Professional Documents
Culture Documents
Introduction xi
iii
iv CONTENTS
This material is built upon the lecture I have given for Topics in Theoretical
Physics I and II (PHYS 3150 and 3160) at Middle Tennessee State University
for more than a decade.
ix
x PREFACE
Introduction
xi
xii INTRODUCTION
Part I
Mathematical Methods in
Physics I
1
Chapter 1
This lecture is based on the assumption that you have some basic knowledge
of sequence, series, and tests for convergence of series. Therefore, I will only
summarize the basics with more emphasis on examples.
Geometric series: Let’s consider two simple idealistic biological and physical
processes that can be described by geometric progressions.
i. bacteria in a culture doubles every hour
3
4 CHAPTER 1. SERIES AND CONVERGENCE
expressions as
Sn = 2 1 + 2 + 22 + 23 + :::2n (1.4)
2 3 n
2 2 2 2
Sn = 1 + + + + ::: (1.5)
3 3 3 3
P
Using the summation sign, ; these can be put in the form
n
X n
X
Sn = 2 2k = 2 rk ;
k=0 k=0
n
X k Xn
2
Sn = = rk ; (1.6)
3
k=0 k=0
where r = 2 for the doubling bacteria and r = 2=3 for the bouncing ball. In
general as n ! 1; the series
1
X
S = lim Sn = a rk (1.7)
n!1
k=0
and
1
X
rk+1 = r + r2 + r3 + r4 + ::: (1.9)
k=0
we can write
1
X 1
X 1
X
k k+1
r r = (1 r) rk
k=0 k=0 k=0
= 1 + r + r2 + r3 + r4 + ::: r + r2 + r3 + r4 + ::: = 1: (1.10)
Example 1.1 Two trains move toward one another with a speed vo . A bird
(an “ideal physics bird”) ‡ies back and forth between the two trains with
a speed vbird = vo , where > 1. Find the total distance traveled by the
bird before the trains crash if they start a distance Do apart.
Solution: Consider the …rst four consecutive back and forth movement of the
bird as shown in the …gure below.The bird starts from train one ( T1 ) and
moves towards train two ( T2 ). Let the time it takes to reach to T2 is t1
and during this time it has traveled a distance x1 as measured from the
initial position of T1 . Then we can express the distance x1
x1 = D0 v0 t1 : (1.13)
Since the bird is moving with a constant speed vo ; the distance the bird
traveled, x1 ; over the time, t1 ; is related to its speed by
x1 x1
vo = ) t1 = : (1.14)
t1 vo
Substituting Eq. (1.14) into Eq. (1.13), we …nd
x1
x1 = D 0 ) x1 = D0 : (1.15)
1+
From Fig. 1.3, we note that
D1 = D0 2 (D0 x1 ) = 2x1 D0 ) D1 = 2 D0 D0
1+
1
) D1 = D0 : (1.16)
+1
6 CHAPTER 1. SERIES AND CONVERGENCE
Figure 1.3: The motion of the bird and the two trains.
Following a similar procedure and referring to Fig. 1.3, the distance the
bird traveled when it ‡ies back to the …rst train, x2 ; can be expressed as
x2 = D1 v0 t 2 (1.17)
so that using
x2 x2
vo = ) t2 = (1.18)
t2 vo
we …nd
x2
x2 = D1 v0 t2 = D1 ) x2 = D1 : (1.19)
1+
Substituting Eq. (1.16) into Eq. (1.19), one …nds
1
x2 = D0 : (1.20)
1+ +1
x3 = D2 : (1.21)
1+
Using
2
1 1
D2 = D1 = D0 (1.22)
+1 +1
1.2. TESTING SERIES FOR CONVERGENCE 7
we …nd
2
1
x3 = D0 : (1.23)
1+ +1
Then the total distance the bird ‡ies before the two trains crash can be
expressed as an in…nite series,
1
X = x1 + x2 + x3 + ::: = D0 + D0
1+ 1+ +1
2
1
+ D0 :::
1+ +1
" #
0 1 2
1 1 1
)X= D0 + + + ::: ; (1.24)
1+ +1 +1 +1
where
1
a= D0 ; r = < 1: (1.26)
1+ +1
Eq. (1.25) is a geometric series and its sum can easily be determined using
the relation in Eq. (1.12)
1
!
X a 1 1
k
X=a r = = D0 1 = D0 : (1.27)
1 r 1+ 1 +1
2
k=0
The result in Eq. (1.27) is the total distance that the bird has traveled
before the two trains crashed.
(2) Absolute convergence test: Suppose the sum of the absolute values of all
the terms in the in…nite series in Eq. (1.28) converges, then the in…nite
series is called absolutely convergent series. An in…nite series that is ab-
solutely convergent is a convergent series. In other words, if
1
X
ja0 j + ja1 j + ja2 j + :::: = jan j (1.30)
n=0
(i.e. bn > 0). Compare the nth term in Eq. (1.31) (bn ) with the cor-
responding nth term in Eq. (1.28) (an ). If jan j bn for all n; then the
series in Eq. (1.28) is a convergent series.
(4) The integral test: In the series in Eq. (1.28), if 0 < an+1 an for n > N ,
then one can use the integral test,
Z 1
I= an dn: (1.32)
an+1
= lim : (1.33)
n!1 an
The convergence of the series will then be determined by the value of the
ratio, ; according to
8
< < 1; the series is convergent,
= > 1; the series is divergent, (1.34)
:
= 1 use a di¤erent test.
(b) Use the ratio test to check the convergence of this series.
Solution:
(a) We …rst rewrite the series using some general expression in terms of n
that generates all the terms in the series. In most cases we look for some
kind of relation that can be expressed using factorial and/or exponential
function. To this end, we note that
1 2 6 1 1 2 1 3 2 4 3 2
1+ + + + ::: = + + + + :::
6 120 5040 0! 3! 5! 7! 9!
0! 1! 2! 3! 4! n!
= + + + + :: :::; (1.36)
1! 3! 5! 7! 9! (2n + 1)!
X1
1 2 6 24
1+ + + + ; ::: = an ; (1.37)
6 120 5040 362; 880 n=0
where
n!
an = : (1.38)
(2n + 1)!
(b) We use the ratio test to determine if the series is convergent or divergent.
Using the result determined in (a), Eqs. (?? and (??, the ratio, ;
(n+1)! (n+1)(n!)
an+1 (2(n+1)+1)! (2n+3)(2n+2)(2n+1)!
= lim = lim n!
= lim n!
n!1 an n!1
(2n+1)!
n!1
(2n+1)!
n+1 n 1
= lim ) ' lim = lim =0
n!1 (2n + 3) (2n + 2) n!1 4n2 n!1 4n
which shows, < 1; and the series is a convergent series according to Eq.
(1.34).
i. Power Series: A series in which the nth term is a constant, bn ; times the a
function, (x a)n ,
1
X n 2
P (x) = bn (x a) = b0 + b1 (x a) + b2 (x a) + :::; (1.39)
n=0
where a is real constant and x is real variable (i.e. x 2 R). Note that a
power series can be convergent or divergent. It depends on the variable x.
ii. Taylor series: Taylor series is a power series in which the coe¢ cient in the
nth term of the series, bn ; is given by the nth derivative of the function
P (x) evaluated at x = a; and divided by n factorial,
1
X n 1 dn P (x)
P (x) = bn (x a) , where bn = : (1.40)
n=0
n! dxn x=a
iii. Maclaurin’s Series: Maclaurin’s Series is a Taylor series in which the con-
stant, a = 0:
1
X 1 dn P (x)
P (x) = bn xn , where bn = (1.41)
n=0
n! dxn x=0
Any function, f (x); that is di¤ erentiable for all values of x in the speci…ed
n
domain, can be expressed in Taylor or Maclaurin’s series. That means if d dx f (x)
n
1
X 1 dn f (x)
f (x) = bn (x a)n , where, bn = (1.42)
n=0
n! dxn x=a
1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 11
Example 1.3
(a) Show that the Maclaurin’s Series expansion for the function
is expressible as
X1
( 1)n+1 n
ln(1 + x) = x : (1.46)
n=1
n
Solution:
(a) Using Eq. (1.41)
1
X 1 dn P (x)
P (x) = bn xn , where bn = (1.47)
n=0
n! dxn x=0
we have
1
b0 =
ln(1 + x)jx=0 ) b0 = 0; (1.48)
0!
1 d 1
b1 = ln(1 + x) = ) b1 = 1; (1.49)
1! dx x=0 1 + x x=0
1 d2 1 d 1
b2 = ln(1 + x) =
2! dx2 x=0 2! dx 1+x x=0
1
1 1 ( 1)
= ) b2 = ; (1.50)
2! (1 + x)2 2
x=0
!
1 d3 1 d 1
b3 = ln(1 + x) = 2
3! dx3 x=0 3! dx (1 + x) x=0
2 2
1 ( 1) (1 2) ( 1)
= 3 ) b3 = ; (1.51)
3 2! (1 + x) 3
x=0
and
!
1 d4 1 d 1
b4 = ln(1 + x) = 2
4! dx4 x=0 4! dx (1 + x) x=0
3 3 2
1 ( 1) (1 2 3) 1 ( 1) 3! ( 1)
= 4 ) b4 = 4 = ;
4 3! (1 + x) 4 3! (1 + x) 4
x=0 x=0
(1.52)
(a) To …nd the interval of convergence we use the ratio test. The series con-
verges when
an+1 (x)
= lim < 1: (1.55)
n!1 an (x)
Using
( 1)n+2 n+1
an+1 (x) = x (1.56)
n+1
and
( 1)n+1 n
an (x) = x ; (1.57)
n
we …nd
( 1)n+2 n+1 n nx
= lim x = lim = jxj < 1
n!1 n+1 ( 1)n+1 xn n!1 n+1
) 1<x<1 (1.58)
and the ratio test can not tell us whether the series is a convergent or
divergent series. So one must use a di¤ erent test. A simple integral test
Z1
dn 1
= ln (x)j = 1 (1.60)
n
E = T + Eo = T + m0 c2 = m0 c2 : (1.64)
Show that the relativistic kinetic energy as deduced from these equations
reduces to the expected (classical) form in the non-relativistic limit— that
is, for v << c. That means we want to show that the kinetic energy of
the object is given by
1
T ' m0 v 2 (1.65)
2
for v << c which is what you know from Intro. Physics.
Solution: The relativistic kinetic energy, in terms of the kinetic energy and the
total energy can be expressed as
h 1=2
i
2
T = m0 c2 m0 c2 = m0 c2 1 1 (1.66)
2
for x = and p = 1=2; we may write
1 1
2 1=2 1 2 2( 2 1) 4
1 =1+ + + ::: . (1.68)
2 2!
In the classical limit, v << c ) = v=c << 1, we can drop all the terms
with n for n 4;
2 1=2 1 2 1 v2
1 '1+ =1+ . (1.69)
2 2 c2
Therefore, the relativistic kinetic energy in the classical limit becomes
1 v2 1
T ' m0 c2 1 + 1 )T ' m0 v 2 : (1.70)
2 c2 2
Example 1.5 Two identical small balls of charge Q and mass m are hung from
two silk threads of length L. The threads are attached at a common point
on the ceiling. The balls hang in equilibrium separated by a distance
1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 15
x. Show that, for small charges Q and large lengths L, the equilibrium
position is given by
1=3
Q2 L
x= : (1.71)
2 0 mg
Solution: If the balls are at equilibrium the net force acting on each ball must
be zero. There are three forces acting on each of the balls. These are
the repulsive electrostatic force ( Fe ), the tension force ( T ), and earth’s
gravitational force ( W ) [Fig. 1.4].The magnitude of the electrical ( Fe )
1 Q2
Fe = and W = mg;
4 0 x2
respectively. Since the balls are at equilibrium, the net force on each balls
must add up to zero. Considering the ball on the right side in Fig. 1.4,
for the net for in the y-direction, we have
and
1 Q2
F1x = Fe Ty = 0 ) T sin = (1.81)
4 0 x2
which leads to
sin Q2
= : (1.82)
cos 4 0 mgx2
Apply the Maclaurin’s series expansion
3 5 7
sin = + ::: (1.83)
3! 5! 7!
2 4 6
cos = 1 +
::: (1.84)
2! 4! 6!
for small angle (which is the case for large x << L which what we should
expect for small charge Q), we have
so that
1=3
Q2 x Q2 Q2 L
sin ' 2
) ' 2
)x' : (1.86)
4 0 mgx 2L 4 0 mgx 2 0 mg
18 CHAPTER 1. SERIES AND CONVERGENCE
(a) Connect the midpoints of the sides (of the equilateral triangle shaped)
chunk of land to form a total of 4 smaller triangularly shaped chunks of
land with equal sides (i.e. 4 equilateral triangles).
(b) Sell the middle triangularly shaped chunk of land.
(c) For each of the other 3 remaining unsold chunk of lands draw lines con-
necting the midpoints. You will form 4 small chuck of land from each.
Again you sell each middle triangular chuck of land for another investor.
Find the in…nite series for the total size (i.e. area) of land sold to the
investors if the process described above continued inde…nitely.
n2 + 5n3
an = p :
2n3 + 3 4 + n6
3. Use the ratio test to …nd whether the following series converge or diverge
(a)
X1 2
5n (n!)
:
n=0
(2n)!
(b) p
1
X (2n)!
:
n=0
n!
5. Prove that
n
X n (n + 1)
i=
i=1
2
1.
x
f (x) = p
1 x2
2.
1 x
g (x) =
1+x
Hint: Express the function as sum or product of two functions.
3. The little boy at the play ground: you wanted to introduce a little boy
to a swing set at the play ground. One sunny summer day you took him
to the play ground. You put him on the swing and secure his sit belt.
Then you pushed his sit in a horizontal direction with a force F so that
the rope of the swing is displaced by an angle and his sit is at a distance
x from the vertical (see the …gure below). At this instant the little boy
is at equilibrium position. Let the weight of the little boy be W and the
length of the rope is l:
(a)
ln (x a)
f (x) = lim p ;
x!1 x a
where a > 0.
(b)
g (x) = lim (x 1) ln [4 (x 1)] :
x!1
Complex Numbers,
functions, and series
z = a + ib; (2.1)
with a and b are real numbers and i is know as the imaginary, which is de…ned
as,
1=2
i2 = 1) i = ( 1) : (2.2)
For any complex number, z = a + ib; a is know as the real part and b is known
as the imaginary part,
Re z = a; Im z = b: (2.3)
In polar coordinate
a = r cos ; b = r sin ; (2.5)
21
22 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
where r is called the modulus (or magnitude) and is called the phase of the
complex number z: Thus one can express the complex number z = a + ib
The Cartesian and Polar coordinates, as shown in Fig. 2.1, are related by the
equations
x = r cos ; y = r sin ; (2.7)
The magnitude, r, (which is also represented as jzj) and the phase angle for
the complex number z are related to the Cartesian coordinates (x; y) by
p q
y
jzj = x2 + y 2 = r2 cos2 ( ) + r2 sin2 ( ) = r; = tan 1 : (2.10)
x
Note that the phase angle is always measured from the positive x-axis in a
counterclockwise direction.
2.1. COMPLEX NUMBERS 23
z =x iy (2.11)
Example 2.1 Find the real and imaginary part of the complex number
3 + 4i
z= (2.13)
3 4i
Solution:
To …nd the real and imaginary parts …rst we need to write it in the form,
z = x + iy. This can be done by making the denominator real. In order to
make it real we multiply it (and also the numerator) by complex conjugate,
Example 2.2 Find the real and imaginary part of the complex conjugate of
the complex quantity p
3 + 2i sin
z= ; (2.15)
4 + 2e4i
where is real.
Solution:
so that p
3 2i sin
z = (2.18)
4 + 2 cos (4 ) 2i sin (4 )
To …nd the real and imaginary part of z we need to make the denominator
in the above expression real. We can do this by multiplying it with its
24 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
Solution:
(a) since f (x) and g(x) are possibly complex functions, the complex conjugate
of Z(x) should be written as
3
Z (x) = 2f (x) + ig (x) (2.26)
2
where f (x) and g (x) are the complex conjugates of f (x) and g(x); re-
spectively.
2.1. COMPLEX NUMBERS 25
Then
3
Z (x) = 2f (x) + ig (x) (2.29)
2
becomes
3 p p
Z (x) = 2 3ix2 + i 2x 4 2i ) Z (x) = 6ix2 + 3xi + 6 2
2
(2.30)
Now substituting x = 1; we …nd
p
Z (x) = 6 2 3i: (2.31)
(5 + i) x 5yi = 4 + 3i (2.32)
Solution: To determine the values of the real quantities x and y; we …rst need
to write the left hand side in the form z = a + ib
Two complex number z1 and z2 are equal if and only if their corresponding
Re and Im parts are equal. Thus
where the amplitudes of the electric (E0 ) and the angular frequency (!)
are real. The wave vector (~k = k^
z ) given by
~k = 2 z^ = 2 n z^ = 2 n !
z^ = n z^; (2.35)
0 c c
in which 0 is the wave length, c is the speed of light in vacuum and
= c=n is the wavelength in the medium, with refractive index n; that
the wave is traveling in. The refractive index of the medium is possibly
26 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
complex for some media such as metals and should be expressed, generally,
as
n = nR + inI ; (2.36)
where both nR and nI are real. Note that Eq. (2.36) shows the wave ~k in
Eq. (2.35) is also possibly complex and should be written as
~ t) = ~k
B(z; ~ (z; t):
E (2.38)
~ t) = 1 E(z;
S(z; ~ t) ~ t):
B(z; (2.39)
0
Solution:
(a) Using
~ t) = E0 exp [i (kz
E(z; !t)] x
^; (2.41)
we may write the complex conjugate as
~ (z; t) = E0 exp [ i (k z
E !t)] x
^; (2.42)
so that
h ! i
~ (z; t)
E = E0 exp i (nR inI ) z !t x ^
c h i
! !
= E0 exp nI z exp i nR z !t x ^ (2.44)
c c
(b) In view of the result in part (a) and Eq. (2.37), the magnetic …eld becomes
~ t) = ~k E
B(z; ~ (z; t) = ! (nR + inI ) z^
! h c! i
E0 exp nI z exp i nR z !t x ^
c c h i
~ t) = ! (nR + inI ) E0 exp
) B(z;
!
nI z exp i
!
nR z !t z^ x ^
c c h c i
~ t) = ! (nR + inI ) E0 exp
) B(z;
!
nI z exp i
!
nR z !t y^
c c c
(2.45)
This result shows that the intensity decays exponentially as the wave pen-
etrates inside the medium.
zn = an + ibn ; (2.53)
and
1
X
bn = b0 + b1 + b2 + b3 :::bn + ::: = B; (2.55)
n=0
(a) You can use any convergence test that involves the absolute value sign,
which in this case will mean the magnitude of a complex quantity. This
is applied to the full complex in…nite series.
(b) Use any convergence test to separately test the convergence of the real and
imaginary parts of the complex series. The full series converges only if
both the real and imaginary parts of the series converge separately.
i 1 Zn+1
= lim = lim ) lim =0<1
n!1 (2n + 2) (2n + 1) n!1 (2n + 2) (2n + 1) n!1 Zn
(2.66)
which means the series is convergent.
30 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
Example 2.7 Find and sketch the circle of convergence in the complex plane
for the following complex in…nite series:
1
X 1
X n
n (z 2 + i)
cn z = (2.67)
n=0 n=0
2n
Then using
n
n (z 2 + i)
an (z z0 ) =
2n
n+1
n+1 (z 2 + i)
an+1 (z z0 ) = (2.70)
2n+1
we can write
,
n+1 n+1
(z 2 + i) (z 2 + i)
= lim <1
n!1 2n+1 2n+1
n+1
(z 2 + i) 2n z 2+i
) = lim n <1) <1
n!1 2n+1 (z 2 + i) 2
) jz 2 + ij < 2 ) jz (2 i)j < 2: (2.71)
This means the series is convergent inside a disk of radius R = 2 centered
at z0 = 2 i:
Example 2.8 Evaluate f (2 i) and …nd the real and imaginary parts if
z 1
f (z) = (2.72)
z
Solution: We can write f (2 i) as
2 i 1 1 i
f (2 i) = = (2.73)
2 i 2 i
which can be rewritten as
(1 i) (2 + i) 3 i
f (2 i) = = (2.74)
(2 i) (2 + i) 5
and the real and imaginary part
Re (f (2 i)) = 3=5; Im (f (2 i)) = 1=5 (2.75)
2.2. COMPLEX INFINITE SERIES AND THE CIRCLE OF CONVERGENCE31
Solution:
(a) Using the series expansion for the complex function
1
X zn z2 z3
exp (z) = =1+z+ + + :: (2.78)
0
n! 2! 3!
we can write
1
X n 2 3
(i )
exp (i ) = =1+i i + :: (2.79)
0
n! 2! 3!
which leads to
z = x + iy = r exp (i ) ; (2.89)
where p
1 y
r= x2 + y 2 ; = tan : (2.90)
x
This complex number to the power n (z n ) can then be written as
n n
zn = (x + iy) = [r exp (i )] = rn exp (in )
n
) z n = (x + iy) = rn (cos (n ) + i sin (n )) : (2.91)
Similarly, for the nth roots to this complex number (z 1=n ), one …nds
1=n
z 1=n = (x + iy) = r1=n exp (i k ) = r1=n (cos ( k ) + i sin ( k )) ; (2.92)
where
+2 k
k = ; for k = 0; 1; 2::n 1 (2.93)
n
Example 2.10 Solve the equation
z 5 = 32 (2.94)
where
r1=5 = 2;
+2 k 2 k 2
k = = ; for k = 0; 1; 2; 3; 4 ) 0 = 0; 1 = = 72 ;
n n 5
4 6 8
2 = = 144 ; 3 = = 216 ; 4 = = 288 (2.99)
5 5 5
Then the solution to the equation are found to be
Figure 2.2: The root on the complex plane. The circle has a radius r = 2
ln z = ln rei = ln r + ln ei = ln r + i ln e ) ln z = ln r + i (2.113)
ei( + )x
e i( + )x
= 2i sin [( + ) x] ; ei( )x
e i( )x
= 2i sin [( ) x] :
(2.119)
Using Eq. (2.119), we can write the integral as
Z
1
I= [sin [( + ) x] + sin [( ) x]] dx: (2.120)
2
36 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
which leads to
1 cos [( + ) x] cos [( ) x]
I= + (2.121)
2 +
Orthonormal Sets of Functions: The set of functions {f1 (x); f2 (x); :::fj (x); :::}
is said to be orthonormal if
Z x2
1 i=j
fi (x)fj (x)dx = ij = (2.124)
x1 0 i 6= j
Example 2.12 Evaluate the following integral, where and are positive,
nonzero integers: Z
I= sin ( x) sin ( x) dx (2.125)
and show that the sine function form an orthonormal set of functions.
Solution: Applying Euler’s formula, we may write
ei x
e i x
ei x
e i x
sin ( x) = ; sin ( x) = (2.126)
2i 2i
so that
ei x
e i x
ei x
e i x
sin ( x) sin ( x) =
2i 2i
ei( + )x
ei( )x
e i( )x
+e i( + )x
= (2.127)
:
4
This can be rewritten as
1 ei( )x
+e i( )x
ei( + )x
+e i( + )x
sin ( x) sin ( x) = :
2 2 2
(2.128)
or
1
sin ( x) sin ( x) = fcos [( ) x] cos [( + ) x]g (2.129)
2
2.4. ALGEBRAIC VS TRANSCENDENTAL FUNCTIONS 37
1 sin [( ) x] sin [( + ) x]
)I= : (2.131)
2 +
We recall sin is an odd function, which means
we have
sin [( + ) ] = 0 (2.135)
which leads to
sin [( ) ]
I= : (2.136)
This leads to
Z
=
I= sin ( x) sin ( x) dx = ; = (2.137)
0 6 =
cos [i ln 2] (2.139)
Solution: Evaluating this expression is much easier if we use the Euler’s for-
mula:
eiz + e iz
cos (z) = (2.140)
2
we can express cos [i ln 2] as
ei(i ln 2) + e i(i ln 2)
e ln 2
+ eln 2
cos [i ln 2] = = (2.141)
2 2
38 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
Solution: To …nd the principal value we …rst put the complex number in the
form rei : Recalling that
p 7
1 i= 2 exp i +2 n (2.150)
4
we have
p i i
i 7
(1 i) = 2 exp i
+2 n (2.151)
4
i 7
i
(1 i) = 21=2 exp i i +2 n
4
i i 7
(1 i) = 2 2 exp +2 n
4
The principal value is when n = 0:That means
i i 7
(1 i) = 2 2 exp (2.152)
4
2.4. ALGEBRAIC VS TRANSCENDENTAL FUNCTIONS 39
Example 2.16 Find the value for the following complex expression
2 !12 3
p
3+i
z = sin 1 4 p 5 (2.153)
3 i
Noting that
1 1 p p
tan p = ; 3 + i = 2 ) 3 + i = 2ei 6 (2.155)
3 6
1 1 11 p 11
tan p = ) 3 i = 2ei 6 (2.156)
3 6
we may write
2ei 6
12 h i12
= e i( 6 )
11
sin ( ) = 11
6 (2.157)
2ei 6
ei12( 6 )=e
11
20i 10i(2 )
sin ( ) = 6 =e =1
) = 2k ; k = 0; 1; 2::: (2.158)
2
40 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
(z1 z2 ) = z1 z2
(z1 z2 ) = z1 z2
Hint: it is easier to prove the statements about the quotient and the product
using the polar coordinates rei form; for the sum or di¤ erence it is easier
to use the rectangular form x + iy:
4.
(a) Find the absolute value of a complex number
3
z = (1 + 2i)
(b) Solve for all possible values of the real number x and y in the equation
2ix + 3 = y i
5. Physical application: A particle moves in the (x; y) plane so that its position
(x; y) as a function of time is given by
(a) Find the velocity (v) and acceleration (a) of the particle
dz d2 z
v= ;a = 2
dt dt
(b) Find the magnitude (absolute value) of the velocity and acceleration
(c) Describe the motion of the particle.
42 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
(b)
1
X n 2 n
1 i 1 i 1 i 1 i
=1+ + + ::: + :::
n=0
1+i 1+i 1+i 1+i
2. Find the disk of convergence for each of the following power series of complex
variable
(a)
X1 3 3 3
(n!) z n z (2!) z 2 (n!) z n
=1+ + ::: :::
n=0
(3n)! 3! 6! (3n)!
(b)
1
X n 2 3 n
(z i) z i (z i) (z i) (z i)
= + + + ::: :::
n=1
n 1 2 3 n
3.
(a) Show that for any real y, eiy = 1: Hence show that jez j = ex for every
complex z
jez j = ex :
(b) For the complex numbers z1 and z2 ; prove that
jz1 z2 j = jz1 j jz2 j
z1 jz1 j
=
z2 jz2 j
Hint: write the numbers in rei form
(c) Using the relations in part (a) and (b) …nd the value for
1 i
W =
1+i
1.
z = sin ( i ln 3)
2.
z = tan i
3.
z = cos (2i ln i)
4. p
1
z = tanh i 3
5. Physical Application: Consider the RLC-ac circuit shown in the …gure below.
The circuit has a resistor and an inductor connected in series and then a
capacitor in parallel with them (R and L in series„and then C in parallel
with them). From introductory physics, the impedance for a resistor (ZR ),
Z12 = Z1 + Z2
44 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES
Figure 3.1: A two-dimensional vector and its components on the x-y plane.
~ is
Magnitude of a vector : the length of the arrow representing a vector A
~ ~
called the length or the magnitude of A (written A or A) and is given by
q
A = A2x + A2y (2-Dimensions)
q
A = A2x + A2y + A2z (3-Dimensions) (3.1)
45
46 CHAPTER 3. VECTORS, LINES, AND PLANES
A unit vector : a unit vector is a vector whose magnitude is one. For any
~ we can …nd a unit vector denoted by A^ (a vector along the direction
vector A
~ with unit magnitude). This unit vector is given by
of A
~
A
A^ = (3.2)
~
A
A. The Scalar Product (The Dot Product): The scalar product of two vectors
~ = Ax^{ + Ay |^+ Az k^ and B
A ~ = Bx^{ + By |^+ Bz k^ (written as A
~ B)
~ is given
by
A~ B
~ = Ax Bx + Ay By + Az Bz : (3.3)
If the magnitude of the two vectors and the angle between the two, ; is
known, the scalar product of these two vectors can also be determined
using
~ B
A ~ = A ~ B~ cos : (3.4)
A dot product can be used to determine the angle between two vectors
0 1 0 1
A~ B~ A B + A B + A B
= cos 1 @ A = cos 1 @ x x y y z zA
(3.5)
~
A B ~ ~
A B ~
B. The Vector Product (The Cross Product): The vector product of two vectors
A~ and B ~ (written as A
~ B) ~ gives a third vector C
~ perpendicular to the
plane formed by the two vectors. If the angle between the two vectors is
; the magnitude of C~ is given by
~
C= A ~ = A
B ~ B
~ sin (3.7)
The Right-Hand Rule: To …nd the direction of C~ =A ~ B ~ put your right hand
~
on the …rst vector (A) curl your …ngers towards the second vector (B) ~
~
then your thumb points in the direction of vector C.
^{ ^ |^
|^ = k; k^ = ^{; k^ ^{ = |^: (3.8)
Cx = Ay B z Az B y (3.9)
Cy = Az Bx Ax Bz
Cz = Ax B y Ay B z
Solution:
48 CHAPTER 3. VECTORS, LINES, AND PLANES
F~ = (4; 0; 0) N = 4N x
^ (3.11)
We use the relation for the magnitude of the cross product of two vectors
given by
j~ j = j~rj F~ sin ( ) (3.13)
where is the angle between the position and the force vector. The mag-
nitude of the position and force vectors are obtained using
q
j~rj = rx2 + ry2 + rz2 = 2m (3.14)
and q
F~ = Fx2 + Fy2 + Fz2 = 4N: (3.15)
= (3.16)
2
the magnitude of the torque would be
To …nd the direction use the right hand rule and you will …nd the direction
to be along the negative-z direction. Then the torque may be expressed
as
~ = 8N m^ z (3.18)
F~ = (4; 0; 0) N = 4N x
^ (3.20)
~r = x (t) x
^ + y (t) y^ + z (t) z^ (3.33)
~r ~r0 = [x (t) 1] x
^ + [y (t) 2] y^ + [z (t) 3] z^ (3.34)
~v = 0^
x + (2m=s) y^ + (1m=s) z^: (3.35)
This means
) x (t) x
^ + y (t) y^ + z (t) z^ = 1^
x + 2^
y + 3^
z + (2m=s) t^
y + (1m=s) t^
z (3.37)
or
~r (t) = ~r0 + ~v t (3.38)
where
~r0 = x
^ + 2^
y + 3^
z ; ~v = (2m=s) y^ + (1m=s) z^ (3.39)
Example 3 Consider the particle in the previous example. What is the dis-
tance of closest approach (the “impact parameter”) of the particle to the
origin?
Solution: Let the distance of the closest approach be described by a point
(x0; y0; z0)
~r0 = x0 x
^ + y 0 y^ + z 0 z^ (3.40)
The closest distance is determined by the magnitude of the vector normal
to the trajectory of the particle. That means we must have ~r0 ? ~r0 ~r0 for
the impact parameter. We recall that when two vectors are perpendicular
their dot product is zero
Then
(~r0 + ~v t0 ) ~v t0 = 0
0 02
(~r0 ~v ) t + (~v ~v ) t = 0 (3.44)
Noting that
2 2
~v ~v = (2m=s) + (1m=s) = 5m=s (3.45)
~r0 ~v = r0x vx + r0y vy + r0z vz = 2 2+3 1=7
Therefore
(~r0 ~v ) t0 + (~v ~v ) t02 = 0 (3.46)
becomes
7
7t0 + 5t02 = 0 ) t0 = 0; t0 = (3.47)
5
The acceptable value is t0 = 0. Thus the shortest distance from the origin
is given by the magnitude of the vector
p p
~r0 = ~r0 + ~v t0 = ~r0 ) r0 = 12 + 22 + 32 = 14 (3.48)
Example 4 Find the shortest distance from the origin to the plane
3x 2y 6z = 7 (3.49)
Solution: Comparing
3x 2y 6z = 7 (3.50)
to the equation of a plane
ax + by + cz = d (3.51)
we have
~ = 3^
N x 2^
y 6^
z (3.52)
which is a vector normal to the plane. The unit vector along normal vector
is given by
~
^ = N = q 3^
N
x 2^ y 6^z ^ = 3x
)N ^
2
y^
6
z^ (3.53)
~
N 2 2 7 7 7
32 + ( 2) + ( 6)
We now pick any point on the plane. This point must satisfy the equation
3x 2y 6z = 7: (3.54)
52 CHAPTER 3. VECTORS, LINES, AND PLANES
~r = 3^
x + 1^
y (3.55)
Then the distance from the origin to the plane is the component of this
~ ; because the distance
vector along the vector perpendicular to the plane, N
is measured by the length of the normal to the plane. We can write this
distance as
d = ~r N^ = 3 3 2 1 ) d = 1units (3.56)
7 7
Example 5 Find the symmetric and parametric equations of the line through
(0; 2; 4) and (3; 2; 1).
~ = (3
A 0) x
^+( 2 2) y^ + (1 ~ = 3^
4) z^ ) A x + 0^
y 3^
z (3.59)
we have
a = 3; b = 0; c = 3 (3.60)
so that the symmetric equations becomes
x 0 z 4
= ;y = 2 (3.61)
3 3
The parametric equations are given by
8
< x = x0 + at
~ or
~r = ~r0 + At y = y0 + bt (3.62)
:
z = z0 + ct
which give 8
< x = 3t
~r = y= 2 (3.63)
:
z = 4 3t
Example 6 Find the equation of the plane containing the three points (0; 1; 1),
(2; 1; 3), and (4; 2; 1).
3.5. HOMEWORK ASSIGNMENT 6 53
Solutions: We …rst construct two vectors A ~ and B~ that lie on the plane formed
by these three points. Consider the vector pointing from (0; 1; 1) to (2; 1; 3)
and the vector from (0; 1; 1) to (4; 2; 1) and let denote these vectors by A ~
~
and B, respectively. The we can write
~ = (2
A 0) x
^ + (1 1) y^ + (3 1) z^ = 2^
x + 2^
z (3.64)
~ = (4
B 0) x
^ + (4 1) y^ + (1 1) z^ = 4^
x + 3^
y
The vector, N ~ , that is normal to the plane formed by these two vectors
(i.e. the three points) can be determined using the cross product of the
two vectors
x
^ y^ z^
~ =A
~ ~ = ~ =x 2 2 0 2 0 2
N B 0 2 2 )N ^ y^ + z^
3 0 4 0 4 3
4 3 0
(3.65)
~ =
N 6^
x + 8^
y 8^
z (3.66)
We recall that equation of a plane is given by
a (x x0 ) + b (y y0 ) + c (z z0 ) = 0 (3.67)
where
~ = a^
N x + b^
y + c^
z (3.68)
and
~r0 = x0 x
^ + y0 y^ z0 z^ (3.69)
is a point on the plane as measured from the origin. Then we can write
the equation of the plane formed by the three points as
6 (x x0 ) + 8 (y y0 ) 8 (z z0 ) = 0 (3.70)
where
~r0 = x0 x
^ + y0 y^ z0 z^ (3.71)
is given by ~r0 = (0; 1; 1), (2; 1; 3), or (4; 2; 1):
3. Show that the lines that join (0; 0; 0) to (1; 2; 1) and (1; 1; 1) to (2; 3; 4)
intersect and …nd the acute angle between them.
54 CHAPTER 3. VECTORS, LINES, AND PLANES
0 1
2 2
2 3 1
A = ; AT = @ 3 1 A (4.3)
2 1 0
1 0
0 1
2 4
2 1 3
B = @ 1 1 A ; BT =
4 1 1
3 1
55
56 CHAPTER 4. MATRIX AND DETERMINANTS
IA = AI = A (4.4)
2 2 3 2 1 2 4 2 4 6 2 8
2A = = (4.6)
2 2 1 2 0 2 5 2 4 2 0 10
Matrix Multiplication: two matrices can be multiplied if and only if the num-
ber of columns of the …rst matrix is equal to the number of rows of the second
matrix. If matrices have the same dimension, then they can be multiplied. From
the above matrices we can make the multiplications:
0 1
2 4
2 3 1 @ (ab)11 (ab)12
AB = 1 1 A= (4.8)
2 1 0 (ab)21 (ab)22
3 1
0 10 1 0 1
2 1 3 2 0 1 (cd)11 (cd)12 (cd)13
CD = @ 4 1 2 A@ 1 1 2 A = @ (cd)21 (cd)22 (cd)23 A
1 0 1 3 1 0 (cd)31 (cd)32 (cd)33
(4.9)
but we can not make the matrix multiplications BC or BD
4.3. MATRIX REPRESENTATION OF A SET OF LINEAR EQUATIONS57
CD 6= DC (4.11)
F (G + H) = F G + F H (4.14)
a1 x + b1 y + c1 z + d1 w = c1 ; (4.15)
a2 x + c2 z + d2 w = c2 ;
b3 y + c3 z = c4 ;
a4 x = c4 :
and 2 3
c1
6 c2 7
k=6 7
4 c3 5 : (4.20)
c4
One can augment the matrices M and k and write a matrix
2 3
a1 b1 c1 d1 c1
6 a2 0 c2 d2 c2 7
A=6 4 0 b3 c3 0
7 (4.21)
c3 5
a4 0 0 0 c4
(c) Replace any row by a linear combination of the rows of the matrix (as long
as the linear combination includes the row being replaced). In other words
you can add to or subtract from a multiple of one row to another,
4.4. MATRIX APPLICATION: SOLVING A SET OF LINEAR EQUATIONS59
4th Repeat step 2 until the augmented matrix is in Row Echelon Form and
entirely row reduced.
Row Echelon Form: a matrix is in Row Echelon Form if all of the following
conditions are satis…ed
All nonzero rows (rows with at least one nonzero element) are above any
rows of all zeroes.
Note: All zero rows, if any, belong at the bottom of the matrix.
The leading coe¢ cient (the …rst nonzero number from the left, also called
the pivot) of a nonzero row is always strictly to the right of the leading
coe¢ cient of the row above it
All entries in a column below a leading entry are zeroes (implied by the
…rst two criteria).
Reduced row echelon form (row canonical form): A matrix that is in row
echelon form and also its every leading coe¢ cient is 1.
is in Row Echelon Form but not in reduced row form. The following
matrix is not in Row Echelon Form
2 3
1 2 3 4
C=4 0 3 7 2 5 (4.24)
0 2 0 0
as the leading coe¢ cient of row 3 is not strictly to the right of the leading
coe¢ cient of row 2.
(a) If (rank M )<(rank A), the equations are inconsistent and there is no solu-
tion.
(c) If (rank M )=(rank A)= R < n, then R unknowns can be found in terms of
the remaining n R unknowns.
Example 2 Consider the circuit shown. The circuit element values are R =
1:0W and V = 2:0V .
(a) Applying Kirchho¤’s voltage and current rules …nd set of linear equations.
(b) Find the coe¢ cient and augmented matrices for the set of linear equations.
(c) Find the values of the three branch currents I1 ; I2 ; and I3 using Gaussian
elimination method.
(d) Find the rank of the coe¢ cient and augmented matrices. Is the ranks
consistent with what we discussed earlier?
Kirchho¤ ’s Voltage Law : Start at one point in a circuit and go around any
closed loop in the circuit. The voltage at the starting point of the loop must be
the same as the voltage at the ending point, since they are the same point in
the circuit. Another way of stating this is simply that the net change in voltage
as you go around any closed loop in the circuit must be zero (so the starting
potential must be the same as the ending potential). This is just a restatement
of the conservation of energy.
Kirchho¤ ’s Current Law :The net current ‡owing into any node of a circuit
must equal the net current ‡owing out of that same node. This law says that
there is no place from which charge magically appears in the circuit, and no
place that the charge disappears. There is a net conservation of charge in an
electric circuit.
Solution:
(a) Applying Kirchho¤’s voltage law for the closed loop abef a we can write
V
V +I1 R+2I2 R 2 V +I1 R = 0 ) 2I1 R+2I2 R = V ) I1 +I2 =
2R
(4.25)
and for closed loop bcdeb
I3 R 2 V + I3 R + 2 V 2I2 R = 0 ) 2I3 R 2I2 R = 0 I2 + I3 = 0:
(4.26)
4.4. MATRIX APPLICATION: SOLVING A SET OF LINEAR EQUATIONS61
I1 = I2 + I3 ) I1 I2 I3 = 0 (4.27)
I1 + I2 = 1 (4.28)
I2 + I3 = 0
I1 I2 I3 = 0
or
I1 + I2 + 0I3 = 1 (4.29)
0I1 I2 + I3 = 0
I1 I2 I3 = 0
M I = K; (4.30)
and 2 3 2 3
I1 1
I = 4 I2 5 ; K = 4 0 5 : (4.32)
I3 0
The augmented matrix is a combination of the coe¢ cient and the constant
matrix given by 2 3
1 1 0 1
A=4 0 1 1 0 5 (4.33)
1 1 1 0
(c) We apply elementary row operations to row reduce the augmented matrix.
2 3
1 1 0 1
A=4 0 1 1 0 5 (4.34)
1 1 1 0
where Mij is called the minor of aij : The minor of aij , Mij ; is the determinant
of the matrix formed by removing the row and the column containing aij : Since
we are given a 2-dimensional square matrix i = 1 or 2. If we chose i = 1, we
have
2
X
jAj = ( 1)1+j a1j M1j = ( 1)1+1 a11 M11 + ( 1)1+2 a12 M12 (4.47)
j=1
Note: The result in Eq. ( 4.56) shows that the determinant of the matrix C is
~ and B,
the cross product of vector A ~
x^ y^ z^
~
A ~ =
B Ax Ay Az : (4.57)
Bx By Bz
We now consider a square, n n; matrix A (a matrix with n rows and n
columns) 2 3
a11 a12 a1j a1n 1 a1n
6 a21 a22 a2j a2n 1 a2n 7
6 7
6 7
6 7
6 7
6 7
6 7
6 7
A=6 6 a i1 ai2 aij ain 1 ain
7
7 (4.58)
6 7
6 7
6 7
6 7
6 7
6 7
4 an 11 an 12 an 1j an 1n 1 an 1n 5
an1 an2 anj ann 1 ann
Then the determinant of this matrix (represented as jAj or det A ) is given by
n
X n
X
jAj = ( 1)i+j aij Mij = ( 1)i+j aij Mij (4.59)
j=1 i=1
where Mij is called the minor of aij : The minor of aij , Mij ; is the determinant
of the matrix formed by removing the row and the column containing aij :
Useful properties of determinants: It is useful to know the following proper-
ties of determinants of a square matrix which can be proved using Eq. (4.59).
1. Multiplying by a constant: Suppose a row or a column of a square matrix A
is multiplied by a constant k, then the determinant of the new matrix is
k times the determinant of the matrix A.
Example 3 For the 2-dimensiona square matrix
a11 a12
A= ; (4.60)
a21 a22
we may write a new matrix, B given by
ka11 a12 ka11 ka12
B= or B = ; (4.61)
ka21 a22 a21 a22
4.5. DETERMINANT OF A SQUARE MATRIX 65
then
det B = k det A; (4.62)
2. Zero determinant value: The value of the determinant of square matrix is
zero if
(a) all elements of a row (or a column) are zero
Example 4
0 a12 a11 a12
A= or A = ) det A = 0; (4.63)
0 a22 0 0
2 2 0 2
jAj = 2 2 (4.70)
1 1 1 1
D = jM j ; Dj = jMj j (4.72)
Dj
xj = (4.73)
D
provided D 6= 0
Example 9 Solve the system of equations from Example 2 using Cramer’s rule.
( D = 12; D1 = 8; D2 = 4; and D3 = 4.)
1 1 0 1
jM1 j = ) jM1 j = 2; (4.80)
1 1 0 1
0 1 0 1
jM2 j = ) jM2 j = 1; (4.81)
0 1 1 1
and
1 0 0 0 0 1
jM3 j = + ) jM3 j = 1: (4.82)
1 0 1 0 1 1
Therefore, the three currents are given by
jM1 j 2 jM2 j 1 jM2 j 1
I1 = = ; I2 = = ; I2 = = : (4.83)
jM j 3 jM j 3 jM j 3
where cof (A) is the cofactor of the matrix A. We recall that the minor of matrix
A (Mij ) is the determinant of the matrix formed from matrix A by removing
the ith row and j th column. For the cofactor matrix the elements are expressed
as
i+j
[cof (A)]ij = ( 1) Mij : (4.85)
1
Inverse of a (Square) Matrix : A
1 1
A A = AA =I (4.86)
We can determine the inverse of an invertible matrix (det jAj 6= 0) using row
reduction or the adjoint matrix.
we start from 2 3
a11 a12 a13 1 0 0
4 a21 a22 a23 0 1 0 5 (4.88)
a31 a32 a33 0 0 1
and do elementary row operation until we end up with
2 3
1 0 0 b11 b12 b13
4 0 1 0 b21 b22 b23 5 (4.89)
0 0 1 b31 b32 b33
b. Using the adjoint matrix: Using the adjoint matrix the inverse can be ex-
pressed as
T
[cof (A)]
A 1=
det jAj
Example 10 Find the inverse of the matrix
0 1
1 2 3
A=@ 2 0 4 A (4.91)
1 1 1
using
In order to get this matrix let’s make the elementary row operation.
Add row 1 to row 3:
2 3
1 2 3 1 0 0
4 2 0 4 0 1 0 5 (4.95)
0 1 4 1 0 1
we can write
0 4 2
M11 = det = 4 ) c11 = ( 1) M11 = 4 (4.110)
1 1
2 4 3
M12 = det = 6 ) c12 = ( 1) M12 = 6 (4.111)
1 1
2 0 4
M13 = det = 2 ) c13 = ( 1) M11 = 2 (4.112)
1 1
2 3 3
M21 = det = 5 ) c21 = ( 1) M21 = 5 (4.113)
1 1
1 3 4
M22 = det = 4 ) c22 = ( 1) M22 = 4 (4.114)
1 1
1 2 5
M23 = det = 1 ) c23 = ( 1) M23 = 1 (4.115)
1 1
2 3 4
M31 = det = 8 ) c31 = ( 1) M31 = 8; (4.116)
0 4
1 3 5
M32 = det = 2 ) c32 = ( 1) M32 = 2; (4.117)
2 4
1 2 6
M33 = det = 4 ) c33 = ( 1) M33 = 4: (4.118)
2 0
There follows that
2 3
4 6 2
cof (A) = 4 5 4 1 5 (4.119)
8 2 4
and 2 3
4 5 8
T
[cof (A)] = 4 6 4 2 5: (4.120)
2 1 4
The determinant of A is given by
1 2 3
det jAj = 2 0 4
1 1 1
0 4 2 4 2 0
= 2 +3 (4.121)
1 1 1 1 1 1
) det jAj = 14
becomes
2 3 0 1
4 5 8 2=7 5=14 4=7
1 4
A 1
= 6 4 2 5 = @ 3=7 2=7 1=7 A
14
2 1 4 1=7 1=14 2=7
(4.123)
~ (I used Mathematica)
The inverse of the matrix R
Therefore, since
1
R = RT (4.129)
and
F (x + y) = F (x) + F (y) (4.131)
If any one or these conditions is not met, then the operator/function is not
linear !
Example 11 Consider the operator that takes its argument (scalar or vector),
squares it, and adds 3:
O(A) = A2 + 3 (4.132)
Solution: We need to check the two conditions stated above. First condition:
2
O(aA) = (aA) + 3; (4.133)
2 2
aO(A) = a A + 3 = aA + 3a ) O(aA) 6= aO(A)
You can concluded the function is not linear because if one of the condi-
tions is violated then the function is not linear. But let’s check also the
second condition,
2
O(A+B) = (A + B) +3 ) O(A+B) = A2 +AB +BA+B 2 +3 (4.134)
which shows
O(A + B) 6= O(A) + O(B) (4.136)
a. If (rank M )<(rank A), the equations are inconsistent and there is no solution.
Solution: Applying Kirchho¤’s voltage law to three di¤erent loops we can write
I1 + I2 = 1; I2 + I3 = 0; I1 + I3 = 1; (4.140)
To …nd the Rank of the Augmented and the Coe¢ cient matrix we carry
out elementary row operation:
Add row two to row one: 0 1
1 0 1 1
@ 0 1 1 0 A (4.144)
1 0 1 1
76 CHAPTER 4. MATRIX AND DETERMINANTS
We note that the Augmented matrix can not be row reduced any further.
The rank of the coe¢ cient matrix is 2 and also the rank of the augmented
matrix is 2. That means (rank M )=(rank A), but the number of unknown
variables is 3. This means only two unknowns can be found in terms of
the third unknown.
~ 1 + k2 A
k1 A ~ 2 + k3 A
~ 3 + ::: + kn A
~ n = 0; (4.148)
where ki is none zero real number, the vectors are said to be linearly dependent
otherwise the vectors are linearly independent. Let’s consider three dimensional,
three vectors
~1
A = a1 x
^ + a2 y^ + a3 z^; (4.149)
~1
B = b1 x
^ + b2 y^ + b3 z^;
~3
C = c1 x
^ + c2 y^ + c3 z^:
k1 a1 k1 a 2 k1 a2
det k2 b1 k2 b2 k2 b2 =0 (4.151)
k3 c1 k3 c2 k3 c2
N.B. This condition is applicable if and only if the dimension of the vector
is the same as the number of the vectors. In other words when the coe¢ cient
matrix forms a square matrix.
4.9. LINEAR INDEPENDENCE 77
Example 13 Consider the set of three mutually orthogonal Cartesian unit vec-
tors. Prove that these three unit vectors form a linearly independent set
of vectors.
Solution: For the three unit vectors we may write that
~1
A = x
^; ) a1 = 1; a2 = a3 = 0 (4.152)
~1
B = y^; b2 = 1; b1 = b3 = 0
~3
C = z^ ) c1 = c2 = 0; c3 = 1
so that
k1 a1 k1 a2 k1 a2 k1 0 0
k2 0
det k2 b1 k 2 b2 k 2 b2 = det 0 k2 0 = k1 = k1 k2 k3 :
0 k3
k3 c1 k3 c2 k 3 c2 0 0 k3
(4.153)
Therefore
k1 a1 k1 a2 k1 a2
det k 2 b1 k 2 b2 k 2 b2 = k1 k2 k3 ; (4.154)
k 3 c1 k 3 c2 k3 c2
can be zero if and only if k1 = 0; k2 = 0; or k3 = 0: This means that the
Cartesian unit vectors are linearly independent.
and the two dimensional vector space (a plane) formed by the two vectors
Solution: If we perform row reduction for the matrix formed by the four vec-
tors, 2 3
1 4 5
6 5 2 1 7
M =6 4 2
7 (4.157)
1 3 5
3 6 11
we …nd 2 3
9 0 7
6 0 9 13 7
M =6
0
4 0
7: (4.158)
0 0 5
0 0 0
All the row operations are reversible and we can write the vectors
~ = (1; 4; 5) ; B
A ~ = (5; 2; 1) ; (4.159)
~ = (2; 1; 3) ; D
C ~ = (3; 6; 11)
2.
~ along the direction of A
(a) Find the component of the second vector, B; ~
~A = B
B ~ A^ A:
^ (4.162)
3.
~ along the directions of A
(a) Find the components of the third vector, C; ~ and
~
B
C~A = C ~ A^ A;^ C~B = C ~ B^ B:^ (4.164)
~
C ~ A^ A^
C ~ B
C ^ B
^
C^ = : (4.165)
~
C ~ A^ A^
C ~ B
C ^ B
^
~
Solution: Following the …rst step we normalize vector A
~
A
A^ = = (0; 1; 0; 0) : (4.167)
~
A
~
B ~ A^ A^
B
^= (3; 4; 0; 0) (0; 4; 0; 0) (3; 0; 0; 0)
B = = = (1; 0; 0; 0)
~
B ~ A^ A^
B j(3; 4; 0; 0) (0; 4; 0; 0)j j(3; 0; 0; 0)j
(4.169)
Using
~ B
C ^ = (1; 2; 3; 4) (1; 0; 0; 0) = 1
~ A^ = (1; 2; 3; 4) (0; 1; 0; 0) = 2
C (4.170)
one …nds
~
C ~ A^ A^
C ~ B
C ^ B
^ = (1; 2; 3; 4) 2 (0; 1; 0; 0) (1; 0; 0; 0) = (0; 0; 3; 4) :
80 CHAPTER 4. MATRIX AND DETERMINANTS
The third orthonormal basis vector, following step 3(a) and (b), becomes
~
C ~ A^ A^
C ~ B
C ^ B
^
(0; 0; 3; 4) 3 4
C^ = = = 0; 0; ;
~
C ~ A^ A^
C ~ B
C ^ B
^ j(0; 0; 3; 4)j 5 5
(4.171)
Function Spaces and Linear Independence of Sets of Functions: If f1 (x) ; f2 (x) ; f3 (x) ;
:::fn (x) have derivatives of order n 1, and if the determinant
then the functions are linearly independent. The determinant W is called the
Wronskian of the functions.
Example 17 Is the set of functions {sin (!t), sin2 (!t)} a linearly independent
set? (! = constant)
Solution: We can write the Wronskian for the two functions as
f1 (x) f2 (x) sin (!t) sin2 (!t)
W = det = det
f10 (x) f20 (x) d
dt [sin (!t)] d
dt sin2 (!t)
x 2y 4 = 0; 2z + w = 5; w + y = 2; 3x 1 = 5z:
(b)
3.
(a) What is meant by the speci…c gravity (s:g) of a substance?
(b) What is Archimedes’Principle? Provide a clear statement of this principle.
(c) An object is composed of x grams of lead (sg = 11) and y grams of tin
(sg = 7). The object has a mass of 82grams in air, and an apparent mass
of 77grams when suspended in oil having a speci…c gravity of 0:5. Find
the values of x and y.
4. The half-life of a radioactive substance is the time it takes for half of it
to decay into some other substances called decay products. Suppose that
a radioactive sample consists of components A and B having half-lives
2hr and 3hr, respectively. Assume that the decay products are gases that
escape at once. At the end of 12hr the sample is found to have a mass of
56g (grams), and is found to have a mass of 12g at the end of 18hr. Find
the masses of A and B that were originally present in the sample.
5. Do all the problems in the sample test. No need of turning your solutions
in!
Z1
() dx
0
(b)
2x + 5y + z = 2
x + y + 2z = 1
x + 5z = 3
4. In problem (a) and (b) show that the given functions are linearly inde-
pendent functions.
