rotate at a given point in the field.

d.) In summary:

i.) The MAGNITUDE of the curl of a vector yields the rate at

which the field circulates at a given point in the field.

ii.) The DIRECTION of the curl of a vector at a particular

point tells you the axis about which the field circulates at that
point. It also denotes the sense of the rotation (i.e., clockwise or
counterclockwise).

Note: The word circulate is being used loosely here. Its significance
depends upon the system in question. In our example, it was associated
with the direction of the torque provided by the stream on a paddle-wheel.
We could as easily have been dealing with electrical phenomena. For
example, a magnetic field B that changes with time induces an electric field
E that circles. The rate at which the magnetic field changes ∂ B/ ∂ t is
related to the curl of E such that:

∂B
∇xE = − .
∂t

The moral of the story? How one interprets the word "circulation" is driven
by the context of the problem.

40
 Ch. 2--Derivatives

QUESTIONS
There are none! This chapter was not intended to be an exercise in
skills-learning (hence, no need for questions to test your understanding). It
was designed as a discussion aimed at giving you a feel for how Calculus
works and, more to the point, how physicists treat Calculus in the analysis
of real-world problems. If the material has made sense, great. If not, don't
worry about it. You will have plenty of time to become acquainted with
Calculus, physics style, as we go.

41
42