(a)
f1 (x) = sin (x) ; f2 (x) = cos (x)
(b)
f1 (x) = x; f2 (x) = ex ; f3 (x) = xex
Introduction to di¤erential
calculus-I
Thus for z(x; y) the partial di¤erentiation with respect to x or y; when x and
y are independent variables, is determined by keeping y or x constant and are
given by
85
86 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
Generally for a function that depends on the independent set of variables fx1 ; x2 ; :::xn g ;
f (x1 ; x2 ; :::xn ) ; the partial di¤erentiation with respect to one of these variables,
xi is given by
(a) Sketch this function to show that it de…nes a surface in a 3-D space described
by the Cartesian coordinates (x; y; z), where the z coordinate depends on
the x and y coordinates.
(b) Evaluate the following partial di¤erentiations at the point (x; y) = (1; 0):
@z @z @ 2 z @ @z @ 2 z @ @z @2z @ @z
; ; = ; 2 = ; = (5.5)
@x @y @x@y @x @y @x @x @x @y@x @y @x
Solution:
(a) The 3-D plot of this function obtained using Mathematica is shown below
5.1. PARTIAL DIFFERENTIATION 87
@z @ @z
= x2 + y 2 = 2x ) = 2; (5.6a)
@x @x @x x=1;y=0
@z @ @z
= x2 + y 2 = 2y ) = 0; (5.6b)
@y @y @y x=1;y=0
@2z @ @ @ @2z
= x2 + y 2 = (2y) = 0 ) = 0;
(5.6c)
@x@y @x @y @x @x@y x=1;y=0
@2z @ @ @ @2z
= x2 + y 2 = (2x) = 2 ) = 2;(5.6d)
@x2 @x @x @x @x2 x=1;y=0
@2z @ @ @ @2z
= x2 + y 2 = (2x) = 0 ) = 0:
(5.6e)
@y@x @y @x @y @y@x x=1;y=0
@z
(5.7)
@r x
mean "the partial of z with respect to r, with x held constant" with z expressed
as a function of r and x.
and
z (x; r) = x2 + 2y 2 = x2 + 2 r2 x2 = 2r2 x2 :
Then one can easily show that
@z @
= 2r2 x2 = 4r: (5.10)
@r x @r x
88 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
Example 2b An ideal gas is de…ned to be a gas that obeys the ideal gas equa-
tion of state
P V = nRT (5.11)
and
3
U= nRT (5.12)
2
where p is the pressure, V is the volume, n mass of the ideal gas mea-
sured in moles, T is the temperature, U is the internal energy, and R
the universal gas constant. Note that P; V; T; and U are the thermody-
namic variables. Obviously, as we can see from these two equations are
not independent set of variables. Find expressions for the following partial
di¤erentiations:
@P @P @P @2P @2P
; ; ; ; and : (5.13)
@V T @T V @V U @V 2 T @T 2 V
3
P V = nRT; U = nRT (5.14)
2
we may write
nRT 3 2U nR 2U 2U
P = ; U = nRT; T = )P = = (5.15)
V 2 3nR V 3nR 3V
so that
@P @P (V; T ) @ nRT nRT
= = = ; (5.16a)
@V T @V T @V V V2
@P @P (V; T ) @ nRT nR
= = = ; (5.16b)
@T V @V V @T V V
@P @P (V; U ) @ 2U 2U
= = = ; (5.16c)
@V U @V U @V 3V 3V 2
@P @P (U; V ) @ 2U 2
= = = ; (5.16d)
@U V @U V @U 3V 3V
Q (x)
Q (x) = x: (5.19)
x
In the limit as the interval length, x; goes to zero, we note
dQ (x) Q (x)
= lim ; (5.20)
dx x!0 x
so that the the corresponding in…nitesimal charge dQ (x) becomes
dQ (x)
dQ = dx; (5.21)
dx
which is the di¤erential of the function Q (x) : Physically, this means the in…n-
itesimal charge on the wire in the interval dx depends on how the change Q
changes with respect to x; dQ(x)
dx :
Consider a mountain with hills and valleys. The elevation of the mountain
(z) depends on where you are located on the mountain (x; y) because of the hills
and valleys. If you hike on this mountain, what you will travel on the surface
of the mountain depends on what you traveled along the directions in x and y
and also the slope of the mountain along these directions. Thus an in…nitely
displacement on the mountain dz is given by the total di¤erential
@z (x; y) @z (x; y)
dz = dx + dy: (5.22)
@x @y
Example 3 Application of the total di¤ erential : The height and radius of a
right-circular cylinder are measured to be
V (R; H) = R2 H (5.24)
so that using these results along with R = 0:2 and H = 0:2; the
uncertainty in the volume becomes
V = 50 (5.30)
The volume is
V (R; H) = R2 H = 326:7 (5.31)
then we may write
V V = 330 50cm3 : (5.32)
We often need to apply the chain rule in one, two, or three dimensions in physics.
Function of two variables: For a function of two variable, z (x; y) ; we recall
that the total di¤erential is given by
@z @z
dz = dx + dy: (5.35)
@x @y
If both x and y depend on another variable, like time, t, dividing this equation
by dt, we …nd the Chain rule for a function of two variables
dz @z dx @z dy
= + : (5.36)
dt @x dt @y dt
This can be generalized for function of multivariable.
z = exp x2 y2 (5.37)
where
x(t) = et and y(t) = e t : (5.38)
Find
dz
: (5.39)
dt
Solution: Since x and y are a function of the same variable, t; we can write
dz @z dx @z dy
= + (5.40)
dt @x dt @y dt
Noting that
@z
= ( 2x) exp x2 y2
@x
@z
= ( 2y) exp x2 y2
@y
dx dy
= et = x; = et = y (5.41)
dt dt
we …nd
dz
= ( 2x) exp x2 y 2 x + ( 2y) exp x2 y 2 ( y)
dt
dz
) = 2 y 2 x2 exp x2 y 2 (5.42)
dt
Example 5 Show that the trajectory of the projectile motion of a ball with
mass m thrown from the ground (x0 = 0; y0 = 0) at an angle (measured
from the horizontal (+x-axis)) with a speed v0 is de…ned by equation of
hyperbola given by
y (x) = bx ax2 ; (5.43)
92 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
where
g
a= ; b = v0 tan ( ) : (5.44)
2v0 cos2 ( )
Note that x is the distance of the ball measured on the ground and y(x)
is the height of the ball from the ground.
Solution: From Newtons second law
where
g
a= ; b = v0 tan ( ) : (5.55)
2v0 cos2 ( )
Example 6 Implicit Di¤ erentiation and application of the total di¤ erential :
Consider the equation
y 3 x2 y = 8 (5.56)
where y is a function of x. Evaluate dy=dx and d2 y=dx2 at the point
(x; y) = (3; 1), where y = y (x; y)
dy d2 y
and (5.57)
dx dx2
is to apply implicit di¤erentiation. To this end, let
z (x; y) = y 3 x2 y = 8 (5.58)
so that
dz @z @z dy dz dy
= + =0) = 2xy + 3y 2 x2 =0
dx @x @y dx dx dx
dy 2xy
) = 2 (5.59)
dx 3y x2
dy
= 1: (5.60)
dx
To …nd the second derivative of the function we use the result for the …rst
derivative. Noting that
dy 2xy
= 2 = f (x; y) (5.61)
dx 3y x2
we can write
d2 y @ @ dy
2
= f (x; y) + f (x; y) : (5.62)
dx @x @y dx
Upon carrying out the partial di¤erentiations, we have
@ 2y 4x2 y
[f (x; y)] = 2 + 2 (5.63)
@x 3y x2 (3y 2 x2 )
94 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
and
@ 2x 12xy 2
[f (x; y)] = 2 2; (5.64)
@y 3y x2 (3y 2 x2 )
so that Eq. (5.62) becomes
!
d2 y 2y 4x2 y 2x 12xy 2 dy
= + 2 + 2 : (5.65)
dx2 3y 2 x2 (3y 2 x2 ) 3y 2 x2 (3y 2 x2 ) dx
d2 y 1 8
= 1 1 1= : (5.69)
dx2 3 3
For a function of two variable, z (x; y) ; the extremum points with coordinates
(xi ; yi ; z (xi ; yi )) represent the coordinates for maxima on the hills or the minima
5.4. EXTREMUM (MAX/MIN) PROBLEMS 95
in the valleys on the surface de…ned by the function z (x; y) : The tangent plane
to these extremum points are horizontal planes (parallel to the x-y plane). For
this plane the slopes along the x and the y must be zero. Mathematically, these
slopes are de…ned by the partial derivative of the function z (x; y) evaluated at
the maxima or minima points with coordinates (xi ; yi ). Thus at the extremum
(maxima or minima) points
@f (x; y) @f (x; y)
= 0 and = 0: (5.72)
@x x=xi @y y=yi
These two equations de…nes the horizontal tangent plane at the extremum points
on the surface de…ned by z (x; y)
Example 7 An aquarium with rectangular sides and bottom has a …xed volume
V0 . Find its proportions so that it will require the least amount of glass
for construction.
Solution: As shown in Fig. 5.1, the length, width, and height of the aquarium
are l; w; and h, respectively. The the volume, which is …xed, is related to
l; w; and h by
V (l; w; h) = lwh = V0 : (5.73)
The aquarium has rectangular sides made of glass. The surface area,
A (l; w; h) of the …ve sides (excluding the top side which is open) is given
by
A (l; w; h) = lw + 2 (lh + wh) : (5.74)
We are interested in constructing an aquarium with least amount of glass.
It means the area of the glass sides that the aquarium is constructed from
must be minimized while the volume remains constant. As we can see
96 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
from Eq. (5.1) l; w; and h are not independent variables. We can reduce
the area from a function of three variable to a three variable by expressing
h in terms of l and w using Eq. (5.1) as
V0
h= : (5.75)
lw
Substituting Eq. (5.75) into Eq. (5.74) we can then express the area in
terms of two independent variables
1 1
A (l; w) = lw + 2V0 + : (5.76)
l w
For the minimum area we must have
@A (l; w)
=0 (5.77)
@l
and
@A (l; w)
=0 (5.78)
@w
so that using Eq. (??), one …nds
2V0
w =0 (5.79)
l2
and
2V0
l =0 (5.80)
w2
respectively. Now combining these two equations, we …nd
w4 w3 1=3
w 2V0 =0)w 1 = 0 ) w = 0; w = (2V0 ) (5.81)
4V02 2V0
But the acceptable solution is
1=3
w = (2V0 ) : (5.82)
Substituting combining this result and Eqs. (??) and (5.1) the correspond-
ing length and height of the aquarium are
2V0 2V0 1=3
l = = = (2V0 ) (5.83)
w2 (2V0 )
2=3
1=3
V0 V0
h = 1=3 1=3
= (5.84)
(2V0 ) (2V0 ) 4
Thus the proportions can be expressed as
l:w:h 1
1=3
=1:1:
(2V0 ) 8
These are the dimension of the Aquarium for least amount of glass for
construction and a volume V0 .
5.5. THE METHOD OF LAGRANGIAN MULTIPLIERS 97
That means the volume of the rectangular aquarium given can not be less or
greater than V0 : We were asked to determine the minimum total surface area,
A (l; w; h)
A (l; w; h) = lw + 2 (lh + wh) : (5.86)
This area appear to be a function three variables. But because of the condition
set for the volume, we have seen that it is actually a function of two variables.
Such kind of conditions are referred as constraints. The function that set the
volume to V0 in Eq. (5.85) is the constraint for the area de…ned by the function
in Eq. (??).
Suppose the function f (x; y; z; :::) describes some physical observable for
some system. In real physical problems the variables x; y; z; :::are subject to
constrains so that they are no longer independent. It is possible, at least in
principle, to use each constraint to eliminate one variable and to proceed with a
new and smaller set of independent variables. The use of Lagrangian multipliers
is an alternate technique that may be applied to determine extremum points for
the function when this elimination of variables is inconvenient or undesirable.
Suppose we are interested in …nding the extremum points for the function
f (x; y; z) under the constraint de…ned by the function (x; y; z) = k; where k is
a constant. We note that for the function to be extremum (including the saddle
point)
and for a function F (x; y; z; :::) = f (x; y; z) + (x; y; z); for a constant, at
the extremum points we may write
Suppose a string is stretched from the origin to a point (x; y) on the curve.
Find the minimum length of the string.
Solution: We need to …nd the minimum value for the distance
p
d (x; y) = x2 + y 2 : (5.91)
with the constraint
y (x) = 1 x2 ) (x; y) = y + x2 = 1: (5.92)
Following the procedure for the Method of Lagrangian multipliers, we have
5.5. THE METHOD OF LAGRANGIAN MULTIPLIERS 99
p
(1) Construct the function F (x; y) = f (x; y)+ (x; y) = x2 + y 2 + y + x2
(2) Compute the partial derivatives of F with respect to each of the variables
x; y and set them equal to zero,
@ x 1
F (x; y) = 0) p +2 x=0) p = 2 (5.93)
@x x2 + y2 x2 + y2
@ y y
F (x; y) = 0) p + =0) p = (5.94)
@y x2 + y 2 x2 + y 2
(x; y) = constant
1 1
) x = p ;y = (5.97)
2 2
Then the minimum distance is found to be
p
dmin = 3=2: (5.98)
@F (l; w; h)
= 0 ) w + 2h + wh = 0; (5.102)
@l
@F (l; w; h)
= 0 ) l + 2h + lh = 0; (5.103)
@w
and
@F (l; w; h)
= 0 ) 2w + 2l + lw = 0: (5.104)
@h
We solve the above three equations along with the constraint
lwh = V0 : (5.105)
w + 2h
= (5.106)
wh
and substitute it into Eqs. (5.103) and (5.104) we …nd
w + 2h w + 2h
l + 2h l=0)l 1 = 2h ) w = l (5.107)
w w
and
w + 2h w + 2h
2w + 2l lw = 0 ) l 2 = 2w ) l = 2h (5.108)
wh h
2 ~ 1 1 1
E (l; w; h) = + 2+ 2 :
8m l2 w h
Find the proportions to sides of the box that minimize the ground state
energy subject to the constraint that the volume is a constant
V (l; w; h) = lwh = V0
5.6. CHANGE OF VARIABLES 101
In the next example we will see how we make change of variables for the two
dimensional Laplace equation in Cartesian to Cylindrical coordinates.
Solution: There are two di¤erent way of transforming this equation. I will use
the short way. We note that
@V (r; ) @V (x; y) @x @V (x; y) @y
= + (5.134)
@r @x @r @y @r
so that using expressions for x and y in Eq. (5.133), one …nds
@V (r; ) @V (x; y) @V (x; y)
= cos ( ) + sin ( ) : (5.135)
@r @x @y
104 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
and
sin ( ) @V (r; ) @V (x; y) @V (x; y)
= sin2 ( ) + sin ( ) cos ( ) : (5.139)
r @ @x @y
and
cos ( ) @V (r; ) @V (x; y) @V (x; y)
= sin ( ) cos ( ) + cos2 ( ) : (5.142)
r @ @x @y
@ @ cos ( ) @
= sin ( ) + ; (5.144)
@y @r r @
@ @ sin ( ) @
= cos ( ) (5.145)
@x @r r @
5.6. CHANGE OF VARIABLES 105
@2 @ cos ( ) @ @ cos ( ) @
= sin ( ) + sin ( ) + (5.146)
@y 2 @r r @ @r r @
and
@2 @ sin ( ) @ @ sin ( ) @
= cos ( ) cos ( ) : (5.147)
@x2 @r r @ @r r @
@2 @2 @ 1 @ cos ( ) @ @
= sin2 ( ) + sin ( ) cos ( ) + sin ( )
@y 2 @r2 @r r @ r @ @r
cos ( ) @ @
+ cos ( ) (5.148)
r2 @ @
@2 @2 @ 1 @
= cos2 ( ) sin ( ) cos ( )
@x2 @r2 @r r @
sin ( ) @ @ sin ( ) @ @
cos ( ) + sin ( ) (5.151)
r @ @r r2 @ @
that leads to
@2 @2 @2 1 @ 1 @2
+ = + + (5.153)
@x2 @y 2 @r2 r @r r2 @ 2
which can be rewritten as
@2 @2 1 @ @ 1 @2
+ 2 = r + (5.154)
@x2 @y r @r @r r2 @ 2
Therefore the Laplace equation in polar coordinates can be expressed as
1 @ @V (r; ) 1 @ 2 V (r; )
r + =0 (5.155)
r @r @r r2 @ 2
where
df
p=
: (5.157)
dx
The Legendre transformation in Eq. (5.156) can also be expressed using Eq.
(5.157) as
df df
g =x f (x) (5.158)
dx dx
It is commonly used in thermodynamics and in the Hamiltonian formulation
of classical mechanics. Here we will consider an example in thermodynamics.
As mentioned previously, in thermodynamics we often have many variables to
choose from in order to specify the state of a given system, but only a small
number of these variables are actually independent variables. So-called ther-
modynamic potentials are functions that result from a transformation of the
internal energy function U from one set of variables to another.
Terminology and Notation:
dU = dQ + dW = T dS P dV (5.159)
@U @U
dU (S; V ) = dS + dV: (5.160)
@S V @V S
dU = T dS P dV (5.161)
we …nd
@U
= T; (5.162)
@S V
@U
= P: (5.163)
@V S
@U @U
F =S U: (5.165)
@S V @S V
If we substitute
@U
=T (5.166)
@S V
we …nd
F = F (T; V ) = ST U (5.167)
@z @z @z
; ; and :
@u @v @w
108 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I
4. If xy 3 yx3 = 6 is the equation of the curve, …nd the slope and the
equation of the tangent line at the point (1; 2) : Computer plot the curve
and the tangent line on the same axis.
5. In problem 4 …nd d2 y=dx2 at (1; 2).
Chapter 6
Introduction to Di¤erential
Calculus II
109
110 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
dy d2 y d3 y d4 y
a0 (x) y + a1 (x) + a2 (x) 2 + a3 (x) 3 + a4 (x) 4
dx dx dx dx
dn y
+:::an (x) n = f (x) (6.1)
dx
or, more concisely,
n
X di y
ai (x) = f (x) : (6.2)
i=0
dxi
Note that in ODE ai (x) are generally function of the independent variable x.
When ai (x) =Constant for all i, the ODE is known as linear ordinary di¤erential
equations (LODE).
In this section we learn how can determine the solutions to LODEs. By a
solution to a DE we mean a function which, when substituted back into the
DE, yields an identity. The general solution to an nth -order LDE contains n
independent, arbitrary constants of integration. A particular solution to a DE
is the general solution evaluated for particular values of its arbitrary constants.
These values for the constants are determined by applying boundary conditions
(BC’s) that pertain to the DE, ‘usually obtained from an analysis of the physical
system being modeled by the DE.
a0 dy (x) 1
y (x) + = f (x) : (6.4)
a1 dx a1
Noting that
a0
x
a0 a0
x a1 d h aa0 x i a0 a0 d h aa0 x i
e a1
e a1 x = 1 ) e a1
e 1 =1) =e a1 x
e 1 (6.5)
a0 dx a1 dx
Eq. (6.3) can be rewritten as
a0
x d h aa0 x i dy (x) 1
y (x) e a1
e 1 + = f (x) : (6.6)
dx dx a1
a0
Multiplying both sides of this equation by e a1 t ; we …nd
d h aa0 x i a0 dy (x) 1 a0
y (x) e 1 + e a1 x = f (x) e a1 t (6.7)
dx dx a1
6.1. LINEAR ORDINARY DIFFERENTIAL EQUATIONS 111
0
y (x 0
) e dx 0
= f (x0 ) e a1 x dx0 : (6.9)
x0 dx a1 x0
which leads to
Z x
a0 a0 1 a0
y (x) e a1 x y (x0 ) e a1 x0 = f (x) e a1 x dx (6.10)
a1 x0
Z x
a0 a0 1 a0 0
) y (x) e a1 x = y (x0 ) e a1 x0 + f (x0 ) e a1 x dx0
a1 x0
In this equation a0 are a1 are constants in the DE and x0 and y (x0 ) are also
constants de…ned by the boundary conditions of the function y (x) that pertain
to the DE. It usually obtained from an analysis of the physical system being
modeled by the DE. We also note that the de…nite integral can be expressed as
Z x Z
1 a0 0 1 a0
f (x0 ) e a1 x dx0 = f (x) e a1 x dx C (x0 ) ;
a1 x0 a1
In the next example we will see how we apply Eq. (6.13) to solve real physical
problems described by a …rst-order LODE.
(b) Evaluate the general solution at the appropriate BCs in order to …nd the
particular solution for the distance fallen by the object described above
as a function of time.
(c) Comment on limiting behavior of your answer, and physically describe the
system being modeled.
Solution:
(a) we recall that the acceleration, a, and the velocity, v are related by the
equation
dv
a= (6.14)
dt
and using the expression for the given acceleration, we …nd
dv
= g exp ( kt) : (6.15)
dt
In view of Eqs. (6.3) and (6.13), for a LODE of the form
dv (t)
a0 v (t) + a1 = f (t) (6.16)
dt
the solution can be expressed as
a0
t Z t
a0
t e a1 a0 0
v (t) = C1 e a1
+ f (t0 ) e a1 t dt0 ; (6.17)
a1 t0
where we set the initial time t0 = 0: Upon carrying out the integration,
we …nd
g
v (t) = C1 + (1 exp ( kt)) (6.19)
k
To …nd the position we note that the velocity is related to the displacement
by
dy dy g
v (t) = ) = C1 + (1 exp ( kt)) (6.20)
dt dt k
Noting that this equation can be put
dy (t)
b0 y (t) + b1 = f (t) (6.21)
dt
6.1. LINEAR ORDINARY DIFFERENTIAL EQUATIONS 113
where
g g
b0 = 0; b1 = 1; f (t) = C1 + exp ( kt)) (6.22)
k k
Applying Eq. (6.17) we can write the solution to Eq. (6.17) as
b0
t Z t
b0
t e b1 b0 0
y (t) = C2 e b1
+ f (t0 ) e b1 t dt0 ; (6.23)
b1 t0
C1 = 0; C2 = y0
(a) Using the di¤erential form of Faraday’s Law of induction …nd the current
induced in the circuit assuming that the resistance of the circuit is R.
(c) Using Newton’s second law …nd the speed of the bar as function of time.
(d) Show that the energy dissipated in the circuit (Joule Heat energy) is equal
to the initial kinetic energy of the bar.
Solution:
In the process the area bounded by part of the rails, the wire, and the
rod increases from an area A = xl at t = t; to an area, A = xl + ldx: The
increase in area, dA; will then be
(b) The bar experiences a magnetic force as a result of the induced current and
the uniform magnetic …eld. This magnitude of this force is given by
l2 vB 2
Fm = Iin lB = (6.43)
R
The direction is opposite to the direction of the velocity as one easily
determine using the right-hand-rule.
(c) Using Newtons second law
dv l2 vB 2 l2 B 2 dv
m = Fm = ) v+ = 0: (6.44)
dt R mR dt
that can be put in the form
dv (t)
a0 v (t) + a1 = f (t) (6.45)
dt
116 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
where
l2 B 2
a0 = ; a1 = 1; f (t) = 0
mR
We recall the solution to Eq. (6.45) is given by
a0
t Z t
a0
t e a1 a0 0
v (t) = C1 e a1
+ f (t0 ) e a1 t dt0 ; (6.46)
a1 t0
which gives
l2 B 2 t
a0
t
v (t) = C1 e a1
= C1 e mR (6.47)
Noting that initial velocity,v (t = 0) = v0 ; the velocity of the rod becomes
l2 B 2 t
v(t) = v0 e mR : (6.48)
(d) The energy delivered to the resistor per unit time (power) is given by
dW 2
P = = Iin (t) R (6.49)
dt
Using the results we found for the induced current in Eq. (6.42) along
with the velocity in Eq. (6.48), Eq. (6.42) can be put in the form
2l2 B 2 t
dW l2 v 2 (t) B 2 B 2 l2 v02
= R = e mR (6.50)
dt R2 R
or
dW
= f (t) : (6.51)
dt
Then the total energy delivered to the resistor is
Z1 2l2 B 2 t
B 2 l2 v02 B 2 l2 v02 mR
W = e mR dt ) W =
R R 2l2 B 2
0
1
) W = mvo2 (6.52)
2
Example 3 Discuss the variation in the intensity of radiation as it passes
through a stellar atmosphere assuming a plane-parallel model for the at-
mosphere.
Starting Point: Consider a di¤erential slab of stellar atmosphere of thickness
dz at the position z above the reference position z = 0. Let (z) be the
density of the atmosphere at that position, and let I(z) be the intensity
of radiation incident on the slab. Let Io be the intensity of radiation at
the reference position z = 0 within the atmosphere. Find an expression
for I(z) for z > 0.
6.2. THE FIRST-ORDER DE AND EXACT DIFFERENTIAL 117
dy (x)
a0 (x) y (x) + a1 (x) = f (x) (6.55)
dx
can be determined when the coe¢ cients are constant. Generally, a …rst-order
DE could be ODE where the coe¢ cients could be function of one variable (only
x) as described by Eq. (6.55). It could also be a DE in which the coe¢ cients
are a function of two variables (x and y) that has the form
dy (x)
Q (x; y) = P (x; y) : (6.56)
dx
In this section we will develop the method to solve these two type of …rst-order
DEs. But we …rst consider the later case. To this end, we note that Eq. (6.56)
can be rewritten as
@F (x; y) @F (x; y)
P (x; y) = ; Q (x; y) = : (6.58)
@x @y
This indicates if one can determine this constant function from the total di¤er-
ential obtained from the …rst-order DE in Eq. (6.56), one can solve the DE by
…nding y from the function. This depends on certain conditions that the total
deferential needs to satisfy that we will discuss. But …rst we will see how to
determine the total di¤erential, generally, for any function F (x; y) ; that is not
necessarily a constant function.
Solution: The total di¤erential, dF; for the function of two variables, F (x; y) ;
is given by
dF = P (x; y) dx + Q (x; y) dy: (6.60)
For the given function in Eq. (6.59), we have
@F @
P (x; y) = = 3x2 + 2x sin (2y) = 6x + 2 sin (2y) (6.61)
@x @x
and
@F @
Q (x; y) = = 3x2 + 2x sin (2y) = 4x cos (2y) (6.62)
@y @y
so that total di¤erential becomes
where
@F @F
P (x; y) = = 6x + 2 sin (2y) and Q (x; y) = = 4x cos (2y) : (6.65)
@x @y
Taking the partial derivative of the function P (x; y) with respect to y and the
function Q (x; y) with respect to x, we …nd
@P @F @
= = [6x + 2 sin (2y)] = 4 cos (2y) (6.66)
@y @y@x @y
@Q @F @
= = [4x cos (2y)] = 4 cos (2y) : (6.67)
@x @x@y @x
This shows that for the total di¤erential
we found
@P (x; y) @Q (x; y)
= : (6.69)
@y @x
Such kind of di¤erential is called an exact di¤ erential. If the di¤erential is an
exact di¤ erential, then there is a function F (x; y) such that
@F @F
P (x; y) = ; Q (x; y) =
@x @y
Example 5 For the …rst-order DE
dy
x3 + 2y = 3x (2 xy) ; (6.70)
dx
6.2. THE FIRST-ORDER DE AND EXACT DIFFERENTIAL 119
(a) Find the total di¤erential for a function of two variable F (x; y) =constant
in form
dF (x; y) = P (x; y) dx + Q (x; y) dy = 0 (6.71)
(b) Show that the total di¤erential is an exact di¤erential and …nd the function
F (x; y) =constant.
(c) Using the function F (x; y) =constant, …nd the solution to the given di¤er-
ential equation.
Solution:
(a) Multiplying the given …rst-order di¤erential equation by dx, we …nd
where
P (x; y) = 3x (2 xy) ; Q (x; y) = x3 + 2y : (6.74)
(b) Now taking the partial derivative of P with respect to y and Q with respect
to x, we …nd
@P @
= [3x (2 xy)] = 3x2 (6.75)
@y @y
and
@Q @
= x3 + 2y = 3x2 : (6.76)
@x @x
There follows that
@P @Q
= (6.77)
@y @x
Thus, the total di¤erential is an exact di¤ erential and there exists a func-
tion F (x; y) =constant, such that
@F @F
P = and Q = : (6.78)
@x @y
Now using
@F
P = = 3x (2 xy) (6.79)
@x
and integrating with respect to x
Z x Z x
@F (x0 ; y) 0
dx = 3x0 (2 x0 y) dx0 (6.80)
x0 @x0 x0
we …nd
F (x; y) F (x0 ; y) = 3x2 x3 y 3x20 + x30 y: (6.81)
120 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
we …nd
C (y) = y 2 + C1 ; (6.87)
where
C1 = y02 + C (y0 ) ;
is constant independent of x and y. Therefore, the solution to the DE is
found to be
F (x; y) = 3x2 x3 y + C (y) = 3x2 x3 y y 2 = C. (6.88)
or
dF (x; y) = P (x; y) dx + Q (x; y) dy = 0; (6.97)
where
P (x; y) = 2xe3y + ex ; Q (x; y) = y2 3x2 e3y (6.98)
or
F (x; y) = x2 e3y + xex + C1 (y) :
where we introduced the function de…ned by
@F (x; y) @F (x; y)
P (x; y) = ; Q (x; y) =
@x @y
2 2
@ F (x; y) @ F (x; y)
) = : (6.110)
@y@x @x@y
Sometimes, we may …nd a …rst order LDE with none constant coe¢ cients that
may lead to a total di¤erential that is not exact. Let’s consider the …rst-order
ODE with none constant function
dy (x)
a0 (x) y (x) + a1 (x) = f (x) : (6.111)
dx
6.3. FIRST-ORDER DE AND NONE EXACT TOTAL DIFFERENTIAL 123
where
a0 (x) f (x)
P (x; y) = y (x) ; Q (x; y) = 1: (6.115)
a1 (x) a1 (x)
Here we note that
@P (x; y) @ a0 (x) f (x) @Q (x; y)
= y (x) = y; = 0; (6.116)
@y @y a1 (x) a1 (x) @x
and
@P (x; y) @Q (x; y)
6= (6.117)
@y @x
that mean the resulting DE from the …rst-order ODE is not an exact di¤erential
and the method we developed in the previous section does not work. However,
this method can be used if we can …nd a function that can make the ODE in Eq.
(6.111) to lead to an exact di¤erential. Next we will determine this function
which is referred as "The integration factor".
The Integration Factor: The …rst-order ODE with none constant coe¢ cient
in Eq. (6.111) can put in the form
dy
+ P (x)y = Q(x); (6.118)
dx
where we introduced the functions de…ned by
a0 (x) f (x)
P (x) = ; Q(x) = : (6.119)
a1 (x) a1 (x)
This is the generic …rst-order ODE. Let’s say this ODE does not lead to an exact
di¤erential. Let the function V (x) be the function that when it multiplied to
the ODE in Eq. (6.118) leads to a total di¤erential that is an exact di¤erential.
Thus multiplying Eq. (6.118) by V (x) ; we have
dy
V (x) + V (x) P (x)y = V (x) Q(x)
dx
dy
) V (x) + V (x) P (x)y V (x) Q(x) = 0 (6.120)
dx
124 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
@ @
[V (x) P (x)y V (x) Q(x)] = V (x) (6.122)
@y @x
that leads
dV (x) dV (x)
V (x) P (x) = ) P (x)dx = : (6.123)
dx V (x)
Upon integrating this equation one …nds
R
P (x)dx
V (x) = e (6.124)
The function V (x) given by the integral expression Eq. (6.124) is the multiplying
factor to the ODE in Eq. (6.118) that leads to an exact di¤erential. It is referred
as the integration factor. Thus multiplying Eq. (6.118) by the integration factor,
we …nd R R R
dy
e P (x)dx + e P (x)dx P (x)y = e P (x)dx Q(x): (6.125)
dx
Applying the product rule for di¤erentiation, we put this equation in the form
d h R P (x)dx i R
ye = e P (x)dx Q(x); (6.126)
dx
so that integrating it with respect to x; one can write
R
Z h R i
ye P (x)dx = e P (x)dx Q(x) dx + C; (6.127)
This is the general solution to the …rst-order ODE in Eq. (6.118) and the
constant C is determined from the given boundary condition set to the speci…c
problem that the DE is representing. In the next examples, we will see its
application. But let’s examine Eq. (6.128) if it gives the general solution we
determined for a …rst-order ODE with constant coe¢ cients. We recall, for a
ODE with constant coe¢ cient
dy (x)
a0 y (x) + a1 = f (x) (6.129)
dt
6.3. FIRST-ORDER DE AND NONE EXACT TOTAL DIFFERENTIAL 125
dy
+ P (x)y = Q(x); (6.131)
dx
where
a0 f (x)
P (x) = :Q (x) = : (6.132)
a1 a1
Substituting this functions into Eq. (6.128), we …nd
R a0 Z R a0 R
a1 dx
f (x) a0
y(x) = e e a1 dx dx + Ce a1 dx
: (6.133)
a1
that leads to a0
x Z
a0
x e a1 a0
y(x) = Ce a1
+ e a1 x f (x) dx: (6.134)
a1
which the same as Eq. (6.130).
dy
+ P (x)y = Q(x); (6.136)
dx
(b) Using the general solution for such kind of …rst-order DE (Eq. (6.118)),
…nd the solution for y (x) subject to the boundary condition,
26
y (x0 = 1) = : (6.137)
3
Solution:
dy 1 dy
+ 2x = y) + P (x) y = Q (x) ; (6.138)
dx 2x dx
we have
1
P (x) = ; Q (x) = 2x: (6.139)
2x
126 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
so that
Z
p 1 p 4 2 p
y(x) = 2 x p xdx + C x ) y(x) = x + C x; (6.142)
x 3
where we used the relations
p p p 1
eln x
= x and eln(1= x)
=p : (6.143)
x
Using the given boundary condition
26
y (x = 1) = (6.144)
3
the constant of integration, C, is found to be
4 26
+C = ) C = 10: (6.145)
3 3
Therefore, the general solution in Eq. (6.142) becomes
4 3=2 p
y(x) = 10 x x:
3
This is the solution to the ODE in Eq. (6.135) under the given boundary
condition.
dy d2 y d3 y dn 1 y
a0 y + a1 + a2 2 + a3 3 ::::an 1 (6.147)
dx dx dx dxn 1
dn y
+an n = 0:
dx
We will introduce to how the general solution to such kind of HODE be deter-
mined by solving what is know as the indicial equation that can be derived from
the HODE.
Indicial Equation: The indicial equation is founded on an educational guess.
We guess the solution to Eq.(6.147) be
a0 + a1 k + a2 k 2 + a3 k 3 ::::an 1k
n 1
+ an k n Aekx = 0: (6.149)
a0 + a1 k + a2 k 2 + a3 k 3 ::::an 1k
n 1
+ an k n = 0: (6.150)
This equation is called the indicial equation. The solution to the indicial equa-
tion in Eq. (6.150) gives n roots that could be real or complex. If these roots
are none degenerate (i.e. no repeated identical roots) and are given by
k = k1 ; k2 :::kn ; (6.151)
the general solution to the HODE with constant coe¢ cients in Eq. (6.147) is
given by
y (x) = C1 ek1 x + C2 ek2 x + :::Cn ekn x (6.152)
As we have seen in the in the solutions to …rst-order ODE, the constants
C1 ; C2 ; :::Cn ; are determined from the boundary conditions pertinent to the
particular problem.
Q (t) dI (t)
+L =0 (6.153)
C dt
128 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
Combining Eqs. (6.160) and (6.161), one can easily show that
1 I0 1 I0
A= Q0 + ;B = Q0 : (6.162)
2 i! 2 i!
Therefore, the charge at a given time t can be expressed as
1 I0 1 I0
Q(t) = Q0 + exp (i!t) + Q0 exp ( i!t) ; (6.163)
2 i! 2 i!
or
exp (i!t) + exp ( i!t) I0 exp (i!t) exp ( i!t)
Q(t) = Q0 + :
2 ! 2i
(6.164)
Applying Euler’s formula Eq. (6.164) can be put in the form
Q(t) = A cos (!t) + B sin (!t) ; (6.165)
where
I0
C = Q0 ; D = : (6.166)
!
Introducing the constants de…ned by
I0
Q0 = E sin '; = E cos '; (6.167)
!
Eq. (6.165) can also be expressed as
Q (t) = E [sin ' cos (!t) + cos ' sin (!t)] = E sin (!t + ') ; (6.168)
with s
2
Q0 I0
tan (') = ;E = Q20 + : (6.169)
I0 =! !
If we de…ne another constant de…ned in a slightly di¤erent way
I0
Q0 = G cos '; = G sin '; (6.170)
!
one can rewrite Eq. (6.165) as
Q(t) = G cos ' cos (!t) + G sin ' sin (!t) = G cos (!t ') ; (6.171)
where s
2
I0 =! I0
tan (') = ;G = Q20 + : (6.172)
Q0 !
The current
dQ
I (t) = (6.173)
dt
can be determined using any one of the equations for the charge in Eqs.
(6.164), (6.165), (6.168), or (6.171). If we chose Eq. (6.168), we …nd
I (t) = I0 cos (!t) Q0 ! sin (!t) (6.174)
for the current in the circuit.
130 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
d2 y
! 2 y = 0; (6.175)
dx2
when ! a real constant. For the plus case,
d2 y
+ ! 2 y = 0; (6.176)
dx2
as we saw in the example above, the roots to the indicial equation
k 2 + ! 2 = 0; (6.177)
k2 ! 2 = 0; (6.180)
are real, k1 = ! and k2 = !: It can easily be shown that the solution can be
expressed as 8
>
> A exp (!x) + B exp ( !x) ;
<
C cosh (!x) + D sinh (!x) ;
y (x) = (6.181)
>
> E sinh (!x + ') ;
:
G cosh (!x + ') :
Problem 1: show the solutions in Eq. (6.181)
Problem 2:Find the current and the charge in the previous example if a re-
sistor R is included in series with C and L:
~2 d2 (x)
+ U (x) (x) = E (x); (6.182)
2m dx2
6.4. HIGHER-ORDER ODE’S 131
Solution: In region I, the potential energy is zero, U (x) = 0; and the Schröedinger
Equation becomes
d2 (x) d2 (x)
= k 2 (x) = + k 2 (x) = 0 (6.184)
dx2 dx2
where
2mE
k2 = : (6.185)
~2
Substituting
x
(x) = Ae (6.186)
132 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
In region II, the potential energy is zero, U (x) = U0 ; and the Schröedinger
Equation becomes
~2 d2 (x)
= (U0 E) (x)
2m dx2
d2 (x) 2m (U0 E)
) = (x) (6.189)
dx2 ~2
which we may write as
d2 (x) d2 (x)
2
= q 2 (x) = q 2 (x) = 0 (6.190)
dx dx2
where
2m (U0 E)
q2 = : (6.191)
~2
Substituting
x
(x) = Ae (6.192)
we …nd the indicial equation
2
q2 = 0 ) = q (6.193)
(x) = A1 eqx + A2 e qx
: (6.194)
k2 2ak + a2 = 0; (6.197)
are degenerate
k1 = k2 = a: (6.198)
When the roots are degenerate the general solution in Eq. (??) can be used
with some changes that depend on the degeneracy in the roots of the indicial
equation. For example, for the HODE in Eq. (6.196), the values for the roots
are identical (k1 = k2 = a), where the degeneracy, m = 2, the general solution
should be written as
y(x) = (A + Bx) eax : (6.199)
Note that for m = 2, the coe¢ cient to eax is a polynomial function, P1 (x) =
A + Bx; of degree l = m 1 = 1. Similarly, for a HODE that leads to an indicial
equation, with m = 3 degenerate roots, (k1 = k2 = k3 = a), the general solution
is given by
y(x) = A + Bx + Cx2 eax = P2 (x) eax : (6.200)
In this case the coe¢ cient polynomial function, P2 (x) ; has a degree, l = m 1 =
2:
Thus we can make the following changes to the general solution in Eq. (??)
in relation to the degenerate roots to the indicial equation:
(a) n-th order di¤erential equation leads to an indicial equation with degen-
erate roots, m = n; the general solution is given by
(b) n-th order di¤erential equation that leads to an indicial equation with
three di¤erent set of roots. Suppose the …rst set of roots has m = m1
degenerate roots (i.e. k1 = k2 = k3 = :::km1 = a), the second set has
m = m2 degenerate roots (i.e. k10 = k20 = k30 = :::km2 = a0 ), and the
third set consist of m3 none-degenerate roots (k100 = a001 ; k200 = a002 ; k300 =
a003 :::km
00
3
= a00m3 ) (Note that m1 + m2 + m3 = n). For this HODE equation,
the general solution is given by
a0 x a0 x 00 00
y(x) = Pm 1 1 (x)e + Pm 2 1 (x)e + C100 ea1 x + C200 ea2 x(6.202)
00
00
+:::Cm 3
eam3 x ;
where
Example 11 In each of the following parts, …nd the indicial equation, list the
roots of the indicial equation {ki }, and write out the general solution to
the DEs:
(a)
d3 y dy
3
+ = 0: (6.204)
dx dx
(b)
d4 s d3 s
2 3 = 0: (6.205)
dt4 dt3
Solution:
dy d3 y
y (x) = Aekx ) = ky (x) ) 3 = k 3 y (x) (6.206)
dx dx
into the given DE, the indicial equation becomes
k 3 + k = 0: (6.207)
k 3 + k = 0 ) k k 2 + 1 = 0 ) k1 = 0; k2 = i; k3 = i: (6.208)
The roots do not show any degeneracy in this case and the solution to the
DE can be written as
y (x) = A + Beix + Ce ix
: (6.209)
or
y (x) = A + D cos (x) + E sin (x) (6.210)
(b) Substituting
d3 s d4 s
s (t) = Aekt ; ) = k 3
s (t) ) = k 4 s (t) (6.211)
dt3 dt4
in the given DE, the roots to the indicial equation are found to be
3
2k 4 3k 3 = k 3 (2k 3) = 0 ) k1 = ; k2 = k3 = k4 = 0: (6.212)
2
Since the the indicial equation has three identical roots, the solution can
be written as
3
2k 4 3k 3 = 0 ) k 3 (2k 3) = 0 ) k1 = ; k2 = k3 = k4 = 0; (6.213)
2
6.4. HIGHER-ORDER ODE’S 135
This shows that the three roots (k2 ; k3 ; k4 ) are degenerate with degeneracy,
m = 3; and one none degenerate root (k1 ). Thus applying the relation in
Eq. (6.202), the general solution can written as
where a is the value for the degenerate roots and k1 is the value for the
none degenerate root. Using these values the general solution to the DE
becomes
3
y (x) = A + Bx + Cx2 + Ae 2 x : (6.215)
N.B. You may use Mathematica to check you results. Here is the result for
the previous two examples
Here C[1], C[2], C[3], and C[4] are constants of integration. note that -1 in
8
the …rst solution can be absorbed in C[2] and 27 in the second solution can also
be absorbed in C[1].
Solution: Since the mass initially was at equilibrium and also consider the dis-
placement from this equilibrium position, we do not consider the buoyant
and gravitational force on the mass. Therefore, the equation of motion for
the mass is given by Newton’s second law
d2 y
Fnet = may = m : (6.216)
dt2
The net force acting on the mass is given by
dy
Fnet = ky bvy = ky b (6.217)
dt
which leads to
d2 y dy d2 y dy
m = ky b ) 2 +2 + !2 y = 0 (6.218)
dt2 dt dt dt
where
k b
; !2 = = : (6.219)
m 2m
The indicial equation would then be
2
+2 + ! 2 = 0; (6.220)
1 = 2 = (6.224)
where p
= !2 2 (6.229)
which means the roots for the indicial equation becomes
1 = +i ; 2 = i (6.230)
or
t
y(t) = e [A0 cos ( t) + B 0 sin ( t)] (6.232)
or 00
t
y(t) = e A cos ( t ) : (6.233)
dy d2 y d3 y dn 1 y dn y
a0 y + a1 + a2 2 + a3 3 ::::an 1 n 1
+ an n = f (x) : (6.234)
dx dx dx dx dx
In this section, we will see the procedure that is built upon on the methods we
developed in the previous sections for HODE and an educated guess to part of
the general solution related to the none homogeneity of the DE. This procedure
consist of the following successive steps.
1-st Find the general solution to the HODE (i.e. with f (x) = 0). This solu-
tion is called the complementary function, or the homogeneous solution,
denoted by yh (x). Note that this solution contains n constants of integra-
tion for the nth -order ODE.
2-nd Find in any way you can a particular solution, yp (x); that satis…es the
full none homogeneous ODE. Note that this function contains no arbitrary
constants of integration.
3-rd The general solution to the none homogeneous ODE is then given by
4-th Any boundary conditions given to the problem must then be applied to
the general solution y(x).
138 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
d2 s ds
3 + 2s = 16: (6.236)
dt2 dt
Solution:
Step 1 : Find the homogeneous solution. That means determine the solution
for
d2 s ds
2
3 + 2s = 0: (6.237)
dt dt
Noting that the indicial equation is
k2 3k + 2 = 0 ) k1 = 2; k2 = 1 (6.238)
sp = 8: (6.242)
Step 4 : Apply the boundary conditions if there are any. Here we are not
given any.
Example 14 Find the general solution to the following di¤erential equation:
d2 R
+ 4R = sin (3x) (6.244)
dx2
Solution:
Step 1 : Find the homogeneous solution for
d2 R
+ 4R = 0: (6.245)
dx2
6.4. HIGHER-ORDER ODE’S 139
k 2 + 4 = 0 ) k1 = 2i; k2 = 2i (6.246)
Note that this guess will be valid because 3i is di¤ erent from the roots of
the indicial equation. See the next example.
Since in the particular solution we are not allowed to have arbitrary constant
of integration, we must determine C: Upon substituting Eq. (6.248) into
Eq. (6.244), this constant can be determined to be
9C sin (3x) + 4C sin (3x) = sin (3x) ) 5C sin (3x) = sin (3x)
1
) C= : (6.249)
5
Thus the particular solution becomes
1
Hp = sin (3x) : (6.250)
5
Step 3 : Write the general solution
1
R (x) = Rh (x) + Rp (x) ) R (x) = A cos (2x) + B sin (2x) sin (3x) :
5
(6.251)
Step 4 : Apply the boundary conditions if there are any. Here there are
no any.
Example 16 Find the general solution to the following di¤erential equation:
d2 y
+ 4y = sin (2x) (6.252)
dx2
Solution:
Step 1 : Find the solution for the homogeneous LDE
d2 y
+ 4y = 0: (6.253)
dx2
The roots to the indicial equation
k2 + 4 = 0 (6.254)
140 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
are
k1 = 2i; k2 = 2i: (6.255)
Then the homogeneous part of the solution can be expressed as
or preferably
yh (x) = A0 e2ix + B 0 e 2ix
: (6.257)
Step 2 : When the non-homogenous part f (x) is a trigonometric or hyper-
bolic function (i.e. sin (ax) ; cos (ax) ; sinh (ax) ; or cosh (ax)), before we
try to …nd the particular solution it is recommended to express the func-
tions as exponential function applying Euler’s formula. Thus for function
in Eq. (6.252), one should write
1 2ix 2ix
f (x) = sin (2x) = e e : (6.258)
2i
Now comparing Eqs. (6.257) and (6.258), we note that values to the ex-
ponents to the Homogeneous solution (the roots of the indicial equation)
are the same as the exponents to the function, f (x) : Under these circum-
stances, one can make an educated guess to the particular solution based
on the property of an exponential function and the order of the di¤eren-
tial equation. For example, for the DE in Eq. (6.252), if one guesses a
particular solution
yp (x) = C1 e2ix + C2 e 2ix
which has the same form as the homogeneous solution in Eq. (6.257), we
…nd
d2 yp
+ 4yp = 4yp + 4yp = 0;
dx2
as one would have guessed. Therefore, we must guess a function that must
have a di¤erent form the homogenous solution. In this case, our educated
guess be
yp (x) = x C1 e2ix + C2 e 2ix ; (6.259)
that leads to
dyp
= C1 e2ix + C2 e 2ix + 2ix C1 e2ix C2 e 2ix
dx
d 2 yp
) = 2i C1 e2ix C2 e 2ix + 2i C1 e2ix C2 e 2ix
dx2
4x C1 e2ix + C2 e 2ix (6.260)
2
d yp
) + 4yp = 4i C1 e2ix C2 e 2ix
: (6.261)
dx2
Substituting this result into Eq. (??), we …nd
1 2ix
4i C1 e2ix C2 e 2ix
= e e 2ix
:
2i
6.5. THE METHOD OF SUPERPOSITION AND THE PARTICULAR SOLUTION.141
x e2ix + e 2ix
x
yp (x) = = cos (2x) : (6.264)
4 2 4
where
yp1 (x) = Pn (x) e2ix ; yp2 (x) = Qn (x) e 2ix
: (6.267)
with Pn (x) = C1 x and Qn (x) = C2 x are polynomial functions with degree,
n = 1: To determine the particular solution, we also wrote the function, f (x) =
sin (2x) ; as a sum of two functions such that
d2 yp (x)
+ 4yp (x) = f (x) = f1 (x) + f2 (x) ; (6.268)
dx2
where
f1 (x) = Pm (x) e2ix ; f2 (x) = Qm (x) e 2ix
; (6.269)
1 1
with Pm (x) = 2iand Qm (x) = are polynomial of degree, m = 0: In a
2i
similar way, the particular solutions to
d2 yp0 1 (x) 0
d2 yp0 2 (x)
+ 4yp1 (x) = f 1 (x) and + 4yp0 2 (x) = f2 (x) ; (6.270)
dx2 dx2
can be guess to be
yp0 1 (x) = Pn0 (x) e2ix ; yp0 2 (x) = Q0n (x) e 2ix
: (6.271)
142 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
with Pn0 (x) = C10 x and Q0n (x) = C20 x are polynomial functions with degree, n =
1; respectively. Upon adding these two equations, we …nd a none homogeneous
ODE in Eq. (6.272) has the same form as Eq. (6.268)
d2 yp0
+ 4yp0 = f1 (x) + f2 (x) = f (x) ; (6.272)
dx2
where
yp0 (x) = 2yp0 1 (x) + 2yp0 2 (x) = Pn00 (x) e2ix + Q00n (x) e 2ix
; (6.273)
with Pn00 (x) = 2C10 x = C100 x and Q00n (x) = 2C20 x = C200 x are still the same
polynomials with the same degree, n = 1 obtained the previous example by
directly considering the function f (x). This can be generalized for an n-th
order none homogeneous ODE
X dn y
an = f (x) (6.274)
n
dxn
where
Qn (x) = a0 + a1 x + a2 x2 + :::an xn : (6.281)
is a polynomial of the same degree as Pn (x) with undetermined coe¢ cients
determined by substituting the particular solution to the given none homogenous
DE.
d2 y
y=0 (6.283)
dx2
the indicial equation
k2 1=0 (6.284)
gives
k1 = 1; k2 = 1: (6.285)
The the solution for the homogenous DE can be written as
yh (x) = C1 ex + C2 e x
: (6.286)
d2 y
y = 2 sin (x) + 4x cos (x) (6.287)
dx2
as
d2 y eix e ix
eix + e ix
y=2 + 4x (6.288)
dx2 2i 2
d d
+1 1 = (2x i) eix + (2x + i) e ix
(6.289)
dx dx
or
d d
+1 1 = f1 (x) + f2 (x) ; (6.290)
dx dx
where
f1 (x) = (2x i) eix ; f2 (x) = (2x + i) e ix
; (6.291)
and
a = 1; b = 1 (6.292)
Then for the non-homogenous DE with the function f1 (x)
d d
+1 1 = f1 (x) = (2x i) eix (6.293)
dx dx
144 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
noting that
c = i 6= a; b (6.294)
we may write the particular solution as
that gives
dyp1
= Beix + i (A + Bx) eix = Beix + iyp1 (x) ; (6.296)
dx
d 2 y p1 dBe ix
dyp1
= +i = iBeix + i Beix + iyp1 (x)
dx2 dx dx
= 2iBeix yp1 (x) (6.297)
d2 yp1 (x)
yp1 (x) = f1 (x) = (2x i) eix : (6.298)
dx2
we …nd
2B = 2; 2Bi 2A = i)B= 1) 2i 2A = i
i
) B= 1; A = (6.301)
2
and the …rst particular solution can be written as
i
yp1 (x) = + x eix : (6.302)
2
Similarly, for the second function we may write the second particular so-
lution
yp2 (x) = (A0 + B 0 x) e ix (6.303)
which gives
dyp2 d2 yp2
) = B0e ix
iyp2 (x) ) = 2iB 0 e ix
yp2 (x) : (6.304)
dx dx2
Then
d2 yp2 (x)
yp2 (x) = f2 (x) : (6.305)
dx2
6.6. THE METHOD OF SUCCESSIVE INTEGRATION 145
becomes
2iB 0 e ix
2 (A0 + B 0 x) e ix
= [ 2B 0 x 2B 0 i 2A] eix
ix
= (2x + i) e (6.306)
There follows that
2B 0 = 2; 2Bi 2A = i ) B 0 = 1 ) 2i 2A = i
i
) B0 = 1; A0 = (6.307)
2
and the second particular solution is
i ix
yp2 (x) = x e (6.308)
2
Therefore, the general solution is given by
i i
y (x) = C1 ex + C2 e x
+ x eix + x e ix
(6.309)
2 2
or
y (x) = C1 ex + C2 e x
+ sin (x) 2x cos (x) : (6.310)
Using Mathematica!
in the form
d2 d d d
a2 + a1 + a0 y = a b y = F (x) ; (6.312)
dx2 dx dx dx
d
u (x) = b y (x) ; (6.314)
dx
d du
a u (x) = au = F (x) : (6.315)
dx dx
ax
Multiplying this equation by e ; we have
ax du ax ax
e ae u=e F (x) ; (6.316)
dx
so that applying the product rule, one can write
d ax ax
ue =e F (x) : (6.317)
dx
Integrating this equation with respect to x, we …nd
Z x
0
ue ax = e ax F (x0 ) dx0 + C1
Z x
0
) u = eax e ax F (x0 ) dx0 + C1 eax ; (6.318)
that leads to Z x
ax0
u (x) = eax e F (x0 ) dx0 + C1 eax ; (6.319)
Once again applying the product rule this can be put in the form
Z x
d bx (a b)x 0
e y =e e ax F (x0 ) dx0 + C1 e(a b)x ; (6.322)
dx
so that integrating with respect to x, we can write
Z x Z x
00 0
e bx y = e(a b)x e ax F (x0 ) dx0 dx00
Z x
00
+C1 e(a b)x dx00 + C2 ; (6.323)
where a and b are given by Eq. (6.313). Noting that the …rst term involving the
function F (x) is the a result of the none homogeneity of the DE and the second
and third terms are independent of this function, we can rewrite Eq. (6.324) as
where Z x
b)x00
yH (x) = C1 ebx e(a dx00 + C2 ebx (6.326)
and " #
Z x Z x00
00 0
yp (x) = ebx e(a b)x
e ax
F (x0 ) dx0 dx00 (6.327)
are the homogeneous and particular solutions, respectively. Next we will use
Eqs. (??) and (??) to derive the solutions we determined in the previous sections
using an educated guess.
Problem 1: Derive the integral expression for 3-rd and 4-th order none ho-
mogeneous ODE. From the results obtained try to write an expression for an
n-th order none homogeneous ODE with constant coe¢ cients.
Problem 2: Applying the general solution derived for a …rst-order ODE (Eq.
(6.128) in section 6.3, derive Eqs. (6.319) and (6.324).
d2 y dy
a2 2
+ a1 + a0 y = 0; (6.328)
dx dx
148 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
a2 k 2 + a1 k + a0 = 0 (6.329)
the roots are given by Eq. (6.313). When these roots are none degenerate
(a 6= b), we saw that the general solution to DE is given by
Next we will derive these particular solutions from Eq. (6.327). To this
end, using the expression for n-th degree polynomial function, Pn (x) ;
n
X
Pn (x) = Ai x i ; (6.336)
n=0
6.6. THE METHOD OF SUCCESSIVE INTEGRATION 149
one can write the function, F (x) ; that makes the ODE none homogeneous,
as
X n
F (x) = ecx Pn (x) = ecx Ai x i : (6.337)
n=0
Z " Z #
x n
X x00
00
yp (x) = ebx e(a b)x
Ai e(c a)x0 0i
x dx0 dx00 (6.338)
n=0
Z " Z #
n
X x x00
bx x00 x0 0i 0
= Ai e e e x dx dx00
n=0
where
=c a and =a b: (6.339)
Next we will evaluate Eq. (6.338) to derive the particular solutions given
in Eq. (6.335).
di x0 x 0i
[e ]=e x ; (6.340)
d i
Substituting this equation into Eq. (6.338), we can write the particular
solution as
n
X Z x
x00 di x00
yp (x) = ebx Ai e e dx00
i=0
d i
n
X Z x
bx di + )x00
= e Ai i e( dx00 : (6.342)
i=0
d
n
X
bx di e( + )x
yp (x) = e Ai : (6.343)
i=0
d i +
150 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
Noting that
d0 e( + )x
e( + )x
= = C 0 e( + )x
;
d 0 + +
" #
d e( + )x
d 1 x
= 2 + e( + )x
d + d ( + ) +
= (C0 + C1 x) e( + )x ;
" #
d2 e( + )x
1 x
= 2 + e( + )x
d 2 + ( + ) +
" #
2 x x x2
= 3 2 2 + e( + )x
( + ) ( + ) ( + ) +
= C0 + C1 x + C2 x2 e( + )x
; (6.344)
one can easily show that
Xn Z
di x ( + )x00 00
Ai i e dx = C0 + C1 x + C2 x2 + :::Cn xn e( (6.345)
+ )x
i=0
d
= Qn (x) e( + )x
;
where Qn (x) is a polynomial of degree n. Note that the constants Ai are
absorbed in the constants Ci :Thus, substituting, = c a and = a b
in Eq. (6.345), the particular solution in Eq. (6.343) becomes
yp (x) = ebx Qn (x) e(c a+a b)x
= Qn (x) ecx : (6.346)
(ii.) When c = a & c 6= b: for this case we have = c a = 0 and Eq.
(6.338) becomes
n
X Z x" Z x00 #
bx x00
yp (x) = Ai e e x dx dx00 :
0i 0
(6.347)
n=0
Substituting this into Eq. (6.327), the particular solution can be expressed
as
Z
Ai bx x h x00 00 i+1 i 00
Xn
yp (x) = i+1
e e (x ) dx (6.348)
i=0
n+1
X Z x
Aj 1 x00 j
= ebx j
e (x00 ) dx00 ;
j=1
Following a similar procedure we used in (ii) earlier, one can show that
n+1
X Aj 1 dj e x
= C1 x + C2 x2 + :::Cn+1 xn+1 e x
j=1
j
d j
= x C0 + C1 x + C2 x2 :::Cn xn e x :(6.350)
Thus the particular solution in Eq. (6.349) becomes
yp (x) = ebx xe x Qn (x) = xecx Qn (x) ; (6.351)
where Qn (x) is a polynomial of degree n. This result is the same as the
solution in Eq. (6.335) when c = a & c 6= b or c = b & c 6= a.
(iii.) When c = a =b : for this case we have =c a = 0 and =a b and
Eq. (6.348) becomes
n+1
X Z x
Aj 1 j
yp (x) = ebx j
(x00 ) dx00 : (6.352)
j=1
the mass is at rest with the spring at its equilibrium position. (i.e. the
spring is neither stretched nor compressed.)
152 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
Solution: In order to …nd the amplitude of oscillation one needs to solve the
equation of motion of the mass. Applying Newtons’s second low one can
write the equation of motion for the mass, m, as
d2 x
m = Fe kx = 2 sin (3t) kx: (6.355)
dt2
Using the values for the spring constant, k = 8N=m; and mass, m = 2kg;
that can be put in the form
d2 x d2 d d
+ 4x = +4 x= + 2i 2i x = sin (3t) : (6.356)
dt2 dt2 dt dt
In view of the general solutions in Eqs. (6.326) and (6.327) that we de-
rived by successively integrating the second-order none homogeneous ODE
given in Eq. (6.312), the homogeneous and the particular solutions to Eq.
(6.356) can be expresses as
Z t
00
bt
xH (t) = C1 e e(a b)t dt00 + C2 ebt (6.357)
and Z t" Z #
t00
bt (a b)t00 at0 0 0
xp (t) = e e e F (t ) dt dt00 : (6.358)
so that one can also carry out similar integration to this equation and …nd
2it
e e5it e it 1 1 e3it e3it
xp (t) = = sin (3t) : =
2i 5i2 5i2 5 5 2i
(6.365)
Using the results in Eqs. (6.362) and (6.365), the general solution, x (t) =
xH (t) + xp (t) ; becomes
1
x (t) = A cos (2t) + B sin (2t) sin (3t) : (6.366)
5
Initially (t = 0), the mass is at rest with the spring at its equilibrium
position. (i.e. the spring is neither stretched nor compressed.). This
means for the position of the mass given by Eq. (6.366), we …nd
x (t = 0) = 0 ) A = 0; (6.367)
and for the velocity, that can easily be determined from x (t) ;
dx (t) 3
v (t) = = 2 ( A sin (2t) + B cos (2t)) cos (3t) ; (6.368)
dt 5
we also obtain
3 3
v (t = 0) = 0 ) 2B =0)B= : (6.369)
5 10
3 1
x (t) = sin (2t) sin (3t) : (6.370)
10 5
Using Mathematica:
154 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
In this last section of this chapter, we will be introduced to the technique for
solving PDE and …nd the unknown function as it applied to real physical systems
or processes. This technique is known as the separation of variables technique
described by the following successive steps that build upon the methods for solve
ODE that we were introduced to in the previous sections.
2nd . Substitute the product function into the original PDE and simplify.
4th . Try to separate the variables, getting only one variable on one side of the
equation.
5th . The two sides of the equation must then be equal to the same constant,
called the separation constant. (Think about what to name this constant—
you can save yourself a lot of needless work by choosing a good name!)
6th . Solve the resulting ordinary DE’s using techniques that you were intro-
duced to in the previous sections.
7th . The general solution is then a linear combination of all possible product-
form solutions to the PDE, including all possible values of the separation
constant!
(a) Show that the string supports a wave motion by deriving the wave equation.
@ 2 u (z; t) 1 @ 2 u (z; t)
= ; (6.375)
@z 2 v2 @t2
156 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
where s
T
v= ; (6.376)
(b) Solve the 1-D traveling wave equation for a string of length L having both
ends …xed. Which means, apply the BC’s
Solution:
(a) Because of the weight of the mass hanging the string experiences a tension
force T . When is displaced from the equilibrium by the vibrator as shown
in the …gure below:If the string is displaced from the equilibrium, the net
6.7. PARTIAL DIFFERENTIAL EQUATIONS 157
Noting that
0 @u @u
tan = ; tan ( ) = ; (6.380)
@y y=y+ y @y y=y
we may write !
@u @u
F =T : (6.381)
@y y=y+ y @y y=y
@u u(y + y) u(y)
= lim (6.382)
@y y=y
y!0 y
@2u
F 'T y (6.385)
@y 2
Noting that for the transverse motion, Newton’s second law can be written
as
@2u
F = ma = m 2 (6.386)
@t
If the string has a linear mass density , we can write m = y so that
@2u
F = y (6.387)
@t2
Combining the two expressions for the force F; we …nd
@2u @2u
T y= y (6.388)
@y 2 @t2
158 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
Which leads to
@2u @2u @2u 1 @2u
= or = (6.389)
@y 2 T @t2 @y 2 v 2 @t2
where s
T
v= : (6.390)
we have
X
u(y; t) = Ak (Ck cos (kvt) + Dk sin (kvt)) = 0 (6.397)
k
This can be zero for all time t; if an only if when Ak = 0: This leads to
simpli…ed expression for the general solution
X
u(y; t) = Bk sin (ky) [Ck cos (kvt) + Dk sin (kvt)] : (6.398)
k
6.8. HOMEWORK ASSIGNMENT 12 159
where
n
kn = ; n = 0; 1; 2:3::: (6.403)
L
This concludes the chapter to the introduction to di¤erential equations.
3. Consider 109 Gallons volume salty lake. Suppose a salty water with a salt
density
sw = 5 103 lb=gal
begins to ‡ow at a rate of
dVin
=4 105 gal=h:
dt
into the salty lake. At the same time the salty water in the lake begin to
‡ow out at a rate of
dVout
= 105 gal=h:
dt
(a) Determine the di¤erential equation describing the volume of the salty
water, V (t) ; in the lake. Solve the di¤erential equation to …nd the volume
of the salty water as a function of time.
(b) Find the di¤erential equation that describes the mass, m (t) of the salt in
the salty lake. Solve the di¤erential equation to …nd the mass of the salt
in the lake as function of time if the mass of the salt in the lake at the
initial time, before the in and out ‡ows begin, is m(0) = 107 lb:
~2 @ 2 (x; t) @
+ U (x) (x; t) = i~ (x; t)
2m @x2 @t
where m is the mass of the particle, ~ is Planck’s constant (1:05 10
34J:s), i is the pure imaginary number, and U (x) is the net potential
energy function of the particle at position x. (Note that, since Fx =
dU (x)=dx, specifying the potential energy function in quantum mechan-
ics is equivalent to specifying all of the forces acting on the particle in
classical mechanics.) The form of the Schröudinger equation given above
is called the time-dependent (1-D) Schröudinger equation (TDSE) since
it involves the full time-dependent wave function, (x; t).
(a) Using the separation of variables technique, and letting the separation
constant be denoted E (which turns out to be the total energy of the
particle), show that the resulting di¤erential equation that is independent
of time— the so-called time-independent Schröudinger equation (TISE)—
has the (1-D) form
~2 d2 (x; t)
+ U (x) (x; t) = E (x; t)
2m dx2
where we have set (x; t) = (x) T (t). In applications, the potential
energy function U (x) must be speci…ed, and the resulting TISE solved for
the spatial wave function (x)
(b) Also, show that the total wave function in this case has the form
iEt=~
(x; t) = (x) e :
162 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
Chapter 7
Integral calculus-scalar
functions
163
164 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
y (x) dy (x)
m (x) = lim = : (7.1)
x!0 x dx
is a constant for all x from the …rst to the second point along the …rst path
but not along the second path. For any of these paths the net distance traveled
along the y-direction, that can be determined from the line integral
Z y2 (x2 ) Z x2
y = y2 (x2 ) y1 (x1 ) = dy = m (x) dx: (7.2)
y1 (x1 ) x1
are the same. Similarly, along the x-direction, the net distance given by
Z y2
x = x2 x1 = m0 (y) dy; (7.3)
y1
where
x (y) dx (y)
m0 (y) = lim = ; (7.4)
y!0 y dy
are the same for the two paths.
However, for the total distances (arc length), l, for these two paths are di¤er-
ent, as one can imagine. For the …rst path this length can easily be determined
using Pythagorean theorem and is given by
q q
2 2 2 2
l = (y2 y1 ) + (x2 x1 ) = ( y) + ( x) : (7.5)
For the second path, since the slope is not constant, one need to divide the
path into an in…nitessimal intervals, li ; …nd the lengths using Pythagorean
theorem, and sum it up all to …nd all. This means
s s
X X yi
2 X xi
2
l= lim li = lim xi +1= lim yi + 1:
i
li !0
i
xi !0 xi i
yi !0 yi
(7.7)
Noting that, generally, integration of a function f (x) is de…ned as
X Z2
lim fi ! df; : (7.8)
fi !0
i 1
and di¤erentiation as
df (x) f (x)
= lim (7.9)
dx x!0 x
one can write
v" # s
Z2 Zx2u
u dy 2 Zy2 2
dx
l= dl = t + 1 dx = 1+ dy: (7.10)
dx dy
1 x1 y1
This is the arc-length (or the total distance) for any path one can follow to go
from P1 to P2 :
B. Double integral-Area
Let’s add two points P3 and P4 with coordinates (x1 ; y2 ) and (x2 ; y1 ), re-
spectively, to form the rectangular region shown in Fig. 7.1.1. The area of this
rectangle depends on the length l (the distance between P1 and P4 ) and the
width w (the distance between P1 and P3 )
A = lw: (7.11)
166 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
The points along the length of the rectangle, l, are de…ned by the equation
and s
Zy2 2 Zy2
dx
w= 1+ dy = dy; (7.16)
dy
y1 y1
respectively. The area of the rectangle can then be expressed using the double
(or surface) integral
Zx2 Zy2 Zx2Zy2
A = dx dy = dxdy; (7.17)
x1 y1 x1 y1
or
Zy2 Zx2 Zy2Zx2
A = dy dx = dydx: (7.18)
y1 x1 y1 x1
Note that the order of integration does not bring any di¤erence in the result.
This result can easily be generalized for any arbitrary shape in which the bound-
aries can be de…ned by a well-de…ne real function. Consider the region bounded
by the two curves shown in Fig. 7.1.1. Let the points on the lower boundary are
de…ned by the function y1 (x) and the upper boundary by the function y1 (x) :
In this case the double integral for the area in Eq. () should be expressed as
2 3
Zx2 yZ2 (x)
6 7
A= 4 dy 5 dx: (7.19)
x1 y1 (x)
On the other hand, if the points in the lower boundary and the upper boundary
are de…ned by the functions x1 (y) and x2 (y), respectively, instead (see Fig. ),
The order of integration needs to be switched and the area should be expressed
7.1. INTEGRATION IN CARTESIAN COORDINATES 167
as 2 3
Zx2 xZ2 (y)
6 7
A= 4 dy 5 dx; (7.20)
x1 x1 (y)
Note that even though the order of integrations are switched, the result should
be the same.
C. Triple integral-Volume
ZZZ
V = dxdydx: (7.21)
Example 1 Line and surface integral : Consider a wire bent into the arc of a
circle of radius R, as shown in the …gure below.
(a) You climbed along this wire. Determine the height of the elevation that
you climbed if you have moved a distance of half of the radius along the
x-direction.
(b) Find the length of the wire
(c) Find the area bounded by the curve, the x and y axes
Solution:
168 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
(b) Find the length of the wire is given by using the are-length in Eq. (7.10)
v" #
Zx2u
u dy 2
l= t + 1 dx: (7.25)
dx
x1
Z=2 Z=2
R2 cos ( ) R
l= q d =R d = : (7.28)
2
R2 1 sin ( ) 2
0 0
x = R sin ( ) ) =0
and x = R
1
x = R sin ( ) ) = sin (1) = =2
Noting that the x and y axes are de…ned by the equations y = 0 and
x = 0; respectively, for the area bounded and integration with respect to
x …rst, we have
0p 1
R2 y 2
ZR Z ZRp
B C
A= B @ dxC
A dy = R2 y 2 dy: (7.31)
0 0 0
170 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
Example 2 Volume integral : A triangle in the x-y plane has vertices at the
points (x; y; z) = (0; 0; 0); (1; 0; 0);and (0; 2; 0). Find the volume of space
that lies above this triangle, but below the plane z = 2+x+y. All distance
units are in centimeters.
Solution: The volume is given by
ZZZ
V = dxdydx: (7.35)
The equation of the line joining the point (1; 0; 0) and (0; 2; 0) is given by
y 2 2 0
= ) y = 2x + 2 (7.36)
x 0 0 1
Then the volume is given by
8 2 3 9
Z < y=Z 2x+2 z=2+x+y
x=1 Z =
V = 4 dz 5 dy dx: (7.37)
: ;
x=0 y=0 z=0
with respect to y
Z
x=1 2x+2 Z
x=1
y2
V = 2y + xy + dx = [2 ( 2x + 2) + x ( 2x (7.39)
+ 2)
2 0
x=0 x=0
# Z
x=1
2
( 2x + 2)
+ dx = [6 6x] dx;
2
x=0
A. Electrostatic Potential
The electrostatic properties of a given charge distribution can all be derived
from the electrostatic potential, V (r). In an introductory physics courses
you have introduced to the electrostatic potential of a point charge. This
electrostatic potential at a point in space described by the position vector;
~r; measured from a point charge with charge, q; that is placed at the origin,
is given by
q
V (r) = ; (7.41)
4 0 j~rj
if the point charge is placed in a free space (with an electrical permittivity,
0 ) in the absence of any other charge. When the charge is not a point
charge and instead the charge is a line, surface, or volume charge with
none uniform distribution, the electrostatic potential, V (r); is determined
by carrying out single or multiple integrations that depends on the given
charge distributions. For the given charge distribution let’s consider an
in…nitely small charge dq 0 at a point described by the position vector, ~r0 ;
from the origin of the coordinate system. The corresponding in…nitessimal
electrostatic potential, dV (r); at a point in space described by the position
vector, ~r; that is also measured from the origin, depends on the distance of
this point from the position of the in…nitesimal charge, dq 0 : This distance
R (r0 ) is determined from the magnitude of the position vector directed
from the position of dq 0 to the position of the point in space that we want
to determine the corresponding electrostatic potential given by
0
R (r0 ) = ~r ~r : (7.42)
dq 0
dV (r) = : (7.43)
4 r ~r0 j
0 j~
Line charge density : Suppose you are given a charge that is distributed on
a …nite length of wire with length, L. Also the charge is not uniformly
distributed on the wire and the charge distribution is described by a line
a line charge density,
dq 0
(~r0 ) = ) dq 0 = (~r0 )dr0 : (7.44)
dr0
Using this equation, the total electrostatic potential due to the entire
charge over the entire length can be determined from Eq. (7.43) using the
172 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
line integral
Z Z
dq 0 1 (~r0 )
V (r) = = dr0 (7.45)
4 r ~r0 j
0 j~ 4 0 L j~r ~r0 j
Surface charge density: Any external charge that one may put on a metal-
lic ball is distributed only on the surface of the sphere no matter the
magnitude of the charge. This charge distribute itself uniformly or none
uniformly depending on the absence or presence of charges in surround-
ing regions of the sphere. let’s consider the general case none uniform
distribution that can be described by a surface charge density,
dq 0
(~r0 ) = ) dq 0 = (~r0 )da0 ; (7.46)
da0
where da0 is the in…nitessimal area over which the in…nitessimal charge
dq 0 is distributed on: In this case the electrostatic potential at a point ~r
due to the entire charge on the surface of the sphere is given
Z Z
1 (~r0 )
V (r) = da0 : (7.47)
4 0 j~r ~r0 j
Area
B. The total charge: Like the total electrostatic potential depends on the
charge densities and determined by single or multiple integrations, so does
the total charge. The total charge, Q, distributed on a line is determined
by single integral Z
Q= (~r0 )dr0 ; (7.51)
Line
on a surface and volume by multiple integrations given by
Z Z
Q= (~r0 )da0 ; (7.52)
Area
7.1. INTEGRATION IN CARTESIAN COORDINATES 173
and
ZZ Z
Q= (~r0 )dv 0 ; (7.53)
V olum
respectively.
where ri (r) is the perpendicular distance from the axis of rotation and
dm dm
(r) = = (7.56)
dv dxdydz
B. The total mass: The total mass, M , in a volume with mass density, (r)
is given by
ZZ Z
M= (r)dxdydz: (7.57)
V ol m
174 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
C. Center of Mass: The center of mass (gravity) for a given mass with uni-
formly or nonuniform distribution is a point at which the mass can be
at rotational equilibrium about an axis passing through this point. The
center mass, in Cartesian coordinates, (x; y; z) can be determined using
Z Z Z Z Z Z
x dm = xdm; y dm = ydm; and z dm = zdm: (7.58)
If mass is distributed along a line on the x-axis with linear mass density,
we have
dm
(x) = ) dm = (x) dx (7.62)
dx
and the center of mass becomes
R
(x) xdx
x= R ; y = 0; and z = 0: (7.63)
dm
Similarly for surface mass distribution on the x-y plane, with surface mass
density
dm dm
(x; y) = = ) dm = (x; y) dxdy (7.64)
da dxdy
the center of mass is given by
RR RR
x (x; y) dxdy y (x; y) dxdy
x= RR ;y = R R ; and z = 0: (7.65)
(x; y) dxdy (x; y) dxdy
The next examples illustrates the application of single and multiple integra-
tions to determine the physical quantities described.
Example 3 A thin rod of length L lies along the positive x-axis with one end
at the origin. It has a linear charge density given by (x) = Ax, where A
is a constant.
7.1. INTEGRATION IN CARTESIAN COORDINATES 175
Using the result in part b for A; we may express the electrostatic potential
as
Q b b+L
V (r) = 1 ln : (7.72)
2 0L L b
Solution:
charge
(a) The MKS unit of A must be the unit of length 5
since we must get a unit
charge
for the surface charge density (x; y) after we multiply A by x2 y
length 2
length3 :
x02 + y 02 = R2 (7.74)
the equation of curved part of the semi-circle in the …gure can be written
as p
y 0 (x0 ) = R2 x02 : (7.75)
7.1. INTEGRATION IN CARTESIAN COORDINATES 177
= A y 4
0 5 dy 0 (7.81)
0 3
Noting that r is perpendicular distance of the mass, dm; (in the volume,
dv = dxdydz, that is centered at a point with coordinates (x; y; z)) from
the axis of symmetry (z-axis), we can write
p
r = x2 + y 2 ; (7.87)
where
0 r R: (7.88)
Furthermore, since the cylinder, with height H; is centered about the z-
axis,
H H
z : (7.89)
2 2
Thus the moment of inertia can be expressed as
Z z2 =H=2 "Z y2 =R "Z x2 =pR2 y2 # #
M 2 2
I= x + y dx dy dz:
R2 H z1 = H=2 y1 = R x1 = pR2 y2
(7.90)
Since the integrand is independent of z we can carry out integration with
respect to z followed by integration with respect to z. This leads to
2 p 2 2 3
Z R 3 R y
M 4 x 5
I = + y2 x p 2 2 dy H
R2 H R 3 R y
Z R
M R2 + 2y 2 p 2
= 2
2 R y 2 dy: (7.91)
R R 3
7.1. INTEGRATION IN CARTESIAN COORDINATES 179
we may write
Z =2
M R2 + 2R2 sin2 ( )
I = 2 cos2 ( ) d
=2 3
Z =2
2M R2
= cos2 ( ) + 2 sin2 ( ) cos2 ( ) d : (7.93)
3 =2
Example 6 Consider a wire of constant linear mass density bent into the arc
of a circle of radius R, as shown in the …gure below.
(a) What is the total mass of the wire?
(b) Find the wire’s center-of-mass.
(c) Find the area under the arc of the wire.
(d) The wire is now rotated about the x-axis. Find the volume inside the
surface thus generated.
(e) Exercise: Find the surface area of this curved surface of rotation.
180 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
Solution:
For uniformly distributed linear mass over curved length the linear mass
density, (l) ; given by
dm
(l) = = ) dm = dl
dl
the total mass becomes Z
M= dl = l (7.97)
Using the result in Example 1(b) for the arc-length of the curve
R
l= ;
2
the total mass becomes
R
M= : (7.98)
2
(b) The mass is distributed uniformly on a curved line on the x-y plane. Thus
coordinates of the center of mass are given by
Z Z Z Z
x dm = xdm; y dm = ydm; and z = 0 (7.99)
and s
Z R 2
2 dx
y= y 1+ dy: (7.104)
R 0 dy
The curve is de…ned by the equation of a circle
p p
x2 + y 2 = R2 ) y = R2 y 2 or x = R2 y2 ; (7.105)
we chose the positive roots for both x and y since the in the …rst quadrant
both are positives. Using the expressions for x and y we have
dx y dy x
= p ; = p (7.106)
dy R 2 y2 dx R2 x2
so the one can write
v" #
Z R uu 2 Z R
2 x 2 xdx
x= xt p + 1 dx = p xdx;
R 0 R 2 x2 0 R 2 x2
(7.107)
and
v2 3
Z u !2 Z
R u R
2 u y 2 ydy
y= y t4 p + 15dx = p : (7.108)
R 0 R 2 y2 0 R2 y 2
2 p R
2 p R
x= R2 x2 ;y = R2 x2 ; (7.109)
0 0
182 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
or (7.20) 2 3
Zx2 xZ2 (y)
6 7
A= 4 dy 5 dx: (7.112)
x1 x1 (y)
Here we chose to use Eq. (7.19). For the boundaries shown in the …gure,
we have p
x1 (y) = 0; x2 (y) = R2 x2 (7.113)
and
x1 = 0; x2 = R: (7.114)
so that the area becomes
Z R " Z pR 2 x2
# Z R p
A= dy dx = R2 x2 dx: (7.115)
0 0 0
(d) The rotation of the wire about the x-axis creates a hemisphere of radius, R;
shown in the …gure. A point on the surface of the sphere with coordinates
(x; y; z) is de…ned by the equation of a sphere
p
x2 + y 2 + z 2 = R2 ) x = R2 (y 2 + z 2 ) (7.121)
But here we will use a short cut. If we consider the in…nitesimal volume
to be a cylinder of radius y and thickness (i.e. height) dx; we can write
dv = y 2 dx (7.123)
da = dxdy; (7.129)
Using
@x @x
dx = dr + d = cos ( ) dr r sin ( ) d (7.130)
@r @
and
@y @y
dy = dr + d = sin ( ) dr + r cos ( ) d (7.131)
@r @
one can write
da = rdrd : (7.135)
da = rdrd (7.141)
where r? is the perpendicular distance of the mass dm from the given axis
of rotation and is the mass density. In cylindrical coordinates this can be
expressed as Z Z Z Z
I = r? dm = r? rdrd'dz (7.148)
ZH ZR Z2
M 1
I= r3 drd'dz = M R2 (7.152)
R2 H 2
z=0x=0 0
R2 H
V = : (7.153)
3
(H z) R
0 < ' < 2 ;0 < r < ; 0 < z < H; (7.155)
H
and the volume can be written as
Z 2 "Z H Z (H
H
z)R ! #
V = rdr dz d': (7.156)
0 0 0
can be expressed as
Z Z
'
Q= 0 sin ( ) cos R2 sin ( ) d d' (7.164)
4
For a full sphere the limits of integration for is (0 ) and for ' is
(0 ' 2 ): Thus the total charge can be determined by evaluating the
integral
Z 2 Z
'
Q= 0R
2
sin2 ( ) cos d d'
0 0 4
Z Z 2
'
= 0R
2
sin2 ( ) d cos d' (7.165)
0 0 4
190 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
Recalling that
Z
sin 2
sin2 ( ) d = = (7.166)
0 2 4 0 2
and Z 2 2
' '
cos d' = 4 sin =4 (7.167)
0 4 4 0
where h is the height of the zone. Thus the area of the surface area of the
zone of a sphere becomes
A = 2 Rh; (7.173)
which is indeed independent of the position of the two planes forming the
zone.
4. For the area (A) bounded by the parabola y = x2 and the straight line
2x y + 8 = 0; evaluate the integral
Z Z
I= xdxdy = 36
A
5. Using double integrals determine the volume bounded above the square
plane formed by the vertices (0; 0),(2; 0),(0; 2), and (2; 2) and below the
plane z = 8 x + y:
192 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS
Vector Calculus
~ B
A ~ = Ax B x + Ay B y + Az B z (8.1)
If the angle between the two vectors is the dot product can be determined
using
~ B
A ~ = A ~ cos
~ B (8.2)
~ and B
Cross product (vector product): the cross product of A ~ results in a
~ given by
third vector C
x^ y^ z^
~ =A
C ~ ~ = det
B Ax Ay Az (8.3)
Bx By Bz
~ = (Ay Bz
C Az B y ) x
^ + (Az Bx Ax Bz ) y^ + (Ax By Ay Bx ) z^: (8.4)
~ is given by
If the angle between the two vectors is ; the magnitude of vector C
~ = A
C ~ ~ = A
B ~ B
~ sin (8.5)
~ is always
and the direction is determined using the right-hand rule. Vector C
perpendicular to the plane formed by the two vectors.
195
196 CHAPTER 8. VECTOR CALCULUS
~ B;
Triple scalar and vector products: Consider the three vectors A; ~ and C
~
in Cartesian coordinates
~ = Ax x
A ~ = Bx x
^ + Ay y^ + Az z^; B ~ = Cx x
^ + By y^ + Bz z^; C ^ + Cy y^ + Cz z^: (8.6)
These three vectors can be multiplied to give a scalar or a vector. The product
of these vectors that leads to a scalar is called a triple scalar product and is
given by
Cx Cy Cz
~ A
C: ~ B ~ = det Ax Ay Az : (8.7)
Bx By Bz
Recalling that interchanging two rows changes only make the value of the de-
terminant negative, one can write
Cx Cy Cz Bx By Bz Ax Ay Az
Ax Ay Az = Ax Ay Az = Bx By Bz : (8.8)
Bx By Bz Cx Cy Cz Cx Cy Cz
Referring to the expression for the triple scalar product, one can easily see that
~ A
C: ~ ~ =
B ~ A
B: ~ ~ = A:
C ~ B~ ~ :
C (8.9)
one can establish the relation for triple scalar product of three vectors
~ A
C: ~ ~ = B:
B ~ C~ ~ = A:
A ~ B~ ~ :
C (8.10)
These same three vectors can also be multiplied to yield another vector instead
of a scalar. Such multiplication is known as triple vector product that can be
expressed as
A~ B~ C ~ =B ~ A~ C
~ C~ A ~ B~ .
The expression for triple vector product can be memorized using the phrase
"BAC-CAB RULE ".
W = F~ d~ cos = F~ d:
~ (8.11)
8.1. VECTOR PRODUCTS 197
For none constant force that depends on the position of the object, F~ = F~ (~r) ;
the work done is determined from an integral form of the scalar product for the
force and displacement vectors given by
Z r2
W = F~ d~r: (8.12)
r1
~v = !
~ ~
d: (8.13)
However, when the angular velocity is not a constant and depends on the relative
position of the object, !
~ =!
~ (~r), the average linear velocity is determined using
the line integral,
Z ~r2
~v = !
~ d~r: (8.14)
~
r1
Angular momentum: Suppose the object for which we described the constant
linear velocity, ~v ; and constant angular velocity, !
~ has a mass, m. This mass
has a linear momentum, p~ = m~v : The angular momentum, J; ~ of this object, is
given by the vector product,
J~ = ~r p~: (8.15)
When the linear momentum depends on the position of the object the average
angular momentum becomes
Z
J~ = d~r p~: (8.16)
Magnetic …eld and force: Suppose an object with charge, q; is moving with a
~ The magnetic
constant velocity, ~v ; in a region with a uniform magnetic …eld, B:
force experienced by the object is given by
F~B = q~v ~
B: (8.20)
For a none uniform magnetic …eld, B ~ (r) ; such as the magnetic …eld due to a
wire of length l that carries a current, I: where the magnetic …eld is given by
~l
~ (r) = 0I ~r
B ; (8.21)
4 r3
the magnetic force can be expressed using a triple vector product
~v ~l ~r
q 0I
F~B (~r) = : (8.22)
4 r3
~ (t) = Ax (t) x
A ~ (t) = Bx (t) x
^ + Ay (t) y^ + Az (t) z^; B ^ + By (t) y^ + Bz (t) z^: (8.24)
Using these vectors we can establish the following derivatives with respect to
time:
Derivative of a vector:
~
dA dAx dAy dAz
= x
^+ y^ + z^: (8.25)
dt dt dt dt
8.1. VECTOR PRODUCTS 199
d ~
cA ~ dc + dA c
~ =A (8.26)
dt dt dt
Derivative of scalar vector product:
d ~ ~ ~ ~
~ dB + dA B
A B =A ~ (8.27)
dt dt dt
Derivative of vector product:
d ~ ~
dB ~
dA
A ~ =A
B ~ + ~
B (8.28)
dt dt dt
Example 8.1 The position of an object at time, t; is described by the two
dimensional vector, ~r (t) : This vector, using Polar coordinates, (r; ) ; can
be expressed as
~r (t) = r cos ( ) x
^ + r sin ( ) y^: (8.29)
Find the general expressions for the velocity, ~v (t) ; and acceleration, ~a (t) ;
and express your answer in Polar Coordinates.
Solution:
The velocity: ~v (t) ; is given by
d~r d d
~v (t) = = (r cos ( )) x
^+ (r sin ( )) y^ (8.30)
dt dt dt
Assuming the general case where both r and are time dependent, upon
di¤ erentiating, we …nd
dr d cos ( ) dr d sin ( )
~v (t) = cos ( ) x
^+r x
^+ sin ( ) y^ + r y^
dt dt dt dt
dr d dr d
) ~v (t) = cos ( ) x
^ r sin ( ) x ^+ sin ( ) y^ + r cos ( (8.31)
) y^
dt dt dt dt
that can put in the form
dr d
~v (t) = (cos ( ) x
^ + sin ( ) y^) + r ( sin ( ) x
^ + cos ( ) y^) : (8.32)
dt dt
Noting that in two-dimensional Polar coordinates, the unit vectors along
the radial direction (i.e. along ~r) is
d~r d~r
r^ = = cos ( ) x
^ + sin ( ) y^; (8.33)
dr dr
and along the angular direction is
^ = d~r d~r
= sin ( ) x
^ + cos ( ) y^; (8.34)
d d
200 CHAPTER 8. VECTOR CALCULUS
Further more, using the equation relating the linear tangential and the
angular speeds,
vt = !r; (8.39)
the instantaneous can also be written as
~v (t) = vr r^ + vt ^: (8.40)
When the magnitude of the radial displacement is constant (i.e. the motion
is in a circular orbit), we have
dr
vr = =0 (8.41)
dt
and the instantaneous velocity becomes
d~v d dvr d^
r dvt ^ d^
~a (t) = = vr r^ + vt ^ = r^ + vr + + vt (8.43)
dt dt dt dt dt dt
Noting that as it can be noted in the expressions, both unit vectors are
time dependent and we have
d^
r d d^ d ^
= (cos ( ) x
^ + sin ( ) y^) = ( sin ( ) x
^ + cos ( ) y^) = = !^
dt dt dt dt
(8.44)
and
d^ d d^
= ( sin ( ) x
^ + cos ( ) y^) = ( cos ( ) x
^ sin ( ) y^)
dt dt dt
d^ d
) = r^ = !^ r: (8.45)
dt dt
8.1. VECTOR PRODUCTS 201
d2 r dr ^;
~a (t) = r! 2 r^ + 2 !+r (8.50)
dt2 dt
where we used
dr d! d2 r
at = ! + r; ar = 2 ; vt = !r (8.51)
dt dt dt
and the angular acceleration de…ned by
d2
= (8.52)
dt2
For a circular orbit where r is constant of motion (i.e. ~r = R^
r), we have
dr d2 r
= 0; 2 = 0 (8.53)
dt dt
and
~v (t) = vt ^ = !R^ (8.54)
so that
vt2
~a (t) = R! 2 r^ + R ^ = r^ + R ^ = ~ac + ~ajj (8.55)
R
where
vt2
~ac = r^ (8.56)
R
and
d!
~ajj = R ^ = R ^ (8.57)
dt
are the centripetal and tangential accelerations, respectively.
202 CHAPTER 8. VECTOR CALCULUS
~ B
A ~ = (Ax x^ + Ay y^ + Az z^) (Bx x
^ + By y^ + Bz z^) = Ax Bx + Ay By + Az Bz ;
(8.59)
one can re-write df as
@f @f @f
df = (dx^
x + dy y^ + dz z^) x
^+ y^ + z^ (8.60)
@x @y @z
Noting that the di¤erential for the position vector in Cartesian coordinates,
~r = x^
x + y y^ + z z^;
is given by
d~r = dx^
x + dy y^ + dz z^; (8.61)
one can put df in the form
@f @f @f @ @ @
df = d~r x
^+ y^ + z^ = d~r x
^+ y^ + z^ f; (8.62)
@x @y @z @x @y @z
where we factored out the function f to the right side of the partial di¤eren-
tial operators. The operators in the bracket are represented by the gradient
operator,O
~;
@ @ @
O
~ =x ^ + y^ + z^ : (8.63)
@x @y @z
The gradient operator operates on a scalar function, f (x; y; z) to generate a
vector function
@f @f @f
O
~f = x
^+ y^ + z^: (8.64)
@x @y @z
Thus the in…nitesimal change, df , in the function f (x; y; z) becomes
df = O
~ f dr: (8.65)
8.2. DIFFERENTIAL OPERATORS 203
The directional derivative: The function, f (x; y; z); de…nes a surface in three
dimensional space. Suppose this function is a constant, f (x; y; z) = C. Then
the total di¤erential for the function, df; becomes
~
df = rf d~r = 0; (8.66)
indicating that the scalar product of the vectors rf ~ and d~r is zero and the
vectors are perpendicular to one another at the point (x; y; z). Since d~r =
dx^
x + dy y^ + dz z^; is an in…nitesimal displacement on the surface de…ned by
the function, f (x; y; z); the vector d~r is tangent to the surface at the point.
~ must be a vector normal to the surface at the same point.
Therefore, rf;
Now let’s consider an in…nitesimal displacement, d~s; on the surface de…ned
by the function, f (x; y; z). This displacement can be expressed as
d~s = ds^
n = ds (nx ; ny ; nz ) ; (8.67)
where
n
^ = nx x
^ + ny y^ + nz z^ (8.68)
is a unit vector along the direction of d~s: Thus one can write
d~s = ds^
n = d~r; ) ds (nx ; ny ; nz ) = (dx; dy; dz)
df @f dx @f dy @f dz
= + + : (8.70)
ds @x ds @y ds @z ds
df @f @f @f
= nx + ny + nz
ds @x @y @z
@f @f @f
= x
^+ y^ + z^ (nx x
^ + ny y^ + nz z^) (8.71)
@x @y @z
df
= rf n
^;
ds
where n^ is the unit vector along the direction of the vector S ~ that is tangent
to the surface de…ned by the function, f (x; y; z) at a point on the surface with
coordinates (x; y; z).
204 CHAPTER 8. VECTOR CALCULUS
@f '
^ @f @f
rf = r^ + + z^ (8.72)
@r r @' @z
@f ^ @f '
^ @f
rf = r^ + + ; (8.73)
@r r@ r sin @'
Example 8.2 A point charge Q is located at the origin. The electrostatic po-
tential at a point P a distance r from the point charge is given by
kQ
(r) = ; (8.74)
r
where k is the Coulomb’s law constant
1
k= ; (8.75)
4 0
(a)
@ 1 @ '
^ @
r = r^ +^ + (8.77)
@r r@ r sin @'
(b)
@ @ @
O= x
^+ y^ + z^ (8.78)
@x @y @z
Solution:
8.2. DIFFERENTIAL OPERATORS 205
(a) Since the electric potential depends only on r, using spherical coordinates
is easier than Cartesian. Using the gradient in spherical coordinates, one
can then write the electric …eld as
@ kQ 1 @ kQ 1 @ kQ
~ (~r) =
E r (r) = r^ +^ +'^
@r r r@ r r sin @' r
~ (~r) = @ kQ kQ
)E r^ = 2 r^ (8.79)
@r r r
kQ
(r) = ; (8.80)
r (x; y; z)
where p
~r = x^
x + y y^ + z z^ ) r = x2 + y 2 + z 3 : (8.81)
Then we may write
Noting that the partial derivative, for example with respect to x; can be
determined as
@ (r) d (r) dr d kQ d hp 2 i
= = x + y2 + z3
@x dr dx dr r dx
@ (r) kQ x kQ x kQ
) = p = = x: (8.83)
@x r 2 x2 + y 2 + z 3 r2 r r3
@ (~r) kQ y kQ @ (~r) kQ z kQ
= = y; = = z: (8.84)
@y r2 r r3 @z r2 r r3
kQ kQ kQ kQ kQ
r (~r) = x^
x y y^ z z^ = (x^
x + y y^ + z z^) = ~r:
r3 r3 r3 r3 r3
(8.85)
Noting that the unit vector along ~r is
~r
r^ = (8.86)
r
the electric …eld is found to be
~ (~r) = kQ
E r (~r) = r^: (8.87)
r2
206 CHAPTER 8. VECTOR CALCULUS
Here n^ is the unit vector along the vector S~ tangent to the surface de…ned
by the electrostatic potential function (r) at a point ( x; y; z). This unit
vector is given by
S~ 2 2 1
n
^= = x ^ y^ + z^: (8.90)
S~ 3 3 3
We are interested in the change in the potential, (r) ; along this direction
at a point, (1; 2; 1) ; on the surface. This is given by the directional
derivative
d (r)
= r (r) n ^ : (8.91)
ds (1;2; 1)
Using the result we obtained in part (a) or (b) and the unit vector, n
^ ;one
can then write
d (~r) kQ 2 2 1
= (x^
x + y y^ + z z^) x
^ y^ + z^ (8.92)
ds r3 3 3 3 (1;2; 1)
d (~r) kQ 2 2 1
= x y+ z (8.93)
ds r3 3 3 3 (1;2; 1)
d (~r) kQ 2 2 1
= 3=2
x y+ z (8.94)
ds (x2 + y2 + z2) 3 3 3
(1;2; 1)
Exercise: Do you thing we can …nd the angle between the electric …eld vector
E~ (~r) at point ( 1; 2; 1) ( if we assume KQ = 1)and S?
~
gravitational …eld due to a given mass, or magnetic …eld due to a current carry-
ing wire. For vector functions the gradient operator acts in two di¤erent ways
to describe the divergence and the curl (rotation) of the vector function. For a
vector function
~ (x; y; z) = Vx (x; y; z) x
V ^ + Vy (x; y; z) y^ + Vz (x; y; z) z^ (8.96)
~ (x; y; z) ; denoted by r V
the divergence of V ~ is given by
x
^ y^ z^
r ~ = det
V @ @ @
(8.98)
@x @y @z
Vx Vy Vz
that leads to
where
@2 @2 @2
r2 = + + (8.102)
@x2 @y 2 @z 2
is called the Laplacian operator. As an example, let’s consider the electric …eld
vector, which is given by
E~ (~r) = r (~r) : (8.103)
~ (~r) leads to
The divergence of E
~ (~r) = @2 @2 @2
r E r r (x; y; z) = + + = r2 (x; y; z) :
@x2 @y 2 @z 2
(8.104)
In classical electrodynamics the divergence of the electric …eld is related to the
volume charge density, (x; y; z) ; by the di¤erential form of Gauss’law,
~ (~r) = (x; y; z)
r E ; (8.105)
0
208 CHAPTER 8. VECTOR CALCULUS
~ (~r) = @2 @2 @2 (x; y; z)
r E r r (x; y; z) = 2
+ 2 + 2 =
@x @y @z 0
2
)r (x; y; z) = F (x; y; z) : (8.106)
where
(x; y; z)
F (x; y; z) = : (8.107)
0
Equation (8.106) is known as Poison’s equation. In a free space where there is
no charge, (x; y; z) = 0; we …nd
r2 (x; y; z) = 0; (8.108)
which is known as Laplace’s equation. Often you encounter Poison’s and Laplace’s
equations in classical electromagnetism and mechanics theories.
You also encounter, more frequently, in most branches of theoretical physics
that could involve the gradient, the divergence, and the curl of a scalar and/or
a vector functions. Thus it worth to remember the following relations:
~ ; multiplied
(a) The divergence of a vector and a scalar : For a vector function, V
by a scalar functions f; the divergence can be expressed as
~
r: f V ~ rf + f r:V
=V ~ (8.109)
(b) The curl of the curl : In view of the relation for triple vector product
~
A ~
B ~ =B
C ~ A
~ C
~ ~ A
C ~ B
~ (8.110)
~ ; multiplied by
(c) The curl of a vector and a scalar : For a vector function, V
a scalar functions f; the curl can be expressed as
r ~
fV = (rf ) ~ +f r
V ~
V =f r ~
V ~
V rf (8.112)
Example 8.3 Consider two functions— a scalar function and a vector function—
de…ned as follows:
~ (x; y; z) = (xy; yz; zx) :
f (~r) = xyz; V (8.113)
At the point (1; 1; 1), evaluate the following quantities :
8.2. DIFFERENTIAL OPERATORS 209
(a) rf
~
(b) r V
(c) r ~
V
(d) r2 f
~
(e) r2 V
~
(f ) r2 f V
Solution:
(a) The gradient of the scalare function, f , given by
@f @f @f
rf = x
^+ y^ + z^: (8.114)
@x @y @z
rf = x
^ + y^ z^: (8.116)
~ = (Vx ; Vy ; Vz ), is given by
(b) The divergence of the vector function, V
~ = 1:
r V (8.119)
x
^ y^ z^
r ~ = det
V @ @ @
(8.120)
@x @y @z
xy yz xz
210 CHAPTER 8. VECTOR CALCULUS
that leads to
r ~ =x
V ^ y^ z^: (8.122)
~
@2V @ 2 (xy^
x + yz y^ + zx^
z)
= =0 (8.129)
@y 2 @y 2
~
@2V @ 2 (xy^
x + yz y^ + zx^
z)
= =0 (8.130)
@z 2 @z 2
8.2. DIFFERENTIAL OPERATORS 211
~ =r r V
r2 V ~ r r ~
V
and
~ =r
U ~ =
V y^
x z y^ x^
z: (8.133)
Noting that
~ @ @
r r V = r (y + z + x) = x
^ (y + z + x) + y^ (y + z + x)
@x @y
(8.134)
@f ~
+^
z (y + z + x) ) r r V =x
^ + y^ + z^
@z
and
x
^ y^ z^
r r ~
V =r ~ = det
U @ @ @
(8.135)
@x @y @z
y z x
that can be simpli…ed into
~ @ (y z) @ ( x) @ ( y) @ ( x)
r r V = x
^+ y^
@y @z @z @x
@ ( z) @ ( y) ~ =x
+ z^ ) r r V ^ + y^ + z^: (8.136)
@x @y
~ =r r V
r2 V ~ r r ~
V =0 (8.137)
(f ) Noting that
~
r2 f V ^r2 (f Vx ) + y^r2 (f Vz ) + z^r2 (f Vz )
=x (8.138)
and
2 2 2
@ 2 (xy) z @ 2 (xy) z @ 2 (xy) z
2
r (f Vx ) = + + = 2y 2 z + 2x2 z
@x2 @y 2 @z 2
(8.139)
212 CHAPTER 8. VECTOR CALCULUS
2 2 2
@ 2 (yz) x @ 2 (yz) x @ 2 (yz) x
2
r (f Vy ) = + + = 2z 2 x + 2y 2 x
@x2 @y 2 @z 2
(8.140)
2 2 2 2 2 2
@ (zx) y @ (zx) y @ (zx) y
r2 (f Vz ) = + + = 2z 2 y + 2x2 y
@x2 @y 2 @z 2
(8.141)
we …nd
~
r2 f V ^ 2y 2 z + 2x2 z + y^ 2z 2 x + 2y 2 x + z^ 2z 2 y + 2x2 y (8.142)
=x
Using Mathematica
The gradient, the divergence, the Laplacian, and the curl that we have seen
so far is for scalar and vector functions expressed in terms of Cartesian coordi-
nates. Depending on the problem we want to solve, sometimes it is convenient
to evaluate these quantities in curvilinear coordinates by expressing the func-
tions using the corresponding transformations for the Cartesian to the conve-
nient curvilinear coordinates. Thus it is important to know the expression for
these quantities in the two curvilinear coordinates we studied (cylindrical and
spherical coordinates). These expressions can be derived from its corresponding
expressions in Cartesian. For now we list these expressions in cylindrical and
spherical coordinates.
(a) Cylindrical coordinates (r; '; z): For a scalar function, f (r; '; z); the gradi-
ent and Laplacian are given by
@f 1 @f @f
rf = r^ +'
^ + z^ (8.143)
@r r @' @z
8.2. DIFFERENTIAL OPERATORS 213
and
1 @ @f 1 @2f @2f
r2 f = r + 2 2
+ 2; (8.144)
r @r @r r @' @z
respectively. For a vector function,
and
r ~ = r^ 1 @Vz
V
@V'
+'
^
@Vr @Vz
+ z^
1 @
(rV' )
@Vr
:
r @' @z @z @r r @r @'
(8.146)
respectively.
(b) Spherical coordinates (r; ; '): For a scalar function, f (r; ; '); the gradient
can be expressed as
@f ^ 1 @f 1 @f
rf = r^ + +'
^ ; (8.147)
@r r@ r sin @'
1 @ @f 1 @ @f 1 @2f
r2 f = r2 + sin ( ) + 2 :
r2 @r @r r2 sin ( ) @ @ r2 sin ( ) @'2
(8.148)
On the other hand for a vector function,
~ = 1 @ r2 Vr +
r V
1 @
(sin ( ) V ) +
1 @V'
(8.149)
r2 @r r sin ( ) @ r sin ( ) @'
and
1 @ @V 1 @Vr 1 @
r ~ = r^
V (sin ( ) V' ) +^ (rV' )
r sin ( ) @ @' r sin ( ) @' r @r
1 @ @Vr
+^' (rV ) ; (8.150)
r @r @
respectively.
214 CHAPTER 8. VECTOR CALCULUS
~ =
B 0 Ir=2 a2 '
^ for r < a (8.151)
(a) Using parametric plot show that the magnetic …eld lines form a concentric
circle.
~ 6=
(b) Show that magnetic …eld vector is a rotational vector …eld (i.e. the r B
0) inside the cylinder
~ = 0:
(c) Show that r B
Solution:
(a)
8.2. DIFFERENTIAL OPERATORS 215
(b) We note that the components of the magnetic …eld in cylindrical coordinates
are
Ir
Br = 0; B' = 0 2 ; Bz = 0:
2 a
Then using the expression for the curl of a vector in cylindrical coordinates
~ = r^ 1 @Bz
r B
@B'
+'
^
@Br @Bz
+ z^
1 @
(rB' )
@Br
r @' @z @z @r r @r @'
(8.152)
we …nd
~ @B' 1 @
r B = r^ + z^ (rB' )
@z r @r
@ 0 Ir 1 @ Ir
= r^ 2
+ z^ r 0 2 (8.153)
@z 2 a r @r 2 a
2
)r ~ = z^ 1 @
B 0 Ir
= 0I
z^ (8.154)
r @r 2 a 2 a2
216 CHAPTER 8. VECTOR CALCULUS
~ (~r) = kQ
E r (~r) = r^: (8.158)
r2
Show that this electric …eld vector is irrotational vector …eld.
Solution:
The electric …eld vector in spherical coordinates has components
kQ
Er = ; E = E' = 0
r2
then the curl of the electric …eld vector in spherical coordinates
~ = r^ 1 @ @E
r E (sin ( ) E' ) (8.159)
r sin ( ) @ @'
1 @Er 1 @ 1 @ @Er
+^ (rE' ) + '
^ (rE ) (8.160)
r sin ( ) @' r @r r @r @
becomes
r ~ = 0:
E (8.161)
This means that the electric …eld vector is irrotational vector …eld.
Exercise: Find the divergence of the electric …eld and try to make a physical
interpretation for the result.
8.3. SINGLE VECTOR FUNCTION INTEGRALS 217
Solution:
The closed line integral can easily be carried out if we divide the integral into
three line integrals corresponding to the three sides ( s1; s2;and s3),
I Z Z Z
~ r) d~r =
A(~ ~ r) d~r +
A(~ ~ r) d~r +
A(~ ~ r) d~r
A(~ (8.164)
C s1 s2 s3
~r = x^
x + y y^ (8.165)
so that
d~r = dx^
x + dy y^ (8.166)
218 CHAPTER 8. VECTOR CALCULUS
x = 0 ) ~r = y y^ ) d~r = dy y^ (8.167)
for s2 p p
y= 2 ) ~r = x^
x+ 2^
y ) d~r = dx^
x (8.168)
and for s3
p p dx
y= x ) ~r = x^
x+ x^ x + p y^:
y ) d~r = dx^ (8.169)
2 x
The function
~ r) = x2 y^
A(~ x + xy 2 y^ + a3 e y
x cos ( ) z^; (8.170)
also takes di¤ erent expressions on these three sides of the curve. On s1
~ r) = 0
x = 0 ) A(~ (8.171)
on s2
p p p
y= ~ r) = x2 2^
2; z = 0 ) A(~ y + a3 e
x + 2x^ 2
x cos ( ) z^ (8.172)
and on s3
p p
~ r) = x5=2 x
y = x; z = 0 ) A(~ ^ + x2 y^ + a3 e x
x cos ( ) z^ (8.173)
on s2
Z Z p p
p Z
2
2 p 2
~ r) d~r = 2 2 2 3 8 2
A(~ x 2^
x + 2x^
y dx^
x= 2 x dx = x =
s2 0 0 3 3
0
(8.175)
and on s3
Z Z 0 Z 0
~ r) d~r = dx x2
A(~ x5=2 x
^ + x2 y^ x + p y^ =
dx^ x5=2 + p dx
s3 2 2 x 2 2 x
Z 0 0
1 2 1
x5=2 + x3=2 dx = x7=2 + x5=2
2 2 7 5 2
Z p p
) ~ r) d~r = 16 2 4 2 :
A(~ (8.176)
s3 7 5
we …nd
I p p p p
~ r) d~r = 8 2 16 2 4 2 44 2
A(~ = (8.177)
C 3 7 5 105
8.3. SINGLE VECTOR FUNCTION INTEGRALS 219
If the line integral over a closed curve is zero independent of the path,
I
~ (x; y; z) d~r = 0:
V (8.179)
C
r ~ (x; y; z) = 0:
V (8.180)
Example 8.7 Consider the moon and the earth: Imagine the origin of a spher-
ical coordinate system coincides with the center of the earth. Then the
position of the moon can be described by ( r; ; ') in spherical coordinates.
The gravitational potential energy of the earth and the moon is given by
GM m
U= (8.182)
r
where r is the distance between the earth and the moon, M is mass of the
earth, and m is mass of the moon. (See Fig. 8.1)
F~ = r U (8.183)
on the moon using the gravitational potential energy.
(b) Show that gravitational force is a conservative force using the work done
by the gravitational force when the moon makes one complete revolution
around the earth. I
W = F~ d~r
C
Solution:
(a) The gravitational force is given by
F~ = r U (8.184)
using the gradient in spherical coordinates
@U ^ 1 @U 1 @U
rU = r^ + +'
^ (8.185)
@r r @ r sin @'
we …nd
@U @ GM m GM m
F~ = r^ = r^ = r^ (8.186)
@r @r r r2
(b) The work done is given by
I I
GM m
W = F~ (x; y; z) d~r = r^ d~r (8.187)
C C a2
for a circular orbit r is a constant. That means in spherical coordinates
the position vector
~r = a^
r (8.188)
becomes
d~r = ad^
r: (8.189)
Recalling that in spherical coordinates
~r = r sin ( ) cos (') x
^ + r sin ( ) sin (') y^ + r cos ( ) z^
the unit vector along the radial direction
@~r @~r
r^ =
@r @r
= sin ( ) cos (') x
^ + sin ( ) sin (') y^ + cos ( ) z^ (8.190)
we may write
d^
r = [cos ( ) cos (') d sin ( ) sin (') d'] x
^
+ [cos ( ) sin (') d + sin ( ) cos (') d'] y^ sin ( ) d z^: (8.191)
8.3. SINGLE VECTOR FUNCTION INTEGRALS 221
r^ d~r = sin ( ) cos ( ) cos2 (') d sin2 ( ) sin (') cos (') d'
+ sin ( ) cos ( ) sin2 (') d + sin2 ( ) sin (') cos (') d'
sin ( ) cos ( ) d :
) r^ d~r = 0 (8.192)
~ d~r ) Vx = @U ; Vy = @U ; Vz = @U :
dU = rU d~r = V (8.195)
@x @y @z
For a conservative vector …eld, we know that
x
^ y^ z^
~ (x; y; z) = det @ @ @ @Vz @Vy @Vx @Vz
r V @x @y @z = x
^+ y^
@y @z @z @x
Vx Vy Vz
@Vy @Vx
+ z^ = 0 (8.196)
@x @y
which leads to
@Vz @Vy @Vx @Vz @Vy @Vx
= 0; = 0; =0
@y @z @z @x @x @y
@Vz @Vy @Vx @Vz @Vy @Vx
) = ; = ; = : (8.197)
@y @z @z @x @x @y
Recalling that we assumed the scalar function U is a function related to a
conservative vector …eld V by
@U @U @U
Vx = ; Vy = ; Vz = (8.198)
@x @y @z
222 CHAPTER 8. VECTOR CALCULUS
we may write
@U @U @U
dU = dx + dy + dz; (8.200)
@x @y @z
where the line integral is counterclockwise around the boundary of the area A
(i.e. on the curve C).
Proof:
8.4. MULTIPLE VECTOR INTEGRALS 223
We note that
ZZ ZZ ZZ
@Q(x; y) @P (x; y) @Q(x; y) @P (x; y)
dxdy = dxdy dydx:
A @x @y A @x A @y
(8.205)
For the area shown in the …gure belowwe may write
ZZ Z b Z yu
@P (x; y) @P (x; y)
dydx = dy dx (8.206)
A @y a yl @y
we …nd
ZZ Z b
@P (x; y)
dydx = [P (x; yu ) P (x; yl )] dx
A @y a
Z b Z b Z a Z b
= P (x; yu )dx P (x; yl )dx = P (x; yu )dx P (x; yl )dx
a
ZaZ b
I a
@P (x; y)
) dydx = P (x; y)dx: (8.208)
A @y C
Note that the line integral over the closed curve C is the boundary of the
area A and it must be integrated in a counterclockwise direction. Because
the limits of integration in the two line integrals above indicate that we
perform the integration …rst on the upper part of the curve from b ! a and
then from a ! b: Similarly if we perform the integration in the reverse
224 CHAPTER 8. VECTOR CALCULUS
order (i.e. …rst with respect to x and then with respect to y) for the
integral
ZZ Z d Z xr
@Q(x; y) @Q(x; y)
dxdy = dx dy (8.209)
A @x c xl @x
as shown in the …gure below,for the same area then we can write
ZZ Z d Z d Z d
@Q(x; y)
dxdy = [Q(xr ; y) Q(xl ; y)] dy = Q(xr ; y)dy Q(xl ; y)dy
A @x c c c
Z d Z c ZZ I
@Q(x; y)
= Q(xr ; y)dy + Q(xl ; y)dy ) dxdy = Q(x; y)dy
c d A @x C
(8.210)
Here also the line integral over the same curve C is performed in the coun-
terclockwise direction like the previous integration. Therefore, combining
the two relations
ZZ I ZZ I
@P (x; y) @Q(x; y)
dydx = P (x; y)dx; dxdy = Q(x; y)dy;
A @y C A @x C
(8.211)
we …nd
ZZ I
@Q(x; y) @P (x; y)
dxdy = [P (x; y)dx + Q(x; y)dy] : (8.212)
A @x @y C
Example 8.8 Let the area A be the region bounded by the curves y = x2 and
x = y 2 . Verify Green’s theorem for the integral
I
I= 2xy x2 dx + x + y 2 dy (8.213)
C
8.4. MULTIPLE VECTOR INTEGRALS 225
Solution:
Comparing the given integral with
ZZ I
@Q(x; y) @P (x; y)
dxdy = [P (x; y)dx + Q(x; y)dy] : (8.214)
A @x @y C
we want to verify that the result of
I I
I1 = [P (x; y)dx + Q(x; y)dy] = 2xy x2 dx + x + y 2 dy
C C
(8.215)
is equal to
ZZ
@Q(x; y) @P (x; y)
I2 = dxdy
A @x @y
ZZ " #
@ x + y2 @ 2xy x2
= dxdy (8.216)
A @x @y
for the area shown in the …gure belowAnother short cut for the line integral
is the following:
Z
I1 = 2xy x2 dx + x + y 2 dy
Cl
Z
+ 2xy x2 dx + x + y 2 dy
Cu
226 CHAPTER 8. VECTOR CALCULUS
For the lower curve y = x2 ) dy = 2xdx and for the upper curve x =
y 2 ) dx = 2ydy and noting that the two curves intersect at two points
(0; 0) and (1; 1), we can write
Z 1 Z 0
I1 = 2x3 x2 dx + x + x4 2xdx + 2y 3 y 4 2ydy + 2y 2 dy
0 1
(8.217)
Z 1 Z 0
I1 = 2x3 x2 + 2x2 + 2x5 dx + 4y 4 2y 5 + 2y 2 dy (8.218)
0 1
Z 1 Z 0
5 3 2
I1 = 2x + 2x + x dx + 4y 4 2y 5 + 2y 2 dy (8.219)
0 1
1 0
x6 x4 x3 4y 5 y6 2y 3 1
I1 = + + + + = (8.220)
3 2 3 0 5 3 3 1 30
For the area integral
ZZ " # Z 1 Z py
@ x + y2 @ 2xy x2
I2 = dxdy = (1 2x) dxdy
A @x @y 0 y2
Z 1 p Z 1
2 y p
= x x y2 dy = y y y 2 + y 4 dy
0 0
2 3 5 1
2 3=2 y y y 2 1 1 1 1
= y + = + = (8.221)
3 2 3 5 0 3 2 3 5 30
Consider the surface bounded by the curve, C, on the x-y plane as shown in the
…gure below.Using an in…nitesimal area dA in this region one can establish the
relation
n
^ dA = z^dxdy; (8.223)
where n^ = z^ is the unit vector normal to this area. On the curve, an in…nites-
simal displacement can be written as
d~r = dx^
x + dy y^: (8.224)
~ (x; y) = Dx (x; y) x
D ^ + Dy (x; y) y^ + Dz (x; y) z^ (8.225)
8.4. MULTIPLE VECTOR INTEGRALS 227
and
~ n @Dy @Dx
r D ^ dA = dxdy:
@x @y
Using these two expressions the Stoke’s theorem
I Z
~
D d~r = r ~ n
D ^ dA: (8.227)
Curve Surf ace A
bounding A
can be expressed as
I Z
@Dy @Dx
(Dx dx + Dy dy) = dxdy: (8.228)
Curve Surf ace A @x @y
bounding A
1. The curve C bounds the open area A. The integral must therefore be the
same for any area A bounded by the curve C. This means that you can
pick an easy area A if you wish to make life simpler!
2. The integral Z
r ~ n
D ^ dA (8.229)
A
3. The integral I
~ d~r
D (8.230)
C
Recall Green’s Theorem in the x-y Plane: Let’s consider two functions P (x; y)
and Q(x; y). If these two functions and their …rst derivatives are continuous for
all x and y in the region bounded by the curve C of area A, then Green’s theorem
states
ZZ I
@Q(x; y) @P (x; y)
dxdy = [P (x; y)dx + Q(x; y)dy] ; (8.231)
A @x @y C
where the line integral is counterclockwise around the boundary of the area A
(i.e. on the curve C). Note that for
~ = Dx (x; y) x
D ^ + Dy (x; y) y^ = P (x; y)^
x + Q(x; y)^
y (8.232)
we …nd ZZ I
@Dy @Dx
dxdy = [Dx dx + Dy dy] : (8.233)
A @x @y C
This shows that Green’s theorem is the two-dimensional form of Stoke’s theorem.
x
^ y^ z^
r ~ = det
D @ @ @
(8.237)
@x @y @z
Dx Dy Dz
becomes
! !
~ = @ yz 2 @ ( xz) @ (3y) @ yz 2
r D x
^+ y^
@y @z @z @x
@ ( xz) @ (3y)
+ z^ = z 2 + x x
^ (3 + z) z^ (8.238)
@x @y
for z = 2, we …nd
r ~ = (4 + x) x
D ^ 5^
z (8.239)
Then the ‡ux becomes
Z Z
r ~ n
D ^ dA = [(4 + x) x
^ 5^
z] n
^ dA
A A
Z 2 Z p
4 x2
= p [(4 + x) x
^ 5^
z ] z^dxdy (8.240)
2 4 x2
which gives
Z Z Z p ! Z
2 4 x2 2 p
r D ~ n
^ dA = 5 dy dx = 10 4 x2 dx
p
A 2 4 x2 2
(8.241)
230 CHAPTER 8. VECTOR CALCULUS
we …nd Z
r ~ n
D ^ dA = 20 : (8.242)
A
we have
~
D(x; y; z) d~r = 3y; xz; yz 2 (dx; dy) = 3ydx xzdy (8.244)
I Z 2 p
~ d~r 2x2
D = 3 4 x2 + p dx
C 2 4 x2
Z 2 p 2x2
+ 3 4 x2 + p dx (8.246)
2 4 x2
I Z 2 p
~ d~r 2x2
D = 3 4 x2 + p dx
C 2 4 x2
Z 2 p 2x2
3 4 x2 + p dx (8.247)
2 4 x2
I Z 2 p
~ d~r = 2x2
D 2 3 4 x2 + p dx (8.248)
C 2 4 x2
By evaluating the integral we …nd
I
~ d~r =
D 20 (8.249)
C
(a) Compute the ‡ux v through the closed rectangular surface of sides a; b;
and c shown in the …gure by direct application of the de…nition of ‡ux.
(b) Compute v through the same surface indirectly by applying the Divergence
theorem.
Solution:
~ (x; y; z) over a given surface area A is
(a) The ‡ux v for the vector …eld V
de…ned as ZZ
v = ~ (x; y; z) n
V ^ da:
The …gure above have three pairs of surfaces normal to the three axis. The
‡ux through the surfaces normal to the x axis
ZZ ZZ
vx = ~
V (a; y; z) x
^ dydz + ~ (0; y; z) ( x
V ^) dydz (8.252)
232 CHAPTER 8. VECTOR CALCULUS
Z c Z b Z c Z b
) vx = ~ (a; y; z) x
V ^dydz ~ (0; y; z) x
V ^dydz (8.253)
0 0 0 0
Z c Z b Z c Z b
) vx = Vx (a; y; z)dydz Vx (0; y; z)dydz (8.254)
0 0 0 0
Z c Z a Z c Z a
) vy = Vy (x; b; z)dxdz Vy (x; 0; z)dxdz: (8.257)
0 0 0 0
and Z Z
c a
a2 bc
vz = cxdxdy = : (8.259)
0 0 2
Therefore, the total ‡ux will be
a2 bc ab2 c abc2
v = + + : (8.260)
2 2 2
8.4. MULTIPLE VECTOR INTEGRALS 233
which leads to
a2 bc ab2 c abc2
v = + + : (8.263)
2 2 2
Example 8.10 Verify the divergence theorem for the vector …eld
~ = ar^
V r + b^
' + cz z^ (8.265)
Solution:
Noting that
Vr = ar; V' = b; V z = cz (8.266)
the divergence
becomes
~ = 1 @ r2 a + 1 @ (b) + @ (cz) = 2a + c
r V (8.268)
r @r r @' @z
234 CHAPTER 8. VECTOR CALCULUS
For the surface integral over the closed cylindrical surface we can write
ZZ ZZ
~
V (r; '; z) n
^ da = ~b (r; '; z) n
V ^ b da
Surface bottom
enclosing
ZZ ZZ
+ ~t (r; '; z) n
V ^ t da + ~s (r; '; z) n
V ^ s da: (8.271)
top side
We note that
n
^b = z^; n
^ t = z^; n
^ s = r^ (8.272)
and
~b (r; '; z) = V
V ~ (r; '; z = 0) = ar^
r + b^
' (8.273)
~t (r; '; z) = V
V ~ (r; '; z = H) = ar^
r + b^
' + cH z^ (8.274)
8.4. MULTIPLE VECTOR INTEGRALS 235
~s (r; '; z) = V
V ~ (r = R; '; z) = aR^
r + b^
' + cz z^ (8.275)
so that
ZZ ZZ
~b (r; '; z) n
V ^ b da = (ar^
r + b^
') ( z^) da = 0 (8.276)
bottom bottom
ZZ ZZ
~t (r; '; z) n
V ^ t da = (ar^
r + b^
' + cH z^) z^da
top top
ZZ
= cHda = c R2 H (8.277)
top
and
ZZ ZZ
~s (r; '; z) n
V ^ s da: = (aR^
r + b^
' + cz z^) r^da
side side
ZZ ZZ
= aRda = aR (2 RH) ) ~s (r; '; z) n
V ^ s da = 2a R2 H:
side side
(8.278)
Therefore the integral over the closed cylindrical surface
ZZ ZZ
~
V (r; '; z) n
^ da = ~b (r; '; z) n
V ^ b da
Surface bottom
enclosing
ZZ ZZ
+ ~t (r; '; z) n
V ^ t da + ~s (r; '; z) n
V ^ s da: (8.279)
top side
becomes
ZZ
~ (r; '; z) n
V ^ da = c R2 H + 2a R2 H: = (2a + c) R2 H
Surface
enclosing
ZZZ
= ~ (x; y; z) dxdydz
r V (8.280)
volum e
~ = 1 @ r 2 Vr +
r V
1 @
(sin ( ) V ) +
1 @
(V' ) (8.281)
r2 @r r sin ( ) @ r sin ( ) @'
Example 8.11 Verify the divergence theorem for the vector …eld
~ = ~r
V (8.282)
and the sphere shown below.
Solution: For the ‡ux, we …nd
ZZ Z2 Z
~ (r; ; ') n
V ^ da = r r^R2 sin ( ) d d'
R^
surf ace
0 0
ZZ
~ (r; ; ') n
V ^ da = 4 R3
surf ace
236 CHAPTER 8. VECTOR CALCULUS
~ = r^
The divergence of the vector V r
~ = 1 @ r 2 Vr +
r V
1 @
(sin ( ) V ) +
1 @
(V' ) (8.283)
2
r @r r sin ( ) @ r sin ( ) @'
becomes
~ = 1 @ r2 Vr = 1 @ r3 = 3
r V (8.284)
r2 @r r2 @r
Then the integral becomes
ZZZ Z2 Z
~ (r; ; ') d =
r V 3r2 dr sin ( ) d d' = 4 R3 :
volum e
0 0
(8.285)
Example 8.12 Verify Stokes’ theorem for the magnetic …eld vector
r + 3^
~ = 4^
B 2^
' (8.286)
Solution:
Br = 4; B = 3; B' = 2 (8.287)
8.4. MULTIPLE VECTOR INTEGRALS 237
~ = 1 @ @
r B (sin ( ) B' ) (B ) r^
r sin ( ) @ @'
1 @ 1 @ 1 @ @
+ (Br ) (rB' ) ^ + (rB ) (Br ) ' ^
r sin ( ) @' r @r r @r @
(8.288)
becomes
1 @ 1 @ 1 @
r ~ =
B ( 2 sin ( )) r^ ( 2r) ^ + (3r) '
^
r sin ( ) @ r @r r @r
2 cos ( ) 2 3 ~ = 2 cot ( ) r^ + 2 ^ + 3 '
= r^ + ^ + ' ^)r B ^ (8.289)
r sin ( ) r r r r r
Noting that the in…nitesimal area on the x-y plane in spherical coordinates
can be expressed as
h i
^ da = z^r sin ( ) drd' = cos ( ) r^ sin ( ) ^ r sin ( ) drd'
n (8.290)
then
2 2 3 h i
~ n
r B ^ da = cot ( ) r^ + ^ + '^ cos ( ) r^ sin ( ) ^ r sin ( ) drd'
r r r
(8.291)
~ n 2 2
)r B ^ da = cot ( ) cos ( ) sin ( ) r sin ( ) drd': (8.292)
r r
Noting that on the x-y plane = =2; we have
r ~ n
B ^ da = 2drd': (8.293)
so that the integral Z
r ~ n
B ^ dA (8.294)
A
238 CHAPTER 8. VECTOR CALCULUS
for the surface bounded by the curve shown in the …gure can be written as
Z Z R Z =2
r ~ n
B ^ dA = 2drd' = R : (8.295)
A 0 0
~r = r^
r ) d~r = dr^
r + rd^
r (8.296)
where
r^ = cos (') x
^ + sin (') y^ ) d^
r = [ sin (') x
^ + cos (') y^] d' = '
^ d'
(8.298)
so that
~r = r^r ) d~r = dr^
r + r'
^ d': (8.299)
Then
~ d~r = Br dr + B' rd' = 4dr
B 2rd' (8.300)
which leads to
I Z Z Z
~
B d~r = (4dr 2rd')+ (4dr 2rd')+ (4dr 2rd') (8.301)
C l1 l2 l3
where l1 and l3 are the straight lines (on this lines d' = 0) and l2 is the
curved line (on this line r = R and dr = 0). Therefore
I Z R Z =2 Z 0
~ d~r =
B 4dr 2Rd' + 4dr = R : (8.302)
C 0 0 R
r + 3^
~ = 4^
B 2^
' (8.303)
is valid.
Gravity
r ~g = 4 G (8.306)
Ampere’s Law : Ampere’s law states that
I
B~ d~r = 0 Ien (8.307)
C
where B~ is the magnetic …eld due to the current Ien passing through the area
enclosed by the curve C and 0 is the magnetic permeability of free space. The
di¤erential form of Ampere’s law can be obtained if we apply Stokes’theorem.
Using Stokes’theorem we can write
I Z
~ d~r =
B r B ~ n^ dA = 0 Ien (8.308)
C A
If the current passing through the area enclosed by the curve C is described by
~ the total current Ien can be expressed as
the current density J,
Z
Ien = J~ n^ dA (8.309)
A
so that
I Z Z
~ d~r =
B r ~ n
B ^ dA = 0 Ien = 0 J~ n
^ dA
C A A
Z Z
) r ~
B n
^ dA = 0 J~ n
^ dA ) r ~ =
B ~
0 J: (8.310)
A A
which leads to
I Z Z
~
B d~r = r ~ n
B ^ dA ) J~ n
^ dA = 0 Ien = 0I (8.312)
C A A
On the other hand, if we consider the balloon shaped surface shown in the …gure
there is no current passing through this surface and therefore the current enclosed
is zero Z
Ienc = J~ d~a = 0: (8.313)
balloon surf ace
240 CHAPTER 8. VECTOR CALCULUS
Mathematical Methods in
Physics II
241
Chapter 9
Introduction to the
Calculus of Variations
Example 9.1 A ball of mass m is kicked from the ground level with an initial
speed, vo ; at an angle o above the horizontal. Find the time, t; at which
the height of the ball, y(t), becomes stationary.
243
244CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
Solution: We recall that from kinematics of a projectile the motion of the ball
along the y direction is determined by Newton’s second law
dvy dvy
m = mg ) = g ) vy (t) = v0y gt (9.2)
dt dt
Noting that
dy (t)
vy (t) = (9.3)
dt
the value of the time after the ball is kicked that makes the height function
of the ball,y(t), stationary is given by
Example 9.2 Geodesic: Consider two points in a x-y plane P1 and P2 . Prove
that the shortest distance between the two points is the distance measured
along a straight line (i.e. show that the geodesic is given by an equation
of a straight line, y (x) = mx + b.)
Solution: Let’s consider two points on the x-y plane. Let P1 be (x1 ; y1 ) and P2
be (x2 ; y2 ) : Then the distance between these points is given by the integral
Z (2)
L= ds; (9.5)
(1)
where
s s
p dy
2
dx
2
ds = dx2 + dy 2 = 1+ dx = 1+ dy: (9.6)
dx dy
9.1. GEODESIC AND STATIONARY POINTS 245
Out of the in…nitely many functions that can be used to connect the two
points, we want to determine the one that would give the minimum dis-
tance. Let these function be denoted by Y (x) : From these in…nite number
of functions there is only one function that gives the minimum distance
between the two points. If this function is y (x) ; then we may write Y (x)
in terms of y (x) as
where
0 dY (x; )
Y = : (9.12)
dx
We are interested in the path that gives the the minimum distance be-
tween the two points (i.e. the geodesic). The necessary condition for the
distance, L ( ) ; to be minimum is that the length function, L ( ) ; must
have a stationary point at ( = 0; L ( = 0)). This requires
dL ( )
= 0; (9.13)
d =0
which leads to
"Z #
(2)
dL ( ) 1 1 dY 0 (x; )
= p (2Y 0 ) dx = 0:
d =0 (1) 2 1 + Y 02 d
=0
(9.14)
Using
Y (x; ) = y (x) + (x) (9.15)
246CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
we may write
dY dy d
= + ) Y 0 (x; ) = y 0 (x) + 0
(x) (9.16)
dx dx dx
so that
dY 0 ( ) d 0
= [y (x) + 0 (x)] = 0
(x) : (9.17)
d d
There follows that
Y 0 (x; )j =0 = y 0 (x) (9.18)
and
dY 0 ( ) 0
= (x) : (9.19)
d =0
for
0
(x) = dv ) v = (x) ; (9.22)
and !
y0 d y0
u= p ) du = p dx; (9.23)
1 + y0 2 dx 1 + y0 2
we may write the integral as
Z x2 Z !
(2) x2
y 0 0 (x) y0 d y0
p dx = p (x) (x) p dx = 0:
(1) 1 + y0 2 1 + y0 2 x1 dx 1 + y0 2
x1
(9.24)
Due to the conditions
(x1 ) = (x2 ) = 0 (9.25)
the …rst term in the above expression becomes zero. Thus one can write
Z (2) Z x1 !
dL ( ) y 0 0 (x) d y0
= p dx = (x) p dx = 0:
d =0 (1) 1 + y0 2 x1 dx 1 + y0 2
(9.26)
9.2. THE GENERAL PROBLEM 247
dy py
x ! t; y (x) ! y (t) ; y 0 (x) ! y 0 (t) = = vy (t) =
dt m
describing the dynamics of a particle mass, m, moving along the y-direction in
py y^
terms of the parameters (time = t; position = y (t) ; velocity = vy (t) y^ = m ),
where py y^ is the momentum: In classical mechanics, the dynamics of a particle
is determined by an equation derived from Newton’s second law. As we shall
see, this equation can be derived from a more general equation know as the
Euler-Lagrange Equation
@ @F @F
= 0; (9.31)
@x @y 0 @y
where F = F (x; y (x) ; y 0 (x)) is the function that de…nes the surface constructed
by the set of points with coordinates, (x; y (x) ; y 0 (x)):The Euler-Lagrange Equa-
tion is derived by applying the calculus of variation. In general, in the problem
248CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
that we want to solve applying the calculus of variation, we know the coordi-
nates of two di¤ erent points (x1 ; y (x1 ) ; y 0 (x1 )) and (x2 ; y (x2 ) ; y 0 (x2 )) on the
surface de…ned by F = F (x; y (x) ; y 0 (x)): From the in…nitely many trajectories
that can connect these two points, there is only one trajectory on this surface
that is the shortest (the Geodesic). Finding the Geodesic is the general problem
that can be solved applying the calculus of variation.
The surface is de…ned by the function F (x; y (x) ; y 0 (x)): The distance be-
tween these two points determined by evaluating the integral
Z x2
I= F (x; y (x) ; y 0 (x)) dx: (9.32)
x1
To determine the equation that the function F is governed by so that we …nd the
shortest length joining the two points, let the function for any path connecting
the two points be Y (x) : From these in…nite number of functions there is only
one function that gives the minimum distance between the two points. If this
function is y (x) ; then we may write Y (x) in terms of y (x) as
We also have
dY (x; ) dY ( )
= (x) ) = (x) ; (9.36)
d d =0
9.2. THE GENERAL PROBLEM 249
and
dY dy d
= + or Y 0 (x; ) = y 0 (x) + 0
(x) ) Y 0 (x; )j =0 = y 0 (x) ; (9.37)
dx dx dx
which gives
dY 0 ( ) d 0 0 0 dY 0 ( ) 0
= [y (x) + (x)] = (x) ) = (x) : (9.38)
d d d =0
Noting that
d @F dY ( ) @F dY 0 ( )
[F (x; Y (x; ) ; Y 0 (x; ))] = +
d =0 @Y d @Y 0 d =0
0
@F dY ( ) @F dY ( )
= + 0
(9.41)
@Y =0 d =0 @Y =0 d =0
and substituting
dY ( )
Y (x; )j =0 = y (x) ; = (x) ; Y 0 (x; )j =0 = y 0 (x) ;
d =0
dY 0 ( ) 0
= (x) ; (9.42)
d =0
we …nd
d @
[F (x; Y (x; ) ; Y 0 (x; ))] = F (x; y (x) ; y 0 (x)) (x)
d =0 @y
@
+ F (x; y (x) ; y 0 (x)) 0 (x) ; (9.43)
@y 0
so that using
(x1 ) = (x2 ) = 0; (9.47)
we …nd
Z x2 Z x2
@ @ @F
F (x; y (x) ; y 0 (x)) 0
(x) dx = (x) (x) dx: (9.48)
x1 @y 0 x1 @x @y 0
Example 9.4 Using the Euler-Lagrange equation solve the brachystochrone prob-
lem, assuming the “material point” starts from rest.
Solution: We are given the two points (x1 ; y1 ) and (x2 ; y2 ); we chose axes
through the point 1 with the y axis positive downward as shown in Figure
below. We want to …nd the curve joining the two points, down which
a bead will slide (from rest) in the least time. That means we want to
minimize time t:In other words we want to …nd the stationary value for
the integral
Z 2 Z 2
ds
I= dt = :
1 1 v
The total energy of the bead is zero since it starts from rest assuming the
zero energy level is the origin. If there is no friction, then we can write
the energy at any point below the origin described by the coordinates (x; y)
as
1 p
mv 2 mgy = 0 ) v = 2gy: (9.52)
2
Then Z Z
2 2
ds ds
I= = p : (9.53)
1 v 1 2gy
Noting that
p
ds = dx2 + dy 2
252CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
we have
r
2
Z 2
p Z 2 1 + dx Z y2
p
dx2 + dy 2 dy 1 + x02 1
I= p = p p dy = p
dy;
1 2gy 1 2gy
y1 y 2g
(9.54)
so that the stationary value of this integral is determined from the Euler-
Lagrange equation
@ @F @F
0
= 0; (9.55)
@y @x @x
where p
1 1 + x02
F (y; x (y) ; x0 (y)) = p p ; (9.56)
2g y
and
dx (y)
x0 (y) = : (9.57)
dy
Noting that
@F
=0 (9.58)
@x
and
@F 1 x0
= p p p (9.59)
@x0 2g 1 + x02 y
so that
!
@ 1 x0 x0 p
p p p =0) p p = c; (9.60)
@y 2g 1 + x02 y 1+x 02 y
where c is a constant. Solving for x0 ; we …nd
x0 p
p p = c ) x02 = c 1 + x02 y
1 + x02 y
r Z yr
dx cy cy
) x02 (1 cy) = cy ) = )x= dy: (9.61)
dy 1 cy 0 1 cy
Introducing the transformation de…ned by
1 1
cy = sin2 = (1 cos ( )) ) dy = sin cos d (9.62)
2 2 c 2 2
we have
Z r Z s
y
cy 1 sin2 2
x = dy = sin cos d
0 1 cy c 0 1 sin2 2
2 2
Z Z
1 sin 1
= 2
sin cos d = sin2 d
c 0 cos 2
2 2 c 0 2
Z
1 1 1
) x= (1 cos ( )) d ) x = ( sin ( )) : (9.63)
c 0 2 2c
9.4. APPLICATIONS: CLASSICAL MECHANICS 253
Therefore, the trajectory of the bead that takes the smallest possible time
is given by
1 1
x= ( sin ( )) ; y = (1 cos ( )) : (9.64)
2c 2c
Cycloid: Consider a circle of radius r rolling along the positive x-axis with a
constant angular velocity staring from the origin. If you mark the point on the
circle coinciding with the origin at the initial time (as shown in the Fig. 9.3)
and follow the trajectory of this point its x and y coordinates of this point are
given by
x = r( sin ( )) ; y = r (1 cos ( )) ; (9.65)
where is the angle that the circle (the point) rotated. For example the …gure
below shows this trajectory for a point on a circle of unit radius (r = 1).
For a given , the circle’s centre lies at
x = r ; y = r: (9.66)
x = r ;y = r: (9.67)
I= F (x; y; z; y 0 ; z 0 ) dx (9.68)
x1
254CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
~ The classical action ( S): the integral of the Lagrangian with respect to time
over a given period of time
Z t2
S= Ldt: (9.72)
t1
has a stationary value for the correct path of the motion. The path actu-
ally followed by a system, as speci…ed in terms of the generalized coordi-
nates qi , is that path that makes the action integral stationary:
Z t2
I= L(qi ; q_i ; t) dt = 0 (9.74)
t1
for i = 1; 2; 3:::n:This means that the Lagrangian must satisfy the set of
equations
@ @L @L
=0 (9.75)
@t @ q_i @qi
for i = 1; 2; 3:::n:
Example 9.5 Use Lagrange’s equations to …nd the equation of motion for a
particle traveling along the x-y plane under the in‡uence of a potential
energy function U (x).
Solution: The kinetic energy of a particle moving in the x-y plane can be ex-
pressed as
1 1
T = m vx2 + vy2 = m x_ 2 + y_ 2 : (9.76)
2 2
Then the Lagrangian
1
L=T U= m x_ 2 + y_ 2 U (x): (9.77)
2
256CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
Since L = L (x; y; x;
_ y;
_ t) which is a function of two variables and we must
have two Euler-Lagrange equations
@ @L @L @ @L @L
= 0; = 0: (9.78)
@t @ x_ @x @t @ y_ @y
@ @L @L @ @U (x)
=0) (mx)
_ + =0
@t @ x_ @x @t @x
@U (x)
) m•
x=
@x
@ @L @L
= 0 ) m•
y = 0 (Zero acceleration) (9.79)
@t @ y_ @y
Example 9.6 The Atwood’s Machine: A string of length, l , passes over a fric-
tionless pulley connecting two masses, m1 and m2 . Find an expression
for the acceleration of the masses in the system.
Solution: Let’s de…ne the origin of the y axis on the surface of the ground also
assume m1 > m2 ; and the length of the string is l: At a given time t let
the position of m1 and m2 be y1 and y2 ; as shown in Fig. 9.4. Then the
kinetic energy of the system can be expressed as
1 1
T = m1 y_ 12 + m2 y_ 22 (9.80)
2 2
and the gravitational potential energy
y2 = C y1 l ) y_ 2 = y_ 1 (9.85)
we have
1
L= (m1 + m2 ) y_ 12 m1 gy1 m2 g (C y1 l) : (9.86)
2
which we may put in the form
1
L= (m1 + m2 ) y_ 12 (m1 m2 ) gy1 C1 ; (9.87)
2
where we replaced, C1 = m2 g (C l) : Now using
@ @L @L
=0 (9.88)
@t @ q_i @qi
for qi = y1
@L @L
= (m1 m2 ) g; = (m1 + m2 ) y_ 1 (9.89)
@y1 @ y_ 1
we …nd
@ m1 m2
[(m1 + m2 ) y_ 1 ] = (m1 m2 ) g ) a1 = y•1 = g (9.90)
@t m1 + m2
258CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
Recalling that
y_ 2 = y_ 1 (9.91)
the acceleration of the second mass becomes
m1 m2
a1 = g: (9.92)
m1 + m2
The minus sign indicates the …rst mass is accelerating in the negative y-
direction.
Example 9.7 Central Forces: Describe the properties of the motion of a mass
m moving under the in‡uence of a central force (that is, a force acting
only along the radial direction) given by
F~ = f (r)^
r (9.93)
for some function f (r). Assume that the motion is con…ned to a plane.
Solution: The kinetic energy
1
T =mv 2 : (9.94)
2
Using polar coordinates the magnitude of the velocity can be expressed as
2
v = r_ 2 + r2 _ (9.95)
F~ = r U (r) (9.97)
where U (r) is the potential energy. Since the force is a central force it is
directed along the radial direction and it depends on r only. Therefore the
potential energy can be expressed as
Z
U (r) = f (r) dr: (9.98)
@ @L @L
=0 (9.100)
@t @ q_i @qi
9.5. PHYSICAL APPLICATION OF THE EULER-LAGRANGE EQUATION259
we have
@ @L @L @ @L @L
= 0; =0 (9.101)
@t @_ @ @t @ r_ @r
so that using
@L @L @L 2 @L
= 0; = mr2 _ ; = mr _ + f (r); = mr_ (9.102)
@ @_ @r @ r_
we …nd
@ @L @L
=0) = const ) mr2 _ = cont ) I! = cons:
@t @_ @_
(Conservation of Ang. Mom.)
@ @L @L 2
r = mr _ + f (r)
= 0 ) m• (9.103)
@t @ r_ @r
260CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS
Chapter 10
Introduction to the
Eigenvalue Problem
2 2 3 2 1 2 4 2
2A =
2 2 1 2 0 2 5 2
4 6 2 8
2A = (10.2)
4 2 0 10
261
262 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM
Matrix Multiplication: two matrices can be multiplied if and only if the num-
ber of columns of the …rst matrix is equal to the number of rows of the second
matrix. If matrices have the same dimension, then they can be multiplied. From
the above matrices we can make the multiplications:
0 1
2 4
2 3 1 @ (ab)11 (ab)12
AB = 1 1 A= (10.3)
2 1 0 (ab)21 (ab)22
3 1
0 10 1 0 1
2 1 3 2 0 1 (cd)11 (cd)12 (cd)13
CD = @ 4 1 2 A@ 1 1 2 A = @ (cd)21 (cd)22 (cd)23 A
1 0 1 3 1 0 (cd)31 (cd)32 (cd)33
(10.4)
but we can not make the matrix multiplications BC or BD
The element in row i and column j of the product matrix AB is equal to
row i of A times column j of B. In index notation
n
X
(ab)ij = aik bkj (10.5)
k=1
1 1
A A = AA =I (10.14)
We can determine the inverse of an invertible matrix (det jAj =
6 0) using
row reduction or the adjoint matrix.
a. Row reduction in this approach for the matrix, for example,
2 3
a11 a12 a13
A = 4 a21 a22 a23 5 (10.15)
a31 a32 a33
we start from 2 3
a11 a12 a13 1 0 0
4 a21 a22 a23 0 1 0 5 (10.16)
a31 a32 a33 0 0 1
and do elementary row operation until we end up with
2 3
1 0 0 b11 b12 b13
4 0 1 0 b21 b22 b23 5 (10.17)
0 0 1 b31 b32 b33
so that the inverse of the Matrix A is given by
0 1
b11 b12 b13
A 1 = @ b21 b22 b23 A : (10.18)
b31 b32 b33
b. Using the adjoint matrix: Using the adjoint matrix the inverse can be ex-
pressed as
T
[cof (A)]
A 1= (10.19)
det jAj
Example 10.1 Find the inverse of the matrix
0 1
1 2 3
A=@ 2 0 4 A (10.20)
1 1 1
using
264 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM
b. Find the cofactor Cij of the element Aij in row i and column j which
is equal to (-1) i+j times the value of the determinant remaining when
we cross o¤ row i and column j. After you obtained all elements of
the cofactor matrix write the cofactor matrix C, transpose and divide
it with the determinant of the matrix A. The resulting matrix is A 1 :
1 1 T
A = C : (10.24)
jAj
Ans: 0 1
4 5 8
1@
A 1
= 2 2 2 A: (10.25)
2
2 3 4
In Cartesian coordinates
0 1 0 10 1
x0 M11 M12 M13 x
@ y 0 A = @ M21 M22 M23 A @ y A : (10.29)
z0 M31 M32 M33 z
Example 10.2 Find the eigenvalues and the corresponding eigenvectors of the
matrix
0 1
0 1 0
M =@ 1 0 0 A (10.36)
0 0 0
10.3. EIGENVALUES AND EIGENVECTORS 267
Similarly for 2 =1
0 10 1
1 1 0 x2
@ 1 1 0 A @ y2 A = 0 ) x2 + y2 = 0; x2 y2 = 0; z2 = 0
0 0 1 z2
) x2 = y2 ; z2 = 0 (10.45)
0 1
1
j 2i = x2 @ 1 A ; (10.46)
0
0 1
1
h 2 j 2i = 1 ) x22 1 1 0 @ 1 A = 1 ) x2 = p1 (10.47)
0 2
1 0
1
1
) j 2i = p @ 1 A (10.48)
2 0
and for 3 = 1
0 10 1
1 1 0 x3
@ 1 1 0 A @ y3 A = 0 ) x3 + y3 = 0; x3 + y3 = 0; z3 = 0
0 0 1 z3
) x3 = y3 ; z3 = 0 (10.49)
1 0 0 1
1 1
1 @
) j 3 i = y3 @ 1 A ) j 3i = p 1 A: (10.50)
0 2 0
Using Mathematica:
10.3. EIGENVALUES AND EIGENVECTORS 269
The masses can slide on a horizontal, frictionless surface. The springs are
at their unstretched/uncompressed lengths when the masses are at their equi-
librium positions. At t = 0, the masses are displaced from their equilibrium
positions by the amounts x10 and x20 and released from rest, as shown in the
…gure above. Completely describe the resulting motion.
L (t; x1 ; x2 ; x_ 1 ; x_ 2 ) = T U
1 1 2 1 2 1 2
) L (t; x1 ; x2 ; x_ 1 ; x_ 2 ) = m x_ 21 + x_ 22 kx kx k (x2 x1 )
2 2 1 2 2 2
(10.59)
which leads to
@L @L
= kx1 + k (x2 x1 ) ; = kx2 k (x2 x1 ) ;
@x1 @x2
@ @L @ @L
= m•
x1 ; = m•
x2 : (10.60)
@t @ x_ 1 @t @ x_ 2
@ @L @L
= 0 for i = 1; 2 (10.61)
@t @ x_ i @xi
we …nd
m•
x1 [ kx1 + k (x2 x1 )] = 0 ) m•
x1 = 2kx1 + kx2 (10.62)
and
m•
x2 [ kx2 k (x2 x1 )] = 0 ) m•
x2 = 2kx2 + kx1 : (10.63)
If we introduce a constant r
k
; != (10.64)
m
then the above two equations can be put in the form
x
•1 = 2! 2 x1 + ! 2 x2 (10.65)
272 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM
and
• 2 = ! 2 x1
x 2! 2 x2 : (10.66)
These two equations describe the equations of motion for the two masses.
As you can see the two equations are coupled second order di¤erential
equations.
x
•1 2! 2 !2 x1
= (10.67)
x
•2 !2 2! 2 x2
or ::
~r = M~r (10.68)
where
~r = x1 e^1 + x2 e^2 ; (10.69)
and
2! 2 !2
M= : (10.70)
!2 2! 2
Note that e^1 and e^2 are column matrices given by
1 0
e^1 = ; e^2 = : (10.71)
0 1
Suppose if one solves the eigen value equation for the matrix M
Mj i= j i; (10.72)
where
a11 a11
j 1i = ;j 2i = (10.74)
a21 a22
are the normalized eigen vectors. Noting that
1 1
C C = CC =I
10.4. PHYSICAL APPLICATIONS 273
x1 y1 :: x
•1 y•1
1 1 1 1
C ~r = C = )C ~r = C = ;
x2 y2 x
•2 y•2
(10.77)
and noting that
1
C MC = D
is a similarity transformation, we …nd
y•1 1 0 y1
= :
y•2 0 2 y2
~ = R;
MR ~ (10.78)
Mj i= j i; (10.79)
2! 2 !2 2 2
det = 0 ) 2! 2 + !2 =0
!2 2! 2
) 2! 2 + !2 2! 2 + + !2 = 0 ) 1 = !2 ; 2 = 3! 2 :
(10.80)
2! 2 1 !2 X1
=0 (10.81)
!2 2! 2 1 X2
and
2! 2 2 !2 X1
= 0: (10.82)
!2 2! 2 2 X2
274 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM
~ 1 = X1 1 ~ 2 = X1 1
R ;R : (10.85)
1 1
Upon normalizing these eigen vectors, we …nd
^ 1 = p1
R
1 ^ 2 = p1
;R
1
; (10.86)
2 1 2 1
or using bra-ket notation
1 1 1 1
j 1i =p ;j 2i =p : (10.87)
2 1 2 1
You can use Mathematica to check you results. Here is an example
how to …nd eigenvalues and unnormalized eigenvectors determined
using Mathematica:
10.4. PHYSICAL APPLICATIONS 275
N.B. You must use the method we studied to …nd the inverse of the matrix
C. But, here I am going to use mathematica to …nd the inverse
We note that
! !
p1 p1 p1 p1 1 0
1 2 2 2 2 1
CC = p1 p1 p1 p1
) CC =
2 2 2 2
0 1
1
) C C=I (10.89)
Recalling that ::
~r = M~r (10.90)
can be written as
y•1 1 0 y1
= :
y•2 0 2 y2
where
x1 y1 :: x
•1 y•1
1 1 1 1
C ~r = C = )C ~r = C = ;
x2 y2 x
•2 y•2
(10.91)
we can express the equations of motion as
y•1 !2 0 y1
= ) y•1 = ! 2 y1 ; y•2 = 3! 2 y2
y•2 0 3! 2 y2
(10.92)
276 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM
)
0
= p!B ) B = 0; D = 0 (10.97)
0 3!D
Then using the results above, one can write
! !
x1 (t) p1 p1 p1 (x10 + x20 ) cos (!t)
= 2 2 2 p (10.98)
x2 (t) p1 p1 p1 (x10 x20 ) cos 3!t
2 2 2
or
1 1 p
x1 (t) = (x10 + x20 ) cos (!t) + (x10 x20 ) cos 3!t
2 2
1 1 p
x2 (t) = (x10 + x20 ) cos (!t) (x10 x20 ) cos 3!t (10.99)
:
2 2
Step 4: The Propagator Matrix, U : If upon simplifying the above expressions,
we …nd
1h p i 1h p i
x1 (t) = cos (!t) + cos 3!t x10 + cos (!t) cos 3!t x20
2 2
1h p i 1h p i
x2 (t) = cos (!t) cos 3!t x10 + cos (!t) + cos 3!t x20
2 2
(10.100)
10.4. PHYSICAL APPLICATIONS 277
Step 5: The Normal Modes of Vibration: Suppose the initial state of the two
masses is described by the …rst eigenvector. That means
x10 1 1
~r (0) = =p (10.104)
x20 2 1
then
~r = U (t)~r (0) (10.105)
gives
p p
x1 (t) 1 cos (!t) + cos p3!t cos (!t) cos p3!t
= p
x2 (t) 2 2 cos (!t) cos 3!t cos (!t) + cos 3!t
1
(10.106)
1
which leads to
x10 1 1
~r (0) = =p : (10.108)
x20 2 1
278 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM
then we have
p p
x1 (t) 1 cos (!t) + cos p3!t cos (!t) cos p3!t
= p
x2 (t) 2 2 cos (!t) cos 3!t cos (!t) + cos 3!t
1
(10.109)
1
which gives
p
x1 (t) 1 cos p3!t
=p ) x1 (t) = x2 (t) : (10.110)
x2 (t) 2 cos 3!t
p
The two masses oscillate with a frequency 3! out of phase by .
The two modes of vibrations we saw above are called Normal Modes of vibration.
Chapter 11
Special functions
where
0! = 1: (11.2)
The integral form of the Factorial function: Consider the integral function
given by Z 1
F (p) = e x dx: (11.3)
0
For any real number, > 0, the value of this integral is
Z 1 1
e x 1
F (p) = e x dx = = : (11.4)
0 0
Now let’s di¤erentiate this integral with respect to as many as we can, say n
n
time (i.e. @@ n ). For the …rst derivative, n = 1
Z 1 Z 1
1 0!
e x dx = ) x0 e x dx = 1
0
Z 1 Z 01
@ 1 @ @ 1
) e x dx = ) e x dx =
@ 0 0 @ @
Z 1
1
xe x dx = 2
(11.5)
0
279
280 CHAPTER 11. SPECIAL FUNCTIONS
Therefore it is not di¢ cult to see for the nth derivative, we …nd
Z 1
n!
xn e x dx = n : (11.9)
0
which is the integral form of the Factorial function which is valid for an integer,
n 0.
The Gamma function: The Gamma function is de…ned by the integral Func-
tion given by Z 1
(p) = xp 1
e x
dx: (11.11)
0
where p > 0 is any positive real number. For p = n + 1; with n 0 (positive
integer or zero), we …nd
Z 1
(n + 1) = xn e x dx; (11.12)
0
which is the Factorial function. Therefore, the factorial function in terms of the
Gamma function can be expressed as
Z 1
n! = (n + 1) = xn e x dx: (11.13)
0
If we denote
u = xp ; dv = e x
dx ) du = pxp 1
;v = e x
; (11.16)
we …nd Z 1
x p 1
(p + 1) = e x 0 pxp 1
e x
dx: (11.18)
0
In the above expression the …rst term is zero
Z 1
(p + 1) = p xp 1
e x
dx: (11.19)
0
We know that Z 1
(p) = xp 1
e x
dx (11.20)
0
hence
(p + 1) = p (p): (11.21)
u2 = x ) 2udu = dx (11.25)
we can write
Z 1 Z 1
1 1 u2 u2
= e 2udu = 2 e du . (11.26)
2 0 u 0
Here the set the upper limit of integration for be =2 since both u and v
are positive and we must integrate only in the …rst quadrant (the region shown
in green). Therefore, integrating with respect to r and leads to
1
r2 p
2 1 e 1
=4 = ) = : (11.30)
2 2 2 2
0
~ Replacing x = y= (1 + y) ; we have
dy ydy dy
dx = 2 = 2
1+y (1 + y) (1 + y)
and
y y
x = 0 ) y = 0; x = 1 ) = 1 , lim =1 (11.36)
1+y y!1 1+y
so that
Z 1 p 1 q 1
y y dy
B(p; q) = 1 2
0 1+y 1+y (1 + y)
Z 1 p 1 q 1
y 1 dy
= 2
0 1+y 1+y (1 + y)
Z 1
yp 1
dy
) B(p; q) = p+q (11.37)
0 (1 + y)
Example 11.2 Prove that the Gamma and the Beta Functions are related by
(p) (q)
B(p; q) = : (11.38)
(p + q)
284 CHAPTER 11. SPECIAL FUNCTIONS
we …nd
Z 1 Z 1
2q 1 u2 v2
(q) = 2 u e du , (p) = 2 v 2p 1
e dv . (11.41)
0 0
we …nd
Z 1 Z =2
2q 1 2p 1 r2
(p) (q) = 4 (r cos ) (r sin ) e rdrd . (11.44)
0 0
and applying
Z =2
B(p; q) = 2 sin2p 1
( ) cos2q 1
( )d (11.47)
0
we note that the second integral can be expressed in terms of the Beta
function as Z =2
2q 1 2p 1 B(q; p)
(sin ) (cos ) d = : (11.48)
0 2
Therefore
(p) (q)
B(p; q) = . (11.49)
(p + q)
This equation relates the beta and gamma functions.
11.3. STIRLING’S FORMULA 285
Next we want to …nd an approximate formula when p is very large. This ap-
proximate formula is known as Stirling’s formula and is given by
p
p! = (p + 1) = pp e p 2 p; (11.52)
ln(p!) = p ln p p: (11.53)
Noting that
p p p p p
(p + y p) = eln[(p+y p) ] = ep ln(p+y p) ; (11.57)
We recall that the Taylor series expansion for f (y) about y = 0 is given
by
1 df (y) 1 d2 f (y)
f (y) = f (0) + y+ y 2 + ::: (11.59)
1! dy y=0 2! dy 2 y=0
p
so that for f (y) = ln p + y p ; using
p p
df (y) p p
f (0) = ln (p) ; = p = ;
dy y=0 p + y p y=0 p
2 p p p
d f (y) d p p p 1
= p = p 2 = ; (11.60)
dy 2 y=0 dy p + y p y=0 p + y p y=0 p
Noting that Z 1 p
y2
e 2 dy = 2 (11.63)
1
we may write
p Z 1 Z p
p Z 1
y2 y2 y2
2 = e 2 dy = e 2 dy + p
e 2 dy
1 1 p
Z 1 Z 1 Z p
p p Z p
p
y2 y2 y2 y2
) p
e 2 dy = e 2 dy e 2 dy = 2 e 2 dy;
p 1 1 1
(11.64)
we can write
Z 1 p Z p
p
p y2 p y2
p! ' pep ln(p) p
p
e 2 dy = 2p ep ln(p) p
pep ln(p) p
e 2 dy:
p 1
(11.65)
11.3. STIRLING’S FORMULA 287
where we dropped the …rst two terms as compared to the last two terms.
Example 11.4 Consider a classroom full of gas molecules. There are approxi-
mately N = 5000NA = 3 1027 molecules in the room. From the Binomial
Theorem, it can be shown that the probability for n of the molecules to
be in the front half and n0 = N n molecules to be in the back half of
the room is given by
N 0 N!
P (n) = pn q n = pn q N n
; (11.69)
n n! (N n)!
where p is the probability that a molecule will be found in the front half
of the room, and q is the probability that it will be found in the back half.
From the symmetry of the problem, we must have
1 1
p= ;q = 1 p= =p (11.70)
2 2
On the average, we would expect to …nd half of the molecules in the front
half of the room and the other half of the molecules in the back half of the
room. Find the probability that 0:1% of the molecules in the room have
shifted from the front to the back half of the room. That is, …nd the value
of P (n) = P (0:499N ).
Solution: Since q = p, we can write
N! N!
P (n) = pn pN n
= pN : (11.71)
n! (N n)! n! (N n)!
On the average there are nave = 0:5N of molecules in the front half of the
room and n0ave = N nave = 0:5N in the back half of the room. Here we
want to …nd the probability that 0:1% of the molecules shifted to the back
half of the room. In other words we want to determine the probability
that the number of molecules in the front half (nave ) is reduced by 0:1%.
Which means we want to …nd P (n) for
which is given by
N!
P (n) = pN : (11.73)
n! (N n)!
Obviously, both N and n are very large number and we can make Stirling’s
approximation
ln n! = n ln n n (11.74)
for the factorial. Thus
N!
ln [P (n)] = ln pN = ln N ! ln n! ln(N n)!+ln pN (11.75)
n! (N n)!
can be approximated as
we …nd
and
P (n) = exp[ 6 1021 ]: (11.78)
There are also other related integrals which sometimes referred as error func-
tions. These includes
11.4. THE ERROR FUNCTION 289
Find an expression in terms of the error function for the probability that
the number of elements of a given system satisfying the criterion, x, will
be in the range
x n x x+n
for some real value of n (usually integral).
1 (x x)2
(x)dx = p e 2 2 dx (11.84)
2
then the probability that the number of elements in the range x n
x x + n satisfying the criterion will be
Z x+n Z x+n
1 (x x)2
Pn (x) = (x)dx = p e 2 2 dx: (11.85)
x n x n 2
we can write
Z n
p
p Z n
p
2 1 y2 1 2
y2
Pn (x) = p e 2 dy = p e dy: (11.88)
pn
2
2 pn
2
we …nd Z n
p
2 2
y2
Pn (x) = p e dy: (11.91)
0
Recalling the de…nition for the error function
Z x
2 2
erf(x) = p e t dt (11.92)
0
we …nd that
n
p
Pn (x) = erf
: (11.93)
2
You can get the values of the error function for di¤erent values of n using,
for example, Mathematica. You will …nd the following results
n Pn (x)
1 68:26%
2 95:44%
3 99:74%
292 CHAPTER 11. SPECIAL FUNCTIONS
Figure 11.2: A simple pendulum. At the initial time, t = 0; the mass m was
displaced by an angle, ; from the vertical.
M EI = M E (11.95)
we can write
Z t1=2 Z
d
T =2 dt = 2 q (11.101)
0 2g
l (cos cos )
Using the
where Z =2
d'
K( )= p (11.109)
2
0 1 sin2 '
is the elliptic integral of the …rst kind.
where
@~
r
@r
r^ = sin ( ) cos (') x
^ + sin ( ) sin (') y^ + cos ( ) z^ = @~
r
(11.111)
@r
is the unit vector along the radial direction. In spherical coordinates we recall
the gradient and the Laplacian for a scalar function, f (r; ; '), are given by
@f ^ 1 @f 1 @f
rf = r^ + +'
^ ; (11.112)
@r r@ r sin @'
and
1 @ @f 1 @ @f 1 @2f
r2 f = r2 + sin ( ) + 2 ; (11.113)
r2 @r @r r2 sin ( ) @ @ r2 sin ( ) @'2
respectively. On the other hand for a vector …eld,
V r + V (r; '; )^ + V' (r; '; )^
~ (r; '; ) = Vr (r; '; )^ '; (11.114)
instead of the gradient, most often, we are interested in the divergence of the
vector …eld. In spherical coordinates the divergence of a vector …eld is given by
~ = 1 @ r2 Vr +
r V
1 @
(sin ( ) V ) +
1 @V'
: (11.115)
r2 @r r sin ( ) @ r sin ( ) @'
Suppose we have some physical quantity described by some scalar function that
depend on the radial distance between two points. A good example of such phys-
ical quantity is electrical and gravitational potentials. Let’s say this function in
296 CHAPTER 11. SPECIAL FUNCTIONS
~ =r 1 @ 1 r^
V = r^ = ; for r > 0: (11.117)
r @r r r2
The divergence of this vector …eld becomes
~ =r r^ 1 @ 1 0; for r > 0
r V = r2 = : (11.118)
r2 r2 @r r2 1 for r = 0;
There follows that
1 1 r^ 0; for r > 0
r2 =r r =r = (11.119)
r r r2 1 for r = 0;
which leads to
ZZZ ZZZ
~ d = r^
r V r d = 0; for r > 0 (11.120)
volum e volum e r2
Z1Z Z2
r^
) r d = 0; for r > 0 (11.121)
r2
0 0 0
we …nd
ZZ
~ (x; y; z) d~a
V
Surface
enclosing
Z Z2 Z Z2
r^ 2
= r^r sin ( ) cos (') d d' = sin ( ) cos (') d d' = 4
r2
0 0 0 0
(11.123)
This none zero value must be a result when r = 0 is included. Therefore, one
can write
Z1Z Z2
r^ 0 for r > 0
r d = (11.124)
r2 4 for r = 0
0 0 0
is known as the Dirac delta function. Then for the Dirac delta function
Z1Z Z2
(~r) d = 1;
0 0 0
Z1Z Z2 Z1Z Z2
) f (0) (~r) d = f (0) ) f (~r) (~r) d = f (0) ; (11.126)
0 0 0 0 0 0
In the usual sense of functions the Dirac delta function does not exist. But
there are various forms of sequence functions in the limiting case display the
properties of the Dirac delta function. These functions include:
See Fig. 1.
2. Lorentz function
n 1
n (x) = ;
1 + n2 x2
See Fig. 2.
298 CHAPTER 11. SPECIAL FUNCTIONS
See Fig. ??.The general form of the Dirac delta function for, x = a
0 for x 6= a
(x a) =
1 for x = a
and
Z1 Z1
(x a) dx = 1; f (x) (x a) dx = f (a) : (11.130)
1 1
11.6. THE DIRAC DELTA FUNCTION 299
For 3D case
Z1Z Z2
(~r) r2 sin ( ) drd d' = 1
0 0 0
Z1Z Z2
f (~r) (~r) r2 sin ( ) drd d' = f (0)
0 0 0
Z1Z Z2
f (~r ~r0 ) (~r ~r0 ) r2 sin ( ) drd d' = f (~r0 ) (11.131)
0 0 0
Example 11.6 Form introductory physics, the electric potential, V (~r) ; due to
a point charge located at the origin (0; 0; 0) (i.e. r = 0) at a point in space
described by the position vector, ~r; is given by
1 q
V (~r) = :
4 0 r
Show that the volume charge density, (~r) ; for this point charge can be
expressed in terms of the Dirac delta function
dq
(~r) = = q (~r) = q (x) (y) (z) ; (11.132)
d
where dq an in…nitessimal charge in an in…nitesimal volume d :
Solution: The electric potential, dV (~r) of an in…nitessimal charge dq 0 in a
volume d 0 as shown in Fig. can be expressed as
ZZZ
1 dq 0 1 (~r0 ) d 0 1 (~r0 ) d 0
dV (~r) = 0j
= 0j
) V (~r) = ;
4 0 j~
r ~
r 4 0 j~
r ~
r 4 0 j~r ~r0 j
V
(11.133)
Using spherical coordinates, we can write
Z1Z Z2
1 (~r0 ) r02 sin ( ) dr0 d 0 d'0
V (~r) = (11.134)
4 0 j~r ~r0 j
0 0 0
Z1Z Z2
0
f (~r0 ) (~r0 ~r0 ) r02 sin dr0 d 0 d'0 = f (~r0 ) (11.136)
0 0 0
1 (~r0 )
f (~r0 ) = ; (~r0 ~r0 ) =
j~r ~r0 j q
1 1
) f (~r0 ) = = ) ~r0 = 0 (11.137)
j~r ~r0 j r
which leads to
(~r0 )
= (~r0 ) ) (~r0 ) = q (~r0 ) : (11.138)
q
Example 11.7 The volume charge density, (~r) ; of a point charge, q, placed
at a point on the z-axis, ~r0 = a^
z ; can be expressed as
where (~r) is the Dirac Delta function. Show that the electric potential,
V (~r) ; due to this point charge is given by
1 q q 1
V (~r) = = q : (11.140)
4 0 j~r ~r0 j 4 0 2
x2 + y 2 + (z a)
Solution: Using the given charge density and the expression for the potential,
one can write
ZZZ ZZZ
q (~r0 ~r0 ) d 0 q (~r0 ~r0 ) d 0
V (~r) = 0
= ; (11.142)
4 0 j~r ~r j 4 0 j~r ~r0 j
V V
~r0 = x0 x
^ + y 0 y^ + z 0 z^; ~r = x^x + y y^ + z z^; ~r0 = a^
z
q
0 2 2 2
) j~r ~r j = (x x0 ) + (y y 0 ) + (z z 0 ) (11.143)
11.6. THE DIRAC DELTA FUNCTION 301
and
so that
Z1 Z1 Z1
q (x0 ) (y 0 ) (z 0 a) dx0 dy 0 dz 0
V (~r) = q
4 0 2 2 2
1 1 1 (x x0 ) + (y y 0 ) + (z z0)
Z1 Z1 Z1
q 0 0 0 0
= (z a) dz (y ) dy f (x0 ; y 0 ; z 0 ) (x(11.145)
0
) dx0 ;
4 0
1 1 1
where
1
f (x0 ; y 0 ; z 0 ) = q : (11.146)
2 2 2
(x x0 ) + (y y 0 ) + (z z0)
0 0 1
= f (0; y ; z ) = q : (11.148)
2 2
x2 + (y y0 ) + (z z0)
Once again using the property of the Dirac delta function, we have
Z1 Z1
0 0 0 0
f (0; y ; z ) (y ) dy = f (0; y 0 ; z 0 ) (y 0 0) dy 0
1 1
1
= f (0; 0; z 0 ) = q : (11.150)
2
x2 + y 2 + (z z0)
302 CHAPTER 11. SPECIAL FUNCTIONS
One last time using the Dirac delta function property, we …nd for the
potential
Z1
q q
V (~r) = f (0; 0; z 0 ) (z 0 a) dz 0 = f (0; 0; a)
4 0 4 0
1
q 1
) V (~r) = q : (11.152)
4 0 2
x2 + y 2 + (z a)
Chapter 12
d2 y (x) dy (x)
+ f (x) + g(x)y(x) = 0: (12.1)
dx2 dx
Such kind of di¤erential equations or even those with constant coe¢ cients can
be solved using series substitution method. The method involves a simple pro-
cedure but a little too much algebra. We assume the solution to the di¤erential
equation can be expressed as a power series
1
X
y (x) = an xn = a0 + a1 x + a2 x2 + a3 x3 + :::an xn + ::: (12.2)
n=0
We then substitute this series into the di¤erential equation and determine the
expansion coe¢ cients. We will demonstrate the application of this method using
the following example which we already determined the solution using a di¤erent
method last semester.
303
304CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
Solution: Using Newton’s second law the equation of motion for the mass, m;
can be written as
d2 x (t)
Fnet = ma ) m = kx; (12.3)
dt2
which we may rewrite as
d2 x (t)
+ ! 2 x = 0; (12.4)
dt2
where r
k
;!= (12.5)
m
is the angular frequency. Let’s assume that the solution to this di¤erential
equation is given by the power series
1
X
x (t) = an tn = a0 + a1 t + a2 t2 + a3 t3 + :::an tn + ::: (12.6)
n=0
so that
1
X 1
dx (t) d (tn ) X
= an = nan tn 1
= a1 + 2a2 t + 3a3 t2 +
dt n=0
dt n=0
X1
:::nan tn 1
+ ::: = (n + 1) an+1 tn (12.7)
n=0
and
1 1
d2 x (t) X d2 (tn ) X
= an = n (n 1) an tn 2
dt2 n=0
dt2 n=0
= 2a2 + 3 2a3 t + 4 3a4 t2 + :::n (n 1) an tn 2
+ :::
X1
= (n + 2) (n + 1) an+2 tn (12.8)
n=0
12.1. POWER SERIES SUBSTITUTION 305
1
X 1
X
(n + 2) (n + 1) an+2 tn + ! 2 an tn = 0
n=0 n=0
1
X
) (n + 2) (n + 1) an+2 + ! 2 an tn = 0: (12.9)
n=0
an
(n + 2) (n + 1) an+2 + ! 2 an = 0 ) an+2 = !2 :
(n + 2) (n + 1)
Now let’s examine the …rst few terms for this recursion relation. First we
consider when n is even
a0 2
n = 0 ) a2 = !
2 1
1
a0 ( 1) 2 1
) n = 0 ) a2 1 = !
2!
2 2
a2 2 a0 ( 1) ! 2 2
a0 ( 1) ! 2 2
n = 2 ) a4 = ! = =
4 3 4 3 2 1 4!
a0 2
) n = 2 ) a2 2 = ( 1) ! 2 2
(2 2)!
3 3
a4 ! 2 a0 ( 1) ! 2 3 a0 ( 1) ! 2 3
n = 4 ) a6 = = =
6 5 6 5 4 3 2 1 6!
3 2 3
a0 ( 1) !
) n = 4 ) a2 3 = : (12.10)
(2 3)!
From the results we see above it is not hard to come up with the relation
m
a0 ( 1) ! 2m
a2m = (12.11)
(2m)!
that generates the values for the even term coe¢ cients in the series for
m = 0; 1; 2; ::.
Next we shall consider the odd terms for the recursion relation
an
an+2 = !2 : (12.12)
(n + 2) (n + 1)
306CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
We recall that the Taylor series expansions for sin (x) and cos (x) are
1
X k X1 n
( 1) x2k+1 ( 1) x2n
sin (x) = ; cos (x) = (12.17)
(2k + 1)! n=0
(2n)!
k=0
so that the solution to the di¤erential equation can be put in the form
a1
x (t) = a0 cos (!t) + sin (!t) (12.18)
!
or
x (t) = A cos (!t) + B sin (!t) ; (12.19)
12.2. ORTHOGONAL VECTORS AND DIRAC NOTATION 307
and
dx (t)
= a0 sin (!t) a1 ! 2 cos (!t) ;
dt
dx (0)
= 0 ) a1 ! 2 = 0 ) a1 = 0: (12.21)
dt
Therefore, the equation for the position of the mass as a function of time,
x(t) is given by
x (t) = xmax cos (!t) ; (12.22)
where r
k
!= ; (12.23)
m
is the angular frequency.
the two vectors are said to be orthogonal. If these vectors are complex and
orthogonal, we must write
3
X
~
A ~ =
B Ai Bi = 0; (12.25)
i=1
Orthonormal Sets of Functions: two di¤erent functions A(x) and B(x) are
said to be orthogonal for all x (a; b) when
Z b
A (x)B(x)dx = 0; (12.28)
a
where A (x) is the complex conjugate of the function A(x): A set of functions
For m = n,
hAn (x) jAn (x)i = Cn (12.33)
so that the for the set of functions de…ned by
An (x)
Fn (x) = p (12.34)
Cn
one can write
Z b
hFn (x) jFm (x)i = Fn (x) Fm (x) dx = nm : (12.35)
a
1 eimx e imx
e inx
einx
Fm (x) = p sin (mx) = p ) Fn (x) = p
2i 2i
(12.36)
12.2. ORTHOGONAL VECTORS AND DIRAC NOTATION 309
so that Z b
hFn (x) jFm (x)i = Fn (x) Fm (x) dx (12.37)
a
becomes
Z inx
e einx eimx e imx
hFn (x) jFm (x)i = p p dx
2i 2i
Z
1
= ei(m n)x e i(m+n)x
ei(m+n)x + e i(m n)x
dx
4
1 ei(m n)x e i(m+n)x ei(m+n)x e i(m n)x
= +
4 i m n m+n m+n m n
1 4 sin ((m n) ) 4 sin ((m + n) )
) hFn (x) jFm (x)i = ;
4 m n m+n
(12.38)
sin [(m n) ]
hFn (x) jFm (x)i = = nm ;
(m n)
that give
1
sin ( ) sin ( ) =[cos ( ) cos ( + )] :
2
In view of these relation, one can write the integral as
Z
hFn (x) jFm (x)i = sin (nx) sin (mx) dx
Z
1
= [cos ((n m) x) cos ((n + m) x)] dx (12.39)
2
1 sin ((n m) x) sin ((n + m) x)
= (12.40)
2 n m n+m
1 sin ((n m) ) sin ((n + m) )
= (12.41)
n m n+m
1 sin ((n m) )
hFn (x) jFm (x)i = = nm
n m
310CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
Next we shall study when such kind of set of functions form a complete set for
all x (a; b) in the function space. But before we do that, we want to review
complete set of vectors in vector space.
Complete sets of vectors in vector space: Consider a three dimensional real
space R3 in Cartesian coordinate system. We recall the unit vectors x ^; y^; and z^
or x^1 ; x
^2 ; and x
^3 form an orthonormal set of vectors in 3-D vector space since
x
^n x
^m = nm (12.43)
3
X
hxj jri = ri hxj jxi i : (12.47)
i=1
3
X
hxj jri = ri ji ) rj = hxj jri : (12.48)
i=1
If one of the vectors, x^i ; is missing, then we can not express any vector ~r R3
in terms of the remaining two vectors. But if we have all these three vectors
any vector ~r R3 can be expressed in terms of these vectors. Then we say that
the set of vectors f^
xi g = f^ x1 ; x ^3 g is a complete orthonormal set in R3 :
^2 ; x
We now consider the function space where we have the in…nite set of or-
thonormal functions fFn (x)g = fF0 (x) ; F1 (x) ; F2 (x) ; F3 (x) ; :::g for all x
(a; b) : These set of functions form a complete set if a function, g (x) ; de…ned
12.3. COMPLETE SETS OF FUNCTIONS 311
where the expansion coe¢ cient, cn ; is determined using the the property for
orthonormal set of functions
Z b
hFn (x) jFm (x)i = Fn (x) Fm (x) dx = nm : (12.50)
a
Multiplying both sides by the bra vector hFm (x)j from the left, we have
1
X
hFm (x) jg (x)i = cn hFm (x) jFn (x)i ; (12.51)
n=0
Example 12.3 Any periodic function g(x) de…ned in the interval ( ; ) can
be expressed in terms of the set of orthonormal function
1 1 1 1
Fn (x) = p sin (nx) = p sin (x) ; p sin (2x) ; p sin (3x) ; :::
(12.53)
as
1
X 1
X c
jg (x)i = cn jFn (x)i ) g (x) = pn sin (nx) ; (12.54)
n=1 n=1
where
Z Z
1
) cn = Fn (x) g (x) dx ) cn = p sin (nx) g (x) dx: (12.55)
312CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
1
4 X sin [(2n + 1) x]
f (x) = : (12.57)
n=0
2n + 1
This series expansion for di¤erent values of terms in the series is shown in
Fig. 12.1. It shows the more terms in the series are considered, the more the
series expansion be close to the actual function.
12.4. THE LEGENDRE DIFFERENTIAL EQUATION 313
1 x2 y 00 2xy 0 + l (l + 1) y = 0 (12.59)
where
a2 (x) = 1 x2 ; a1 (x) = 2x; a0 = l (l + 1)
and l = 0; 1; 2; 3::: . The legendre di¤erential equation is derived from the
Laplace’s equation in spherical coordinates. For now we focus on how to …nd
the solution to this di¤erential equation. To this end we may want to rewrite
the Legendre di¤erential equation in the form
l (l + 1) y 2xy 0 x2 y 00 + y 00 = 0:
Suppose the solution to this di¤erential equation, y(x); exists, one must be able
to express the function, y(x); as a convergent power series given by
1
X
y (x) = an xn = a0 + a1 x + a2 x2 + a3 x3 + :::an xn + :::
n=0
1
X
) l (l + 1) y = l (l + 1) an xn (12.60)
n=0
n=0
dx n=0
1
X
00
) x2 y (x) = an n (n 1) xn : (12.62)
n=0
314CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
00
Upon examining the …rst few terms in the series for y (x) ;
1
X
00
y (x) = an n (n 1) xn 2
= 0 + 0 + 2a2 + 3:2a3 x1 + ::: (12.63)
n=0
we note that
1
X
00
y (x) = am m (m 1) xm 2
; (12.64)
m=2
n=m 2 ) m = n + 2;
we have
m=2)n=m 2=0
so that one can rewrite the second derivative as
1
X
00
y (x) = (n + 2) (n + 1) an+2 xn : (12.65)
n=0
l (l + 1) y 2xy 0 x2 y 00 + y 00 = 0;
we …nd
1
X 1
X
(n + 2) (n + 1) an+2 xn + [l (l + 1) n (n 1) 2n] an xn = 0:
n=0 n=0
an+2 (n + 2) (n + 1) + an [l (l + 1) n (n + 1)] = 0
l (l + 1) n (n + 1)
) an+2 = an (12.67)
(n + 2) (n + 1)
and noting that
l (l + 1) n (n + 1) = l2 n2 + l n = (l n) (l + n + 1)
This the the recursion relation we will next use to determine each an ; when n
is even or odd.
(a) Even n = 2; 4; 6:::
(l 0) (l + 0 + 1) l (l + 1)
a2 = a0 ) a2 = a0 (12.69)
(0 + 2) (0 + 1) 2!
(l 2) (l + 2 + 1) (l 2) (l + 3)
a4 = a2 = a2
(2 + 2) (2 + 1) 4 3
l (l + 1) (l 2) (l + 3)
) a4 = a0 (12.70)
4!
(l 4) (l + 4 + 1) (l 4) (l + 5)
a6 = a4 = a4
(4 + 2) (4 + 1) 6 5
l (l + 1) (l 2) (l + 3) (l 4) (l + 5)
) a6 = a0 (12.71)
6!
(b) Odd n = 1; 3; 5:::
(l 1) (l + 1 + 1) (l 1) (l + 2)
a3 = a1 = a1 (12.72)
(1 + 2) (1 + 1) 3!
(l 3) (l + 3 + 1) (l 3) (l + 4)
a5 = a3 = a3
(3 + 2) (3 + 1) 5 4
(l 1) (l + 2) (l 3) (l + 4)
) a5 = a1 (12.73)
5!
(l 5) (l + 5 + 1) (l 5) (l + 6)
a7 = a5 = a5
(5 + 2) (5 + 1) 7 6
(l 1) (l + 2) (l 3) (l + 4) (l 5) (l + 6)
a7 = a1 : (12.74)
7!
Therefore, the function y (x) is a sum of two series
where
l (l + 1) 2 l (l + 1) (l 2) (l + 3) 4
ye (x) = a0 1 x + x
2! 4!
l (l + 1) (l 2) (l + 3) (l 4) (l + 5) 6
x + ::: (12.75)
6!
316CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
and
(l 1) (l + 2) 3 (l 1) (l + 2) (l 3) (l + 4) 5
yo (x) = a1 x x + x
3! 5!
(l 1) (l + 2) (l 3) (l + 4) (l 5) (l + 6) 7
x + ::: (12.76)
7!
Note that a0 and a1 are constants determined from the boundary conditions.
Furthermore, we do not know if this series is convergent or divergent. For this
function to be a solution to the di¤erential equation it must be a convergent
series. The next section focuses on the conditions that must be ful…lled for the
series to be a convergent series for x [ 1; 1].
(l n) (l + n + 1) an+2 (l n) (l + n + 1)
an+2 = an ) lim = lim
(n + 2) (n + 1) n!1 an n!1 (n + 2) (n + 1)
(12.78)
If we determine the interval of convergence for
an+2 2
lim x <1 (12.79)
n!1 an
it means we have determined the interval of convergence for the even series
l (l + 1) 2 l (l + 1) (l 2) (l + 3) 4
ye (x) = a0 1 x + x
2! 4!
l (l + 1) (l 2) (l + 3) (l 4) (l + 5) 6
x + ::: (12.80)
6!
would be the intervals of convergence that is common to both series. Since each
of the coe¢ cients in both the even and odd series satisfy the condition
an+2 2 (l n) (l + n + 1) 2
lim x = lim x <1
n!1 an n!1 (n + 2) (n + 1)
(l n) (l + n + 1) 2
) lim x = lim jxj < 1 (12.83)
n!1 (n + 2) (n + 1) n!1
is convergent in the interval 1 < x < 1: However, the ratio test fails at x = 1;
we need to examine the series at the boundaries x = 1: For x = 1; we …nd
for the even series
l (l + 1) l (l + 1) (l 2) (l + 3)
ye (x) = a0 1 +
2! 4!
l (l + 1) (l 2) (l + 3) (l 4) (l + 5)
+ ::: (12.85)
6!
and the odd series
(l1) (l + 2) (l 1) (l + 2) (l 3) (l + 4)
yo (x) = a1 1 +
3! 5!
(l 1) (l + 2) (l 3) (l + 4) (l 5) (l + 6)
+ ::: (12.86)
7!
where we included the factor 1 in the odd series in the constant a1 : In the
Legendre di¤erential equation,we recall that, l = 0; 1; 2; 3:::s so that if we set
these values one …nds for l = 0;
ye (x) = a0 (12.87)
and
1 2 1 2 3 4 1 2 3 4 5 6
yo (x) = a1 1 + + + + :::
3! 5! 7!
1
X
1 1 1 1 1
= a1 1 + + + :::: + + ::: = a1 (12.88)
3 5 7 2n + 1 n=0
2n + 1
which means the series is a divergent series. Since we are looking for a well
de…ned function for which the series is convergent for all x in the interval,
1 x 1; we must set, a1 = 0; so that
1
X 1
yo (x) = a1 =0 (12.90)
n=0
2n +1
For l = 1; at x = 1; we have
1 (1 + 1) 1 (1 + 1) (1 2) (1 + 3)
ye ( 1) = a0 1 +
2! 4!
1 (1 + 1) (1 2) (1 + 3) (1 4) (1 + 5)
+ :::
6!
2 1 (2) (4) 1 (2) (4) (3) (6)
= a0 1 + :::
2! 4! 6!
1 1 1 1
= a0 + + + :::
3 5 7 2n + 3
X1
1
) ye ( 1) = a0 (12.92)
n=0
2n +3
ye ( 1) = 0: (12.93)
On the other hand, for the odd series for all x (i.e. 1 x 1) at l = 1; we
…nd
(1 1) (1 + 2) (1 1) (1 + 2) (1 3) (1 + 4)
yo (x) = a1 1 +
3! 5!
(1 1) (1 + 2) (l 3) (1 + 4) (1 5) (1 + 6)
+ :::
7!
yo (x) = a1 x; (12.94)
which leads to
y (x) = yo (x) = a1 x for l = 1: (12.95)
For the same reason discussed above it can be shown that the solution to the
Legenedre di¤erential equation must then be given by
ye (x) ; l = even
y (x) = (12.96)
yo (x) ; l = odd
12.6. THE GENERATING FUNCTION FOR THE LEGENDRE POLYNOMIALS319
where
l (l + 1) 2 l (l + 1) (l 2) (l + 3) 4
ye (x) = a0 1 x + x
2! 4!
l (l + 1) (l 2) (l + 3) (l 4) (l + 5) 6
x + ::: ; (12.97)
6!
(l 1) (l + 2) 3 (l 1) (l + 2) (l 3) (l + 4) 5
yo (x) = a1 x x + x
3! 5!
(l 1) (l + 2) (l 3) (l + 4) (l 5) (l + 6) 7
x + ::: : (12.98)
7!
5 3
l = 2 ) y (x) = a0 1 3x2 ; l = 3 ) y (x) = a1 x x (12.99)
3
l = 0 ) y0 (x) = P0 (x) = 1;
l = 1 ) y1 (x) = P1 (x) = x;
1
l = 2 ) y2 (x) = P2 (x) = 3x2 1 ;
2
1
l = 3 ) y3 (x) = P3 (x) = 5x3 3x : (12.100)
2
1 1
P0 (x) = 1; P1 (x) = x; P2 (x) = 3x2 1 ; P3 (x) = 5x3 3x : (12.101)
2 2
In this section we are interested in …nding a function that generates these Poly-
nomials. To this end, we shall consider the function, which is referred as the
Generating function,
1=2
(x; h) = 1 2xh + h2 : (12.102)
l=2
" #
d2 d x h
(x; h) =
dh2 dh (1 2xh + h2 )
3=2
2
1 3 (x h)
= 3=2
+ 5=2
(1 2xh + h2 ) (1 2xh + h2 )
d2 1
(x; h) = 1 + 3x2 = 2! 3x2 1 = 2!P2 (x) ; (12.107)
dh2 h=0 2
l=3
" #
2
d3 d 1 3 (x h)
(x; h) = +
dh3 dh (1 2xh + h2 )
3=2
(1 2xh + h2 )
5=2
3
3 (x h) 3 2 (x h) 3 5 (x h)
= 5=2 5=2
+ 7=2
(1 2xh + h2 ) (1 2xh + h2 ) (1 2xh + h2 )
d3 1
) (x; h) = 15x3 9x = 3! 5x3 3x = 3!P3 (x) ; (12.108)
dh3 h=0 2
The function (x; h) is the generating function for the Legendre polynomials.
12.6. THE GENERATING FUNCTION FOR THE LEGENDRE POLYNOMIALS321
Example 12.4 Consider a point charge Q located at the position ~r0 (the source
point). Find an expression for the electrostatic potential V (r) at the point
P located at the position ~r (the …eld point). (Fig. 12.6
Solution: For a point charge located at the position described by the vector ~r0
the electric potential at the point at position ~r is inversely proportional
to the distance between the charge position and the point p (j~r ~r0 j) and
directly proportional to the charge. It can be expressed as
Q 1
V (r) = : (12.112)
4 0 j~
r ~r0 j
Noting that
1 1=2
= r2 + r02 2rr0 cos (12.113)
j~r ~r0 j
For r0 < r we may write
" # 1=2
2
1 1 r0 r0
= 1+ 2 cos : (12.114)
j~r ~r0 j r r r
for
r0
; x = cos h= (12.118)
r
the expression for the electric potential becomes
1
X l
Q 1 Q r0
V (r) = = Pl (cos ) (12.119)
4 0 j~r ~r0 j 4 0r r
l=0
2 1=2
1 1 r r
= 1+ 0 2 cos : (12.120)
j~r ~r0 j r 0 r r0
or
1 1 1=2
= 1 2xh + h2 (12.121)
j~r ~r0 j r0
where
r0
; x = cos : h= (12.122)
r
In this case the potential is given by
1
X l
Q 1 Q r
V (r) = = Pl (cos ) (12.123)
4 0 j~r ~r0 j 4 0 r0 r0
l=0
1 dl l
Pl (x) = x2 1 (12.124)
2l l! dxl
Example 12.5 Use Rodrigues’formula to …nd the equation for P2 (x).
Solution: Using Rodrigues’ Formula P2 (x) can be expressed as
1 d2 2
P2 (x) = x2 1 (12.125)
22 2! dx2
which can be simpli…ed as follows:
1 d2 1 d
P2 (x) = x4 2x2 + 1 = 4x3 4x
23 dx2 23 dx
1 4 1
) P2 (x) = 12x2 4 = 3x2 1 = 3x2 1 (12.126)
23 23 2
XN
dN N dn u dN n v
(uv) = (12.127)
dxN n dxn dxN n
n=0
12.6. THE GENERATING FUNCTION FOR THE LEGENDRE POLYNOMIALS323
For u = x2 ; we have
d0 u 1
2 d u d2 u dn u
= x ; = 2x; = 2; = 0 for n = 3 (12.130)
dx0 dx1 dx2 dxn
so that
X8
d8 8 dn u d8 n v 8 d0 u d8 v
(uv) = =
dx8 n dxn dx8 n 0 dx0 dx8
n=0
d1 u
8 d7 v 8 d2 u d6 v
+ +
1
dx1 dx7 2 dx2 dx6
8 7
8! d v 8! d v 8! d6 v
= x2 + 2x + 2
8!0! dx8 7!1! dx7 6!2! dx6
d8 d8 v d7 v d6 v
) 8 (uv) = x2 8
+ 16x 7
+ 56 : (12.131)
dx dx dx dx6
Noting that for v = e2x
dn v
= 2n e2x (12.132)
dxn
we …nd
d8
(uv) = x2 28 e2x + 16x 27 e2x + 56 26 e2x = 28 x2 + 8x + 14 e2x
dx8
d8
) 8 (uv) = 256x2 + 2048x + 3584 e2x (12.133)
dx
A Recursion Relation : We recall the generating function
1
X
1=2
1 2xh + h2 = Pl (x) hl (12.134)
l=0
so that
1
d 1=2 d X
1 2xh + h2 = Pl (x) hl
dh dh
l=0
1
X
(x h)
) p = lPl (x) hl 1
: (12.135)
(1 2xh + h2 ) 1 2xh + h2 l=0
324CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
If we use
X 1
1
p = Pl (x) hl (12.136)
1 2xh + h2 l=0
l=0 l=0
1
X 1
X 1
X 1
X
)x Pl (x) hl Pl (x) hl+1 = lPl (x) hl 1
2x lPl (x) hl
l=0 l=0 l=0 l=0
1
X
+ lPl (x) hl+1
l=0
1
X 1
X 1
X
) lPl (x) hl 1
x (1 + 2l) Pl (x) hl + (1 + l) Pl (x) hl+1 = 0
l=0 l=0 l=0
(12.137)
Noting that
1
X
lPl (x) hl 1
= 1P1 (x) + 2P2 (x) h1 + 2P3 (x) h2 + 4P4 (x) h3 + :::
l=0
1
X
= (l + 1) Pl+1 (x) hl (12.138)
l=0
and
1
X
(1 + l) Pl (x) hl+1 = P0 (x) h + 2P1 (x) h2 + 3P2 (x) h3 + 4P3 (x) h4
l=0
+5P4 (x) h5 + ::: = 0 P 1 (x) + P0 (x) h + 2P1 (x) h2 + 3P2 (x) h3
1
X 1
X
+4P3 (x) h4 ) (1 + l) Pl (x) hl+1 = lPl 1 (x) hl ; (12.139)
l=0 l=0
we can write
1
X 1
X 1
X
(l + 1) Pl+1 (x) hl x (1 + 2l) Pl (x) hl + lPl 1 (x) hl = 0
l=0 l=0 l=0
1
X
) [(l + 1) Pl+1 (x) x (1 + 2l) Pl (x) + lPl 1 (x)] hl = 0 (12.140)
l=0
12.7. LEGENDRE SERIES 325
The expression for the expansion coe¢ cients is determined using the orthogo-
nality property of the Legendre polynomials. Multiplying both sides from the
left by hPl (x)j ; we can write
1
X 1
X 2
hPl (x)j f (x)i = am hPl (x) jPm (x)i = am lm
m=0 m=0
2l + 1
2 2l + 1
) hPl (x)j f (x)i = al ) al = hPl (x)j f (x)i : (12.144)
2l + 1 2
Recalling that
Z1
hPl (x)j f (x)i = Pl (x) f (x) dx (12.145)
1
and Pl (x) is real, the expression for the expansion coe¢ cients is expressible as
Z1
2l + 1
al = Pl (x) f (x) dx: (12.146)
2
1
1 1 x 0
f (x) = (12.147)
+1 0 x 1
326CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
f(x)
0
-1.0 -0.5 0.0 0.5 1.0
x
-1
1
X
f (x) = am Pm (x) (12.148)
m=0
where
Z1
2l + 1
al = Pl (x) f (x) dx: (12.149)
2
1
The function has di¤erent values in the interval ( 1; 0) and (0; 1). Hence
12.7. LEGENDRE SERIES 327
1 dl l
Pl (x) = x2 1 (12.151)
2l l! dxl
if we replace x by x, we have
1 dl l l 1 dl l
Pl ( x) = x2 1 = ( 1) x2 1 (12.152)
2l l! d ( x)l 2l l! dxl
we leads to
Pl (x) l = even
Pl ( x) = (12.153)
Pl (x) l = odd
Applying this result we …nd for al
8
>
> Z1
<
(2l + 1) Pl (x) dx; l = odd
al = (12.154)
>
> 0
:
0; l = even
Using
1 1
P0 (x) = 1; P1 (x) = x; P2 (x) = 3x2 1 ; P3 (x) = 5x3 3x
2 2
(12.155)
328CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
we …nd
Z1 Z1
3 7 7
a1 = 3 xdx = ; a3 = 5x3 3x dx = (12.156)
2 2 3
0 0
3 7
f (x) = P1 (x) P2 (x) + ::: (12.157)
2 3
n
X
gn (x) = an xn + an 1x
n 1
+ :::a0 = a k xk (12.158)
k=0
that can best …t to the curve described by the function f (x):Then the polynomial
gn (x) can best …t the curve described by f (x) when the integral
Z 1
2
[f (x) gn (x)] dx (12.159)
1
a minimum. This polynomial is nothing but the nth order Legendre polynomial.
12.8. THE ASSOCIATED LEGENDRE DIFFERENTIAL EQUATION 329
m2
1 x2 y 00 2xy 0 + l (l + 1) y = 0:
1 x2
1 x2 y 00 2xy 0 + l (l + 1) y = 0
y (x) = Pl (x) :
We can solve this DE using series substituting method. However, here we use
a di¤erent approach. To this end we introduce a transformation of variable
de…ned by
m=2
y (x) = 1 x2 u (x) : (12.160)
Upon di¤erentiating this with respect to x
m=2 (m 2)=2
y 0 (x) = 1 x2 u0 (x) m 1 x2 xu (x) (12.161)
and
00 m=2 (m 2)=2
y (x) = 1 x2 u00 (x) m 1 x2 xu0 (x)
(m 2)=2 (m 4)=2
m 1 x2 u (x) + m (m 2) 1 x2 x2 u (x)
(m 2)=2
m 1 x2 xu0 (x)
00 m=2 (m 2)=2
) y (x) = 1 x2 u00 (x) 2m 1 x2 xu0 (x)
(m 2)=2 (m 4)=2
m 1 x2 u (x) + m (m 2) 1 x2 x2 u (x) (12.162)
We recall
m=2
y (x) = 1 x2 xu (x) (12.164)
so that
m2 m=2 m=2
l (l + 1) y = l (l + 1) 1 x2 u (x) m2 1 x2 u (x) :
1 x2
(12.165)
330CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
and using
m=2 (m 2)=2
y 0 (x) = 1 x2 u0 (x) m 1 x2 xu (x) (12.166)
we have
m=2 (m 2)=2
2xy 0 (x) = 2 1 x2 xu0 (x) 2m 1 x2 x2 u (x) : (12.167)
m2
1 x2 y 00 2xy 0 + l (l + 1) y=0 (12.168)
1 x2
as
(m+2)=2 m=2 m=2
1 x2 u00 (x) 2m 1 x2 xu0 (x) m 1 x2 u (x)
2 (m 2)=2 2 m=2
+m (m 2) 1 x x2 u (x) 2 1 x xu0 (x)
(m 2)=2 m=2
+2m 1 x2 x2 u (x) + l (l + 1) 1 x2 u (x)
m=2
m2 1 x2 u (x) = 0: (12.169)
m=2
Dividing the entire equation by 1 x2 ; we …nd
x2
1 x2 u00 (x) 2mxu0 (x) u (x)
mu (x) + m (m 2)
1 x2
x2 1
2xu0 (x) + 2m u (x) + l (l + 1) u (x) m2 u (x) = 0
1 x2 1 x2
) 1 x2 u00 (x) 2mxu0 (x) 2xu0 (x) + l (l + 1) u (x) mu (x)
m (m 2) x2 2mx2 m2
+ 2
u (x) + u (x) u (x) = 0 (12.170)
1 x 1 x2 1 x2
To …nd the general solution for the values of m, we shall examine the di¤erential
equation for the …rst three values of m.
12.8. THE ASSOCIATED LEGENDRE DIFFERENTIAL EQUATION 331
Z1 Z1
hPlm j Plm
0 i = hPlm j xi hx Plm
0 i dx = Pl m (x) Plm
0 (x) dx
1 1
Z1
2 (l + m)!
) hPlm j Plm
0 i = Plm (x) Plm
0 (x) dx = ll0 ; (12.190)
2l + 1 (l m)!
1
where we used that fact that the associated Legendre polynomials are real func-
tions.
d2 (')
+ m2 (') = 0; (12.191)
d'2
d2 y (x) dy (x) m2
1 x2 2x + l (l + 1) y (x) = 0 (12.193)
dx2 dx 1 x2
we have
dy (x) d d 1 d
= = (12.196)
dx d dx sin ( ) d
so that
d2 y (x) d 1 d
=
dx2 dx sin ( ) d
d d 1 d 1 d2 d
=
d d sin ( ) dx sin ( ) d 2 dx
d2 y (x) cos ( ) d ( ) 1 d2 ( )
) = + (12.197)
dx2 sin3 ( ) d sin2 ( ) d 2
where we used the product rule. Then using the results above the associated
Legendre di¤erential equation can be expressed as
cos ( ) d ( ) 1 d2 ( ) cos ( ) d
sin2 ( ) 3 + 2 2 +2
sin ( ) d sin ( ) d sin ( ) d
m2
+ l (l + 1) =0 (12.198)
sin2 ( )
d2 ( ) cos ( ) d m2
) 2 + + l (l + 1) =0 (12.199)
d sin ( ) d sin2 ( )
1 d d ( ) m2
sin + l (l + 1) ( ) = 0: (12.200)
sin d d sin2
The solution to this form of the associated Legendre di¤erential equation can
then be written as
dm
( ) = Plm (cos ( )) = sinm ( ) m Pl (cos ( )) : (12.201)
d (cos ( ))
Note that
l = 0; 1; 2; 3::: (12.202)
and
m2 l2 ) m = l; l + 1; l + 2::::0; 1; ; 3:::l:
The Spherical harmonics form an orthonormal set of vectors/functions and the
orthonormality condition is given by
Z Z 2
Ylm ( ; ') Yl0 m0 ( ; ') sin ( ) d d' = ll0 mm0 : (12.203)
=0 '=0
Spherical harmonics form a complete set of vectors, which we may express, use
Dirac’s notion, as
X1 X l
jYlm i hYlm j = 1
l=0 m= l
1 X
X l Z Z2
0 0 0
; '0 jYlm i hYlm j ; '0 sin d 0 d'0 = 1
l=0 m= l 0 0
1 X
X l Z Z2
0 0 0
) Ylm ; '0 Ylm ; '0 sin d 0 d'0 = 1:
l=0 m= l 0 0
f ( ; ') = h ; ' jf i
that is de…ned in the domain [0; ] and ' [0; 2 ] can be expressed as linear
combination of the spherical harmonics
1 X
X l
f ( ; ') = h ; ' jf i = h ; ' jYlm i hYlm j f i (12.204)
l=0 m= l
where
0 0
h j = ; h'j '0 i = (' '0 ) (12.206)
we can rewrite
1 X
X l
f ( ; ') = h ; ' jf i = h ; ' jYlm i hYlm j f i (12.207)
l=0 m= l
336CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
as
f ( ; ') = h ; ' jf i
1 X
X l Z Z2
0 0 0
= h ; ' jYlm i hYlm j ; '0 ; '0 f i sin d 0 d'0
l=0 m= l 0 0
1 X
X l Z Z2
0 0 0
= h ; ' jYlm i hYlm j ; '0 ; '0 f i sin d 0 d'0 (12.208)
l=0 m= l 0 0
f ( ; ') = h ; ' jf i
2 2 3
1 X
X l Z Z
= 4 Ylm 0
; '0 f 0
; '0 sin 0
d 0 d'0 5 Ylm ( ; ') (12.210)
l=0 m= l 0 0
or
1 X
X l
f ( ; ') = Alm Ylm ( ; ') ; (12.211)
l=0 l
where Z Z 2
Alm = Ylm ( ; ') f ( ; ') sin d d': (12.212)
0 0
X1 l
1 1 r<
=p = Pl (cos ( )) (12.213)
j~r ~r0 j r2 + r02 2rr0 cos ( ) rl+1
l=0 >
where (
l r 0l
r< r l+1
; for r0 < r
l+1
= rl
r> r 0l+1
for r < r0
12.9. SPHERICAL HARMONICS AND THE ADDITION THEOREM 337
Figure 12.3: Two point described by two vectors, ~r and ~r0 : In spherical coordi-
nates the two points have coordinates (r; ; ') and r0 ; 0 ; '0 ;respectively. The
angle between these two vectors is :
cos (' '0 ) = cos (') cos ('0 ) + sin (') sin ('0 )
cos2 (') + sin2 (') = 1; sin2 ( ) + cos2 ( ) = 1 (12.221)
one …nds
0 0
j~r ~r0 j = r2 +r02 2rr0 cos ( ) cos + sin ( ) sin cos (' '0 ) : (12.222)
where
0 0
cos ( ) = cos ( ) cos + sin ( ) sin cos (' '0 ) (12.224)
and is the angle between the two vectors, ~r and ~r0 ; shown in Fig. 12.3:
dq 0 0
= ) dq 0 = d 0
= r02 dr0 sin d 0 d'0 (12.226)
d 0
12.9. SPHERICAL HARMONICS AND THE ADDITION THEOREM 339
The relation
X1 l
1 r<
= P (cos ) ;
l+1 l
(12.228)
j~r ~r0 j r
l=0 >
1
X
1 r0l
= Pl (cos ) : (12.229)
j~r ~r0 j dl+1
l=0
1
X Z RZ Z 2
r0l 0
V (~r) = Pl (cos ) r02 dr0 sin d 0 d'0 :
4 0 dl+1 0 =0 '=0
l=0
(12.230)
Now applying the addition theorem for the spherical harmonics
l
X
4 0
Pl (cos ) = Ylm ( ; ') Ylm ; '0 (12.231)
2l + 1
m= l
340CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
we …nd
1 X
X l
4
V (~r) =
4 0 2l + 1
l=0 m= l
Z R Z Z 2 0l
r 0 0
Ylm ( ; ') Ylm ; '0 r02 dr0 sin d 0 d'0 :
0 =0 '=0 dl+1
(12.232)
Since the integration is with respect to the prime variables we can rewrite
the above expression in the form
1 X
X l
r0l 1
V (~r) = Ylm ( ; ')
0 dl+1 2l + 1
l=0 m= l
Z R Z Z 2
r0l+2 0 0 0
dr Ylm ; '0 sin d 0 d'0 : (12.233)
0 dl+1 =0 '=0
gives
r
1 p
Y00 ( ; ') = ) 4 Y00 ( ; ') = 1: (12.235)
4
Substituting this into the expression for the potential, one can then
1 X
X l
1
V (~r) = Ylm ( ; ')
0 2l + 1
l=0 m= l
p Z R Z Z 2
r0l+2 0 0 0 0
4 dr Y00 ; '0 Ylm ; '0 sin d 0 d'0 :
0 dl+1 =0 '=0
(12.236)
we have
Z Z 2
0 0 0
Y00 ; '0 Ylm ; '0 sin d 0 d'0 = 0l 0m (12.238)
=0 '=0
12.10. THE METHOD OF FROBENIUS AND THE BESSEL EQUATION341
3=2
Noting that y0 x0 = a1 (constant), the solution can be written as
x2 y 00 + xy 0 + x2 p2 y = 0; (12.251)
1
X 1
X
y (x) = xs an xn = an xn+s (12.252)
n=0 n=0
we have
1
X 1
X
x2 p2 y = an xn+s+2 p2 an xn+s ; (12.253)
n=0 n=0
1
X 1
X
y 0 (x) = (n + s) an xn+s 1
) xy 0 (x) = (n + s) an xn+s ; (12.254)
n=0 n=0
1
X
y 00 (x) = (n + s) (n + s 1) an xn+s 2
n=0
1
X
2 00
) x y (x) = (n + s) (n + s 1) an xn+s ; (12.255)
n=0
x2 y 00 + xy 0 + x2 p2 y = 0
1
X 1
X 1
X
n+s n+s
) (n + s) (n + s 1) an x + (n + s) an x + an xn+s+2
n=0 n=0 n=0
1
X 1
X
2
p an xn+s = 0 ) (n + s) (n + s 1) + (n + s) p2 an xn+s
n=0 n=0
1
X X1 h i 1
X
2
+ an xn+s+2 = 0 ) (n + s) p2 an xn+s + an xn+s+2 = 0
n=0 n=0 n=0
(12.256)
344CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
we …nd
h i
2
s2 p2 a0 xs + (1 + s) p2 a1 xs+1
1 h
X i 1
X
2
+ (n + 2 + s) p2 an+2 xn+s+2 + an xn+s+2 = 0 (12.261)
n=0 n=0
h i
2
s2 p2 a0 xs + (1 + s) p2 a1 xs+1
1 nh
X i o
2
+ (n + 2 + s) p2 an+2 + an xn+s+2 = 0: (12.262)
n=0
12.10. THE METHOD OF FROBENIUS AND THE BESSEL EQUATION345
which leads to
h i
2
s2 p2 a0 = 0; (1 + s) p2 a1 = 0;
h i
2
(n + 2 + s) p2 an+2 + an = 0 (12.264)
s= p (12.265)
we …nd h i
2
(1 p) p2 a1 = 0 ) [1 2p] a1 = 0: (12.267)
a1 = 0: (12.268)
First solution ( s=p): For this case the recursion formula becomes
an an an
an+2 = 2 = 2 =
(n + 2 + p) p2
(n + 2) + 2 (n + 2) p (n + 2) (n + 2 + 2p)
(12.271)
Since we found a1 = 0 all the odd terms vanish and the recursion relation can
be expressed, for the even terms using n + 2 = 2m ) n = 2m 2, as
a2m 2 a2m 2
a2m = = : (12.272)
(2m) (2m + 2p) 22 m (m + p)
346CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
(p + 3) = (p + 2) (p + 2) = p (p + 1) (p + 2) (p) ; (12.275)
we can rewrite
a0 a0 p (p) a0 (p + 1)
m = 1 ) a2 = = =
22 1! (1 + p) 22 1!p (1 + p) (p) 22 1! (p + 2)
(12.276)
a0 a0 p (p)
m = 2 ) a4 = = 4
24 2! (1 + p) (2 + p) 2 2!p (1 + p) (2 + p) (p)
a0 (p + 1)
) a4 = 4 (12.277)
2 2! (p + 3)
a0
m = 3 ) a6 =
26 3! (1+ p) (2 + p) (3 + p)
a0 p (p) a0 (p + 1)
= ) a6 = (12.278)
:
26 3!p (1 + p) (2 + p) (3 + p) (p) 26 3! (p + 4)
Therefore, the solution becomes
1
X 1
X
y (x) = xs an xn = xs a2m x2m
n=0 m=0
p a0 (p + 1) 2 a0 (p + 1) 4 a0 (p + 1) 6
y (x) = x a0 x + 4 x x + :::
22 1! (p + 2) 2 2! (p + 3) 26 3! (p + 4)
(12.279)
This function Jp (x) is called the Bessel function of the …rst kind of order p.
Second Solution s=-p: The second solution, when s = p can easily be
obtained from the …rst solution by replacing p with p,
1
X n 2n p
( 1) x
J p (x) = : (12.285)
n=0
(n + 1) (n + 1 p) 2
x2 y 00 + xy 0 + x2 p2 y = 0 (12.289)
and have found the solution is the Bessel function Jp (x) : We now determine
the orthonormality of these functions. To this end, noting that
d dy
x2 y 00 + xy 0 = x (xy 00 + y 0 ) = x x (12.290)
dx dx
we can rewrite the Bessel Equation as
d dy
x x + x2 p2 y = 0 (12.291)
dx dx
the solution of which is given by the bessel functions, y(x) = Jp (x) : Thus for a
the Bessel function, Jp (u) ;one can write
d dJp (u)
u u + u2 p2 y = 0 (12.292)
du du
We introduce the transformation of variable de…ned by
u= ix ) du = i dx;
12.11. THE ORTHOGONALITY OF THE BESSEL FUNCTIONS 349
where
i =f 1; 2; 3 ; :::g
are the zeros of the Bessel function. Then one can write for = i
d dJp ( i x) 2 2
x x + ix p2 Jp ( i x) = 0; (12.293)
dx dx
and = j
d dJp ( j x) 2 2
x x + jx p2 Jp ( j x) =0 (12.294)
dx dx
Multiplying Eq. (12.293) by Jp ( j x) and Eq. (12.294) by Jp ( i x), we have
d dJp ( i x) 2 2
xJp ( j x) x + ix p2 Jp ( j x) Jp ( i x) = 0; (12.295)
dx dx
and
d dJp ( j x) 2 2
xJp ( i x) x jx p2 Jp ( i x) Jp ( j x) =0 (12.296)
dx dx
so that adding these two equations leads to
d d d d
Jp ( j x) x Jp ( i x) Jp ( i x) x Jp ( j x)
dx dx dx dx
2 2
+ i j Jp ( i x) Jp ( j x) x = 0: (12.297)
350CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
x = 0 ) xJp ( i x) = 0; x = 1 ) Jp ( i x) x = Jp ( i ) = 0: (12.302)
one …nds
Z 1 Z 1
d d d d
Jp ( j x) x Jp ( i x) dx = x (Jp ( i x)) (Jp ( j x)) dx
0 dx dx 0 dx dx
(12.303)
and
Z 1 Z 1
d d d d
Jp ( i x) x Jp ( j x) =
(Jp ( i x)) dx: x (Jp ( j x))
0 dx dx0 dx dx
(12.304)
Now substituting Eq. (12.303) and Eq. (12.304) into Eq. (12.298), one …nds
Z 1
2 2
i j Jp ( i x) Jp ( j x) xdx = 0: (12.305)
0
we must have Z 1
Jp ( i x) Jp ( j x) xdx = 0: (12.306)
0
But for i = j; since
2 2 2 2
i j = i i = 0; (12.307)
we should generally expect that
Z 1 Z 1
Jp ( i x) Jp ( j x) xdx = Jp2 ( i x) xdx = Constant (12.308)
0 0
This is the orthogonality condition for the Bessel functions, which we may write
as Z 1
0; i 6= j
Jp ( i x) Jp ( j x) xdx = = ij C: (12.309)
0 C; i = j
To …nd the constant C, let
2 2 2 2 2
i = j + ) i j =( j + ) j =2 j + (12.310)
where j is the zeros of the bessel function
Jp ( j) = 0;
but i is not the zeros of the Bessel function but it becomes the zeros of the
Bessel function when ! 0;
lim Jp ( i ) = lim Jp ( j + ) = Jp ( j) = 0:
!0 !0
Noting that
d dJp (x) 0
(Jp ( j x)) = j = j Jp ( j) (12.313)
dx x=1 dx x= j
we can write
Z 1
j Jp (( + )) Jp0 (
j j)
Jp (( j + ) x) Jp ( j x) xdx = : (12.314)
0 (2 j + )
f ( )j =0 = Jp ( j + )j =0 = Jp ( ) ;
df ( ) dJp ( j + )
= = Jp0 ( j) (12.316)
d =0 d =0
so that
2 00
Jp ( + ) = Jp ( ) + Jp0 ( ) + Jp x + ::: (12.317)
2!
Since j is the zero’s of the Bessel function we know Jp ( j) = 0; one …nds
h 00
i
Jp ( j + )= Jp0 ( j) + Jp ( j) x + ::: (12.318)
2!
Substituting this into Eq. (12.314), we …nd
h i
Z 0
j Jp ( ) Jp0 ( j) +
00
2! Jp ( ) x + :::
1 j j
Jp (( j + ) x) Jp ( j x) xdx =
0 2 j +
(12.319)
Now we set = 0, so that we …nd
Z 1 Jp0 ( 0
j ) Jp ( j)
Jp ( j x) Jp ( j x) xdx = : (12.320)
0 2
Therefore, the orthogonality condition for the Bessel function can be written as
Z 1 Jp0 ( i ) Jp0 ( j)
Jp ( i x) Jp ( j x) xdx = ij : (12.321)
0 2
Limiting (Asymptotic) Forms for the Bessel Functions: The Bessel functions
takes the following approximate expression for di¤erent limiting cases:
12.11. THE ORTHOGONALITY OF THE BESSEL FUNCTIONS 353
1 2a 1 2 a2 p2 c2
y 00 + y0 + bcxc + y=0 (12.329)
x x2
has the solution
y = xa Zp (bxc ) ; (12.330)
where Zp stands for Jp or Np or any linear combination, and a; b; c are constants.
See the lengthening pendulum example.
354CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS
When this is the case the solutions to the di¤erential equation consist of either
two independent Frobenius series
1
X 1
X
y1 (x) = An xn+s ; y2 (x) = Bn xn+s (12.334)
n=0 n=0
or one solution, y1 (x) ; which is a Frobenius series, and the second solution
which is given by
y2 (x) = y1 (x) ln x + y (x) ; (12.335)
where y (x) is a Frobenius series. That means
1
X 1
X
n+s
y1 (x) = An x ; y2 (x) = y1 (x) ln x + Bn xn+s (12.336)
n=0 n=0
The second case occurs only when the roots of the indicial equation are equal or
di¤ er by an integer.
Example 12.12 Consider the di¤erential equation
1 0 2
y 00 + y y=0 (12.337)
x2 x3
which has the solution
y(x) = e1=x : (12.338)
Discuss the applicability of a Frobenius-type generalized power-series so-
lution to this di¤erential equation.
12.12. FUCH’S THEOREM FOR 2-ND DE 355
where
1 2
f (x) = ; g (x) = (12.340)
x2 x3
Using the the given solution
we see that
1 d 1 1=x
F (x) = f (x) y 0 = e1=x = e (12.342)
x2 dx x4
and
2 1=x
G(x) = g (x) y = e : (12.343)
x3
These two functions, F (x) and G(x); have a singular point at x = 0
and can not be expressed as Frobenius series and therefore the Frobenius
Method can not be used to …nd the solution of the di¤erential equation.
Example 12.13 Bessel Functions: An Application (The lengthening Pendu-
lum): Consider a simple pendulum that consists of a mass m attached to
string of length l0 at the initial time t = 0: The length of the pendulum is
increasing at a steady rate,
dl
l_ = = v:
dt
Find the equation of motion and determine the solution for small oscilla-
tion.
Solution: At a given time, t; let the length of the pendulum be,
q
2 2
l (t) = x (t) + y (t) : (12.344)
d2 1 2a d 1 2 a2 p2 c2
2
+ + bclc + =0 (12.355)
dl x dl l2
so that
(l) = la Zp (blc ) : (12.356)
Comparing this with the DE we derived for the lengthening pendulum, we
note that
1 2a 2
= ; (12.357)
l l
1 2 a2 p2 c2 g
bclc + = (12.358)
l2 lv 2
There follows that
r
1 1 g 2 a2
a= ;c = ;b = 2 2
;p = 2 ) p = 1 (12.359)
2 2 v c
A partial di¤erential
equation
If we use
~ (x; y; z)
E = Ex (x; y; z) x^ + Ey (x; y; z) y^ + Ez (x; y; z) z^;
@ @ @
r = x
^+ y^ + z^ (13.2)
@x @y @z
~ (x; y; z) in terms of
Poisson’s Equation: If we express the electric …eld, E
the electric potential V (x; y; z) which is given by
~ (x; y; z) =
E rV (x; y; z) (13.4)
359
360 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
or
@ 2 V (x; y; z) @ 2 V (x; y; z) @ 2 V (x; y; z) (x; y; z)
+ + = : (13.6)
@x2 @y 2 @z 2 0
@2f 1 @2f
= (13.9)
@z 2 v 2 @t2
where r
T
v= : (13.10)
Helmholtz Equation: For three dimensional case the wave equation has the
form
@2f @2f @2f 1 @2f
2
+ 2 + 2
= 2 2 (13.11)
@x @y @z v @t
The space dependent part that we would obtain using separation of variables
has the form
@2f @2f @2f
2
+ 2 + + k2 f = 0 (13.12)
@x @y @z 2
this is called Helmholtz equation.
Schrödinger Equation: in quantum mechanics the state of a particle is rep-
resented by the wave function (x; y; z; t) which satis…ed the time dependent
Schrödinger equation
~2 2 @
V (x; y; z; t) (x; y; z; t) r (x; y; z; t) = i~ (x; y; z; t) (13.13)
2m @t
We will study a technique commonly used to solve the PDEs discussed above.
This technique is called the separation of variables technique. We will learn this
technique by solving problems.
d2 f (x)
= k 2 f (x) : (13.14)
dx2
1 = k; 2 = k (13.17)
For the minus case the solution to the indicial equation are
1 = ik; 2 = ik (13.19)
This equation consists of three independent terms and the sum of these terms
must be zero. This is possible if and only if these terms are constants. Therefore,
we can write
so that
kx2 + ky2 kz2 = 0: (13.25)
These equations are 2-nd order ODE and can be solved using any of the tech-
niques we have studied so far. Considering the x dependent part, we have
1 d2 X (x) d2 X (x)
= kx2 ) + kx2 X (x) = 0: (13.26)
X (x) dx2 dx2
d2 Y (y)
+ ky2 Y (y) = 0: (13.27)
dy 2
13.2. LAPLACE’S EQUATION IN CARTESIAN COORDINATES 363
Assuming kx and ky are real, the solutions to the above ODEs, using the results
in Ex. 23, may be expressed in any one of the following form
8
>
> Akx eikx x + Bkx e ikx x ;
<
Ckx cos (kx x) + Dkx sin (kx x) ;
Xkx (x) = ;
>
> Ekx cos kx x kx ;
:
Fkx sin kx x + kx ;
8
>
> Aky eiky y + Bky e iky y ;
>
>
< ky cos (ky y) + Dky sin (ky y) ;
C
Yky (y) = Eky cos ky y : (13.28)
> ky ;
>
>
>
: Fky sin ky y + ky
d2 Z (z)
kz2 Z (z) = 0; (13.29)
dz 2
assuming kz is real, the solution can be expressed as
8
>
> Akz ekz z + Bkz e kz z ;
<
Ckz cosh (kz z) + Dkz sinh (kz z) ;
Zkz (z) = ;: (13.30)
>
> Ekz cosh kz z + kz ;
:
Fkz sinh kz z + kz ;
Note that the constants kx ; ky ; are kz along with the other constants determined
from the boundary condition set for the given problem and the condition
We shall see how these constants are determined in the next example.
Figure 13.1: A rectangular wave guide with width w and height h. It is in…nitely
long along the z direction.
where
kx2 + ky2 kz2 = 0; (13.34)
Xkx (x) ; Yky (y) ; and Zkz (x) given by Eqs. (13.28) and (13.30). The
boundary conditions are
And also the potential must …nite everywhere inside the waveguide. Im-
posing these boundary conditions
X
V (0; y; z) = 0) Xkx (0) Yky (y) Zkz (z) = 0 ) Xkx (0) (13.36)
= 0;
kx ;ky ;kz
X
V (w; y; z) = 0) Xkx (w) Yky (y) Zkz (z) = 0 ) Xkx (w)(13.37)
= 0;
kx ;ky ;kz
X
V (x; 0; z) = 0) Xkx (x) Yky (0) Zkz (z) = 0 ) Yky (0) (13.38)
= 0;
kx ;ky ;kz
13.2. LAPLACE’S EQUATION IN CARTESIAN COORDINATES 365
Since the cylinder is in…nite along the z-direction and the potential must
be …nite, we must have
0 1
X
lim @ Xkx (x) Yky (y) Zkz (z)A = Constant
z! 1
kx ;ky ;kz
we …nd
n
ky = i : (13.43)
w
At this stage we note
n
Xkx (x) = Dkx sin (kx x) = Dn sin x ;
w
n n
Yky (y) = Dky sin (ky y) = Dky sin i y = Dn sinh y ;
w w
Zkz (z) = A: (13.44)
ei(ix)
e i(ix) ex e x
ex e x
sin (ix) = = =i
2i 2i 2
) sin (ix) = i sinh (x)
366 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
and included the i into the constant Dn . Then the general solution can
be expressed as
X 1
X n n
V (x; y; z) = Xkx (x) Yky (y) Zkz (x) = Hn sin x sinh y
n=1
w w
kx ;ky ;kz
(13.45)
Now applying the last boundary condition, V (x; h; z) = V0 ; we have
1
X n h
Vkx ky (x; y = h; z) = V0 ) Hn sin x sinh n = V0 (13.46)
n=1
w w
and
Z W w
m w m w
sin x dx = cos x = [1 ( 1)m ] (13.49)
0 w m w 0 m
we …nd
1
X h w w
Hn sinh n mn = V0 [1 ( 1)m ]
n=1
w 2 m
h 2
) Hm sinh m = V0 [1 ( 1)m ]
w m
(
0; m = even
) Hm = 4V0
; m = odd (13.50)
m sinh(m h
w )
@2 @2 @2
r2 V (x; y; z) = V (x; y; z) + V (x; y; z) + V (x; y; z) = 0 (13.51)
@x2 @y 2 @z 2
1 @ @ 1 @ @ 1 @2
r2 = 2
r2 + 2 sin + 2 2 = 0:
r @r @r r sin @ @ r sin @'2
1 @ @V 1 @ @V 1 @2V
) r2 V r = 2 r2 + 2 sin + 2 2 =0
r @r @r r sin @ @ r sin @'2
(13.52)
we have
1 @ @ 1 @ @
2
r2 (R (r) ( ) (')) + sin (R (r) ( ) ('))
r @r @r r2 sin @ @
1 @2
+ (R (r) ( ) (')) = 0: (13.53)
r2 sin2 @'2
where
1 d dR (r)
F1 (r) = r2 ;
R (r) dr dr
1 d d ( ) 1 d2 (')
F2 ( ; ') = sin + = 0: (13.58)
sin ( )d d sin2 (') d'2
We note that F1 (r) depends on r and F2 ( ; ') depends on and ' which are
independent. Therefore, each of these function must be a constant,
and
1 d d ( ) 1 d2 (')
sin + 2 =
l (l + 1)
sin ( )d d sin (') d'2
sin d d ( ) 1 d2 (')
) sin + = l (l + 1) sin2
( )d d (') d'2
sin d d ( ) 1 d2 (')
) sin + l (l + 1) sin2 + = 0: (13.61)
( )d d (') d'2
We note that in the above expression the …rst two terms depend on and the
third term depend on ': Therefore, we must have
1 d2 (')
= m2 ;
(') d'2
sin d d ( )
sin + l (l + 1) sin2 = m2 ; (13.62)
( )d d
d2 (')
+ m2 (') = 0; (13.63)
d'2
the solution of which are given by
and
d d ( ) m2
sin sin + sin2 l (l + 1) ( ) = 0: (13.64)
d d sin2
Introducing the transformation of variables de…ned by
p
x = cos ) dx = sin d ; sin = 1 x2
dx d dx dx
) d = ) = = (13.65)
sin sin sin2 1 x2
we may write the di¤erential equation
d d ( ) m2
sin sin + sin2 l (l + 1) ( )=0 (13.66)
d d sin2
as
d d (x) m2
1 x2 1 x2 + 1 x2 l (l + 1) (x) = 0
dx dx 1 x2
d2 (x) d (x) m2
) 1 x2 2x + l (l + 1) (x) = 0 (13.67)
dx2 dx 1 x2
370 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
Next we proceed to …nd the solution for the di¤erential equation for the r
dependent part
d dR (r)
r2 l (l + 1) R (r) = 0
dr dr
d2 R (r) dR (r)
) r2 + 2r l (l + 1) R (r) = 0
dr2 dr
d2 R (r) 2 dR (r) l (l + 1)
) + R (r) = 0: (13.69)
dr2 r dr r2
The above di¤erential equation is singular at r = 0. In such cases we know that
the solution is determined by using Frobenius method. Let’s assume a solution
given by
X1
R (r) = an rn+s (13.70)
n=0
so that
1 1
dR (r) X d2 R (r) X
= an (n + s) xn+s 1
; = an (n + s) (n + s 1) xn+s 2
:
dr n=0
dr2 n=0
d2 R (r) 2 dR (r) l (l + 1)
+ R (r) = 0 (13.71)
dr2 r dr r2
becomes
1
X 1
X
an (n + s) (n + s 1) xn+s 2
+2 an (n + s) xn+s 2
n=0 n=0
X
n+s 2
l (l + 1) an x =0
n=0
1
X
) an [(n + s) (n + s 1) + 2 (n + s) l (l + 1)] xn+s 2
=0
n=0
1
X
) an [(n + s) (n + s + 1) l (l + 1)] xn+s 2
= 0: (13.72)
n=0
Then the general solution for the radial part of the DE given by a linear com-
bination of these two functions which we can write as
Cl
Rl (r) = Bl rl + l+1 : (13.77)
r
Now we can write the general solution for the Laplace equation in spherical
coordinates as
1 X
X l
V (r; ; ') = Rl (r) lm ( ) m (') (13.78)
l=0 m= l
or
1 X
X l
Cl
V (r; ; ') = Bl r l + Plm (cos ) [Am cos (m') + Bm sin (m')] :
rl+1
l=0 m= l
(13.79)
Example 13.3 All of space is initially …lled with a uniform electric …eld of
magnitude, Eo ; pointing in the positive z-direction. A grounded conducting
sphere of radius, a; is introduced into the space with its center at the
origin of coordinates. We wish to …nd the electrostatic potential at all
points outside of the sphere.
Solution: Since there is no free charge outside the sphere, the electrical po-
tential V (~r) satis…es the Laplace’s equation. Because of the spherical
symmetry, it is easy to solve the Laplace’s equation in spherical coordi-
nates,
r2 V (~r) = 0
1 @ @ 1 @ @
) r2 V (r; ; ') + 2 sin V (r; ; ')
r2 @r @r r sin @ @
1 @2
+ 2 2 V (r; ; ') = 0: (13.80)
r sin @'2
While solving this partial di¤erential equation I will outline the basic
steps which we used in the previous example. It can be used as a general
guideline to solve partial di¤erential equations.
372 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
Identify and write down the boundary conditions for the given problem: For
this problem we know that the sphere is grounded. Hence, the electric
potential at any point on the surface of the sphere must be zero. That
means
V (r = a; ; ') = 0: (13.81)
In addition we should expect that the presence of the conducting sphere
far away from the sphere (in…nity) is negligible. In other words at in…nity
the electric potential is just the electric potential of the constant electric
…eld pointing along the z-direction. Recalling that the electric potential
V and electric …eld E~ are related by
~ =
E rV; (13.82)
Apply the boundary conditions to the general solution: The general solu-
tion is given by
1 X
X l
Cl
V (r; ; ') = Bl r l + Plm (cos ) [Am cos (m') + Bm sin (m')] :
rl+1
l=0 m= l
(13.85)
The sphere is grounded
V (r = a; ; ') = 0: (13.86)
which means
1 X
X l
Cl
V (a; ; ') = Bl al + Plm (cos )
al+1
l=0 m= l
Cl
(Am cos (m') + Bm sin (m')) = 0 ) Bl al + = 0 ) Cl = Bl a2l+10
al+1
(13.87)
we …nd
1 X
X l
V (r; ; ') = lim Bl rl Plm (cos ) (Am cos (m') + Bm sin (m'))
r!1 r!1
l=0 m= l
= E0 r cos (13.90)
Since there is no ' dependence on the right hand side, we must have
m=0 (13.91)
which leads to
1
X
lim A0 Bl rl Pl0 (cos ) = E0 r cos : (13.92)
r!1
l=0
comparing the right and left hand side of this equation we …nd
Substituting this result into the reduced expression for the potential
1
X a2l+1
V (r; ; ') = Bl r l Pl0 (cos ) (13.95)
rl+1
l=0
we …nd
a3
V (r; ; ') = 1 E0 r cos : (13.96)
r3
1 @ @V 1 @2V @2V
r2 V = s + 2 2
+ =0 (13.97)
s @s @s s @' @z 2
we have
(') Z (z) d R (s) Z (z) d2 (')
dR (s) d2 Z (z)
s + + R (s) (') =0
s ds ds s2 d'2 dz 2
(13.98)
so that multiplying by s2 =R (s) (') Z (z), we …nd
that leads to
1 d dR (s) 1 d2 Z (z)
s2 s + = p2 ;
sR (s) ds ds Z (z) dz 2
1 d2 (')
= p2 : (13.100)
(') d'2
1 d dR (s) 1 d2 Z (z)
s2 s + = p2 (13.101)
sR (s) ds ds Z (z) dz 2
13.4. LAPLACE’S EQUATION IN CYLINDRICAL COORDINATES 375
as
1 d dR (s) 1 d2 Z (z) p2
s + =
sR (s) ds ds Z (z) dz 2 s2
2 2
1 d dR (s) p 1 d Z (z)
s 2
+ =0 (13.102)
sR (s) ds ds s Z (z) dz 2
so that
2
1 d dR (s) p2 q
s = ;
sR (s) ds ds s2 r0
2
1 d2 Z (z) q
= ; (13.103)
Z (z) dz 2 r0
where q is a constant and r0 is the radius of the cylinder. The di¤erential
equation
2
1 d dR (s) p2 q
s = (13.104)
sR (s) ds ds s2 r0
can be rewritten as
2
d dR (s) q
s s p2 R (s) = s2 R (s)
ds ds r0
" #
2
d dR (s) qs 2
)s s + p R (s) = 0
ds ds r0
" #
2 2
2 d R (s) dR (s) qs 2
)s +s + p R (s) = 0 (13.105)
ds2 ds r0
d2 y (x) dy (x)
x2 2
+x + x2 p2 y (x) = 0: (13.111)
dx dx
exists only for 0 x 1: This is the Bessel DE and its solutions are given by
the Bessel functions
y (x) = ap Jp (x) + bp Np (x) : (13.112)
Noting that the solution for the ' dependent of the di¤erential equation
1 d2 (')
= p2 (13.113)
(') d'2
As one can see from Fig. , this requires p must be an integer, p = n = 1; 2; 3:::.
Therefore the general solution for the Laplace equation is given by
XX q q qs
V (s; '; z) = Cq e r0 z + Dq e r0 z
(An cos (n') + Bn sin (n')) Jn :
q n=1
ro
Note that we have taken into account that, when the constant p is an integer,
p = n; Jn (x) and J n (x) are not independent
using
2
1 d2 Z (z) q 1 d2 (')
= ; = p2
Z (z) dz 2 r0 (') d'2
13.4. LAPLACE’S EQUATION IN CYLINDRICAL COORDINATES 377
we have " #
2
1 d dR (s) q
s2 s + p2 = 0; (13.120)
sR (s) ds ds r0
so that for p = 0 and q = 0; we …nd
s d dR (s) d
dR (s) dR (s)
s = 0) s = 0; ) s =A
R (s) ds ds ds
ds ds
ds
) dR (s) = A ) R (s) = A ln (s) (13.121)
s
Therefore the general solution to Laplace’s equation can be expressed as
V (s; '; 0) = s cos (') ; V (s; '; L) = 0; and V (r0 ; '; z) = 0; (13.125)
we …nd
q q q
Cq e r0 L + Dq e r0 L
= 0 ) Dq = Cq e2 r0 L : (13.130)
or
q q !
X q
L e r0 (L z)
e r0 (L z)
qs
V (s; '; z) = 2Cq e r0
cos (') Jp
q
2 ro
X q qs
) V (s; '; z) = F q sinh (L z) cos (') J1 : (13.132)
q
r0 ro
this requires
J1 (q) = 0 (13.134)
which means
q = qi i = 1; 2; 3::: (13.135)
are the zeros of the Bessel function. Thus the potential can be written as
X qi qi s
V (s; '; z) = Fi sinh (L z) cos (') J1 : (13.136)
i
r0 ro
which leads to
X qi qi s
V (s; '; z = 0) = Fi sinh L cos (') J1 = s cos '
i
r0 ro
X qi qi s
) Fi sinh L J1 = s: (13.139)
i
r0 ro
q s
Multiplying both sides by J1 rjo rso d rso and integrating over s, we
have
X Z r0
qi qi s qj s s s
Fi sinh L J1 J1 d
i
r0 0 ro ro ro ro
Z r0
qj s s s
= sJ1 d : (13.140)
0 ro ro ro
s
and replacing, ro = x; we …nd
X Z 1 Z 1
qi
Fi sinh L J1 (qi x) J1 (qj x) xd (x) = r0 xJ1 (qj x) xd (x)
i
r0 0 0
X Z
qi Jp0 (qi ) Jp0 (qj ) 1
) Fi sinh L ij = r0 xJ1 (qj x) xdx:
i
r0 2 0
Z 1
qj Jp02 (qj )
) Fj sinh L = r0 x2 J1 (qj x) dx
r0 2 0
R1
r0 0 x2 J1 (qj x) dx
) Fj = Jp02 (qj )
(13.141)
qj
sinh r0 L 2
we may write
r0 J2 (qj )
Fj = (13.142)
qj J102 (qj )
qj sinh r0 L 2
or applying
Jp0 (qj ) = JP +1 (qj ) (13.143)
380 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
we …nd
2r0
Fj = : (13.144)
qj
qj sinh r0 L J2 (qj )
Therefore the potential becomes
X 2r0 qj s qj
V (s; '; z) = J1 cos (') sinh (L z) ;
qj sinh
qj ro r0
j r0 L J2 (qj )
where 0 is the electrical permittivity of free space. I will discuss the basic
procedures for solving Poisson’s Equation and applying the boundary conditions
using an example in spherical coordinates.
Approach to Solving Poisson’s Equation: when we studied how to solve in-
homogenous second order linear di¤erential equations, for example
d2 y dy
2
+a + y = f (x) (13.149)
dx dx
13.5. POISSON’S EQUATION 381
we have seen that the solution is sum of the homogenous yh and the particular
yp solutions given by
y = yh + yp : (13.150)
The Homogenous solution, yh ; satis…es the equation
d2 yh dyh
2
+a + yh = 0 (13.151)
dx dx
and the particular solution yp is determined using the di¤erent techniques we
discussed. The same principle is applied in solving Poisson’s equation. For the
Poisson’s equation
(~r)
r2 V (x; y; z) = (13.152)
0
the homogenous solutions, Vh ; is basically is the solution to the Laplace equation
r2 Vh (x; y; z) = 0 (13.153)
We can easily show that this solution satis…es the Poisson’s equation
r2 V h = 0 (13.156)
(~r)
r2 Vp = (13.157)
0
Example 12.5 Point Charge and Grounded Conducting Sphere Note: A point
charge Q is located a distance, a; from the center of a grounded, con-
ducting sphere of radius R, where R < a. Find the electrostatic potential
everywhere outside of the sphere. Assume that the center of the sphere is
at the origin of coordinates, and that the point charge is on the positive z
axis.
382 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
Solution: We want to …nd the electric potential at some point outside the
sphere r > R: Since there is a point charge along the z axis at a distance
a (> R) which is in the region outside the sphere we obviously see that
the charge density is not zero in this region and the Laplace’s equation is
not valid in all regions outside the sphere. Therefore, to …nd the electric
potential we need to …nd the solution to the Poisson’s Equation. From
the nature of the problem it is better to use spherical coordinates
(~r)
r2 V (r; ; ') =
0
1 @ @ 1 @ @
) 2 r2 V (r; ; ') + 2 sin V (r; ; ')
r @r @r r sin @ @
1 @2 (~r)
+ 2 2 V (r; ; ') = : (13.159)
r sin @'2 0
The Homogenous solution for this equation is the general solution to the
Laplace’s equation in spherical coordinates
1 @ @ 1 @ @
r2 Vh (r; ; ') + 2 sin Vh (r; ; ')
r2 @r @r r sin @ @
1 @2
+ 2 2 Vh (r; ; ') = 0 (13.160)
r sin @'2
13.5. POISSON’S EQUATION 383
1 Q
VQ (r; ; ') = : (13.162)
4 0 j~r ~r0 j
For
~r = r sin cos '^
x + r sin sin '^
y + r cos z^
and
~r0 = a^
z;
we have
1=2
j~r ~r0 j = r2 + a2 2ra cos (13.163)
so that
1 Q
VQ (r; ; ') = 1=2
: (13.164)
4 0 (r2 + a2 2ra cos )
To …nd the particular solution we note that the charge outside the sphere
is zero except the point charge at a point on the z axis a distance a from
the origin. Therefore, the particular solution for the Poisson’s equation
(~r)
r2 V (r; ; ') = (13.165)
0
1 Q
Vp (r; ; ') = VQ (r; ; ') = 1=2
: (13.166)
4 0 (r2 + a2 2ra cos )
Bl = 0: (13.170)
we may write
1
1 Q Q X Rl
= Pl (cos ) (13.174)
4 0 (R2 + a2 2Ra cos )
1=2 4 0 al+1
l=0
13.5. POISSON’S EQUATION 385
for R < a and the electric potential on the surface of the sphere becomes
1 X
X l
Cl m
V (R; ; ') = P (cos ) (Am cos (m') + Bm sin (m'))
Rl+1 l
l=0 m= l
1
Q X Rl
+ Pl (cos ) = 0: (13.175)
4 0 al+1
l=0
and write
1
Q X 1 R2l+1 q0 1
l+1 l+1
Pl (cos ) = + p :
4 0 r a 4 0 r +r
2 02
2rr0 cos
l=0
(13.183)
386 CHAPTER 13. A PARTIAL DIFFERENTIAL EQUATION
q0 1 Q 1
V (r; ; ') = p + ; p
4 0 r2 + r02 2rr0 cos 4
2ra cos 0 r2 + a2
(13.184)
0 R2 0 R
where r = a and q = Q a :The above result shows that the electric
potential due to a point charge Q placed outside the sphere at a distance
a from the center of a grounded conducting sphere can be imagined as
sum of the potential due to the point charge Q and a negative "image
R2
charge" q 0 = Q R 0
a located inside the sphere at a distance r = a from
the center of the sphere. In reality this image charge is not a charge
existing at this position. It is the total induced charge on the grounded
conducting sphere due to the electric …eld produced by the point charge Q.
The method of replacing this induced charge by an image charge to …nd
the electric potential is called the method of images. It is a useful method
in determining the electric potential of charges placed around conducting
sphere, in…nitely long conducting cylinder, or in…nitely wide conducting
plate.
Chapter 14
Functions of Complex
Variables
z = a + ib; (14.1)
are called complex numbers. In any complex number, z = a + ib; a is the real
part and b is the imaginary part
Re z = a; Im z = b (14.3)
where r is called the modulus or the magnitude and is called the phase of the
complex number z
Euler’s Equation and Exponential Representation of Complex Numbers :
Any complex number z = x + iy can be expressed in an exponential form
387
388 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
z =x iy (14.8)
2z + 1
f (z) = (14.10)
z i
Solution: To …nd the real and imaginary part of a complex function for the
complex variable
z = x + iy (14.11)
Example 13.2 Find the real and imaginary parts of the complex function
z = r exp (i ) ; (14.16)
1=4 i
f (z) = [r exp (i )] = r1=4 exp = r1=4 cos + ir1=4 sin
4 4 4
(14.18)
so that the real and imaginary parts can be expressed as
@f @u @v @f @u @v
= +i ; = +i (14.27)
@x @x @x @y @y @y
and substituting these expressions into
@f @f
= i (14.28)
@x @y
one …nds
@u @v @u @v @u @v @v @u
+i = i +i ) +i = i (14.29)
@x @x @y @y @x @x @y @y
Two complex function are equal if and only if their real part are equal and their
imaginary part are equal. Therefore,
@u @v @v @u
= ; = : (14.30)
@x @y @x @y
which is The Cauchy-Riemann Conditions.
14.3. IMPORTANT TERMINOLOGIES 391
(a) Regular point: The point zo is said to be a regular point of the function
f (z) if the function f (z) is analytic at that point.
(b) Singular point: If f (z) is not analytic at a point, zo , then zo is said to be
a singular point, or singularity of f (z).
(c) Isolated Singularity: If zo is the only singularity of a function f (z) within
an arbitrarily small region surrounding zo , then zo is said to be an isolated
singularity of the function f (z):
(d) Harmonic and conjugate Harmonic Functions: Consider the …rst Cauchy-
Riemann Condition
@u @v
= : (14.31)
@x @y
Upon di¤erentiating with respect to x; …nd
@2u @2v
2
= : (14.32)
@x @x@y
Similarly, di¤erentiate the second condition
@v @u
= (14.33)
@x @y
with respect to y; we have
Example 13.3 Expand the complex function given below in a Taylor series
about the origin, and …nd the circle of convergence for this series.
p
f (z) = z + 2i (0 2 ) (14.36)
X1
1 dn f (z0 ) n
f (z) = (z z0 ) : (14.37)
n=0
n! dz n
p
For the complex function f (z) = z + 2i, we have
p df (z) 1 1
f (z) = z + 2i; = p
dz 2 z + 2i
d2 f (z) 1 1 d3 f (z) 3 1
= ; = 3 (14.38)
dz 2 2
2 (z + 2i) 3=2 dz 3 2 (z + 2i)5=2
Therefore
p 1 1 1 1 2
f (z) = z0 + 2i + (z z0 ) 3
(z z0 )
2 (z0 + 2i)1=2 2 (z0 + 2i)3=2
1 1 3
+ 4
(z z0 ) ::: (14.39)
2 (z0 + 2i)5=2
where k is the wave number, ! is the angular frequency, E0x and E0y are
the amplitude of the electric …eld in the x and y; direction, respectively.
(a) Express the complex electric …eld using Euler’s formula assuming E0x and
E0y are complex.
14.3. IMPORTANT TERMINOLOGIES 393
(b) Assume the phase of the x component 'x = 0 and that of the y component
'y = ', …nd the components of the electric …eld for' = 0; =2.
(c) Find the real part of the Electric …eld vector for ' = 0; =2 and plot
the corresponding results for Ey vs Ex for a …xed value of z (for example
z = 0) and 0 !t 2 .
Solution:
(a) For complex electric …eld, using Euler’s formula, we may write
so that
~ = (E0x x
E ^ + E0y y^) ei(kz !t)
; (14.42)
becomes
~ = jE0x j ei'x x
E ^ + jE0y j ei'y y^ ei(kz !t)
= jE0x j ei(kz !t+'x )
x
^
+ jE0y j ei(kz !t+'y )
y^:; (14.43)
(b) If we assume the phase of the x component 'x = 0 and that of the y
component 'y = '; as measured relative to the phase of the x component
of the electric …eld, then we can write
~ = jE0x j ei(kz
E !t)
^ + jE0y j ei(kz
x !t+')
y^ (14.44)
which leads to
cos ( =2) = cos ( ) cos ( =2) sin ( ) sin ( =2) = sin ( ) (14.50)
394 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
and
sin ( =2) = sin ( ) cos ( =2) cos ( ) sin ( =2) = cos ( ) (14.51)
and light is linearly polarized. Under this condition if jE0y j = 0 the light
is x-polarized; and if jE0x j = 0; the light is y-polarized.
Ex = jE0x j cos ( !t) + i jE0x j sin ( !t) = jE0x j cos (!t) i jE0x j sin (!t) ; ;
(14.53)
Ey = jE0y j cos ( !t) + i jE0y j sin ( !t) = jE0y j cos (!t) i jE0y j sin (!t) :
Ey = jE0y j sin ( !t) i jE0y j cos ( !t) = jE0y j sin (!t) i jE0y j cos (!t)
) Ey = (jE0y j sin (!t) + i jE0y j cos (!t)) : (14.56)
396 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
Proof: Using
z = x + iy (14.59)
for a complex function
we may write
I I I
f (z) dz = [(u(x; y)dx v (x; y) dy)] + i [(v (x; y) dx + u(x; y)dy)]
C C C
ZZ ZZ
@ [ v] @u @v @u
= dxdy + i dxdy (14.63)
@x @y @x @y
A A
ZZ ZZ
@v @u @u @v
= + dxdy + i dxdy: (14.64)
@x @y @x @y
A A
z zo = (x x0 ) + i (y y0 ) = jz zo j ei : (14.67)
where q
2 3 1 y y0
jz zo j = (x x0 ) + (y y0 ) ; = tan :
x x0
Figure 14.3: A circle centered about z0 : The function, f (z) ; is analytic every-
where inside and on the closed curve, C:
where C 0 is the curve shown in the Fig. 14.4. This integral can be expressed as
I Z Z Z Z
g (z) dz = g (z) dz + g (z) dz + g (z) dz + g (z) dz = 0 (14.71)
C1 l2 C2 l1
C0
Since we are interested in the limit ! 0; in this limit integrating along l1 and
l2 leads to same function integrated in opposite direction over the same line.
400 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
Figure 14.4: A region excluding the singular point z0 for the function, g (z) :
Consequently, Z Z
g (z) dz = g (z) dz (14.72)
l1 l2
so that I Z Z
g (z) dz = g (z) dz + g (z) dz = 0 (14.73)
C1 C2
C0
or I I
f (z) f (z)
dz = dz: (14.74)
z a z a
C1 C2
Note that from Fig. 14.4 the integration over C1 is counterclockwise where
as the integration over C2 is clockwise. We can make the integration over C2
counterclockwise and drop the minus sign.
I I
f (z) f (z)
dz = dz: (14.75)
z a z a
C1 C2
Now using equation of a circle on a complex plane we saw earlier, we may write
z z0 = ei ) z = z0 + ei ) dz = i ei d (14.76)
so that
I Z2 Z2
f (z) f z0 + e i
dz = i ei d = i f z0 + ei d (14.77)
z z0 ei
C2 0 0
14.5. CAUCHY’S INTEGRAL FORMULA 401
which leads to
I I Z2
f (z) f (z)
dz = dz = i f z0 + ei d : (14.78)
z z0 z z0
C1 C2 0
We are interested for the integration over the original curve C. One can recover
this curve in the limit ! 0, where C1 becomes C and f z0 + ei = f (z0 ) :
Thus in this limit one can write
I Z2 I
f (z) 1 f (z)
dz = i f (z0 ) d = 2 if (z0 ) ) f (z0 ) = dz: (14.79)
z z0 2 i z z0
C 0 C
Z 2+4i Z 2+4i
2 2
= x y dx 2xydy + 2i xydx + x2 y 2 dy
1+i 1+i
so that replacing
y = 3x 2 ) dy = 3dx (14.85)
we …nd
Z 2+4i Z 2 h i
2
z 2 dz = x2 (3x 2) dx 2x (3x 2) 3dx
1+i 1
Z 2 h i
2
+i 2x (3x 2) dx + x2 (3x 2) 3dx
1
Z 2 Z 2
2
= 26x + 24x 4 dx + i 18x2 + 32x2 12 dx
1 1
86
= 6i (14.86)
3
(b) the integral along the …rst line from the point z1 = 1 + i to zo = 2 + i, and
the second line from the point zo to the point z2 = 2 + 4i can be expressed
as
Z 2+4i Z 2 Z 4
z 2 dz = x2 1 + 2ix dx + 22 y 2 + 4iy idy
1+i 1 1
2 4
x3 y3 4
= + ix2 x +i + 2iy 2 + 4y = + 3i 30 9i
3 1 3 1 3
Z 2+4i
86
) z 2 dz = 6i (14.87)
1+i 3
Noting that
I Z z0 Z z2 Z z1 Z z0 Z z2 Z z2
2 2 2 2 2 2
z dz = z dz+ z dz+ z dz = z dz+ z dz z 2 dz::
z1 z0 z2 z1 z0 z1
C
(14.91)
and using the result we obtained in part a and b,
I
86 86
z 2 dz = 6i 6i = 0: (14.92)
3 3
C
so that I
ez 1
I= dz = 2 i 1 + e : (14.98)
C z (z + 1)
404 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
2 b1 b2
f (z) = a0 + a1 (z z0 ) + a2 (z z0 ) + ::: + + 2 + ::: (14.99)
z z0 (z z0 )
1
X 1
X
n bn
= an (z z0 ) + n;
n=0 n=1
(z z0 )
we can write [applying the procedure we followed when we proof the Cauchy
integral formula] that I I
f (z) dz = f (z) dz: (14.100)
C C2
We can expand the function f (z) in Laurent series in the shaded region since
the function is analytic in this region
2 b1
f (z) = a0 + a1 (z zk ) + a2 (z zk ) + ::: +
z zk
1
X 1
X
b2 n bn
+ 2 + ::: = an (z zk ) + n (14.101)
(z zk ) n=0 n=1
(z zk )
406 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
This leads to
I I I I
2
f (z) dz = a0 dz + a1 (z zk ) dz + a2 (z zk ) dz
C C2 C2 C2
I I I
b1 b2 b2
+::: + dz + 2 dz + 3 dz::: (14.102)
C2 z zk C2 (z zk ) C2 (z zk )
n
Since the function (z z0 ) is analytic everywhere in the region bounded by
the curve C2 for n = 0, we have
I
n
(z zk ) dz = 0 (14.103)
C2
so that
I I I I
b1 b2 b2
f (z) dz = dz + 2 dz + 3 dz::: (14.104)
C C2 z zk C2 (z zk ) C2 (z zk )
Now using
z zk = ei ) z = zk + ei ) dz = i ei d (14.105)
we have
I I I I
1 b2 b2
f (z) dz = b1 i ei d + 2i
i
e d + 3i ei d :::
C C2 ei C2 ( ei ) C2 ( ei )
(14.106)
14.8. METHODS OF FINDING RESIDUES 407
I 1
X I
ibn i(n 1)
f (z) dz = 2 ib1 + n 1
e d
C n=2 C2
1
X Z 2
ibn
= 2 ib1 + n 1
fcos [(n 1) ] i sin [(n 1) ]g d (14.107)
n=2 0
Since
Z 2 Z 2
cos [(n 1) ] d = i sin [(n 1) ] d = 0 (14.108)
0 0
where
R (zk ) = b1 (14.110)
I n
X
f (z) dz = 2 i R (zk ) : (14.111)
k=1
dn f (x)
If exists for n 0 and x 2 Re domain
dxn
X1
n 1 dn f (x)
) f (x) = bn (x a) , where bn = (14.112)
n=0
n! dxn x=a
408 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
This can be applied for complex z. Taylor Series for some common functions:
X1
( 1)n 2n+1 x3 x5 x7
sin x = x =x + ::: convergent for all x
n=0
(2n 1)! 3! 5! 7!
X1
( 1)n 2n x2 x4 x6
cos x = x =1 + :::convergent for all x
n=0
(2n)! 2! 4! 6!
X1
1 n x2 x3
ex = x =1+x+ + :::convergent for all x
n=0
n! 2! 3!
X1
( 1)n+1 n x2 x3 x4
ln(1 + x) = x =x + :::convergent for 1<x 1;
n=1
n 2 3 4
1
X p p(p 1) 2 p(p 1)(p 2)
(1 + x)p = xn = 1 + px + x + x3 :::
n 2! 3!
n=1
convergent for all jxj < 1 (14.113)
where
p p (p 1) (p 2) (p 3) ::: (p n + 1)
= ; (14.114)
n n!
is called binomial coe¢ cient.
Power series expansions of functions are unique. That is, if you use di¤erent
methods for …nding a power-series expansion of a function, then you must get
the same result. A power series expansion for a function must be the power
series expansion for that function!
Example 13.7 Find R(0) for the complex function
1
f (z) = z cos (14.115)
z
cos (z)
R ( ) = lim [(z ) f (z)] ) R ( ) = lim (z ) (14.137)
z! z! sin (z)
cos (z)
R ( ) = lim [(z ) f (z)] ) R ( ) = lim (z )
z! z! sin (z)
" #
d
[(z ) cos (z)] cos (z) (z ) sin (z)
) R ( ) = lim dz d = lim
z!
dz sin (z) z! cos (z)
cos ( ) ( ) sin ( )
) R( ) = =1 (14.138)
cos ( )
Now if we make a series expansion for the function
cos (z)
f (z) = cot (z) = : (14.139)
sin (z)
14.8. METHODS OF FINDING RESIDUES 411
we …nd
1 z 1 3 2 5
f (z) = (z ) (z ) + :::: (14.140)
z 3 45 945
which indeed shows the function has a simple pole at z0 = :
Method 3 (Generalization of method 2): To …nd the residue of f (z) at zo
when f (z) has a pole of order n at zo , compute
1 dm 1
m
R (zo ) = f(z z0 ) f (z)g ; (14.141)
(m 1)! dz m 1
z=zo
where m > n:
Example 13.12 Find R (3) for
zezt
f (z) = 2 (14.142)
(z 3)
where t is a parameter (possible complex).
Solution: We note that f (z) has a pole at zo = 3 of order n = 2. Thus we can
apply the third method to …nd the residue at zo which we write as
" ( )#
1 d2 1 2 zezt d
R (3) = (z 3) 2 = zezt
(2 1)! dz 2 1 (z 3) dz z=3 z=3
) R (3) = (1 + zt) ezt z=3
= (1 + 3t) e3t : (14.143)
where R (zk ) is the residue of the function f (z) at z = zk : We are going to use
this theorem and the methods of …nding residues to evaluate several di¤erent
types of de…nite integrals. The methods are best shown by examples.
Example 13.14 Evaluate the real integral
Z1
dx
I= (14.147)
x4+1
0
I Z R Z
dz dx dz
= + (14.149)
z4 + 1 R x4 + 1 C z4 + 1
14.9. APPLICATIONS OF THE RESIDUE THEOREM 413
where 0 ; then
I Z R Z
dz dx iR exp[i ]d
4
= 4
+ : (14.151)
z +1 R x +1 0 R4 exp[4i ] + 1
Now if we let R ! 1, we can see that the integral over the curved part
(the second integral) approaches to zero and we …nd that
I Z 1
dz dx
= : (14.152)
z4 + 1 1 x4 + 1
1
f (z) = (14.154)
z4 + 1
on the closed curve de…ned by this semicircle of in…nite radius, using the
Residue theorem, can be expressed as
I n
X
f (z) dz = 2 i R (zk ) : (14.155)
k=1
In order to …nd the residues we …rst need to get the poles. Noting that
h i1=4
1=4
z 4 + 1 = 0 ) z = ( 1) = ei( +n(2 ))
; (14.156)
where m > n: We note that the complex function f (z) can then be ex-
pressed as
1 1
f (z) = =
z4 + 1 (z z1 ) (z z2 ) (z z3 ) (z z4 )
1
= 3 5 7
(14.160)
z ei 4 z ei 4 z ei 4 z ei 4
3
This shows that the two poles at z1 = ei 4 and z2 = ei 4 are simple poles
(of order one) and we can use the second method to …nd the residues at
these poles
R (zo ) = lim [(z z0 ) f (z)] : (14.161)
z!zo
i4
For z1 = e , we have
z ei 4
R ei 4 = lim
z!ei 4 z ei 4 z e i 34 z ei
5
4 z ei
7
4
1
=
z e i 34 z ei
5
4 z ei
7
4
so that using
3 p 5 p 7 p
ei 4 ei 4 = 2; ei 4 ei 4 = 2 (1 + i) ; ei 4 ei 4 = i 2 (14.162)
the residue becomes
1 1
R ei 4 = p p p = p : (14.163)
2 2 (1 + i) i 2 2 2 (1 i)
3
Similarly, the residue at the second pole zo = ei 4 is given by
3
z ei 4
i 34
R e = lim3 3 5 7
z!ei 4 z ei 4 z ei 4 z ei 4 z ei 4
1
= 5 7
(14.164)
z ei 4 z ei 4 z ei 4
14.9. APPLICATIONS OF THE RESIDUE THEOREM 415
and using
3 p 3 5 p 3 7 p
ei 4 ei 4 = 2; ei 4 ei 4 = i 2; ei 4 ei 4 = 2 ( 1 + i)
(14.165)
we …nd
3 1 1
R ei 4 = p p p = p : (14.166)
2i 2 2 ( 1 + i) 2 2 (i + 1)
Therefore
Z 1 I Z 1
dx 1 dz dx
4
= 4
) = p : (14.169)
0 x +1 2 z +1 0 x4 +1 2 2
Mathematica Result:
Z2
cos (3 )
I= d (14.170)
5 4 cos ( )
0
1 1 1
cos ( ) = (exp (i ) + exp ( i )) = exp (i ) +
2 2 exp (i )
1 1 1 z2 + 1
= z+ = (14.171)
2 z 2 z
416 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
and
" #
1 1 3 1
cos (3 ) = (exp (3i ) + exp ( 3i )) = (exp (i )) + 3
2 2 (exp (i ))
1 1 1 z6 + 1
= z3 + ) cos (3 ) = ; (14.172)
2 z3 2 z3
1 z 6 +1
cos (3 ) 2 z3 z6 + 1 z6 + 1
= 2(z 2 +1)
= =
5 4 cos ( ) 5 2z 2 [5z 2 (z 2 + 1)] 2z 2 [2z 2 5z + 2]
z
(14.173)
For z = exp (i ) ; noting that
dz
dz = i exp (i ) d ) d = i exp ( i ) dz ) d = i ; (14.174)
z
the integral becomes
Z2 I
cos (3 ) z6 + 1 dz
I = d = i
5 4 cos ( ) 2z 2 [2z 2 5z + 2] z
0
I Z2 I
i z 6 + 1 dz cos (3 )
= ) d = f (z) dz;
(14.175)
2z 3 [2z 2 5z + 2] 5 4 cos ( )
0
where
i z6 + 1
f (z) = (14.176)
2z 3 (2z 2 5z + 2)
and the closed curve is a unit circle centered about the origin shown in
the …gure below. We can then use the Residue theorem
I n
X
f (z) dz = 2 i R (zk ) (14.177)
k=1
to …nd the integral. To this end, we …rst …nd out the residues inside the
unit circle. Noting that
2z 2 5z + 2 = (2z 1) (z 2) ) 2z 2 5z + 2 = 2 (z 1=2) (z 2)
(14.178)
we may write
i z6 + 1
f (z) = 3 : (14.179)
4 (z 0) (z 1=2) (z 2)
This expression shows that the function f (z) has poles z1 = 0; z2 =
1=2; z3 = 2: However, only the …rst two poles (z1 = 0; z2 = 1=2) are
14.9. APPLICATIONS OF THE RESIDUE THEOREM 417
R=1
inside the unit circle. Thus we need only the residue of these two poles.
The second pole z2 = 1=2 is a simple pole since the order is one and the
residue can be expressed as
" #
1
i z6 + 1 i 64 +1
R (1=2) = lim [(z 1=2) f (z)] = lim 3
= 1 1
z!1=2 z!1=2 4z (z 2) 2 2 2
65 4 65
=i ) R (1=2) = i : (14.180)
64 3 48
1 dm 1
m
R (zo ) = f(z z0 ) f (z)g ; (14.181)
(m 1)! dz m 1
z=zo
( )
d3 1
1 z3i z6 + 1
R (0) =
(3 1)! dz 3 1 2z 3 (2z 2 5z + 2)
z=0
( ) ( )
1 d2 i z6 + 1 1 d2 i z6 + 1
= =
2 dz 2 2 (2z 2 5z + 2) 2 dz 2 2 (2z 2 5z + 2)
z=0 z=0
(14.182)
418 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
"
1 6iz 5 (4z 5) 15iz 4
R (0) = 2 +
2 (2z 2 5z + 2) 2z 2 5z + 2
2
#
i 1 + z 6 (4z 5) 2i 1 + z 6
+ 3 2
(2z 2 5z + 2) (2z 2 5z + 2) z=0
!
2
i ( 5) 2 i 25 1 21i
R (0) = = = (14.183)
2 23 2 2 2 8 2 16
Mathematica result:
where R (z0 ) is the residue of the function f (z) at z = zk : If the are more than
one residue, the residue theorem states that
I Xn
f (z) dz = 2 i R (zk ) (14.186)
k=1
Here the poles zk must be inside the closed curve over which the contour integral
is being carried on. What if there are poles exactly on the curve. Which means
the function f (z) is not analytic at zk0 that is/are exactly located somewhere on
the curve. For example, let’s say we are interested to evaluate the real integral
Z 1
dx
I1 = 2
: (14.187)
1 x 1
To solve this integral we need to consider the closed curve shown in the …gure
below and evaluate the contour integral
I
I2 = f (z) dz; (14.188)
where
1
f (z) = : (14.189)
z2 1
420 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
But we know that f (z) has a simple poles at z10 = 1 and z20 = 1 which are
exactly on the contour integral not inside the contour integral. So how can we
evaluate I1 by directly using the residue theorem. We can evaluate such kind of
integral using two di¤erent approaches. We need to either put the poles inside
or outside the contour by modifying the semicircular contour if we are going to
use the residue theorem directly. As we will see either of these approach lead to
the same result which result in a modi…cation of the residue theorem.
where zk is the pole inside the contour and zk0 is the simple pole on the contour.
Proof :: First Approach (Put the pole inside the curve): The contour shown
in the …gure below keeps the pole inside the contour. For now let’s assume there
are no other poles inside this contour except z0 = z10 and z0 = z20 so that
I X2
f (z) dz = 2 i R (zk0 ) : (14.191)
k=1
which gives
eiz
R (0) = lim z =1 (14.198)
z!0 z
Therefore the integral becomes
I
eiz
dz = i (14.199)
z
we may write
Z1 Z1
cos (x) sin (x)
dx + i dx = i (14.201)
x x
1 1
which leads to
Z1 Z1
sin (x) cos (x)
dx = ; dx = 0 (14.202)
x x
1 1
sin(x)
Since x is an even function, we …nd
Z 1
sin (x)
dx = : (14.203)
0 x 2
Fs = kx:
In addition to this force, the electron can also experience some damping force
Fd which is proportional to the speed of the electron
dx
Fd = m :
dt
If the electron is exposed to an EM wave with frequency ! and polarized in the
x-direction
E~ = E0 cos (!t) x
^
14.12. THE REVISED RESIDUE THEOREM 423
Then using newtons second law, the net force acting on the electron can be
written as
d2 x dx
m = qE0 cos (!t) kx
dt2 dt
d2 x dx
)m 2 + m + kx = qE0 cos (!t)
dt dt
d2 x dx k qE0
) 2 + + x= cos (!t)
dt dt m m
or
d2 x dx qE0
+ + ! 20 x = cos (!t)
dt2 dt m
where r
k
!0 =
m
is the natural frequency of the electron. In terms of the complex variables x
~
~ we may write this equation as
and E,
d2 x
~ d~
x qE0
+ + ! 20 x
~= exp ( i!t) (14.204)
dt2 dt m
The Homogenous solution to this di¤erential equation is given by
t=2
x
~H (t) = e (A cos ( t) + B sin ( t)) :
x
~p (t) = x
~0 exp ( i!t)
m! 2 x
~0 i !m~
x0 + k~
x0 = qE0 (14.205)
which gives
m! 2 x
~0 x0 + m! 20 x
im !~ ~0 = qE0
qE0 =m
)x
~0 = (14.206)
! 20 !2 i !
qE0 =m
x
~p (t) = exp ( i!t)
! 20 !2 i !
424 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
x
~ (t) = x
~H (t) + x
~p (t)
t=2
x
~ (t) = e (A cos ( t) + B sin ( t))
qE0 =m
+ 2 exp ( i!t)
!0 !2 i !
But we are interested in the steady state of the electron which happen if we
waited long enough (i.e. t ! 1). Thus for steady state the complex displace-
ment of the electron becomes
qE0 =m
x
~ (t) ' exp ( i!t)
! 20 !2 i !
which shows at steady state the electron begins to oscillates with frequency of
the EM …eld as we would expect. The complex dipole moment of the electron
can then be expressed as
q 2 E0 =m
p~ (t) = q~
x (t) = exp ( i!t)
! 20 !2 i !
This can also be put in the form
q 2 E0 =m ! 20 !2 + i !
p~ (t) = 2 2
exp ( i!t)
(! 20 ! 2 ) + ( !)
) p~ (t) = j~
p0 j exp [ (i!t ')]
where
q 2 E0 =m
p0 j = q
j~
2 2
(! 20 ! 2 ) + ( !)
and
1 !
' = tan :
! 20 !2
This means that p is out of phase by ' with respect to the electric …eld E. ~
Lagging behind by ' which is very small when ! ! 0 and rises to when
! ! 0 . If there are N molecules per unit volume and fj electrons per molecule.
If these electrons in the j th molecule is oscillating with the natural frequency
! j and damped by j ; then the total dipole moment of the electrons per unit
volume (the polarization P~ ) can be expressed as
N q2 X fj
P~ = E0 exp ( i!t)
m j=1 ! 2j !2 i j !
N q2 X fj
P~ = ~
E
m j=1 ! 2j !2 i j !
14.12. THE REVISED RESIDUE THEOREM 425
P~ = 0 eE
~
where
N q2 X fj
e = 2 :
0 m j=1 ! j !2 i !
~= 0 (1 + e)
~ N q2 X fj
~r = =1+ e =1+ 2 (14.207)
0 0 m j=1 ! j !2 i j !
Eq. (14.207) shows the dielectric constant (the permittivity of the medium)
depends on the frequency of the EM wave-the medium is dispersive. Still wave
equation is satis…ed by the electric and magnetic …elds. For a dispersive medium
it is given by
2 ~
r2 E ~ = ~ 0@ E
@t2
with a plane wave h i
E~=E ~0 exp i kz~ !t :
If we express it as
k~ = kRe + ikIm
The electric …eld may be put in the form
~=E
E ~0 exp ( kIm z) exp [i (kRe z !t)] :
and the intensity which is proportional to the square of the amplitude of the
electric …eld becomes
I/ E ~0 exp ( 2kIm z)
which shows a damping in the …eld amplitude due to the absorption by the
medium. For that reason the absorption coe¢ cient of the medium is given by
= 2kIm :
For gases the second terms is very small and we can make a binomial expansion
for the square root which gives
0 1
2 X
! N q f j
k~ ' @1 + A:
c 2 0 m j=1 ! 2j ! 2 i j !
Then we can …nd the real and imaginary part of the wave number to be
0 1
2 X f ! 2
! 2
!@ Nq j j A
kRe ' 1+
c 2 0 m j=1 ! 2 ! 2 2 + 2
j j!
N q2 !2 X fj j
kIm ' 2 2
2c 0 m j=1 ! 2 !2 +
j j!
so that the absorption coe¢ cient and the refractive index of the medium can be
expressed as
N q2 !2 X fj j
= 2 2
0 m j=1 ! 2 !2 + !
j j
N q2 X fj ! 2j ! 2
n=1+
2 0 m j=1 ! 2 ! 2 2 + 2
j j!
N q2 !2 fj j
= 2 2
0m ! 2j ! 2 + j!
N q2 fj ! 2j ! 2
n=1+
2 0 m !2 !2 2 + 2
j j!
we can write
N q 2 ! 3j 2
= 4
0 m! j 1 2 2 2
+( )
2
N q 2 ! 2j fj 1
n=1+
2 0 m! 4j 1 2 2
+(
2
)
2
N q2
=
0 m! j 1 2 2 2
+( )
2
N q 2 fj 1
n=1+
2 0 m! 2j 1 2 2
+(
2
)
2
= 1
2 2 2
1 +( )
2
1
n=1+ 2
2 2 2
1 +( )
where
N q2 N q 2 fj 1 fj
1 = ; 2 = =
0 m! j 2 0 m! 2j 2! j
The plot for such particular molecule the absorption coe¢ cient is shown in
…gure below. It shows that as the damping constant increases the range of
the spectrum that would be absorbed by the medium increases with a peak
absorption at resonance when the frequency of the EM wave is the same as the
natural frequency of the electrons (i.e. = 1 ) ! = ! j ).
On the other hand the refractive index shown in the …gure below shows
unusual behavior near the resonant region. In this region the refractive index
decreases in contrary what we know in optics. The range of the spectrum where
this unusual behavior is observed increases with increase in damping e¤ect.
Comparison of the absorption coe¢ cient and the refractive index is shown
in the …gure below. The curves shown by the dotted lines are for the absorption
and the solid line are for the refractive index. In the immediate neighborhood
of a resonance the index of refraction drops sharply. The material may be
practically opaque in this frequency range since it coincides with the region of
maximum absorption. This is because of the electron are forced to oscillate
with their favorite frequency and the amplitude of oscillation is maximum and
correspondingly high amount of energy is dissipated by the damping mechanism.
Because this behavior is atypical, it is called anomalous dispersion (the region
! 1 < ! < ! 2 ).
Refractive index increases for ! < ! 1 and ! > ! 2 which is consistent with
our experience from optics (dispersion). The most familiar example of dispersion
is probably a rainbow, in which dispersion causes the spatial separation of a
white light into components of di¤erent wavelengths (di¤erent colors).
428 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
N q2 X fj
n=1+
2 0 m j=1 ! 2j ! 2
For transparent materials, the nearest signi…cant resonance typically lie in the
ultraviolet, then ! < ! j : Taking this into account and noting that we are very
far from resonance we can make the approximation
!
1 1 1 !2
= ' 2 1+ 2
! 2j ! 2 ! 2 1 !2
2 !j !j
j !j
14.12. THE REVISED RESIDUE THEOREM 429
so that
!
N q2 X 1 !2
n=1+ 1+ 2
2 0 m j=1 ! 2j !j
0 1 0 1
2 X 2 X
N q f j N q f j
)n=1+@ A + !2 @ A
2 0 m j=1 ! 2j 2 0 m j=1 ! 4j
! = 2 c=
we can write
0 1 0 1
Nq X2
fj A 1 2 2 2
c Nq X
2
fj A
n=1+@ + 2@
2 0 m j=1 ! 2j 0 m !
j=1 j
4
or
B
n=1+A 1+ 2
This is known as Cauchy’s Formula: the constant A is called the coe¢ cient of
Refraction and B is called the coe¢ cient of Dispersion. It applies in most gases,
in the optical region.
The Kramers-Kronig Relations: Consider the complex function
which is a complex function !; where u (!) and v (!) are real. Suppose this
function is analytic in the upper half-plane of ! and it vanishes faster than
1= j!j as j!j ! 1: The Kramers-kronig relations are given by
Z 1
1 f (!)
f (! 0 ) = P d!
i 1 ! !0
Z 1
1 u (!) + iv (!)
) u (!) + iv (!) = P d!
i 1 ! !0
) u (!) + iv (!)
Z 1 Z 1
1 v (!) 1 u (!)
= P d! i P d!
1 ! !0 1 ! !0
which leads to
Z 1 Z 1
1 v (!) 1 u (!)
u (!) = P d!; v (!) = P d!
1 ! !0 1 ! !0
Laplace transform
The function F (p) is called the (integral) transform of f (t) by the kernel
K (p; t) : Here we are interested in two types of integral transforms: the Laplace
and Fourier transforms. Integral transform operation may also be described as
mapping a function f (t) in t-space into another function F (p) in p-space. This
interpretation takes on physical signi…cance in the time-frequency relation of
Fourier transforms.
Laplace Transforms: the Laplace transform F (p) or L of a function f (t) is
de…ned by
Z b
F (p) = L [f (t)] = lim e pt f (t) dt: (15.2)
b!1 0
Note:
(i) The Laplace transform for the function f (t) exists even if the integral
for this function, Z 1
I= f (t) dt (15.3)
0
does not exist.
(ii) For a function f (t) the Laplace transform to exist there must be a
positive constant M such that
s0 t
e f (t) M (15.4)
433
434 CHAPTER 15. LAPLACE TRANSFORM
(iii) The Laplace transform fails to exist when the function has a strong
singularity as t ! 0
(iv Laplace Transform is linear. That means if the Laplace transform for
the two functions f (t) and g (t) exist, then we can write that
Example 14.2 Find the Laplace transform of the following functions and spec-
ify the conditions [if there is any speci…c condition that must be satis…ed
to do the transform]
(a)
f (t) = 1 (15.10)
(b)
f (t) = e!t (15.11)
(c)
f1 (t) = cosh (!t) ; f2 (t) = sinh (!t) (15.12)
(d)
f (t) = cos (!t) ; f (t) = sin (!t) (15.13)
Solution:
435
Example 14.2 For the unit step function, also called the Heaviside function,
de…ned by
0 for t < a
U (t a) = : (15.24)
1 for t > a
Find
L [U (t a)] : (15.25)
Example 14.3 Find the Laplace transform for the …rst, second, and third
derivative of the function, dfdt(t)
(a)
df (t)
L [f 0 (t)] = L ; (15.28)
dt
(b)
d2 f (t)
L [f 00 (t)] = L ; (15.29)
dt2
(c)
h 000 i d3 f (t)
L f (t) = L ; (15.30)
dt3
Solution:
we have
1 Z 1
df (t) df (t) df (t)
L [f 00 (t)] = e pt +p e pt dt =
dt 0 0 dt dt t=0
Z 1
df (t)
+p e pt dt ) L [f 00 (t)] = pL [f 0 (t)] f 0 (0) (15.35)
0 dt
so that using the result we found for the …rst derivative of the function in
part (a),
L [f 0 (t)] = pL [f (t)] f (0) (15.36)
we …nd
L [f 00 (t)] = p2 L [f (t)] pf (0) f 0 (0) : (15.37)
1 4p 3 4p 3
1 4p 3
L =L =L 1 2
p2 + 4 p2 + 22 p + 22 p2 + 22
4p 3
=L 1 2 L 1 2
p + 22 p + 22
1 4p 3 p 3 1 2
)L = 4L 1 2 L : (15.46)
p2 + 4 p + 22 2 p2 + 2 2
Recalling that
p !
L [cos (!t)] = ; L [sin (!t)] = 2 (15.47)
p2 + ! 2 p + !2
so that
1 4p 3 3
L = 4 cos (2t) sin (2t) (15.48)
p2 + 4 2
Mathematica Result:
15.1. INVERSE LAPLACE TRANSFORM 439
1 1 1
L [L [f (t)]] = L (15.53)
2 p
which leads to
Z 1
1 sin (tx)
L [L [f (t)]] = ) f (t) = dx = for t > 0 (15.54)
2 0 x 2
Noting that for
sin ( tx) = sin (tx) (15.55)
440 CHAPTER 15. LAPLACE TRANSFORM
where U (t) is the unit step function (Heaviside function) we saw in above,
which can be expressed as
0 for t < 0
U (t) = :
1 for t > 0
Example 14.5 Solve the following di¤erential equation given the initial condi-
tions y (0) = 1 and y 0 (0) = 0
d2 y
+ y (x) = 1 (15.60)
dx2
we obtain
d2 y
L = p2 L [y (x)] p: (15.64)
dx2
Then the di¤erential equation becomes
1 1
p2 L [y (x)] p + L [y (x)] = ) p2 + 1 L [y (x)] = + p
p p
1
) L [y (x)] = (15.65)
p
There follows that
1 1
y (x) = L =1 (15.66)
p
Example 14.6 Solve the following di¤erential equation given the initial condi-
tions, y (0) = 2 and y 0 (0) = 1
d2 y dy t
+3 + 2y (t) = 2e (15.67)
dt2 dt
Solution: Take the Laplace transform of each term and apply the given initial
conditions
d2 y
L = p2 L [y (t)] py (0) y 0 (0) = p2 L [y (t)] 2p + 1
dt2
dy
L = pL [y (t)] y (0) = pL [y (t)] 2
dt
1
L e t =L e t = : (15.68)
p+1
The Laplace transform of the di¤erential equation can then be expressed
as
d2 y dy
L + 3L + 2L [y (t)] = 2L e t (15.69)
dt2 dt
442 CHAPTER 15. LAPLACE TRANSFORM
becomes
1
p2 L [y (t)] 2p + 1 + 3 (pL [y (t)] 2) + 2L [y (t)] = 2
p+1
2
) p2 + 3p + 2 L [y (t)] 2p 5=
p+1
2
) (p + 2) (p + 1) L [y (t)] = 5 + 2p +
p+1
5 + 2p 2
) L [y (t)] = + (15.70)
(p + 2) (p + 1) (p + 2) (p + 1)2
2p2 + 7p + 7
L [y (t)] = 2: (15.71)
(p + 2) (p + 1)
Noting that
2p2 + 7p + 7 A B C
2 = + +
(p + 2) (p + 1) p + 2 p + 1 (p + 1)2
A p2 + 2p + 1 + B p2 + 3p + 2 + C(p + 2)
= 2
(p + 2) (p + 1)
(A + B) p2 + (2A + 3B + C) p + A + 2B + 2C
= 2 (15.72)
(p + 2) (p + 1)
A + B = 2; 2A + 3B + C = 7; A + 2B + 2C = 7 (15.73)
we can write
2p2 + 7p + 7 1 1 2
L [y (t)] = 2 = + + : (15.74)
(p + 2) (p + 1) p + 2 p + 1 (p + 1)2
Using Mathematica:
15.2. APPLICATIONS OF LAPLACE TRANSFORMS 443
Noting that
1 1 2t 1 1 t
L =e ;L =e (15.76)
p+2 p+1
and
2 d 1
2 = 2
(p + 1) d p+ =1
" #
1 2 1 d 1
)L 2 =L 2
(p + 1) d p+ =1
t
d 1 1 de t
= 2 L = 2 = 2 2te =1
d p+ =1 d =1
" #
1 2 t
)L 2 = 2te (15.77)
(p + 1)
becomes
2t t t 2t
y (t) = e +e + 2te =e 1 + et (1 + 2t) (15.79)
Using Mathematica:
Example 14.7 Consider the circuit shown in the …gure below. At t = 0 the
switch S is closed to position A. Find the resulting current, I(t). Use
R = 10 ; L = 2H; " (t) = 50 sin (5t)
dI dI
" (t) IR L = 0 ) 50 sin (5t) 10I 2 = 0: (15.80)
dt dt
The Laplace transform can be written as
dI
25L [sin (5t)] 5 [I] L = 0: (15.81)
dt
Using
5 dI
L [sin (5t)] = ;L = pL [I (t)] I (0) (15.82)
p2 + 52 dt
we …nd
125
5L [I] pL [I (t)] + I (0) = 0
p2
+ 52
1 125 125 I (0)
) L [I (t)] = 2 2
+ I (0) = 2 2
+ :
p+5 p +5 (p + 5 ) (p + 5) p + 5
(15.83)
15.2. APPLICATIONS OF LAPLACE TRANSFORMS 445
Noting that
125 A BP + C Ap2 + 25A + BP 2 + 5Bp + Cp + 5C
= + 2 =
(p2 2
+ 5 ) (p + 5) p+5 p +5 2 p2 + 5 2
(A + B) p2 + (5B + C) p + 25A + 5C
=
p2 + 52
A + B = 0; 5B + C = 0; 25A + 5C = 125
5 5 25
) A = ;B = ;C = (15.84)
2 2 2
one can express
125 5 1 5 p 25 1
= + (15.85)
(p2 + 52 ) (p + 5) 2p+5 2 p2 + 52 2 p2 + 52
Just to check using Mathematica:
1 5 1 5 P
I (t) = L [L [I (t)]] = + I (0) L 1 L 1
2 p+5 2 p2 + 52
5 5
+ L 1 2 (15.87)
2 p + 52
Referring to out table
f (t) 1 e!t sin (!t) cos (!t) sinh (!t) cosh (!t) U (t a)
p p ap
1 1 ! ! e
L [f (t)] p p ! p2 +! 2 p2 +! 2 p2 ! 2 p2 ! 2 p
we …nd
5 5 5
I (t) = + I (0) e 5t cos (5t) + sin (5t) : (15.88)
2 2 2
In RL circuit the inductor acts as an open circuit at the initial time and
we can set I (0) = 0
5 5t 5
I (t) = e + [sin (5t) cos (5t)] (15.89)
2 2
Using Mathematica:
446 CHAPTER 15. LAPLACE TRANSFORM
where
k b2
! 21 = (15.92)
m 4m2
448 CHAPTER 15. LAPLACE TRANSFORM
Chapter 16
f (x) = f (x T)
for all x.
Average Values: The average value of the function f (x) in the interval,
a < x < b, is given by
Z b
1
hf (x)i = f (x) dx
b a a
for a periodic function with a period T , we can write the average value in one
period as
Z
1 T =2
hf (x)i = f (x) dx:
T T =2
449
450 CHAPTER 16. FOURIER SERIES AND TRANSFORM
Example 16.1 Compute hf (x)i for f (x) = sin(x) over the interval 0 < x < .
Solution: Using the expression
Z b
1
hf (x)i = f (x) dx (16.1)
b a a
we …nd
Z
1 1 2
hf (x)i = sin(x)dx = cos(x) ) hf (x)i = : (16.2)
0 0
Example 16.2 Compute hf (x)i for f (x) = sin(x) over the interval <x<
. That is, …nd hf (x)i over one complete period.
Solution: Here we have changed the interval to <x< and the average
should be expressed as
Z
1 1
hf (x)i = sin (x) dx = cos(x) =0 (16.3)
( ) 2
Example 16.3 The current in an RLC circuit with an oscillating voltage source
(“emf”) of the form
V (t) = Vm sin (!t) (16.4)
is given by
I (t) = Im sin (!t ') (16.5)
where Vm = Im Z. The quantity Z, called the total impedance of the
circuit, is given by s
2
1
Z = R2 + !L (16.6)
!C
for a series circuit. Find an expression for the root-mean-square current,
Irms , over one full period.
Solution: The root mean square current, Irms ; is the square root of the average
value for I 2 (t) in one full period, T . That means
p
Irms = hI 2 (t)i: (16.7)
So …rst we need to …nd I 2 (t) :To this end, we recall the angular frequency
! = 2 f = 2T ; and the period, T , can be expressed in terms of the angular
frequency, !; as
2 1 !
T = ) = : (16.8)
! T 2
Then the average, I 2 (t) in the interval 0 < t < T , given by
Z T
1
I 2 (t) = I 2 (t) dt (16.9)
T 0
16.1. INTRODUCTION TO FOURIER SERIES 451
can be expressed as
Z T
2 ! 2
I (t) = [Im sin (!t ')] dt: (16.10)
2 0
t = 0)x= ';
2
t = T ) x = !T '=! '=2 ' (16.12)
!
we have
Z 2 ' 2 Z 2 '
! dx Im
I 2 (t) = 2
Im sin2 (x) = sin2 (x) dx (16.13)
2 ' ! 2 '
Using the
we may write
2 2 '
Im x 1
I 2 (t) = sin (2x) (16.14)
2 2 4 '
we have
2
Im 2 ' 1
I 2 (t) = + sin (' 2 ) cos (' 2 )
2 2 2
2
Im ' 1
+ sin (') cos (') : (16.18)
2 2 2
so that using the periodicity of the sine and cosine function with period
2
sin (') = sin (' 2 ) ; cos (') = cos (' 2 ) (16.19)
we …nd
2
Im
I 2 (t) =
: (16.20)
2
Then the root mean square current, Irms ; is the square root of the average
value for I 2 (t) in one full period, T , becomes
p Im
Irms = hI 2 (t)i = p : (16.21)
2
Z
1
bn = f (x) sin (nx) dx; n = 1; 2; 3::: (16.24)
a0 = a0 ei0 x
(16.26)
1
X 1
X 1
X
1 an ibn an + ibn
f (x) = a0 ei0 x
+ e inx
+ e inx
= cn einx ;
2 n=1
2 n= 1
2 n= 1
(16.27)
where
8 1
< 2 a0 n=0
an ibn
cn = 2 n 1 (16.28)
: an +ibn
2 n 1
Z Z
an ibn 1 1 inx
= an = f (x) [cos (nx) i sin (nx)] dx = f (x) e dx;
2 2 2
n = 0; 1; 2; 3:::;
Z Z
an + ibn 1 1
= f (x) [cos (nx) + i sin (nx)] dx = f (x) einx dx;
2 2 2
n = 1; 2; 3::: (16.29)
which leads to
Z
1 inx
cn = f (x) e dx; where n = ::: 3; 2; 1; 0; 1; 2; 3:::
2
Example 16.4 Find the Fourier series expansion for the function
Solution: Using
Z
1
an = f (x) cos (nx) dx; n = 0; 1; 2; 3::: (16.31)
454 CHAPTER 16. FOURIER SERIES AND TRANSFORM
we have
2 3
Z0 Z
14
an = cos (nx) dx + cos (nx) d5
0
2 3 2 0 3
Z0 Z Z Z
1 1
= 4 cos ( nx) d ( x) + cos (nx) d5 = 4 cos (nx) dx + cos (nx) d5
0 0
2 3
Z Z
14
= cos (nx) dx + cos (nx) d5
0 0
) an = 0; n = 0; 1; 2; 3::: (16.32)
And using
Z
1
bn = f (x) sin (nx) dx; n = 1; 2; 3:::
2 0 3
Z Z
14
= sin (nx) dx + sin (nx) d5
0
2 3 2 3
Z0 Z Z0 Z
14 14
= sin ( nx) d ( x) + sin (nx) d5 = sin (nx) dx + sin (nx) d5
0 0
2 3
Z Z
14
= sin (nx) dx + sin (nx) d5
0 0
Z n
2 2 cos (nx) 2 (1 ( 1) )
= sin (nx) dx = =
n 0 n
0
4 4
n n = 1; 3; 5:::
) an = = ; n = 0; 1; 2; 3:: (16.33)
0 n = 2; 4; 6::: (2n + 1)
is given by
1
4 X sin [(2n + 1) x]
f (x) = : (16.35)
n=0
2n + 1
16.3. DIRICHLET CONDITIONS 455
(a) f (x) must have a …nite number of extrema (maxima, minima) in any given
interval (period)
(b) f (x) must have a …nite number of discontinuities in any given interval
(c) f (x) must be absolutely integrable over a period.
(d) f (x) must be bounded within that period (that is, it doesn’t become in…-
nite)
to the given function for all values of x where the function is continuous, and
will converge to the average value of the function at any value of x where the
function has a discontinuity.
Dirichlet’s Theorem for 1-Dimensional Fourier Series: If f (x) satis…es Dirich-
let conditions, the Fourier series will converge to the given function for all values
of x where the function is continuous, and will converge to the average value
of the function at any value of x where the function has a discontinuity. This
means
X1 X1
1 1
a0 + [an cos (nx) + bn sin (nx)] = cn einx = (f (x+ ) + f (x ))
2 n=1 n= 1
2
(16.36)
where the notation
Example 15.5 Verify Dirichlet’s Theorem for the function in the previous ex-
ample
1 for 0 < x <
f (x) = (16.38)
1 for <x<0
(a) f (x) have a …nite number of extrema (maxima, minima) in any given in-
terval, <x< :
Recalling that
1
X 1
inx 4 X sin [(2n + 1) x]
cn e = (16.40)
n= 1 n=1
2n + 1
for x = 0; we have
1
X
cn einx = 0: (16.41)
n= 1
16.4. FOURIER SERIES WITH SPATIAL AND TEMPORAL ARGUMENTS457
Noting that
f (x+ ) = 1; f (x ) = 1 (16.42)
we …nd
1
(f (x+ ) + f (x )) = 0: (16.43)
2
Therefore, we can conclude that
1
X 1
cn einx = (f (x+ ) + f (x )) : (16.44)
n= 1
2
Z
1
bn = f (x) sin (nx) dx; n = 1; 2; 3::: (16.46)
Z=2
2
bn = f (x) sin (nkx) dx; n = 1; 2; 3::: (16.48)
=2
we can write
1
cos (n!t) sin (!t) = fsin [(n + 1) !t] sin [(n 1) !t]g (16.57)
2
1
sin (n!t) sin (!t) = fcos [(n 1) !t] cos [(n + 1) !t]g
2
so that
8 9
>
<ZT =2 T =2
Z >
=
Vm
an = sin [(n + 1) !t] dt sin [(n 1) !t] dt ;
T >
: >
;
0 0
8 9
>
<ZT =2 T =2
Z >
=
Vm
bn = cos [(n 1) !t] dt cos [(n + 1) !t] dt (16.58)
:
T >
: >
;
0 0
For n = 0
T =2
Z
2 Vm 2 Vm Vm
a0 = sin [!t] dt = = (16.59)
T T!
0
for n = 1
T =2
Z
Vm Vm Vm
a1 = sin [2!t] dt = = (16.60)
T 2T !
0
T =2
Z
Vm Vm Vm Vm Vm
b1 = cos [2!t] dt = = (16.61)
:
2T T 2T 2T ! 4
0
where we used
2
!= ) T! = 2 (16.62)
T
For n > 1
T =2
Vm cos [(n + 1) !t] cos [(n 1) !t]
an = +
T (n + 1) ! (n 1) ! 0
Vm 1 cos [(n + 1) !T =2] 1 cos [(n 1) !T =2]
= +
T! n+1 n+1 n 1 n 1
Vm 2 cos [(n + 1) ] cos [(n 1) ]
= +
2 n2 1 n+1 n 1
n
Vm 2 cos (n ) cos (n ) Vm 1 ( 1)
= + = +
2 n2 1 n+1 n 1 n2 1 n2 1
n
Vm 1 + ( 1) 0 for odd n > 1
= = Vm 2
n2 1 n2 1 for even n > 0
Vm 2
) an = 2 ; n = 1; 2; 3:: (16.63)
(2n) 1
460 CHAPTER 16. FOURIER SERIES AND TRANSFORM
and
8 9
>
<ZT =2 T =2
Z >
=
Vm
bn = cos [(n 1) !t] dt cos [(n + 1) !t] dt
T >
: >
;
0 0
T =2
Vm sin [(n 1) !t] sin [(n + 1) !t]
=
!T n 1 n+1 0
Vm sin [(n 1) ] sin [(n + 1) ]
) bn = =0 (16.64)
2 n 1 n+1
Using the results
Vm Vm Vm 2
a0 = ; a1 = ; a2n = 2 ; n = 1; 2; 3::
4 (2n) 1
Vm Vm
b1 = ; bn = 0 (16.65)
2T 4
we can write the Fourier series as
Vm Vm ! Vm Vm
f (t) = + cos (!t) + sin (!t)
2 4 4 4
1
X
2 Vm cos (2n!t)
: (16.66)
n=1
4n2 1
The Inverse Fourier Transform: We may develop the inverse Fourier Transform
starting from the Fourier series expansion. Suppose we are considering the
interval [ T; T ] instead of [ ; ] the Fourier series expansion for the function
f (t) can be expressed as
1 h
1 X n n i
f (t) = a0 + an cos t + bn sin t (16.68)
2 n=1
T T
where the Fourier Coe¢ cients are given by
ZT
1 n 0
an = f (t0 ) cos t dt0 ; n = 0; 1; 2; 3:::
T T
T
ZT
1 n 0
bn = f (t0 ) sin t dt0 ; n = 1; 2; 3::: (16.69)
T T
T
16.5. THE FOURIER TRANSFORM 461
1
X ZT
1 n n 0
+ sin t f (t0 ) sin t dt0
T n=1
T T
T
ZT 1 ZT n
1 0 0 1 X n n 0 n n 0 o 0
= f (t ) dt + f (t0 ) cos t cos t + sin t sin t dt
T T n=1 T T T T
T T
ZT X1 ZT h n i
1 1
) f (t) = f (t0 ) dt0 + f (t0 ) cos (t t0 ) dt0 ; (16.70)
T n=1
T T
T T
where we used the double angle formula for cosine. Now introducing the variable
de…ned by
n ! ! n+1 ! n 1
!n = ) = = (16.71)
T T
so that
ZT 1 ZT
1 0 0 1X
f (t) = ! f (t ) dt + ! f (t0 ) cos (! n (t t0 )) dt0 (16.72)
T n=1 T
or
1 ZT
1X
f (t) = ! f (t0 ) cos (! n (t t0 )) dt0 : (16.73)
n=0 T
and
Z 1 Z1
i
d! f (t0 ) sin [! (t t0 )] dt0 = 0 (16.77)
2 1
1
so that
Z 1 Z1
1
f (t) = d! f (t0 ) cos [! (t t0 )] dt0
0
1
Z 1 Z1
1
= d! f (t0 ) fcos [! (t t0 )] + i sin [! (t t0 )]g dt0
2 1
1
Z1 Z1
1 i!t 1 i!t0
) f (t) = p d!e p f (t0 ) e dt0 (16.78)
2 2
1 1
We recall Z 1
1 i!t
F (!) = F [f (t)] = p f (t) e dt (16.79)
2 1
so that
Z1
1
f (t) = p F (!) ei!t d! = F 1
[F (!)] (16.80)
2
1
Example 15.7 Find and describe the Fourier transform of the function f (t)
de…ned by
1 jtj < 1
f (t) = (16.81)
0 jtj > 1
which is an even function of t. This is the single slit di¤raction problem
of physical optics. The slit is described by f (t) :The di¤raction pattern
amplitude is given by the Fourier transform F (!) :
Solution: We recall Z 1
1 i!t
F [f (t)] = p f (t) e dt (16.82)
2 1
so that for the function given we can write
Z 1 1
1 1 e i!t
F [f (t)] = p e i!t dt = p
2 1 2 i! 1
r
2 ei! e i! 2 sin (!)
=p ) F [f (x)] = (16.83)
2 ! 2i !
16.5. THE FOURIER TRANSFORM 463
and using
" #
2 2
2 2 2 i! ! 2 i! !2
a x + i!x = a x+ 2 + = a2 x + +
2a 2a2 2a2 4a2
(16.87)
one can write
!2 Z 1
e 4a2 a2 (x+ 2a
i!
2)
2
F [f (x)] = p e dx: (16.88)
2 1
464 CHAPTER 16. FOURIER SERIES AND TRANSFORM
i! i!
for the rectangular closed loop with vertices T; T; T + 2a 2; T + 2a 2 . For
(16.91)
I Z1 Z1 2
Z1
a2 z 2 a2 x2 a2 (x+ 2a
i!
2) a2 x2
e dz = 0= e dx + e dx = e dx
1 1 1
Z1 2
Z1 2
Z1
a2 (x+ 2a
i!
2) a2 (x+ 2a
i!
2) a2 x2
e dx = 0) e dx = e dx (16.93)
1 1 1
We may write
Z1 Z1
a2 x2 1 y2
e dx = e dy (16.94)
a
1 1
16.6. THE DIRAC DELTA FUNCTION AND THE FOURIER TRANSFORM465
we …nd
Z1 2
p
a2 (x+ 2a
i!
2)
e dx = : (16.96)
a
1
becomes
!2
e 4a2
F [f (x)] = p = F (!) : (16.98)
a 2
so that substituting
Z 1
1 i!t0
F (!) = F [f (t)] = p f (t0 ) e dt0 (16.100)
2 1
466 CHAPTER 16. FOURIER SERIES AND TRANSFORM
we have
2 3 2 3
Z1 Z1 Z1 Z1
1 4 p1
0 1 i! (t t 0
) d! 5 dt0
f (t) = p f (t0 ) e i!t
dt0 5 ei!t d! = f (t0 ) 4 e
2 2 2
1 1 1 1
(16.101)
Recalling the property of the Dirac Delta function
Z1
f (t) = f (t0 ) (t t0 ) dt0 (16.102)
1
we …nd
Z1
0 1 i! (t t0 )
(t t)= e d! (16.103)
2
1
The Dirac Delta function has di¤erent forms. One of this form is given by
Za " 0 0
#
0 1 i! (t t0 ) 1 eia(t t ) e ia(t t )
(t t ) = lim e d! = lim
2 a!1 2 a!1 i (t t0 )
a
sin [a (t t0 )]
) (t t0 ) = lim : (16.104)
a!1 (t t0 )
Suppose the function
f (t) = 1 ) f (t0 ) = 1 (16.105)
for
Z1
f (t) = f (t0 ) (t t0 ) dt0 (16.106)
1
we then …nd
Z1
(t t0 ) dt0 = 1: (16.107)
1
Now let’s examine the graph of the Dirac Delta function
sin [a (t t0 )]
(t t0 ) = lim : (16.108)
a!1 (t t0 )
for di¤erent value of a at t0 = 0:
From the above graphs it is easy to conclude that
1 t = t0
(t t0 ) = (16.109)
0 otherwise
which also leads to
Zb
f (t) a < t < b
f (t0 ) (t t0 ) dt0 = (16.110)
0 otherwise
a
16.7. APPLICATIONS OF THE FOURIER TRANSFORM 467
~2 2 @ (~r; t)
r (~r; t) + U (~r) (~r; t) = i~ (16.111)
2m @t
The Special Case of a Free Particle Traveling in the Positive-x Direction: For
a free particle, we have
U (~r) = 0 (16.112)
and the time independent part of the S.E. for a particle traveling in the positive
x-direction can be written as
~2 d2 (x) d2 (x) 2mE
= E (x) ) = (x) : (16.113)
2m dx2 dx2 ~2
Using De Broglie wavelength of the particle, we can relate the momentum with
the wave vector
h h 2 h
= )p= = ) p = k~: (16.114)
p 2
468 CHAPTER 16. FOURIER SERIES AND TRANSFORM
p2
E= (16.115)
2m
can be expressed as
k 2 ~2
E= : (16.116)
2m
Using this result one …nds
Since for a free particle k takes continuous value, we may write the general
solution as
Z1
(x) = A (k) eikx dk: (16.119)
1
We do not know anything about about A(k). What we do know, from a pure
mathematical point of view, is a constant resulting from integrating a second
order di¤erential equation. From a physical point of view it is determined by
quantum mechanical conditions that need to be met for a free particle. In
quantum mechanics everything is probabilistic which is determined by the wave
function. The probability of …nding the particle anywhere ( 1 < x < +1) is
one. That means
Z1 Z1
2
j (x)j dx = (x) (x) dx = 1: (16.120)
1 1
we may write
2 3
Z1 Z1
1
A (k) 4 A (k 0 ) (k k 0 ) dk 0 5 dk = (16.123)
2
1 1
Z1
f (t) = f (t0 ) (t t0 ) dt0 (16.124)
1
we …nd
Z1 Z1
1 2 1
A (k) A (k) dk = ) jA (k)j dk = : (16.125)
2 2
1 1
Expressing A (k) as
1
A (k) = p (k) ; (16.126)
2
where (k) is known as the momentum eigen function (which you will be intro-
duce to in quantum mechanics class), we …nd
Z1 Z1
2
j (k)j dk = (k) (k) dk = 1: (16.127)
1 1
Then in terms of the momentum eigen function, the wave function can be ex-
pressed as
Z1 Z1
ikx 1
(x) = A (k) e dk = p (k) eikx dk: (16.128)
2
1 1
This shows that the wave function for a free particle is the inverse Fourier
Transform of the momentum eigenfunction (k)
Uncertainties
q q
2 2
x = x = hx2 i hxi ; k = k = hk 2 i hki (16.129)
Expectation Values
Z1 Z1
hf (x)i = (x) f (x) (x) dx; hg (k)i = (k) g (k) (k) dk (16.130)
1 1
470 CHAPTER 16. FOURIER SERIES AND TRANSFORM
A k2 =4
(k) = p e (16.133)
2
hki = 0; k 2 = : (16.134)
~
x px = (16.135)
2
is satis…ed.
Solution:
(a) We recall
Z1
1
(x) = p (k) eikx dk (16.136)
2
1
Z1
1 ikx
(k) = p (x) e dx: (16.137)
2
1
Then substituting
1=4
x2 2 x2
(x) = Ae = e (16.138)
16.7. APPLICATIONS OF THE FOURIER TRANSFORM 471
we …nd
Z1 Z1
A x2 ikx A (x2 + ik ) dx:
(k) = p e e dx = p e (16.139)
2 2
1 1
Noting that
2
ik ik k2
x2 + = x+ + (16.140)
2 4 2
we can write
Z1
A k2
(x+ 2ik ) dx:
2
(k) = p e 4 e (16.141)
2
1
we …nd
2 p p
A 3=2 1=2
k2 = p (2 ) = A2 (2 ) : (16.148)
2 2 2
Substituting
1=4
2
A= (16.149)
we …nd p p
2 2 p
k = p 2 = : (16.150)
2
(c) It is easier to …nd these expectation values in position space rather than in
momentum space, which we may write as
Z1 Z1
2
hxi = (x) x (x) dx; ; x = (x) x2 (x) dx: (16.151)
1 1
Replacing
p dy
y= 2 x ) dx = p (16.154)
2
we have
Z1
2 A2 y2
x = p 3=2
y2 e dy (16.155)
2 1
which gives p
2 A2
x = p 3=2
(16.156)
2 2
or p p
2 1 2 1
x = p 3=2
p = (16.157)
2 2 4
16.8. FOURIER TRANSFORMS AND CONVOLUTION 473
Noting that Z 1
i!y
F2 (!) = f2 (y) e dy (16.168)
1
Solution: We note that the two functions shown above can be expressed as
1 10 < x < 10
f (x) = (16.170)
0 otherwise
1
g (x) = 10 x 0 < x < 10
(16.171)
0 otherwise
For g (x y) ; we may write
1
10 (x y) 0 < x y < 10
g (x y) = (16.172)
0 otherwise
or
1
10 (x y) x 10 < y < x
g (x y) = (16.173)
0 otherwise
Applying the convolution formula
Z x Z x
1 1 1
f g=p f (y) (x y) dy = p f (y) (x y) dy
2 x 10 10 10 2 x 10
(16.174)
this can be rewritten as
Z x
1
f g= p f (y) (x y) dy (16.175)
10 2 x 10
Case 1 : x 10
For this case the range of y would be ( 1; 20) :In this interval we have
f (y) = 0; and we …nd
f g=0 (16.176)
Z x x
1 1 y2
f g= p (x y) dy = p xy
10 2 10 10 2 2 10
2 2
1 x 1 x
= p + 10x + 50 )f g=p +x+5 (16.177)
10 2 2 2 20
Z x x
1 1 y2
f g= p (x y) dy = p xy
10 2 x 10 10 2 2 x 10
" !#
2 2
1 x (x 10)
= p x (x 10)
10 2 2 2
1 x2 x2 100 5
= p )f g=p (16.178)
10 2 2 2 2
Z 10 10
1 1 y2
f g= p (x y) dy = p xy
10 2 x 10 10 2 2 x 10
1 20x 100 x2 + 100
= p
10 2 2
1 x2
)f g=p x (16.179)
2 20
2.0
1.5
f*g
1.0
0.5
0.0
-20 -10 0 10 20 30
Mathematical Methods in
physics III
477
Chapter 17
Manifolds
begins to rotate. It is free to rotate about the x-axis, y-axis, or z-axis. Let’s say
we want to describe the angular position of the center of mass of the meter stick
over a period of time, = 10s; with a time interval of 2s. How many independent
parameters, that depend on time, do we need to describe the angular position
of the center of mass of the meterstick relative to its initial position at t = 0?
Well the answer is simple. We need three independent parameters, the Euler
479
480 CHAPTER 17. MANIFOLDS
angles, ( 1 (t) ; 2 (t) ; 3 (t)) which describes the rotation about the x, y, and
z axes at a given instant of time. Then over the 10 second interval we have a
set that consist of 5 points
1 2 3 1 2 3 1 2 3
(2) ; (2) ; (2) ; (4) ; (4) ; (4) ; (6) ; (6) ; (6) ;
1 2 3 1 2 3
(8) ; (8) ; (8) ; (10) ; (10) ; (10)
We can make the time interval in…nitesimal to continuously describe the angular
position of the center of mass of the meterstick. The resulting set of points form
a Manifold of dimension three.
Let’s consider another example of a Manifold. In classical mechanics you
may have studied what is called the phase space. In this space you can describe
the state of a particle over a period of time using the three coordinates of
space locating the position of the particle and the three coordinates of speed
(or momentum) describing how fast the particle is moving at a given instant of
time. In Cartesian coordinate system we use the independent parameters (x (t) ;
y (t) ; z (t)) for position and (x_ (t) ; y_ (t) ; z_ (t)) for the speed of the particle. I can
represent these independent parameters by (x1 (t) ; x2 (t) ; x3 (t) ; x4 (t) ; x5 (t) ;
x6 (t)). So when you describe the state of the particle say from t = 0 to t = t0 ;
you can use in…nitesimal time interval so that you will have a set of points that
can be parameterized continuously in terms of (x1 (t) ; x2 (t) ; ::: x6 (t)). These
set of points form a Manifold of dimension six.
Now let’s apply this to the Minkowski spacetime where we have three space
coordinates (x (t) ; y (t) ; z (t)) and time, t. This forms a four dimensional man-
ifold with four coordinates each parametrized by the proper time ; (x1 ( ) ;
x2 ( ) ; x3 ( ) ; x4 ( )). Therefore, a manifold is any set that can be continu-
ously parameterized. Therefore, an N dimensional manifold, M; of points is
one for which N independent real coordinates (x1 ; x2 ; x3 ; :::; xN ) are required
to specify any point completely.
A manifold is Continuos: if you pick any point, p, on the Manifold and you
can …nd another points whose coordinates di¤er in…nitesimally from the point
p.
A manifold is di¤ erentiable: if you pick any point, p, on the Manifold and
you can …nd a scalar function that is di¤erentiable at that point p.
Coordinates of a Manifold M : a point in an N -Dimensional Manifold is
represented by the coordinates (x1 ; x2 ; x3 ; ::: xN ) which we represent by xa
where it is understood that a = 1; 2; 3::: N:
Degeneracy in a Manifold : sometimes it may not be possible to cover the
whole manifold with only one none-degenerate coordinate system. Example is
a plane in polar coordinate system ( ; '). A plane is a two dimension Manifold.
(called R2 ). A plane in polar coordinates has a degeneracy at the origin since
' is indeterminate at the origin. (Fig. 17.2)
Coordinate patches: these are coordinate systems that covers a portion of
the Manifold where we have degeneracy. For example the surface of a sphere is
a two dimensional Manifold (called S 2 ). It can be described by two independent
coordinates (x1 = ; x2 = ') except at two points on the Manifold. These are
17.2. CURVES AND SURFACES IN A MANIFOLD 481
Figure 17.3: The north and south pole on the surface of sphere are degenerate
in polar coordinates.
the north and south pole where ' is indeterminate (or there is degeneracy).
(Fig.17.3) There is no coordinate system that covers the entire sphere without
running into these two degenerate points. In this case the smallest number of
patches we need is two.
Atlas: an atlas is a set of coordinate patches that covers the whole Manifold.
f x1 ; x2 ; x3 ; :::xN = 0: (17.5)
The surface shown in blue in Fig. 17.4 in the 3D manifold needs two parameters
to de…ne it
x1 = 2 cos(u1 ); x2 = sin (u2 ) ; x3 = u2 : (17.6)
Note that this surface is in 3D manifold and it parameterized by two coordinates
(u; v), (M = N 1 = 3 1 = 2; it is a hypersurface).
17.2. CURVES AND SURFACES IN A MANIFOLD 483
Example 2.1 Let’s consider the 3-D Euclidean Manifold. A sphere is a High-
persurface since M = 2; (Fig.17.5). A point on a sphere is de…ned by
x2 + y 2 + z 2 = a2 ; (17.7)
where a is the radius of the sphere. We note that in this case the surface
of the sphere is a Hypersurface that can be de…ned by the equation
2 2 2
g x1 ; x2 ; x3 = x1 + x2 + x3 a2 = 0; (17.8)
and u3 de…ned by
x1 = u1 sin (u2 ) cos (u3 ) ; x2 = u1 sin (u2 ) sin (u3 ) ; x3 = u1 cos (u2 )
we can write the equation that de…ne the surface of the sphere (M = 2)
that is embedded in a 3D manifold (N = 3) using only one parameter (by
eliminating the N 1 = 2; parameters) as
2 2
g x1 ; x2 ; x3 = (u1 sin (u2 ) cos (u3 )) + (u1 sin (u2 ) sin (u3 ))
2
+ (u1 cos (u2 )) a2 = 0 ) g x1 ; x2 ; x3 = u21 a2 = 0 (17.9)
of dimension N (M < N ).
g1 x1 ; x2 ; x3 ; :::xN = 0; (17.10)
1 2 3 N
g2 x ; x ; x ; :::x = 0;
1 2 3 N
g3 x ; x ; x ; :::x = 0
:
:
:
1 2 3 N
gN M x ; x ; x ; :::x = 0
so that 2 3 2 32 3
@g @x01 @x02 @x03 @g
@x1 @x1 @x1 @x1 @x01
4 @g 5=6
4 @x01 @x02 @x03 74
5 @g 5 (17.18)
@x2 @x2 @x2 @x2 @x02
@g @x01 @x02 @x03 @g
@x3 @x3 @x3 @x3 @x03
can be written as
2 @g 3 2 @x01 @x02 @x03
32 @x1 @x2 @x3
32
@g 3
@x1 @x1 @x1 @x1 @x01 @x01 @x01 @x1
4 @g2 5 = 6 76 74
1 2 3
4 @x01 @x02 @x03
54 @x @x @x
5 @g 5: (17.19)
@x @x2 @x2 @x2 @x02 @x02 @x02 @x2
@g @x 01
@x02 @x03 1 2 3 @g
@x @x @x
@x3 @x3 @x3 @x3 @x03 @x03 @x03 @x3
so that
1 1
A A = AA = 1: (17.23)
T
We note that the transpose, A ; is given by
2 1 1
3
@x1 @x @x
6 @x01 @x02
2
@x03
2 7
AT = 6
4
@x2 @x @x 7:
5 (17.24)
@x01 @x02
3
@x03
3
@x3 @x @x
@x01 @x02 @x03
Similarly for the inverse matrix, the transpose matrix which we express as
2 01 01 01
3
@x @x @x
@x0a 6 @x1
@x02
@x2
@x02
@x3
@x02 7
=4 @x1 @x2 @x3 5 (17.25)
@xb @x03 @x03 @x03
@x1 @x2 @x3
486 CHAPTER 17. MANIFOLDS
We note that
0a 0a
X @x @xb 0a N 0a
@x0a @x @x1 @x @x2 @x @xN
01
= 1 01
+ 2 01
+ :: N 01
= (17.30)
@x @x @x @x @x @x @x @xb @x01
b=1
17.3. COORDINATE TRANSFORMATIONS AND SUMMATION CONVENTION487
where the summation is evaluated at P . Similarly for the interval between x0a
and x0a + dx0a ; in the none-primed coordinate system, we can write
X @xa N
@xa 01 @xa 02 @xa
dxa = 01
dx + 02
dx ::: 0N dx0N = dx0b ; (17.34)
@x @x @x @x0b
b=1
and
N
X @xa 0b @xa 0b
dxa = 0b
dx = gb (x0 ) dx0b = dx : (17.36)
@x @x0b
b=1
488 CHAPTER 17. MANIFOLDS
The index a; which is known as the free index, can take any value from 1 to
N . The index b, which is known as the dummy index and it must be summed
up from 1 to N .
Let’s reconsider the worldline of the alien in the spaceship in the previous chap-
ter. Imagine there is a surface de…ned by the two coordinates (x; ct) represented
as x1 ; x2 : Let’s consider two points P and Q; as shown in Fig. 17.7. The
coordinates for the points are x1 ; x2 and x1 + dx1 ; x2 + dx2 ; respectively.
Suppose this surface is de…ned by the function s x1 ; x2 The interval between
@s 1 @s 2
ds = dx + dx
@x1 @x2
@s 1 @s 2 @s 1 @s 2
) ds2 = 1
dx + 2
dx 1
dx + dx
@x @x @x @x2
@s @s 1 1 @s @s 1 2 @s @s 1 2
) ds2 = dx dx + dx dx + dx dx
@x1 @x1 @x1 @x2 @x2 @x1
@s @s
+ 2 2 dx2 dx2 (17.38)
@x @x
17.4. THE RIEMANNIAN GEOMETRY 489
ds2 = g11 x1 ; x2 dx1 dx1 + g12 x1 ; x2 dx1 dx2 + g21 x1 ; x2 dx1 dx2
2 X
X 2
1 2 2 2
+g22 x ; x dx dx = gab (x) dxa dxb (17.39)
a=1 b=1
where we replaced
@s @s @s @s @s @s @s @s
g11 = ; g12 = ; g21 = ; g22 =
@x1 @x1 @x1 @x2 @x2 @x1 @x2 @x2
Consider a curved surface in 3-D Euclidean space. We know that this surface
can be de…ned by a function g, that depends on (x; y; z) in Cartesian, (r; ; ')
in spherical, or (r; '; z) in cylindrical coordinates. Suppose we represent these
coordinates by x1 ; x2 ; x3 ; then we may write the function that de…nes the
surface as
s (x) = g (x) = g x1 ; x2 ; x3 : (17.40)
Now let’s consider a point P on this surface that has coordinates x1 ; x2 ; x3 :
Suppose we consider another point Q with coordinates x1 + dx1 ; x2 + dx2 ; x3 + dx2 ;
we may de…ne the surface between these two points as ds2 . This displacement
is just the di¤erential of Eq. (17.40)
@g @g @g @g @g @g
ds2 = dx1 + dx2 + dx3 : dx1 + dx2 + dx3
@x1 @x2 @x3 @x1 @x2 @x3
(17.41)
or using Einstein’s summation convention, the geometry of the surface between
the two points
3
! 3
!
X @g X @g b
2 a
ds = (ds) (ds) = dx dx
a=1
@xa @xb
b=1
3 X
X 3
@g @g a b
= dx dx : (17.42)
a=1 b=1
@xa @xb
@g @g @g @g
a b
= :
@x @x @xb @xa
Suppose the coordinates for an orthonormal set like the Cartesian, spherical, or
cylindrical, we note that
2 2
3
@g
6 @x1 0 0 7
6 2 7
G=6 6 0 @g
@x2 0 7
7 (17.46)
4 3 5
@g
0 0 @x3
N X
X N
ds2 = gab (x) dxa dxb ; (17.47)
a=1 b=1
@xa @xb 0c 0d
ds2 = gab (x) dx dx ; (17.50)
@x0c @x0d
or
0
ds2 = gcd (x0 ) dx0c dx0d ; (17.51)
where
a b
0 0 @x @x
gcd (x ) = gab (x) 0c 0d : (17.52)
@x @x
Note that x = x (x0 ) :
17.5. INTRINSIC AND EXTRINSIC GEOMETRY AND THE METRIC 491
(a) Intrinsic: when the geometry remains unchanged as viewed in the higher
dimensional manifold.
(b) Extrinsic: when the geometry is di¤erent as viewed in the higher dimen-
sional manifold.
~r ~r0 = (x x0 ) x
^ + (y y0 ) y^ + (z z0 ) z^ (17.54)
is zero,
~ (~r
N ~r0 ) = a (x x0 ) + b (y y0 ) + c (z z0 ) = 0: (17.55)
ax + by + cz = d; (17.56)
492 CHAPTER 17. MANIFOLDS
where
d = ax0 + by0 + cz0 : (17.57)
Using the notation x1 ; x2 ; x3 for (x; y; z) , we may write
where
d = ax10 + bx20 + cx30 : (17.59)
1 2
For a plane that depends on only (x ; x )we may write the interval between
Now let’s consider the 2-D cylindrical surface which we can construct using
our 2-D plane. Suppose the cylinder has radius, a; with its axis along the
z-axis (which we call it x03 axis): Using cylindrical coordinates a point
on the surface of the cylinder can be described by (a; '; z) or using our
coordinates notations (a; x02 ; x03 ), we can de…ne the surface by the function
we may write the interval between two points on this surface, point P
and Q with coordinates (x1 ; x2 ; x3 ) and point Q with coordinates (x1 +
dx1 ; x2 + dx2 ; x3 + dx3 ) just using geometrical visualization, as
2 2 2
ds2 = dx1 + dx2 + dx3 : (17.62)
we …nd
2 2 2 2 2
ds2 = dx1 + dx2 + dx3 = a2 dx02 + dx03 : (17.63)
17.5. INTRINSIC AND EXTRINSIC GEOMETRY AND THE METRIC 493
x1 = ax02 ; x2 = x03
Such kind of geometry is not intrinsic and it is called extrinsic. Its cur-
vature is extrinsic and is a result of the way it is embedded in the three
dimensional space..
This is a 2-D surface embedded in a 3-D manifold. You can not obtain
this geometry from the plane geometry like the 2-D cylindrical geometry.
Such kind of geometry is intrinsic. This means the geometry of a sphere
is intrinsically curved because we can not transform Eq.(17.65) to the
Euclidean form
2 2
ds2 = dx1 + dx2 (17.66)
over the whole surface by any coordinate transformation. Note that this
can be done locally but not for the whole spherical surface.
Example 2.4 Find the metric for a two-dimensional sphere of radius, a; em-
bedded in a 3-D Euclidean space both in Cartesian coordinates (x1 ; x2 ; x3 ).
Refer to Fig. 17.5
Solution: We recall that the line element in a 3-D Euclidean space is given by
2 2 2
ds2 = dx1 + dx2 + dx3 ; (17.67)
so that
x1 dx1 + x2 dx2
dx3 = r :
2 2
a2 (x1 ) + (x2 )
Then for a 2-D sphere embedded in a 3-D Euclidean space, the metric is
given by
0 12
2 2 B x1 dx1 + x2 dx2 C
ds2 = dx1 + dx2 +B
@ r C :
A (17.70)
2 2
a2 (x1 ) + (x2 )
x1 = x2 ' 0
Let’s use the coordinates (x01 ; x02 ; x03 ). de…ned by the transformation
Note that the origin is set at the north pole of the sphere at point as shown
in Fig. 17.5,. Then the interval can be written as
2
ds2 = dx01 cos x02 x01 sin x02 dx02
0 12
01 01
2 x dx
+ dx01 sin x02 + x01 cos x02 dx02 + @ q A (17.73)
2
a2 (x01 )
or
2 2
ds2 = g11 dx01 + g22 dx02 (17.76)
where
a2 2
g11 = 2 ; g22 = x01 (17.77)
a2 (x01 )
are the none zero elements of the metric tensor. We will see the use of
these elements of the metric tensor in the next section to determine length
and area of a 2-D sphere in a 3-D Euclidean manifold.
so that
x1 dx1 + x2 dx2 + x3 dx3
dx4 = r :
2 2 2
a2 (x1 ) + (x2 ) + (x3 )
We introduce the coordinates (r; ; ') which we represent (x01 ; x02 ; x03 ).
and de…ned by the transformation
or
x1 = x01 sin x02 cos x03 ; x2 = sin x02 sin x03 ; x3 = cos x02 :
Then the distance between P and Q shown by the red line (hypotenuse
side) can be expressed as
2
ds02 = dr2 + ds2s = dr2 + r2 sin2 ( ) d'2 + r2 d (17.82)
ds02 = dx2 + dy 2 + dz 2
so that
rdr
dx4 = p
: (17.85)
a2 r2
Then the metric for a 3-D sphere in a 4-D Euclidean manifold is given by
2 r2 dr2
ds2 = ds02 + dx4 = dr2 + r2 sin2 ( ) d'2 + r2 d + ; (17.86)
a2 r2
498 CHAPTER 17. MANIFOLDS
where we used the notation x01 ; x02 ; x03 for the coordinates (r; ; ') and
identify the none zero elements of the metric tensor
a2
g11 = ; g22 = r2 ; g33 = r2 sin2 ( )
a2 r2
We will see the use of these elements of the metric tensor in the next
section to determine length, area, and volume of a 3-D sphere in a 4-D
Euclidean manifold.
the absolute value is because of the fact that for pseudo-Riemannian manifolds it
can be negative as is the case for spacelike in the Minkawski spacetime manifold.
For xa = xa (u) ; we recall write
dxa dxb
dxa = du; dxb = du (17.90)
du du
so that s
Z Q
dxa dxb
LP Q = gab (u) du: (17.91)
P du du
dxp pq 1; p=q
= = (17.93)
dxq 0; p 6= q
For a hypersurface, like the 2D sphere in the 3D manifold, the surface is de…ned
by the equation
x3 = constant ) dx3 = 0:
and the expression for the area reduces to
ZZ p ZZ p
A= jg22 (x) g11 (x)jdx2 dx1 + jg11 (x) g22 (x)jdx1 dx2
where one can absorb the factor 2 into the metric elements. For such coordinates
the metric is diagonal
2 2 2
ds2 = g11 dx1 + g22 dx2 ::: + gN N dxN : (17.94)
xa = constant; (17.95)
p
dA = jg11 g22 jdx1 dx2 (17.96)
xa = constant (17.97)
(a) The distance, D, from the origin to the perimeter on this surface along
a line of constant ' i:e: x02 = cons The origin is at the north pole as
shown in Fig. 17.6.
(b) The circumference of the circle with radius, = R i:e: x01 = R = cons :
(c) The area of the spherical surface enclosed by the perimeter with radius,
= R; (i.e. the surface shaded green in Fig. 17.6).
Solution:
(a) We recall that the metric for a 2-D sphere in a 3-D Euclidean manifold is
given by
2 2
ds2 = g11 dx01 + g22 dx02 ; (17.99)
where we used the notation x01 ; x02 for the coordinates ( ; ') and identify
the none zero elements of the metric tensor
a2 2
g11 = ; g22 = : (17.100)
a2 2
17.6. LENGTH, AREA, AND VOLUME 501
(c) The surface is de…ned by x03 =constant: Then according to Eq. (17.96),
the in…nitesimal area is given by
p
dA = jg11 g22 jdx01 dx02 : (17.106)
To cover the area (colored Aqua) shown in Fig 17.6, we should have for
the limits of integrations [0; R] for and [0; 2 ] for '
ZRZ2 " r #
a R2
A= p d d = 2 a2 1 1 : (17.108)
a2 2 a2
0 0
(a) the distance from its center to the surface of this sphere along constant
and constant .
becomes Z Q p
LP Q = jg11 (x0 ) dx01 dx01 j (17.114)
P
or
Z s Z
R R
a2 a 1 R
D= dr2 = p dr = a sin ; (17.115)
0 a2 r 2
0 a2 r2 a
which is the same as the result we obtained in the previous example.
(b) Across the equator, we have x01 = r = R; x02 = = =2 and obviously
dx01 = 0 and dx02 = 0: Then the length
Z Qq
LP Q = jgab (x0 ) dx0a dx0b j; (17.116)
P
g22 = r2 r=R
= R2
2
g33 = 2
r sin ( ) r=R
= R2 sin2 ( ) (17.119)
and the limit of integration for x02 is (0; ) and for x03 is (0; 2 ) ; the
surface area would be become
Z Z 2 q Z Z 2
2
A= 4
R sin ( ) d d = R2 sin ( ) d d = 4 R2 :
0 0 0 0
(17.120)
(d) In this case we are considering a 3-D sphere embedded in a 4-D Euclidean
space. For this space the volume is de…ned by x04 =constant. Therefore,
applying the relation in Eq. (17.98), an in…nitessimal volume in this 4-D
space is given by
p
dV = jg11 g22 g33 jdx01 dx02 dx03 : (17.121)
Thus using
a2
g11 = ; g22 = r2 ; g33 = r2 sin2 ( ) (17.122)
a2 r2
the volume bounded by the 2-D spherical surface of radius x01 = r = R
becomes
Z RZ Z 2 s 2 4 2
a r sin ( )
V = drd d
0 0 0 a2 r 2
Z R Z Z 2
r2 dr
= a p sin ( ) d d (17.123)
0 a2 r2 0 0
or
Z R 2 Z Z 2
x01 dx01 sin x02 dx02 dx03
V =a q q ;
0 2 0 0 2
a2 (x01 ) a2 (x01 )
Let’s evaluate the integral
Z
r2 dr
I= p (17.124)
a2 r2
Introducing the transformation de…ned by
r = a sin ( ) ) dr = a cos ( ) d
17.6. LENGTH, AREA, AND VOLUME 505
we have
Z Z 2 2 Z
r2 dr a sin ( ) a cos ( ) d
I= p = q = a2 sin2 ( ) d
a2 r2 a2 a2 sin2 ( )
Z
a2 a2 sin (2 )
= (1 cos (2 )) d =
2 2 2
Z 2 2
r dr a
)I= p = [ sin ( ) cos ( )] (17.125)
a2 r2 2
so that using
R
r = a sin ( ) ) sin ( ) = a; for r = R
(17.126)
0; for r = 0
and
q ( q
R 2
cos ( ) = 1 2
sin ( ) = 1 a ; for r = R (17.127)
0; for r = 0
one …nds
2 s 3
Z R 2 2 2
r dr a 4 1 R R R 5
p = sin 1 (17.128)
0 a2 r2 2 a a a
One must be able to recover the 3-D Euclidean space for a ! 1: This
means Ra << 1 the result for the volume of a sphere with radius, R
must be that of the volume of a sphere with radius R in 3D Euclidean
space (V =. 43 R3 ). One can easily …nd from Eq. (17.129) using the
approximations for Ra << 1;
s
3 2 2
R R 1 R R 1 R
sin 1 ' ; 1 =1 (17.130)
a a 3 a a 2 a
that gives
( !)
3 2
3 R 1 R R 1 R
V =2 a 1
a 3 a a 2 a
( )
3 3
3 1 R 1 R 4 3
=2 a + )V = R (17.131)
3 a 2 a 3
506 CHAPTER 17. MANIFOLDS
Homework :
(a) Express the circumference, the area, and the volume in terms of D in the
previous example and show that all have maximum values at
a
D= (17.132)
2
(b) Show that the total volume of this space is …nite and is equal to
2 3
V =2 a : (17.133)
(a) Show that circumference, C, and the area, A are still C = 2 R and A =
4 R2 :
(b) The distance D from the center of the sphere to the surface is
1
D = b sinh (R=b) (17.134)
and in this case show that A and V of the sphere are monotonically in-
creasing functions
for all points on the manifold. However, it is possible to …nd coordinates x0a
0
such that at the point P the new metric functions gab (x) satisfy the conditions
0
gab (xaP ) = ab ; (17.137)
0
@gab (x0 )
= 0: (17.138)
@x0c xa
P
17.7. LOCAL CARTESIAN COORDINATES AND TANGENT SPACE 507
We recall from Mathematical methods for any function, g (x) ; that is di¤ er-
entiable for all values of x in the speci…ed domain,
dn g(x)
exists for all n 0 and x 2 R;
dxn
one can write the series expansion about xp in the domain (Taylor series)
X1
1 dn g(x0 )
g(x) = (x xp ) n : (17.139)
n=0
n! dx0n x=xp
1 2
For a function of two variable g (x) = g x ; x ; this becomes
@g(x01 ; x02 )
g x1 ; x2 = g x1p ; x2p + (x1 x1p )
@x01 0 1 2
x =xp ;xp
8
@g(x01 ; x02 ) 1 < @ 2 g(x01 ; x02 )
+ (x2 x2p )2 + (x1 x1p )2
@x02 x0 =x1p ;x2p 2! : @ (x01 )2 0 x =x1p ;x2p
2 01 02
@ g(x ; x )
+2 (x1 x1p )(x2 x2p )
@x01 @x01 x0 =x1p ;x2p
9
@ 2 g(x01 ; x02 ) =
+ 2 (x2 x2p )2 + ::: (17.141)
@ (x02 ) ;
x0 =x1p ;x2p
From this expression you can imagine how it gets nasty for a function of N
variables, like the metric, gab (x) = gab x1 ; x2 ; x3 ; :::xN :
Example 2.8 Let’s reconsider the metric for 2D sphere embedded in a 3D
manifold in Cartesian coordinates
0 12
2 2 B x1 dx1 + x2 dx2 C
ds2 = dx1 + dx2 +B
@r
C :
A (17.142)
2 2
a2 (x1 ) + (x2 )
We note that for the 2D sphere in the 3D manifold, using the metric we
…nd
g11 x1p ; x2p = g22 x1p ; x2p = 1; g12 x1p ; x2p = 0
and
" 2 #
@g11 x01 ; x02 @ x01 x1p
= 2 2
@x01 @x01 a2 x01 x1p x02 x2p
x1p ;x2p
2 3
01 3
6 2 x x1p 2 x 01
x1p 7
=4 2 2 +h i 5
a2 x01 x1p x02 x2p 2 2 2
a2 x01 x1p x02 x2p
x1p ;x2p
=0
Similarly one can easily show that
" 2 #
@g11 x01 ; x02 @ x01 x1p
= 2 2
@x02 @x02 a2 x01 x1p x02 x2p
x1p ;x2p x1p ;x2p
2 3
2
6 2 x01 x1p x02 x2p 7
= 4h i 5 =0
2 2 2
a2 x01 x1p x02 x2p
x1p ;x2p
17.7. LOCAL CARTESIAN COORDINATES AND TANGENT SPACE 509
" 2 #
@g22 x01 ; x02 @ x02 x2p
= 2 2
@x01 @x01 a2 x01 x1p x02 x2p
x1p ;x2p x1p ;x2p
2 3
2
6 2 x01 x1p x02 x2p 7
= 4h i 5 =0 (17.145)
2 2 2
a2 x01 x1p x02 x2p
x1p ;x2p
" 2 #
@g22 x01 ; x02 @ x02 x2p
= 2 2
@x02 @x02 a2 x01 x1p x02 x2p
x1p ;x2p
2 3
02 3
6 2 x x2p 2 x 02
x2p 7
=4 2 2 +h i 5
a2 x01 x1p x02 x2p 2 2 2
a2 x01 x1p x02 x2p
x1p ;x2p
=0 (17.146)
" 2 #
@g12 x01 ; x02 @ 2 x01 x1p x02 x2p
= 2 2 =0
@x01 @x01 a2 x01 x1p x02 x2p
x1p ;x2p x1p ;x2p
" 2 #
@g12 x01 ; x02 @ 2 x01 x1p x02 x2p
= 2 2 (17.147)
=0
@x02 @x02 a2 x01 x1p x02 x2p
x1p ;x2p x1p ;x2p
as you can see from the rectangular shape of the grids. We can show this
quantitatively using the function that de…nes the surface. The surface shown in
Fig. 17.10 is de…ned by the function
Figure 17.10: Three di¤erent tangent spaces (in this case tangent plane) at three
di¤erent points on the manifold.
so that
@g @g @g
dg = dx1 + dx2 + dx3 = 0
@x1 @x2 @x3
@g @g @g
) dg = x
^+ y^ + z^ dx1 x
^ + dx2 y^ + dx3 z^ = 0
@x1 @x2 @x3
~ x1 ; x2 ; x3 d~r x1 ; x2 ; x3 = 0;
dg = A (17.151)
where
A~ x1 ; x2 ; x3 = @g x ^+
@g
y^ +
@g
z^
@x 1 @x 2 @x3
@g @g @g
1
= 2 sin(x1 ) cos(x1 ); 2 = sin(x2 ); 3 = 1 (17.152)
@x @x @x
Note Eq. (17.151) shows that the vector A ~ x1 ; x2 ; x3 is normal to the surface
1 2 3
at the point with coordinates x ; x ; x :Now let’s pick a point, P , on the
surface with coordinates, ~rp = x1p ; x2p ; x3p and another neighboring point Q
with coordinates ~r = x1 ; x2 ; x3 that are at the same plane that is tangent to
the surface at point P, then we have for the vector on this plane given by
~r = ~r ~rp = x1 x1p x
^ + x2 x2p y^ + x3 x3p z^ (17.153)
and the vector normal to this tangent plane at point P is given by
~ x1 ; x2 ; x3 = @g @g @g
A p p p x
^+ y^ + z^ = m1 x
^ + m2 y^ z^ (17.154)
@x1 @x2 @x3 x1p ;x2p ;x3p
17.8. THE SIGNATURE OF A MANIFOLD 511
where
m1 = 2 sin(x1p ) cos(x1p ); m2 = sin(x2p ) (17.155)
The equation of the tangent plane at this point is determined by
~ x1 ; x2 ; x3
g=A ~r x1 ; x2 ; x3
= m1 x1 x1p + m2 x2 x2p x3 x3p = 0
) x3 x1 ; x2 = m1 x1 x1p + m2 x2 x2p + sin2 (x1p ) + cos(x2p ) (17.156)
or
a b
0 @x @x
gcd (xp ) = gab (xp ) ; (17.161)
@x0c xp @x0d
xp
G0 = X T GX: (17.166)
XT = X 1
) G0 = X 1
GX: (17.168)
G = X0 1
G0 X 0 = X 0 1
X 1
GXX 0 (17.169)
XX 0 = I ) X 0 = X 1
(17.170)
a a
This can be true if the transformation from x ! x0 is linear. The transforma-
tion matrix, X; with element Xba ; at point P each element must be a constant
and we can write
a
x0 = Xba xb : (17.171)
17.8. THE SIGNATURE OF A MANIFOLD 513
is positive for strictly Riemannian. This means gcd (x) must be positive de…nite
and the eigenvalues, a ; for G are also positive de…nite. On the other hand
for Pseudo Riemannian since the metric can be negative the eigenvalues can
a
be negative.p Now if we scale the coordinates x0 by these eigenvalues (i.e.
0a 0a 0
x ! x = j a j), for the metric tensor, G , we can easily show
2 3
1 0 ::: 0
6 0 1 ::: 0 7
6 7
6 : : 7
0
G =66 7: (17.174)
6 : : : 7 7
4 : : 5
0 0 ::: 1
Thus , at any arbitrary point, P; in a pseudo-Riemannian manifold, it is always
a
possible to …nd a coordinate system x0 such that in the neighborhood of P we
have h i
0 2
gab (x0 ) = ab + O x0 x0p ; (17.175)
where [ ab ] = diag ( 1; 1; :::; 1) : The number of positive entries (N+ ) minus
the number of negative entries (N ) in [ ab ] is called the signature of the man-
ifold. For example for the Minkowski spacetime manifold where the metric is
given by
2
ds2 = d (ct) dx2 dy 2 dz 2 (17.176)
2 3
1 0 0 0
6 0 1 0 0 7
=6 4 0 0
7: (17.177)
1 0 5
0 0 0 1
the signature is 2.
N-dimensional volume without a constraint: In an N-dimensional (pseudo)
Riemannian manifold with orthogonal coordinates system where the metric ten-
sor is diagonal, the full N-dimensional volume element dN V is
p
dN V = jgjdx1 dx2 dx3 :::dxN (17.178)
514 CHAPTER 17. MANIFOLDS
Vector Calculus on
manifolds
~t = lim ~s ~s (u + u) ~s (u)
= lim ; (18.1)
u!0 u u!0 u
where s is the in…nitessimal separation vector between the point P with coor-
dinate (u; s (u)) and some nearby point Q with coordinate (u + u; s (u + u))
on the curve, C, in the manifold corresponding (see Fig.??).
515
516 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
where x1p ; x2p ; x3p are the coordinates for point P . We note that a point on the
surface can de…ned by a vector
^ + x2 y^ + sin2 (x1 ) + cos(x2 ) z^
~s = x1 x (18.10)
18.3. THE METRIC FUNCTION AND COORDINATE TRANSFORMATIONS517
so that
~s = x1 x
^ + x2 y^ + 2 sin(x1 ) cos(x1 ) x1 sin(x2 ) x2 z^ (18.11)
The two basis vectors, e^1 and e^2 ; at point P , in Fig. 18.2 are given by
~s x2 1 1 2 x
2
e^1 = lim = lim x
^ + y
^ + 2 sin(x ) cos(x ) sin(x ) (18.12)
z^ ;
1
x !0 x1 x1 !0 x1 x1
~s x1 1 1 x
1
e^2 = lim = lim x
^ + y
^ + 2 sin(x ) cos(x ) sin(x2 )(18.13)
z^
x2 !0 x2 x2 !0 x2 x2
can be obtained from the inner product of the in…nitessimal vector separation.
We note that
ds2 = d~s d~s = e^a (x) dxa e^b (x) dxb = e^a (x) e^b (x) dxa dxb
) ds2 = gab (x) dxa dxb ; (18.17)
where
gab (x) = e^a (x) e^b (x)
is the metric function
(a) Find the basis vectors e^01 and e^02 in the tangent space at point P.
18.3. THE METRIC FUNCTION AND COORDINATE TRANSFORMATIONS519
(b) Re-derive the metric elements for a 2D sphere from the basis vectors.
Solution:
(a) We note that the tangent vector connecting point P with coordinate ( ; ')
with its neighboring point Q with coordinates ( + ; ' + ') (See Fig.
18.4) is expressible as
p
~s = ( cos (')) x
^ + ( sin (')) y^ + a2 2 z^
= [cos (') sin (') '] x
^ + [sin (') + cos (') '] y^
d
p z^ (18.18)
a2 2
" #
= cos (') x
^ + sin (') y^ p z^ +[ sin (') x
^ + cos (') y^] '
a2 2
(18.19)
(18.20)
and
~s
e^' = lim
'!0 '
" #
= cos (') x
^ + sin (') y^ p z^ +[ sin (') x
^ + cos (') y^] :
a2 2 '
(18.21)
Noting that
= =0 (18.22)
' '
we …nd
e^ = cos (') x
^ + sin (') y^ p z^: (18.23)
a2 2
and
e^' = sin (') x
^ + cos (') y^: (18.24)
520 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
(b) The metric elements can be determined using the basis vectors
(a) Find the basis vectors e^r ; e^ ; and e^' in the tangent space at point P .
(b) Re-derive the metric elements for a 3D sphere from the basis vectors.
we have
b
e^b (x) d~s = e^b (x) e^a (x) dxa = a dx
a
= dxb : (18.29)
Expressing the in…nitessimal vector separation using the dual coordinates basis
vectors as
d~s = e^a (x) dxa (18.30)
we also …nd
b
e^b (x) d~s = e^b (x) e^a (x) dxa = a dx
a
= dxb : (18.31)
Thus the metric
ds2 = d~s d~s = e^a (x) dxa e^b (x) dxb = g ab (x) dxa dxb ; (18.32)
Using
@xa 0b
dxa = dx ; (18.37)
@x0b
we have
@xa
d~s = e^a dxa = e^a dx0b = e^0c dx0c : (18.38)
@x0b
Noting that
e^b e^b = 1
522 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
we have
@xa
e^a dx0b = e^0b e^0b e^0c dx0c = e^0c e^0b e^0b dx0c = e^0b bc dx0c (18.39)
@x0b
@xa
) e^a dx0b = e^0b dx0b : (18.40)
@x0b
so that
@xa
e^0b =
e^a : (18.41)
@x0b
Homework: Show that for the dual basis vector
@x0a c
e^0a = e^ : (18.42)
@xc
Components of a vector in coordinate transformations: In coordinate
transformations the vector components are di¤erent but the vector itself is un-
changed. Suppose the vector, ~v ; is a vector at point P in the xa coordinate
system and ~v 0 is the vector in the x0a coordinates at the same point on the
manifold. These vectors may be expressed in terms of the basis vectors in the
two coordinates di¤erently
But the vector is the same since it describes a geometrical entity that is inde-
pendent of the coordinate system. Therefore, we must have
@x0b c
e^0b = e^ ; (18.47)
@xc
so that
@x0b c @x0b @x0b a
c
e^ e^a = v a c ca =
v 0b = v a v : (18.48)
@x @x @xa
Homework: Show that for the covariant components of a vector transformed by
the equation
@xa
vb0 = va : (18.49)
@x0b
18.3. THE METRIC FUNCTION AND COORDINATE TRANSFORMATIONS523
~v = va e^a : (18.54)
and
~ = wb e^b :
w (18.55)
so that the inner product becomes
~ = va e^a wb e^b = g ab va wb :
~v w (18.56)
where
g ab = e^a e^b : (18.57)
We can use the covariant and controvariant components of the vectors to deter-
mine the inner products
a a
~ = v a e^a wb e^b = e^a e^b v a wb =
~v w b v wb = v a wa (18.58)
or
a
~ = va e^a wb e^b = e^a e^b va wb =
~v w b va w
b
= va w a : (18.59)
Whichever way we determine the inner products we must get the same values.
Thus from Eqs. (18.52) and (18.58), we …nd
~ = gab v a wb = v a wa ) gab wb = wa :
~v w (18.60)
~ = g ab va wb = va wa ) g ab wb = wa :
~v w (18.61)
From Eq. (18.60), we note that the covariant form of the metric tensor can be
used to lower an index and from Eq. (18.61) we also see that the controvariant
form of the metric tensor can be used to raise an index. Applying Eqs. (18.60)
and (18.61), we can express the basis vectors
and
e^c = g cb e^b : (18.63)
Then noting that the inner product
a a
e^a e^c = c ) gad e^d g cb e^b = gad g cb e^d e^b = c
d a
) gad g cb e^d e^b = gad g cb b = c
a
) gab g cb = g cb gab = c: (18.64)
gab v a wb = va wa = g ab va wb = v a wa (18.65)
gab v a v b = g ab va vb = v a va = va v a ; (18.66)
and it can be zero without the vector being actually be a zero vector. We can
see this if we recall the pseudo-Riemannian manifold metric
~v = v a e^a : (18.69)
and
~ = wb e^b :
w (18.70)
are said to be orthogonal when
gab v a wb = g ab va wb = v a wa = va wa = 0: (18.71)
P~ = pa e^a (18.72)
dP~ d a dpa d^
ea
= (p e^a ) = e^a + pa (18.73)
d d d d
For the coordinates xa = xa ( ) ; we have
d^
ea ea dxb
@^
= (18.74)
d @xb d
so that
dP~ dpa dxb @^
ea
= e^a + pa : (18.75)
d d d @xb
In order to better understand the origin of the a¢ ne connections we shall re-
reconsider the 2D sphere embedded in a 3D manifold. We saw that the tangent
space is a plane with basis vectors de…ned by
e^ = cos (') x
^ + sin (') y^ p z^: (18.76)
a2 2
and
e^' = sin (') x
^ + cos (') y^: (18.77)
There follows that
@^
e @^
e' 1
= = sin (') x
^ + cos (') y^ = e^' = f ' ( ; ') e^' : (18.78)
@' @
526 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
where
1
f ' ( ; ') = ;
is a function that connects the change in the basis vectors with respect to the
coordinates to the basis vectors. Here we note that we are still in the tangent
space. However, if we switch the variables for the derivatives, we …nd
" #
2
@^
e 1 a2
= p + z
^ = z^;
@ a2 2
(a2 2 )3=2 (a2 2 )3=2
@^
e'
= cos (') x
^ sin (') y^ = (cos (') x
^ + sin (') y^) :
@'
that we can not tell whether it belongs to the tangent space or not. In order to
…nd out that we introduce a basis vector normal to the tangent space in terms
of the basis vector in the tangent space as
e^? = e^ e^' :
This normal basis vector is found to be
2
e^? = p (cos (') x
^ + sin (') y^) + z^: (18.79)
a2 2
e^ = cos (') x
^ + sin (') y^ p z^: (18.80)
a2 2
one …nd !
a2 2
1
z^ = e^? p e^ (18.81)
a2 a2 2
and r
2 2
1
cos (') x
^ + sin (') y^ = 1 e^? e^ : (18.82)
a a2 a2
Using these relations, we can then write
r
2 3
@^
e 1
@^
e'
= p e^ + e^? ; = 1 e^? + 2 e^ : (18.83)
@ a2 2 @' a a2 a
or
@^
e @^
e'
= [f ( ; ')]jj e^ +[f ( ; ')]? e^? ; = [f'' ( ; ')]jj e^ +[f'' ( ; ')]? e^? :
@ @'
(18.84)
where
1
[f ( ; ')]jj = p [f ( ; ')]? =
a2 2
r
2 3
[f'' ( ; ')]jj = 1 [f'' ( ; ')]? = (18.85)
a a2 a2
18.5. THE AFFINE CONNECTIONS 527
These results show that the derivatives for the basis vectors can lead to compo-
nents that does not belong to the tangent space. Let’s see where these compo-
nents are located for a speci…c point for a 2D sphere with radius,
p a = 2 units.
The point we shall consider has coordinates (x = 1; y = 0; z = 3) which is on
the surface of the 2D sphere. At this point the two basis vectors (^ e ; e^' ) and
the vector normal to the tangent space, are shown in Fig. 18.5.
p
@^
e 1 @^
e 3 1
= p e^ + e^? ; ' = e^? + e^ : (18.86)
@ 3 @' 4 4
Therefore, generally for none spherical geometry or none Euclidean manifold,
the change in the basis vectors with respec to the coordinate at a point on a
manifold (derivative of the basis vector at that point) can have both normal and
tangential components to the tangent space. However, since we are con…ned to
the tangent space, for example in the 2D sphere of radius a, we are con…ned to
the tangent space that is a plane and the normal component does not belong
to the tangent space at that particular point. Therefore, we shall consider only
the projection parallel to the tangent space at point P;
@^
ea e^a
= lim : (18.87)
@xc c
x !0 xc jjTp
Suppose we represent the coe¢ cients resulting from the derivative of the basis
vectors that are components that belong to the tangent space at the point
(e.g.[fac ( ; ')]jjTp ) by bac ; then we can write
@^
ea 1 2 3 N b
= ac e
^1 + ac e
^2 + ac e
^3 ::: + ac e
^N = ac e
^b ; (18.88)
@xc
528 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
where N is the dimension of the tangent space. The N 3 coe¢ cients bac are
known collectively as the a¢ ne connection or in older textbooks, the Christo¤el
symbol (of the second kind) at point P .
Homework:
(1) Find all the elements for the a¢ ne connection, bac ; for a point on a 2D
sphere embedded in a 3D Euclidean space. Note that in the expressions
@^e ' @^
e '
= e^ + e^' ; = 'e
^ + 'e
^' ;
@ @'
@^
e' ' @^e' '
= ' e^ + ' e^' ; = '' e
^ + '' e
^' ; (18.89)
@ @'
you are going to determine
' ' ' '
; ; '; '; ' ; ' ; '' ; and '' : (18.90)
we …nd
@^
ea @^
ea
e^d= bac db ) d
ac = e^d (18.93)
@xc @xc
Since d is a dummy index, we can write
b @^
ea
ac = e^b = e^b @c e^a (18.94)
@xc
From now on we will use the notation @c e^a ;
@^
ea
@c e^a = : (18.95)
@xc
Di¤erentiating Eq. (18.92) with respect to the coordinate, xc ; and applying the
notation, we …nd
b
e^a @c e^b + e^b @c e^a = @c a = 0 ) e^b @c e^a = e^a @c e^b (18.96)
Using the de…nition of the derivative of the basis vectors and our notation
@^
ea b d
@c e^a = = bac e^b ) e^d @c e^a = b
ac e
^b e^d = ac b = d
ac (18.97)
@xc
d
) ^d @c e^a
ac = e (18.98)
18.5. THE AFFINE CONNECTIONS 529
e^b @c e^a = a
bc e
^b e^b = e^b a b
bc e
^
a a b
) @c e^ = bc e
^ (18.101)
b @^
ea
ac = e^b : (18.102)
@xc
we have
0b e0a
@^
ac = e^0b : (18.103)
@x0c
so that applying the relations
@xf @x0b d
e^0a = 0a
e^f ; e^0b = e^ : (18.104)
@x @xd
we may write
0b @x0b d @ @xf
ac = e^ e^f
@xd @x0c @x0a
@x0b d ef @xf
@^ @ 2 xf
= e^ + e
^ f
@xd @x0c @x0a @x0c @x0a
@x0b @xf d @^
ef @ 2 xf @x0b d
= e
^ + e^ e^f : (18.105)
@xd @x0a @x0c @x0c @x0a @xd
Using the chain rule one can write
@^
ef ef @xg
@^
=
@x0c @xg @x0c
so that Eq. (18.105) becomes
@x0b @ 2 xd
@xd @x0c @x0a
we will …nd
@x0b @ 2 xd @xd @xf @ 2 x0b
=
@xd @x0c @x0a @x0a @x0c @xd @xf
show that we arrive at an alternative expression for the a¢ ne connect under
coordinates transformations
b @^
ea
ac = e^b (18.109)
@xc
in a similar manner we can also write
b @^
ec
ca = e^b (18.110)
@xa
b
The di¤erence between Eq. (18.109) and (18.110), Tac
b b b
Tac = ac ca (18.111)
so that
@gab
= @c gab = @c (^
ea e^b ) = e^a @c e^b + e^b @c e^a : (18.113)
@xc
Applying the relation
d
@c e^b = bc e
^d (18.114)
we can rewrite Eq. (18.113) as
d d
@c gab = e^b ac e
^d + e^a ^d = dac e^b
bc e e^d + d
bc e
^a e^d
d d
) @c gab = ac gbd + bc gad : (18.115)
18.5. THE AFFINE CONNECTIONS 531
Similarly
@b gca = @b (^
ec e^a ) = e^c @b e^a + e^a @b e^c
d d
) @b gca = e^c ab e
^d + e^a cb e
^d
d d
) @b gca = ab gcd + cb gad (18.116)
and
d d
@a gbc = ba gdc + ca gbd (18.117)
Now combining Eqs. (18.115)- (18.117), we can write
d d d d
@c gab + @b gca @a gbc = ac gbd + bc gad + ab gcd + cb gad
d d
ba gdc ca gbd (18.118)
b b
Recalling that we will be interested in a torsionless manifold where ac = ca ;
Eq. (18.118)
@c gab + @b gca @a gbc = 2 dbc gad : (18.119)
Multiplying Eq. (18.120) by g pa ;
a g ad
bc = (@b gcd + @c gdb @d gbc ) (18.124)
2
The right hand side of Eq. (18.124) is known as the metric connection and is
a
often represented by bc
Useful Formulae:
d
abc = gad bc (18.125)
Multiplying by g f a
f f
gf a abc = g f a gad d
bc = d
d
bc = bc
f fa
) bc =g abc (18.126)
532 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
f gf d
bc = (@b gcd + @c gdb @d gbc ) (18.127)
2
Relabeling the index d by a in Eq. (18.127), we have
f gf a
bc = (@b gca + @c gab @a gbc ) (18.128)
2
so that equating Eqs. (18.126) and (18.128), we …nd
1
abc = (@b gca + @c gab @a gbc ) (18.129)
2
The quantity abc is traditionally known as a Christo¤ el symbol of the …rst kind.
Noting that
1
bac = (@a gcb + @c gba @b gac ) (18.130)
2
we can write
1
abc + bac = (@b gca + @c gab @a gbc + @a gcb + @c gba @b gac ) (18.131)
2
taking into account the symmetry of the metric tensor, from Eq. (18.131), we
…nd
@c gab = abc + bac (18.132)
Eq. (18.132) allows us to express the partial derivative of the metric components
in terms of the connection coe¢ cients.
We recall that the determinant of the matrix G can be expressed as
a+b
det G = gab ( 1) Mab = gab cof [G]ab ; (18.133)
where
a+b
[cof [G]]ab = ( 1) Mab (18.134)
is the the cofactor matrix to G which is determined from the minor, Mab and it
is a constant matrix. We recall that the minor of matrix G denoted by Mab is
the determinant of the matrix formed from matrix G by removing the ath row
and bth column. Let det G = g; so that
we may write
@c g = cof [G]ab @c gab = gg ab @c gab : (18.137)
18.6. LOCAL GEODESIC AND CARTESIAN COORDINATES 533
where we used the fact that abc (P ) ; xaP ; xbP ; and xcP are constant at point P.
The coordinates xa are independent coordinates,
we can write
@x0a
= ad + adc (P ) (xc xcP ) (18.151)
@xd
We are interested in the a¢ ne connection at point P so that when we are
evaluating Eq. (18.151) at xc = xcp , we …nd
@x0a a
= d (18.152)
@xd P
@xa a
= d (18.153)
@x0d P
@x0a @ a @
= + adc (P ) e (xc xcP )
@xe @xd @xe d @x
@x0a
) = adc (P ) ce = a
de (P ) (18.154)
@xe @xd
Using the results in Eqs. (18.152)-(18.154), the transformation equation for the
a¢ ne connection
at point P becomes
0b b f g d d f b
ac (P ) = d a c fg (P ) a c df (P ) (18.156)
0b b f d d b b d b
ac (P ) = d a f c (P ) a dc (P ) = d ac (P ) ac (P ) (18.157)
0b b b
) ac (P ) = ac (P ) ac (P ) = 0:
The result in Eq. (18.157) shows that for the coordinate transformation de…ned
by Eq. (18.144), the a¢ ne connection becomes zero. Such coordinates where
the a¢ ne connection becomes zero at a point P on a manifold is known as local
geodesic coordinates about P .
In chapter 2 we have shown that the conditions for local Cartesian coordi-
nates at a given point P in a seudo-Riemmanian manifold are
0
gab (P ) = ab ; (18.158)
0
@gab (x0 )
= 0 (18.159)
@x0c P
0 0d 0 0d 0
@c gab = ac gbd + bc gad (18.160)
and for a geodesic coordinates the connection is zero at point P , and therefore
0
@gab (x0 )
= 0:
@x0c P
It is important to note that for geodesic coordinates the metric does not nec-
essarily satisfy Eq. (18.158). But we can …nd coordinates x00a that satisfy Eq.
(18.158) by making a linear transformation to the x0a
~v = v a e^a ; (18.163)
where e^a are the coordinate basis vectors. The derivative of this vector with
respect to the coordinate, xb ; can be expressed as
@~v
= @b~v = @b (v a e^a ) = e^a @b (v a ) + v a @b (^
ea ) (18.164)
@xb
We recall
c
@b (^
ea ) = ab e
^c
so that
@b~v = (@b v a ) e^a + c a
ab v e
^c (18.165)
Switching the places for ther indices c and a one can write
c a a c
ab v e
^c = cb v e
^a (18.166)
so that
@b~v = (@b v a ) e^a + v c a
cb e
^a = (@b v a + a c
cb v ) e
^a (18.167)
The quantity in the bracket which is represented as
rb v a = @b v a + a c
cb v (18.168)
is known as the covariant derivative of the vector components. Thus the deriv-
ative of a vector can be expressed as
rb v a = @b v a (18.170)
which is just the ordinary derivative that we are very familiar with!
Homework: Suppose the vector is expressed in terms of its covariant compo-
nents
~v = va e^a (18.171)
show that
c
rb va = @b va ab vc (18.172)
18.7. THE GRADIENT, THE DIVERGENCE, THE CURL ON A MANIFOLD537
rb v a = @b v a + a c
cb v (18.177)
r ~v = ra v a = @a v a + a b
ba v : (18.178)
r2 = r r : (18.181)
But in order to apply the relation we derive for the divergence in Eq. (18.180),
we need the vector
~v = (@a ) e^a = va e^a : (18.183)
in terms of its controvariant components. We have seen that the index can be
raised or lowered using the metric tensor. In this case we want to raise it, so we
have
g ab vb = v a
and we can express the vector ~v as
d d a d^
ea (u) dv a (u)
~v (u) = [v (u) e^a (u)] = v a (u) + e^a (u) (18.192)
du du du du
d ea (u) dxb
d^ dv a (u)
) ~v (u) = v a (u) + e
^a (u) (18.193)
du dxb du du
Using the relation
d^
ea d
= ab e
^d (18.194)
dxb
18.9. PARALLEL TRANSPORT 539
we have
d dxb d dv a
~v = v a e
^ d + e
^ a (18.195)
du du ab du
and replacing the index a by c in the …rst term
d dxb dv a
~v = v c d
cb e
^d + e^a (18.196)
du du du
and d by a
d dxb dv a dv a a c dx
b
~v (u) = v c a
cb e
^a + e^a = + cb v e^a ; (18.197)
du du du du du
which we put in the form
d Dv a
~v (u) = e^a ; (18.198)
du Du
where
Dv a dv a dxb
= + acb v c (18.199)
Du du du
is called the intrinsic (or absolute) derivative of the component v a : Substituting
dv a dv a dxb
= (18.200)
du dxb du
into Eq. (18.199), we …nd
Dv a @v a dxb a c dx
b
dxb a c dx
b
dxb
= + cb v = @b v a + cb v = (@b v a + a c
cb v )
Du @xb du du du du du
Dv a dxb
) = (rb v a ) ; (18.201)
Du du
where we used the relation in Eq. (18.168).
Homework: Suppose the vector, ~v ; depends on the parameter u on a curve
de…ned by xa (u) is expressed in terms of its covariant components
~v = va (u) e^a (u) : (18.202)
Show that the intrinsic derivative of this vector is given by
c
Dva dva b dx
= ac vb (18.203)
Du du du
we have
s p
q
dxa dxb jg ab dxa dxb j
jtj = jgab ta tb j = g ab = (18.213)
du du du
we recall the metric or the interval (the distance squared along the curve on the
manifold between the two points P and Q) is
ds
jtj = =0 (18.216)
du
at all points on the curve and the distance, s; does not depend on the parameter,
u; and we clearly can not use it as a¢ ne parameter since it does not satisfy the
condition in Eq. (18.215). But it is possible to …nd a privileged family of a¢ ne
parameter.
542 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
where
s s
p dy
2
dx
2
ds = dx2 + dy 2 = 1+ dx = 1+ dy (18.218)
dx dy
We may rewrite this distance as
s
Z (2) 2
dy
I= 1+ dx: (18.219)
(1) dx
Out of the in…nitely many functions that can be used to connect the two points,
we want to determine the one that would give the minimum distance. Let these
function be denoted by Y (x) : From these in…nite number of functions there is
only one function that gives the minimum distance between the two points. If
this function is y (x) ; then we may write Y (x) in terms of y (x) as
Y (x) = y (x) + (x) (18.220)
where (x) is an arbitrary function which must satisfy the condition
(x1 ) = (x2 ) = 0 (18.221)
so that
Y (x) = y (x) (18.222)
at the two points (x = x1 = x2 ); is the constant of variation that means it is
constant that determines by how much Y (x) di¤ers from y (x). Now in terms
of Y , we may write
Z (2) p
I( )= 1 + Y 0 2 dx: (18.223)
(1)
18.11. REFRESHMENT FROM THEORETICAL PHYSICS: THE CALCULUS OF VARIATION543
where
dY (x) 0
: Y = (18.224)
dx
For the distance to be a minimum, the necessary condition is that the integral
must be stationary. This requires
dI ( )
=0 (18.225)
d =0
which leads to
Z (2)
dI ( ) 1 1 dY 0 ( )
= p (2Y 0 ) dx = 0:
d =0 (1) 2 1 + Y 02 d
=0
Using
Y (x) = y (x) + (x) (18.226)
we may write
0 0
Y (x) = y 0 (x) + 0
(x) ) Y (x) = y 0 (x) (18.227)
=0
so that
dY 0 ( ) 0 dY 0 ( ) 0
= (x) ) = (x) (18.228)
d d =0
which leads to
Z (2) Z (2)
1 1 dY 0 ( ) y 0 0 (x)
p (2Y 0 ) dx = p dx = 0:
(1) 2 1 + Y 02 d
=0
(1) 1 + y0 2
(18.229)
Using integration by parts
Z Z
udv = uv vdu (18.230)
for
!
0 y0 d y0
(x) = dv ) v = (x) ; u = p ) du = p dx (18.231)
1 + y0 2 dx 1 + y0 2
we …nd
Z Z !
(2) x1
dI ( ) y 0 0 (x) d y0
= p dx = (x) p dx = 0
d =0 (1) 1 + y0 2 x1 dx 1 + y0 2
(18.233)
Since (x) is an arbitrary function, we must have
!
d y0 d p d 0
p = 0) 1 + y0 2 = 0 ) 1 + y 2 =(18.234)
0
dx 1+y 02
dx dx
d dy 0
) y 02 = 0 ) = 0 ) y0 = m (18.235)
dx dx
where m is a constant. Thus
dy
y0 = = m ) y (x) = mx + b
dx
which is equation of a straight line.
The Problem: We consider some unknown function y(x). We assume that
this function is known at two …xed points, y(x1 ) and y(x2 ). We wish to …nd the
function y(x) that makes the integral
Z x2
I= F (x; y; y 0 ) dx
x1
stationary for some known function F (x; y; y0). It can be shown that this func-
tion satis…es the Euler-Lagrange Equation:
d @F @F
=0
dx @y 0 @y
The basis vector with respect to the coordinates xc does not change along the
geodesic curve. This means for the tangent vector
a
~t = dx e^a (18.238)
du
18.12. THE GEODESIC 545
to the curve at least the direction remains the same. In the Euclidean space
this makes the curve to be a straight line where the tangent vectors has same
direction along the line. Thus for Euclidean space the geodesic is a straight line.
For general curve de…ned by xa = xa (u) ; on a manifold, if the curve is
geodesic then the tangent curve must have the same direction at all points on
the curve. This mean the change in the tangent vector with respect to the
parameter u is then only the magnitude that changes and it is given by
d~t dxa
= (u) ~t = (u) e^a ; (18.239)
du du
where (u) is some function of u: Using our result for intrinsic derivative of a
vector ~v (u)
d Dv a dv a dxb
~v (u) = e^a = + acb v c e^a ; (18.240)
du Du du du
for the tangent vector intrinsic derivative, we have
dta a c dx
b
dxa
+ cb t = (u) (18.242)
du du du
Noting that
a
~t = dx e^a = ta e^a (18.243)
du
Eq. (18.242) can be written as
d2 xa a dx dxb
c
dxa
+ cb = (u) (18.244)
du2 du du du
The result in Eq. (18.244) is valid for both non-null and null geodesics parame-
terized in terms of some general parameter u. For a¢ ne parameter u where it
is related to the distance s on the curve
u = as + b (18.245)
we have
du = ads (18.246)
and
ds
jtj = =a (18.247)
du
which is a tangent vector with a constant length that is independent of the
parameter u: This means
d~t
= (u) ~t = 0 ) (u) = 0 (18.248)
du
546 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
d2 xa a dx
c
dxb
+ cb =0 (18.249)
du2 du du
Eq. (18.249) is a parallel transport for the tangent vector that we discussed in
the previous section.
Dta dta dxb
= + acb tc =0 (18.250)
Du du du
which can be shown by replacing
dxa dxc
= ta ; = tc (18.251)
du du
in Eq. (18.249).
Homework:
If we change the a¢ ne parameter u to u0; the coordinates that de…ne the
geodesic curve would change from xa (u) to xa (u0 ) : Show that in terms of the
new a¢ ne parameter, u0; the geodesic in Eq. (18.249) becomes
!
d2 u
d2 xa a dxc
dxb
du 02 dxa
+ cb = du
(18.252)
du02 du0 du0 du0
du0
where q
ds
F = jgab x_ a x_ b j = s_ = (18.256)
du
18.13. STATIONARY PROPERTY OF THE NON-NULL GEODESIC 547
Using the principle of variation that you have learned in Theoretical Physics Part
I and revised in the previous section the curve along a surface which marks the
shortest distance between two neighboring points is the geodesic. Using the
principle of variation it can be shown that the integral must be stationary for
the geodesic and the function satis…es the Euler-Lagrange Equation:
d @F @F
= 0: (18.257)
du @ x_ c @xc
d gac x_ a x_ a x_ b @a gab
= 0: (18.259)
du s_ 2s_
For the …rst terms we may write
1 s• x_ a x_ b @a gab
(@b gac ) x_ a x_ b + gac x
•a gac x_ a = 0: (18.263)
s_ s_ 2s_
548 CHAPTER 18. VECTOR CALCULUS ON MANIFOLDS
x_ a x_ b @a gab s•
•a + (@b gac ) x_ a x_ b
) gac x = gac x_ a : (18.264)
2 s_
Noting that we can interchange the indices and write
so that
1
(@b gac ) x_ a x_ b + (@a gbc ) x_ b x_ a
(@b gac ) x_ a x_ b = (18.266)
2
Substituting Eq. (18.266) into Eq. (18.264), we have
1 a b s•
•a +
gac x x_ x_ @b gac + x_ b x_ a @a gbc x_ a x_ b @a gab = gac x_ a : (18.267)
2 s_
1 s•
•a +
gac x [@b gac + @a gbc @a gab ] x_ a x_ b = gac x_ a : (18.268)
2 s_
and multiplying by g dc ;
1 s• dc
g dc gac x
•a + g dc [@b gac + @a gbc @a gab ] x_ a x_ b = g gac x_ a : (18.269)
2 s_
d
Now applying the relationg dc gac = a; we …nd
d a 1 s• d a
ax
• + g dc [@b gac + @a gbc @a gab ] x_ a x_ b = ax
_ : (18.270)
2 s_
1 s•
•d + g dc [@b gac + @a gbc
x @a gab ] x_ a x_ b = x_ d : (18.271)
2 s_
a 1 dc
bc = g [@b gac + @a gbc @a gab ] (18.272)
2
we …nd
s•
•d +
x a a b
bc x
_ x_ = x_ d : (18.273)
s_
Comparing the result in Eq. (18.273) with Eq. (18.252)
!
d2 u
d2 xa c
a dx dx
b
du02 dxa
02
+ cb 0 0
= du
(18.274)
du du du du0
du0
Tensor Calculus on
manifolds
~t (~v ) = ~t ~v : (19.1)
This maps the vector …eld, ~v ; to the tangent space, Tp : In this case one vector (~v )
is linearly mapped into the tangent space by the tangent vector ~t (i.e. ~t ! ~t (~v )).
Therefore the tangent vector ~t forms a …rst rank tensor t.
A tensor : Based on the notion of a vector on a manifold, a tensor is de…ned
by the precise set of operations applied to the a set of vectors to produce a
scalar and the number of vectors in the set determines the rank of the tensor.
If there are N number of vectors in the set, the tensor is N th rank tensor (See
the table below)
Tensor Operation Rank
t (~u; ~v ) ~u ~v 2
t (~u; ~v ; w)
~ ~u ~v w
~ 3
t (~u; ~v ; w;
~ ~x) ~u ~v w
~ ~x 4
t (~u; ~v ; w;
~ ~x; ~y ) ~u ~v w
~ ~x ~y 5
From this table we can easily see that a scalar …eld can be classi…ed as a
zero ranked tensor …eld since it does not depend on a vector …eld.
549
550 CHAPTER 19. TENSOR CALCULUS ON MANIFOLDS
A tensor is a linear map of the vectors into the real and therefore any ranked
tensor is linear. This means, for example, for 1st rank tensor, we must have
~t ( ~u + ~v ) = ~t ( ~u) + ~t ( ~v ) = ~t (~u) + ~t (~v ) (19.2)
Example 4.1 Let’s reconsider the 2D sphere in the 3D manifoldFor the vector
we have
t1
t1 = t (~ea ) ~e1 ; t2 = t (~ea ) ~e2 ) t (~ea ) =
t2
~e1 ~e1 + ~e2 ~e1 g11 + g21
) t (~ea ) = =
~e1 ~e2 + ~e2 ~e2 g12 + g22
and we recall for 2D sphere g21 = g12 = 0;
g11
t (~ea ) =
g22
which is a 1-st rank tensor. Now let’s consider a quantity de…ned by by
an operation on set that consist of two vectors as
t (~ea ; ~eb ) =(~e1 ~e1 ) + (~e2 ~e1 ) + (~e1 ~e2 ) + (~e2 ~e2 )
g11 g12
) t (~ea ; ~eb ) =
g21 g22
which forms a 2-nd rank tensor. Using the dual basis vector, we can also
express this 2-nd rank tensor as
t ~ea ; ~eb = ~e1 ~e1 + ~e2 ~e1 + ~e1 ~e2 + ~e2 ~e2
g11 g12
) t (~ea ; ~eb ) =
g21 g22
or
t ~e;a ~eb = ~e1 ~e1 + ~e2 ~e1 + ~e2 ~e2 + ~e2 ~e2
g 12 g 12
) t (~ea ; ~eb ) =
g 21 g 22
1
t(a1 a2 :::aN ) = (addition over all permutations of the indices a1 a2 :::aN )
N!
(19.18)
and
1
t[a1 a2 :::aN ] = (Alternating subtraction and addition over
N!
all permutations of the indices a1 a2 :::aN ) (19.19)
1
t[a1 a2 a3 ] =
t[a1 a2 a3 ] (19.20)
3!
For example for 3rd rank tensor, we have
1
t(a1 a2 a3 ) = (ta a a + ta2 a1 a3 + ta1 a2 a3 + ta3 a2 a1 + ta1 a2 a3 + ta1 a3 a2 ) (19.21)
3! 1 2 3
and
1
t(a1 a2 a3 ) = (ta a a ta2 a1 a3 + ta1 a2 a3 ta3 a2 a1 + ta1 a2 a3 ta1 a3 a2 ) (19.22)
3! 1 2 3
Particular subset of indices permutation: we have a di¤erent notations when
the permutation is to particular subset of indices. This is described using a 4th
19.1. TENSORS FIELDS AND RANK OF A TENSOR 553
rank tensor.
1
t(ab)cd = (tabcd + tbacd ) (19.23)
2
symmetric permutation to indices a and b only
1
t[ab]cd = (tabcd tbacd ) (19.24)
2
antisymmetric permutation to indices a and b only
1
ta[bjcjd] = (tabcd tbdca )
2
antisymmetric permutation to indices b and d only (19.25)
Note that the symbol jj are used to exclude unwanted indices from the sym-
metrization () antisymmetrization [] implied.
Example 4.2 Let’s reconsider the 3D sphere in the 4D manifoldWe recall the
tangent space at a point on this 3D sphere form three basis and dual basis
vectors, ~e1 ; ~e2 ; and ~e3 (or ~er ; ~e ; and ~e' ). Write the symmetric, t(123) and
antisymmetric components t(123)
We recall that covariant and controvartiant components of the metric tensor are
given by
gab = g (~ea ; ~eb ) = ~ea ~eb ; g ab = g ~ea ; ~eb = ~ea ~eb (19.28)
and the mixed components
a
gab = g ~ea ; ~eb = g (~ea ; ~eb ) = b (19.29)
Raising and lowering tensor indices. Consider the third rank tensor, t; expressed
in terms of its covariant components
or mixed components
tcab = ~ea ~eb ~ec (19.31)
We recall from chapter 3 that the covariant form of the metric can be used to
lower and the controvariant can be used to raise the index. Thus to lower the
index in Eq. (19.31) we multiply by the covariant form of the metric tensor,
gdc ;
gdc tcab = gdc~ea ~eb ~ec (19.32)
noting that
c
gdc = ~ed ~ec ; and ~eb ~ec = b
we have
c
gdc tcab = ~ed ~ec ~ea b = ~ed ~eb ~ea = tdba (19.33)
so that
tdba = gdc tcab (19.34)
Since the indices are dummy variables and tensor tabc is symmetric, one can
write
tabc = gcd tdab (19.35)
Homework: Raise the tensor tabc to tabc :
and the second rank tensor t (~u; ~v ) maps the vectors ~u and ~v into real numbers
in the tangent space
Now the question is can we map a tensor into another tensor of a di¤erent rank.
Consider a third rank tensor t (~u; ~v ; w)
~ which maps the three vectors into real
numbers. Now instead if we replace the two vectors (~u; ~v ) by the basis vectors
(~ea ; ~eb ) on the tangent plane, we have
t (~ea ; ~eb ; w)
~ = ~ea ~eb w;
~ (19.37)
~ = wc~ec
w (19.38)
we …nd
which is a 2-nd rank tensor. We see that the 3-rd rank tensor t maps the vector,
w;
~ into a second rank tensor s: As another example let’s consider the third rank
tensor
t (~v ; ~eb ; w)
~ = ~v ~eb w:
~ (19.40)
Suppose we express the vectors, ~v and ~u; in terms of its controvariant compo-
nents
~ = wc~ec ; ~v = v a~ea ;
w (19.41)
then
where we …nd a 1st rank tensor, sb : This means the 3rd rank tensor mapped the
two vectors into 1st rank tensor (a vector).
sab = s (~ea ; ~eb ) = t (~ea ; ~eb ) r (~ea ; ~eb ) = tab rab : (19.43)
Now let’s consider 2-nd rank tensor, t (~ p; ~q), and 1-st rank tensor (a vector),
s (~r) ; the outer product of these two tensors give another tensor, h, of a di¤erent
rank given by
t s (~p; ~q; ~r) = t (~
p; ~q) s (~r) = h: (19.46)
556 CHAPTER 19. TENSOR CALCULUS ON MANIFOLDS
The tensor h is a 3-rd rank tensor. Using the basis and dual basis vectors we
may express the components as
Example 4.2 Let’s consider two vectors in the tangent space for a point in a
3D sphere embeded in 4D manifold geven by
p~ = pa ea ; ~q = q b eb
where a; b = 1; 2; 3, and ~e1 ; ~e2 ; and ~e3 (or ~er ; ~e ; and ~e' ). Find the
components of the 2-nd rank tensor for
u v (~
p; ~q) = u (~
p) v (~q) ; (19.51)
u p; ~q) = pa q b ea eb
v (~
= p 1 q 1 e1 e1 + p 1 q 2 e1 e2 + p 1 q 3 e1 e3
+p2 q 1 e2 e1 + p2 q 2 e2 e2 + p3 q 3 e3 e3
+p3 q 1 e3 e1 + p3 q 2 e3 e2 + p3 q 3 e3 e3
t s = tab sb (19.52)
and
s t = tab sa (19.53)
are not necessarily the same.
19.4. TENSORS AND COORDINATE TRANSFORMATIONS 557
Like vectors (1-st rank tensor) tensors of higher rank are geometrical objects
too: We already know that vectors are geometrical objects that can be made
up from a linear combination of the basis vectors
t = ta ea = ta ea : (19.54)
The vector that de…nes a given geometry on a manifold, does not depend on how
it is represented. The geometry that a vector de…nes remain the same geometry
whatever representation we used to describe the vector. The same is true for
higher ranked tensors. As an example, let’s consider a 2-nd rank tensor with
component tab constructed from the outer product of two basis vectors
t = ea eb : (19.55)
Now suppose we have some general 2-nd rank tensor, t, whose controvariant
components are tab (ea eb ) : The action of this second rank tensor on two
basis vectors
tab (ea eb ) ec ; ed = tab ca db = tcd (19.57)
Therefore what is true for example for 2-d rank tensor
t (ea ) = ta ea = ta ea : (19.61)
@x0a @x0b cd
t0 e0a ; e0b = t (19.64)
@xc @xd
@xc @x0b d
t0 e0a ; eb0 = t (19.65)
@x0a @xd c
@xc @xd
t0 (e0a ; e0b ) = tcd (19.66)
@x0a @x0b
3-nd rank tensor: Suppose we have a mixed 3-rd rank tensor, tcab ! t (ea ; eb ; ec )
in the x0a coordinate system is given by
@xc @xd
@x0a @x0b
we have
@xc @xd @xc @xd
0a 0b
tcd = scd (19.69)
@x @x @x0a @x0b
so that using Eq. (19.66), we …nd
Well Eq. (19.70) shows that the equality holds under the coordinate transfor-
mation. However, the question is that these components (set of quantities) are
actually form a tensor. The quotient theorem set’s the condition for a set of
quantities actually represent a tensor component.
The quotient theorem: if a set of quantities when contracted with a tensor
produces another tensor, then the original set of quantities are also a tensor.
Suppose in an N dimensional manifold you are given a 3-rd rank tensor,
t; and 1-st rank tensor, v: The tensor, t; has a set of N 3 quantities tabc and
the tensor v has N quantities of v a : Suppose we form a set of N 2 quantities
by contracting the 4-th rank tensor formed by the outer product of these two
19.6. COVARIANT DERIVATIVES OF A TENSOR 559
rb v a = @b v a + a c
cb v (19.81)
rv = (rb v a ) ea : = @b v a + a c
cb v (19.82)
rb v a forms a mixed 2-nd rank tensor. We denote this second rank tensor by
rv. Noting that
r = ea @ a ; v = v b eb
we can express
rv = ea @a v b eb = ea @a v b eb = ra v b ea eb (19.83)
Let’s consider the the covariant derivative of a second rank tensor t with ex-
pressed in terms of its controvariant components tab
rc t = rc tab ea eb (19.84)
@c t = @c tab (ea eb ) + tf b a
cf (^
ea eb ) + taf b
cf (ea e^b ) (19.89)
@c t = @c tab + a fb
cf t + b af
cf t (^
ea eb ) = rc tab e^a eb (19.90)
where
rc tab = @c tab + a db
cd t + b ad
cd t ; (19.91)
is the covariant derivative.
Homework: Show that for the covariant derivatives of the mixed and covari-
ant component of a 2-nd rank tensor t are given by
rc tab = @c tab + a d
dc tb
d a
bc td (19.92)
d d
rc tab = @c tab ac tdb bc tad (19.93)
19.7. INTRINSIC DERIVATIVE 561
Useful relation
@c e^a = a b
bc e
^ (19.94)
Homework: Show that the covariant derivative of the metric tensor is zero
rg = 0
Suppose we represent the metric tensor in terms of its controvariant components,
g ab ; then you must show that the covariant derivative expressed as
rc g ab = @c g ab + a db
cd g + b ad
cd g =0 (19.95)
Useful relations, for example, the a¢ ne connection and the metric are related
by
f gf d
bc = (@b gcd + @c gdb @d gbc ) (19.96)
2
Application of the property of the metric tensor: Suppose we have a 2-nd rank
tensor, t, for which we want to …nd the covariant derivative from its components,
for example, in mixed form. That means we want to …nd rc tab from tad : We can
express the covariant components of this tensor using contraction as
tab = g bd tad :
Note that the metric tensor is symmetric (g bd = g db ). Then
tab = g bd tad
rc tab = rc g bd tad = rc g bd tad + g bd rc tad (19.97)
since
rc g bd = 0
we can express
rc tab = g bd rc tad : (19.98)
dea (u) dea (u) dxc deb (u) deb (u) dxb
= ; =
du dxc du du dxc du
and using the a¢ ne connection
dea (u) f
@c eb = = bc e
^f (19.99)
dxc
we have
dea (u) f dxc deb (u) f dxb
= ac e
^f (u) ; = bc e
^f (u) (19.100)
du du du du
so that
dt (u) dtab dxc f dxb
= ea (u) eb (u) + tab f
ac e
^f (u) eb (u) + tab ea (u) bc e
^f
du du du du
(19.101)
dt (u) dtab f dx
c
f dx
b
= ea (u) eb (u)+tab ac ef (u) eb (u)+tab bc ea (u) ef (u) :
du du du du
(19.102)
Replacing the dummy index a by f in the …rst term, we can write
dt (u) dtf b f dx
c
f dx
b
= ef (u) eb (u)+tab ac ef (u) eb (u)+tab bc ea (u) ef (u) :
du du du du
(19.103)
Noting that by making the following dummy index change b ! d followed by
a!b
b d d
f dx f dx f dx
tab bc ea (u) ef (u) = tad dc ea (u) ef (u) = tbd dc eb (u) ef (u)
du du du
we …nd
dt (u) dtf b f dx
c
f dx
d
= + tab ac + tbd dc ef (u) eb (u) : (19.104)
du du du du
that we expressed as
dt (u) Dtf b
= ef (u) eb (u) : (19.105)
du Du
where
Dtf b dtf b dxc dxd
= + tab fac + tbd fdc ; (19.106)
Du du du du
is called the intrinsic (absolute) derivative of the component tf b along the curve
de…ned by xa (u) : For the sake of convenience we make change of dummy indices
(f ! a) in Eq. (19.105) and (a ! d) in Eq. (19.106) as
dt (u) Dtab
= ea (u) eb (u) : (19.107)
du Du
19.7. INTRINSIC DERIVATIVE 563
where
Dtab dtab dxc dxd
= + tdb adc + tbd adc : (19.108)
Du du du du
We can switch the dummy indices c and d in the third term as the a¢ ne is
symetric for Torsionless
Dtab dtab a dx
c
a dx
c
= + tdb dc + tbc cd :
Du du du du
It can be easily shown that
Dtab dxc a dx
c
a dx
c
dxc
= @c tab + tdb dc + tbc cd = @c tab + tdb a
dc + tbc a
cd
Du du du du du
Using the result in Eq. (19.91) we can write
Dtab dxc
= rc tab (19.111)
Du du